.

Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP001

Detection Challenges in Clinical Diagnostics

View Online

RSC Detection Science Series

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP001

Editor-in-Chief:
Professor Mike Thompson, University of Toronto, Canada

Series Editors:
Dr Subrayal M Reddy, University of Surrey, Guildford, UK
Dr Damien Arrigan, Tyndall National Institute, Cork, Ireland

Titles in the Series:

1: Sensor Technology in Neuroscience
2: Detection Challenges in Clinical Diagnostics

How to obtain future titles on publication:

A standing order plan is available for this series. A standing order will bring
delivery of each new volume immediately on publication.

For further information please contact:

Book Sales Department, Royal Society of Chemistry, Thomas Graham House,
Science Park, Milton Road, Cambridge, CB4 0WF, UK
Telephone: +44 (0)1223 420066, Fax: +44 (0)1223 420247
Email: [email protected]
Visit our website at www.rsc.org/books

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP001

View Online

Detection Challenges in Clinical

Diagnostics
Edited by

Pankaj Vadgama
Queen Mary University of London, UK
Email: [email protected]
and

Serban Peteu
National Institute for R&D in Chemistry and Petrochemistry,
Bucharest, Romania
Email: [email protected]

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP001

View Online

RSC Detection Science Series No. 2

ISBN: 978-1-84973-612-1
ISSN: 2052-3068
A catalogue record for this book is available from the British Library
r The Royal Society of Chemistry 2013
All rights reserved
Apart from fair dealing for the purposes of research for non-commercial purposes or for
private study, criticism or review, as permitted under the Copyright, Designs and Patents
Act 1988 and the Copyright and Related Rights Regulations 2003, this publication may not
be reproduced, stored or transmitted, in any form or by any means, without the prior
permission in writing of The Royal Society of Chemistry or the copyright owner, or in the
case of reproduction in accordance with the terms of licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of the licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries concerning
reproduction outside the terms stated here should be sent to The Royal Society of
Chemistry at the address printed on this page.
The RSC is not responsible for individual opinions expressed in this work.
Published by The Royal Society of Chemistry,
Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, UK
Registered Charity Number 207890
For further information see our web site at www.rsc.org

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP005

With deep appreciation to

Dixa, Reena, Rooshin and Preeya
Mihaela and Mihai
for all their support

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP005

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP007

Preface
There have been quite a number of topical texts and reviews published recently
dealing with clinical diagnostics. So, one question to ask is why another book?
This Preface hopefully provides a rationale for this. In a parallel vein, there was
a belief in the idea of a Fountain of Youth; an aged-adult body enters the
Fountain Service and exits with new body components, with as good-as-new
regeneration. To bring this up to date, even if this was achievable, say by tissue
engineering the question arises would the insurance companies foot the bill?
So not only are there residual uncertainties, even in utopia, but context is ever
changing.
To return to the present topic, our aim was to bring the reader up to date
within the context of rapidly evolving technology and to communicate this
through the eyes of research leaders. A broad range of approaches is scoped
and the diagnostics needs and bottlenecks surveyed. Both academic and industrial experts are included, all addressing robust tools for dealing with the
world of real biological measurement especially from the perspective of a
commonly neglected expert: the end user.
Chapter 1 (by Thompson et al.) and Chapter 2 (by Vadgama et al.) deal with
Clinical Diagnostics, both in the laboratory and at the bedside, from the
broader picture down to some details. There have been powerful advances in
extralaboratory testing enabled by new solid-state technology encompassing
reagent immobilisation and miniaturization. This is a dicult area to monitor
and set standards for, because of the distributed nature of such testing across a
variety of clinical sites and even the home. Laboratory analysis has taken on
major advances with high throughput and small sample volumes, so polarisation between technologies that are aimed at laboratory testing vs. those for
extra laboratory testing are inevitable.

RSC Detection Science Series No. 2

Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

vii

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP007

viii

Preface

Analytical specicity remains a vital issue, still not fully resolved, especially
for low-concentration analytes, regardless of technique, and where sample
separation cannot be part of the assay system, as in say in vivo sensors, quite
signicant eort is required to redesign the basic construct. This latter owes as
much to materials science and engineering as to chemistry.
Selectivity in general is discussed in Chapter 1 (by Thompson et al.) with
specic regard to Clinical Chemistry. In the in vivo context, sensors need to
function selectively given their exposure to unmodied samples. There is the
added complication of high local protein concentration and cellular ingress at
the implant site. Returning to the selectivity challenge, there are helpful, published techniques for electrochemical bio/sensors reviewed in Chapters 1 and 6
(Thompson et al.; Peteu and Szunerits), including membranes to address solute
transport and interfacing issues. Ultimately, one needs to be mindful of the
trade-o between analytical complexity and slower processing as against the
goal of high specicity.
This issue of damaging nonspecic adsorption (NSA), represents a true
Achilles heel for direct contact sensors, and is examined in detail. With biological matrices constituting highly complex solute mixtures, it becomes clear
they could well prevent the detection/quantication of target analytes present at
considerably lower concentration, outlined in Chapter 1 (Thompson et al.).
Early advances in this regard, though not always seen as such, are the dry
reagent systems developed for glucose as illustrated in Chapter 3 (Wang and
Hu). Here, unless the integrated laminates are not tailored to whole-device
function and can be produced in mass numbers, the overall transduction value
cannot be realized. This chapter examines the progress and challenges of the
blood-glucose biosensors. Managing ones diabetes also decreases the occurrence of its serious complications such as nephropathy, neuropathy and retinopathy. The pathogenesis of diabetes and its complications seem to be
correlated with the presence of nitro-oxidative species including peroxynitrite
potentially implicated in beta-cells destruction, as highlighted in Chapter 6
(Peteu and Szunerits).
Chapter 4 (Gaspar et al.) critically assess recent progress and many challenges in electrochemical detection of disease-related diagnostic biomarkers.
Mostly, we have relied on biomolecules, but aptamer technology shows how
such synthetic structures can be harnessed to give stable readers for biochemical targets. These are early days still for the technology, and designer
aptamers bred via SELEX (selective evolution of ligands by exponential enrichment), should extend their repertoire. It may also be that here and elsewhere, with use of arrays, absolute selectivity will not be a necessity.
As with sampling integrity, so with continuous use sensors for monitoring
in vivo, Chapter 5 (by Meyerho et al.), shows there is great need to reduce bioincompatibility and resultant surface fouling; biouids are not tolerant of foreign
surfaces. In vivo sensors are, however, uniquely positioned to provide siterelated information, in particular at specic extravascular, tissue locations, but
face a huge safety and biocompatibility challenge. The fact that this has been
resolved in some cases raises the possibility of broader forms of monitoring

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP007

Preface

ix

systems. Even using early proof-of-concept systems, it may yet be possible to

pick up biological signatures that arise from wider disease sets.
Chapter 5 (Meyerho et al.) especially scrutinizes the challenges for sensors
long-term biocompatibility. In spite of the great sensors advances in vitro, the
commercial development of implantable chemical sensors has reached a
bottleneck. So much so, that, even with the FDA-required recalibration the
output of devices, is still not considered reliable enough. Special coating materials able to say attenuate the activation of platelets or materials able to inhibit
the inammatory response will become important.
In the case of more exotic short-lived radical species such as peroxynitrite
(often r1 s lifetime) featured in Chapter 6 (by Peteu and Szunerits) measurement in real environments will be dicult. For species such as peroxynitrite,
quantication poses a whole new level of measurement uncertainties, yet they
are important: cell signalling, reactivity and tissue damage are mediated by such
short-lived radicals. Mitochondrial oxidation is itself a free-radical generator.
Interestingly, nitric oxide, superoxide, the precursors of peroxynitrite, and peroxynitrite itself have been dubbed the good, the bad and the ugly because of
their tissue-level eects. This chapter illustrates the chemical diversity of such
reactive species and the way in which electrochemical interfaces and sensor
chemistry could track some of these, as a glimpse into future clinical use.
The biomachinery resulting in ineective haematopoiesis and augmented
leukaemia risk in myelodysplastic syndromes is largely known. However, one
major challenge illustrated in Chapter 7 (by McNamara et al.) is how to correctly classify and risk stratify the patients. Molecular biology assessment
could help as diagnostic tools. So often, the right diagnosis oered early in the
game will aect morbidity and survival.
Chapter 8 (by Barr et al.) describes Raman for noninvasive early cancer
diagnosis. Early histological appearances may be dicult to categorise, and
there is less interobserver agreement. Some changes may not even be evident by
traditional histology, moreover; diagnosis is expensive, time consuming and
requires required tissue biopsy. For the specic case of oesophageal neoplasia,
there is evidence that a Raman signature can identify molecular change prior to
morphological aberrations.
Raman spectroscopy can deliver high sensitivity for degenerating, premalignant, oesophageal epithelium. The gain would be objectivity, speed and a
real-time assessment for early removal of dysplastic tissue and follow up.
Raman-based bre-optic interrogation has potential as an in situ surgical
adjunct.
Signal handling with arrays is well exemplied in Chapter 9 (by Kendall
et al.). Any opportunity to create multiple arrays should be taken, as this adds
depth to measurement. Here, for volatiles analysis the power of pattern recognition is well demonstrated. Undoubtedly the principles can be extrapolated
to other modes of measurement, including where selectivity and drift are
challenges, be it in vivo or in vitro.
Another overarching challenge in clinical diagnostics: the critical need to
provide information not just data. With ease of data generation, it could be said

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP007

Preface

that there is too much data for the clinician to deal with, especially if it is real
time as with in vivo monitoring devices. Herein lies the problem of what the
data is really for; if it is a medical defence strategy it becomes a waste of nances, but if behind the data there is a genuine quest for what the bio/
pathological implications might be, then the data becomes of considerably
greater value.
As with any scientic quest, one might consider this to be analogous to a
person searching for their home keys under a streetlamp, even though these
might have been dropped somewhere else, because thats where the light
is.. . . This book has tried to shed some light on some of the important
detection challenges impeding future progress in clinical diagnostics. It may be
the light is currently in the wrong place, but eventually the home keys will
be found. Finally, we would like to thank the team at Royal Society of
Chemistry who guided us so patiently through the publication maze and
without whom there would be no book: Merlin Fox, Rosalind Searle and Alice
Toby-Brant (Commissioning); Lois Bradnam and Sarah Salter (Production).
Pankaj Vadgama
London
Serban Peteu
Bucharest

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

Contents
List of Contributors
Chapter 1

xvii

Biosensor Technology and the Clinical Biochemistry

Laboratory Issue of Signal Interference from the Biological
Matrix
Michael Thompson, Sonia Sheikh, Christophe Blaszykowski
and Alexander Romaschin
1.1
1.2

Laboratory Clinical Biochemical Assays

Biosensor Technology
1.2.1 Biosensor Architecture
1.2.2 Probe Attachment to Device Surfaces
1.2.3 Devices and Transduction
1.3 Biosensors and Measurement of Clinical Targets
1.4 Signal Interference and the Non-specic Adsorption
Problem
1.5 A Look at Surface Chemistries to Solve the NSA Issue
1.6 A Final Comment
Acknowledgements
References
Chapter 2

Integrated Chemistries for Analytical Simplication

and Point of Care Testing
Pankaj Vadgama, Salzitsa Anastasova and
Anna Spehar-Deleze
2.1
2.2

Introduction
Fluidics for POCT

1
6
6
7
11
21
23
29
30
31
32

35

35
36

RSC Detection Science Series No. 2

Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

xi

View Online

xii

Contents

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

2.3

Lateral Flow Techniques From Simple Colorimetric

Strips to 3D Flow
2.3.1 Colorimetric Strips
2.3.2 Nanoparticle Labels
2.3.3 New Opportunities
2.4 Biological Recognition
2.4.1 Immobilisation Techniques
2.4.2 Enzyme-Based Systems
2.4.3 Aptamers as Biorecognition Elements
2.5 Electrochemical Sensors
2.5.1 Ion-Selective Electrodes
2.5.2 Amperometric Enzyme Biosensors
2.5.3 Immunosensors
2.5.4 DNA Sensors
2.6 Micromechanical Transduction
2.7 Commercialisation
2.8 Conclusions
Acknowledgement
References

Chapter 3

Blood-Glucose Biosensors, Development

and Challenges
Yuan Wang and Madeleine Hu
3.1
3.2

Introduction
History of Blood-Glucose Biosensor
3.2.1 First Generation of Biosensors
3.2.2 Second Generation of Glucose
Biosensors
3.2.3 Third Generation of Glucose
Biosensor
3.3 Sensors for BGMS
3.3.1 Electrode
3.3.2 Reaction Chamber
3.3.3 Chemistry
3.3.4 Detection Method
3.3.5 Correction Algorithm
3.3.6 Calibration of BGMS
3.3.7 Performance Validation of BGMS
3.3.8 Manufacturing Process
3.4 Clinical Utility and Potential Problems
3.5 Continuous Glucose Monitoring System
(CGMS)
References

40
40
41
43
43
43
47
48
50
50
52
54
56
56
57
58
59
59

65

65
66
67
68
69
69
70
72
73
74
74
76
77
79
80
81
82

View Online

xiii

Contents

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

Chapter 4

Recent Progress in the Electrochemical Detection of

Disease-Related Diagnostic Biomarkers
Alina Vasilescu, Wolfgang Schuhmann and Szilveszter Gaspar
4.1
4.2

Introduction
Electrochemical Sensors for Detection of Cancer
Biomarkers
4.2.1 Electrochemical Sensors for
Carcinoembryonic Antigen
4.2.2 Electrochemical Sensors for Prostate-Specic
Antigen
4.2.3 Electrochemical Sensors for Other Protein
Cancer Biomarkers
4.2.4 Electrochemical Sensors for Simultaneous
Detection of Several Protein Biomarkers
4.2.5 Electrochemical Sensors for Genetic Markers
of Cancer
4.2.6 Electrochemical Sensors for Detection of
Cancer Cells
4.3 Electrochemical Sensors for Detection of Cardiac
Biomarkers
4.4 Electrochemical Sensors for Detection of Acquired
Immunodeciency Syndrome
4.5 Electrochemical Sensors for the Detection of
Hepatitis Biomarkers
4.5.1 Electrochemical Sensors for Hepatitis B
4.5.2 Electrochemical Sensors for Hepatitis C
4.6 Electrochemical Sensors for the Detection of
Rheumatoid Arthritis Biomarkers
4.7 Electrochemical Sensors for the Detection of Celiac
Disease Biomarkers
4.8 Electrochemical Sensors for the Detection of Urinary
Tract Infection Biomarkers
4.9 Challenges in the Use of Electrochemical Sensors in
Diagnostics
4.10 Conclusions
References
Chapter 5

In Vivo Sensors for Continuous Monitoring of Blood Gases,

Glucose, and Lactate: Biocompatibility Challenges and
Potential Solutions
Megan C. Frost, Alexander K. Wolf and Mark E. Meyerho
5.1
5.2

Introduction
Design of In Vivo Sensors

89

89
92
92
93
94
96
98
100
103
107
110
110
113
115
116
117
121
122
122

129

129
130

View Online

xiv

Contents

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

5.2.1
5.2.2

Sensing PO2/PCO2/pH in Blood

Sensing Glucose in Subcutaneous
Tissue
5.2.3 Sensing Lactate in Blood
5.3 Biocompatibility Issues that Inuence In Vivo Sensor
Performance and Reliability
5.3.1 Details of Biological Response in Blood
5.3.2 Details of Biological Response in
Subcutaneous Tissue
5.4 Strategies for Mediating Biological Response to
Implanted Sensors
5.4.1 Materials Development for Improved
Biological Response to Implanted
Sensors
5.4.2 Active Releasing Materials for Controlling
Biological Response
5.4.3 Materials to Mimic Biological Form and
Function
5.5 Summary
References

Chapter 6

Peroxynitrite Electrochemical Quantication: Recent

Advances and Challenges
Serban F. Peteu and Sabine Szunerits
6.1
6.2

Introduction
Challenges Confronted in the Accurate Quantication
of Peroxynitrite
6.3 Electrochemical Interfaces and Methods for
Peroxynitrite Detection
6.4 Strategies to Increase Response Sensitivity:
Electroactive Polymers or Graphene
6.5 Conclusions and Perspectives
Acknowledgements
References

Chapter 7

130
134
137
138
139
140
143

143
144
150
150
151

156

156
159
163
169
174
176
176

The Diagnosis of Myelodysplastic Syndromes

Alison S. Thomas and Christopher McNamara

182

7.1

183
183
184
184

Morphological Features of Myelodysplasia

7.1.1 Erythroid Dysplasia
7.1.2 Granulocytic Dysplasia
7.1.3 Megakaryocytic Dysplasia

View Online

xv

Contents

Dierential Diagnosis of Myelodysplasia

7.2.1 Metabolic Abnormalities
7.2.2 Drugs and Toxins
7.2.3 Infectious Diseases
7.2.4 Congenital Disorders
7.2.5 Acute Myeloid Leukaemia and Other
Haematopoietic Neoplasms
7.3 Initial Assessment of a Patient with Possible MDS
7.4 Additional Diagnostic Tools for MDS
7.4.1 Bone Marrow Aspirate Cytochemistry
7.4.2 Bone Marrow Trephine Biopsy
7.4.3 Cytogenetics
7.4.4 Immunophenotyping
7.5 Diagnostic Diculties in MDS
7.5.1 Unexplained Cytopenias in the Presence of
Minimal Dysplasia
7.5.2 Myelioproliferative/Myelodysplastic Overlap
Syndromes
7.5.3 Cytopenias in the Presence of a Hypocellular
Bone Marrow
7.5.4 Dysplasia in the Presence of Bone Marrow
Fibrosis
7.6 Subclassication of MDS
7.6.1 Original FAB Classication
7.6.2 WHO Classication 3rd Edition and
4th Edition
7.7 Evolving Understanding of the Pathophysiology
of MDS
7.7.1 Altered Methylation
7.7.2 Abnormalities in RNA Splicing Machinery
References

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

7.2

Chapter 8

The Prospects for Real-Time Raman Spectroscopy for

Oesophageal Neoplasia
Max Almond, Gavin Rhys-Lloyd, Jo Hutchings,
Geeta Shetty, Neil Shepherd, Catherine Kendall,
Nicholas Stone and Hugh Barr
8.1
8.2
8.3
8.4
8.5
8.6

The Clinical Problem

Endoscopic Recognition of Dysplasia and Cancer
Raman Spectroscopy
Instrumentation
Raman Spectroscopy for Medical Diagnosis
Raman Spectroscopy Assessment of Oesophageal
Pathology

186
186
186
187
187
187
188
188
188
188
190
191
191
191
192
192
192
193
193
194
194
195
196
196

201

201
202
204
206
207
208

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP011

xvi

Chapter 9

Contents

8.7 Fibre-Optic Raman Spectroscopy

8.8 Novel Raman Probe
8.9 Discussion
References

210
213
215
218

Volatile Analysis for Clinical Diagnostics

Catherine Kendall, Hugh Barr and Naresh Magan

222

9.1

222
223
224
227
227
228
231
231
233
233

Volatile Diagnostics
9.1.1 GC-MS and SIFT-MS Techniques
9.1.2 Electronic Noses
9.2 Applications of Volatile Fingerprinting
9.2.1 Food
9.2.2 Medical Applications
9.2.3 Environmental
9.3 Ventilator-Associated Pneumonia (VAP)
Acknowledgements
References
Subject Index

238

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP017

List of Contributors
L. Max Almond Surgical Fellow, Biophotonics Research Unit, Leadon
House, Gloucestershire Royal Hospital, Great Western Road, Gloucester
GL13NN, United Kingdom
Dr. Salzitsa Anastasova Post Doctoral Research Assistant, School of
Engineering and Material Sciences, Queen Mary University of London, Mile
End Road, E14N, United Kingdom
Prof. Hugh Barr Professor of Surgery, Consultant General & Gastrointestinal
Surgeon, Biophotonics Research Unit, Gloucestershire Hospitals NHS
Foundation Trust, Great Western Road Gloucester GL1 3NN & Craneld
Health, Craneld University, Craneld MK43 0AL, United Kingdom
Dr. Christophe Blaszykowski Research Associate, Department of Chemistry,
University of Toronto, 80 St. George Street, Toronto, Ontario M5S
3H6, Canada
Dr. Megan C. Frost Associate Professor, Department of Biomedical Engineering, Michigan Technological University, 1400 Townsend Dr., Houghton,
MI 49931-1295, United States of America
Dr. Szilveszter Gaspar Senior Researcher, International Centre of Biodynamics, 1B Intrarea Portocalelor, 060101 Bucharest, Romania
Madeleine Hu Premed Student, the College of New Jersey, 2000 Pennington
Road, Ewing Township, NJ 08628, USA, [email protected]
Dr. Joanne Hutchings National Institute for Health Research Post-Doctoral
Researcher, Biophotonics Research Unit, Leadon House, Gloucestershire
Royal Hospital, Great Western Road, Gloucester GL13NN, United
Kingdom
Dr. Catherine Kendall Senior Lecturer, Biophotonics Research Unit,
Gloucestershire Hospitals NHS Foundation Trust, Great Western Road
Gloucester GL1 3NN & Honorary Senior Clinical Lecturer Craneld
Health, Craneld University, Craneld MK43 0AL, United Kingdom
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

xvii

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP017

xviii

List of Contributors

Prof. Naresh Magan Professor, Personal Chair, Applied Mycology, Research

Director and Lead Academic for Doctoral Training Centre for Health and
Biosciences, Craneld Health, Vincent Building, Craneld University,
Craneld MK43 0AL, United Kingdom
Dr. Christopher McNamara Consultant Haematologist, The Department
of Haematology, Royal Free Hospital, London NW3 2QG, United
Kingdom
Prof. Mark E. Meyerho Philip P. Elving Professor of Chemistry, Department of Chemistry, University of Michigan, 940 N. University, Ann Arbor,
MI 49109-1055, United States of America
Dr. Serban F. Peteu Senior Scientist, National Institute for Research and
Development in Chemistry and Petrochemistry, 202 Splaiul Independentei,
060021 Bucharest, Romania
Dr. Gavin Rhys-Lloyd Post-Doctoral Chemometrics, Biophotonics Research
Unit, Leadon House, Gloucestershire Royal Hospital, Great Western Road,
Gloucester GL13NN, United Kingdom
Dr. Alexander Romaschin Director, Clinical Biochemistry and Keenan
Research Centre, St. Michaels Hospital, 30 Bond Street, Toronto, Ontario
M5B 1W8, Canada
Prof. Wolfgang Schuhmann Professor, Analytische Chemie-Elektroanalytik &
Sensorik, Ruhr-Universitat Bochum, Universitatsstr. 150, Bochum D-44780,
Germany
Sonia Sheikh Graduate Student, Department of Chemistry, University of
Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada
Prof. Neil Shepherd Professor of Histopathology, Biophotonics Research
Unit, Leadon House, Gloucestershire Royal Hospital, Great Western Road,
Gloucester GL13NN, Department of Histopathology, Gloucestershire
Hospitals NHS Foundation Trust, Faculty of Bioscience and Medicine,
Craneld Health, Craneld University, Bedfordshire, United Kingdom
Geeta Shetty Biophotonics Research Unit, Leadon House, Gloucestershire
Royal Hospital, Great Western Road, Gloucester GL13NN, United
Kingdom
Dr. Anna Spehar-Deleze Post Doctoral Research Assistant, School of
Engineering and Material Sciences, Queen Mary University of London,
Mile End Road, E14N, United Kingdom
Prof. Nicholas Stone Professor of Biomedical Imaging and Biosensing, Biophotonics Research Unit, Leadon House, Gloucestershire Royal Hospital,
Great Western Road, Gloucester GL13NN, United Kingdom, Biomedical
Imaging and Bio Sensing, School of Physics, University of Exeter, EX4 4QL,
United Kingdom
Prof. Sabine Szunerits Professor, Institut de Recherche Interdisciplinaire,
Universite Lille 1, Parc de la Haute Borne, 50 Avenue de Halley, BP 70478,
59658 Villeneuve dAscq, France
Dr. Alison S. Thomas Clinical Research Fellow, The Department of
Haematology, The Royal Free Hospital, London NW3 2QG, United
Kingdom

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP017

List of Contributors

xix

Prof. Michael Thompson Professor of Bioanalytical Chemistry, Department

of Chemistry and Institute for Biomaterials and Biomedical Engineering
University of Toronto, Canada
Prof. Pankaj Vadgama Director of the IRC in Biomedical Materials;
Professor of Clinical Biochemistry, Queen Mary University of London,
Mile End Road, London, E14N, United Kingdom; Consultant Chemical
Pathologist in the Department of Clinical Biochemistry at Barts Health
NHS Trust
Dr. Alina Vasilescu Senior Researcher, International Centre of Biodynamics,
1B Intrarea Portocalelor, 060101 Bucharest, Romania
Yuan Wang Senior Scientist, Siemens Healthcare Diagnostics, 511 Benedict
Avenue, Tarrytown, New York, NY 10591, United States of America
Alexander K. Wolf Graduate Student, Department of Chemistry, University
of Michigan, 940 N. University, Ann Arbor, MI 49109-1055, United States
of America

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-FP017

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

CHAPTER 1

Biosensor Technology and the

Clinical Biochemistry
Laboratory Issue of Signal
Interference from the Biological
Matrix
MICHAEL THOMPSON,*a SONIA SHEIKH,a
CHRISTOPHE BLASZYKOWSKIa AND
ALEXANDER ROMASCHINb
a

Department of Chemistry and Institute for Biomaterials and Biomedical

Engineering, University of Toronto, 80 St. George Street, Toronto, Ontario
M5S 3H6, Canada; b Keenan Research Centre and Clinical Biochemistry,
St. Michaels Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8, Canada
*Email: [email protected]

1.1 Laboratory Clinical Biochemical Assays

Assays for biochemicals of clinical interest in biological uids, tissues or feces
can be arbitrarily divided into tests performed in either the clinical biochemistry
laboratory or via a point-of-care device.1 The latter technology often takes the
form of a one-use structure that yields a result in a localized environment such
as the home, hospital bedside or general practitioners facility. The blood
glucose and pregnancy assay devices, which involve discardable test strips, will
be familiar to many. This type of structure has attracted a high level of interest
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Chapter 1

in recent years, not least from the commercial standpoint. Often quoted
advantages are said to lie in convenience, speed of analysis and cost savings.
There are of course attractive features in certain cases for procuring a measurement in a particular location. Obviously, biosensor devices might be
expected to play a major role in this arena. However, the present discussion
concentrates on the potential for use of biosensors in the dedicated central
biochemistry laboratory, which is typically located in a major hospital for
obvious reasons. Assays performed in this type of facility are very wide ranging
in scope with samples originating from the areas of diagnosis of medical
conditions, toxicology or monitoring of drug therapy. Sera or plasma dominate
the types of samples analyzed but many assays also involve various tissues,
feces and other bodily uids such as urine.
Test technology associated with all the various aspects of medicine is a large
and expensive operation. In terms of the cost of performing assays, the
Province of Ontario, for example, spends well over 1 billion Canadian (CDN)
dollars annually, the tests being conducted in both private laboratories and
hospital facilities. A large hospital such as St. Michaels in Toronto performs
a vast array of tests such as those involving immunology, hematology,
bacteriology, mycology, cytology, genetics and various aspects of pathology
in addition to biochemical measurements. With respect to the latter, some
3 million tests are performed annually with reagent costs alone being in the
region of 2 million CDN dollars. Although not intended to be exhaustive,
Table 1.1 provides an overview of the type of biochemical and immuno-assays
conducted by the clinical biochemistry on an annual basis together with total
numbers comprised of both in- and out-patient tests.2 The winner is the
thyroid stimulating hormone (TSH) test, which alone costs several million
dollars annually. The equipment employed in the portfolio of clinical measurements ranges from conventional chromatography to sophisticated mass
spectrometry, and combinations thereof. As would be expected given the sheer
volume of samples, the laboratory is highly automated and robotized. Typically, a blood sample will be bar-coded then centrifuged to extract serum before
entering an automated analytical train. Samples may then be directed automatically to their various analytical stations. A good example of a subsequent
analysis would be the prevalent magnetic-bead enzyme-linked immunosorbent
assay (ELISA) method for targets where immuno-assay is feasible. Despite the
high level of automation incorporated into the analytical train, it should be
emphasized that measurements are still performed in a batch-based fashion.
Having said that, it is certainly the situation that a number of protocols oer
multi-analyte capability, the aforementioned ELISA assay being a case in point.
Biosensor technology in principle oers rapid, label-free measurements,
which can be conducted in highly automatable congurations. Furthermore,
the possibility for multiplying tests using generic sensing physics is certainly an
attractive strategy. However, the clear reality is that biosensor devices do not
gure prominently in the clinical biochemistry laboratory. Use of ion-selective
electrode technology for certain cations might represent a contradiction to this
statement, but it could justiably be argued that these devices do not constitute

View Online

Biosensor Technology and the Clinical Biochemistry Laboratory

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Table 1.1

Overview of the type of biochemical and immuno-assays conducted

by a clinical biochemistry laboratory on an annual basis (source: St.
Michaels Hospital, Toronto, Canada).

Type of Test
Acetone, quantitative*
Albumin, quantitative*
Alcohol, ethyl-quantitative*
Alcohol, fractionation and quantication*
Amylase*
Barbiturates, quantitative*
Barbiturates, fractionation and quantication (serum) includes
other drugs requiring similar methodology (e.g. tricyclic
antidepressants)*
Bilirubin, total*
Bilirubin, conjugated*
pH*
pCO2, pO2 and pH in combination*
Carbamazepine, quantitative (Tegretol)*
Chlordiazepoxide, quantitative (Librium)*
Calcium*
Calcium ionized*
Catecholamines, fractionated*
Chloride*
Cholesterol, total*
Acetaminophen*
Carboxyhemoglobin*
CO2 content, CO2 combining power, bicarbonate
(measured not calculated)*
Creatine phosphokinase*
Creatinine*
Creatinine clearance*
Target drug testing, urine, qualitative or quantitative*
Diazepam, quantitative (Valium, Vivol)*
Drugs of abuse screen, urine*
Broad spectrum toxicology screen, urine includes conrmatory
testing*
Electrophoresis, serum including total protein
Electrophoresis, other than serum including total protein*
Glycosylated hemoglobin Hgb A1*
Flurzepam, quantitative (Dalmane)*
Glucose tolerance test in pregnancy*
Glucose tolerance test*
Gamma glutamyl transpeptidase*
Glucose, quantitative (not by dipstick)*
Glucose, semiquantitative (dipstick if read with reectance meter)*
High density lipoprotein cholesterol*
5H1AA quantication U*
Hemoglobin A2 by chromatography*
Iron, total with iron binding capacity and percent saturation*
Lactic acid (lactate)*
Lactic dehydrogenase (L.D.H), total*
Lipase

Number performed
(per year)
448
125 360
5010
5889
21477
13 898
704
72 000
22 545
20
62 757
504
10
91 723
9676
2485
227 057
34 663
5036
4
219 783
63 914
257 139
4884
37 691
37
48
7399
7556
599
19 394
0
2694
973
7705
188 863
46 538
32 179
1554
2030
21 054
13 733
19 245
1475

View Online

Table 1.1

Chapter 1

(Continued)

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Type of Test
Lipoprotein, electrophoresis*
Lipoprotein, ultracentrifugation*
Lithium*
Lidocaine*
Methotrexate (amethopterin)*
N-acetylprocainamide*
Magnesium*
Metanephrines, total U*
Methemoglobin*
Myoglobin, quantitative U*
Occult blood*
Osmolality (osmolarity)*
Oxalic acid (Oxalate) U*
Phenothiazines, quantitive U*
Phosphatase, alkaline*
Phosphatase, alkaline fractionation*
Phosphorus (inorganic phosphate)*
Plasma hemoglobin*
Potassium*
Protein, total*
Primidone, quantitative (Mysoline)*
Procainamide*
Quinidine*
Salicylate, quantitative*
SGOT (AST)*
SGPT (ALT)*
Sodium*
Thiocyanates*
Triglycerides*
Urea Nitrogen (B.U.N.)*
Uric Acid (urate)*
Urinalysis, routine chemical (any of S.G., pH, protein,
sugar, hemoglobin, ketones, urobilinogen, bilirubin, leukocyte
esterase, nitrite)*
Urinalysis microscopic examination of centrifuged specimen*
Valproic acid (valproate)*
VMA, Vanillylmandelic acid (Vanillylmandelate)*
Biochemical assays not included above*
Prostate specic antigen (PSA), total
Alpha Glycoprotein Subunit
Amylase isoenzymes
Apolipoproteins A and B
FK506 (Tacrolimus)
Lipoprotein (a)
Prostate specic antigen (PSA), total
Troponin
Homocysteine
Chylomicrons
Oligoclonal banding
Citrate
Sirolimus (Rapamycin)

Number performed
(per year)
3
499
866
86
115
74
84 408
2712
77
176
4293
10 044
3104
69
82 037
62
86 956
44
235 045
110 937
420
71
18
4875
88 095
80 304
237 209
15
33426
201 706
23 038
37 534
11 497
1154
1212
7844
1063
179
535
12 478
8577
792
4258
44 477
1585
4
1437
2967
526

View Online

Biosensor Technology and the Clinical Biochemistry Laboratory

Table 1.1

(Continued)

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Type of Test
Natriuretic peptide Brain (BNP)
Foetal bronectin
Bioavailable testosterone
Aldosterone*
Cortisol*
Aminoglycosides (e.g. Gentamicin, Tobramycin)*
Androsternedione*
Digoxin*
ACTH (andrenocorticotrophic hormone)*
Folate, in red cells, to include hematocrite and if requested, serum
folate*
Estradiol*
FSH (pituitary gonadotropins)*
Growth hormone*
HCG (human chorionic gonadotropins)*
Hepatitis associated antigen or antibody immuno-assay - per assay
(e.g. hepatitis B surface antigen or antibody, hepatitis B core
antibody, hepatitis A antibody)*
Aminophylline (Theophylline)*
Anti-DNA*
Diphenylhydantoin (Phenytoin), quantitative (Dilantin)*
Insulin*
LH (luteinizing hormone)
Ferritin*
Parathyroid hormone*
Progesterone*
Prolactin*
17-OH Progesterone*
IgE* - not to be billed for RAST test
T4, free - absolute (includes T-4 total)*
Testosterone
TSH (thyroid stimulating hormone)*
Phenobarbitone*
Vitamin B12*
C-peptide immunoreactivity*
Dehydroepiandrosterone sulfate (DHEAS)*
25-hydroxy vitamin D*
T-3, free*
Thyroglobulin*
Alphafetoprotein
Hormone receptors for carcinoma (to include estrogen and/or
progesterone assays)*
Total

Number performed
(per year)
3201
16
12 886
3059
3470
1108
1215
1103
1520
7710
1914
2807
2343
9848
11 469
167
1822
2424
802
3091
17 240
7256
908
3104
308
1131
11 888
12 767
30 406
363
10 897
1084
664
7703
6350
2466
942
793
3 274 901

biosensors anyway! The dearth of biosensor devices in the clinical biochemistry

laboratory also appears to be matched by the lack of employment of so-called
lab-on-a-chip3,4 devices, which are generally characterized by microuidic
sample handling. Time will tell but at present neither of these technologies
appears to have fullled the promise oered. This chapter attempts to evaluate

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Chapter 1

possible reasons for this including a detailed look at one major problem that
of interference by adsorbed species from the biological matrix. In order to
provide a backcloth to discuss the issues at stake, we take a prior concise
look at biosensor technology. The chapter has something of an emphasis on
label-free methodology.

1.2 Biosensor Technology

1.2.1 Biosensor Architecture
A biosensor is composed of chemical recognition sites (probes) attached to a
substrate surface that, in turn, is in close proximity and union with a transducer. Ideally, this conguration responds sensitively and selectively to the
presence of (bio)chemicals, we term the target or analyte, usually present in the
liquid phase. The technology is based on the recognition of species through
selective binding of the target to the probe at the substrate surface of the device
with such surface presence being converted into an electrical signal. The overall
architecture is depicted in Figure 1.1. Note that probes are invariably composed
of biological entities such as cells or biochemicals such as antibodies. The
sample itself may or may not be of biological origin. An important aspect of
biosensor technology is the nature of the response of the device with respect to
time. The eld is often considered to include the aforementioned one-test
disposable systems, where there is no attempt to conduct a measurement over a
period of time. This type of device is more often than not at the heart of the
point-of-care assay. From the clinical biochemistry point of view with regard to
the central laboratory, both time-responsive and one-shot structures might be
applicable. This issue will be considered in more detail later in the chapter.
The signal obtained from a selective biosensor response will take the
following form (Eq. (1.1)):5
Xn Sn1 C1 Sn2 C2       Snn Cn

Figure 1.1

1:1

Schematic representation of a biosensor architecture featuring the biosensing interface, transducer and output system.

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

where C and S values represent the concentration of analyte and interferents

in proximity to the device, and the response sensitivities, respectively. An
enormous number of components would be expected to be involved in this
equation in terms of samples such as serum or urine. A sensitive response
implies maximization of one S value and, for a selective signal, a minimization
of all other S values. Clearly, the sensor signal will be a composite of the
chemistry of the attachment process and the physical perturbation caused by
the probeanalyte complex. In this respect, it is crucial that, when designing a
sensor for particular applications, the physics of the complete transduction
process be thoroughly characterized and understood. In fact, some devices
simply respond to the presence of the analyte, whereas others detect structural
shifts caused by targetprobe binding. The distinction between these mechanisms is not always evident.

1.2.2 Probe Attachment to Device Surfaces

A mandatory aspect of biosensor fabrication is the attachment of the biochemical probe to the surface of the transducer of choice. Over the years, a host
of methods have been developed with this goal in mind.6,7 A number of criteria
are relevant to this component of sensor production:
 It will be vital to retain the binding activity of the probe in terms of its
ability to bind the target. It is anticipated that certain proteins, for
example, may be partially denatured and lose anity upon attachment to
the sensor surface, whereas others such as antibodies may be more robust.
 The proper spatial orientation of the probe with respect to the plane of the
device surface will be an important element to ensure exposure of binding
sites for target capture.
 In terms of sensitivity, it is crucial to maximize the density of probe
molecules per surface area unit. Simply stated, the more probe available
on the device surface, the greater will be the sensor signal.
 The spatial characteristics of the probe with regard to the surface plane are a
factor that is important but not studied widely. If probes are in too close
proximity on the surface, this may result in steric hindrance to target
binding. In practice, there exists a maximal (optimal) eective probe loading.
 There are other more mundane factors such as the cost of reagents and the
potential longevity of the modied sensor surface. These will clearly be
critical from a commercial standpoint.
Some examples of methods for probe attachment follow. This short compendium is not intended to be an exhaustive treatment.

1.2.2.1

Noncovalent Attachment

The probe of interest can be chemically or physically adsorbed from solution

directly, in most cases, on the substrate surface of the device. Various protocols

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Chapter 1

are employed to achieve such an eect such as dip casting, painting, spraying
and spin coating, where these terms will be self-explanatory to the reader. The
interaction of the probe with the surface will be characterized primarily by
hydrogen bonding and/or hydrophobic interactions. Adsorbed molecules can
also act as a linker for eventual biomolecule attachment. A system that is widely
employed is avidin/streptavidin/neutravidin chemistry.8 Avidin is a tetrameric
protein that contains four binding sites for the ligand, biotin. Interaction of
the protein with this molecule results in a particularly strong binding
(Kd 1015 M). Accordingly, various biomolecules can be modied with relative facility by biotin incorporation in order to link them to surface-attached
avidin or a sister molecule such as the aglycosylated version, neutravidin. An
analogous strategy uses protein A.9 The latter is a polypeptide (MWB42 kDa)
isolated from Staphylococcus Aureus that is capable of binding specically to
the Fc region of various antibody molecules and has thus been employed in
immunosensor technology.
Another noncovalent approach is the trapping of the recognition molecule
within the three-dimensional structure of a specic chemical matrix. The matrix
may simply act as a holding moiety or be modied in some way to take part
in the transduction process. For the former, polymers such as polyacrylamide
have been employed in a number of experimental protocols. However, there are
a number of potential disadvantages to this type of approach for placing the
probe on the device surface including the possibility of biomolecule leakage
and/or denaturation. An additional consideration is the necessity for the target
to diuse into the polymer matrix in order for the biochemical interaction to
take place. An analogous procedure uses encapsulation by sol-gel technology,10
wherein a solution of a monomer such as an alkoxide (the sol) is induced to
polymerize into a biphasic conguration (the gel) that incorporates both liquid
and solid. A typical monomer among many would be tetraethylorthosilicate
(TEOS EtO4Si), which is readily hydrolyzed by water to produce a siloxanebased polymeric structure with a gel consistency. A cavity can be formed
around the probe of interest in somewhat the same manner as for the polymers
mentioned above.

1.2.2.2

Covalent Binding

Attachment via covalent bonds has been by far the most used approach. A very
wide variety of chemistries have been employed with modest success in terms of
the criteria outlined above.11,12 Many functional groups, whether directly
present on the device substrate or obtained by modication, have been utilized
to form a probe partial monolayer. Groups are available on biomolecules to
instigate the surface link and examples of these for proteins are presented in
Table 1.2.
A common protocol to bind enzymes, antibodies or molecular receptors to
the substrate is to initially functionalize the surface, which is then followed by a
second reaction of activation. In essence, this process simply allows a
convenient, highly reactive linker moiety to be introduced to the system.

View Online

Biosensor Technology and the Clinical Biochemistry Laboratory

Table 1.2

Target functional groups for biosensor

immobilization of proteins.

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Functional Group

Amino Acid

Cysteine
Thiol

Lysine (side chain)

Amine

Tyrosine
Phenol

Histidine
Imidazole

Tryptophan

Indole

Arginine

Guanidine

Glutamic acid, Aspartic acid

Carboxylic acid

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

10

Figure 1.2

Chapter 1

Schematic representation of a common strategy (EDC/NHS chemistry) to

covalently biofunctionalize surfaces via a preactivation process.
(Reprinted with kind permission of the Royal Society of Chemistry.)

The resulting interface is then normally allowed to react in turn with one of the
protein nucleophilic groups mentioned above. The literature is replete with
many examples of this sort of approach and, in Figure 1.2, we provide a
schematic of one strategy. If groups already evident on the device surface are
not used directly, functionalization of surfaces is often achieved with species
such as aminopropyltriethoxysilane (APTES), which reacts with interfacial
hydroxyl groups on whatever substrate they are present.

1.2.2.3

Self-Assembled Monolayer Chemistry

The introduction of close-packed monolayers has been used widely in biosensor

development. Two very dierent strategies are employed the rst being the
LangmuirBlodgett lm technique.13 In this experiment, close-packed monolayers are imposed on a surface by transferring lipid or lipid-like (amphiphilic)
lms from the Langmuir trough, under the correct surface pressure conditions,
by a dipping process. In principle, several layers can be deposited in sequence
using this dipping approach. The very important advantage oered by this
technique is the possibility to combine articial lipid membrane congurations
directly and in situ with integral membrane proteins (IMPs). Such membrane
systems are, of course, widely prevalent as signaling devices in biology.
An important chemistry, which has been signicantly developed in recent
years, is the self-assembled monolayer (SAM). This approach relies on the use
of linking molecules that are engineered to spontaneously form ordered molecular assemblies on solid substrates. In this case, the most common strategy
has been the assembly of relatively long chain thiols on the surfaces of clean
gold substrates,14,15 as shown in Figure 1.3. The distal end of bifunctional thiols
can then be employed for conventional covalent binding of proteins and oligonucleotides as described above (Figure 1.2).16 An alternative technique is the
use of trichlorosilanes rather than thiols, which can be bound to surfaces that
have been functionalized with or naturally possess OH groups.17 Examples of

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

11

Figure 1.3

Long-chain alkyl thiol deposition onto a gold substrate to generate a selfassembled monolayer.
(Reprinted with kind permission of the Royal Society of Chemistry.)

Figure 1.4

Preparation of an organosilane adlayer on quartz and subsequent covalent

functionalization.
(Reprinted with kind permission of the American Chemical Society.)

substrates in this case would be silicon dioxide (e.g. quartz Figure 1.4)18 and
indium-tin oxide (ITO).19

1.2.3 Devices and Transduction

A very large amount of work has been performed on a variety of biosensor
structures. A bibliography detailing this eort up to the end of 2012 is provided
herein.20 Three distinct areas of physics have been employed with respect to the
transduction process, these being electrochemistry, acoustic wave technology
and electromagnetic radiation. There is considerable variety in terms of the
nature of measurement, for example, possible label-free operation and secondly
dipstick or response with time methodology. Moreover, a number of adjunct
techniques have been employed in order to enhance the level of information
obtained from the sensor determination.

View Online

12

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

1.2.3.1

Chapter 1

Electrochemical Devices

A number of dierent approaches are possible via electrochemistry,21,22 which

involves the transfer of electrons or charge at solid/liquid or liquid/liquid
interfaces. Those involving solid structures unsurprisingly dominate the eld,
which is summarized in Table 1.3. A particularly useful feature of many of these
techniques is that they can be combined naturally with various forms of
integrated electronic circuitry. Indeed, the necessary chemistry can be imposed
directly on the surface of such an electronic device (see below).
Historically, one of the earliest devices generated from the world of electrochemistry was the sensor based on potentiometry. In this technique, an indicating electrode for the analyte of interest is incorporated in an
electrochemical cell (with reference electrode) for which minimal or no current
is passed. In this category, the ion-selective (indicating) electrode (ISE) forms
the basis of systems, which are capable of detecting ions and other species of
biochemical interest.23,24 By far the most common in use is the glass electrode,
which is sensitive to changes in hydronium-ion concentration. Other systems
available for selective ion detection include electrodes that employ crystalline
matrices, e.g. LaF3 for F, or incorporate liquid ion exchangers in polymer
matrices such as systems for sensing Ca21. The ISE has also been the basic
structure used to develop enzyme-based molecular sensors where analytes are
allowed to interact with immobilized enzymes and transform into products
detected by the electrode.
An alternative transducer to the conventional electrode outlined above is the
eld-eect transistor (FET), which was introduced by Bergveld25 in the 1980s
for ion sensing. Figure 1.5 shows the structure of a typical n-channel enhancement mode insulated-gate FET (IGFET). A conductive pathway, termed
a channel, can be instigated between n-type regions. One of the n-type contacts
is called the source and the other the drain; a potential is applied that drives
current in the channel. The contact on the surface of the center part of the
insulating layer is called the gate. It is in this location that the device is employed to detect charge changes connected to the chemistry conducted at the
gate/solution interface (Figure 1.6). Note that certain regions of the device must
be encapsulated and that the system as mentioned above needs an adjunct
reference electrode. The device has been employed in various formats including

Table 1.3

Electrochemical techniques available for application in biosensor

technology.

Technique

Methodology

Amperometry
Conductometry
Coulometry
Potentiometry
Voltammetry

Measures current versus concentration; potential is controlled

Measures conductance ( 1/resistance) at controlled concentration
Measures change over time with controlled potential
Measures electrochemical potential at current 0
Measures current at applied potential; controlled concentration of
electroactive species

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

13

Figure 1.5

The n-channel enhancement mode insulated-gate eld eect transistor

IGFET.
(Reprinted with kind permission of the Royal Society of Chemistry.)

Figure 1.6

IGFET typical immunosensor set-up. Note the insulating protection of

source and drain contacts.
(Reprinted with kind permission of the Royal Society of Chemistry.)

the detection of ion concentrations (ISFET)26 or immunochemical interactions

(IMMUNOFET).27
Techniques that are based on measurement of the resulting current versus
applied potential are called voltammetric methods. Anodic stripping and cyclic
voltammetry would be two examples of such techniques. Amperometry has
many denitions in the literature but is more often than not simply regarded to
involve the measurement of current associated with an oxidationreduction
reaction at an electrode. It is worth noting that the genesis of biosensor
technology was the oxygen-mediated electrode for the amperometric assay of
glucose. Over the years, the electrochemical detection of this analyte, given its
importance, has been the subject of literally thousands of studies, the history of

View Online

14

Chapter 1

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

28,29

which has been reviewed by Wang.

One of the most important advances
in the technology was the introduction of organometallic mediators such as
ferrocene. This iron p-arene complex composed of a cyclopentadiene sandwich
of Fe acts as a convenient electron acceptor for glucose oxidase, thus avoiding
the dependence on the role of O2.
The detection of nucleic acid interactions and duplex formation at an
electrode surface by an amperometric protocol has been the subject of intensive
research for some time.30 One example among several is the use of the redox
properties of, for example, ruthenium complexes to enhance the electroactivity
of nucleic acid species (the intrinsic redox behavior of nucleic acid moieties
is insuciently sensitive for analytical purposes). The electrochemistry community often describes this protocol as a label-free detection strategy, which
is obviously nonsense given the necessary use of the Ru adjunct agent.
Finally, we mention electrochemical impedance spectroscopy (EIS). In
this technique, a sinusoidally-varying voltage is applied to an electrochemical
system and the resulting current is measured, usually over a range of angular
frequencies. Recording of these parameters can be employed to compute the
real and imaginary components of the electrical impedance (Z). Primarily, the
system has been used to detect biomolecular recognition events taking place at
an electrode much as the situation with the other electrochemical arrangements
outlined above.31

1.2.3.2

Acoustic Wave-Based Biosensor Technology

Acoustic wave sensors are generally very much associated with piezoelectric
physics. (It should be noted, however, that a number of available structures do
not employ piezoelectric components in a direct sense.) Piezoelectricity is the
electric polarization produced by mechanical strain in crystals belonging to
certain classes, the polarization being proportional to the strain and changing
sign with it.32 Of course, we now recognize that the reverse is true, that is, a
specic crystal can be mechanically strained when subjected to an electric
polarization. The devices used in biosensing have in common the deformation
caused on a piezoelectric crystal by the application of an electric eld. The
origin of this eect lies in the interaction between electric charge and elastic
restoring forces in the crystal. All importantly, the phenomenon cannot take
place in a crystal possessing central symmetry. In order to maximize the
coupling of electrical and mechanical eects, it is conventional for practitioners
to use particular slices of crystals or cuts. For example, with respect to the
predominant piezoelectric material, quartz, these are AT-, ST- and BT-cuts.
Movement of particles in a piezoelectric crystal caused by an oscillating
electric eld-induced stress, restored by elastic forces, leads to standing or
travelling waves in the material and the phenomenon of piezoelectric
resonance.
Several types of acoustic wave devices have been used as biosensors, but the
simple thicknessshear mode (TSM) has been by the most employed for
detection purposes. The TSM is composed of an electroded (often gold)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

Figure 1.7

15

Working principle of the thickness-shear mode (TSM) acoustic wave

device featuring the standing wave in the substrate with propagation of
a damped acoustic shear wave into the liquid.
(Reprinted with kind permission of the Royal Society of Chemistry.)

piezoelectric wafer (usually AT-cut quartz), as depicted in Figure 1.7. An

oscillating electrical potential is applied via the electrodes in order to drive
mechanical motion in the device. A resonant acoustic shear wave is generated,
which travels through the piezoelectric material with little energy dissipation.
The wave is reected at the device/surroundings interface in order to maintain a
standing wave (Figure 1.7). In chemical and biosensor applications, material is
generally added to the device surface, for example, species involved in biomolecular interaction. On a simple theoretical basis, this scenario was the
subject of very early work by Sauerbrey,33 who showed that when material is
deposited on the device surface the resonant frequency is changed according to
his famous equation that is expressed for quartz as the piezoelectric material
(Eq. (1.2)):
2f02
Df  p
Dm
A rq mq

1:2

where Df is the frequency change, f0 is the primary resonant frequency of the

sensor, A is the eective surface area associated with the piezoelectric process,
rq is the density of quartz, mq is the shear modulus of the particular cut of
quartz employed and Dm is the change in mass. The idea of mass detection has
spawned the ubiquitous term, quartz crystal microbalance or QCM. Unfortunately, this argument has been widely extended to operation of the device in the
liquid phase and has become something of a dogma, especially in the community of electrochemists. In such a medium however, it is a reality that
acoustic energy is transferred to the surrounding liquid resulting in a damped
wave via the eects of bulk phase viscosity (Figure 1.7).
Experimentally, a number of approaches have been employed in order to
measure the resonant frequency and other acoustic parameters, in some case in
a static fashion, and in others, in owing liquid through a ow-injection conguration. The most rigorous approach is that provided by what is often

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

16

Chapter 1

termed acoustic network analysis. In this method, the magnitude and phase of
impedance of the sensor is determined at a set of frequencies under resonance
conditions. A network analyzer is employed to record the ButterworthVan
Dyke equivalent circuit, which essentially relates the physical properties of the
device to electrical parameters. In terms of applications, recent years have
seen a rapid increase in the development of TSM technology, which has
been employed to detect surface-induced protein conformational changes,34
immunochemical interactions,35 nucleic acid hybridization36 and cell-surface
attachment.37
A recent acoustic wave device development is the introduction of the
electromagnetic piezoelectric acoustic sensor (EMPAS) structure.38 In this
technology, acoustic waves are instigated in an electrodeless quartz wafer by an
electromagnetic eld produced in close proximity to the wafer by a at spiral
coil (Figure 1.8). The secondary electric eld associated with the coil drives
the piezoelectric eect in the device. The conguration possesses a number of
important advantages:
 It is not necessary to operate the device with contact metal electrodes
in place and with electrical connections. Acoustic resonance is driven
remotely. This renders advantages in terms of ow-through design and
surface chemistry can be studied directly in an unhindered fashion.
 Crucially, it is possible to operate the sensor at ultra-high frequencies, e.g.
1 GHz, via bulk acoustic wave overtones. This leads to higher analytical
sensitivity.
 It is possible to tune the device with ease to specic frequencies, which
could potentially lead to important interfacial chemical information.
 The surface chemistry for biomolecule attachment involves SiO2, which is
a more developed area of chemistry than is the case for binding to metals
such as gold.

Figure 1.8

Working principle of the electromagnetic piezoelectric acoustic wave

sensor (EMPAS): acoustic resonance is remotely driven in the electrodefree quartz substrate by a secondary electric eld induced by the electromagnetic eld associated with an AC-powered at spiral coil.
(Reprinted with kind permission of the Royal Society of Chemistry.)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

17

Surface-launched wave devices have also been employed in biosensing, where

waves are generated in a piezoelectric substrate by transducers that are placed
on the surface of the material. Particle movement, which is generally detected
by separate transducer(s) imposed on the same surface, is often restricted to the
near surface of the substrate, unlike the case for the TSM discussed previously. For an introduction to several of these devices, please refer to reference
39. The surface acoustic or Rayleigh wave (SAW) sensor has an interdigital
transducer (IDT) fabricated on the piezoelectric substrate (e.g. ST-cut quartz)
as shown in Figure 1.9. The chemistry related to the sensing processes is conducted on the area between the IDTs and it is generally the case that the
electrodes have to be isolated from the liquid if the device is operated in that
medium, in a similar fashion to that described above for FETs. An example of

Figure 1.9

Upper: schematics of a surface acoustic wave device. Center: Rayleigh

waves contain both shear and compressional components. Lower: schematics of a surface transverse sensor.
(Reprinted with kind permission of the Royal Society of Chemistry.)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

18

Chapter 1

another device employed via this technology is also portrayed in Figure 1.9.
This is the surface transverse wave (STW).40 The use of these biosensors, which
all involve very similar biomolecule immobilization strategies to those outlined
above, have been reviewed by Rapp and coworkers.41

1.2.3.3

Electromagnetic Radiation and Optical Biosensor

Technology

Virtually, the full gamut of physics oered by optical science has been employed over the years for the detection of fundamental biophysical processes,
biochemical binding events or species of bioanalytical interest such as biomarkers for disease. Techniques include those based on measurements of absorption, luminescence, interference, reectance, scattering (including Raman
spectroscopy) and refractive-index phenomena. The wide variety of techniques
employed in optical sensing have been nicely reviewed.42 One of the main
features of this area has been the concentration on label-free detection.43
One of the earlier devices was based on optical waveguide technology and, in
particular, the optical-ber sensor. In essence, light transmission along a ber is
produced via a guided wave through an integral process of internal reection.
Light can transmit along a ber with complete internal reection or via some
loss of electromagnetic energy through the eect of reection/refraction at
the interface where materials of dierent refractive index are involved. Both
these processes have been employed extensively in the development of biosensors and in this context the term optrode has appeared, as obviously
spawned by the older concept of an electrode in the eld of electrochemistry. In
the rst methodology, the ber is simply used as a delivery system for light
interaction with an optical conguration placed at the distal end of the structure,44 as shown in Figure 1.10a. In the second scenario called the intrinsic
structure, some light energy nds its way under certain conditions into the
medium outside the ber in the form of a penetrative evanescent wave
(Figure 1.10b). Importantly, the interaction of surface chemistry at this interface with the evanescent wave can result in perturbation of the light phase,
intensity and polarization. There are many examples of such an arrangement in
biosensor detection.45 This type of device has been employed successfully in the
assay of real samples such as for the detection of pathogens.46
Surface plasmon resonance (SPR) has become one of the leading technologies
with respect to detection in bioanalytical and biophysical chemistry. The
physics eect is based on the fact that valence electrons of metals such as gold
and silver exhibit oscillations in their density. If these oscillations, in the form
of surface waves, are present at the interface of the metal with a material of a
dierent dielectric constant, they can be excited by the introduction of electromagnetic radiation. The classical approach to the study of this phenomenon
was initially introduced by Kretchmann47 and is shown schematically in
Figure 1.11a. The incident radiation is reected at the metal/dielectric
boundary resulting in an evanescent wave in the metal, much as described

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

19

above for optical bers. At the correct (resonant) angle, this wave can couple in
a resonant fashion with the frequency of the incoming radiation, resulting in the
excitation of electron density oscillation mentioned above, leading to the term
surface plasmon resonance. Experimentally, in sensor operation, biochemical
binding events are conducted at the metal surface and detected through
measurement of shifts in the SPR angle, observed wavelength of absorption or
change in the position of reectivity. A resonant transfer of electromagnetic
energy into the surface plasmon wave is observed through a minimum in the
plot of incident angle versus reected intensity (Figure 1.11b).

Figure 1.10

(a) Schematics of extrinsic ber-optic delivery of radiation to a cell for

absorption measurement. (b) Evanescent radiation penetrating to the
exterior of a ber in an intrinsic conguration.
(Reprinted with kind permission of the Royal Society of Chemistry.)

Figure 1.11

(a) Kretschmann conguration for surface plasmon resonance (SPR). (b)

Incidence angle versus reected intensity of radiation as a result of SPR.
(Reprinted with kind permission of the Royal Society of Chemistry.)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

20

Chapter 1

A simplied schematic of a typical SPR experiment is shown in Figure 1.12.

This type of instrument designed for work in the biosensor arena is generally
capable of analyzing several channels (gold surfaces) using microuidic sample
introduction in a real-time, label-free fashion through ow-injection technology. The system generates a response plot, which is often referred to in the eld
as a sensorgram, presumably in the light of chromatograms and the like!
With respect to applications, the SPR technique has been employed in a very
wide variety of cases such as adsorption of proteins, cells and nucleic acids on
gold and modied metal surfaces, detection of biomolecular interactions such
as found in immunochemistry and nucleic acid duplex formation.4853 The
method has proven to be particularly helpful for epitope determination in the
former area.
Finally in this section, we mention the technique of interferometry. The latter
is based on the superposition of, usually, electromagnetic waves that yields
information concerning the original nature of the waves. In biosensing technology, several types of device designs have been employed in order to attempt
the detection of biochemical species, examples being MachZehnders, Youngs
and Hartmans sensors (these are depicted in Figures 1.13ac). The rst of these

Figure 1.12

Typical SPR experiment involving ow injection of target analyte into

the ow cell.
(Reprinted with kind permission of the Royal Society of Chemistry.)

Figure 1.13

(ac) MachZehnders, Youngs (multichannel) and Hartmans interferometer sensors, respectively.

(Reprinted with kind permission of Elsevier.)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

21

involves two dierent light pathways, as is typical in interferometry, with one

being subjected to passage through the sample. The laser radiation is then
combined with a reference beam resulting in the interference-based signal. The
analytical arm incorporates an evanescent eld, which interacts with, in this
application, a biochemical moiety. This technology was employed in the earlier
years for the detection of protein chemistry, where analytical detection limits of
around 50 pM were found.54
An additional interference-based structure is Youngs device.55 The overall
principle is similar to the sensor mechanism employed in MachZehnders
device, the detection mode is dierent, however. In this case, the output from
the sample and reference channels is combined to form interference fringes on a
CCD screen. The method involves Fourier transform of the spatial intensity
measure at the detector screen. In Hartmans interferometer,56 electromagnetic
radiation is coupled into a waveguide using gratings, which allows the interference phenomenon to occur between reference and sample strips. As with the
other two devices outlined above, protein chemistry was detected as imposed on
the waveguide strip.

1.3 Biosensors and Measurement of Clinical Targets

It is very evident from Table 1.1 that clinical biochemical detection and
measurements involve a vast array of targets in biological matrices. Although it
is not the aim of this chapter to detail the fundamental requirements of the
laboratory these are provided elsewhere57 we do summarize some of the
main necessary criteria. These are then compared in the context of what is
oered by biosensor technology. Essential properties and protocols of clinical
determinations are:
 High throughput. As outlined above, the clinical biochemistry laboratory
faces literally thousands of samples on a daily basis, which may have one
particular analyte determination required or perhaps several on the same
sample. The sheer numbers involved mandate a very high level of automation and robotic technology. This means that considerable technical
eort is required to keep equipment in operation. For the most part,
modern-day clinical analysis incorporates the capability to handle samples
in a batch-wise process.
 Speed of analysis. Obviously, the requirements in this criterion will vary
enormously. However, it would be desirable to possess the capability to
perform rapid analysis, for example, as requested by a physician in an
emergency situation.
 Generic methodology. Where possible, it is advantageous to possess the
possibility to perform the determination of dierent targets employing the
same detection physical chemistry. It goes without saying that this aspect,
by the very nature of target analyses required, will be somewhat limited.
However, there exist several technologies possessing this attribute, such as
magnetic-bead ELISA.

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

22

Chapter 1

 Validation of doseresponse and limit-of-detection (LOD) characteristics.

The former parameter refers to the meaningfulness of the signal obtained
from the target in a biological matrix in terms of its concentration. This
would be more familiar to the analytical chemist as the calibration
curve for the particular method involved. The LOD would also be well
known to the analyst being the minimum concentration value that a
method can detect (in analytical chemistry, generally 3 times the standard
deviation of the background noise).
 Cost per analysis. Given the great expense involved in the operation of a
central clinical laboratory, it is hardly surprising that the calculated cost
per assay will always be a factor under consideration. Factored into this
are also the costs associated with salaries of operators, reagents, equipment and its maintenance.
 Functionality in complex media. It is clearly crucial for a particular method
to function in samples such as blood, serum, plasma or urine. We have left
this aspect to the end for reasons that will become apparent.
We now turn to the ideal biosensor. The relevance of these criteria to the
clinical chemistry operation will be apparent to the reader.
 High sensitivity with resulting low value of limit-of-detection. This parameter will be governed by the physics of the detection technique employed the S values outlined above.
 High accuracy in terms of concentration measurement for the analyte.
Signaling conducted with high reproducibility a precision issue.
 The dynamic range with respect to analyte response should be as wide as
possible. For certain targets, indeed, clinical values may vary
considerably.
 Signal approach dipstick/batch or real-time sensing. It would be advantageous for a particular device to be capable of operation in the dipstick/batch manner or as a real-time sensor. The speed of response will
obviously be important in both situations although for dierent reasons.
Real-time operation oers the potentially attractive possibility to incorporate the biosensor into an automated ow-injection scenario. This is
not the well-known segmented situation, but a technique based on analyte
dispersions allowed to ow past the device surface. The ow-injection
analysis (FIA) will require that the biosensor signal be reversible or the
target can be washed from the surface within the ow system.
 Device response calibration in terms of concentration. This is mandatory
and especially important for the real-time sensor. This feature has constituted an extremely dicult problem when it comes to, for example, the
operation of a corporeal implantable structure. Early solutions to this
issue may well lie in a strategy of device self-referencing.
 Selectivity or even specicity (see reference 5 for an excellent denition of
these parameters) with respect to response to the analyte. In this respect
with specic regard to clinical chemistry, it is essential that the device

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

23

function selectively in media that contain very high concentration of

various biomolecules such as proteins and even cells. These species will
tend to adsorb to the sensor surface potentially interfering with the required signal. This is one of the reasons why the legendary closed-loop
implantable glucose device is not available despite many years of research.
The issue of nonspecic adsorption (NSA) represents a true Achilles Heel
for application of biosensor technology in clinical biochemistry.
Finally, it is worth noting that there have been several attempts to assess the
capability of biosensors to detect molecules of clinical interest. Table 1.4 shows
an example where the authors place performance criteria in the context of
classical analytical chemistry.58

1.4 Signal Interference and the Non-specic Adsorption

Problem
In biosensor technology, the undesired non-specic adsorption (NSA) of
adversary species (sometimes also referred to as non-specic binding) as
opposed to the specic adsorption of target analytes is a serious and prevailing concern for many detection platforms intended to perform analysis in
extremely challenging biological samples, more often than not blood serum or
plasma. In fact, even cleared of the cellular components of blood, these biological matrices still consist of highly complex mixtures of potentially interfering biomolecules in particular various types of proteins at high
concentration (6080 g/L)59 that have the propensity to adsorb nonspecically to the sensing surface of devices thereby preventing the detection,
not to mention the quantication, of target analytes present at considerably
lower concentration (down to ng/L or a dierence of nine orders of magnitude).60 Indeed, during the biorecognition phase (wherein analytes are expected
to site-specically bind to complementary receptors immobilized on the sensing
platform), non-specically adsorbing species also generate physicochemical
stimuli that are indiscriminately detected by the biosensor and interfere with the
specic response of the target analyte (Figure 1.14).
The most unfortunate immediate consequence of NSA besides altering the
performance of the biosensor, may it even be operational in biological samples
in the rst place is the occurrence of false positives, i.e. a response incorrectly interpreted as a genuine binding event, which understandably renders
these biosensors irrelevant for real-world applications.61 Arguably, NSA is the
single most important reason why biosensors still have not found a prominent
place as alternative diagnostic tools in clinical analysis at the beginning of
the 21st century despite tremendous promise notably in terms of cost/ease
of operation and miniaturization for point-of-care applications the ultimate aim of biosensor technology.62 More generally, NSA generates a high,
often overwhelming background signal that may lower the sensitivity of the
biosensor (poor signal-to-noise ratio) to clinically irrelevant levels.60,63

Analytical gures of merit of dierent types of biosensor detecting various molecules of clinical interest. (Adapted from
reference 58.)

Analyte detected
(matrix)
Glucose (buer)
Glucose (blood
serum)
Glucose (human
serum)
Glucose (buer)
Glucose (buer)
Glucose (buer)

Sensor type

Linear range

LOD

Sensitivity (slope) Additional information

CNT enzyme-based biosensor

Amperometric enzyme-based
biosensor
Amperometric enzyme-based
biosensor
Amperometric CNT-based
biosensor
CNT enzyme-based biosensor
CNT enzyme-based biosensor

Up to 10 mM
1.0 mM0.8 mM

0.25 mM
0.5 mMa

1.0 mM1.6 mM

0.69 mMa

0.0050.3 mM

3 mMa

Up to 2 mM
15.0 mM6.0 mM

4 mM
7 mMa

0.025.7 mM

8.2 mMa

0.0040.7 mM

1.0 pMa

0.0040.10 mM

1.4 mMb

1.2 mM1 mM

0.12 mMc

Up to 12 mM

0.18 mM

30.2  0.5 mA/mM r2 0.999; n 10

2.3 mA/M
r2 0.999; n 7
CV(%) 4.8
69.26 mA/M.cm2 r2 0.999; n 25
CV(%) 1.1
80  4 mA/M.cm2 r2 0.996; n 5
CV(%) o5%
0.33 mA/mM
r2 0.998
r2 0.992; n 8
CV(%) 1 .6
8.8  0.2 mA/
CV(%) 2.1
M.cm2
1.55 mA/mM
r2 0.9954; n 8
CV(%) 4.2%
y 14.23x 0.2602
r2 0.999; n 11
CV(%) 2.5
r2 0.9998; n 11
CV(%)o4.0
1.61 nA/mM

0.26.0 mM

0.2 mMa

0.559 mA/mM

50400 mg/dL

25 mg/dL

25400 mg/dL

25 mg/dL

50500 mg/dL

50 mg/dL

Glucose (buer)

Amperometric enzyme-based
biosensor
Cholesterol
Amperometric CNT-based
(human serum)
biosensor
Cholesterol
CNT enzyme-based biosensor
(human serum)
Amperometric enzyme-based
biosensor
Amperometric enzyme-based
biosensor
CNT amperometric biosensor
Impedance enzyme-based
biosensor
Impedance enzyme-based
biosensor
SPR enzyme-based biosensor
Voltammetric enzyme-based
biosensor

0.010.06 mM

r2 0.9946
7.76105 Abs/
mg.dL
1.04 m1/mg.dL

y 0.03546x 7.76105

0.85 nA/mM

r2 0.99
y 0.6054x 0.2411

Chapter 1

Cholesterol
(buer)
Cholesterol
(buer)
Cholesterol
(buer)
Cholesterol
(buer)
Cholesterol
(buer)
Cholesterol
(human serum)
Cholesterol
(human blood)

24

.
Published on 23 September 2013 on http://pubs.rsc.org |

Table 1.4

.
Published on 23 September 2013 on http://pubs.rsc.org |

Glutamatc
(buer)
CEA (human
serum)
CEA (human
serum)
CEA (human
serum)
CEA (human
serum)
Nitric oxide
(buer)
Nitric oxide
(buer)
Nitric oxide
(buer)

CNT enzyme-based biosensor

0.2250 mM

0.01 mMa

433 mA/mM.cm2

Amperometric
biosensor
Amperometric
biosensor
Amperomctric
biosensor
Amperometric
biosensor
Amperometric
biosensor

enzyme-based

0.5 mM500 mM

0.01 mMa

enzyme-based

0.25 mM

2.0 mMa

enzyme-based

10 mM1.5 mM

5 mM

100  10 mA/
M.cm2
80  10 nA/
mM.cm2
88.8 nA/mM

enzyme-based

5 mM0.5 mM

5 mM

0.62 mA/mM.cm2

enzyme-based

20 mM0.75 mM

20 mM

CNT enzyme-based biosensor

25 mM

25 mM

Amperomctric immunosensor

0.55.0 ng/mL

0.2 ng/mLa

SPR immunosensor
Potentiometric immunosensor

4.485.7 ng/mL

n5
y 88.832x
r2 0.9945; n 3
r2 0.9968
y 5.0145x 0.1638
r2 0.9950; n 3
CV(%) 9.6
y 1.29x 0.088

0.5 ng/mL

S/N 2.7

1.2 ng/mLc

y 24.36 log x  10.37

r 0.997; n 5
CV(%) 2.2
r2 0.9975

1.73 U/mLa

Amperometric CNT-based
sensor
Hemoglobin-based
electrochemical biosensor
Hemoglobin-based
amperometric biosensor

0.2150 mM

0.08 pM

150 mM

0.003 pMa

0.011.0 mM

5.0 pMa

0.04 nM5.0 mM

20 pMa

0.0252.5 mg/mL
(28.5 pM2.9 nM)

25.0 ng/mLd

3.2  0.1 mL/U

0.0424 mA/mM

y 0.0424x 2.441
r 0.9978
y 1.356x 1.168102
r2 0.9989; n 8
CV(%) 4.2
y 0.2108x 72.31
r2 0.989

25

Electrochemical immunosensor 214 U/mL

Nitric oxide
Hemoglobin-based
(buer)
amperometric biosensor
C-Reactive
Magnetic immunosensor
Protein (human
blood)

1.29  0.18 nA/

mM.cm2

r2 0.9984
CV(%) 4.8
r2 0.991

Biosensor Technology and the Clinical Biochemistry Laboratory

Glutamate
(buer)
Glutamate
(buer)
Glutamate
(buer)
Glutamate
(buer)
Glutamate
(buer)
Glutamate
(buer)

.
Published on 23 September 2013 on http://pubs.rsc.org |

26

Table 1.4

(Continued)

Analyte detected
(matrix)
C-Reactive
Protein (human
blood)
C-Reactive
Protein (human
blood)
C-Reactive
Protein (human
blood)
Catecholamines:
DA, EPI, NEPI
(buer)
Catecholamines:
DA, EPI, NEPI
(human urine
and plasma)
EPI (buer)
NEPI (buer)
DA (buer)
DA (buer)

Linear range

SPR Immunosensor

25 mg/mL
(2.35.75 nM)

Capacitive immunosensor

0.0251 mg/mL

SPR immunosensor

525 mg/mL

Amperometric enzyme-based
biosensor

DA: up to 40 mM
EPI: up to 55 mM
NEPI: up to 55 mM
5125 pg/mL

Optical ber enzyme-based

biosensor
Fiber-optic enzyme-based
biosensor
DNA-based biosensor
Electrochemical enzyme-based
CNT biosensor
Amperometric enzyme-based
biosensor
CNT-ISFET sensor
Amperometric enzyme-based
biosensor
Carbon ber-based biosensor

LOD
1 mg/mL

Sensitivity (slope) Additional information

e

y 10.675x16.995
r2 0.9867
y 2.78x 1.29
r2 0.9255
CV(%) 1.72

0.2 mMa
75 nA/mM
0.4 mMa
45 nA/mM
0.3 mMa
60 nA/mM
DA: 2.1 pg/mLc
EPI: 2.6 pg/mLc
NEPI: 3.4 pg/
mLc

y  0.016533x  1.30882
r2 0.988
y 0.85x 8.87; r2 0.993
y 0.61x  1.19

0.20.9 mM
0.580 mM
1.030.0 mM

CV(%) 5.3
CV(%) 6.0
CV(%) 6.1
y 0.344x 0.4; r2 0.9998
y 0.140x 0.04; r2 0.9996
y 0.252x 0.38; r2 0.9997

5 nM
400 nMa

68.6 mA/mM.cm2 r2 0.999

5120 mM
1010 000 mM

0.01 mM

11500 mM

0.6 mMa

1100 mM

1 mMf

378.2 mA/decade

r 0.998

Chapter 1

Acetylcholine
(buer)
Acetylcholine
(buer)
Acetylcholine
(buer)

Sensor type

Human ferritin
(serum)
Choline (buer)
Diphtheria
antigen
(buer)
hCG (human
serum)
hCG (human
serum and
urine)
Insulin (serum)
Vibrio cholerae
(buer)

.
Published on 23 September 2013 on http://pubs.rsc.org |

Bulk acoustic wave impedance 2103310 7

biosensor
cells/mL

2103 cells/mL

log y 11.23457  0.00763x

r2 0.9458; n 5

QCM immunosensor

105108 cells/mL

105 cells/mL

QCM sensor

102107 cfu/mL

102 cfu/mL

QCM immunosensor

0.1100 ng/mL

SPR immunosensor

0.2200 ng/mL

Carbon ber-based biosensor

Potentiometric
CNT-based
immunosensor
Electrochemical CNT-based
immunosensor
QCM immunosensor

1100 mM
241280 ng/mL

1 mMf
7.8 ng/mL

0.55.0 mlU/mL

0.3 mlU/mLc

2.5500 mlU/mL

2.5 mlU/mLa

CV(%)o5.0

6200 ng/mL
31053109
cells/mL

6 ng/mL
105 cells/mL

r 0.9984

SPR immunosensor
SPR immunosensor

y  2.90x  64.29
r2 0.997

5.38  0.10
mlU/mL

y 89.52 log x 117.68

r2 0.9944
CV(%): 4.9720.4
y 152.13 log x 168.16
r2 0.9997
r2 0.9837
y 74 log x  98.6
r 0.9978; n 13
CV(%) 2.3
r 0.9998

27

CNT, Carbon nanotube; CEA, Carcinoembryonic antigen; DA, Dopamine; EPI, Epinephrine; LOD, Limit of detection; NEPI, Norepinephrine; QCM, Quartz crystal
microbalance; SPR, Surface plasmon resonance; hCG, Human chorionic gonadotropin.
a
LOD is three times the signal-to-noise ratio.
b
LOD is 3s/k with s the standard deviation of 11 determinations and k the slope of calibration plot.
c
LOD is three times the residual standard deviation.
d
LOD is 10 times the background noise.
e
LOD is higher than four times the background (0.5 AU, Arbitrary Units).
f
LOD is approximately twice the noise level.

Biosensor Technology and the Clinical Biochemistry Laboratory

Mycobacterium
tuberculosis
(simulation
sample)
Mycobacterium
tuberculosis
(saliva)
Mycobacterium
tuberculosis
(buer)
Human ferritin
(serum)

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

28

Figure 1.14

Chapter 1

The recurring problem in biosensor technology: the non-specic adsorption (NSA) of adversary species interfering with the specic target
analyte response.

Admittedly, the conventional enzyme-linked immunosorbent assay (ELISA)

which displays remarkable sensitivity down to the low pM range64 but remains a
non-domestic test to be performed in a laboratory setting still is promised a
bright future.
To reliably alleviate the detrimental eect of NSA, it is necessary to better
understand the complex mechanism(s) and dynamics of interaction involved
once the biosensing surface comes into contact with the biological environment
and the dierent types of potentially interfering species present within (e.g.
proteins). In essence, the process of non-specic protein adsorption is quite
complex and governed by several dierent factors, which includes: (i) the nature
of the proteins contained in the biological uid notably with respect to their
structure (e.g. globular), size (molecular weight) or distribution of charge/
polarity (isoelectric point); (ii) the physicochemical properties of the biosensing
surface (e.g. charge, topography and morphology, surface energy); and (iii) the
environmental conditions (i.e. pH, ionic strength and temperature).65 To add to
the complexity of the situation, proteins are exible entities that can assume a
variety of dierent conformational states, which results in the possibility for
them to unfold and adopt the adequate geometry to t the underlying surface,
decrease the energy of interaction and irreversibly adsorb.66 Furthermore, the
energetics of adsorption itself is typically composed of diverse molecular forces
(hydrophobic/electrostatic interactions, hydrogen bonding) that interplay to
various degrees in a complex overall manner.65 In multicomponent samples,
proteins compete for surface binding sites, which triggers a series of collisions
and results in a cascade of adsorptiondesorption/exchange processes,66
governed by the aforementioned conformational phenomena that occur at
the more fundamental molecular level. Empirically, it has long been observed
that protein adsorption from mixtures follows a temporal pattern, wherein
higher mobility, more abundant proteins rst adsorb transiently before being
gradually replaced by less motile, scarcer ones with higher surface anity.66
This general, well-established phenomenon of sequential protein adsorption is
known as the Vroman eect.66

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

29

Undoubtedly, many factors aect the interactions of proteins with materials

(of which the NSA of adversary species to biosensing surfaces is a particular
case) and excellent reviews pertaining to both the theoretical and experimental eorts made to model/probe the complex molecular events at play can
be found elsewhere in the literature.66,67 Nevertheless, despite the mechanistic
complexity of (non-specic) protein adsorption, some general rules relating the
properties of proteins with their ability to adsorb to surfaces have been devised.
For instance, the larger a protein, the more likely it is to possess multiple adhesion sites and readily adsorb to surfaces.65 The same holds true when considering the exibility of a protein (taken here as its ability to unfold), especially
when (more) binding sites are exposed during conformational remodelling. This
can occur in particular when a hydrophobic pocket (may it be a large inner
protein core or simply a more limited surface domain) previously hidden from
the surrounding aqueous medium (wherein folded proteins reside fully hydrated) to minimize energetically unfavored polar/nonpolar interactions is
revealed upon adsorption.66
NSA is a particular case of a more general, ubiquitous phenomenon surface
fouling that spontaneously occurs when exogenous biomaterials are exposed
to biological environments. For many situations plagued by fouling that
extend well beyond the eld of biosensors tremendous eorts have been devoted over several decades to engineer antifouling surfaces, most traditionally
through the imposition of organic lms able to ecaciously resist protein adsorption.68 As will be discussed in some more detail in the following section,
numerous types of such coatings have been reported in countless literature
publications and ultralow fouling (o5 ng/cm2) has now been achieved for
biotechnologically irrelevant, single-protein buered solutions.68 Regrettably,
few coatings present/retain in actuality such a remarkable level of performance
when exposed to highly complex biological media such as blood serum or
plasma (even diluted), these being otherwise more demanding.68 Fortunately,
recent times have witnessed an increase in studies entirely focused on
minimizing or ideally eliminate entirely fouling for real-world biological
samples.68 Assuredly, this trend should benet the eld of biosensor technology
and help eradicate the NSA plague.

1.5 A Look at Surface Chemistries to Solve the NSA

Issue
In Section 1.1, we have seen that the actual surface presented by a given biosensor structure to a biological uid can take many forms. This will include
inorganic material such as silica or indium-tin oxide in addition to polymers
and organic linker molecules. It is a reality that the NSA or fouling discussed in
the previous section can occur on all these materials to varying degrees. Accordingly, it is clear that there will be no panacea in terms of avoiding or even
reducing the adsorption phenomenon. Although there has been a level of research aimed at the NSA issue with biosensors, it is certainly the case that most

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

30

Chapter 1

relevant work has been published on the study of materialbiological uid

interactions. This topic has recently been reviewed in detail by the present
authors.68 The majority of the research described therein deals with adsorption
of simple protein solutions and the like, with a much more modest eort devoted to biological uids. A brief scan at fouling follows with some emphasis on
our own eorts to avoid the phenomenon.
By far the majority of past related work describes the use of a great variety of
organic coatings such as peptides, polyethylene glycols (PEG), zwitterionic
sulfo- and carboxybetaines, methacrylates, acrylamide and biomimetic surface
chemistries.68 A vast literature on such coatings, often relatively thick in dimension, deals with the minimization, as mentioned above, of protein adsorption from simple buer solutions. Whether in the latter type of solutions or
more complex media, the PEG molecular family constitutes one of the most
studied, but despite the high level of interest in this system the mechanism that
lies behind the PEG eect remains obscure. One prevalent theory is that a
kosmotropic water barrier to biological macromolecule adsorption is instigated by the polymer.68
In our research, we have focused on attempts to combine antifouling behavior with crosslinkers that are capable of immobilizing biochemical probes.17
Covalently-bound adlayers based on silanization methodology, much as described in Section 1.2, have been employed in conjunction with the ultra-high
frequency (GHz) EMPAS sensor. This is in order to study the surface eects
caused when devices are subjected to a dispersion of goat serum.69 The adlayerforming molecules are custom-designed with several properties in mind. The
molecules contain the ether linkage as present in PEG; they form ultrathin lms
(B5-A thick) and are capable of merging with sister crosslinking molecules that
form similar surface covalent bonds. The latter contain a distal functionalizable
moiety that can be used to attach probes in a subsequent, preactivation-free
step.17 The behavior of MEG-OH was compared directly with related
molecules with respect to fouling by the components of serum (Figure 1.15).69
The remarkable result from this work is that a single ether oxygen atom in the
monoethylene moiety of MEG-OH very signicantly alters the surface behavior of the sensor in terms of interfacial adsorption from this matrix. On
a speculative level, we believe that an intercalated water-based structure
stabilized by the distal -OH groups is responsible for this eect. Further studies
are underway both from a fundamental point of view and also with employment of the antifouling eect in biosensor experiments.

1.6 A Final Comment

A glance into a clinical biochemistry laboratory in a major hospital such as St.
Michaels located in Toronto, Canada reveals groups of large instruments
connected together with sizeable trains composed of robotic equipment. Much
of this is orchestrated around batch-based assays and technology. The set-up is
very far removed from biochip, biosensor and lab-on-a-chip devices, all of
which have been touted for potential application in clinical assays. As outlined

View Online

31

Biosensor Technology and the Clinical Biochemistry Laboratory

F

F
O

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Si
O

MEG-TFA

Figure 1.15

Si
O

MEG-OMe

Si
O

MEG-OH

Si
O

OTS-OH

Si
O

OTS

Modication of quartz with various structurally related organosilane

surface modiers.
(Reprinted with kind permission of the Royal Society of Chemistry.)

herein, the main reasons for this appear to lie in speed of analysis, convenience
and saving with respect to reagent cost, etc. The latter appears to be a non-issue
since such cost constitutes a minor component for the laboratory budget. In
view of the requirements for the clinical laboratory, one may wonder how
biosensor technology could realistically contribute to the operation of this sort
of facility. In this respect, biosensor devices can function particularly well in a
ow-injection scenario, especially those systems based on label-free detection
involving acoustic wave and SPR physics. This approach can lead to the
avoidance of batch measurements and lengthy trains of machines. Another
advantage would be the possibility to perform generic detection, that is, the
sensing of dierent targets with the same aforementioned physics/technologies.
However, there exist two main problems that remain to be solved. The rst
would be to render a particular sensing platform reusable in the FIA conguration. A second would be to nd a solution to the all-prevalent fouling
problem encountered with biological uids as discussed in some detail above.
Lastly, we would like to mention a nal word about microuidics and labon-a-chip devices, which appear to be ideal for combination with biosensor
detection. These technologies seem to be more suited for point-of-care systems
rather than the central clinical laboratory. However, despite the frequently
emphasized promises oered by these devices, there is no doubt that NSA still
remains a serious obstacle to be tackled. For example, NSA will not only occur
at the detector employed in the device but also in channels of the microuidic
set-up.70

Acknowledgements
The authors are grateful to the Collaborative Health Research Program of
BSERC and CIGR for support of their research. SS thanks the Province of
Ontario for the award of an Ontario Graduate Scholarship.

View Online

32

Chapter 1

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

References
1. C. P. Price, A. St. John and J. M. Hicks (ed.), Point-of-Care Testing, 2nd
edn. American Association for Clinical Chemistry Press, Washington DC,
2004.
2. Source: St. Michaels Hospital, Toronto, Canada.
3. K. E. Herold and A. Rasooly (ed.), Lab-on-a-Chip Technology Vol. 1:
Fabrication and Microuidics, Caister Academic Press, Norfolk, UK,
2009.
4. See Lab-on-a-Chip, Published by the Royal Society of Chemistry,
Cambridge, UK.
5. H. Kaiser, Pure. Appl. Chem., 1973, 34, 35.
6. V. Dugas, A. Elaissari and Y. Chevalier, Recognition Receptors in Biosensor M. Zourob (ed.), Springer-Verlag, Berlin, Heidelberg Publisher,
2010, pp 47134.
7. D. C. Kim and D. J. Kang, Sensors, 2008, 8, 6605.
8. E. A. Bayer and M. Wilchek (ed.), Avidin-Biotin Technology Vol. 184,
Academic Press Inc., San Diego, CA, 1990.
9. G. P. Anderson, M. A. Jacoby, F. S. Ligler and K. D. King, Biosens.
Bioelectron., 1997, 12, 329.
10. R. Gupta and N. K. Chaudhury, Biosens. Bioelectron., 2007, 22, 2387.
11. R. A. Williams and H. W. Blanch, Biosens. Bioelectron., 1994, 9, 159.
12. K. Jans, K. Bonroy, G. Reekmans, R. De Parma, S. Peeters, H. Jans, T.
Stakenborg, F. Frederix and W. Laureyn, Sensors for Environment, Health
and Security M.-I. Baraton (ed.), Springer-Verlag, Berlin, Heidelberg
Publisher, 2009, pp 277298.
13. A. Ulman, An Introduction to Ultrathin Organic Films From LangmuirBlodgett to Self-Assembly, Academic Press Inc., San Diego, CA, 1991.
14. J. C. Love, L. A. Estro, J. K. Kriebel, R. G. Nuzzo and G. M. Whitesides,
Chem. Rev., 2005, 105, 1103.
15. C. Vericat, M. E. Vela, G. Benitez, P. Carro and R. C. Salvarezza, Chem.
Soc. Rev., 2010, 39, 1805.
16. H. Hakkinen, Nature Chem., 2012, 4, 443.
17. S. Sheikh, J. C.-C. Sheng, C. Blaszykowski and M. Thompson, Chem. Sci.,
2010, 1, 271.
18. C. Blaszykowski, S. Sheikh, P. Benvenuto and M. Thompson, Langmuir,
2012, 28, 2318.
19. C. Blaszykowski, L.-E. Cheran and M. Thompson, Can. J. Chem., 2011,
89, 1512.
20. M. Thompson, L.-E. Cheran and S. Sadeghi. Sensor Technology in
Neuroscience, Royal Society of Chemistry, Cambridge, UK, 2013.
21. N. J. Ronkainen, H. B. Halsall and W. B. Heineman, Chem. Soc. Rev.,
2010, 39, 1747.
22. J. Wang, Biosens. Bioelectron., 2006, 21, 1887.
23. J. Korita and K. Stulik, Ion Selective Electrodes, 2nd edn. Cambridge
University Press, 1983.

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

Biosensor Technology and the Clinical Biochemistry Laboratory

33

24. C. H. Fry and S. E. N. Langley, Ion-Selective Electrodes for Biological

Systems, CRC Press, 2002.
25. P. Bergveld, Biosensors, 1986, 2, 15.
26. P. Bergveld, Sens. Actuators B-Chem., 2003, 88, 1.
27. M. J. Schoning and A. Poghossian, Analyst, 2002, 127, 1137.
28. J. Wang, Electroanal., 2001, 13, 983.
29. J. Wang, Chem. Rev., 2008, 108, 814.
30. J. Wang, Perspectives in Bioanalysis Vol. 1, Elsevier, 2005, pp 175194.
31. P. De Marco, A. Ng and D. Panduwinata, Electroanalysis, 2008, 20, 313.
32. W. G. Cady, Piezoelectricity Vol. 1, Dover Publications, New York,
1964, p. 4.
33. G. Sauerbrey, Z. Phys., 1959, 155, 206.
34. X. Wang, J. S. Ellis, P. Sundaram, E.-L. Moore and M. Thompson, Molec.
Biosys., 2006, 2, 184.
35. D. H. Ather and V. Relpa, J. Biophys. Chem., 2012, 3, 211.
36. H. Su, K. M. R. Kallury, M. Thompson and A. Roach, Anal. Chem., 1994,
66, 769.
37. X. Wang, J. S. Ellis, C.-D. Kan, R.-K. Li and M. Thompson, Analyst,
2008, 133, 85.
38. S. M. Ballantyne and M. Thompson, Analyst, 2004, 129, 219.
39. M. Thompson, S. M. Ballantyne, L.-E. Cheran, A. C. Stevenson and
C. R. Lowe, Analyst, 2003, 128, 1048.
40. M. Thompson and D. C. Stone, Surface-Launched Acoustic Wave Sensors:
Chemical Sensing and Thin Film Characterization, Wiley & Sons Inc., 1997.
41. K. Lange, B. R. Rapp and M. Rapp, Anal. Bioanal. Chem., 2008,
391, 1509.
42. X. Fan, I. M. White, S. J. Shopova, H. Zhu, J. D. Suter and Y. Sun, Anal.
Chim. Acta, 2000, 620, 8.
43. B. T. Cunningham, Label-Free Biosensors Techniques and Applications
M. A. Cooper (ed.), Cambridge University Press, 2008.
44. Y. Li, Y. Fan, L. Zhang, S. Ma and J. Zheng, Proc. SPIE, 2000, 4077, 242.
45. P. A. E. Piunno, U. J. Krull, R. H. E. Hudson, M. J. Damha and
H. Cohen, Anal. Chem., 1995, 67, 1635.
46. A. M. Valasex, C. A. Lana, S. Tu, M. T. Morgan and A. S. K. Bhunia,
Sensors, 2009, 9, 5819.
47. E. Kretschmann, Z. Phys, 1971, 241, 313.
48. J. Homola, Surface Plasmon Resonance Based Sensors, Springer-Verlag,
Berlin, Heidelberg Publisher, 2006.
49. J. Homola, S. S. Yee and G. Gauglitz, Sens. Actuators B-Chem, 1999, 4, 3.
50. X. D. Hoa, A. G. Kirk and M. Tabrizian, Biosens. Bioelectron., 2007,
3, 151.
51. R. L. Rich and D. G. Miszka, Curr. Opin. Biotech., 2000, 11, 54.
52. X. Guo, J. Biophoton., 2012, 5, 483.
53. J. Homola, Anal. Bioanal. Chem., 2003, 377, 528.
54. R. G. Heideman, R. P. H. Kooyman and J. Greve, Sens. Actuators BChem., 1993, 10, 209.

View Online

.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00001

34

Chapter 1

55. A. Brandenburg, Sens. Actuators B-Chem., 1997, 39, 266.

56. B. H. Schneider, E. L. Dickinson, M. D. Vach, J. V. Hoijer and
L. V. Howard, Biosens. Bioelectron., 2000, 15, 13.
57. M. L. Turgeon, J. J. Linnae and K. M. Ringsrud, Linne & Ringsruds
Clinical Laboratory Science: the Basics and Routine Techniques, 5th edn.
M. L. Turgeon (ed.), Mosby Elsevier, 2007.
58. C. I. L. Justino, T. A. Rocha-Santos and A. C. Duarte, Trends Anal.
Chem., 2010, 29, 1172.
59. J. N. Adkins, S. M. Varnum, K. J. Auberry, R. J. Moore, N. H. Angell,
R. D. Smith, D. L. Springer and J. G. Pounds, Mol. Cell. Proteomics, 2002,
1, 947.
60. M. J. A. Shiddiky, P. H. Kithva, D. Kozak and M. Trau, Biosens. Bioelectron., 2012, 38, 132.
61. J. Tamayo, P. M. Kosaka, J. J. Ruz, A. S. Paulo and M. Calleja, Chem.
Soc. Rev., 2013, 42, 1287.
62. J. Breault-Turcot and J.-F. Masson, Anal. Bioanal. Chem., 2012, 403, 1477.
63. S. Choi and J. Chae, J. Micromech. Microeng., 2010, 20, 1.
64. D. A. Giljohann and C. A. Mirkin, Nature, 2009, 462, 461.
65. K. C. Dee, D. A. Puleo and R. Bizios. An Introduction to Tissue-Biomaterial Interactions, Wiley & Sons Inc., Hoboken, NJ, 2002, pp 3751.
66. R. A. Latour, Jr., Encyclopedia of Biomaterials and Biomedical Engineering,
Taylor & Francis, 2005, pp 115 and references therein.
67. J. J. Gray, Curr. Opin. Struc. Biol., 2004, 14, 110.
68. C. Blaszykowski, S. Sheikh and M. Thompson, Chem. Soc. Rev., 2012, 41,
5599 and references therein.
69. S. Sheikh, D. Y. Yang, C. Blaszykowski and M. Thompson, Chem.
Commun., 2012, 48, 1305.
70. R. Mukhopadhyay, Anal. Chem., 2005, 77, 429A.

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

CHAPTER 2

Integrated Chemistries for

Analytical Simplication
and Point of Care Testing
PANKAJ VADGAMA,* SALZITSA ANASTASOVA AND
ANNA SPEHAR-DELEZE
Queen Mary University of London, Mile End Road, E14N, London,
United Kingdom
*Email: [email protected]

2.1 Introduction
In an era of individualised therapy and patient empowerment, it is a natural
progression for assays that were once the sole preserve of the central laboratory
to now fall within reach the of the nonlaboratory professional or the patient
with assays typically undertaken out with the laboratory. The area of extralaboratory testing has, accordingly, seen major expansion in recent years, but
has also become the subject of considerable debate as regards its true value to
the patient. It is certainly a more demanding and expensive way of undertaking
tests, but the champions of such testing immediacy are clear that a wide
range of benets accrue. These include a more rapid return of data to the
clinician to institute earlier therapy, reduced hospital stay, reduced medical
complication rates, especially in chronic disease such as diabetes, and an overall
improvement in quality of life at the population level.1
This form of decentralised testing, more commonly referred to as pointof-care testing (POCT), largely remains under the aegis of laboratory medicine.
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

35

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

36

Chapter 2

Moreover, through technological advances, it is evolving rapidly both in analytical scope and the diversity of clinical application; monitoring of both ambulatory and nonambulatory patients is now readily feasible. POCT
applications range from basic blood glucose measurement to complex enzyme
assays. The organisational benets include the elimination of specimen transport and sample preparation and certainty over when the assay data is available.2 POCT may be undertaken using systems as contrasting as single-use
hand-held devices and bench-top analysers, the latter being more allied to
traditional laboratory analysers. Whilst the growth of POCT has had considerable impact on how monitoring is undertaken, it cannot replace clinical
judgment. It is best used when a clinician is in a position to respond rapidly to
test results.3
The wider patient pathway and organisational issues fall outside the scope of
this review. However, the ability to roll out a measurement capability to the
POCT setting in the rst place, especially for complex assays, relies heavily on
the harnessing of new technologies. These especially involve the integration of
biochemical reagents as immobilised components of measuring platforms. Such
platforms might be sensor based, creating a combined, monolithic, bio-/
articial structure, or involve an independent biological phase and detector.
Overall, quite disparate technologies need to be functionally integrated into
bioengineered, ergonomically simple, constructs.4 This review covers design
issues for the various subcomponents and how these can satisfy the needs of
selective, sensitive measurement, excluding in vivo monitoring. The broad
topics are: routes to sample presentation using controlled ow, the way multiple uid movements are achieved, e.g. in porous structures using lateral ow,
how biochemical reactive and recognition systems are congured and how this
biochemically mediated response triggers the measurement pathway. In the
quest for operationally simple and miniaturised point of care systems, there
has been considerable leveraging of bioreagent immobilisation chemistries
and MEMS (microelectromechanical systems) technologies as well as a
re-engineering of optical and electrochemical transducers. Along with this have
been leading-edge developments in nanotechnology, to achieve extreme miniaturisation,5 and better sample interfacing materials to control what actually
reaches the measuring element of the analytical system. The latter is essentially
a form of packaging to protect the vulnerable functional elements.6
Typically, whole blood is used for analysis, and it may be necessary to use
barrier layers to screen out the formed elements of blood for say an optical or
colorimetric assay.

2.2 Fluidics for POCT

More complex assays for POCT, or where bench-top analysers are used, necessitate some form of sample and/or reagent ow in a uidic channel. The
adaptation of traditional uidics has been particularly evident in the case of
multiparameter blood gas and electrolyte systems. This contrasts with singleparameter reagent strips such as for urine albumin and blood glucose that

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

37

require only the application of sample to the measuring surface. Bench-top

analysers, as smaller versions of laboratory analysers, simplify vulnerable
operator-dependent steps, mainly through automation.
MEMS may be used to achieve ow miniaturisation and sample conservation, and though microuidics dimensions are not reached, other than in
experimental systems, many of the logistics of uid handling are now established. In the Alere Triages System, which comprises a meter and various test
panels, e.g. for cardiac marker measurement (CK-MB, myoglobin, TroponinI),7,8 an immunoassay reaction mixture, after controlled incubation, ows over
dierent capture surfaces for selective assay. The i-STAT meter for blood gases,
ions and metabolites has individualised uidics in self-contained cartridges that
include of the i-STAT calibrants; in one interesting veterinary use, synovial
uid was assayed for pH, glucose and lactate.9 In a system where sample selfcontainment is paramount, the Pima CD41 analyser for HIV testing, the
sample is taken up by an integrated capillary with all subsequent immunobinding and uidic movement occurring within a cartridge.10 High-number
parallel assays would be dicult to integrate with a uidic system. In the
Evidence microchip immunoassay array, more than 20 anthelmintic drugs
could be screened simultaneously, but each component constituted a single well
into which preset reagents were added as the route to a simplied assay.11
Advances in printing, lamination and micromoulding technology have enabled greater miniaturisation along with the creation of more complex and
versatile ow conduits that can carry sample and carrier phases to a detector
phase in a time-controlled manner. In one experimental conguration, a
microuidic circuit was formulated with an open ow component serving as
sampler for precision sample uptake and delivery.12 However, the majority of
ow devices utilise fully closed ow congurations. As well as sample conservation, an additional benet of microuidics is the generation of laminar ow
that can allow formation of stable liquid/liquid interfaces within a single
channel with only diusion and convection determining solute exchanges. This
is an ideal way of transferring an analyte from a sample stream to a mobile
recipient phase. The principle of this is shown in Figure 2.113 where stable
parallel ows are seen. Where the liquids ow together, solute exchange occurs
at a rate determined by molecular weight, aiding solute selection; indeed this
can be extrapolated to the separation of colloidal and cellular elements from
the sample.
An immunoassay has been demonstrated where drug binding to a selective
antibody retards its transport across the liquid/liquid interface thereby providing a measure of drug concentration. The assay principle here is that of
competitive binding between sample and labelled drug with optical tracking.14
The principle of size-based separation has been further developed in a protein
separation assay for salivary MMP (matrix metalloproteinase) (Figure 2.2).15
Here, the target protein binds to a uorescently labelled antibody, and separation by size-based selectivity achieved using a porous gel and a size exclusion
membrane; sample was moved electrophoretically rather than by pressure
driven ow. We have developed in situ formed protein based membranes at

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

38

Chapter 2

Figure 2.1

Laminar ow in an H-Filters system (Micronics). Sample, e.g. whole

blood, is introduced from the bottom left and ows in parallel with a
receiver solution, e.g. buer, introduced through top left. Smaller sample
molecules diuse from the sample to the receiver stream to be harvested
from the top right outlet, while bigger molecules and particles are harvested from bottom right.
(Reproduced from ref. 13 with permission.)

Figure 2.2

Microchip electrophoretic immunoassay (mCEI) device for the detection

of MMP-8 from saliva samples. S-sample, B-buer, SW-sample waste,
BW-buer waste, mAb*-uorescently labelled monoclonal antibody to
MMP-8. Polyacrylamide gel composition variation is indicated by
grayscale shading (%T and %C are percentage of total acrylamide and
bis-acrylamide crosslinker, respectively). Inset shows 40 size-exclusion
membrane bright-eld image, scale bar 100 mm.
(Reproduced from ref. 15 with permission.)

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

39

parallel uid ow interfaces using a rapid crosslinking with terephthaloyl

chloride that then established a stable barrier between parallel streams;16 such
an arrangement holds the prospects for pore size-based control of lateral solute
transport as well as serving as a support for bioreagent immobilisation.
A practical consideration with uid ows is how to generate the initial ow.
Precision pumps have been developed with appropriate miniaturisation, however, even with traditional pumps, complex patterns of microow can be generated. Versatility of ow design was demonstrated in one example where
agglutinated red cells were detected during blood crossmatching; here, agglutinated cells reduced light transmission across a uidic channel as compared
with non-agglutinated cells.17 Micropumps can have drawbacks of reliability as
well as the added need for powering and machining; passive pumps may oer a
simpler alternative. In one example, an array of hexagonal microfabricated
pillars with dened spacing achieved controlled uid lling giving pumpless
ow.18 The system had minimal internal dead space and ow resistance and
allowed stable ow during a one-step immunoassay. This particular ow
conguration also incorporated various ow constraints and delay elements
coupled with sample collection. With integration of heterogeneous components
in this way, it becomes feasible to build lab-on-a-chip capabilities to achieve
sequential sampling, separation, dilution, reagent mixing/incubation and
readout steps that go to make up a typical clinical assay procedure for a
laboratory analyser.
Particularly at low cross section channels, as exist in microuidic systems,
the lack of surface compatibility of cell- and protein-loaded uids with
standard biomaterials, polymers and inorganics such as silicon, remains a
problem. One way to mitigate this is to coat or surface modify the materials
with known colloid tolerant interfaces such as those based on poly(ethylene
oxide) or biomimetic phosphoryl choline. An alternative has been the use of
microdroplet ow where the relevant biocomponent is retained within a containment phase surrounded by an oil or air continuum.19 The discretisation of
both reagent and sample in this way can also reduce volume requirements and
allow manipulative steps with reduced need for complex channels and valves.
Electric eld generation at insulator covered electrode arrays can allow
microdroplet transport to be readily controlled (electrowetting). By this means,
also, necessary microdroplet division and fusion can be instituted to allow for
interactive reactive processes. An ambitious target reported for this approach
was multiparameter newborn screening.20 A dedicated laboratory for this requires to undertake immunoassays, enzyme assays and even DNA identication, but a droplet-manipulating platform capable of dispensing and washing
droplets after binding/reactive steps appears to show promise with multiple
model targets. Sample volume requirements here dropped to B1% and high
throughput assays for screening, e.g. for hypothyroidism and G-6-PD deciency
looked feasible It may be possible through innovations in materials for microuidics to be able to customise and simplify production; using hot embossed cyclic
polyolen Bhattacharrya and Klapperich21 were able to simplify microchannel
fabrication for the immunoassay of C-reactive protein, an inammation marker.

View Online

40

Chapter 2

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

2.3 Lateral Flow Techniques From Simple

Colorimetric Strips to 3D Flow
2.3.1 Colorimetric Strips
Lateral ow provides the least technologically complex and demanding platform for the qualitative or semiquantitative analysis of an analyte. It has its
origins in paper-based regent strips where an adsorbed reagent enables
single-use colorimetric measurement, typically by just visual inspection and
comparison with a colour chart In the production process, the reagent is deposited on adsorbent paper later split into square pads, in the case of urine
testing providing test panels on a single dipstick. The core assays here were for
urine albumin, ketones, white cells, specic gravity and nitrite. Though simple
to use, they did depend in some instances on reaction timing, and of course the
visual discrimination of the observer to assess the sometimes marginal colour
changes. A switch to instrumentation with use of light-emitting diodes (LEDs),
charge-coupled device (CCD) arrays and CCD cameras has allowed for greater
objectivity and the avoidance of observer error. Later renements included the
use of enzymes, glucose oxidase being the classic example, and the mitigation of
interference through added reagent, e.g. to scavenge or identify an interferent in
the sample.22 The transition from urine to blood chemistries utilised traditional
regents, but in addition to the basic functional elements of a uid sample uptake layer, a reagent layer and a light-reecting layer,23 there now appeared
masking, trapping and separation layers. This was especially necessary given
the extension to enzyme-based assay (e.g. for glucose, urea) and to the assay of
enzymes (e.g. creatine kinase, amylase).
Quite sophisticated layer formatting has been possible as evidenced by an early
phenytoin immunoassay.24 In this instance, antibody to phenytoin constituted a
dry reagent anity component. Labelled and sample phenytoin competed for
binding to the antibody; the label used was FAD, and only as a part of unbound
phenytoin was it able to attach as a prosthetic group to activate glucose oxidase
apoprotein. Resulting holoenzyme activity was then measured in the strip by a
conventional colorimetric indicator reaction. An innovative use of a nonpolar
phase coexisting with a polar (sample) phase involved use of the ionophore
valinomycin as a K1 binder within the nonpolar layer.25 The valinomycin was
associated within the apolar phase with a pH indicator dye; chelation of K1 by
valinomycin led to its becoming cationic, with ejection of H1 into the polar phase
from the indicator dye and a resulting dye colour change.
Though not a chromogen, the lanthanide ion terbium(III) forms a usable
luminescent complex with p-aminobenzoic acid (PABA). This was exploited in
a paper-strip format to assay the local anaesthetics benzocaine and procaine;
their hydrolysis in alkali generated the necessary PABA end product leading to
a luminescent chelate. The reagent strip is more stable than one using bio/organic reagents, but does require special luminescence detection apparatus.
Albumin dye binding and the resulting colour change has been the mainstay
of urine albumin measurement by a dry reagent pad. However, there is a

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

41

serious sensitivity limit that precludes detection in the microalbuminuria

range relevant to detecting and monitoring diabetic nephropathy.
A tetrabromosulfonphthalein dye (DIDNTB) was found to have a specially
high anity for albumin26 and allowed its detection in urine down to 10 mg/L;
the constituted reagent strip maintained an optimum pH for the dye using
impregnated citrate buer. A counterpart to the high albumin anity was the
reduced binding competition from IgG/IgA and immunoglobulin light chains
found in urine. A non-chemistry strip, but one that is highly successful is that
used to monitor anticoagulant treatment. In a recent commercial embodiment,
the CoaguCheckXS,27 the test strip takes up venous or capillary blood, hydrating a thromboplastin layer that generates a blood-coagulation sequence to
generate thrombin from the sample. The thrombin as indicator of the prothrombin time then cleaves a peptide substrate; here an electrochemical rather
than an optical end point was monitored.
In contrast to the vertical transmission of sample through a porous phase,
lateral ow devices involve horizontal ow and have the facility for topographic
resolution of bound vs free analyte, and provided a visible label is used, allow for
instrument free detection.28 Their primary use has been in the detection of
proteins using antibody binding. The rst and foremost use has been detection of
urine human chorionic gonadotrophin (hCG) for pregnancy testing. Essentially,
two epitopes are recognised on a protein by preloaded antibodies. After the liquid sample is applied to a loading pad, it moves laterally within the porous
support and is exposed to these antibodies in turn. An absorbent pad at the
opposite end of the analytical strip draws up the liquid sample to drive the
controlled lateral ow. First, the target protein binds to labelled antibody, and
the two move onto an immobilised capture antibody zone that in turn immobilises the antigen labelled antibody complex to generate a visible, concentrated
zone. Visibility is generally achieved by particulate labelling of the rst antibody,
e.g. using gold nanoparticles or latex beads.

2.3.2 Nanoparticle Labels

Scaling of sizes, the availability of new materials, variation of surface proles
and novel bulk properties open up new assay avenues with nanoscale structures. In one example, magnetic nanoparticles were used as immobilisation
surfaces for antibody;29 In this model, it was shown that orientation of antibody was highly relevant to sensitivity; particle binding through the carbohydrate component of the antibody, away from the binding site, allowed for
freer access to hCG target antigen. Magnetic particle binding has also been
used for sample washing and separation of free from bound antigen.30 In this
example, a possible basis for a commercial device, cardiac troponin I was assayed
by rst mixing antibody-loaded particles with sample, then activated electromagnets were used to drive the particles to a second antibody-binding surface,
and nally diametrically positioned electromagnets separated unbound particles
from the antibody-bound layer (Figure 2.3). Frustrated total optical internal
reection nally enabled imaging of the antibody-bound particulate layer.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

42

Figure 2.3

Chapter 2

A schematic of magnetic-bead-based immunoassay: analyte binding by

antibody-functionalised magnetic nanoparticles with top and bottom
magnets o (left), nanoparticles bind to the sensor surface, lower magnet
on (middle) and magnetic removal of free and weakly bound nanoparticles, upper magnet on (right).
(Reproduced from ref. 30 with permission.)

Further labels include selenium, carbon, liposomes, phosphors and quantum

dots to add to the repertoire of traditional uorescent and bioluminescent dyes.
With limited antibody, it is possible to devise competitive immunoassays where
the labelled and unlabelled antigens compete for the antibody binding.31 In one
assay of serum albumin, a uorescent label was used, where competition was for
antibody immobilised on a strip. Increased analyte concentration reduced label
binding and therefore uorescent dye accumulation at the strip surface;32 here,
however, a uorescence reader was necessary for the assay, adding to instrumentation complexity. In order to detect multiple analytes simultaneously, it is
necessary to have individually distinguishable labels. Thus, to detect antibodies
to Hepatitis A, B and RSV, respectively, metallic nanoparticles capable of
surface-enhanced Raman scattering were dierentially functionalised to give
resolvable Raman signatures and activated by specic antibodies to dierent
targets;33 lateral ow to a single capture strip holding a second antibody for each
target then enabled selective, simultaneous quantication.
Carbon nanoparticles have proven to be versatile surfaces for label as well as
bioreagent binding, e.g. nucleic acids, antibodies and other proteins, potentially
conferring detection sensitivity greater than gold or latex particles.34 Rayev and
Shmagel35 preadsorbed anity protein and subsequently crosslinked this to
give a stabilised binding surface enabling the attachment of observable carbon
nanoparticles to topologically localised target biomolecules. Thus, by coating
with protein G, which has a general anity for immunoglobulin, it was possible
to identify antibodies from seropositive samples, where antibody had bound to
HIV antigen retained at a porous support. A lateral ow arrangement was also
demonstrated for hCG detection by pre-loading with antibody to the b-subunit
of hCG. A nucleic acid lateral ow assay has also been developed exploiting
this readily visible label. In one example, a Plasmodium falciparum DNA
template in PCR was labelled with primers bearing two tags, digoxigenin and
biotin;36 carbon nanoparticles loaded with avidin bound to the biotin end and

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

43

the digoxigenin end was captured by immobilised antibodies. Amplied nucleic

acid sequences have been detected using dye-loaded nanoscale liposomes; here
lateral ow of a target sequence led to its capture by immobilised complementary strand DNA; this was subsequently made visible by binding to a reporter sequence on the surface of a liposome37 and practical illustration has
been given, with RNA amplication, of the detection of dierent serotypes of
the Dengue fever virus.38 For these serotypes, the sensitivity achieved ranged
from the detection of 50 000 down to just 50 RNA molecules. Also, with an
optical readout, quantitation was possible. A further anity system was
demonstrated by Kulagina et al.39 who immobilised antimicrobial peptides
(AMPs) as capture agents to conrm the targeting of model pathogens.
Venezuelan equine encephalitis virus, vaccinia virus and Coxiella burnetti (an
intracellular bacterium) were successfully identied with an eectiveness that
proved even greater than with selective antibody. Further extension of sensitivity for this uorescence-based assay could be possible using nanoparticles.

2.3.3 New Opportunities

In addition to single planar ow, 3D multiphase ows have been explored by
Whitesides Group40 using paper/brous supports. These microuidic paper
analytical devices (mPADs), channel solutions horizontally as well as vertically
and could deliver to multiarray reagent pads. The principle of the system was
demonstrated for glucose and protein measurement.
With newer porous materials, which allow for reduced nonspecic binding and
greater control over the lateral ow rate, improved selectivity and concentration
resolution could eventually be achieved. In this way, the quantitative capabilities
of lateral ow will become manifest. The key question will be how far it will be
necessary to integrate optical or other quantitative readers into the newer systems. While a denitively quantitative measurement may be obtainable, this will
also add costs to any commercial device, reduce portability and detract from its
being a consumer item for home use. Unlike home meters for say glucose, where
quantitative measurement is imperative, in many lateral ow applications, the
usually macromolecular target need not be determined with exact quantitation.
What is much more important is the ability to dierentiate so-called normal from
abnormal, or the qualitative identication of antibodies to pathogens. Nevertheless, if very cheap or even disposable readers can be developed, it may become
possible to combine disposability with reliable quantitation, a further, important
outcome would be the avoidance of observer error.

2.4 Biological Recognition

2.4.1 Immobilisation Techniques
Biological modication of a device or the transformation of a soluble bioreagent into a solid phase has innumerable analytical benets. Such approaches

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

44

Chapter 2

have been the formal basis of biosensor construction as well as of standalone

bioreactor and bioanity phases. The biomodication approach has allowed
exploitation of the special binding selectivity and anity of enzymes, antibodies
and aptamers, respectively. However, the immobilisation of such recognition
systems needs careful tailoring; without this the bioanity molecule may either
not survive the retention process or be so poorly orientated at a surface that it is
unable to achieve recognition capability. In the case of readily diusible targets
such as enzyme substrates, orientation may not be an issue as the target molecule can reach the active site, but for macromolecular targets, it may not be
possible to have eective binding unless the binding subcomponent, e.g. the
Fab region of an antibody, has sucient exposure at the sample interface, and
is not buried in the subjacent solid phase. The standard repertoire of physical
adsorption, chemical crosslinking, covalent attachment and polymer entrapment remains the core means of immobilisation to this day, but of relevance to
this chapter, these have been adapted to diagnostic platforms. However, the
special needs of analysis have meant that there has been signicant renement
of the basic immobilisation repertoire.
Porous supports provide high surface area matrices for biomolecule
adsorption and retention. As such, they are attractive for direct adsorptive
deposition of biomolecules. Whilst leaching is a potential issue, for single-use
devices, an analysis can be completed well before such loss of the bioactive
moiety. The usual physisorption binding forces are in operation, notably
hydrophilic and hydrophobic interactions, van der Waals forces and ionic and
hydrogen bonding. The local inuence of these aggregated forces can be to
distort biomolecule shape and so render the biomolecule inactive, more of a
problem with enzymes than with antibodies. Kong and Hu41 have reviewed the
general strategies for biomolecule immobilisation on paper surfaces. They
highlight the fact that cellulose is slightly anionic and so binding tends to involve cationic domains on biomolecules. Inevitably, the solution environment
(pH, pI) modies the eectiveness of binding. Special interactions also play a
part, such as binding that is mediated through tyrosine residues to the edge
plane of cellulose crystals.42 Horseradish peroxidase was used as a general
adsorptive model by Di Risio and Yan43 who showed that with this globular
protein, net positive charge at near-neutral solution conditions promoted
adsorption, but there was also a further dominant eect, an enhancement of
adsorption at hydrophobic surfaces. These hydrophobic interaction forces,
however, also led to protein unravelling with loss of activity.
The attractive forces posed by a surface are, of course, not selective for the
bioreagent, but furnish a locus for any extraneous protein deposition. This
leads to masking of the measurement surface, sensitivity drift and loss of signal.
Surface-blocking agents such as milk and bovine serum albumin (BSA) are well
established in standard multiwell immunoassays for blocking nonspecic
binding and could be used in lateral ow assays, but there is also the possibility
of incorporating agents such as surfactants and BSA into the solution phase.
An immunoassay where a liquid crystal was the reporter system showed
reduced nonspecic binding through addition of Tween 20.44

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

45

Provided the biomolecule is not denatured on initial contact with a surface,

it is likely to be stabilised. Thus, for alkaline phosphatase and horseradish
peroxidase with adsorption on paper, storage at elevated temperatures,
even up to 90 1C, demonstrated retained activity for many hours compared
with activity loss in solution within minutes.45 Degradation dynamics
indicated two sequential, rst-order, reactions, suggesting two distinct pathways, the second, slower one dominated half-life. In practical terms, alkaline
phosphatase showed retention of 63% activity at 14 days storage at 23 1C. The
study also showed the resistance of enzyme to shear forces generated during
printing.
Inkjet printing of enzymes has been a route to high precision, rapid deposition, facilitating array formation. Computer-controlled printing, delivering
nanolitre volumes can avoid crossover and the process also allows the biomolecule to be in solution; these are attractive features for device miniaturisation. Commercial printers have been used, for example, to deposit to a
feature size of 100 micrometres with deposition of up to four dierent
solutions.46 Inkjet printing uses thermal or mechanical forces to drive solution
droplets from a nozzle, while e-jet printing uses electric-eld-gradient generation between the delivery nozzle and the substrate, driving out minute drops,
depending on nozzle size. Thus, it has been possible to deliver droplet sizes in
the 0.1 micrometre range;47 with streptavidin as a model protein, activity was
retained, based on the measured anity behaviour of biotin.
Antibodies are generally more robust than enzymes. A study with anti-rabbit
IgG adsorbed on cellulose paper47 showed that at 40 1C there was an activity
half-life of 10 days with no added degradative eect of humidity. Interestingly,
precoating the paper with a crosslinked polyamide-epichlorohydrin lm, conferring a cationic property to the cellulose, did not signicantly enhance the
loading of this anionic protein, nor did it improve thermal stability, possibly
because thick layers of protein, not in direct contact with the surface, were
involved. The implication of this study was that strip manufacture with hot-roll
processing is feasible. Loading of antibodies has been increased through use of
high porosity, high surface area structures. Porous silicon has the interesting
dual eect of providing a low fouling, highly hydrophilic surface at the molecular level, while creating a hydrophobic macrostructure due to its nanoporosity, thereby helping to retain polar solution during assay conditions.48 As
an alternative to a cavity, a high-curvature microsurface using particles can
enable antibody adsorption by simple electrostatic attraction; as an illustration
of this, sulfonate functionalised polystyrene microspheres have been used
and co-attachment of quantum dots as labels enabled design of a
uoroimmunoassay.49
Protein A binds to the Fc region of an antibody, and so preadsorbed protein
A has been frequently used to orientate antibodies, allowing for an outfacing
Fab region for optimum binding exposure. In one report, Protein A preadsorbed to nanogold enabled both a greater surface area for antibody
binding and the desired antibody orientation for a higher sensitivity immunoassay, targeted to haptoglobin.50 The impact of antibody orientation will

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

46

Figure 2.4

Chapter 2

Proposed mechanism for electrostatic, preadsorptive antibody orientation

at polystyrene due to ionisation under alkaline conditions: (a) random
orientation at pH 7.4 in the absence of charge repulsion, (b) antibody
reorientation due to charge repulsion between the Fab region and polystyrene surface at pH 9.5.
(Reproduced from ref. 54 with permission.)

depend both on the nature of the solid support and that of the antibody. Thus,
in a study of antibody and antibody fragments (Fab) using selective region
bridging via biotin to surface-immobilised streptavidin, orientation led to increased binding,51 with up to 610-fold greater binding capacity following the
orientation, depending on the surface used. PEG was the anchor point at the
support surfaces, and the length of this tether aected the binding sensitivity to
orientation, with longer PEG chains generating greater orientation dependence. Orientation eects partly worked through delivering a higher surface
packing density, further reinforced by high packing of streptavidin at the
surface. The additional step of creating Fab fragments may not justify the eort
unless a further chemistry is designed to give surface orientation. Thus, in a study
of Fab and whole antibody binding to an IgG target, intact antibody loading was
already greater than for the smaller Fab, unless the latter had been orientated.52
Regardless of whether covalent immobilisation or adsorption are the end
outcome of the antibody retention process, noncovalent (e.g. hydrophobic,
electrostatic) interactions at the interface are important in driving some of the
initial orientation.53 In a study of antibody immobilisation on polystyrene
beads,54 high pH was used to rst generate a preorientation step before
surface attachment. At high pH, mutual anionic repulsion retarded antibody
adsorption, giving time for reorientation to occur near the surface; at this early
stage the high anionic nature of the Fab region led to its facing away from the
surface (Figure 2.4). The more hydrophobic Fc region was then able to attach
to the surface by hydrophobically driven attachment. The stability of the attached antibody layer was maintained by steric restriction at the surface using
PEG as a co-immobilised packing layer.

View Online

Integrated Chemistries for Analytical Simplication and Point of Care Testing

47

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

2.4.2 Enzyme-Based Systems

The ability of substrate to permeate macroscale, reactive enzyme layers has
enabled the use of thick-lm techniques for their construction. The only real
limitation has been the signicant diusive distances for product/reactant that
limit the speed of response. However, even this is only of importance when a
steady-state output is needed as in the case of continuous monitoring devices.
For biosensors specically, one of the most popular routes to immobilisation
has been use of the homo-bifunctional reagent glutaraldehyde to crosslink the
enzyme to give a matrix that also incorporates an inert protein, typically albumin. Glutaraldehyde itself has a somewhat protean structure, with thirteen
dierent forms apparently present depending on solution conditions,55 and
though the exact reaction with a protein is not certain, linkage with amine
groups via the aldehyde function looks to be the most important and, moreover, involves an oligomeric form of the reagent, rather than the monomer. The
tendency has been to take an empirical approach to enzyme immobilisation
based on trial and error, and to generate enzyme membranes of the requisite
thickness and activity, but based on an unknown crosslinking density; this
becomes a more challenging approach to design when the enzyme in question is
unstable or vulnerable to the distorting eect of the crosslinking process. Unless a quencher is introduced, the crosslinking, moreover, may be dicult to
make reproducible or controllable. One general outcome is that at low enzyme
concentration, damaging intramolecular crosslinking is promoted, whereas the
preferred intermolecular crosslinking is only promoted at high enzyme concentration. One way in which the latter is facilitated, of course, is through the
inclusion of an inert protein.
Surface retention has been explored using glutaraldehyde bridging, but leads to
limited surface layers and insucient loading if only mono- or oligolayers are
formed. A recent example with glutaraldehyde involved an aminated surface,
where enzyme was preadsorbed, and subsequent crosslinking led to high loading as
well as bridging of enzyme to the surface.56 Enhanced stabilisation of enzyme after
crosslinking has also been achieved by initial precipitation of enzyme.57 By creating
insoluble aggregates of enzyme through desolvation and then crosslinking, higher
thermal and storage stability has been demonstrated. A further potential advantage of this crosslinked enzyme aggregate (CLEA) approach is that the source
enzyme does not have to be of high purity; this is aspect conferred by the aggregation step.58 The chemical crosslinking component is important for stabilisation,
not just as a means of avoiding dissolution, but in modulating structural stability.
Compared with solution enzyme, there are greater constraints to conformational
uctuations, and the ability to transition into nonactive conformers is thereby
diminished; ultimately it is a balance between intrinsic stabilisation energies within
the enzyme molecule vs the dynamics of thermodenaturation.59
Enzyme mobility and binding to dierent surfaces is aected both by
the nature of the surface and by the bridging molecule used, as was shown
in stability/loading studies of glucose oxidase, lactate oxidase and horseradish
peroxidase, respectively, on aminated glass and cellulosic surfaces.60

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

48

Chapter 2

Interestingly, bridging molecule structure, where an aromatic reagent was used,

had a greater inuence on stability as enzyme size increased, and in the case of
glucose oxidase there was a substantially eect on eective Km. Cellulose is
an attractive matrix for dry reagent chemistries, and the potential for
photopatterning on it was demonstrated by its functionalisation with an
azidobenzene bridge; the latter, photoactivated to a reactive nitrene, enabled
ready enzyme binding.61 Cellulose can also act as a host for other polymeric
supports, as shown with a preformed glycidyl methacrylate linkage that then
allow for epoxy group mediated binding to enzyme (urease) via amino side
groups.62 The general use of paper-based supports and their bioactivation
for diagnostics has been reviewed by Pelton.63 Now, with the advent of
microarrays for proteomics, high-density arrays are featured at a wide range of
polymeric and glass substrates, the associated chemistries have diversied and
are frequently based on photoactivatable bridging compounds.64

2.4.3 Aptamers as Biorecognition Elements

Aptamers are synthetic oligonucleotides selected for their specic target anities and have shown a broad propensity to bind to ions, small organics and
proteins.65 Their development and tailoring is a burgeoning scientic eld,
likely to contribute eventually to clinical diagnostics in the way that antibodies
have done. They are, however, more stable than antibodies, and can be
chemically generated, rather than needing a biological system for their production. Their immobilisation for diagnostics essentially follows the route already used for genomic sensors and arrays.
Adsorption of aptamers is less eective than that of proteins; there are fewer
functional groups for noncovalent interactions. An aptamer for ATP, for example,
retained activity when adsorbed to microcrystalline cellulose, but also showed
ready, ionic-strength-dependent, desorption.66 DNA-loaded gold nanoparticles
that had been interlinked by an aptamer able to bind to adenosine disaggregated
in the presence of adenosine and furnished a semiquantitative paper colorimetric
assay.67 Here, whilst the aptamer itself served as a noncovalent bridge to DNA
strands on the nanogold, the latter needed to be attached covalently.
Entrapment of aptamer within a sample-accessible phase can allow for its
ready retention, while allowing a sucient degree of conformational exibility
such that a binding event can trigger transduction. One such arrangement involved the entrapment of RNA aptamer within sol gels68 and binding to a
target led to structural change inducing uorescence; the reporter mechanism
involved induced separation of a uorophore/quencher combination. The polarity and charge of the matrix had important eects on mobility and the access
of the target molecule. The balance of these features provided for the necessary
level of conformational freedom of RNA, whilst allowing it to be stabilised;
ssRNA is less stable relative to ssDNA, and is also more susceptible to enzymic
degradation. By adsorbing ssRNA to graphene oxide, resistance to nuclease
digestion was demonstrated for a theophylline aptamer.69 Mechanisms for this
are unclear, but possibilities considered were altered local ionic composition,
conformational remodelling and steric hindrance. In this regard this may be a

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

49

general nanostructure eect as was suggested by the parallel eect of singlewalled carbon nanotubes at DNA.70
Covalent attachment of aptamers has been by far the most popular route to
diagnostic device design, and has been reviewed by Balamurugan et al.71 The
techniques established are anticipated by those used for nucleic acid detection.
Thiol-derivatised DNA is a classic approach that readily allows for selfassembled monolayers on gold, but consideration needs to be given to both the
mutual charge repulsion of anionic single-strand DNA and the complicating
aspect of nonspecic adsorption. With these provisos, reproducible immobilisation on both planar and particulate gold is achievable under mild conditions.
As with other immobilised biomolecules, the direct exposure of aptamer to the
solution environment is critical to its function. Whilst the structural distortions
that arise at multitethered proteins are not an issue with aptamers, there is
nevertheless a need to design appropriate spacers and spacer lengths to the
surface to optimise binding capacity and anity. Ethylene glycol repeat units as
spacers have had the benecial eect of reducing nonspecic binding as well as
improving binding eectiveness.72 A variant of this has been to immobilise the
ethylene glycol motif on to a gold surface after aptamer attachment; this made
for higher aptamer loading then when simultaneous attachment was carried
out.73 Aromatic spacers can create more compact, stacked, arrays at a surface,
and this could also assist with higher loading.
Conducting polymers such as polypyrrole oer an impedimetric means of
registering anity binding, but also provide a more general and easily deposited
surface with useful functional side groups. Thus, with electropolymerised biotinylated pyrrole,74 an avidin bridge to biotinylated aptamer was readily formed.
The anity system here was an aptamer for prion protein (Figure 2.5); the
spacer DNA sequence used to immobilise the aptamer aected binding anity,

Figure 2.5

Synthetic steps in aptasensor functionalisation of a gold surface: (a)

electropolymerisation of pyrrole with a hydroxyphthalimido derivative
of pyrrole to create a composite layer, (b) functionalisation with biotin
using biotin hydrazide, (c) surface attachment of biotinylated aptamer for
human cellular prion protein (PrPC) through a streptavidin-biotin linkage.
(Reproduced from ref. 74 with permission.)

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

50

Chapter 2

dsDNA, instead of ssDNA, with its more rigid spacer properties giving improved
surface organisation and ultimately a higher binding anity.
In another electropolymerisation procedure, pyrene with pendant lysines
was deposited on Pt;75 the latter allowed for copper (II) binding and through
this linkage to an aptamer that had been functionalised with histidines for histag attachment. Inorganic surfaces may also be used for aptamer attachment.
Thus, an aminated diamond surface functionalised with terephthalic
acid enabled binding of an amino-terminated RNA aptamer;76 the
diamond here formed the sensing interface of a eld eect device for HIV-Tat
protein. One versatile inorganic phase is the quantum dot, usable for
high-sensitivity colorimetric assay, and many of the traditional bridging
methods applied to polymeric surfaces are now being reported for quantum
dots.77,78

2.5 Electrochemical Sensors

2.5.1 Ion-Selective Electrodes
Ion-selective electrodes (ISEs), established with the glass pH electrode, have
been a forerunner of many reagent-free, miniaturised, reversibly responsive
systems advanced for extra-laboratory testing. The original principles enunciated, those of interfacial charge generation at an ion interactive interface, have
not changed, nor has the perennial dependence on a stable reference electrode,
needed to maintain a stable emf in the face of a variable solution electrolyte
composition.79 Special problems arise when using ISEs in biological uids, and
these remain an issue even with advanced, contemporary devices. So whilst
fundamental chemistry has not changed greatly, what has advanced the eld for
clinical use has been the re-engineering of ISEs such that they are now miniaturised, robust, solid state and use integrated reference electrodes.
With new anity ligands, some ISEs have achieved improved selectivity and
sensitivity.80 However, despite constant eorts towards new ligand synthesis, in
the clinical context at least, it has been dicult to extend ISEs beyond the
traditional repertoire of H1, Na1, K1, Ca21, Cl and possibly Li1, Mg21 and
NH41. Certainly, trace metal ion detection and measurement of organic ions
remains elusive. There, however, remains an unassailable market for ISEs for
the core repertoire, key factors being their ability to operate in whole blood
without sample separation or dilution, key advantages for POCT. Dilution is
possible, and does permit better handling in a ow system as well as providing
convergence with a more buer-like matrix; the practical benet of this is
greater assay reproducibility and extension of the assay concentration range,
e.g. as might be required for urine analysis. A drawback to dilution, as with any
assay, is dilution error due to major variation in protein content, often seen in
critical-care patients.81
Sodium measurement originally used modied glass, a variant on the glass
pH electrode, but now PVC membranes loaded with ionophore, notably
monensin, an antibiotic, or its derivatives, are used. In samples such as urine,

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

51

the lipophilic PVC membrane can extract nonpolar solutes that then interfere
with the response, leading to a loss of selectivity as well as to drift. However,
urine assay is possible provided a nonpolar pre-extraction is carried out.82 An
interesting assay variant has been the use of sonication to create PVC microspheres loaded with a sodium ionophore. Within the same phase, a uorescent
chromionophore for H1 was incorporated; uptake of Na1 led to chromionophore deprotonation, simultaneous ejection of H1 and thereby a change in
microsphere uorescence.83 Such an arrangement was considered for Na1 to be
of use in high-throughput analysis. Newer ionophores continue to emerge, one
example was based on a calyx[4]arene ester.84 This sensor system was
congured with a solid-state internal electrolyte comprising polypyrrole
stabilised with a cobalt organic counteranion. The resulting electrode had
enhanced sensitivity and represents a general solid-state ISE model. Crown
ethers such as cyclic ionophores have been frequently studied, but adequate
binding for clinical use has been a challenge. One variant where a silicon atom
formed a part of the ether ring has given improved performance with low-pH
eects and detection limits in the micromolar range for sodium.85
The microbial ionophore valinomycin has provided for high-selectivity
measurement of potassium, with manipulation of internal electrolyte in traditional electrodes enabling extension of sensitivity limits.86 An alternative solid
internal contact electrode was based on poly(3-octylthophene).87 This lipophilic
electron conductor avoids the response instabilities associated with development of intervening aqueous lms with the ionophore layer. Graphene has also
been tried as a solid contact material.88
Various synthetic ionophores have been developed for ionised calcium,
which give relatively trouble-free measurement, and commercial versions of
magnesium sensors have also emerged, though these electrodes, based on
diketone and amide antibiotics, have calcium ion interference and need a
correction step for background calcium levels.89 Chloride sensing has been
achieved through the use of a quaternary ammonium ion loaded membrane,
but there can be interference from other halogen ions and bicarbonate.90
Lithium ion measurement was subject to Na1 interference with earlier ionophores. An improvement has been possible with a redesigned crown ether
chromionophore (Figure 2.6), admittedly for optical sensing, but design considerations for size selectivity are better understood.91 Thus, bulky side groups

Figure 2.6

Li1 selective uoroionophore comprising a crown ether with bulky side

groups for selectivity modied with coumarin as uorophore. R is CH3 or
(CH2)4-CH CH2).
(Reproduced from ref. 91 with permission.)

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

52

Chapter 2

on the crown ether prevented formation of sandwich complexes, avoiding

the bivalent chelation of larger ions, whilst enhancement of ring rigidity prevented entry of smaller ions. pH electrodes using ionophores have, to a degree,
been able to replace their glass counterpart; in one example tridodecylamine
was used as part of a pH-responding ion-selective eld eect transducer
(ISFET).92
There remains a challenge to extrapolate from core measured ions in
medicine to other diagnostically important ions, such as those that are
present in trace amounts or that are organics. Design of anity molecules
based on modelling and simulation may secure the needed extension of the
measurement span.

2.5.2 Amperometric Enzyme Biosensors

One reason for the diculty in extending the analytical repertoire of ISEs, is the
need for a new synthetic anity molecule for each new target ion. Large
numbers of ionophores have been reported, but relatively few have delivered on
the key practical criteria of selectivity, sensitivity, rapid binding and the equally
necessary rapid dissociation. By contrast, amperometric sensors do not need
the invention of a new binding agent for each new analytical target. Not only
can tuning of the polarising voltage provide for initial selectivity, but this can
be augmented through generic barrier membranes to screen out interferents.
Most importantly, redox enzymes and electron mediators are able to, respectively, bind to the solute of interest and facilitate the translation of this into
an analytical signal.93 The early days of amperometric biosensors, moreover,
were considerably aided by a combination of a strong need for clinical blood
glucose measurement (market pull) and the availability of an exceptionally
robust enzyme (glucose oxidase) capable of providing stable biosensors (technology push). This favourable set of circumstances has, alas, not been repeated
for any other biological target, though a high degree of commercial eort and
innovation have underpinned developments in the last couple of decades.
An example of the conversion to practical realisation of a redox enzymebased system has been the glucose and lactate sensors advanced by Roche
Diagnostics.94 Here, the oxidase enzymes for the respective metabolites were
formulated in acrylate polymer to give stable, adherent thick lms over a
screen-printed working electrode on one face, the other face protected by a
biocompatible membrane comprising a PVC copolymer with vinyl acetate/vinyl
maleinate. Another practical adaptive step here was a MnO2carbon ink
composite electrode that regenerated O2 from the enzyme reaction to help replenish the co-substrate, so extending the linear range while also allowing for
measurements at reducing polarising voltages, the latter minimising electrochemical interference.
A combination of enzyme, electron mediator and selectivity barrier was reported for a disposable glucose sensor that used SiO2 nanoparticles as a means
of controlling transmembrane diusion.95 While permeability is a key factor
determining sensor performance, it may not be constant for a given membrane,

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

53

and hydration and ageing eects warrant more detailed analysis. In an attempt
to follow solution eects, Hsu et al.96 monitored thickness and refractive-index
changes at a thin glucose oxidase layer; refractive index altered in solution in a
nonlinear relationship with pH, but thickness was unaected. This conrms
that quite subtle structural changes do occur in enzyme layers and that their
tracking may identify better immobilisation strategies. Regardless of this, the
continued emphasis on substrate/O2 diusion into the enzyme layer may give a
one-sided picture of performance variables. In a modelling study using controlled layer-by-layer deposition of barrier lms at a glucose electrode,97 it was
found that a greater barrier could actually lead to a bigger response. This was
concluded to be due to a parallel reduced permeability to H2O2, and therefore
its accumulation within the sensing layer.
Nanoscale electrodes, especially combined with new materials, provide a
further means of modifying baseline sensor performance, even without use of
separate barrier membranes. At a single-wall carbon-nanotube electrode, the
entrapment of glucose oxidase within a polypyrrole layer considerably extended the upper linear range, while also allowing for detection limits down to
the micromolar range.98 Alternatively, carboxylated graphene enabled
self-assembly of glucose oxidase, and with this a direct electron exchange with
the enzyme prosthetic group, radically lowering the polarisation voltage needed
for a response.99 A more natural direct electron transfer is achieved using a
PQQ-glucose dehydrogenase; here the PQQ (pyrroloquinline quinone) cofactor
is able to relay electrons to a mediator, or directly to a polarised working
electrode, and does not have an O2 requirement; ruthenium hexamine and an
osmium complex has been utilised as the electron relays to construct commercial glucose strips.100 In an example of direct electron transfer, a nanoporous carbon electrode was used.101 The PQQ-glucose dehydrogenase enzyme
is not as selective as glucose oxidase, but the problem has been overcome using
genetically engineered enzyme. In a similar way an FAD-glucose dehydrogenase has been modied to give greater glucose selectivity.102 Even in the
case of classical glucose oxidase, mutagenesis of the amino acid residues involved in the reaction with O2 has allowed for a downregulation of oxidative
activity with parallel upscaling of intrinsic dehydrogenase activity, so enabling
use of articial mediator for an oxygen-independent response.103
Lactate sensors are based on lactate oxidase, a low-stability enzyme with a
high lactate anity, leading, respectively, to low operational stability and a
limited analytical range. Sensor design, as with any O2-dependent system needs
to promote a high transport rate for O2 vs the substrate to maximise the linear
range of operation. The intrinsic high diusion coecient of O2 is advantageous in this regard, as is its regeneration from the H2O2 electrode reaction.
One modelling study showed that enzyme layer concentrations of O2 were
minimally altered in the lactate oxidase layer compared with bulk solution
[O2].104 A taylored screen-printed lactate electrode with cobalt phthalocyanine
as electron mediator for hydrogen peroxide enabled, the operating voltage was
reduced.105 Lactate oxidase can be stabilised by chemical crosslinking, provided the enzyme survives this initial binding step. Further improved storage

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

54

Chapter 2

stability is achievable, and may be eected via the resulting membrane matrix;
an interesting enhancement of stability was seen when mucin was used in the
crosslinking layer.104 Electrode architecture and surface area aect sensitivity,
as was shown using a 3D macroporous gold construct to immobilise the lactate
oxidase.106 The general construction approaches for lactate sensors has been
reviewed by Nikolaus and Strehlitz.107
The advantage of surface area increase was demonstrated for a cholesterol
oxidase-based cholesterol electrode where the enzyme was adsorbed onto a
nanostructured ZnO lm.108 The practical use, challenges of sample delivery,
electrode conguration, timing and temperature control have been described
for a cholesterol strip sensor that used oxidase coupled with horseradish peroxidase and a ferricyanide mediator.109 Even direct electrochemical oxidation
of cholesterol at platinum nanoparticle assemblies on carbon nanotubes has
been demonstrated;110 response was not aected by interferents, apparently;
whether such a system can provide enzyme-free clinical assay remains to
be seen.
Creatinine measurement requires a more complex, multienzyme, sequence.
Despite the challenges of multienzyme functional compatibility under common
reaction conditions, it has been possible to combine the classic sequence of
creatinine amidohydrolase, creatine amidinohydrolase and sarcosine oxidase to
achieve the respective conversion of creatinine to creatine and then to sarcosine
and nally H2O2 end product.111 Typically, an inner membrane barrier is used
to reduce electrochemical interference, but in an alternative approach, Shin
et al.112 used a PbO2 oxidising prelayer to inactivate such redox active compounds. In addition, a hydrophilic polyurethane barrier enabled transport of
cationic creatinine, while limiting entry of creatine. Gel entrapment of the
enzymes avoids denaturation, but retention of activity was also demonstrated
by facile enzyme attachment to porous polypropylene functionalised with
poly(acrylic acid),113 and even chemical crosslinking seems possible.114 A factor
that has not received the level of attention it needs is the inhibition of enzyme
activity by Ag1 from Ag/AgCl reference electrodes. In the case of creatinine
amidohydrolase, rapid loss of activity was seen with 1 mM Ag1;115 a covering
barrier of cellulose acetate over the reference was able to protect against Ag1
leaching.

2.5.3 Immunosensors
Direct electrochemical detection of an antigenic target, through binding to an
immobilised antibody, will generate surface-charge eects, and these may be
registered as impedimetric or potentiometric change.116 This is an attractive,
simple means of detection, but in reality the strong eects of biological sample
solution variables and nonspecic binding of macromolecules make it dicult
to consider such systems for clinical use. Practical progress has been achieved,
therefore, mainly by the use of tried and tested electrochemical and redox enzyme labels, utilising traditional immunoassay formats classied, as with any
other immunoassay, as competitive or noncompetitive.

View Online

55

Integrated Chemistries for Analytical Simplication and Point of Care Testing

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

117

A classical noncompetitive immunoassay is represented in an IgA assay where

horseradish peroxidase (HRP) was used as the label on a binding second antibody
(Figure 2.7). An immobilised rst antibody on a gold electrode provided the
anity surface for initial IgA attachment, and nonspecic protein binding was
reduced by near-surface poly(ethylene glycol). HRP, in the presence of H2O2,
oxidised o-dianisidine (ODS) to give a product that was detectable at the electrode.
This arrangement shows both the feasibility of using amperometric formats and the
diculty of converting any such approach into a single-step assay.
One way of achieving assay simplicity is to automate reagent presentation to
the electrode surface. A microuidic arrangement was devised by Liang et al.118
for brain natriuretic peptide (BNP) where Prussian blue as electron mediator
was used to amplify the response from the second antibody label.
A multiparameter (carcinoembryonic antigen, a-fetoprotein, b-human choriogonadotropin, carcinoma antigen 125) competitive immunoassay has been
reported. This used limited antibody in solution for competitive binding between sample antigen and membrane immobilised antigen. The resulting surface binding of antibody provided a measure of target antigen level; HRP was
the tag on the antibody, and membrane-entrapped mediator enhanced the
enzyme signal.119 A similar competitive assay was realised for D-dimer in
serum, but here the immobilised antigen had been retained on magnetic
beads, later captured at a working electrode surface.120 The advantage of the

Figure 2.7

Schematic of an electrochemical immunoassay for IgA. Bipodal carboxy

terminated alkyl thiols provide a surface function for immobilising capture antibody via carbodiimide/N-hydroxy-succinimide bridging. The
second antibody is horseradish peroxidase (HRP) labelled to catalyse an
indicator reaction using O-dianosidine as mediator.
(Reproduced from ref. 117 with permission.)

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

56

Chapter 2

distributed, particulate phase was that it accelerated binding exchanges, while

the facility for magnetic transfer meant both a simplied separation of free
from bound antibody and a way of carrying out electrochemical detection
under better controlled solution conditions. In an alternative format, a competitive immunoassay for Staphylococcus aureus involved the limited antibody
being immobilised on magnetic particles, with labelled antigen in solution.121
Because binding here involves only one antibody, two epitopes are not needed,
and so small molecules can be estimated. Oestradiol was measured in this way
using monoclonal antibody attached to gold nanoparticles;122 competitive
binding between HRP labelled and unlabelled oestrodiol was measured electrochemically via enzymatically generated benzoquinone.

2.5.4 DNA Sensors

Many types of DNA sensors have been reported, but they generally require
PCR amplication to generate sucient target DNA and depend on some form
of labelling, e.g. electrochemical, for determining the hybridisation reaction.
There are, therefore, parallels with immunosensors. Intrinsic electrochemical
properties of DNA bases have been utilised for sensing DNA, as have DNAinduced impedimetric and potentiometric changes at surfaces, but these are
invariably basic studies in highly controlled solution.
An unconventional means of generating an amplied signal at captured
DNA was to use a gold nanoparticle labelled probe DNA to hybridise to
cytomegalovirus DNA.123 Here, oxidative dissolution of the nanoparticles released high quantities of Au(III) ions that were then determined by anodestripping voltammetry. Duplex DNA intercalating agents are frequently used to
identify hybridisation. In one example, methylene blue was used as the agent, and
was able to reduce ferricyanide which served as a diusible redox reporter.124 A
common intercalater is the electrochemically active dye Hoechst 33258, which
can be oxidised at modest polarising voltages. This mode of reporting has been
used to detect single base-pair mismatches,125 though sensitivity is limited. The
binder was used for detection of PCR-amplied DNA mixtures using an electrode array.126 New forms of label continue to be developed; thus, noncovalent
labelling was achieved using Zr(IV).127 The highly charged ion has strong anity
for dsDNA phosphate groups and was able to bridge to an anionic ferrocene
carboxylate reporter. Redox cycling of the latter led to electrochemical signal
generation at hybridised DNA. The exploitation of labels has been extensively
reported at DNA sensors, and optical methods will add to the eort on electrochemical approaches, but the challenge will be of how to simplify procedures
suciently for POCT. There remain a set of irreducible steps, so the most
plausible strategy would appear to be their automation.

2.6 Micromechanical Transduction

Microgravimetric and surface acoustic wave (SAW) sensors oer two
distinct mass-based transduction modalities that have the ability to register

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

Figure 2.8

57

Schematic of surface acoustic wave (SAW) immunosensor in contact with

(1): arrows indicate ow direction. Device components are: (2) a piezoelectric crystal (3) interdigitated transducers to set up the acoustic wave,
(4) the surface acoustic wave, (5) capture antibodies (6) target molecules.
The driving electronics (7) generate change in the output signal (8) when
the target molecule binds to the sensor surface.
(Reproduced from ref. 130 with permission.)

protein/macromolecule binding. Variables such as deposited surface mass, layer

thickness and post-binding surface microviscosity change all contribute to the
response, but a simple quantitative link can, nevertheless, be made with the
amount bound. Typically, antibodies are used to functionalise the surface for
selectivity. Low molecular weight species such as drugs can be measured, but a
label is needed to generate a high-mass product. High-sensitivity detection of
protein breast cancer markers was demonstrated, for example, at a SAW device
functionalised with antibody;128 coupling chemistry that delivered an orientated
antibody gave better sensitivity. In another study, precoating of the SAW surface with a glass coating reduced acoustic velocity, so increasing surface wave
connement and therefore sensitivity.129 Additional use of a poly(vinylamine)
phase reduced nonspecic binding, allowing assay in diluted whole blood.
A schematic of an immunosensor based on a SAW device is shown in Figure 2.8.
The general eld and the potential for commercialisation have been reviewed.130

2.7 Commercialisation
POCT systems as consumer products demand considerable multidisciplinary
input. Only in this way is the integration of multifunctional components possible. Final design goes well beyond the original chemistry concept, and the
often perceived advantages of the original chemistry become submerged in
quite unexpected problems not amenable to chemical manipulation. Overall,
engineering and analytical principles require to converge in order to address the
special fabrication and design needs of robust analysis. The associated development costs are not trivial, and are orders of magnitude greater than those of

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

58

Chapter 2

the original research. This ramp-up is as much of a barrier as the technical issues
and a reason why many promising systems do not reach the market place.
There have been notable commercial successes beyond glucometers. One
well-established system is the iSTAT; the cartridge devices of this deskilled
analytical platform make for a highly attractive analytical package.131 While
not achieving a matched simplicity, the Vantixt sandwich immunoassay system
with its antigen-coated potentiometric sensor goes a long way to simplifying
this type of assay;132 immunoassay is inherently dicult to simplify. The Alpha
DX, a STAT whole-blood analysis system alternatively approaches immunoassay by uorescent sandwich immunoassay via target molecule capture at a
test disc.133 In a further example, a radial geometry immunoassay (Stratus
CSs) has been developed for cardiac markers134 where a dendrimer bridging
molecule improves the presentation of capture antibody. For the Biosite
Triages there is linear tracking of reagent and antigen, a similar platform has
also been advanced for brain natriuretic peptide.135
A commercial ABL800 FLEX analyser used for measurements of pH, blood
gas, electrolyte, oximetry and metabolite parameters has also been developed.136

2.8 Conclusions
Transformation to the solid state has been an important route to producing
POCT devices, enabling incorporation of both biological and articial functional components. The technology for immobilisation of delicate biological
reagents, especially, has allowed the development of a diverse range of measurement platforms. In addition, the wider range of transduction and registration methods have extended the ways in which biorecognition can be
followed. While there are still no real alternatives to biological receptor systems
for the measurement of complex organics, their conversion to robust reagents,
either through chemistry or genetic engineering, is likely to expand.
Parallel developments in miniaturisation and lamination technologies
have enabled radical design changes. The increased use of micro- and
nano-constructs, along with multilayer deposition techniques has allowed
creation of ever more complex devices with greater dimensional compression.
A parallel exists with the functional compression of electronic devices. There
will be a time, also, when device incorporation into electronics will become
seamless and the added value of data interpretation and transmission fully
realised. This would not be just a technological goal for its own sake, but a
means of data sharing to enhance clinical interpretation.
Recent miniaturisation eort has led on to newer materials such as CNTs
and graphene. These may eventually deliver practical devices that allow the
utilisation of new engineering principles. Tomorrows POCT devices may well
become less intrusive than those of today with the newer materials. If automated preanalytical processes and sample handling can be similarly re-engineered, then convergence to the lab-on-a-chip concept will be possible in the
real world. Some of the subcomponents set to make up such platforms have

View Online

Integrated Chemistries for Analytical Simplication and Point of Care Testing

59

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

been described in this chapter. What remains as an intriguing next phase is their
integration into seamless systems.

Acknowledgement
The authors wish to thank EPSRC for generous funding through Project EP/
H009744/1 (ESPRIT).

References
1. C. P. Price, Brit. Med. J., 2001, 322, 1285.
2. P. B. Luppa, C. Muller, A. Schlichtiger and H. Schlebusch, Trac-Trends
Analyt. Chem., 2011, 30, 887.
3. J. Y. Wick, The Consultant Pharmacist: J. Am. Soc. Consultant Pharmacists, 2010, 25, 416.
4. C. P. Price and L. J. Kricka, Clin. Chem., 2007, 53, 1665.
5. T. Vo-Dinh, B. M. Cullum and D. L. Stokes, Sens. Actuators B-Chem.,
2001, 74, 2.
6. R. Pauliukaite, M. Schoenleber, P. Vadgama and C. M. A. Brett, Anal.
Bioanal. Chem., 2008, 390, 1121.
7. S. M. Ng, P. Krishnaswamy, R. Morissey, P. Clopton, R. Fitzgerald and
A. S. Maisel, Am. J. Cardiol., 2001, 88, 611.
8. J. McCord, R. M. Nowak, P. A. McCullough, C. Foreback, S. Borzak,
G. Tokarski, M. C. Tomlanovich, G. Jacobsen and W. D. Weaver,
Circulation, 2001, 104, 1483.
9. J. E. Dechant, W. A. Symm and J. E. Nieto, Vet. Surg., 2011, 40, 811.
10. Y. C. Manabe, Y. P. Wang, A. Elbireer, B. Auerbach and B. Castelnuovo,
Plos One, 2012, 7, e34319.
11. J. Porter, N. OLoan, B. Bell, J. Mahoney, M. McGarrity,
R. I. McConnell and S. P. Fitzgerald, Anal. Bioanal. Chem., 2012,
403, 3051.
12. M. A. Qasaimeh, S. G. Ricoult and D. Juncker, Lab on a Chip, 2013,
13, 40.
13. T. H. Schulte, R. L. Bardell and B. H. Weigl, Clin. Chim. Acta, 2002,
321, 1.
14. P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M. R. Tam and
B. H. Weigl, Nature, 2006, 442, 412.
15. A. E. Herr, A. V. Hatch, D. J. Throckmorton, H. M. Tran, J. S. Brennan,
W. V. Giannobile and A. K. Singh, Proc. Nat. Acad. Sci. USA, 2007, 104,
5268.
16. H. Chang, R. Khan, Z. Rong, A. Sapelkin and P. Vadgama, Biofabrication, 2010, 2, Art No 035002.
17. M. K. Ramasubramanian and S. P. Alexander, Biomed. Microdevices,
2009, 11, 217.
18. L. Gervais and E. Delamarche, Lab on a Chip, 2009, 9, 3330.
19. S. Hsiung, C. Chen and G. Lee, Lab on a chip, 2009, 9, 3330.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

60

Chapter 2

20. D. S. Millington, R. Sista, A. Eckhardt, J. Rouse, D. Bali, R. Goldberg,

M. Cotten, R. Buckley and V. Pamula, Sem. Perinatol., 2010,
34, 163.
21. A. Bhattacharyya and C. M. Klapperich, Biomed. Microdevices, 2007,
9, 245.
22. M. H. Zweig and A. Jackson, Clin. Chem., 1986, 32, 674.
23. B. Walter, Method Enzymol., 1998, 137, 794.
24. R. Sommer, C. Nelson and A. Greenquist, Clin. Chem., 1986, 32,
1770.
25. R. H. Ng, K. M. Sparks and B. E. Statland, Clin. Chem., 1992, 38, 1371.
26. M. J. Pugia, J. A. Lott, J. A. Prott and T. K. Cast, J. Clin. Lab. Anal.,
1999, 13, 180.
27. A. Tripodi, V. Chantarangkul, M. Primignani, A. DellEra, M. Clerici,
F. Iannuzzi, A. Aghemo, M. Cazzaniga, F. Salerno and P. M. Mannucci,
J. Hepatol., 2009, 51, 288.
28. A. Warsinke, Anal. Bioanal. Chem., 2009, 393, 1393.
29. S. Puertas, M. Moros, R. Fernandez-Pacheco, M. R. Ibarra, V. Grazu
and J. M. de la Fuente, J. Phys. D Appl. Phys., 2010, 43.
30. W. U. Dittmer, T. H. Evers, W. M. Hardeman, W. Huijnen, R. Kamps,
P. de Kievit, J. H. M. Neijzen, J. H. Nieuwenhuis, M. J. J. Sijbers, D. W.
C. Dekkers, M. H. Hefti and M. F. W. C. Martens, Clin. Chim. Acta,
2010, 411, 868.
31. G. A. Posthuma-Trumpie, J. Korf and A. van Amerongen, Anal. Bioanal.
Chem., 2009, 393, 569.
32. S. Choi, E. Y. Choi, H. S. Kim and S. W. Oh, Clin. Chem., 2004, 50, 1052.
33. W. E. Doering, M. E. Piotti, M. J. Natan and R. G. Freeman, Adv.
Mater., 2007, 19, 3100.
34. G. A. Posthuma-Trumpie, J. H. Wichers, M. Koets, L. B. J. M. Berendsen
and A. van Amerongen, Anal. Bioanal. Chem., 2012, 402, 593.
35. M. Rayev and K. Shmagel, J. Immun. Methods, 2008, 336, 9.
36. P. F. Mens, A. van Amerongen, P. Sawa, P. A. Kager and H. D. F.
H. Schallig, Diagnos. Microbiol. Infect. Dis., 2008, 61, 421.
37. K. A. Edwards and A. J. Baeumner, Anal. Bioanal. Chem., 2006,
386, 1613.
38. N. V. Zaytseva, R. A. Montagna, E. M. Lee and A. J. Baeumner, Anal.
Bioanal. Chem., 2004, 380, 46.
39. N. V. Kulagina, K. M. Shaer, F. S. Ligler and C. R. Taitt, Sens.
Actuators B-Chem., 2007, 121, 150.
40. A. W. Martinez, S. T. Phillips and G. M. Whitesides, Proc. Natl. Acad.
Sci. USA, 2008, 105, 19606.
41. F. Kong and Y. F. Hu, Anal. Bioanal. Chem., 2012, 403, 7.
42. J. Lehtio, J. Sugiyama, M. Gustavsson, L. Fransson, M. Linder and
T. T. Teeri, Proc. Natl. Acad. Sci. USA, 2003, 100, 484.
43. S. Di Risio and N. Yan, Colloids Surf. B-Biointerfaces, 2010, 79, 397.
44. W. Zhang, W. T. Ang, C.-Y. Xue and K.-L. Yang, ACS Appl. Mater.
Interfaces, 2011, 3, 3496.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

61

45. M. Khan, X. Li, W. Shen and G. Garnier, Colloids and Surfaces BBiointerfaces, 2010, 75, 239.
46. L. Pardo, W. C. Wilson and T. J. Boland, Langmuir, 2003, 19, 1462.
47. J. Wang, B. Yiu, J. Obermeyer, C. D. M. Filipe, J. D. Brennan and
R. Pelton, Biomacromolecules, 2012, 13, 559.
48. H. Yan, A. Ahmad-Tajudin, M. Bengtsson, S. Xiao, T. Laurell and
S. Ekstrom, Analyt. Chem., 2011, 83, 4942.
49. Q. Ma, T.-Y. Song, P. Yuan, C. Wang and X.-G. Su, Colloids Surf.
B-Biointerfaces, 2008, 64, 248.
50. H. Wang, Y. L. Liu, Y. H. Yang, T. Deng, G. L. Shen and R. Q. Yu,
Anal. Biochem., 2004, 324, 219.
51. P. Peluso, D. S. Wilson, D. Do, H. Tran, M. Venkatasubbaiah,
D. Quincy, B. Heidecker, K. Poindexter, N. Tolani, M. Phelan, K. Witte,
L. S. Jung, P. Wagner and S. Nock, Anal. Biochem., 2003, 312, 113.
52. K. Bonroy, F. Frederix, G. Reekmans, E. Dewolf, R. De Palma,
G. Borghs, P. Declerck and B. Goddeeris, J. Immun. Methods, 2006,
312, 167.
53. J. Buijs, J. W. T. Lichtenbelt, W. Norde and J. Lyklema, Colloids Surf.
B-Biointerfaces, 1995, 5, 11.
54. X. Yuan, D. Fabregat, K. Yoshimoto and Y. Nagasaki, Colloids Surf.
B-Biointerfaces, 2012, 99, 45.
55. I. Migneault, C. Dartiguenave, M. J. Bertrand and K. C. Waldron, Biotechniques, 2004, 37, 790.
56. F. Lopez-Gallego, L. Betancor, C. Mateo, A. Hidalgo, N. AlonsoMorales, G. Dellamora-Ortiz, J. M. Guisan and R. Fernandez-Lafuente,
J. Biotech., 2005, 119, 70.
57. M. N. Gupta and S. Raghava, Methods Mol. Biol. (Clifton, N.J.), 2011,
679, 133.
58. R. A. Sheldon, Biochem. Soc. Trans., 2007, 35, 1583.
59. A. Haouz, J. M. Glandieres and B. Alpert, Febs Lett., 2001, 506, 216.
60. J. C. Tiller, R. Rieseler, P. Berlin and D. Klemm, Biomacromolecules,
2002, 3, 1021.
61. U. Bora, P. Sharma, K. Kannan and P. Nahar, J. Biotech., 2006,
126, 220.
62. C. Tyagi, L. K. Tomar and H. Singh, J. Appl. Polym. Sci., 2009,
111, 1381.
63. R. Pelton, Trac-Trends Analyt. Chem., 2009, 28, 925.
64. P. Jonkheijm, D. Weinrich, M. Koehn, H. Engelkamp, P. C. M.
Christianen, J. Kuhlmann, J. C. Maan, D. Nuesse, H. Schroeder,
R. Wacker, R. Breinbauer, C. M. Niemeyer and H. Waldmann, Angew.
Chem.-Int. Ed., 2008, 47, 4421.
65. D. H. J. Bunka and P. G. Stockley, Nature Rev. Microbiol., 2006, 4, 588.
66. Su, R. Nutiu, C. D. M. Filipe, Y. Li and R. Pelton, Langmuir, 2007,
23, 1300.
67. W. Zhao, M. M. Ali, S. D. Aguirre, M. A. Brook and Y. Li, Anal. Chem.,
2008, 80, 8431.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

62

Chapter 2

68. C. Carrasquilla, P. S. Lau, Y. Li and J. D. Brennan, J. Am. Chem. Soc.,

2012, 134, 10998.
69. L. Cui, Z. Chen, Z. Zhu, X. Lin, X. Chen and C. J. Yang, Analyt. Chem.,
2013, 85, 2269.
70. Y. Wu, J. A. Phillips, H. Liu, R. Yang and W. Tan, ACS Nano, 2008,
2, 2023.
71. S. Balamurugan, A. Obubuafo, S. A. Soper and D. A. Spivak, Anal.
Bioanal. Chem., 2008, 390, 1009.
72. S. J. Hurst, A. K. R. Lytton-Jean and C. A. Mirkin, Analyt. Chem., 2006,
78, 8313.
73. X. Zhang and V. K. Yadavalli, Biosens. Bioelectron., 2011, 26, 3142.
74. A. Miodek, A. Poturnayova, M. Snejdarkova, T. Hianik and H. KorriYoussou, Anal. Bioanal. Chem., 2013, 405, 2505.
75. Y. Wenjuan, A. Le Go, N. Spinelli, M. Holzinger, G.-W. Diao, D. Shan,
E. Defrancq and S. Cosnier, Biosens. Bioelectron., 2013, 42, 556.
76. M. Levy, S. F. Cater and A. D. Ellington, Chem. Biochem., 2005,
6, 2163.
77. X.-C. Chen, Y.-L. Deng, Y. Lin, D.-W. Pang, H. Qing, F. Qu and
H.-Y. Xie, Nanotechnology, 2008, 19.
78. G. Mayer, Angew. Chem.-Int. Ed., 2009, 48, 2672.
79. R. P. Buck and E. Lindner, Acc. Chem. Res., 1998, 31, 257.
80. E. Bakker, P. Buhlmann and E. Pretsch, Electroanalysis, 1999, 11, 915.
81. G. Dimeski, T. J. Morgan, J. J. Presneill and B. Venkatesh, J. Crit. Care,
2012, 27.
82. F. Phillips, K. Kaczor, N. Gandhi, B. D. Pendley, R. K. Danish,
M. R. Neuman, B. Toth, V. Horvath and E. Lindner, Talanta, 2007,
74, 255.
83. R. Retter, S. Peper, M. Bell, I. Tsagkatakis and E. Bakker, Analyt. Chem.,
2002, 74, 5420.
84. I. A. M. de Oliveira, D. Risco, F. Vocanson, E. Crespo, F. Teixidor,
N. Zine, J. Bausells, J. Samitier and A. Errachid, Sens. Actuators BChem., 2008, 130, 295.
85. S. Chandra and H. Lang, Sens. Actuators B-Chem., 2006, 114, 849.
86. W. Qin, T. Zwickl and E. Pretsch, Analyt. Chem., 2000, 72, 3236.
87. K. Y. Chumbimuni-Torres, N. Rubinova, A. Radu, L. T. Kubota and
E. Bakker, Analyt. Chem., 2006, 78, 1318.
88. F. Li, J. Ye, M. Zhou, S. Gan, Q. Zhang, D. Han, and Li Niu, Analyst,
2012, 137, 618.
89. G. Dimeski, T. Badrick and A. St John, Clinica Chim. Acta, 2010,
411, 309.
90. G. Dimeski and A. E. Clague, Clin. Chem., 2004, 50, 1106.
91. D. Citterio, J. Takeda, M. Kosugi, H. Hisamoto, S.-i. Sasaki,
H. Komatsu and K. Suzuki, Analyt. Chem., 2007, 79, 1237.
92. N. Abramova and A. Bratov, Talanta, 2010, 81, 208.
93. P. Vadgama and P. W. Crump, Analyst, 1992, 117, 1657.
94. B. P. H. Schaar, Anal. Bioanal. Chem., 2002, 372, 254.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

Integrated Chemistries for Analytical Simplication and Point of Care Testing

63

95. Z. Chen, C. Fang, H. Wang and J. He, Electrochimica Acta, 2009,

55, 544.
96. C.-C. Hsu, W.-L. Lan, Y.-C. Chen and Y.-Y. Sung, Opt. Lasers Eng.,
2013, 51, 388.
97. R. A. Croce, Jr., S. Vaddiraju, F. Papadimitrakopoulos and F. C. Jain,
Sensors, 2012, 12, 13402.
98. F. Valentini, L. G. Fernandez, E. Tamburri and G. Palleschi, Biosens.
Bioelectron., 2013, 43, 75.
99. B. Liang, L. Fang, G. Yang, Y. Hu, X. Guo and X. Ye, Biosens. Bioelectron., 2013, 43, 131.
100. A. Heller and B. Feldman, Chem. Rev., 2008, 108, 2482.
101. V. Flexer, F. Durand, S. Tsujimura and N. Mano, Analyt. Chem., 2011,
83, 5721.
102. Y. Yamashita, S. Ferri, M. L. Huynh, H. Shimizu, H. Yamaoka and
K. Sode, Enzyme Microbial. Technol., 2013, 52, 123.
103. Y. Horaguchi, S. Saito, K. Kojima, W. Tsugawa, S. Ferri and K. Sode,
Int. J. Mol. Sci., 2012, 13, 14149.
104. M. R. Romero, F. Ahumada, F. Garay and A. M. Baruzzi, Analyt.
Chem., 2010, 82, 5568.
105. T. Shimomura, T. Sumiya, M. Ono, T. Ito and T.-a. Hanaoka, Anal.
Chim. Acta, 2012, 714, 114.
106. M. Gamero, J. L. G. Fierro, E. Lorenzo and C. Alonso, Electroanalysis,
2013, 25, 179.
107. N. Nikolaus and B. Strehlitz, Microchimica Acta, 2008, 160, 15.
108. N. Batra, M. Tomar and V. Gupta, Analyst, 2012, 137, 5854.
109. R. Foster, J. Cassidy and E. ODonoghue, Electroanalysis, 2000,
12, 716.
110. J. Yang, H. Lee, M. Cho, J. Nam and Y. Lee, Sens. Actuators B-Chem.,
2012, 171, 374.
111. T. Tsuchida and K. Yoda, Clin. Chem., 1983, 29, 51.
112. J. H. Shin, Y. S. Choi, H. J. Lee, S. H. Choi, J. Ha, I. J. Yoon, H. Nam
and G. S. Cha, Analyt. Chem., 2001, 73, 5965.
113. G. H. Hsiue, P. L. Lu and J. C. Chen, J. Appl. Polym. Sci., 2004, 92, 3126.
114. M. B. Madaras, I. C. Popescu, S. Ufer and R. P. Buck, Anal. Chim. Acta,
1996, 319, 335.
115. J. A. Berberich, A. Chan, M. Boden and A. J. Russell, Acta Biomater.,
2005, 1, 193.
116. A. Warsinke, A. Benkert and F. W. Scheller, Fresenius J. Analyt. Chem.,
2000, 366, 622.
117. L. Carlos Rosales-Rivera, J. Lluis Acero-Sanchez, P. Lozano-Sanchez,
I. Katakis and C. K. OSullivan, Biosens. Bioelectron., 2012, 33, 134.
118. W. Liang, Y. Li, B. Zhang, Z. Zhang, A. Chen, D. Qi, W. Yi and C. Hu,
Biosens. Bioelectron., 2012, 31, 480.
119. J. Wu, F. Yan, J. Tang, C. Zhai and H. Ju, Clin. Chem., 2007, 53, 1495.
120. M. Gamella, S. Campuzano, F. Conzuelo, A. Julio Reviejo and
J. M. Pingarron, Electroanalysis, 2012, 24, 2235.

View Online

14/10/2013 09:27:45.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00035

64

Chapter 2

121. B. Esteban-Fernandez de Avila, M. Pedrero, S. Campuzano,

V. Escamilla-Gomez and J. M. Pingarron, Anal. Bioanal. Chem., 2012,
403, 917.
122. X. Liu and D. K. Y. Wong, Talanta, 2009, 77, 1437.
123. L. Authier, C. Grossiord, P. Brossier and B. Limoges, Analyt. Chem.,
2001, 73, 4450.
124. E. M. Boon, D. M. Ceres, T. G. Drummond, M. G. Hill and J. K. Barton,
Nature Biotechnol., 2000, 18, 1096.
125. Y. S. Choi and D. H. Park, J. Korean Phys. Soc., 2004, 44, 1556.
126. M. Takahashi, J. Okada, K. Ito, M. Hashimoto, K. Hashimoto,
Y. Yoshida, Y. Furuichi, Y. Ohta, S. Mishiro and N Gemma, Clin.
Chem., 2004, 50, 658.
127. P. Wipawakarn, H. Ju and D. K. Y. Wong, Anal. Bioanal. Chem., 2012,
402, 2817.
128. F. J. Gruhl and K. Laenge, Anal. Biochem., 2012, 420, 188.
129. S. Rupp, M. von Schickfus, S. Hunklinger, H. Eipel, A. Priebe, D. Enders
and A. Pucci, Sens. Actuators B-Chem., 2008, 134, 225.
130. K. Laenge, B. E. Rapp and M. Rapp, Anal. Bioanal. Chem., 2008,
391, 1509.
131. I. R. Lauks, Acc. Chem. Res., 1998, 31, 317.
132. J. Cork, R. M. Jones and J. Sawyer, J. Immun. Methods, 2013, 387, 140.
133. F. S. Apple, F. P. Anderson, P. Collinson, R. L. Jesse, M. C. Kontos,
M. A. Levitt, E. A. Miller and M. M. Murakami, Clin. Chem., 2000,
46, 1604.
134. S. Altinier, M. Mion, A. Cappelletti, M. Zaninotto and M. Plebani, Clin.
Chem., 2000, 46, 991.
135. Y. Fischer, K. Filzmaier, H. Stiegler, J. Graf, S. Fuhs, A. Franke,
U. Janssens and A. M. Gressner, Clin. Chem., 2001, 47, 591.
136. http://www.radiometer.com.

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

CHAPTER 3

Blood-Glucose Biosensors,
Development and Challenges
YUAN WANG*a AND MADELEINE HUb
a

Siemens HealthCare Diagnostics, 511 Benedix Ave. Tarrytown, NY 10591,

United States; b The College of New Jersey, 2000 Pennington Road, Ewing,
NJ 08628-0718, United States
*Email: [email protected]

3.1 Introduction
Diabetes mellitus is the most common endocrine disorder of carbohydrate
metabolism. It is one of the major causes of premature illness and death
worldwide. The World Health Organization (WHO) estimated that 285 million
people, corresponding to 6.4% of the worlds adult population, lived with
diabetes in 2010.1 The number is expected to increase to 439 million by 2030,
corresponding to 7.8% of the worlds adult population. A sedentary lifestyle
combined with changes in eating habits and increasing obesity is thought to be
the main cause of such an increase. The National Health and Nutrition
Examination Survey III shows that two-thirds of adults in the US diagnosed
with Type 2 diabetes have a body mass index (BMI) of 27 or greater, which is
classied as overweight and unhealthy.
With an increasingly diabetic population, the Blood Glucose Monitoring
System (BGMS) is becoming an ever-important tool for diabetes management.28 The goal of BGMS is to help the patients achieve and maintain normal
blood glucose concentrations in order to delay or even prevent the progression
of microvascular (retinopathy, nephropathy and neuropathy) and macrovascular (stroke and coronary artery disease) complications. The ndings of the
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

65

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

66

Chapter 3

Diabetes Control and Complications Trial (DCCT) and the United Kingdom
Prospective Diabetes Study (UKPDS) clearly showed that intensive control of
elevated levels of blood glucose in patients with diabetes decreases the occurrence of complications such as nephropathy, neuropathy and retinopathy and
may also reduce the occurrence and severity of large blood vessel disease.911 In
addition, the system can also be useful for detecting hypoglycemia and
providing real-time information for immediate adjustment in medications,
dietary, and physical activities.6,12 Regular and frequent measurement of
blood glucose can provide valuable information for optimizing and/or
changing patient treatment strategies.
Due to the aforementioned rapid increase in diabetes population and the
growing awareness for good diabetes management, the growth of BGMS
market is accelerating. In 2004, BGMS accounted for approximately 85% of
the world market for biosensors, which had been estimated to be around
$5 billion USD.13 According to a recent report by Global Industry Analysts,
Inc., the global market for glucose biosensors and strips would reach
$11.5 billion USD by 2012.
In this chapter, we will review the development history of the BGMS, explain
the principal chemical mechanisms on which various systems are based, discuss
issues relating to accurate and robust detection, including minimizing of
interference, and, nally, look at challenges and future trends.

3.2 History of Blood-Glucose Biosensor

In 1932 Warburg and Christian reported the yellow enzyme from yeast
changed to colorless upon oxidizing its substrate and reassumed the yellow
color after reoxidation by oxygen.14 In 1939, Franke and Dener isolated
glucose oxidase (GOx) from Aspergillus niger and revealed that the GOx reaction center containing riboavin that was involved in its reduction/oxidation
reaction.15 The mechanism of the glucose and glucose oxidase (GOx) reaction
has been established and is well understood. Glucose reacts with yellow FADGOx to generate FADH2-GOx which is then oxidized by oxygen, producing
FAD-GOx and hydrogen peroxide.
Glucose FAD-GOx ! gluconolactone FADH2 -GOx

3:1

FADH2 -GOx O2 ! FAD-GOx H2 O2

3:2

The rst glucose assay using electrochemical methods was based on hydrogen peroxide oxidation of I to I2 (dark blue), where the concentration of H2O2
was determined by I/I2 potentiometry.16,17
H2 O2 2H 2I ! H2 O I2 :

3:3

The rst portable blood glucose meter, the Ames Reectance Meter (ARM),
was introduced by Miles Laboratories (Elkhart, IN, USA) in 1969. A drop of

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

67

blood was added onto a Dextrostix and left for one minute before being washed
o. The Dextrostix was then placed in the ARM and a beam of light was shined
on it.1821 A darker blue strip would be generated by a higher H2O2 content and
would reect back less light, therefore indicating higher blood glucose content.
However, this is not how a typical glucose biosensor works today.

3.2.1 First Generation of Biosensors

Considerable advances in glucose biosensors were made after the introduction
of the oxygen electrode by Clark and Lyons in 1962.22 However, problems were
encountered with various background oxygen levels that led to irreproducible
results. Updike and Hicks stabilized GOx in a polyacrylamide gel on an oxygen
electrode that corrected for background oxygen levels and signicantly
simplied the electrochemical glucose assay.23,24 The rst amperometric
enzyme glucose sensor was developed in 1973, in which hydrogen peroxide
levels were analyzed instead of the highly variable oxygen reduction current.25
Nevertheless, the necessary oxidation of hydrogen peroxide required a relatively higher potential, at which common endogenous reducing substances,
such as ascorbic acid and uric acid, etc., would be oxidized, causing a falsely
elevated glucose reading.
The rst generation of biosensors was dependent on the presence of oxygen
as a cofactor to ensure the catalytic regeneration of the FAD center (as in reaction 3.2). This method of sensor generation faced two major problems:
interference from electroactive species of endogenous reducing substances and
common drugs such as acetaminophen, and a dependence on free oxygen as a
catalytic mediator.
Although GOx is generally a very specic and selective enzyme, the potential
range at which the hydrogen peroxide is oxidized coincides with the oxidation
potential of a number of compounds found in blood. These compounds include
both endogenous and exogenous substances (Table 3.1). In addition, blood
protein and chloride ions can be absorbed onto the electrodes surface and leave
the sensor susceptible to fouling.26
Oxygen dependence is a problem specic to this generation. The oxygen
content in various blood samples can be considerably dierent. In arterial
blood samples, the normal partial pressure of oxygen (pO2) is in a range of
Table 3.1

Common endogenous and exogenous

interference substances.

Endogenous substances

Exogenous substances

Ascorbic acid
Bilirubin
Chloride
L-Cystine
Protein
Uric acid

Acetaminophen
Alcohol
L-Dopa
Salicylic acid
Tolazamide

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

68

Chapter 3

75100 mmHg, whereas the pO2 level in venous blood is about 40 mmHg. Since
the regeneration of FAD-GOx requires one oxygen molecule per glucose
molecule (reaction 3.2), depending on the dissolved oxygen level in the blood
sample, the errors surrounding this high oxygen dependence is quite signicant.
To overcome the problems encountered in the rst generation glucose biosensors, a number of approaches have been developed. A comprehensive review
paper describing dierent methods used for elimination of electrochemical
interference in glucose biosensors was published by Jia et al.27 These methods
include coating the sensor with a preoxidizing layer to convert the interference
substances into electrochemically inert forms before reaching the electrode
surface;2834 the use of a permselective membrane to block electroactive substances reaching electrode surface; utilizing a modied layer, a mediator, or
nanomaterial to allow a reduced potential where the eects of electroactive
substances are negligible; and use of scanning electrochemical microscopy
(SECM) to detect depletion of electroactive species in the diusion layer.
Many methods were also proposed to minimize the errors caused by various
background oxygen levels. These include the use of a polymer membrane that
limits glucose diusion, allowing more oxygen input relative to glucose which
eases the oxygen dependence issue;3538 the use of oxygen-rich electrode materials that act as an internal source of oxygen;39,40 and replacement of GOx
with an oxygen insensitive enzyme, glucose dehydrogenase (this will be discussed further later).

3.2.2 Second Generation of Glucose Biosensors

The limitations of the rst generation of blood glucose biosensors were overcome by the second generation of biosensors, for which mediated blood glucose
biosensors were developed. Varieties of nonphysiological electron acceptors
have been synthesized and can be utilized to facilitate electron transfer from the
enzyme reaction center to the electrode surface. These mediators include ferrocene derivatives,41,42 ferrocyanide,43 quinones,44,45 methylene blue,14 thionine, pyocyanine, safranine T15 and transition-metal complexes.46 A
comprehensive review paper by Heller and Feldman was published in 2008.47 In
the review paper, mediators are categorized into organic, inorganic, and metal
mediators. Each of these mediators possess its own attributes such as low
molecular weight, low redox potential to avoid interfering substances, high
stability, resistance to forming side compounds, and low toxicity, which is especially important for in vivo glucose biosensors.
In addition to the mediator, redox hydrogels were also used for shuttling the
electron from the enzyme reaction center to the electrode. In this case, enzymes
are immobilized in the redox hydrogels and the electron transfer is realized by
self-exchange between rapidly reduced and rapidly oxidized redox functions
tethered to backbones of a crosslinked polymer network.48 Because the redox
hydrogels envelop the redox enzymes, they electrically connect the enzyme
reaction centers to electrodes regardless the spatial orientation of the enzymes
at the electrode surface. They also connect multiple enzyme layers that lead to a

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

69

10-fold higher, or in some cases, even 100-fold higher current densities, than an
enzyme monolayer on the electrode surface.4952 The electron conductivity is at
its highest when the redox hydrogel is poised at its redox potential.
The rst research on the amperometric determination of blood glucose using
a redox couple-mediated and GOx-catalyzed reaction was reported in 1970.53
However, this study did not lead to a rapid adaptation of the technology in
home testing.54 Although the mediator reacts with the enzyme at a considerably
faster rate than oxygen, the dissolved oxygen may also compete with the mediator when FAD-GOx enzymes are present. In recent years, GDH, an oxygeninsensitive enzyme, has been widely used in the BGMS industry.
The rst electrochemical blood glucose monitor for diabetic patients home
testing was launched in 1987 as ExacTech by Medisense Inc. It was a pen-sized
monitor and used PQQ-GDH and a ferrocene derivative as a mediator.55 Its
success led to a revolution in diabetes management. Various BGMSs use redox
mediators coupled with either GOx or GDH enzymes. This strategy has been
widely commercialized.

3.2.3 Third Generation of Glucose Biosensor

The third-generation glucose biosensors are based on direct electron transfer
between the enzyme and the electrode without the need for toxic mediators.
Because the FADH2 is buried at a deep center, about 1315 A below the
electrode-contacting periphery of its glycoprotein,56 direct electron transfer
from FADH2 of GOx to an electrode surface is much slower. In recent years,
rapid advances in the development of nano and porous materials have greatly
increased electrode surface area and dynamics.5759 In 2003, Xiao et al.60 bound
the FAD site of GOx to 1.4 nm gold nanocrystals, showing that it is possible to
electro-oxidize glucose when the electrons are relayed to an electrode through a
gold nanoplug in the GOx. However, the contact to the electrodes was poor
and a high potential was required.
Other approaches used for direct electron transfer between enzyme and
electrode include the use of an overoxidized boron-doped diamond electrode
(BDDE)61 in which the carboxyl group on the electrode surface is chemically
crosslinked to the enzyme through glutaraldehyde.
Although the complications of tailored mediators are avoided in this
third generation of glucose sensors, the eciency of the electron conductivity,
selectivity and sensitivity of the sensors remain as challenges. Therefore, the
third-generation glucose biosensors are not yet adopted widely in the current
BGMS market.

3.3 Sensors for BGMS

Though many techniques are available for BGMS, including electrochemical,
optical, thermometric, piezoelectric and magnetic,62 the majority of current
glucose biosensors are electrochemistry based, using second generation blood
glucose biosensor technology. Enzymatic amperometric glucose biosensors are

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

70

Chapter 3

Figure 3.1

Commercially available enzymatic amperometric blood glucose monitoring Systems from major manufacturers (Bayer, Johnson and Johnson,
Roche and Abbott, left to right, top to bottom).

Figure 3.2

The structure of a basic amperometric BGMS sensor.

the most common devices commercially available, and have been widely
studied over the last few decades (Figure 3.1). As a result, we will focus on this
type of sensors in this section, and discuss in details of the construction of
modern BGMSs.

3.3.1 Electrode
Electrode substrate, working electrode, reference electrode and/or counterelectrode can be used as base electrodes for BGMSs. The structure of a BGMS
sensor is demonstrated in Figure 3.2.

3.3.1.1

Electrode Substrate

The base electrode for a BGMS is normally constructed on a thin piece of

plastic. This plastic material has to be thermally and chemically stable. It serves

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

Figure 3.3

71

Commercialized amperometric BGMS sensors (left to right: Bayer, Roche,

Johnson & Johnson and Abbott. Top: sensor strips; bottom: after sensors
are opened to view the inside.

as a foundation for the electrode on which the conductive materials will be either
screen printed or vapor deposited. As these processes may involve high temperature and organic solvents, the substrate materials should have a high glasstransition temperature and be inert to common organic solvents. The common
electrode substrates are polyethylene terephthalate (PET), polycarbonate,
polyimide, polyethylenenapthalate (PEN). polyethersulfone (PES) and cyclic
polyolen. The BGMS commercially available are illustrated in Figure 3.3.63

3.3.1.2

Working Electrode

The working electrode is the place on which the reaction of glucose oxidation
occurs. The glucose-derived electrons exit the sample and transfer to the meter.
Common working electrodes can consist of precious metals such as gold (Au)
and palladium (Pd), or carbon. While Au and Pt are normally coated onto a
plastic substrate by vapor deposition, a carbon working electrode is screen
printed from a carbon paste a mixture of carbon particles and a polyester
binder. Because the numbers of electrons generated from a working electrode
are directly related to its area, the working electrode area must be known and
kept constant during the manufacturing process. This area is generally dened
by an insulating dielectric layer or by laser abrasion (for metal electrodes). The

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

72

Chapter 3

distance between the working electrode and the reference/counterelectrode

is normally kept to a minimum in achieving for a small sample volume and
reduced interelectrode electrolytic resistance. Another design is to have working and reference electrodes positioned face to face for reduced sample
volume.64

3.3.1.3

Reference/Counterelectrode

Most commercially available BGMS test strips are two-electrode systems. The
reference and counterelectrodes are combined into a single electrode. The reference electrode ensures a constant potential applied onto the working electrode and the counterelectrode provides an electrical path for glucose-derived
electrons generated from the working electrode to be transferred to the meter.
The common reference/counterelectrode can consist of Au, Pd or Ag/AgCl.
The Au or Pd reference/counterelectrodes are normally made of using the same
material and process as the working electrode, and the area is dened by
removing a ne line of metal material at the edge of desired area using laser
abrasion. This practice minimizes the manufacturing step and thus reduces the
cost, as the manufacturer can make metal-coated plastic sheets in one step
and dene the working and reference/counterelectrode area by laser abrasion.
Ag/AgCl reference/counterelectrodes are usually formed by screen printing an
Ag/AgCl ink, which is made of Ag and AgCl particles with polyester binding
material.

3.3.1.4

Fluid-Detection Electrode

It is important that the same amount of blood sample reacts with the chemistry
within the reaction chamber. This is because the quantity of the enzyme inside
the reaction chamber is xed. If the sample volume was dierent every time, a
variable glucose reading would be expected. To monitor the sample volume
inside the reaction chamber, uid detector electrodes are often placed at the end
of the chamber.65,66 When these electrodes do not detect any uid, an error
code is typically displayed on the device to inform the end user that there is
insucient sample and there is a need to repeat the test with a new sensor.

3.3.2 Reaction Chamber

The reaction chamber is a miniature electrochemical cell. It is formed on the
top of the electrode with a spacer dening the edges and a cover plate facing the
electrode surface. The spacer used is normally a double-sided sticky material
that keeps the cover plate and the electrode together. The cover plate is normally transparent and can be treated with a wetting agent to ensure a fast
capillary reaction by which the blood sample can be drawn into the reaction
chamber quickly. The size of the reaction chamber corresponds to the product
sample volume.

View Online

Blood-Glucose Biosensors, Development and Challenges

73

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

3.3.3 Chemistry
Chemistry is at the heart of a glucose biosensor glucose is oxidized and the
signal generated is proportional to its concentration in a blood sample. In
general, the chemistry consists of electrolyte, polymer, surfactant, mediator and
enzyme. Buer salt is normally used as an electrolyte to provide conductivity
between electrodes. Hydrophilic polymer is present in the chemistry as a
thickening agent to adjust the viscosity of the chemistry so that it can be easily
and reproducibly dispensed. The polymer also prevents the chemistry from
being washed away from its deposition position when the blood sample is
pulled into the reaction chamber by capillary reaction. Surfactant as a wet
agent is often used to ensure a rapid chemical hydration. All commercialized
blood glucose sensors contain a mediator in the chemistry, which shuttles the
electron from the reaction center to the electrode surface. As a result, these
types of sensors can work at a relatively low potential that minimizes the eect
from common interference substances. The common redox mediators are potassium ferricynide and ruthenium complexes.
Enzymes are the main component of the chemistry. There are two types of
enzymes that have been used for the commercialized BGMS: glucose oxidase
and glucose dehydrogenase. Glucose oxidase is a very stable, specic and costeective enzyme. However, as mentioned in Section 3.2, the response of a
glucose sensor made of glucose oxidase is subject to inuence from the dissolved oxygen concentration in the sample. Because the BGMS is designed for
patient home testing and the oxygen-level dierence in a normal persons capillary blood is in a relatively narrow range, glucose oxidase is still a common
enzyme for some commercialized BGMSs. Nevertheless, this type of sensor
normally uses capillary or oxygenized venous blood to calibrate the system
during the product development. It needs extra precautions if the system is used
in a hospital or a clinical lab where venous or arterial blood samples are
available and could be tested.
The mechanism of the reaction using glucose oxidase is plotted in Figure 3.4.
Glucose is oxidized to gluconolactone, and glucose oxidase is reduced
to FADH-GOx. Then ferricyanide is reduced to ferrocyanide, and reduced
FADH-GOx is oxidized to FAD-GOx. Oxygen as an oxidizing agent competes

Figure 3.4

Mechanism of the reaction using glucose oxidase.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

74

Figure 3.5

Chapter 3

Mechanism of the reaction using GDH.

for the electron, thus articially reducing the glucose signal and providing a
false low glucose reading.
To reduce the oxygen eect, oxygen-insensitive enzymes such as pyrroloquinoline quinine glucose dehydrogenase (PQQ-GDH),6776 nicotinamide
adenine dinucleotide GDH (NAD-GDH)7784 and avin adenine dinucleotide
GDH (FAD-GDH)8595 have been widely used for glucose biosensors.
However, PQQ-GDH has been found to lack specicity. It not only reacts with
glucose but also with maltose, galactose and xylose, and produces falsely high
results. It is therefore advised that patients need extra caution to check the
medicine they take to make absolutely sure it will not interfere with their bloodglucose measurement. Dialysis patients should not use the blood-glucose sensor
that uses PQQ-GDH technology. This is because isosmotic peritoneal dialysis
solution contains glucose polymer (icodextrin) as the primary osmotic agent,
icodextrin is metabolized by alpha-amylase into oligosaccharides with a lower
degree of polymerization, including maltose, maltotriose and maltotetraose,
etc. In addition to the problem with the specicity of PQQ-GDH enzyme,
GDH-type enzyme is overall more expensive than GOx enzyme. The mechanism of the reaction using GDH is shown in Figure 3.5.

3.3.4 Detection Method

The most common electrochemical detection methods used in BGMSs are
amperometry and coulometry. The amperometric method measures the current
generated from the reaction at a potential, this current is proportional to the
glucose concentration in the sample. The coulometric method measures the
amount of coulombs produced from the reaction; the quantity of the coulombs
is related to the glucose concentration in the sample.

3.3.5 Correction Algorithm

The BGMS product is designed to test capillary blood samples. Depending on
the patients age, gender and health condition, there are variations in
hematocrit level, dissolved oxygen content and electrochemical interferents, etc.
Each of these variations can have a signicant eect on test accuracy and lead

View Online

Blood-Glucose Biosensors, Development and Challenges

75

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

to a biased result. To improve the accuracy, manufacturers normally employ a

series of algorithms to minimize these variations.

3.3.5.1

Hematocrit Correction

The hematocrit levels for male and female adults are in the range of 4254%
and 3846%, respectively. Because the BGMs measure the glucose that is only
in the liquid phase, and the higher the hematocrit the lower the liquid content,
samples with a higher hematocrit level will have a lower glucose reading than
those with a lower hematocrit level, even when the actual glucose concentrations in these samples are the same.96114
To correct the hematocrit variation, some BGMs systems have hematocrit
electrodes on board to simultaneously measure the glucose and hematocrit
levels of a testing sample; they apply correction algorithms to correct the
hematocrit eect before displaying a nal glucose result.115117 Other BGMs
systems rely on electrochemical methods118 and polymers in the chemistry119 to
correct or reduce hematocrit eect. However, the majority of commercial
available BGMs do not have hematocrit correction algorithm, instead, they
calibrate to a middle hematocrit level (i.e. 45%). In this case, samples with 45%
hematocrit give very accurate test results. Those with either higher or lower
hematocrit level give less accurate results. In general, samples with hematocrit
levels greater than 45% will generate a low glucose response and have a
negative bias, while samples with hematocrit levels less than 45% will have a
high glucose response and have a positive bias. Manufacturers usually optimize
the chemistry formulation to make the BGMs less sensitive to the hematocrit
level and maintain the assay accuracy within regulatory requirement.

3.3.5.2

Temperature Correction

In addition to hematocrit-correction algorithms, temperature-correction algorithms are also widely used to correct the temperature eect on the glucose
measurement as generically illustrated in Figure 3.6. The glucose biosensor is
an enzyme-based assay. The enzymatic reaction is very dependent on temperature. As a result, BGMs are normally equipped with temperature-detection
capacities. One common approach is to have one or two thermometer(s) embedded inside the handheld meter. When the meter is in thermal equilibrium
with the atmosphere where blood testing is performed, the temperature reading
from the thermometer is very close to the temperature of the glucose test strip.
However, when the meter is not equilibrated to the testing environment, temperature oset can cause inaccurate readings.120122
To overcome this problem, multiple approaches have been proposed. One
approach is to have an IR sensor near the glucose strip insertion port so that
the temperature of the glucose test strip can be monitored instantaneously, and
there will be no equilibrium time needed.123,124 However, the accuracy of the IR
sensor is subject to environmental interference such as breezes and vibration.
Another approach is to have a temperature sensor printed on the glucose strip

View Online

76

Chapter 3
40
Before correction
After Correction

20
% bias to 25 C

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

30

10
0
10

10

15

20

25

30

35

40

20
30
40

Temperature (degree)

Figure 3.6

Percentage bias before and after temperature correction.

and measure the temperature at the same time as the glucose. The nal
blood glucose results are calculated based on the temperature measurement
at the time.125,126 In addition, a detachable temperature measurement unit,127
AC/DC (alternative current/direct current) measurement128 and measuring
current dierence between the rst and second voltage129 have also been
proposed.

3.3.6 Calibration of BGMS

Millions and millions of glucose biosensors are manufactured every year. These
glucose test strips are randomly sampled and tested by the manufacturer before
they enter the market to characterize their analytical performance, including
linearity, coecient of variation, hematocrit and oxygen dependence, response
to electrochemical interferents and other factors. The manufacturing release
test is normally performed with spiked whole blood of low, medium and high
glucose concentrations covering the manufacturer claimed detection range. The
performance of the glucose test strips is compared with a reference method and
the dierence between the two is plotted. The Yellow Spring Instrument (YSI)
glucose analyzer is a commonly used reference method in the BGMS industry.
The glucose test strips will be released only if their performance meets the
acceptance criteria. Each glucose test strip batch has an assigned value that is
associated with its calibration information, such as slope and intercept. This
information is programmed into a chip that is shipped with the test strips. Users
are advised to check and make sure the number (codes) on a test strip bottle
matches with that of the calibration chip and always apply a new chip into the
meter when a new glucose test strip bottle is used.
Some BGMS systems have no coding technology. In this case, the
manufacturing process and incoming raw materials need to be extremely well

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

77

controlled. Manufacturing quality control is a key element in ensuring production of reproducible glucose test strips.
There are two types of BGMS systems: plasma calibrated and whole-blood
calibrated. This is because some countries measure glucose concentration in
whole blood while others measure glucose concentration in plasma. In theory,
plasma glucose readings are higher than whole-blood glucose readings for a
specic sample. The dierence coecient is normally in a range of 1.12 to 1.15
(i.e. plasma glucose 1.12 whole blood glucose), depending on the chemistry
of a glucose test strip. Because plasma glucose and whole-blood glucose readings
are interchangeable, the manufacturers usually perform glucose testing in either
plasma or whole-blood samples and convert to other readings when needed.

3.3.7 Performance Validation of BGMS

The BGMS is a user-friendly device. It is easy to operate, requires a very small
blood sample and provides test results in a few seconds. Even though the BGMS
is a home-monitoring device, the quality of its measurement is very important,
because many patients, especially Type 1 diabetic patients, rely on the glucose
reading to adjust their insulin intake. Therefore, there are a number of regulatory
recommendations, and analytical performance criteria are available to guide the
manufacturers during the development of their products. The International
Organization for Standardization (ISO) 15197:2003 recommends an allowable
error of  15 mg/dL for blood glucose levels o75 and  20% for blood glucose
levels Z75 mg/dl.130 In 1987, the American Diabetes Association (ADA) recommended a goal of total error of o10% at glucose concentrations between 30
and 400 mg/dL 100% of the time when compared to the reference laboratory
measurements.131 On March 1617 2010 the FDA and the toxicology devices
unit of the Center for Devices and Radiological Health (CDRH) held a public
meeting to discuss the issues with the current BGMS manufacturers and health
care professionals regarding whether the accuracy standard for the current systems was acceptable. At the end of a two-day meeting, FDA and other stakeholders appeared to be coming together; the consensus was that the overall
accuracy of BGMS should be at least  15% total error at the upper range
of glucose results, tighter than the current  20% standard presented in the
10-year-old ISO standard used by FDA to clear the BGMS products for market.
The ISO 15197:2003 has been widely used to evaluate BGMS analytical
performances including repeatability, precision, reliability and accuracy.
A detailed evaluation plan is listed below.

3.3.7.1

Repeatability

According to ISO 15197:2003, repeatability of a BGMS is evaluated at ve

blood glucose levels: 3050, 51110, 111150, 151250, and 251400 mg/dL.
The evaluation requires ten glucose meters to be used and equilibrated to 25 
11C and 50  10% RH (relative humidity). On each meter, 10 measurements
are taken. The mean value, the standard deviation (SD) and the coecient of

View Online

78

Chapter 3

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

variation (CV) for each meter from the ten measurements are calculated. The
grand mean, the pooled variance, the pooled SD (with 95% condence interval)
and the pooled CV are also calculated.

3.3.7.2

Precision

The precision of a BGMS is evaluated using control materials at three glucose

concentrations with 10 glucose meters. It requires one measurement per day per
sample over a 10-day period for each of 10 meters. The mean, SD and CV for
each meter from the ten measurements are calculated. The grand mean, the
pooled variance, the pooled SD (with 95% condence interval) and the pooled
CV are calculated. The pooled SD and CV are measures of the intermediate
precision of a single system over multiple days.

3.3.7.3

Reliability

The reliability evaluation is mainly to check the robustness of the BGMS. The
glucose meters are exposed to extreme conditions that the system claims to
operate under, such as lowest and highest temperatures and humidity. The
analytical performances before and after harsh conditions are compared, and
the dierences should be statistically insignicant. The glucose meters are also
evaluated for their mechanical strength using drop testing and vibration.

3.3.7.4

Accuracy

Based on the ISO 15197:2003 requirement, the accuracy of a BGMs is evaluated in seven dierent glucose categories using fresh-blood samples from diabetic patients, see Table 3.2. Because there is a strict requirement for the
number of samples per glucose category, it is very hard to obtain fresh-blood
samples at extreme glucose levels. Blood samples at very high and very low
glucose concentrations were created by either glycolyzing a blood sample to
lower the glucose concentration or supplementing a blood sample with a
concentrated glucose solution to increase the glucose level. All glucose measurements are compared with a reference value and the dierence between the
two is calculated.
Table 3.2

ISO 15197 required glucose distribution for

performance evaluation.

Glucose concentration (mg/dL)

Number of samples

o 50
5180
81120
121200
201300
301400
4400

5
15
20
30
15
10
5

View Online

Blood-Glucose Biosensors, Development and Challenges

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

3.3.7.5

79

Hematocrit Independence

A BGMS normally has its hematocrit (HCT) claim range stated in the product
insert. Depending on the chemistry and correction algorithm, hematocrit eects
can be signicantly dierent among dierent systems. In general, the higher the
HCT of a blood sample, the lower its glucose reading. To evaluate the HCT
independence, blood samples are altered by either removing the plasma to increase the HCT level or adding the plasma to decrease the HCT level to reach
the target HCT levels. Within the claimed HCT levels, all results are expected to
be within 15% of the reference value.

3.3.7.6

Interference

There are a number of variables that can interfere with the performance of a
BGMS. These variables can be categorized as related to the environment, patient and endogenous and exogenous substances. The environment-related
variables include temperature, humidity and altitude, while patient-related
variables include hypertension, hypoxemia, HCT, hydration, icterus, lipemia
and other health conditions. Endogenous and exogenous substance-related
variables are normally electrochemically active substances such as acetaminophen, ascorbic acid, uric acid, salicylic acid, bilirubin, tetracycline, dopamine,
ephedrine, ibuprofen, L-DOPA, methy-DOPA, tolazamide and other frequently administered diabetes drugs.
In addition to the aforementioned interference, enzymes used in the glucose
biosensor can induce inaccuracy as well. PQQ-GDH is known to be a nonspecic enzyme. It oxidizes not only glucose but also maltose, maltotriose,
maltotetraose, and icodextrin. Icodextrin is an osmotic agent, commonly used
for patients undergoing peritoneal dialysis.132,133 It is metabolized during systemic circulation into dierent glucose polymers, mainly maltose, which leads
to a signicant overestimation of glucose results.

3.3.8 Manufacturing Process

Millions of test strips are manufactured each year. Many dierent manufacturing methods have been explored and adopted to produce high-performance, cost-eective glucose test strips. Screen printing is the most common
manufacturing method. The conductive raw material is printed on the substrate
via a patterned screen.134137 The screen is made of either nylon or polyester134,135 or silk136 material having certain mesh count with patterns from
light-sensitive emulsion. Metal sputtering is another method to produce a
conductive layer on the substrate.138145 Conductive metals such as gold and
platinum, etc. are sputtered onto the substrate to cover a target area or the
entire sheet. A laser beam is used to create the electrodes as desired. There are
alternative methods of constructing testing strips. For example, one method
uses injection molding to make the substrate with holes, and conductive raw
materials are inserted into the holes to create electrodes.146,147

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

80

Chapter 3

Once the electrodes are formed, an insulation layer is printed on to dene the
sample reaction area. An enzyme solution is then deposited onto the surface of
the electrodes. Dierent methods have been reported for enzyme deposition,
including screen printing,147153 slot-die coating,154 spraying,155 dispensing156,157 and inkjet printing.158,159 Among these methods, screen printing and
dispensing are the most widely used methods in the BGMS industry.
A transparent lid with a reclosable adhesive backing is normally used to make a
complete BGMS glucose test strip. This reclosable adhesive backing is patterned
so that adhesives only exist in regions not corresponding to the reaction area.160
It is of the essence that the performance across glucose test strips is consistent
between inter- and intramanufacturing batches. Since a raw material lot can
only produce certain quantities of glucose test strips, it is inevitable that there
will be raw-material changes during production. Therefore, inspection of incoming materials is very important. Manufacturers are expected to set up tight
specications for each raw material prior to mass production. An inline
automatic inspection is normally conducted during production and defective
products are crossed out and removed from the good ones before the products
are packaged. During the production, if the response slope and intercept are
found to be outside of acceptable ranges, a quick adjustment can be made by
changing the working electrode area and reagent mediator amount.150

3.4 Clinical Utility and Potential Problems

Although the BGMS is designed only to monitor the patients blood glucose at
home and it is not best suited for the diagnosis of the diabetes or to make a
clinical decision, millions of diabetic patients constantly rely on the results of
the BGMS to adjust their insulin dosage, or to treat hypoglycemia. An error in
the BGMS reading can sometimes cause dreadful outcomes. It is the manufacturers responsibility to develop the most accurate BGMS that meets all
regulatory requirements. It is the healthcare personnels responsibility to recommend a BGMS system and to provide proper training to patients based on
individual circumstances. It is patients responsibility to operate the BGMS
system following the manufacturers use guide.
The performance of a BGMS system is extensively evaluated by the manufacturer before its commercialization. Clinical trials are also conducted to ensure the system is easy to use and provide reliable blood glucose results in the
hands of both healthcare professionals and patients. However, controversial
results were reported on the real-world accuracy of the BGMS.161166 The
studies conducted by the manufacturers or healthcare professionals are normally carried out in a controlled environment representing the best situation,
while extreme environmental conditions or abnormal situations could occur in
the real world.
As we know, the embedded temperature sensor in the BGMS only measures
the temperature inside the meter, not the testing environment. A biased result
can occur when the BGMS is not equilibrated to the ambient temperature. For
example, some patients like to keep their BGMS in their pocket, or leave it in

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

81

the car during hot summer or cold winter days and then conduct a test right
away without waiting a required 1520 min for the meter to be equilibrated to
the ambient temperature, as suggested by the manufacturer.
The BGMS glucose test strips have a nite lifetime when stored at
manufacturer-specied temperature and humidity. Their shelf life can be
shortened if the vial is left open for a certain period of time or the test strips are
exposed to extreme temperature and humidity. A biased result can be obtained
when expired or environmentally stressed test strips are used.
Most BGMSs have alternative site-testing options, in which blood testing
can be done in areas such as the palm, forearm and upper arm. Because the skin
of these areas contains fewer nerves than the ngertip, alternative site testing
could be less painful and may be more acceptable to patients. However, the
ngers contain a more concentrated network of blood vessels as compared to
the forearm, which allows the blood in the nger to adjust more quickly to
rapid changes in glucose. This means that readings can be delayed at alternate
sites during times of rapidly changing blood glucose, making it more dicult to
identify hyperglycemia or hypoglycemia. A 1530-min lag is often observed at
these sites when the glucose level is not in a stable status.
In addition, miscoding the meter, testing without washing ones hands,
leaving the lid open after taking the test strip out and using expired test strips
are all common mistakes that could happen in the real world. Certain substances in patients blood specimen such as triglycerides, uric acid, oxygen,
ascorbic acid (vitamin C) and hematocrit as well as drug residues and
the products of their metabolism will also interfere with the accuracy of
the BGMS.

3.5 Continuous Glucose Monitoring System (CGMS)

The BGMS has been a very useful tool in diabetes management. However, it
can only provide discreet glucose information at the time of the test; the glucose
trend in the next few minutes is still unknown. As a result, eective actions
cannot be taken to correct an upcoming hyperglycemia or treat an impending
hypoglycemia. To compensate the shortage of BGMSs, continuous monitoring
of blood glucose was proposed in 1974.167 Shichiri et al.168 demonstrated the
rst in vivo needle-type enzymatic Continuous Glucose Monitoring System
(CGMS) in 1982. Due to biofouling issues from the protein and coagulation
factors and the risk of thromboembolism, subcutaneous glucose monitor approaches that reect the blood glucose level have been adapted over the continuous blood glucose monitor.169173 The rst commercial implantable needletype biosensor that measured the glucose concentration in interstitial uid was
marketed by Minimed. However, it did not provide real-time data; the results
could only be viewed retrospectively at a physicians oce after downloading.174 Today, there are three FDA-approved needle-type real-time CGMS
devices available, including the Minimed Guardian by Medtronic, SEVEN by
Dexcom and Freestyle Navigator by Abbott (although Abbott discontinued
Freestyle Navigator in 2011). These devices display the real-time glucose

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

82

Chapter 3

concentration every one to ve minutes and use the disposable sensors that can
be changed every three to seven days.175
Although the CGMS displays real-time glucose concentrations and provides
a glucose trend, the FDA has not approved use of the CGMS in place of the
BGMS. Many improvements are required, including accuracy, biocompatibility, calibration, long-term stability, specicity, linearity and miniaturization.
Despite the improvements needed for the CGMS, researchers are looking into
closed-loop systems where a CGMS and an insulin pump are connected together to form an articial pancreas. The insulin pump will deliver extra insulin
when the patients blood glucose level is elevated, and switch o and send an
alert when the patient experiences hypoglycemia. Though it may take a great
deal of eort and many years to realize the articial pancreas, future prospects
for diabetes management are very exciting.

References
1. J. E. Shaw, R. A. Sicree and P. Z. Zimmet, Diabet. Res. Clin. Pract., 2010,
87, 4.
2. N. Poolsup, N. Suksomboon and S. Rattanasookchit, Diabet. Technol.
Ther., 2009, 11, 775.
3. G. H. Murata, J. H. Shah, R. M. Homan, C. S. Wendel, K. D. Adam,
P. A. Solvas, S. U. Bokhari and W. C. Duckworth, Diabet. Care, 2003,
26, 1759.
4. S. Skeie, G. B. Kristensen, S. Carlsen and S. Sandberg, J. Diabet. Sci.
Technol., 2009, 3, 83.
5. S. L. Tunis and M. E. Minshall, Curr. Med. Res. Opin., 2010, 26, 151.
6. E. I. Boutati and S. A. Raptis, Diabet. Care, 2009, 32(Suppl. 2), S205.
7. L. G. Jovanovic, Endocr. Pract., 2008, 14, 239.
8. M. J. OKane and J. Pickup, Ann. Clin. Biochem., 2009, 46, 273.
9. The Diabetes Control and Complications Trial Research Group, N. Engl.
J. Med., 1993, 329, 977.
10. R. R. Holman, S. K. Paul, M. A. Bethel, D. R. Matthews and H. A. Neil,
N. Engl. J. Med., 2008, 359, 1577.
11. I. M. Stratton, A. I. Adler, H. A. Neil, D. R. Matthews, S. E. Manley,
C. A. Cull, D. Hadden, R. C. Turner and R. R. Holman, BMJ, 2000,
321, 405.
12. American Diabetes Association, Diabet. Care, 2010, 33, S11.
13. J. D. Newman and A. P. Turner, Biosens. Bioelectron., 2005, 20, 2435.
14. O. Warburg and W. Christian, Biochem. Z., 1932, 254, 438.
15. W. Franke and M. Dener, Anal. Chem., 1939, 541, 117.
16. H. V. Malmstadt and H. L. Pardue, Analyt. Chem., 1961, 33, 1040.
17. H. L. Pardue, R. K. Simon and H. V. Malmstadt, Analyt. Chem., 1964,
36, 735.
18. M. I. Drury, F. J. Timoney and P. Delaney, J. Ir. Med. Assoc., 1965,
56, 52.
19. M. S. Jensen, Ugeskr-Laeger, 1965, 127, 709.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.

42.
43.
44.
45.
46.
47.
48.

83

W. Korp, Wien. Med. Wochenschr., 1965, 115, 435.

V. Schmidt, Ugesk-Laeger, 1965, 127, 706.
B. C. Clark and C. Lyons, Ann. New York Acad. Sci., 1962, 556, 46.
S. J. Updike and G. P. Hicks, Nature, 1967, 214, 986.
S. J. Updike and G. P. Hicks, Science, 1967, 158, 270.
G. G. Guilbault and G. J. Lubrano, Anal. Chim. Acta, 1973, 64, 439.
K. E. Toghill and R. G. Compton, Int. J. Electrochem. Sci., 2010, 5,
1246.
W. Z. Jia, K. Wang and X. H. Xia, Trends Analyt. Chem., 2010, 29, 306.
J. Anzai, H. Takeshita, Y. Kobayashi, T. Osa and T. Hoshi, Analyt.
Chem., 1998, 70, 811.
I. Mao, P. G. Osborne, K. Yamamoto and T. Kato, Analyt. Chem., 2002,
74, 3684.
A. Collins, E. Mikeladze, M. Bengtsson, M. Kokaia, T. Laurell and
E. Csoeregi, Electroanalysis (NY), 2001, 13, 425.
G. Cui, S. J. Kim, S. H. Choi, H. Nam and G. S. Cha, Analyt. Chem.,
2000, 72, 1925.
S. H. Choi, S. D. Lee, J. H. Shin, J. Ha, H. Nam and G. S. Cha, Anal.
Chim Acta, 2002, 461, 251.
J. H. Shin, Y. S. Choi, H. J. Lee, S. H. Choi, J. Ha, I. J. Yoon, H. Nam
and G. S. Cha, Analyt. Chem., 2001, 73, 5965.
I. J. Xu, X. L. Luo, Y. Du and H. Y. Chen, Electrochem. Commun., 2004,
6, 1169.
J. C. Armour, J. Y. Lucisano, B. D. McKean and D. A. Gough, Diabetes,
1990, 39, 1519.
R. Pauliukaite, M. Schoenleber, P. Vadgama and C. Brett, Anal. Bioanal.
Chem., 2008, 390, 1121.
P. Vadgama, I. Christie, Y. Benmakroha and S. Reddy, WO 9402585 A1
1994.
P. Vadgama, S. Reddy and M. Kyrolainen, WO 9717607 A1 1997.
J. Wang and F. Lu, J. Am. Chem. Soc., 1998, 120, 1048.
J. Wang, J. W. Mo, S. Li and J. Poter, Anal. Chim. Acta, 2001, 441, 183.
A. E. G. Cass, G. Davis, G. D. Francis, H. A. O. Hill, W. J. Aston,
I. J. Higgins, E. V. Plotkin, L. D. L. Scott and A. P. F. Turner, Analyt.
Chem., 1984, 56, 667.
A. Mulchandani and S. Pan, Anal. Biochem., 1999, 267, 141.
S. Tsujimura, S. Kojima, K. Kano, T. Ikeda, M. Sato, H. Sanada and
H. Omura, Biosci., Biotechnol., Biochem., 2006, 70, 654.
M. G. Loughran, J. M. Hall and A. P. F. Turner, Electroanalysis, 1996,
8, 870.
K. Lau, S. A. L. de Fortescu, L. J. Murphy and J. M. Slater, Electroanalysis, 2003, 15, 975.
C. Taylor, G. Kenausis, I. Katakis and A. Heller, J. Electroanal. Chem.,
1995, 396, 511.
A. Heller and B. Feldman, Chem. Rev., 2008, 108, 2482.
A. Aoki and A. Heller, J. Phys. Chem., 1993, 97, 11014.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

84

Chapter 3

49. L. Ye, M. Haemmerle, A. J. J. Olsthoorn, W. Schuhmann, H. L. Schmidt,

J. A. Duine and A. Heller, Analyt. Chem. Soc., 1993, 65, 238.
50. F. Mao, N. Mano and A. Heller, J. Am. Chem. Soc., 2003, 125, 4951.
51. N. Mano, F. Mao and A. Heller, Chem. Commun., 2004, 18, 2116.
52. N. Mano, F. Mao and A. Heller, J. Electroanal. Chem., 2005, 574, 347.
53. D. L. Williams, A. R. Doig and A. Korosi, Analyt. Chem., 1970, 42, 118.
54. J. Hu, Biosens. Bioelectron., 2009, 24, 1083.
55. D. R. Matthews, R. R. Holman, E. Bown, J. Steemson, A. Watson,
S. Hughes and D. Scott, Lancet, 1987, 1, 778.
56. H. J. Hecht, H. M. Kalisz, J. Hendle, R. D. Schmid and D. Schomburg,
J. Mol. Biol., 1993, 229, 153.
57. Y. Wang and F. Caruso, Chem. Commun. (Cambridge, UK), 2004, 1528.
58. S. Wu, H. Ju and Y. Liu, Adv. Funct. Mater., 2007, 17, 585.
59. S. Bao, C. M. Li, J. Zang, X. Cui, Y. Qiao and J. Guo, Adv. Funct. Mater.,
2008, 18, 591.
60. Y. Xiao, F. Patolsky, E. Katz, J. F. Hainfeld and I. Willner, Science, 2003,
299, 1877.
61. J. Wu and Y. Qu, Anal. Bioanal. Chem., 2006, 385, 1330.
62. J. D. Newman and A. P. Turner, Biosensors: Principles and Practice;
Portland Press: London, UK, 1992, 27, 147.
63. R. Heydarpour, M. K. Shi and K. Kian and K. Tsai, US 20020110673,
2002.
64. J. F. Benjamin, A. Heller, E. Heller, F. Mao, J. A. Vivolo, J. V.
Funderburk, F. C. Colman and R. Krishnan, US 6299757 B1, 2001.
65. I. Harding, S. Diamond, R. Williams and S. G. Iyengar, US 7547382 B2
2009.
66. B. W. Bae, B. S. Kang, S. G. Park, S. D. Lee and M. J. Kwon, US 7297248
B2, 2007.
67. W. Jin, U. Wollenberger and F. W. Scheller, Biol. Chem., 1998, 379, 1207.
68. M. Zayats, E. Katz, R. Baron and J. Willner, Am. Chem. Soc., 2005,
127, 12400.
69. O. A. Raitman, F. Patolsky, E. Katz and I. Willner, Chem. Commun.
(Camb.), 2002, 17, 1936.
70. M. Alkasrawi, I. C. Popescu, B. Mattiasson, E. Csoregi and
V. Laurinavicius, Anal. Commun., 1999, 36, 395.
71. A. Malinauskas, J. Kuzmarskyte, R. Meskys and A. Ramanavicius, Sens.
Actuators, B: Chem., 2004, B100, 387.
72. A. Suzumura, T. Koyama and S. Maeda, JP 2005333872 A, 2005.
73. C. Zhao and G. Wittstock, Biosens. Bioelectron., 2005, 20, 1277.
74. J. Razumiene, J. Barkauskas, V. Kubilius, R. Meskys and
V. Laurinavicius, Talanta, 2005, 67, 783.
75. G. Li, H. Xu, W. Huang, Y. Wang, Y. Wu and R. Parajuli, Measure. Sci.
Technol., 2008, 19, 065203/1.
76. T. Yoshioka, Chem. Sens., 2009, 25(Suppl. B), 25.
77. J. Kulys, L. Tetianec and I. Bratkovskaja, Biotech. J., 2010, 5, 822.
78. P. N. Bartlett and R. G. Whitaker, Biosensors, 1987, 3, 359.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

79.
80.
81.
82.
83.

84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.

102.
103.

104.

85

P. N. Bartlett, E. Simon and C. S. Toh, Bioelectrochemistry, 2002, 56, 117.

L. Gorton and E. Dominguez, J. Biotechnol., 2002, 82, 371.
E. J. DCosta, I. J. Higgins and A. P. Turner, Biosensors, 1986, 2, 71.
M. Zhang, C. Mullens and W. Gorski, Analyt. Chem., 2007, 79, 2446.
L. I. Boguslavsky, L. Geng, I. P. Kovalev, S. K. Sahni, Z. Xu,
T. A. Skotheim, V. Laurinavicius, B. Persson and B and L. Gorton,
Biosens. Bioelectron., 1995, 10, 693.
S. Kojima, S. Tsujimura, K. Kano, T. Ikeda, M. Sato, H. Sanada and
H. Omura, Chem. Sens., 2004, 20(Suppl. B), 768.
A. Hiratsuka, K. Fujisawa and H. Muguruma, Anal. Sci., 2008, 24,
483.
H. Kawaminami, M. Kitabayashi and Y. Nishiya, US 20120171708 A1,
2012.
H. Shimizu and W. Tsugawa, Electrochemistry (Tokyo, Japan), 2012,
80, 375.
H. Inoue, Y. Akasaka and K. Shirosaki, WO 2011118758 A1, 2011.
K. Nishio and H. Ido, WO 2011007792 A1, 2011.
K. Yuuki and Y. Nakanishi, PCT Int. Appl., WO 2009069381 A1, 2009.
W. Tsugawa and K. Sode, PCT Int. Appl., WO 2009037838 A1, 2009.
M. Kitabayashi, Y. Tsuji, Y. Nishiya and T. Kishimoto, US 20080090278
A1, 2008.
N. Tanaka and K. Yuuki, WO 2007139013 A1, 2007.
M. Kitabayashi, Y. Tsuji, H. Aiba, H. Kawaminami, T. Kishimoto and
Y. Nishiya, WO 2007116710 A1, 2007.
S. Tsujimura, S. Kojima, K. Kano, T. Ikeda, M. Sato, H. Sanada and
H. Omura, Biosci. Biotech. Biochem., 2006, 70, 654.
S. Kos, A. van Meerkerk, J. van der Linden, T. Stiphout and R. Wulkan,
Clin. Chem. Lab. Medic., 2012, 50, 1573.
R. W. A. Peake, M. S. Devgun and I. A. D. OBrien, Brit. J. Biomed. Sci.,
2012, 69, 41.
W. J. Cheng, C. W. Lin, T. G. Wu, C. S. Su and M. S. Hsieh, Clin. Chim.
Acta, 2013, 415, 152.
K. Wiener, Diabet. Med. and J. Brit., Diabet. Assoc., 1991, 8, 172.
M. Daves, R. Cemin, B. Fattor, G. Cosio, G. L. Salvagno, F. Rizza and
G. Lippi, Biochem. Med., 2011, 21, 306.
I. Jday-Daly, C. Augereau-Vacher, C. De Curraize, M. Fonfrede,
G. Lefevre, B. Lacour and C. Hennequin-Le Meur, Annal. Biol. Clin.,
2011, 69, 55.
S. Sano, T. Kishi and M. Gotoh, Tonyobyo (Tokyo, Japan), 2009,
52, 895.
K. Hirose, S. Inoue, S. Miyahara, K. Fujisawa, K. Misawa, S. Kubota,
T. Sakuma, W. Yumita and K. Nakajima, Iryo to Kensa Kiki-Shiyaku,
2009, 32, 423.
F. Yakushiji, H. Fujita, H. Suzuki, R. Joukyu, M. Yasuda, Y. Terayama,
K. Nagasawa, A. Ohwada, K. Taniguchi, K. Fujiki, M. Shimojo and
H. Kinoshita, Diabet. Tech. Therapeut., 2009, 11, 369.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

86

Chapter 3

105. R. J. Slingerland, W. Muller, J. T. Meeues, I. van Blerk, C. Gouka-Tseng,

R. Dollahmoersid, C. Witteveen and K. Vroonhof, Nederlands Tijdschrift
voor Klinische Chemie en Laboratoriumgeneeskunde, 2007, 32, 202.
106. Y. Mitsuyoshi, M. Noguchi, K. Ohtake, A. Shibata, H. Kikuchi and
M. Murata, Iryo to Kensa Kiki-Shiyaku, 2008, 31, 411.
107. M. J. Kwon and S. Y. Lee, Daehan Jindan Geomsa Ui Haghoeji, 2008,
28, 8.
108. E. S. Kilpatrick, A. G. Rumley, H. Myint, M. H. Dominiczak and
M. Small, Annals Clin. Biochem., 1993, 30, 485.
109. W. J. Cheng, C. W. Lin, T. G. Wu, C. S. Su and M. S. Hsieh, Clin. Chim.
Acta; Int. J. Clin. Chem., 2013, 415, 152.
110. B. Solnica, J. Skupien, B. Kusnierz-Cabala, K. Slowinska-Solnica,
P. Witek, A. Cempa and M. T. Malecki, Clin. Chem. Lab. Med.: CCLM/
FESCC, 2012, 50, 361.
111. M. Daves, R. Cemin, B. Fattor, G. Cosio, G. L. Salvagno, F. Rizza and
G. Lippi, Biochemia medica: casopis Hrvatskoga drustva medicinskih
biokemicara/HDMB, 2011, 21, 306.
112. B. H. Ginsberg, J. Diabet. Sci. Tech., 2009, 3, 903.
113. Z. Tang, J. H. Lee, R. F. Louie and G. J. Kost, Arch. Path. Lab. Med.,
2000, 124, 1135.
114. P. B. Barreau and J. E. Buttery, Diabet. Care, 1988, 11, 116.
115. L. V. Rao, F. Jakubiak, J. S. Sidwell, J. W. Winkelman and M. L. Snyder,
Clin. Chim. Acta, 2005, 356, 178.
116. E. Yamanishi and H. Tokunaga, JP 2005147990 A, 2005.
117. M. Teodorczyk, M. Cardosi and S. Setford, J. Diabet. Sci. Tech., 2012,
6, 648.
118. P. B. Musholt, C. Schipper, N. Thome, S. Ramljak, M. Schmidt, T. Forst
and A. Pfutzner, J. Diabet. Sci., 2011, 5, 1167.
119. Y. S. Yu, US 5789255 A, 1998.
120. K. Nerhus, P. Rustad and S. Sandberg, Diabet. Tech. Therapeut., 2011,
13, 883.
121. M. J. Haller, J. J. Shuster, D. Schatz and R. J. Melker, Diabet. Tech.
Therapeut., 2007, 9, 1.
122. Y. Takahashi, M. Tomioka, H. Ohkawa, M. Sanaka and Y. Iwamoto,
Igaku to Yakugaku, 2000, 44, 109.
123. M. Z. Kermani, S. Noujaim and T. Jetter, US 20100130838 A1, 2010.
124. H. Haar, WO 2001033214 A2, 2001.
125. S. Chinnayelka, J. Fei, Y. Wang, N. Parasnis, H. C. S. Sun, R. Giord
and S. Charlton, US 20110312004 A1, 2011.
126. J. H. Eom, KR 980316 B1, 2010.
127. M. S. Park, M. S. Cha, Y. C. Ahn and Y. B. Choi, WO 2010032911 A1,
2010.
128. D. W. Burke, L. S. Kuhn, T. A. Beaty and V. Svetnik, US 20040157338
A1, 2004.
129. M. Z. Kermani and M. Teodorczyk, US 2012012341 A1, 2012.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

Blood-Glucose Biosensors, Development and Challenges

87

130. In vitro diagnostic test systems Requirements for blood-glucose monitoring

systems for self-testing in managing diabetes mellitus, International Organization for Standardization: Geneva, Switzerland, 2003.
131. American Diabetes Association. Consensus statement on self-monitoring
of blood glucose. Diabet. Care, 1987, 10, 93.
132. W. Janssen, G. Har, M. Caers and A. Schellekens, Clin. Chem., 1998,
44, 2379.
133. S. O. Oyibo, G. M. Pritchard, L. McLay, E. James, I. Laing, R. Gokal
and A. J. Boulton, Diabet. Med., 2002, 19, 693.
134. Yasuo Nippon Tokushu Fabric Corp. Emori Ltd; Kenji Shin-Etsu
Chemical Co. Fushimi Ltd; KeiChi Akebono Sangyo Co. Ishikawa Ltd;
Toru Shin-Etsu Chemical Co. Nakanishi Ltd and Susumu Shin-Etsu
Chemical CO., EP 0208616 A1,1986.
135. U. Susumu and N. Toru, EP 0220121 A2, 2012.
136. Y. S. Choi, US 2003/0024811 A1, 2003.
137. A. Joseph, WO 2012025208 A1, 2012.
138. M. Chira, N. Jiko, H. Matsuzaki and K. Takagi, EP 2479312 A1, 2012.
139. T. Hariu, M. Inoue and K. Watanabe, EP 0376709 A3, 1990.
140. K. Inoue and S. Matsuzaki, EP 1431414 A1, 2004.
141. R. A. Emigh and W. B. Willett US 5687600, 1994.
142. T. A. Lee, J. S. Chen, C. H. Yeh, L. H. Chen, C. H. Chao and H. H. Chen,
US 2002/0083571 A1, 2002.
143. K. Hasegawa and N. Ishii, US 8252127 B2, 2011.
144. A. Inoue, H. Kimura, A. Nakamura, K. Sasamori, H. Takahashi and
M. Yahagi, EP 1652960 A1, 2004.
145. C. M. Huang, US 20040256228 A1, 2004.
146. C. T. Hsu, H. H. Chung, D. M. Tsai, M. Y. Fang, H. C. Hsiao and
J. M. Zen, Analyt. Chem., 2009, 81, 515.
147. M. F. Cardosi and M. OConnell, EP 2243841 A1 2010.
148. W. Tuan, EP 2365329 A1, 2011.
149. Z. C. Chen, C. Fang, H. Y. Wang and J. S. He, Electrochim. Acta, 2009,
55, 544.
150. G. Mace, R. Marshall, M. Alvarez-Icaza, K. Delaney, K. Duus,
T. Jarvis, M. Lamacka, N. Phippen and R. Russell, WO 2009090392 A1,
2009.
151. J. Marcinkeviciene and J. Kulys, Biosens. Bioelectron., 1993, 8, 209.
152. C. A. Galan-Vidal, J. Munoz, C. Dominguez and S. Alegret, Sens.
Actuators, B: Chem., 1998, B52, 257.
153. O. W. H. Davies, J. F. McAleer and R. M. Yeudall, WO 2001073109 A2,
2001.
154. C. D. Wilsey and D. Mosoiu, WO 2004113917 A2, 2004.
155. H. Yoshida, T. Baba and K. Shirasaki, JP2011099693 A, 2011.
156. J. M. Gleisner, CA 2142026 A1, 1996.
157. H. Sakamoto, Y. Takahashi, Y. Higuchi and T. Yamaguchi, JP 09101297
A, 1997.

View Online

14/10/2013 09:27:49.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00065

88

Chapter 3

158. O. Frey, P. D. van der Wal, N. F. de Rooij and M. Koudelka-Hep,

Procedia Chem., 2009, 1, 281.
159. Y. S. Chao, TW 291900 B, 2008.
160. S. Wooldridge and C. Bolam, US 20100051455 A1, 2010.
161. N. Lalic, T. Tankova, M. Nourredine, C. Parkin, U. Schweppe and
I. Amann-Zalan, Diabet. Tech. Therap., 2012, 14, 338.
162. G. E. Ritchie, A. P. Kengne, R. Joshi, C. Chow, B. Neal, A. Patel and
S. Zoungas, Diabet. Care, 2011, 34, 44.
163. G. V. McGarraugh, W. L. Clarke and B. P. Kovatchev, Diabet. Tech.
Therap., 2010, 12, 365.
164. A. Pfutzner, M. Mitri, P. B. Musholt, D. Sachsenheimer, M. Borchert,
A. Yap and T. Forst, Curr. Med. Res. Opin., 2012, 28, 525.
165. B. Hunsley and W. Ryan, J. Diabet. Sci. Tech., 2007, 1(2), 173.
166. B. H. Ginsberg, J. Diabet. Sci. Tech., 2009, 3, 903.
167. A. M. Albisser, B. S. Leibel, T. G. Ewart, Z. Davidovac, C. K. Botz,
W. Zingg, H. Schipper and R. Gander, Diabetes, 1974, 23, 397.
168. M. Shichiri, R. Kawamori, Y. Yamasaki, N. Hakui and H. Abe, Lancet,
1982, 2, 1129.
169. D. S. Bindra, Y. Zhang, G. S. Wilson, R. Sternberg, D. R. Thevenot,
D. Moatti and G. Reach, Analyt. Chem., 1991, 63, 1692.
170. E. Csoregi, D. W. Schmidtke and A. Heller, Analyt. Chem., 1995,
67, 1240.
171. C. Henry, Analyt. Chem., 1998, 70, 594A.
172. D. W. Schmidtke, A. C. Freeland and A. Heller, Proc. Natl. Acad. Sci.
USA, 1998, 95, 294.
173. K. Rebrin and G. M. Steil, Diabet. Tech. Ther., 2000, 2, 461.
174. T. M. Gross, B. W. Bode, D. Einhorn, D. M. Kayne, J. H. Reed,
N. H. White and J. J. Mastrototaro, Diabet. Techn. Ther., 2000, 2, 49.
175. M. Cox, J. Pediatr. Health Care, 2009, 23, 344.

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

CHAPTER 4

Recent Progress in the

Electrochemical Detection
of Disease-Related Diagnostic
Biomarkers
ALINA VASILESCU,a WOLFGANG SCHUHMANNb AND
SZILVESZTER GASPAR*a
a

International Centre of Biodynamics, 1B Intrarea Portocalelor,

060101 Bucharest, Romania; b Analytische Chemie-Elektroanalytik & Sensorik,
Ruhr-Universitat Bochum, Universitatsstr. 150, Bochum D-44780, Germany
*Email: [email protected]

4.1 Introduction
The European market of in vitro diagnostics has grown 24% annually between
2006 and 2010, and 0.9% from 2010 to 2011. The market revenues generated by
the European in vitro diagnostics industries were of h 10847 million in 2011 (see
European In Vitro Diagnostic Market Statistics Report 2011 by the European
Diagnostic Manufacturers Association). Although a decline was anticipated for
2012 (due to the economic recession), it is clear that an increasing number of
in vitro diagnostic tests are available and used in order to correctly diagnose
and manage diseases. Many of these tests are carried out in specialized
laboratories, on a wide range of complex instruments, and with high costs
(in terms of both money and time). Great eort is currently directed towards
making diagnostic tests faster and cheaper and also towards making some of
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

89

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

90

Chapter 4

them directly usable by doctors (for point-of-care diagnostics) or even patients

(at home). Electrochemical sensors are usually characterized by high sensitivities, they can easily be miniaturized and then integrated into microuidic
systems, and they require no complex instrumentation. Therefore, there is great
enthusiasm concerning the use of electrochemical sensors in disease diagnostics
(at least when judged by the large number of papers published on this matter).
Electrochemical sensors are solid/liquid interfaces characterized by electrical
currents or potentials that are dependent on the presence and the concentration
of an analyte of interest. Simple interfaces (such as those between bare electrodes and solutions) have a current or potential output that is dependent on
many factors. Therefore, in order to develop an electrochemical sensor, the
solid/liquid interface must be carefully modied to maximize the eect of the
analyte of interest and minimize the eect of other factors. One elegant way of
achieving the desired selectivity for electrochemical sensors is based on proteins
that are able to selectively bind and/or convert their substrate. As we will see in
the next sections, the selectivity of antibodies and/or enzymes can be added to
electrodes by immobilizing these biomolecules onto the electrodes. Even
though immobilized on the electrode, these biomolecules perform almost as in
their natural environment, that is, they recognize their substrate even from
mixtures with other compounds. Having a good and selective electrochemical
biosensor is thus possible if the physicochemical changes associated with the
recognition event are sensitively converted into currents or potentials.
In order to detect the minute (physical or chemical) changes associated with
the recognition event, several electrochemical methods are in use. These
methods are identied by names that might not say a lot for nonelectrochemists
(e.g. square-wave voltammetry, SWV) but in the end all of them consist of
applying a potential or current program (such as sweeps, pulses, etc.) and
measuring the resulting current or potential response (continuously or intermittently). Fundamentals of electrochemical methods will not be very much
detailed in the coming paragraphs. Details can be found elsewhere.1 Details
about electrochemical impedance spectroscopy (EIS, a method that will be
often mentioned in the coming sections), are found also in a very recent review
together with details about its use in bioanalysis.2
Often, the sensitivity of the electrochemical method alone is not sucient to
observe the interaction between the analyte and the selected biorecognition
element. In these situations dierent amplication methods are used. The most
common of these methods is the use of an enzyme label in the anity interaction between an antibody and an antigen (exactly as in enzyme linked
immuno sorbent assay, ELISA). By using the enzyme label, because each
immunocomplex will generate hundreds of thousands of detectable molecules,
the detection of really low concentrations of analyte is possible.
The nature of the sensor material also signicantly impacts the sensitivity of
electrochemical detection. Advantages brought in recent years by nanomaterials
were rationalized in terms of high surface/volume ratio and very good electrocatalytic properties (e.g. carbon nanotubes3,4) or high adsorption capacity
of biomolecules with preservation of their activity (e.g. Au nanoparticles4).

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

Figure 4.1

91

Schematic representation of the main elements and working principle of

electrochemical biosensors for disease diagnostics. Such sensors are often
quite similar to a sandwich-type ELISA. However, the capture recognition element is immobilized onto electrodes instead of the plastic plates
used in ELISA. Moreover, the anity binding between the analyte and the
recognition element is electrochemically observed with or without the use
of a second, labelled, detection biorecognition element. Higher sensitivities and shorter analysis times are usually obtained as compared to
ELISA.

The schematic representation of the elements of the most common electrochemical biosensors developed for disease diagnostics is shown in Figure 4.1.
The present chapter will detail representative examples of electrochemical
sensors for the diagnosis of seven, often-targeted diseases: cancer, heart diseases, acquired immunodeciency syndrome (AIDS), hepatitis, rheumatoid
arthritis (RA), celiac disease, and urinary tract infection (UTI). It will also
highlight the challenges still faced by such sensors. A relatively new and
thorough review on electrochemical sensors and biosensors, but not necessarily
for disease diagnostics, can be found elsewhere.5 Biosensors in clinical chemistry, but not necessarily electrochemical, were also reviewed elsewhere.6,7
Electrochemical sensors that make use of nanomaterials and target biomedical applications (including diagnostics) were also recently reviewed elsewhere.810 Some other reviews put the emphasis on practical aspects of the
research on electrochemical biosensors towards point-of-care diagnostics,
including mainly data from published clinical trials.1113

View Online

92

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

4.2 Electrochemical Sensors for Detection of Cancer

Biomarkers
There is no single test that can be used to undoubtedly diagnose cancer.
Examination of tissue and cell samples (e.g. collected by biopsy), scans (X-ray,
magnetic resonance imaging, etc.), various blood tests (e.g. complete blood
count), and history and physical condition examination are combined with
tumor marker tests instead. Regarding the tumor marker tests, signicant
eorts are currently directed towards identifying new markers as well as
towards new analytical methods for the detection of these markers. The search
for new markers is motivated by the lack of specicity of many current
markers. Some of the markers show up in conditions other than cancer and/or
are not specic for a certain cancer. The search for new analytical methods is
driven by the lack of sensitivity, the high cost, the high complexity and/or the
long response time of the currently existing methods to detect markers. The
present section oers an overview of cancer biomarkers for which electrochemical sensors were proposed in the recent literature. There are other
excellent reviews on the electrochemical detection of cancer biomarkers.1416

4.2.1 Electrochemical Sensors for Carcinoembryonic Antigen

Carcinoembryonic antigen (CEA) is a 180200 kDa glycoprotein that is found
in concentrations of less than 2.5 ng/mL (510 ng/mL) in the blood of healthy
nonsmoker (smoker) adults.17 However, high concentrations of CEA have
been reported in patients with dierent cancers (e.g. gastrointestinal cancer).
Serum CEA values larger than 25 ng/mL are highly suggestive of metastatic
cancer.17
In order to electrochemically detect CEA, anti-CEA antibodies were
immobilized onto glutathione-modied Au nanoparticles, and then the
so-obtained particles were entrapped into a lm of poly(ortho-aminophenol)
electrodeposited onto Au electrodes.18 When incubated with samples containing CEA, the Faradaic impedance of such modied electrodes increased
proportionally with the concentration of CEA. The approach is characterized
by an important advantage: it is among the very few approaches that use no
label and thus it is simple. This label-free electrochemical sensor for CEA detection is characterized by a detection limit of 0.1 ng/mL and a linear range of
0.520 ng/mL. However, unspecic binding to this sensor requires additional
investigations. It was shown that the poly(ortho-aminophenol) lm resists well
to unspecic binding when tested with lysozyme (1 mg/mL) or with serum
(diluted 50 times). However, lysozyme is far from being the most abundant
protein in serum. Moreover, if serum samples are diluted 50 times (as in the
tests of unspecic binding), normal CEA levels will be not measured with this
sensor because they fall below its detection limit.
A lower CEA detection limit (0.01 ng/mL) was achieved using comb-like,
interdigitated carbon electrodes (as working electrodes), a glassy carbon electrode (GCE, as counterelectrode), polystyrene microparticles and a set of three

View Online

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

93

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

19

antibodies. Mouse anti-CEA antibody (1st antibody) was immobilized

onto the polystyrene microparticles and used to capture CEA from samples.
Subsequently, the microparticles were incubated with rabbit polyclonal
anti-CEA antibody (2nd antibody) and goat polyclonal anti-IgG antibody
b-galactosidase conjugate (3rd antibody). b-galactosidase is an enzyme that
produces p-aminophenol when p-aminophenyl-b-D-galactopyranoside is
available. Therefore, the concentration of CEA in the sample can be revealed
by electrochemically measuring how much p-aminophenol is produced. The
novelty of the approach is represented by the way p-aminophenol is quantied
by using a dual amplication system. P-aminophenol is redox cycled in between the interdigitated electrodes (i.e. each p-aminophenol molecule is several
times oxidized on the anode and then immediately reduced on the cathode).
This redox cycling represents the rst amplication system. Each time
p-aminophenol is oxidized to quinoneimine on the anode of the interdigitated
electrodes, silver ions are reduced onto the glassy carbon counterelectrode
(that communicates with the anode through a salt bridge). The higher the
concentration of p-aminophenol, and the longer the redox cycling is allowed to
proceed, the more silver is deposited onto the counterelectrode. This silver
accumulation is thus the second amplication system. The problems associated
with nonspecic binding are eliminated by incubation of microparticles with
bovine serum albumin, BSA, (immediately after immobilization of the mouse
anti-CEA antibody) and by the fact that the counterelectrode (where the silver
deposit was formed) is never exposed to the sample.
Yet another possibility is to carry out the sandwich assay of CEA in a
microtiter plate (i.e. not on the surface of the electrode) and use an electrochemical sensor as the detector to quantify the activity of the enzyme label
(after it is eluted from the immunocomplex).20 The activity of the horseradish
peroxidase (HRP) label linked to the detection anti-CEA antibody was
quantied using a thionine-modied Au electrode and dierential pulse
voltammetry. A detection limit of 0.2 ng/mL CEA was achieved based on this
approach. Good agreement was also observed between results obtained with
this electrochemical approach and immunoradiometric assay while analyzing
(a relatively low number) of human serum samples. However, the approach
retains some of the drawbacks of ELISA (e.g. the quite long incubation times).

4.2.2 Electrochemical Sensors for Prostate-Specic Antigen

Prostate-specic antigen (PSA) is a glycoprotein produced in the prostate gland
with a molecular mass of 26 kDa (without its carbohydrate moieties).21 The
more elevated the blood PSA concentration of a man is, the more likely it is
that he has prostate cancer. However, there are benign conditions that cause
PSA levels to increase as well as prostate cancers that do not induce a PSA level
increase.22 Therefore, diagnosis of prostate cancer involves also investigations
other than tumor marker tests.23
The electrochemical detection of PSA was achieved with a sandwich-type
immunoassay carried out with i) a graphite electrode modied rst with Au

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

94

Chapter 4

nanoparticles and then with capture anti-PSA antibodies, and ii) magnetic
beads modied with both detection anti-PSA antibodies and about 7500 HRP
molecules.24 Nonspecic binding to the sensor was eliminated by blocking the
sensor surface with 0.4% Casein and 0.05% Tween-20 in phosphate buer.
Immunocomplexes formed on the sensor (after incubation with sample) were
quantied by measuring the current resulting from the bioelectrochemical (i.e.
HRP-mediated) reduction of hydrogen peroxide at  0.3 V. The approach
allowed a detection limit of 0.5 pg/mL in 10 mL undiluted serum sample to be
obtained. This detection limit represents a 2000-fold increase compared
to the detection limit achieved by the same approach carried out with single
HRP-labeled antibodies. Measurements of PSA in real samples gave excellent
correlations with standard ELISA.
In a very similar approach the electrochemical sandwich-type immunoassay
was carried out using i) a GCE modied rst with iron-oxide nanoparticles and
then with capture anti-PSA antibodies, and ii) iron-oxide nanoparticles decorated with detection anti-PSA antibodies and HRP.25 Nonspecic binding to
the sensor was this time eliminated by blocking the sensor surface with 1% BSA
solution. Immunocomplexes formed on the sensor (after incubation with
sample) were quantied by measuring the current resulting from the bioelectrochemical (i.e. HRP-mediated) reduction of hydrogen peroxide at  0.2 V.
The immunosensor was characterized by a linear range from 0.005 to 50 ng/mL
and a detection limit of 4 pg/mL. Experiments made with spiked full-blood
serum allowed recoveries from 92.1 to 112.5% to be observed. There was a
good agreement between results obtained with the electrochemical sensor and
ELISA in the same serum samples.
GCEs were sequentially modied with silver-coated mesoporous silica
nanoparticles (to increase the active surface area and improve the electrochemical properties of the electrode), with anti-PSA antibodies (to capture PSA
from samples), and BSA (to eliminate nonspecic binding).26 The resulting
sensors were interrogated with cyclic voltammetry and hydroquinone as indicator (the more PSA present in the sample the smaller were the current peaks
for hydroquinone). The sensor detected PSA with a linear range from 0.05 to
50.0 ng/mL, and a detection limit of 15 pg/mL. When used to detect PSA in
spiked human serum samples, recoveries between 98% and 101% were obtained. When results obtained with human serum samples were compared to
ELISA, a maximum relative error of 6.4% was observed. However, only a
small number of tests were carried out with real samples, a shortcoming that
characterizes many studies on electrochemical sensors for diagnostics.

4.2.3 Electrochemical Sensors for Other Protein Cancer

Biomarkers
A smaller number of dierent sensors were developed to detect other biomarkers than CEA and PSA. These sensors are shortly described in this section.
The 62-kDa c-Myc protein, the product of the c-myc proto-oncogene, is a
nuclear phosphoprotein with DNA binding properties.27 Elevated expression

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

95

of the c-Myc correlates with poor prognosis in head and neck squamous cell
carcinoma.28 The electrochemical detection of c-Myc was only recently reported.29 The capture antibody was immobilized onto cysteamine-modied Au
electrodes, while the detection anti-c-Myc antibody was immobilized onto Au
nanoparticles. The formation of the immunocomplex (between the capture
antibody, the antigen, and the detection antibody) was detected using EIS. The
sensor shows a linear range from 4.3 pmol/L to 43 nmol/L and an estimated
detection limit of 1.5 pmol/L. Recovery between 96.2 and 109% was achieved
in 1% serum samples.
Mucin 1 is a 122-kDa molecular weight protein that is expressed on the
apical surface of epithelial cells of many dierent tissues. Abnormalities in
mucin 1 expression are associated with many cancer types (e.g. breast, ovarian
and lung).30,31 Mucin 1 was electrochemically detected with an aptamer-based
sensor.32 Using aptamers as a biorecognition element instead of the more
commonly used antibodies has several advantages.33 Aptamers are small,
chemically stable, and their production is cost eective. In this case (as in many
other cases) the aptamer used as the biorecognition element had a methylene
blue label and a hairpin conformation. When the aptamer is immobilized on
the electrode surface, the methylene blue label is positioned very close to the
electrode, leading to a high current signal in cyclic voltammetry (CV) or squarewave voltammetry (SWV) due to the short electron-transfer distance. Once the
aptamer binds mucin 1 its conformation changes and the electron-transfer
distance between the methylene blue label and the electrode increases. The
higher the concentration of mucin 1 is the smaller the current signal of the
methylene label becomes. The aptamer-based mucin 1 sensor displayed a
detection limit of 50 nM and a dynamic response range up to 1.5 mM.
Squamous cell carcinoma antigen (SCCA) belongs to the serine protease
inhibitor family of proteins and is a useful tumor marker for diagnosis and
management of squamous cell carcinoma.34,35 This antigen was electrochemically detected based on a sandwich-type immunoassay.36 The GCE to be
used in the assay was rst decorated with nitrogen-doped graphene sheets.
These sheets are characterized by high active surface areas and good conductivities that are two important properties for the development of electrochemical sensors. The rst anti-SCCA antibody was immobilized onto these
sheets with glutaraldehyde. The second antibody was adsorbed onto Pt and
Fe2O3 nanoparticles. High amounts of antibodies can be loaded onto such
particles due to their large surface area. Moreover, the two components of the
nanoparticles (Pt and Fe2O3) show a synergistic eect in catalyzing the reduction of hydrogen peroxide. Therefore, such nanoparticles can replace peroxidase that is commonly used in immunoassays and can improve the
robustness of the assay due to the better stability as compared with the commonly used enzymes. The immunosensor displayed a linear range from 0.05 to
18 ng/mL and a detection limit of 15.3 pg/mL.
a-enolase (ENOA) is a metabolic enzyme that also acts as a plasminogen
receptor and by this mediates activation of plasmin and extracellular matrix
degradation. In tumor cells, ENOA is upregulated and supports anaerobic

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

96

Chapter 4

proliferation, promotes cancer invasion, and is subjected to specic posttranslational modications.37 ENOA was detected using two antibodies, an
anti-ENOA monoclonal antibody that was adsorbed onto the electrode surface
and an anti-ENOA polyclonal antibody that was labelled with Au nanoparticles.38 After the usual incubation and washing steps, the Au nanoparticles
of the formed immunocomplexes were oxidized in 0.1 M HCl at 1.2 V for 120 s,
and the resulting gold ions were quantied using SWV. The immunosensor was
characterized by a linear range from 108 to 1012 g/mL and a detection limit of
11.9 fg (equivalent to 5 mL of a 2.38 pg/mL solution). Compared to the more
commonly used enzyme labels, the Au nanoparticles used in this approach are
more stable and do not require substrates or long incubation times to accumulate signicant amounts of reaction products.
Murine double minute 2 (MDM2) is a protein that is encoded by the MDM2
gene, and an important negative regulator of the p53 tumor suppressor. It
mediates the ubiquitination of p53 leading to its degradation by the proteasome. Elevated expression of MDM2 is observed in a signicant proportion of
dierent types of cancer.39 A label-free electrochemical sensor for the detection
of MDM2 was built by immobilizing anti-MDM2 antibodies onto Au electrodes.40 Such modied electrodes are able to bind MDM2 and binding of
MDM2 increases the faradaic impedance of the modied electrodes in a
concentration-dependent manner. The selectivity of the sensors was tested by
challenging them with recombinant human epidermal growth factor receptor
(HEGFR). However, 1 ng/mL was the highest HEGFR concentration tested
while higher concentrations of proteins can be expected in real samples. The
storage stability of the sensors was also tested. After 2 weeks at 4 1C the sensor
retained 81.8% of its initial signal to 1 ng/mL MDM2. The sensors were also
used to detect MDM2 in normal mouse tissue homogenates and cancerous
mouse tissue homogenates. In samples of 0.1% normal mouse tissue homogenate spiked with dierent concentrations of MDM2, a linear range from 1
to 10 ng/mL and a detection limit of 1.3 pg/mL were found. When analyzing the
same cancerous tissue homogenate the sensor reported a MDM2 concentration
of 7.8 ng/ml while the ELISA kit reported a MDM2 concentration of 10.9 ng/
mL. While the ELISA required 5 h to provide results, the electrochemical
sensor reported on the MDM2 concentration in less than 1 h.

4.2.4 Electrochemical Sensors for Simultaneous Detection of

Several Protein Biomarkers
Over the years many of the cancer biomarkers have been discovered to be not
very specic to a certain cancer and/or to have concentrations aected also by
benign disorders. One very convenient approach to achieve a better disease
diagnosis in such conditions is based on analytical tools simultaneously reporting on several cancer-related biomarkers in a single run. Such tools could
also simplify the working procedure, increase throughput, and reduce the cost
per test. It is relatively easy to make electrodes very small due to the readily
available technology from the microelectronics industry, and to have several

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

97

electrode systems integrated into the same device. Such multielectrode systems
can then be used for the simultaneous electrochemical detection of several
biomarkers. The analysis of signals from such devices might overcome the
mentioned lack of specicity, and thus reliability of individual tests.
The simultaneous electrochemical detection of cancer antigen 125 (CA 125),
CA 15-3, and CA 19-9 was recently reported.41 Magnetic beads were decorated
with monoclonal antibodies for all three antigens and used to capture the
antigens from samples. Polymer dendrimers were decorated with metal sulde
quantum dots and polyclonal rabbit antibodies (detection antibodies) for the
three antigens. For detection of CA 125, CA 15-3, and CA 19-9 the polymer
dendrimers were decorated with Zn sulde, Cd sulde, and Pb sulde quantum
dots, respectively. After carrying out a sandwich-type immunoassay with the
modied magnetic beads and dendrimers, and after dissolving the quantum
dots, anodic stripping voltammetry (ASV) was used to quantify each of the
three released metal ions and thus to quantify the three antigens from the
sample. This multiplexed immunoassay enabled the simultaneous detection of
the three cancer biomarkers in a single run with dynamic ranges of 0.0150 U/
mL and detection limits of 0.005 U/mL.
The simultaneous electrochemical detection of interleukin 6 (IL-6), interleukin 8 (IL-8), vascular endothelial growth factor (VEGF), and vascular
endothelial growth factor-C (VEGF-C) was also reported.42 Magnetic beads
were decorated with antibodies against the target proteins (B100 000 antibodies per bead) as well as with horseradish peroxidase (HRP,B400 000 enzyme molecules per bead). These beads were then used to carry out
immunomagnetic separation of the target proteins from samples before being
injected into a microuidic device with an array of microelectrodes as detectors.
Each microelectrode was modied with antibodies against one of the four
target proteins and thus could bind only those magnetic beads that came out
from the separation step bearing the respective protein. The concentration of
such beads (and thus the concentration of the analyte) was nally determined
by electrochemically detecting the activity of the HRP label with hydroquionone. The assay duration was 50 min. The analysis of 78 oral cancer patient serum samples and 49 controls gave clinical sensitivity of 89% and
specicity of 98% for oral-cancer detection.
In another publication by the same authors the simultaneous electrochemical detection of PSA, prostate specic membrane antigen (PSMA),
platelet factor-4 (PF-4), and IL-6 in the same sample is described.43 Antibodies against these four biomarkers for prostate cancer were immobilized
onto four dierent electrodes of the same array. Before being modied with
the antibodies, the four electrodes were modied with a forest of single-wall
carbon nanotubes in order to increase their active surface area and improve
their electrochemical behavior. The antigenantibody complexes were electrochemically detected and quantied using antibodies labelled with HRP,
hydrogen peroxide and an electrochemically active cosubstrate of HRP
(a quinone). The electrochemical detection presented excellent correlation
with results obtained by ELISA.

View Online

98

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

4.2.5 Electrochemical Sensors for Genetic Markers of Cancer

Detection of DNA (or RNA) sequences with electrochemical sensors requires
approaches quite similar to those described above for detecting protein biomarkers. A single-strand DNA (or a peptide nucleic acid, PNA) sequence is
immobilized onto the electrode (as the biorecognition element). If its complementary DNA (or RNA) sequence is present in the sample, binding between the
two strands occurs. The event is then electrochemically observed.
TP53 gene encodes p53, a 53-kDa tumor-suppressor protein. Mutations in
this gene are associated with a variety of human cancers.44,45 The discrimination between intact and G-A single nucleotide polymorphism (SNP)containing TP53 cancer biomarker sequences was demonstrated with an
electrochemical sensor using a short DNA hairpin molecular beacon (5 0 -GT
TGT GCA GCG CCT CAC AAC-3 0 ) as biorecognition element.46 The DNA
hairpin molecular beacon was labelled with methylene blue before being immobilized onto the electrode. When complementary DNA from the sample
binds to the DNA hairpin molecular beacon, it signicantly changes its conformation and thus alters the distance between the methylene blue label and the
electrode surface. Currents due to oxidation/reduction of methylene blue are
very sensitive to the distance between methylene blue and electrode. Therefore,
such sensors can be used to observe dierences between hybridization with
intact and SNP-containing sequences. Using a truncated DNA hairpin molecular beacon (only 20 nucleotides, instead of the previously used 26 nucleotides) proved to be advantageous in several ways. When using the truncated
DNA hairpin molecular beacon, the sensor was characterized by improved
selectivity for SNP, and the way it works switched from having a high
current signal in the absence of any hybridization to having a high current
signal exclusively in the presence of hybridization (i.e. switched from ON-OFF
to an OFF-ON mode). This switch is advantageous because it avoids false
positives due to desorption of the biorecognition element from the electrode
surface.
A more complex electrode was also used to detect TP53 gene-based DNA
sequences.47 Aiming for electrodes with higher surface area that allows immobilization of larger amounts of the biorecognition element and thus
achieving higher sensitivities, Au electrodes were modied with a network of
electrospun nylon 6 bres entrapping carboxylated carbon nanotubes and
covered with polypyrrole. ssDNA was adsorbed onto such electrodes and used
as recognition element just as in the above example. The ability of methylene
blue to bind to the adsorbed ssDNA decreases after hybridization with the
target DNA. SWV can then be used to quantify how much methylene blue
binds to the interface and thus to quantify the extent of hybridization. Such
sensors were able to detect 50 fM of the wild type TP53 sequence and to distinguish between the wild type and the three base mismatched sequence of
TP53. However, the precise role of each nanomaterial was not investigated, a
shortcoming that characterizes several recent papers mixing together many
dierent nanomaterials.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

99

Up to this point, blocking the sensor surface with an inert protein (before
being used to make measurements in biological samples) was the only method
mentioned to eliminate nonspecic binding. This method is the most often used
one to eliminate nonspecic binding. However, there are also other options to
minimize nonspecic binding. Carefully designing the solid/liquid interface of
the sensors is another option. Three polyethylene glycol (PEG)-based thiol
derivatives were recently used to build sensors for the electrochemical detection
of oestrogen receptor-a (ESR1, a genetic marker involved in breast cancer).48
The sensors were based on PEG alkanethiol (structure no. 1), a mixture of PEG
alkanethiol and mercaptohexanol (structure no. 2) or a bipodal aromatic PEG
alkanethiol (structure no. 3, see Figure 4.2). Best results were obtained with
sensors built by the coassembly of ESR1 capture probes and bipodal aromatic
PEG alkanethiol in a ratio of 1 : 100. These sensors were used for the analysis of
a PCR product resulting from the amplication of the genetic material
extracted from 20 MCF7 cells.

Figure 4.2

Polyethylene glycol and alkanethiol-based surfaces used in the electrochemical detection of a genetic marker of breast cancer. Structure no. 3
proved to perform best by providing both sensitivity and ability to
eliminate nonspecic binding.
Reprinted from Biosensors and Bioelectronics, Vol. 26, Henry, O.Y.F.,
Lluis Acero Sanchez, J., OSullivan, C.K., Bipodal PEGylated alkanethiol
for the enhanced electrochemical detection of genetic markers involved in
breast cancer, 15001506, (2010) with permission from Elsevier.

View Online

100

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

4.2.6 Electrochemical Sensors for Detection of Cancer Cells

Approaches to detect cancer cells with electrochemical sensors are not very
much dierent than the above-described approaches to detect proteins and
DNA sequences. A biorecognition element is again immobilized onto the
electrode surface. However, this element now recognizes features that are on/in
the cell membrane and not freely diusing in the solution. Among the used
recognition elements one nds aptamers, small molecules (with corresponding
receptors on cancer cells), and of course antibodies. Representative examples of
sensors based on each type of biorecognition element will be given below.
Aptamers are especially interesting biorecognition elements when it comes to
detection of cells because their selection procedure can nowadays be carried out
with whole cells.49,50 Detection limits as low as 1000 cells were reported for a
sensor combining graphene-modied electrodes with a 26-mer DNA aptamer
that binds nucleolin that is overexpressed on the cancer-cell surface.51 The
sensor is interrogated using impedance spectroscopy and regenerated through
DNA hybridization that releases the aptamer-bound cells (Figure 4.3).
Another aptamer-based electrochemical sensor involves the use of the
anticancer drug daunomycin.52 This drug interacts with the cell membrane of
cancer cells. Daunomycin-decorated cancer cells are then selectively captured
onto aptamer-modied electrodes. Impedance spectroscopy is used to detect
the presence of the cells on the electrode surface. In order to increase the active
surface area, the number of immobilized aptamers and nally the sensitivity of
the sensor, the aptamers were immobilized on a conducting polymerAu
nanoparticle composite lm.
The detection of MCF7 cancer cells with electrochemical sensors based on an
aptamercellaptamer sandwich architecture was also recently demonstrated.53
Mucin 1-binding aptamer was both immobilized onto 2-mm diameter Au disk
electrodes to act as capture aptamer and labeled with horseradish peroxidase to
act as detection aptamer. The enzyme activity of the aptamercellaptamer
structures that are only formed when cancer cells are present in the sample, was
detected using CV. The sensor was characterized by a linear range from 100 to
1  107 cells, and a detection limit of 100 cells (Obs.: Comparison with other
sensors is dicult because no concentrations are mentioned). When challenged
with 1  107 SP2/0 cells that are expected to not express mucin 1, the sensor
generated a current signal corresponding to 100 MCF7 cells. This indicates that
either SP2/0 cells express at least a certain amount of mucin 1 on their surface
or the nonspecic binding is not suciently eliminated.
Immobilization of folic acid onto the electrode surface was also reported to
assure the selective capture of cancer cells (e.g. HeLa or KB cells).54,55 This was
possible due to the fact that cancer cells overexpress folate receptors, while
normal cells do not. The captured cancer cells were then detected using differential pulse voltammetry (DPV) carried out with surface-immobilized ferrocene. The concept assured a detection limit of 10 HeLa cells per mL.54 Yet
another possibility is to interrogate the sensor surface with impedance
spectroscopy. This concept assured a detection limit of 20 KB cells per mL.55

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

Figure 4.3

101

Cancer-cell detection based on graphene-modied electrodes and nucleolinbinding aptamer (AS1411) as biorecognition element. Graphene is modied
with parylene tetra-carboxylic acid (PTCA) before being immobilized onto
graphite electrodes using Naon. AS1411 is then covalently attached to the
carboxyl groups of the PTCA. The anity binding between cancer cells and
AS1411 (having G-quadruplex structure) was observed using EIS, CV, and
uorescence microscopy. Regeneration of the electrode surface using a
second aptamer, complementary to AS1411, was possible.
Reprinted from Biomaterials, Vol. 32, Feng, L.Y., Chen, Y., Ren, J.S., and
Qu, X.G. A graphene functionalized electrochemical aptasensor for selective
label-free detection of cancer cells, 29302937, (2011) with permission from
Elsevier.

No interferences from HEK 293 cells up to a concentration of 104 cells/mL

were observed for the detection of the HeLa cells54 but binding of HFL-I cells
to the surface of the sensor for the KB cells was reported.55
The detection of low-abundance tumor cells (HepB3) using more traditional
recognition elements, i.e. antibodies, was also demonstrated.56 The authors of
the study make use of graphene, which is immobilized onto the electrode surface using chitosan in order to assure larger active surface areas and improved
electrochemical behavior of the electrode. However, the electrochemical
properties of the graphene were not exploited because detection was carried out
with a bismuth-lm-modied GCE. Anti-EpCAM antibodies were immobilized

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

102

Chapter 4

onto graphene to act as capture antibodies. The electrodes modied in this way
were incubated rst with the sample (30 min) and then further treated in two
dierent ways. As a rst possibility, the electrodes were incubated for 50 min
with SiO2 nanoparticles bearing both CdTe quantum dots and anti-EpCAM
antibodies and with SiO2 nanoparticles bearing both ZnSe quantum dots and
anti-GPC3 antibodies. EpCAM and GPC3 are both antigens (proteoglycans)
from the surface of HepB3 cells. The formation of the immune complexes
between the surface-immobilized antibodies, cells, and nally the nanoparticleimmobilized antibodies was detected and quantied using SWV. SWV was
detecting the Zn and Cd ions resulting from dissolving the quantum dots accumulated onto the sensor surface with HNO3. The quantication of HepB3
cells was possible based on the current peaks of both Zn and Cd. The detection
limit was 5 cells/mL and 10 cells/mL when using a detection scheme based on
Cd ions and Zn ions, respectively. The linear range of the sensor was up to 106
cells/mL. As a second possibility, the electrodes were incubated with SiO2
nanoparticles bearing the same antibodies as in the rst approach but also
quantum dots that allow the uorescent detection of the two immunocomplexes at two dierent wavelengths. Fluorescence microscopy was then used to
visualize the HepB3 cells captured on the electrode surface and labelled with the
nanoprobes.
Anti-EpCAM antibodies were also immobilized onto platinum microelectrodes. The so-obtained antibody-modied microelectrodes were used to detect
ovarian cancer cells, SKOV3.57 Binding of the cancer cells to the surfaceimmobilized anti-EpCAM antibodies was observed as a change in the impedance of the sensor. A detection limit of 4 SKOV3 cells/mL was obtained.
Considering this and the previous example of electrochemical sensors for
cancer-cell detection, it becomes obvious that anti-EpCAM antibody-modied
electrodes will not be able to distinguish between dierent cancer cells.
The electrochemical detection of HEGFR 2-overexpressing breast cancer
cells was also recently reported.58 Unlike in the above approaches, which make
use of either aptamers or antibodies, in this study the cancer cells were detected
by a sandwich-type assay using both an antibody (anti-HEGFR 2, immobilized
onto the electrode surface) and an aptamer (anti-HEGFR 2, immobilized onto
hydrazine-modied Au nanoparticles). The role of the aptamer- and hydrazinemodied Au nanoparticles was to allow detection of the formed immunocomplexes through silver reduction (i.e. through silver staining). Silver staining
could be observed both with an optical microscope and stripping voltammetry.
When used to detect SK-BR-3 breast cancer cells in human serum samples, the
method was characterized by an excellent detection limit of 26 cells/mL.
In addition to the above-described detection schemes employing aptamers,
small molecules, or antibodies as biorecognition elements, cancer cells can also
be detected based on the quantication of specic mRNA. Detection of circulating tumor cells in prostate cancer patient blood samples based on the
electrochemical detection of mRNA from such cells was recently reported.59
First, Au fractal microelectrodes with a nanostructured coating of palladium
were fabricated by thin-lm technology and electrodeposition. These

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

103

microelectrodes were then modied with the following PNAs as capture probes:
NH2-C-G-D-gtc-att-gga-aat-aac-atg-gag-D-CONH2 allowing detection of PSA
mRNA, and NH2-C-G-ata-agg-ctt-cct-gcc-gcg-ct-CONH2 allowing detection
of TMP/ERG Type III mRNA. The rst probe enables unambiguous conrmation that the collected cells originate from a prostate tumor, while the
second probe oers information for the classication of dierent types of
prostate tumors because the TMPRSS2-ERG Type III gene fusion is a marker
found in a subset of prostate tumors. Binding between the capture probe and
target mRNA was observed using voltammetry and a system of two redox
species, Ru(NH3)631 and Fe(CN)63. Statistically signicant changes in the
current signal of the sensor were observed when it was exposed to mRNA at
concentrations as low as 1 pg/mL or extracts from 100 cells/40 mL of cultured
prostate cancer cells (VCaP). Finally, when analyzing blood samples from
prostate cancer patients as well as controls the electrochemical sensor produced
results that were consistent with established PCR-based techniques.

4.3 Electrochemical Sensors for Detection of Cardiac

Biomarkers
Time delay has a major impact on the outcome of patients with cardiac
problems. Unfortunately, heart diseases share symptoms with other diseases
(e.g. acute myocardial infarction and pulmonary embolism were found to share
more than chest pain60), and thus their diagnosis requires several investigations
and time. Therefore, there is a great interest in improving both the accuracy
and speed with which chest pain patients are diagnosed. A fast and reliable
point-of-care device reporting on several cardiac biomarkers is seen as an appropriate tool to obtain such an improvement in diagnosis accuracy and
speed.61 This section will present a couple of electrochemical sensors reported
in the literature for the detection of cardiac biomarkers. Most of these sensors
are suitable to be integrated in point-of-care devices. Biosensors (not necessarily based on electrochemical principles) for the detection of cardiac biomarkers have also been reviewed elsewhere.62
Cardiac troponin T (35 kDa) and troponin I (29 kDa) are cardiac regulatory
proteins that control the calcium-mediated interaction between actin and myosin. Raised serum concentrations of these proteins (4 0.1 mg/L for troponin T,
and from 0.1 to 2 mg/L for troponin I) are now accepted as the standard biochemical marker for the diagnosis of myocardial infarction.63,64 The electrochemical, label-free detection of cardiac troponin I was recently achieved using
a carbon nanobre electrode array modied with anticardiac troponin I antibodies.65 Binding of the antigen to the antibody-modied electrode was
observed with EIS. The sensor was characterized by a detection limit of
0.2 ng/mL, which is 25 times lower than that obtained with conventional methods.
About the same detection limit (148 pg/mL) was achieved for human cardiac
troponin I detection using a sandwich-type immunoassay carried out with i)
mouse antitroponin I monoclonal antibody as capture antibody immobilized

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

104

Figure 4.4

Chapter 4

System of on-chip, planar electrodes (A) and electrochemical detector

obtained by xing a PDMS microchannel onto electrodes (B). On-chip,
planar electrode systems are usually fabricated by means of thin-lm
processing and consist of counterelectrode (CE), reference electrode (RE),
and working electrode (WE). The PDMS part of the detector also hosts
the inlet and outlet for a convenient loading of samples. The depicted
system was used to detect human cardiac troponin I.
Reprinted from Biosensors and Bioelectronics, Vol. 23, Ko, S., Kim, B.,
Jo, S.-S., Oh, S. Y., and Park, J.-K. Electrochemical detection of cardiac
troponin I using a microchip with the surface-functionalized poly(dimethylsiloxane) channel, 5159, (2007) with permission from Elsevier.

onto the walls of a polydimethylsiloxane (PDMS) microchannel, ii) alkaline

phosphatase-labeled mouse antitroponin I monoclonal antibody as detection
antibody, and iii) an array of interdigitated microelectrodes that form the
bottom of the PDMS microchannel (see Figure 4.4) and allow the electrochemical detection of the enzymes reaction product.66 It is worth noting that in
this approach the interface for the immobilization of the capture antibody is
advantageously separated from the interface where the electrochemical detection occurs. Thus, both the immobilization and the detection can be separately optimized without aecting each other. The PDMS wall, where the
capture antibodies are immobilized, is still kept close to the detection electrode
(only 50 mm from each other) in order to eciently detect the product of the
enzyme reaction. In order to eliminate unspecic binding to the PDMS walls of
the microchannel, a 10-min incubation with 1 mg/mL BSA was performed.
Troponin I-proportional signals were obtained for concentrations from
0.2 ng/mL to 10 mg/mL (i.e. for the clinically relevant concentration domain) in
about 8 min from the injection of the analyte.
The simultaneous detection of troponin I and human-heart-type fatty-acidbinding protein (another important cardiac biomarker), by using an electrochemical sandwich-type immunoassay was also recently demonstrated.67
Excellent detection limits (13 fg/mL) were obtained mainly due to the use of

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

105

several nanomaterials. GCEs used in the assay were modied with both graphene nanoribbons (in order to increase active surface area and improve
electrochemical behavior) and antibodies against the two biomarkers. The
sandwich immunoassay was completed with titanium phosphate nanospheres
modied in two dierent ways. Titanium phosphate nanospheres were modied
with either antitroponin I antibodies and Cd or antifatty-acid-binding protein
antibodies and Zn. Cd and Zn show dierent potentials in SWV. Therefore, the
presence and the quantity of the markers could be determined by detecting and
quantifying the two metals with SWV. Results obtained with this electrochemical immunoassay in the analysis of human serum samples correlated well
with results obtained by ELISA.
A special category of electrochemical sensors is represented by sensors using
electrochemiluminescence as the detection principle. Related sensors are special
because, although they involve electrochemistry, electrochemiluminescence delivers an optical signal that is proportional to the concentration of the analyte of
interest. Electrochemiluminescence-based detection of cardiac troponin I (from
10 pg/mL to 5.0 ng/mL, and with a detection limit of 4.5 pg/mL) was recently
reported.68 The approach makes use of a short linear peptide (FYSHSFHENWPSK) that selectively binds troponin I. This peptide is immobilized both
onto magnetic particles and onto liposomes containing bis(2,2 0 -bipyridine)-4,4 0 dicarboxybipyridine ruthenium-di(N-succinimidyl ester) bis(hexauorophosphate) as an electrochemiluminescent reagent. The modied magnetic particles
and liposomes are then used to capture troponin I from samples and to label
it, respectively. The particletroponinliposome complexes are subsequently
separated from the sample, the liposomes disrupted, and the amount of released
electrochemiluminescent reagent quantied by using a GCE poised to 1.15 V
vs. Ag/AgCl. The sensor was successfully used to detect troponin I in four human
serum samples. Another electrochemiluminescence-based sensor for troponin I
detection made use of Au nanoparticles modied with both luminol and antitroponin I antibodies.69 The modied particles were immobilized onto Au electrodes, which thus exhibited stable and strong electrochemiluminescence. In the
presence of troponin I the intensity of this electrochemiluminescence decreased in
a concentration-dependent manner. The sensor was characterized by a dynamic
range from 0.1 to 1000 ng/mL and was used to detect troponin I in plasma
samples.
The detection of troponin T was also demonstrated using an electrochemical
sandwich-type immunoassay carried out with i) screen-printed electrodes
obtained by printing a mixture of graphite powder, silver epoxy, and tetracyanoquinodimethane as redox compound for mediating the electron transfer
between HRP and electrode, ii) antitroponin T capture antibodies that were
immobilized onto polystyrene beads, and iii) HRP-labeled antitroponin T detection antibodies.70 The antibody-modied beads were immobilized onto the
screen-printed electrodes with glutaraldehyde. The so-modied electrodes were
incubated with sample and then with HRP-labelled detection antibodies. The
HRP of the formed immunocomplexes was determined by CV. The sensor was
characterized by a linear range from 0.1 and 10 ng/mL and a detection limit of

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

106

Chapter 4

0.2 ng/mL. The sensor was also calibrated with undiluted human-serum samples with known troponin T content as determined using ELISA. A linear
dependence of the current signal on the troponin T concentration was again
observed. An improved detection limit for troponin T detection (0.017 ng/mL)
was obtained when using i) magnetic beads modied with antitroponin T
antibodies to capture troponin T from samples, ii) HRP-labeled antitroponin T
antibodies as detection antibodies, iii) a magnet to collect the magnetic beads
onto the detection electrode, and iv) a screen-printed electrode to detect the
3,3 0 ,5,5 0 -tetramethylbenzidine oxidized by the HRP label in the presence of
H2O2.71 This sensor was successfully tested also with serum samples.
Myoglobin is an oxygen-binding protein that is released into serum as early
as 1 h after acute myocardial infarction. However, because it is found in a high
concentration in both cardiac and skeletal muscle, myoglobin has poor clinical
specicity and is not used by most hospitals to evaluate chest pain.72 It was also
documented that adding myoglobin measurements to troponin I measurements
has a limited value for exclusion of myocardial infarction.73 In spite of this
limited diagnostic value of myoglobin, some electrochemical sensors for myoglobin detection have been reported. In a recent example, screen-printed electrodes were modied with didodecyldimethylammonium bromide-stabilized
Au nanoparticles and with mouse antihuman myoglobin antibodies and then
used to detect myoglobin in standard solutions and human plasma.74 The
presence of a heme prosthetic group in myoglobin opened up the possibility to
detect and quantify myoglobin using SWV and direct electron transfer between
the surface-immobilized heme and the electrode. The electrochemical immunosensor detected myoglobin in 30 min up to concentrations of 1780 ng/mL,
and with a detection limit of 10 ng/mL. The current signal was found to linearly
increase with the myoglobin content of human plasma samples from healthy
donors and patients that were previously analyzed with a commercially available immunoassay. It is not clear whether the sensor is aected by nonspecic
binding or not and whether the sensor can be used to determine the quantity of
myoglobin in plasma samples or not. In a quite dierent approach, single
polyaniline nanowires were modied with antimyoglobin monoclonal antibodies and used to detect myoglobin.75 Nonspecic binding was eliminated by
soaking the nanowires into 2 mg/mL BSA immediately after immobilization of
the antibodies. Binding of myoglobin to the modied nanowires was detected in
a few seconds as an increase in nanowire conductivity. The single nanowire
biosensor detected myoglobin down to 1.4 ng/mL and showed no signicant
signal when challenged with 500 ng/mL BSA. However, if one takes into
account the concentration of albumin in serum (B40 mg/mL) as well as the
concentration of myoglobin in patients with documented acute myocardial
infarction (B500 ng/mL76), it becomes obvious that the tested BSA concentration was just too small.
An increased concentration of serum c-reactive protein is an indicator of an
inammation in the body. Although this inammation can be due to several
reasons, c-reactive protein levels seem to be also correlated with levels of heartdisease risk. Concentrations higher than 3 mg/mL are indicators of a high risk

View Online

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

107

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

77

of cardiovascular diseases. A very simple, label-free electrochemical sensor

for detection of c-reactive protein consisted of a macroporous Au electrode
modied with anti-c-reactive protein antibodies.78 The sensor detected
c-reactive protein in concentrations from 0.1 to 20 ng/mL when interrogated
with EIS. It was also used to detect c-reactive protein in three serum samples
with acceptable accuracy.

4.4 Electrochemical Sensors for Detection of Acquired

Immunodeciency Syndrome
Acquired Immuno Deciency Syndrome (AIDS) is the disease caused by
human immune deciency virus (HIV).79 Recently developed electrochemical
sensors, with potentials to detect an infection with HIV, fall into two categories.
The rst category of sensors detects the presence of HIV by detecting its proteins (e.g. p24 HIV capsid protein, HIV enzymes, etc.). The second category of
sensors detects virus DNA (after reverse transcription PCR) based on hybridization between capture and complementary DNA sequences. These categories add to the classic possibility to detect infection with HIV by detecting
the antibodies the body is producing as a response to the infection. While these
classic tests are useful 28 weeks after infection, i.e. when the body
produces sucient antibodies to be detected, the new tests including their
electrochemical versions are able to detect infection already in the acute phase
of the infection.
The electrochemical detection of p24 antigen was achieved with capture
antibodies immobilized onto Au nanoparticle-modied GCEs, detection antibodies labelled with HRP, and hydroquinone as redox mediator that is converted by HRP and then electrochemically detected.80 Before use the sensor was
immersed into 3% BSA in order to block the surface and reduce unspecic
binding. DPV was used as a sensitive method to detect the oxidized redox
mediator. The sensor displayed a linear range from 0.01 to 100 ng/mL and a
sensitivity twice the sensitivity of the conventional ELISA.80 p24 antigen was
detected at much lower concentrations (subattogram per millilitre) using a
label-free approach in which anti-p24 antibodies were immobilized onto Au
electrodes and the binding between antigen and immobilized antibody was
observed as a change in the capacitance of the resulting sensor.81 In addition to
the impressive detection limit this capacitive immunosensor is characterized by
a short assay time (B20 min) and a selectivity good enough to allow detection
in p24-spiked human serum samples.
The electrochemical detection of HIV-1 integrase, HIV-1 reverse transcriptase, and HIV-1 protease was achieved using Au electrodes modied with
ferrocene-conjugated peptides (see Figure 4.5).82 The sensors detected the HIV
enzymes in nanomolar concentrations based on the fact that the electrochemical behavior of the ferrocene moiety is strongly altered following binding
of the HIV enzyme to the surface-immobilized peptide. Binding aects both the
potential and the amplitude of the current peaks recorded for the oxidation and

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

108

Figure 4.5

Chapter 4

HIV enzyme detection using a Au electrode modied with a ferroceneconjugated peptide as biorecognition element. The peptides used to construct the sensors were VVStaASta (for detection of HIV-1 protease),
VEAIIRILQQLLFIH (for detection of HIV-1 reverse transcriptase), or
YQLLIRMIYKNI (for detection of HIV-1 integrase).
Reproduced from ref. 84.

reduction of the ferrocene moiety in CV scans. The nonspecic binding to the

sensors was also investigated. However, the tests were carried out at relatively
high concentrations of the target analyte (100 nM) and low concentrations of
nontarget proteins of 300 nM, which is a low concentration if one considers the
protein content of human serum.
When using this sensor design in combination with impedance spectroscopy,
HIV-1 protease was detected in picomolar concentrations.83 Sensors for HIV-1
protease detection were also obtained by immobilizing the ferroceneconjugated peptide (pepstatin) onto a composite material obtained from singlewalled carbon nanotubes and Au nanoparticles.84 The detection limit for
HIV-1 protease detection with the composite material was 0.8 pM.
HIV-1 protease was also detected using its peptide substrate (GSGSARVLAEAGSGSC or GGGSGSGSARVLAEAGGGSGSGSC) that was rst
conjugated to a magnetic bead (30 nm in diameter) and then immobilized onto
a Au electrode.85 The impedance of such peptide-modied electrodes decreased
proportionally with the concentration of active HIV-1 protease in the sample.
The detection limit of the approach was 10 pg/mL. Selectivity was tested by
injection of 1 ng/mL of BSA that did not produce signicant changes in the
impedance of the peptide-modied electrodes. However, this concentration is
very small compared to the protein concentrations that might be present in real
samples.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

109

A HIV DNA biosensor was developed by covalently immobilizing HIV

ssDNA fragments (5 0 -ACT GCT AGA GAT TTT CCA CAT-3 0 ) to a GCE.86
The hybridization of the immobilized HIV ssDNA fragment with its complementary DNA fragment (5 0 -ATG TCG AAA ATC TCT AGC AGT-3 0 ) was
evidenced by using DPV and [Co(phen)2IP]21 as a novel electrochemical indicator/intercalator (where phen 1,10-phenantroline and IP imidazo[f]
[1,10] phenanthroline). A detection limit of 27 pmol and a linear range from
1.61010 to 6.2109 mol was obtained. This sensor was not tested with real
samples.
A much lower detection limit (2 aM) was obtained when a HIV DNA
fragment was detected after making two changes to this approach.87 The rst
change was to use two auxiliary DNA probes in addition to the capture and
target DNAs. These auxiliary probes initiated a cascade of hybridizations
once the initial hybridization between the target probe and the surface-bound
capture probe occurred. The cascade eventually generated micrometer-long
one-dimensional DNA structures on the sensor surface. The second change was
to use [Ru(NH3)6]31 instead of [Co(phen)2IP]21 as redox indicator for quantifying the formed DNA structures. When used in cell lysates or human serum
(diluted 1 : 1), the sensor was still characterized by a good detection limit (5 aM).
HIV has two types, HIV-1 and HIV-2, but electrochemical sensors were
mainly developed for HIV-1, because HIV-2 is relatively rare. However, an array
of planar sensors suitable for detection of both HIV-1 and HIV-2 oligonucleotides was already reported.88 Hairpin DNA probes (GCGAGCCTGGGATTAAATAAAATAGTAAGAATGTATAGCGCTCGC-(CH2)6-SH for HIV-1
and GCGAGCAAAGGACCAGGCGCAACTAAATTCAGCTCGC-(CH2)6SH for HIV-2) were immobilized onto separate Au electrodes having dimensions
of 1 mm2 mm and used as capture probes. These probes bind a lot
more methylene blue in the single strand state than in the hybridized state.
This property allows detection of hybridization by electrochemically (e.g. by
SWV) observing the amount of methylene blue bound to the sensor
surface before and after incubation with the investigated sample. The sequences
GCTATACATTCTTACTATTTTATTTAATCCCAG and TGAATTTAGTTGCGCCTGGTCCTTT were used as target probes specic for HIV-1 and
HIV-2, respectively. Selectivity tests were also carried out with two sequences
displaying single mismatch. The sensor array allowing the detection of oligonucleotides for both HIV-1 and HIV-2 was characterized by a linear range from
20 to 100 nM, a detection limit of 0.1 nM, and by the ability to discriminate
between the complementary and single-base mutation oligonucleotides.
The detection of a polypurine tract sequence conserved in all HIV-1 strains
with an electrochemical sensor was also demonstrated.89 Unlike most of the
mentioned DNA sensors that use ssDNA both as capture probe and as target
probe, this sensor uses triplex-forming oligonucleotides as capture probe and
target double-stranded (ds) DNA. The sensor employs the already described
methylene blue- and SWV-based ON-OFF mode. The sensor detects dsDNA
at concentrations as low as 10 nM and is selective enough to be used in complex
sample matrices such as blood serum.

View Online

110

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

4.5 Electrochemical Sensors for the Detection of

Hepatitis Biomarkers
One of the major causes of mortality worldwide is due to infection with
hepatitis. There are ve known hepatitis viruses: A (HAV), B (HBV), C (HCV),
D (HDV) and E (HEV). Diagnosis of hepatitis viremia is based on the detection
of the virus antigen or of the nucleic material of the virus. Recent progress in
the eld was pushed by the importance of early detection in asymptomatic
patients suering from chronic disease and by the need for personalized
treatment. This has prompted the appearance of a new generation of commercial kits and new technologies. As more nanomaterials and technologies
became available, electrochemical biosensors have emerged as viable alternatives to current standard tests, promising to surpass them in terms of cost,
ease of operation and sensitivity. The vast majority of scientic literature is
centred on Hepatitis B, while a smaller number of reports is focused on
Hepatitis C.

4.5.1 Electrochemical Sensors for Hepatitis B

A variety of electrochemical sensors for diagnosis of hepatitis B have been
reported based on amperometric,90 voltammetric,91 potentiometric,92 conductometric,93 and impedimetric methods.3,94
Hepatitis B surface antigen (HBsAg) is the earliest indicator of acute hepatitis B and it is frequently used for identifying infected people before apparition
of the actual symptoms.95 The current standard method is based on ELISA.96
Sensitive detection of HBsAg was achieved with electrochemical immunosensors of various architectures,97,98 using amplication strategies based on
nanomaterials that allowed improvement of the detection limit by a factor of
100 compared to the standard ELISA test.91 An immunosensor for Hepatitis B
antigen was developed by modifying a nanoporous Au electrode with HRPlabeled antibodyAu nanoparticles bioconjugates.91 Increased sensitivity was
due to the high electroactive area of the nanoporous Au electrode and Au
nanoparticles, allowing for ecient immobilization of HRP-labelled secondary
antibodies (one Au nanoparticle could be covered by several HRP-labelled
antibody molecules). This sensor was characterized by a detection limit of 2.3
pg/mL and a linear range between 0.01 and 1.0 ng/mL (see Figure 4.6).
A similar approach was reported by other authors,93 while others preferred to
use ionic liquids for enhancing the performances of immunosensors for
Hepatitis B antigen.90
A novel, portable and integrated electrochemical biosensor for the sensitive
and selective detection of HBsAg was recently described.95 The device is based
on a sandwich immunoassay put in place by combining immunochromatographic separation on a test strip with electrochemical detection by SWV using
a screen-printed electrode (see Figure 4.7). After optimization, the sensor was
characterized by a detection limit of 0.3 ng/mL (S/N 3) HBsAg and a wide

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

Figure 4.6

111

Analysis of dierent concentrations of HBsAg in the range 0.011.0 ng/mL

by DPV with an electrochemical immunosensor (A). Calibration curves
for sandwich immunoassay of HBsAg (B). Plot of DPV peak current vs.
HBsAg concentration from 0 to 0.1 ng/mL (inset in B).
Reprinted from Talanta, Vol. 30, Ding, C., Li, H., Hu, K., and Lin, J.-M.
Electrochemical immunoassay of hepatitis B surface antigen by the
amplication of Au nanoparticles based on the nanoporous Au electrode,
13851391, (2010), with permission from Elsevier.

linear range (1500 ng/mL). The sensor was applied to the detection of HBsAg
in spiked human-plasma samples. The average recovery was 91.3% and the
device was proven to be more sensitive than a colorimetric immunochromatographic test-strip assay.
Besides biosensors for HBsAg, signicant research was focused on the
electrochemical sensing of hepatitis B virus DNA (after reverse transcription
PCR), by monitoring the hybridization event between the target and a probe
DNA.94,99,100
A label-free DNA sensor was fabricated by coating a single-walled carbonnanotube (SWCNT) array with Au nanoparticles followed by self-assembly of
single-stranded probe DNA.3 Sensitive detection of complementary hepatitis B
ssDNA was achieved by EIS. The principle exploits the variation in the chargetransfer resistance of the modied electrode as a result of DNA hybridization at
the sensor surface. The authors have investigated both aligned and random
SWCNT arrays, which led to similar results.
Magnetic beads are increasingly used in bioassays as they bring a huge
simplication in sample preparation and allow for a high degree of exibility in
the assays.92,101 An electrochemical genomagnetic assay for the detection of
specic HBV DNA included a synthetic 20-mer probe immobilized onto
magnetic beads.101 After incubation with the target DNA, the DNA hybrids
were removed from the magnetic beads by washing with NaOH and were
analysed by DPV using a pencil graphite electrode. The detection was focused
on the guanine oxidation signal at 1.0 V by DPV. The signal only appears
following hybridization, as guanine was only contained by the target DNA
sequence and not by the synthetic probe.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

112

Figure 4.7

Chapter 4

Portable electrochemical detection system for HBsAg consisting in an

integrated electrochemical biosensor, a portable electrochemical analyzer
and a laptop (A). The integrated electrochemical biosensor contains an
immunochromatographic strip for the specic recognition (B) and an
electrochemical cell with screen-printed electrodes for sensitive detection (C).
Reprinted from Talanta, Vol. 94, Zou, Z.-X., Wang, J., Wang, H., Li, Y.-Q.,
and Lin, Y. An integrated electrochemical device based on immunochromatographic test strip and enzyme labels for sensitive detection of diseaserelated biomarkers, 5864, (2012) with permission from Elsevier.

Currently, one direction of research in DNA biosensors, exemplied by

sensors for HBV DNA, focuses on the use of electroactive compounds such as
cobalt phenanthroline,102 methylene blue103 and [Os(bpy)2Cl2]1,104 as well as
novel redox probes such as bis(benzimidazole)cadmium(II) dinitrate
(Cd(bzim)2(NO3)2)105 for monitoring the hybridization event.
Multiplexed detection of hepatitis virus antigens was recently reported using
an integrated automatic electrochemical immunosensor array.106 It allows for
the simultaneous detection of 5 types of hepatitis virus antigens (i.e. hepatitis A,
hepatitis B, hepatitis C, hepatitis D, and hepatitis E) in 5 min. The design of the
immunoassay relies on a direct format where the corresponding antibodies are
immobilized on a protein A/Au nanoparticle support and the detection of
antigens is done by potentiometry. The analytical signal is represented by the
variation in the electrical potential (DE) between the antibody-modied sensor
and a reference electrode as result of the antigenantibody interaction. A cuto value of 2 mV was assigned to discriminate between positive and negative
results based on testing serum samples that were negative for hepatitis by
standard ELISA. The detection limit of the sensor array was around 1.0 ng/mL

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

113

for all virus antigens and the linear range extended up to 350 ng/mL. The intraand interassay precisions of the sensor array were r7.3% and 8.1%, respectively. A very good correlation was found between the results provided by the
sensor and standard ELISA for the analysis of 43 serum samples.
Detection limits and linear range reported in the literature for HBV DNA
vary for dierent lengths of target DNA sequence.92,101,107 Thus, for a 21-bp of
a Hepatitis B DNA sequence the charge-transfer resistance measured by EIS 3
was linearly correlated with the logarithm of target HBV ssDNA concentration
over the range between 1.01018 M and 1.01012 M with a detection limit of
1 attomole/mL.
Analysis time ranges from under half an hour90,106 to several hours,91 depending on the testing format (direct versus sandwich) and the number of steps
involved. In an eort to make the leap from a laboratory concept to practical
applications, an increasing number of studies are focusing on the development
of electrochemical biosensors aiming on improved reproducibility and precision of the method as well as applications to real samples. The intra- and
interassay coecients of variation reported for dierent levels of HBsAg were
below 8%, which is adequate considering the manual steps involved in sensor
preparation.90,91,93 The selectivity of the immunoassay for HBsAg was studied
by incubating the sensor with solutions of antigen in the presence of potentially
interfering compounds in various concentrations.91,93
Electrochemical sensors developed in recent years have been challenged with
serum samples and compared to commercial ELISA tests to establish their
usefulness for the diagnosis of Hepatitis B. Careful design of the sensor
architecture with consideration of nonspecic adsorption allowed observation
of a high degree of correlation between the results provided by the biosensors
and current standard methods.90,91,93,95

4.5.2 Electrochemical Sensors for Hepatitis C

Infection with HCV aects approximately 170 million people worldwide and
represents the major cause of parentally transmitted non-A, non-B hepatitis
infection.108,109 Many people suering from chronic Hepatitis C infection are
asymptomatic. If left untreated, chronic infection leads to liver cirrhosis and
hepatocellular carcinoma. There are 6 known genotypes of HCV and multiple
subtypes, with Genotype 1 (subtypes 1a and 1b) being the most prevalent
genotypes worldwide. Since the discovery of the virus in 1989, the progress in
diagnosing Hepatitis C has been impressive.
Currently, diagnosis of hepatitis C is performed by serological testing of
hepatitis C antibodies (rst by ELISA, followed by conrmation test with recombinant immunoblot assay RIBA) and molecular assays that detect HCV
RNA. Antibody-based tests indicate past as well as present infections, while the
detection of HCV RNA is performed to conrm a current infection and to follow
the response to a treatment. Genotyping and viral load determination are also
very important at this stage in order to assign the adequate treatment and monitor
patients response. HCV genotype and subtype can be determined by various

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

114

Chapter 4

commercially available assays based on dierent methods such as direct sequence

analysis, reverse hybridization, and genotype-specic real-time PCR.110,111 On the
other hand, several ELISA and RIBA assays are commercialized and approved
by regulatory agencies in dierent countries. For example, the OraQuick HCV
rapid antibody test was approved by the FDA in 2010. It uses whole blood
samples, it is portable, easy to use and provides an answer in 40 min.112 Biosensors are mainly intended for the same market of point-of-care tests, therefore
having to compete with these more traditional and robust tests.
Detection of HCV with biosensors relies on the use of antibodies,106 DNA/
PNAs and whole cells engineered to contain antibodies or antigens related to the
virus. The advantage of using PNAs versus DNA probes relies in the uncharged,
exible structure of PNA and their resistance to nuclease and protease activity.
PNA hybridizes with complementary nucleic acids. Intercalators as well as
electroactive molecules (e.g. methylene blue) have been used to detect the formation of a hybrid with target DNA or of a triplex with double-stranded
DNA.113,114
A novel amperometric genosensor for qualitative HCV RNA detection and
genotyping (HCV 1, 2A/C, 2B and 3) was obtained by immobilizing a 18-mer
DNA probe on functionalized graphite.115 The sensor was incubated with target
biotin-labelled complementary DNA, obtained from HCV RNA in serum by
reverse transcriptase-linked polymerase chain reaction (RT-PCR). Detection of
the hybridization event was achieved by further incubation of the sensor with an
avidinperoxidase conjugate and amperometry at 0.45 V versus Ag/AgCl using
the H2O2-KI system. The current intensity was proportional to the amount of
HCV amplicons in the samples. In this architecture, the anity system biotin(strept)avidin was used to immobilize the DNA probe on the modied graphite
electrode and to attach the enzyme to the hybridized target DNA. By using
anity-based immobilization, the authors achieved both a controlled orientation
and strong binding of DNA probes on the electrode surface. To anchor streptavidin on the electrode, the protein was entrapped within a siloxanepoly(propylene oxide) (PPO) nanocomposites prepared by the solgel method. The
detection limit reported for all virus genotypes was around 1 ng/mL115 and the
results were highly reproducible (RSD of 0.21.2% for positive and negative
controls tested ve times on dierent days).
To establish a correlation between a positive/negative result and the absolute
value of the analytical response obtained with the genosensor, a cut-o value
needs to be dened. For example, this value was set at the current intensity
average plus twice the standard deviation obtained for samples deemed as HCV
RNA negative and for negative controls analysed by the standard kit.115
Simultaneous detection and genotyping of HCV RNA (genotypes 1, 2A/C,
2B and 3) was demonstrated with an amperometric genosensor for eight serum
samples from HCV-infected patients.115 By immobilizing dierent probes
(HCV 1, 2A/C, 2B and 3 oligonucleotides) on the PPO electrode it was possible
to distinguish positive and negative sera samples. The amperometric sensor
gave comparable results to the commercial test with spectrophotometric detection (Amplicors, from Roche) and proved more sensitive.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

115

Another approach reported in the literature refers to the bioelectric recognition

assay (BERA) that registers the electrical response of whole cultured cells upon
binding viruses such as HCV inducing a change of their physiology.116 Vero
broblast cells modied by electroinsertion of HBV specic antigens (HBsAg) in
their membranes responded in a fast (45 s) and specic manner to HBV-positive
samples from a pool of 133 clinical blood serum samples. The response was due
to changes in the cell membrane potential that were measured by appropriate
microelectrodes immobilized in an alginate matrix. This assay is a promising onestep, fast and label-free assay, however, it is not as sensitive as the DNA sensors
yet and the response does not correlate in a linear manner with the viral load.117

4.6 Electrochemical Sensors for the Detection of

Rheumatoid Arthritis Biomarkers
Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease that causes
the inammation of joints and surrounding tissue. It aects about 1% of worlds
population. The diagnosis of RA is based on clinical symptoms supported by Xrays, patient history and serological tests. Various autoantibodies have been
proposed as biomarkers for RA: rheumatoid factor, antikeratin, anticitrullinated
peptides, anti-RA33, anti-Sa, and anti-p68 autoantibodies.118,119
At present, RA is diagnosed based on the criteria developed in 2009 by the
American College of Rheumatology/European League Against Rheumatism
collaborative initiative.120 According to these criteria, rheumatoid factor (RF)
and the anticyclic citrullinated peptide (CCP) antibody are the most important
main serological markers.
RF is an autoantibody directed against the Fc region of human IgG and it is
typically detected by nephelometry using a latex agglutination assay. Autoantibodies to citrullinated proteins or peptides (laggrin, brin or vimentin) are
considered the best biomarkers for the early detection of RA as they are present
in patients years before the apparition of clinical symptoms.121,122 Citrulline is
formed enzymatically by deimination of arginine. The CCP antibody is an
antibody directed against citrulline that appears early on in RA patients. AntiCCP assays are performed by plate-based ELISA available from several
manufacturers. A lateral ow immunoassay test for the rapid detection of
RA is commercialized by Euro-Diagnostica AB. Automated, faster analysis of
anti-CCP antibodies (e.g. by Electrochemiluminescence (ECL) on Elecsyss
analysis systems) is also commercially available nowadays.123
It should be noted that new autoantibodies are being discovered and studied
as potential markers of the disease. They hold promise for the future as currently about one-third of RA patients are seronegative for both RF and CCP
antibody.124,125
Besides RF and anti-CCP antibodies, the macrophage migration inhibitory
factor (MIF) has been suggested as a potential biomarker for RA.126
There are not many biosensors for the detection of RA. The presence of
anticitrullinated protein antibodies specic to RA was studied with SPR

View Online

116

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

127

microarray imaging
down to a detection limit of 0.5 pM. A quartz crystal
microbalance (QCM) method, employing a single-walled carbon-nanotubebased immunosensor proved very sensitive (detection limit in the femtomole
range).128 Fluorescence-based and sensitive diagnostic assay of RA was
successfully performed using ZnO nanorods.118 Practical applications of such
biosensors are hampered by the ease of operation and by portability issues.
An electrochemical immunosensor based on the detection of the anti-CFFCP1
antibody RA biomarker was recently reported.129 The device is composed of a
thin-lm metal electrode coated with a multiwalled-carbon-nanotubepolystyrene (MWCNTPS) composite, on top of which a citrullinated specic peptide
receptor was covalently attached. After incubation of the modied electrodes
with serum, the sensors were further incubated with anti-IgG secondary antibodies labelled with HRP and the detection was done by amperometry using the
3,3 0 ,5,5 0 -tetramethylbenzidine (TMB)/H2O2 system. Compared to ELISA, the
detection limits achieved with this biosensor were similar and the analysis time
was shorter. Otherwise, the sensor fabrication and the incubation steps specic to
a sandwich assay are limiting the practical applications of such a device.
A remarkable fact about this research is the step forward made by its authors to
characterize the reproducibility of sensors from dierent batches over a 3-month
period. They emphasized that for 146 electrodes the coecient of variation of
the intensity of current versus a classic redox species (1.0 mM ferricyanide) was
27.5%. This was correlated with the high degree of manual intervention in sensor
fabrication. The yield of sensors within 10% variation was about 70%, which is
still far from adequate reproducibility for practical applications.
A novel electrochemical immunosensor for MIF was developed by functionalizing a Au electrode coated with Au nanoparticles, titanium dioxide nanoparticles and thionine (NGP-NTiP-Thi) with anti-MIF antibodies of the IgM
type.130 The IgM immunosensor was characterized by a linear range between
0.03 and 230 ng/mL and a detection limit of 0.02 ng/mL being also applied for
testing MIF in sera of RA patients.

4.7 Electrochemical Sensors for the Detection of Celiac

Disease Biomarkers
Celiac disease or gluten enteropathy aects many people worldwide (as much
as 1% of the population of US). Gluten and similar proteins from wheat, rye
and barley provoke an exacerbated immune response in genetically predisposed
individuals that ultimately leads to chronic inammation and damage to the
upper part of the small intestine.
The only standard accepted worldwide for diagnosing celiac disease remains
intestinal biopsy with conrmation of the diagnosis by retesting after a glutenfree diet. However, serological tests for IgA antitransglutaminase 2 (IgA antiTG2) or antiendomysial antibody, total IgA and IgG anti-tTG2 (antitissue
transglutaminase 2), antiendomysial or antigliadin antibody are performed rst
as they indicate an increased probability of celiac disease.131,132 In the initial

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

117

screening, IgA isotypes are preferred due to their higher sensitivity, however,
some of the patients present IgA deciency and IgG-based tests need to be
performed instead. Best serological markers of celiac disease are antibodies (IgA
anti-TG2 or antiendomysial) directed to the same antigen. Due to their lower
sensitivity and specicity antigliadin antibodies were not used in the diagnosis of
celiac disease until recently, when tests based on deamidated gliadin peptides
instead of whole gliadin protein have emerged. These allow for similar results as
those using IgA anti-TG2 tests. Novel antibodies against deamidated gliadin
peptide (DGP) are considered especially valuable for the diagnosis of celiac
disease, in those individuals with selective IgA deciency. A recent study133
compared 4 dierent strategies for the serologic diagnosis of celiac disease and
concluded, by using two commercial assays, that determining IgA anti-tTG and
IgG anti-DGP in all patients performed better than the other strategies.
Several electrochemical biosensors for the diagnosis of celiac disease have
been described so far based on impedimetric,134 amperometric135 and voltammetric techniques.4,136 The main characteristics of these sensors are summarized in Table 4.1.
A variety of sensor architectures has been considered in electrochemical
devices intended for diagnosis of celiac disease. While some authors embedded
a relatively high degree of complexity in the materials and technique used136,137
others tried to rationalize the design of the sensor to achieve best analytical
performances. For example, coating a Au electrode with a bipodal thiol containing ethylene glycol groups and carboxylic end groups allowed for a controlled, covalent immobilization of the bioreceptor135,138 while insuring both
low nonspecic adsorption and good diusion of electroactive substrates to
electrode surface.139
Multiplexing and a high degree of exibility of the assay were reported
using a dual screen-printed electrode, where one electrode was modied with
gliadin and the other one with tTG. By using IgA or IgG as secondary antibody
in a sandwich-type assay, this device allowed determination of the IgA and
IgG-type antigliadin and anti-tTG antibodies in real patients samples.136
In order to compare these new sensors with commercial ELISA the researchers used calibration solutions of antibodies from the ELISA kit to build
the calibration curve with their electrochemical biosensor.4,135,138 A cut-o
value was dened for the electrochemical signal, allowing to classify samples as
positive/negative.4,140
Correlation with ELISA was generally good141 even excellent in some
cases.4,135,138 A more rigorous analysis including more serum dilutions and a
higher number of patient samples is needed for a clear conclusion.

4.8 Electrochemical Sensors for the Detection of

Urinary Tract Infection Biomarkers
Urinary tract infection (UTI) is one of the most common infections, causing
severe distress to aected individuals and being associated with high healthcare

View Online

118

Table 4.1

Chapter 4

Electrochemical biosensors for the diagnosis of celiac disease.

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Biorecognition
element
Immobilization matrix
TG
tTG
tTG

gliadin

tTG

tTG
gliadin
gliadin

IgA
DGP

On screen-printed Au
electrode coated with a
polyelectrolyte
Covalent attachment to
Au coated with DT2b
thiol
Physical adsorption on
screen-printed carbon
electrodes modied
with MWCNT and Au
nanoparticles
Physical adsorption on
screen-printed carbon
electrodes modied
with MWCNT and Au
nanoparticles
Physical adsorption on
screen printed carbon
electrodes modied
with MWCNT and Au
nanoparticles
Physical adsorption on
graphiteepoxy
composite electrodes
Covalent attachment to
Au coated with DT2
thiol
Physical adsorption on
screen-printed carbon
electrodes modied
with MWCNT and Au
nanoparticles
Covalent attachment to
Au coated with DT2
thiol
Physical adsorption on
screen-printed carbon
electrodes modied
with MWCNT and Au
nanoparticles

Electrochemical
method

Detection limit
a

Ref.

EIS (3-amino-9- n/a

ethylcarbazole)

136

390 ng/mL for

rabbit anti-tTG
antibodies
CV, anodic
2.45 U/mL for tTG
redissolution of IgA; 2.95 U/mL
silver
for tTG IgG

137

CV, anodic
3.16 U/mL for
redissolution of AGAc IgA;
silver
2.82 U/mL for
AGA IgG

138

n/a
CV, anodic
redissolution of
silver

Amperometric,
n/a
hydroquinone/
H2O2
Amperometric,
46 ng/mL
TMBc/H2O2

139

Amperometric,
TMB/H2O2

138

140

9.1 U/mL for AGA 141

CV, anodic
redissolution of IgA; 9.0 U/mL for
AGA IgG
silver
Amperometric,
142 ng/mL
o-dianisidine/
H2O2
n/a
CV, anodic
redissolution of
silver

140
142

Not available.
(22-(3,5-bis((6 mercaptohexyl)oxy)phenyl)-3,6,9,12,15,18,21-heptaoxadocosanoic acid (DT2).
Antigliadin antibody.

b
c

and social costs. It is estimated that UTI aects about 40% of women at some
point in their lives.
These infections represent 2050% of hospital-acquired infections in ICU
units and in the same time the most common disease for which antibiotics are

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

119

prescribed in long-term care facilities. UTI is a common source of fever in

children and, if left untreated, can induce renal scarring, hypertension, preeclampsia, and end-stage renal disease.142 Diagnosis of UTI is made based on
symptoms (e.g. frequency of urination, dysuria, and pyuria) and on the presence of more than 104 cfu/mL uropathogens in urine. Escherichia coli is by far
the most common uropathogen causing more than 50% of the cases.143 Other
uropathogens, causing less than 10% of infections each are Klebsiella pneumonia, Enterococcus faecalis, Proteus mirabilis, Staphylococcus saprophyticus
and Pseudomonas aeruginosa.
Conventional methods for the detection of E. coli include plate count,
multiple-tube fermentation (MTF), membrane lter (MF), etc., which require a
few hours up to 2 days to be completed.
Diagnosis of UTI is one area where electrochemical biosensors for pointof-care devices could bring a signicant contribution. A recent review144
provides an optimistic prognosis on the use of electrochemical biosensors to
improve UTI diagnosis. The optimism stems from the compatibility of such
devices with microuidics and the demonstrated applicability of some electrochemical sensors for point-of-care testing (e.g. glucose sensor or the I-STAT
handheld blood analyzer developed by Abbott).
The main areas of development of biosensors for UTI diagnosis include:
uropathogen identication,145 testing the antimicrobial susceptibility146 to select the most appropriate antibiotics and dierentiating actual infections from
asymptomatic bacteriuria.147 The latter was recently achieved with integrated
platforms containing both nucleic acid probes and protein biomarkers on the
same sensor array.
Most recent progress regarding biosensors that have been applied to the
analysis of clinical samples for UTI diagnosis will be further discussed.
In a 2005 study, a novel electrochemical 16-sensor array was combined with a
set of oligonucleotide probes applied to nylon membranes in a dot-blot format.148 The probes allowed discrimination among 16S genes from 11 dierent
species of uropathogenic bacteria. An oligonucleotide labeled with peroxidase
was used as detector probe and the detection relied on a reaction catalysed by
HRP. This earlier research showed that such molecular hybridization approaches are amenable to fast room-temperature determinations.
An integrated platform has been described combining a step of bacterial lysis
with sensitive electrochemical detection using a nanostructured electrode on an
analysis chip. Unpuried lysates in buer and urine were used to determine
bacteria in a total time of 30 min, with a reported detection limit of
103 cfu/mL.149
The UTI Sensor Array145 is an example of a biosensor for uropathogen
identication based on an electrochemical 16-sensor array containing immobilized DNA probes targeting the 16S rRNA of the most common uropathogens. The eight probes (immobilized in duplicate) include universal as well as
genus- and species-specic ones. Specic recognition of bacterial 16S rRNA
was achieved by sandwich hybridization using capture probes (immobilized on
the electrode) and pair detector oligonucleotides (free in solution). To amplify

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

120

Chapter 4

the hybridization signal and to translate it into an electrochemical signal the

uorescein-labelled detector probe is incubated with an HRP-conjugated
antiuorescein antibody. Compared to urine culture, the detection limit reported for this biosensor is within the clinical cut-o (104 cfu/mL), the overall
sensitivity was 92% and the specicity 97%. The biosensor provided results
in 1 h.
The same group further developed the UTI sensor array for application to
antimicrobial susceptibility testing.146 An integrated sensor platform able of
multiplex detection of uropathogens and lactoferrrin (as biomarker of the host
immune response) illustrates further the versatility of this concept and the huge
potential for UTI diagnosis.147 Lactoferrin is a protein derived from white
blood cells (WBC) and was researched as a biomarker of pyuria (the presence
of white blood cells in urine). As a characteristic of UTI, pyuria is generally
determined by urine microscopy, although there is also a dipstick test measuring the enzymatic activity of leukocyte esterase (LE).
A sandwich amperometric immunoassay for detection of lactoferrin achieved
a detection limit of 145 pg/ml. Based on the analysis of 111 clinical samples, the
concentration of lactoferrin was shown to be correlated with the results of urine
microscopy for WBC,147 supporting the choice of this protein as a biomarker of
pyuria.
In the new generation of biosensors, the detection of pathogens is integrated
with that of protein biomarker (lactoferrin) on the same analytical chip. The
principles behind this approach rely on sandwich hybridization of capture and
detector oligonucleotides to the target 16S rRNA from uropathogens and on a
sandwich assay with capture and detector antibodies for the detection of protein. Electrochemical detection is achieved by using detector probes and antibodies labelled with HRP. A classic amperometric detection of the TMB/H2O2
reaction catalysed by HRP is employed to translate the specic recognition
events into quantiable signals.150
A prospective clinical study conducted in 2009 on 116 patients emphasized
the possibility to identify uropathogens using an electrochemical biosensor
platform, with similar results as standard urine cultures145 within the short time
frame characteristic to point-of-care testing. A study aiming to clinical validation was reported recently with an integrated sensor array for determining
both nucleic acids and biomarker lactoferrin.150 This sensor array is considered
an example of the next generation of electrochemical devices for UTI diagnosis.
Compared to the commercial glucose sensor or the pregnancy test, developing
a rapid test for diagnosing UTI is deemed to be more complicated, as the matrix
is complex and parameters such as pH and cell count could vary in large limits.
Sample preparation remains a limiting factor for the development of disposable
biosensors for UTI, a problem that could nd a solution by using electrokinetics
or microuidic technology. For example, a multifunctional electrode platform
was used for both electrochemical sensing and to enhance the signal by electrokinetics by in situ stirring and heating. By this approach the sensitivity of the
16S rRNA hybridization assay for UTI diagnosis was increased by an order of
magnitude while the analysis time was drastically reduced.151

View Online

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

121

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Versatility, low cost and the potential for multiplexing have all been enumerated as advantages that drive the development of biosensors for urinarytract infections.144

4.9 Challenges in the Use of Electrochemical Sensors in

Diagnostics
Although many papers are published on electrochemical sensors for diagnostics, relatively few sensors have real chances to reach end users as commercially available products. The main challenges still faced by electrochemical
sensors in their way towards a signicant acceptance in clinical settings are
detailed below.
Electrochemical sensors detecting biomarkers must reliably deal with real
clinical samples. The literature on such electrochemical sensors reects a continuous race for better detection limits and only scarce interest in assuring a
selectivity that allows use with real samples. There are many sensors tested only
with standard solutions or tested with too low concentrations of possible
interfering species, but only few sensors tested with a reasonable number of real
samples. Nonspecic binding is probably the biggest problem that must be
more carefully addressed because samples commonly contain very low concentrations of biomarker in the presence of very high concentrations of other
proteins.
Electrochemical sensors detecting biomarkers must display stability; Stability
is another poorly investigated sensor parameter. Single-use sensors must have
reasonable storage stability, while sensors for multiple uses must also display
operational stability. Yet another possibility is to have sensors simple enough
to be assembled and then calibrated by end users. However, most of the sensors
are far from such simplicity.
Electrochemical sensors must be appropriate for mass production; The electrochemical sensors described in the scientic literature are most often manually made through processes that are not amenable for automation and thus
mass production.
Electrochemical sensors must compete with existing methods in terms of sample
throughput (and convenience in use, in general). Unlike sensors to be used for
point-of-care testing (when a response time of a few minutes is desired next to
reliability), sensors to be used in clinical laboratories must have competitive
sample throughput. There are two possibilities to achieve this. The rst is to
have electrochemical sensors with very short response times (few minutes instead of hours). The second possibility is to use arrays of several sensors reporting on several markers or on the same marker but in several samples. Even
the most advanced sensor arrays described in the literature up to now are not
able to equal the sample throughput of experiments made on common ELISA
microplates.
New success stories will be added to that of the portable, electrochemical
glucose meters, once these challenges are eliminated.

View Online

122

Chapter 4

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

4.10 Conclusions
Adding simplicity, speed, and cost eectiveness to the reliability of clinical laboratory tests is often very challenging. In the attempt to overcome this challenge, many electrochemical sensors were developed for the detection of
biomarkers associated to some of the major human diseases. Specicity of these
sensors was ensured by using biorecognition elements of very dierent nature
ranging from small molecules (e.g. folic acid to recognize cancer cells through
their folate receptors) to biomacromolecules (e.g. antibodies and more recently
aptamers). Sensitivity of these sensors was warranted by using dierent electrochemical methods and (more recently) nanomaterials. There are several
challenges still to overcome by electrochemical sensor in order to be used in
diagnostics. On the positive side, results are extremely promising judged on the
achieved response times, detection limits, and sensitivities (often surpassing
those of classic methods). The correlation between results provided with electrochemical biosensors and standard methods (e.g. ELISA) was generally good
(but in some cases the number of clinical samples was too low to draw a rm
conclusion). Combining dierent concepts on the same analytical platform (e.g.
DNA hybridization and immunosensing), and lab-on-a-chip approaches integrating sample preparation and detection steps are proof of impressive progress
in the eld and also of huge potential. Collaborative projects bringing together
sensor developers, clinicians and end users (clinical laboratories) could be the
driving force for the future development of electrochemical sensors for clinical
diagnosis.

References
1. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals
and Applications, John Wiley & Sons, Inc., Hoboken, NJ, 2nd edn, 2001.
2. E. P. Randviir and C. E. Banks, Anal. Methods, 2013, 5, 10981115.
3. S. Wang, L. Li, H. Jin, T. Yang, W. Bao, S. Huang and J. Wang, Biosens.
Bioelectron., 2013, 41, 205210.
4. M. M. P. S. Neves, M. B. Gonzalez-Garc a, H. P. A. Nouws and
A. Costa-Garc a, Biosens. Bioelectron., 2012, 31, 95100.
5. D. W. Kimmel, G. LeBlanc, M. E. Meschievitz and D. E. Cliel, Analyt.
Chem., 2012, 84, 685707.
6. P. DOrazio, Clin. Chim. Acta, 2011, 412, 17491761.
7. F. J. Gruhl, B. E. Rapp, and K. Lange, in Advances in Biochemical
Engineering/Biotechnology 2011, Springer, Berlin, Heidelberg, 2011, pp.
134.
8. A. Chen and S. Chatterjee, Chem. Soc. Rev., 2013, DOI: 10.1039/
C3CS35518G.
9. A. Merkoci, Electroanal., 2013, 25, 1527.
10. G.-J. Zhang and Y. Ning, Anal. Chim. Acta, 2012, 749, 115.
11. J. Wang, Biosens. Bioelectron., 2006, 21, 18871892.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

123

12. F. Ricci, G. Adornetto and G. Palleschi, Electrochim. Acta, 2012, 84, 74

83.
13. A. Schlichtiger, P. B. Luppa, D. Neumeier and M. Thaler, Bioanal. Rev.,
2012, 4, 7586.
14. B. V. Chikkaveeraiah, A. A. Bhirde, N. Y. Morgan, H. S. Eden and
X. Chen, ACS Nano, 2012, 6, 65466561.
15. S. K. Arya and S. Bhansali, Chem. Rev., 2011, 111, 67836809.
16. J. Li, S. Li and C. F. Yang, Electroanal., 2012, 24, 22132229.
17. V. L. W. Go, Cancer, 1976, 37, 562566.
18. H. Tang, J. Chen, L. Nie, Y. Kuang and S. Yao, Biosens. Bioelectron.,
2007, 22, 10611067.
19. T. Yasukawa, Y. Yoshimoto, T. Goto and F. Mizutani, Biosens. Bioelectron., 2012, 37, 1923.
20. Z. Dai, J. Chen, F. Yan and H. Ju, Cancer Detect. Prev., 2005, 29,
233240.
21. A. Belanger, H. van Halbeek, H. C. Graves, K. Grandbois, T. A. Stamey,
L. Huang, I. Poppe and F. Labrie, Prostate, 1995, 27, 187197.
22. I. M. Thompson, D. K. Pauler, P. J. Goodman, C. M. Tangen,
M. S. Lucia, H. L. Parnes, L. M. Minasian, L. G. Ford, S. M. Lippman,
E. D. Crawford, J. J. Crowley and C. A. Coltman Jr, N. Engl. J. Med.,
2004, 350, 22392246.
23. R. S. Porter and J. L. Kaplan, ed., The Merck Manual of Diagnosis and
Therapy, Merck Sharp & Dohme Corp., a subsidiary of Merck & Co.,
Inc., Whitehouse Station, N.J., online edition., 2010.
24. V. Mani, B. V. Chikkaveeraiah, V. Patel, J. S. Gutkind and J. F. Rusling,
ACS Nano, 2009, 3, 585594.
25. H. Li, Q. Wei, G. Wang, M. Yang, F. Qu and Z. Qian, Biosens.
Bioelectron., 2011, 26, 30443049.
26. H. Wang, Y. Zhang, H. Yu, D. Wu, H. Ma, H. Li, B. Du and Q. Wei,
Anal. Biochem., 2013, 434, 123127.
27. G. J. Kato and C. V. Dang, FASEB J., 1992, 6, 30653072.
28. J. K. Field, D. A. Spandidos, P. M. Stell, E. D. Vaughan, G. I. Evan and
J. P. Moore, Oncogene, 1989, 4, 14631468.
29. J.-L. He, Y.-F. Tian, Z. Cao, W. Zou and X. Sun, Sens. Actuators, B,181,
835841.
30. R. Singh and D. Bandyopadhyay, Cancer Biol. Ther., 2007, 6, 481486.
31. E. Szabo, in Lung Cancer, B. Driscoll and J. M. Walker, Humana Press,
2003, vol. 74, pp. 251258.
32. F. Ma, C. Ho, A. K. H. Cheng, and H.-Z. Yu, Electrochim. Acta, http://
dx.doi.org/10.1016/j.electacta.2013.02.088.
33. S. Song, L. Wang, J. Li, J. Zhao and C. Fan, Trac-Trends Analyt. Chem.,
2008, 27, 108117.
34. J. M. Catanzaro, J. L. Guerriero, J. Liu, E. Ullman, N. Sheshadri,
J. J. Chen and W.-X. Zong, PLoS ONE, 2011, 6, e19096.
35. Murakami, Int. J. Oncol., 2010, 36, 13951400.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

124

Chapter 4

36. D. Wu, H. Fan, Y. Li, Y. Zhang, H. Liang and Q. Wei, Biosens. Bioelectron., 46, 9196.
37. M. Capello, S. Ferri-Borgogno, P. Cappello and F. Novelli, FEBS J.,
2011, 278, 10641074.
38. J. A. Ho, H.-C. Chang, N.-Y. Shih, L.-C. Wu, Y.-F. Chang, C.-C. Chen
and C. Chou, Analyt. Chem., 2010, 82, 59445950.
39. Y. Levav-Cohen, S. Haupt and Y. Haupt, Growth Factors, 2005, 23,
183192.
40. R. Elshafey, C. Tlili, A. Abulrob, A. C. Tavares and M. Zourob, Biosens.
Bioelectron., 2013, 39, 220225.
41. D. Tang, L. Hou, R. Niessner, M. Xu, Z. Gao and D. Knopp, Biosens.
Bioelectron.,46, 3743.
42. R. Malhotra, V. Patel, B. V. Chikkaveeraiah, B. S. Munge, S. C. Cheong,
R. B. Zain, M. T. Abraham, D. K. Dey, J. S. Gutkind and J. F. Rusling,
Analyt. Chem., 2012, 84, 62496255.
43. B. V. Chikkaveeraiah, A. Bhirde, R. Malhotra, V. Patel, J. S. Gutkind
and J. F. Rusling, Analyt. Chem., 2009, 81, 91299134.
44. P. A. J. Muller and K. H. Vousden, Nature Cell Biol., 2013, 15, 28.
45. M. Hollstein, D. Sidransky, B. Vogelstein and C. C. Harris, Science, 1991,
253, 4953.
46. E. Farjami, L. Clima, K. Gothelf and E. E. Ferapontova, Analyt. Chem.,
2011, 83, 15941602.
47. X. Wang, X. Wang, X. Wang, F. Chen, K. Zhu, Q. Xu and M. Tang,
Anal. Chim. Acta, 2013, 765, 6369.
48. O. Y. F. Henry, J. L. A. Sanchez and C. K. OSullivan, Biosens.
Bioelectron., 2010, 26, 15001506.
49. D. Shangguan, Y. Li, Z. Tang, Z. C. Cao, H. W. Chen, P. Mallikaratchy,
K. Sefah, C. J. Yang and W. Tan, PNAS, 2006, 103, 1183811843.
50. Z. Tang, D. Shangguan, K. Wang, H. Shi, K. Sefah, P. Mallikratchy,
H. W. Chen, Y. Li and W. Tan, Analyt. Chem., 2007, 79, 49004907.
51. L. Y. Feng, Y. Chen, J. S. Ren and X. G. Qu, Biomaterials, 2011, 32,
29302937.
52. P. Chandra, H.-B. Noh and Y.-B. Shim, Chem. Commun., 2013, 49, 1900
1902.
53. X. Zhu, J. Yang, M. Liu, Y. Wu, Z. Shen and G. Li, Anal. Chim. Acta,
2013, 764, 5963.
54. J. Liu, Y. Qin, D. Li, T. Wang, Y. Liu, J. Wang and E. Wang, Biosens.
Bioelectron., 2013, 41, 436441.
55. C. Hu, D.-P. Yang, Z. Wang, P. Huang, X. Wang, D. Chen, D. Cui,
M. Yang and N. Jia, Biosens. Bioelectron., 2013, 41, 656662.
56. Y. Wu, P. Xue, Y. Kang and K. M. Hui, Analyt. Chem., 2013, 85, 31663173.
57. A. Venkatanarayanan, T. E. Keyes and R. J. Forster, Analyt. Chem.,
2013, 85, 22162222.
58. Y. Zhu, P. Chandra and Y.-B. Shim, Analyt. Chem., 2013, 85, 10581064.
59. I. Ivanov, J. Stojcic, A. Stanimirovic, E. Sargent, R. K. Nam and
S. O. Kelley, Analyt. Chem., 2013, 85, 398403.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

125

60. G. T. Wilson and F. A. Schaller, J. Am. Osteopath. Assoc., 2008, 108,

344349.
61. U. Friess and M. Stark, Anal. Bioanal. Chem., 2009, 393, 14531462.
62. A. Qureshi, Y. Gurbuz and J. H. Niazi, Sens. Actuators, B, 2012, 171172,
6276.
63. L. Babuin and A. S. Jae, Can. Med. Assoc. J., 2005, 173, 11911202.
64. S. Sharma, P. G. Jackson and J. Makan, J. Clin. Pathol., 2004, 57,
10251026.
65. A. Periyakaruppan, R. P. Gandhiraman, M. Meyyappan, and J. E.
Koehne, Analyt. Chem., 2013.
66. S. Ko, B. Kim, S.-S. Jo, S. Y. Oh and J.-K. Park, Biosens. Bioelectron.,
2007, 23, 5159.
67. L.-N. Feng, Z.-P. Bian, J. Peng, F. Jiang, G.-H. Yang, Y.-D. Zhu,
D. Yang, L.-P. Jiang and J.-J. Zhu, Analyt. Chem., 2012, 84,
78107815.
68. H. Qi, X. Qiu, D. Xie, C. Ling, Q. Gao, and C. Zhang, Analyt. Chem.,
2013, DOI: 10.1021/ac4005259.
69. F. Li, Y. Yu, H. Cui, D. Yang and Z. Bian, Analyst, 2013, 138, 1844
1850.
70. B. V. M Silva, I. T. Cavalcanti, A. B. Mattos, P. Moura, M. D. P.
T. Sotomayor and R. F. Dutra, Biosens. Bioelectron., 2010, 26, 1062
1067.
71. B. E.-F. de Avila, V. Escamilla-Gomez, S. Campuzano, M. Pedrero and
J. M. Pingarron, Electroanal., 2013, 25, 5158.
72. K. Lewandrowski, A. Chen and J. Januzzi, Am. J. Clin. Pathol., 2002,
118, S93.
73. K. M. Eggers, J. Oldgren, A. Nordenskjold and B. Lindahl, Am. Heart J.,
2004, 148, 574581.
74. E. Suprun, T. Bulko, A. Lisitsa, O. Gnedenko, A. Ivanov,
V. Shumyantseva and A. Archakov, Biosens. Bioelectron., 2010, 25,
16941698.
75. I. Lee, X. Luo, X. T. Cui and M. Yun, Biosens. Bioelectron., 2011, 26,
32973302.
76. M. J. Stone, M. R. Waterman, D. Harimoto, G. Murray, N. Willson,
M. R. Platt, G. Blomqvist and J. T. Willerson, Br. Heart J., 1977, 39,
375380.
77. P. M. Ridker, Circulation, 2003, 107, 363369.
78. X. Chen, Y. Wang, J. Zhou, W. Yan, X. Li and J.-J. Zhu, Analyt. Chem.,
2008, 80, 21332140.
79. K. A. Sepkowitz, N. Engl. J. Med., 2001, 344, 17641772.
80. L. Zheng, L. Jia, B. Li, B. Situ, Q. Liu, Q. Wang and N. Gan, Molecules,
2012, 17, 59886000.
81. K. Teeparuksapun, M. Hedstrom, E. Y. Wong, S. Tang, I. K. Hewlett
and B. Mattiasson, Analyt. Chem., 2010, 82, 84068411.
82. K. Kerman and H.-B. Kraatz, Analyst, 2009, 134, 24002404.
83. K. A. Mahmoud and J. H. T. Luong, Analyt. Chem., 2008, 80, 70567062.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

126

Chapter 4

84. K. A. Mahmoud, S. Hrapovic and J. H. T. Luong, ACS Nano, 2008, 2,

10511057.
85. C. Esseghaier, A. Ng and M. Zourob, Biosens. Bioelectron., 2013, 41,
335341.
86. S.-Y. Niu, S.-S. Zhang, L. Wang and X.-M. Li, J. Electroanal. Chem.,
2006, 597, 111118.
87. X. Chen, C.-Y. Hong, Y.-H. Lin, J.-H. Chen, G.-N. Chen and
H.-H. Yang, Analyt. Chem., 2012, 84, 82778283.
88. D. Zhang, Y. Peng, H. Qi, Q. Gao and C. Zhang, Biosens. Bioelectron.,
2010, 25, 10881094.
89. A. Patterson, F. Caprio, A. Vallee-Belisle, D. Moscone, K. W. Plaxco,
G. Palleschi and F. Ricci, Analyt. Chem., 2010, 82, 91099115.
90. D. Tang, H. Li and J. Liao, Microuid. Nanouid., 2009, 6, 403409.
91. C. Ding, H. Li, K. Hu and J.-M. Lin, Talanta, 2010, 80, 13851391.
92. H. Hanaee, H. Ghourchian and A.-A. Ziaee, Anal. Biochem., 2007, 370,
195200.
93. H. Liu, Y. Yang, P. Chen and Z. Zhong, Biochem. Eng. J., 2009, 45, 107
112.
94. W. M. Hassen, C. Chaix, A. Abdelghani, F. Bessueille, D. Leonard and
N. Jarezic-Renault, Sens. Actuators, B, 2008, 134, 755760.
95. Z.-X. Zou, J. Wang, H. Wang, Y.-Q. Li and Y. Lin, Talanta, 2012, 94, 58
64.
96. J. Pal, L. Palinkas, Z. Nyarady, T. Czompoly, I. Marczinovits,
G. Lustyik, Y. Saleh Ali, G. Berencsi, R. Chen, R. Varro, A. Par and
P. Nemeth, J. Immunol. Methods, 2005, 306, 183192.
97. D. Tang, R. Yuan, Y. Chai, Y. Fu, J. Dai, Y. Liu and X. Zhong, Biosens.
Bioelectron., 2005, 21, 539548.
98. T. Konry, B. Hadad, Y. Shemer-Avni, S. Cosnier and R. S. Marks,
Talanta, 2008, 75, 564571.
99. C. Ding, F. Zhao, M. Zhang and S. Zhang, Bioelectrochem., 2008, 72,
2833.
100. N. Liu, G.-J. Li, S.-F. Liu and S.-S. Zhang, Sens. Actuators, B, 2008, 133,
582587.
101. A. Erdem, D. Ozkan Ariksoysal, H. Karadeniz, P. Kara, A. Sengonul,
A. A. Sayiner and M. Ozsoz, Electrochem. Commun., 2005, 7, 815820.
102. A. Erdem, K. Kerman, B. Meric, U. S. Akarca and M. Ozsoz, Electroanal., 1999, 11, 586587.
103. A. Erdem, K. Kerman, B. Meric, U. S. Akarca and M. Ozsoz, Anal. Chim.
Acta, 2000, 422, 139149.
104. H.-X. Ju, Y.-K. Ye, J.-H. Zhao and Y.-L. Zhu, Anal. Biochem., 2003, 313,
255261.
105. X. Li, Y. Chen and X. Huang, J. Inorg. Biochem., 2007, 101, 918924.
106. J. Tang, B. Su, D. Tang and G. Chen, Biosens. Bioelectron., 2010, 25,
26572662.
107. A. Erdem, M. Muti, H. Karadeniz, G. Congur and E. Canavar, Colloids
Surf., B, 2012, 95, 222228.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

Electrochemical Detection of Disease-Related Diagnostic Biomarkers

127

108. European Association for the Study of the Liver, J. Hepatol., 2011, 55,
245264.
109. D. Lavanchy, Clin. Microbiol. Infect., 2011, 17, 107115.
110. M. G. Ghany, D. B. Strader, D. L. Thomas and L. B. See, Hepatology,
2009, 49, 13351374.
111. S. Chevaliez and J.-M. Pawlotsky, Best Pract. Res., Clin. Gastroenterol.,
2008, 22, 10311048.
112. S. R. Lee, G. D. Yearwood, G. B. Guillon, L. A. Kurtz, M. Fischl,
T. Friel, C. A. Berne and K. W. Kardos, J. Clin. Virol., 2010, 48, 1517.
113. M. S. Hejazi, M. H. Pournaghi-Azar and F. Ahour, Anal. Biochem., 2010,
399, 118124.
114. M. H. Pournaghi-Azar, F. Ahour and M. S. Hejazi, Anal. Bioanal. Chem.,
2010, 397, 35813587.
115. C. dos, S. Riccardi, K. Dahmouche, C. V. Santilli, P. I. da Costa and
H. Yamanaka, Talanta, 2006, 70, 637643.
116. S. Kintzios, E. Pistola, J. Konstas, F. Bem, T. Matakiadis,
N. Alexandropoulos, I. Biselis and R. Levin, Biosens. Bioelectron., 2001,
16, 467480.
117. A. Perdikaris, N. Alexandropoulos and S. Kintzios, Sensors, 2009, 9,
21762186.
118. K.-Y. Ahn, K. Kwon, J. Huh, G. T. Kim, E. B. Lee, D. Park and J. Lee,
Biosens. Bioelectron., 2011, 28, 378385.
119. K. Conrad, D. Roggenbuck, D. Reinhold and T. Dorner, Autoimmun.
Rev., 2010, 9, 431435.
120. D. Aletaha, T. Neogi, A. J. Silman, J. Funovits, D. T. Felson,
C. O. Bingham, N. S. Birnbaum, G. R. Burmester, V. P. Bykerk and
M. D. Cohen, Arthritis Rheum., 2010, 62, 25692581.
121. G. Pruijn, A. Wiik and W. van Venrooij, Arthritis Res. Ther., 2010,
12, 203.
122. A. S. Wiik, W. J. van Venrooij and G. J. M. Pruijn, Autoimmun Rev.,
2010, 10, 9093.
123. B. Cai, L. Wang, J. Liu and W. Feng, Clin. Biochem., 2011, 44, 989993.
124. K. Somers, P. Geusens, D. Elewaut, F. De Keyser, J.-L. Rummens,
M. Coenen, M. Blom, P. Stinissen and V. Somers, J. Autoimmun., 2011,
36, 3346.
125. L. A. Trouw and M. Mahler, Autoimmun. Rev., 2012, 12, 318322.
126. E. F. Morand, R. Bucala and M. Leech, Arthritis Rheum., 2003, 48, 291299.
127. A. M. C. Lokate, J. B. Beusink, G. A. J. Besselink, G. J. M. Pruijn and
R. B. M. Schasfoort, J. Am. Chem. Soc., 2007, 129, 1401314018.
128. K. A. Drouvalakis, S. Bangsaruntip, W. Hueber, L. G. Kozar, P. J. Utz
and H. Dai, Biosens. Bioelectron., 2008, 23, 14131421.
129. M. de Gracia Villa, C. Jimenez-Jorquera, I. Haro, M. J. Gomara,
R. Sanmart , C. Fernandez-Sanchez and E. Mendoza, Biosens. Bioelectron., 2011, 27, 113118.
130. S. Li, R. Zhang, P. Li, W. Yi, Z. Zhang, S. Chen, S. Su, L. Zhao and
C. Hu, Int. Immunopharmacol., 2008, 8, 859865.

View Online

14/10/2013 09:27:56.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00089

128

Chapter 4

131. C. Briani, D. Samaroo and A. Alaedini, Autoimmun. Rev., 2008, 7,

644650.
132. N. K. Gatselis, K. Zachou, G. L. Norman, G. Tzellas, M. Speletas,
S. Gabeta, A. Germenis, G. K. Koukoulis and G. N. Dalekos, Clin. Chim.
Acta, 2012, 413, 16831688.
133. P. Vermeersch, K. Geboes, G. Marien, I. Homan, M. Hiele and
X. Bossuyt, Clin. Chim. Acta, 2012, 413, 17611767.
134. T. Balkenhohl and F. Lisdat, Anal. Chim. Acta, 2007, 597, 5057.
135. S. Dulay, P. Lozano-Sanchez, E. Iwuoha, I. Katakis and
C. K. OSullivan, Biosens. Bioelectron., 2011, 26, 38523856.
136. M. M. P. S. Neves, M. B. Gonzalez-Garc a, C. Delerue-Matos, and A.
Costa-Garc a, Sens. Actuators, B, 2013, DOI: 10.1016/j.snb.2012.09.019.
137. M. Ortiz, A. Fragoso and C. K. OSullivan, Analyt. Chem., 2011, 83,
29312938.
138. L. C. Rosales-Rivera, J. L. Acero-Sanchez, P. Lozano-Sanchez, I. Katakis
and C. K. OSullivan, Biosens. Bioelectron., 2011, 26, 44714476.
139. L. C. Rosales-Rivera, J. L. Acero-Sanchez, P. Lozano-Sanchez,
I. Katakis, and C. K. OSullivan, Bios. Bioelectron., 2012, 33 , 134138.
140. M. Pividori, A. Lermo, A. Bonanni, S. Alegret and M. del Valle, Anal.
Biochem., 2009, 388, 229234.
141. M. M. P. S. Neves, M. B. Gonzalez-Garc a, A. Santos-Silva and A. CostaGarc a, Sens. Actuators, B, 2012, 163, 253259.
142. V. A. Hener and M. H. Gorelick, Clin. Ped. Emerg. Med., 2008, 9,
233237.
143. N. S. Sheerin, Medicine, 2011, 39, 384389.
144. K. E. Mach, P. K. Wong and J. C. Liao, Trends Pharmacol. Sci., 2011, 32,
330336.
145. K. E. Mach, C. B. Du, H. Phull, D. A. Haake, M.-C. Shih, E. J. Baron
and J. C. Liao, J. Urol., 2009, 182, 27352741.
146. K. E. Mach, R. Mohan, E. J. Baron, M.-C. Shih, V. Gau, P. K. Wong and
J. C. Liao, J. Urol., 2011, 185, 148153.
147. Y. Pan, G. A. Sonn, M. L. Y. Sin, K. E. Mach, M.-C. Shih, V. Gau,
P. K. Wong and J. C. Liao, Biosens. Bioelectron., 2010, 26, 649654.
148. C.-P. Sun, J. C. Liao, Y.-H. Zhang, V. Gau, M. Mastali, J. T. Babbitt,
W. S. Grundfest, B. M. Churchill, E. R. B. McCabe and D. A. Haake,
Mol. Genet. Metab., 2005, 84, 9099.
149. B. Lam, Z. Fang, E. H. Sargent and S. O. Kelley, Analyt. Chem., 2012, 84,
2125.
150. R. Mohan, K. E. Mach, M. Bercovici, Y. Pan, L. Dhulipala, P. K. Wong
and J. C. Liao, PLoS ONE, 2011, 6, e26846.
151. M. L. Y. Sin, T. Liu, J. D. Pyne, V. Gau, J. C. Liao and P. K. Wong,
Analyt. Chem., 2012, 84, 27022707.

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

CHAPTER 5

In Vivo Sensors for Continuous

Monitoring of Blood Gases,
Glucose, and Lactate:
Biocompatibility Challenges and
Potential Solutions
MEGAN C. FROST,*a ALEXANDER K. WOLFb AND
MARK E. MEYERHOFFb
a

Department of Biomedical Engineering, Michigan Technological University,

Houghton, MI 49931-1295, USA; b Department of Chemistry, The University
of Michigan, Ann Arbor, MI 48109-1055, USA
*Email: [email protected]

5.1 Introduction
Advances in electrochemical and optical sensors for specic ions (H1, Na1,
K1, Ca12, Cl, gases (PO2, PCO2), and low molecular weight biomolecules
(e.g., glucose, lactate, creatinine, urea, etc.) over the past 40 years have revolutionized the measurement of such species in the clinical setting, enabling
rapid, near patient/point-of-care testing of such critical care analytes in
undiluted blood samples.1,2 However, nearly all existing sensor-based
measurement systems require discrete samples of blood. Only a few commercial
systems are designed for continuous monitoring capability either in vivo
(for glucose monitoring in subcutaneous uid3,4) or for short periods in blood
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

129

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

130

Chapter 5

within the bypass blood loops used during open-heart surgery (where blood
clotting is prevented with high levels of anticoagulant).5,6 Even the existing
implantable sensors for glucose that are commercially available have signicant
limitations in their accuracy, requiring frequent calibration and conrmation of
results via separate in vitro measurements of blood glucose using conventional
glucometer test strips.710
Although there have been enormous eorts over the past 25 years by many
companies to develop catheter-type sensors that could monitor blood gases
(oxygen, carbon dioxide and pH) within arteries of critically ill patients using
miniaturized electrochemical or optical sensors,1114 all of these eorts failed to
produce a successful product. The lack of success in developing implantable
chemical sensors that can provide reliable analytical results, especially for the
blood gas test panel (PO2, PCO2, and pH), as well as for glucose and lactate
(note: lactate levels are a major indicator of status for intensive-care patients1517)
can be attributed primarily to the biological response to the indwelling sensors.
In this chapter, we summarize the status of implantable chemical sensor technology for these analytes, describe the biocompatibility problems inherent to
both intravascular (placed in the blood stream) and subcutaneous (implanted
under the skin) chemical sensors, and discuss new approaches aimed at solving
these fundamental problems that might enable greatly enhanced analytical performance for indwelling chemical sensing devices.

5.2 Design of In Vivo Sensors

The basic requirements for sensors potentially useful for continuous real-time,
in vivo measurements are: a) be of very small size (o1 mm o.d.) for easy
placement within a blood vessel (i.e., artery for blood gases/pH; vein for glucose/lacate) or under the skin in subcutaneous space (primarily for glucose); b)
exhibit long-term analytical signal stability (limiting frequency of recalibration); c) cannot exude toxic chemicals into the patient; and d) provide reliable
analyte values that correlate with conventional in vitro tests that use discrete
blood samples, ideally such that analyte values reported by the implanted
sensor can be used in place of discrete blood-test measurements to initiate
appropriate medical therapy.

5.2.1 Sensing PO2/PCO2/pH in Blood

Miniaturized versions of both electrochemical (potentiometric and amperometric) and optical sensors have been employed to prepare intravascular
devices. Figure 5.1 illustrates the overall concept of intravascular blood-gas/
pH sensors to monitor patients continuously. Optical sensors have certain
advantages over electrochemical devices, including (1) ease of miniaturization
using modern optical bers; (2) potentially less electronic noise (no transduction wires); (3) potential long-term stability using ratiometric measurements at
multiple wavelengths; and (4) no need for a separate reference electrode. These
advantages promoted the development of optical sensor technology initially for
the design of intravascular blood gas sensors using multiple optical bers.11 In

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

Figure 5.1

131

Overall concept of in vivo blood gas/pH patient monitoring system based

on use of intravascular chemical sensors. Sensor inserted in to artery
would be able to continuously monitor oxygen, carbon dioxide and pH
levels in the atrial blood.
Light

Oxygen sensor

Silicone coating
Hydrophilic coating

pH sensor
Carbon dioxide sensor
Oxygen sensor

Carbon dioxide
sensor

Figure 5.2

pH sensor
End view

Schematic of intravascular optical blood-gas/pH sensing catheter conguration that uses three optical bers to carry light to and from uorescent dyes in sensing layers that enables optical detection of pH, PCO2
and PO2. This conguration is based on US Patent # 5,047,627 (1991).

such systems, light is passed to and from the sensing site that contains immobilized dyes sensitive to the analyte species by the optical bers and measures
(by reectance) the uorescence or phosphorescence originating from the immobilized indicator at the distal end of the bers (see Figure 5.2 for multiber
design for intravascular blood gas/pH sensing originally patented by Abbott

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

132

Chapter 5

Laboratories). Optical ber sensors for PO2 measurements are typically based
on the immobilization of certain organic dyes, such as pyrene, diphenylphenanthrene, phenanthrene, uoranthene, or metal ligand complexes, such as
ruthenium[II] tris[dipyridine], or Pt and Pd metalloporphyrins within hydrophobic polymer lms (e.g., silicone rubber) in which oxygen is very soluble.18,19
The uorescence or phosphorescence of such species at a given wavelength is
often quenched in the presence of paramagnetic species, including molecular
oxygen. In the case of embedded uorescent dyes, the intensity of the emitted
uorescence of such lms will decrease in proportion to the PO2 of the sample
in contact with the polymer lm. Optical-ber oxygen sensors have also
been developed using a luminophore platinum (II) mesotetrakis(pentauorophenyl)porphyrin (PtTFPP) complex embedded in dierent sol-gels.20,21
The uorescence of the PtTFPP is quenched in the presence of oxygen.
Optical sensors suitable for the determination of PCO2 employ optical pH
transducers (with immobilized indicators) as inner transducers in an arrangement quite similar to the classic StowSeveringhaus style electrochemical PCO2
sensor design. In this design, the pH-sensing layer is behind an outer gas
permeable coating (e.g., silicone rubber) that prevents sample proton access to
the pH sensing layer (see Figure 5.3).22 The addition of bicarbonate salt within
the pH-sensing hydrogel layer creates the required electrolyte lm layer, which
varies in pH depending on the partial pressure of PCO2 in equilibrium with the
lm. As the partial pressure of PCO2 in the sample increases, the pH of the
bicarbonate layer decreases, and the corresponding decrease in the concentration of the deprotonated form of the indicator (or increase in the
concentration of protonated form) is optically sensed. An alternate optical
intravascular PCO2 sensor was developed by Jin, et al.23 using the uorescence
indicator 1-hydroxypyrene-3,6,8-trisulfonate that directly interacts with CO2.
Optical pH sensors require immobilization of appropriate pH indicators
within thin layers of hydrophilic polymers (e.g., hydrogels) because equilibrium
access of protons to the indicator is essential. Fluorescein or 8-hydroxy-1,2,6pyrene trisulfonate (HPTS) have often been used as the indicators and the
uorescence of the protonated or deprotonated form of the dyes can be used for

Figure 5.3

Basic in vivo PCO2 sensing conguration that can be employed to detect

carbon dioxide levels in blood using inner electrochemical or optical pH
sensor.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

Figure 5.4

133

General design of implantable electrochemical PO2 sensing catheter,

where oxygen in sample permeates outer tubing wall of the catheter and
is reduced electrochemically at working Pt wire electrode. The current ow
is directly proportional to the PO2 level in the blood stream.

sensing purposes.24 Jin and coworkers report the fabrication of an intra-arterial

pH sensor based on uorescence using N-allyl-4-piperazinyl-1,8-napthalimide
and 2-hydroxyethyl methacrylate.25 The sensors were tested in rabbits and
showed appropriate response times (t90 13.7  0.8 s to 109.61  4.06 s,
depending on the conditions of the pH change) that are quite useful for realtime monitoring. One problem with respect to using immobilized indicators for
accurate physiological pH measurements is the eect of ionic strength on the
pKa of the indicator.26 Because optical sensors measure the concentration of
protonated or deprotonated dye as an indirect measure of hydrogen ion
activity, variations in the ionic strength of the physiological sample have been
known to inuence the accuracy of the pH measurement.
While many of the prototypes for implantable blood gas monitoring
developed in the 1980s/1990s were based on optical sensors that could be easily
inserted in the radial artery of patients, miniaturized electrochemical sensors
have also been described for in vivo sensing of blood gases.1113 Electrochemical
oxygen sensing catheters are based on the Clark-style oxygen sensing conguration (Figure 5.4), in which a platinum wire along with a Ag/AgCl or
other reference electrode are placed within a narrow gas permeable tubing
lled with an electrolyte solution.11 Oxygen within the blood stream diuses
through the walls of the tubing and is reduced electrochemically at the Pt
working electrode via application of a 0.6 to 0.7 V vs. the Ag/AgCl reference
electrode. The resulting current in the circuit is proportional to the PO2 levels in
the blood.
Electrochemical PCO2 sensors suitable for intravascular sensing are typically
prepared with miniature pH sensors, ideally nonglass potentiometric devices,
made with metal oxides27 or ionophore-based polymer membrane devices28,29

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

134

Chapter 5

that contain modied lling solutions. Carbon dioxide diusion through the
walls of a catheter tubing changes the pH within an inner bicarbonate lling
solution. In one proposed in vivo sensor conguration, the pH ionophore
tridodecylamine was impregnated into the walls of a dual lumen silicone rubber
catheter tubing, with one of the lumens lled with buer and the other with
bicarbonate solution. The voltage between Ag/AgCl electrodes placed in each
lumen is dependent on the PCO2 level in the sample surrounding the tubing,
since the wall between the two lumens is pH sensitive and detects the pH change
within the lumen lled with the bicarbonate solution. If an external reference
electrode is also employed, then the voltage between the Ag/AgCl wire in the
buered lumen vs. the external reference is related to the pH of the sample
phase. Other recent work focused on electrochemical pH sensing includes a
report by Makos, et al.30 who developed a modied carbon-ber electrode by
the reduction of the diazonium salt Fast Blue RR on the ber. The resultant
sensor was capable of monitoring pH changes as small as 0.005 pH units.
Huang and coworkers31 report the fabrication of a exible pH sensor utilizing a
polymer substrate with an iridium oxide sensing layer, a sol-gel matrix, and an
electroplated AgCl reference electrode. The ability to conform to a variety of
shapes oers a wide range of potential applications.

5.2.2 Sensing Glucose in Subcutaneous Tissue

The vast majority of sensors that can be implanted subcutaneously for
continuous glucose monitoring (CGM) and lactate monitoring are electrochemical devices.7,3240 For glucose sensors, the enzyme glucose oxidase is most
often employed as an immobilized catalyst on the surface of a suitable electrochemical transducer. Some congurations use amperometric oxygen sensors with
glucose oxidase immobilized as layer on the outer surface. Glucose present in the
bathing solution at the sensor reacts with oxygen via the enzyme, lowering the
partial pressure of oxygen immediately adjacent to the outer gas-permeable
membrane of the oxygen sensor. The decrease in PO2 and concomitant current
decrease is proportional to the level of glucose in the surrounding blood or tissue.
However, typically, a second nonenzyme covered PO2 sensor is required to
monitor the level of PO2 present in the sample phase. Similar congurations
based on optical PO2 sensors have also been reported.41
More often, implantable glucose sensors are fabricated using electrochemical
oxidation of hydrogen peroxide (H2O2), produced by the immobilized glucose
oxidase catalyzed reaction of glucose and oxygen. This is normally accomplished using a small surface area of platinum wire as the anode and an applied
voltage of 0.6 to 0.7 V vs. an external Ag/AgCl electrode. Other physiological species (e.g., ascorbic acid, uric acid, etc.) can be oxidized at this
potential, so additional layers of membranes/coatings that decrease access of
these species to the platinum surface are required. Further, the outermost layer
of such sensors requires a polymeric coating (illustrated in Figure 5.5) that
restricts glucose diusion42 so that the glucose concentrations within the
underlying immobilized enzyme layer are always less than the lowest levels of

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

Figure 5.5

135

One design of subcutaneous or intravascular glucose sensors

(diametero0.6 mm), with glucose oxidase (GOx) immobilized over
surface of Pt/Ir wire and a coiled Ag/AgCl reference electrode. The
outer polyurethane layer decreases the rate of glucose access to enzyme
layer, while allowing rapid diusion of oxygen, that enables the sensor to
exhibit a linear response to high levels of glucose (to 20 mM). The Pt/Ir
electrode is poised at 0.65 V vs. Ag/AgCl to oxidize hydrogen peroxide
generated from the enzymatic reaction.
With permission from reference 11.

oxygen that might be present in the in vivo sample phase. This requirement
greatly decreases the sensitivity of the devices to variations in localized oxygen
levels and also enables the sensors to respond linearly to abnormally high
levels of glucose (up to 20 mM) that might be present in blood or tissue of
patents.
Another approach taken to prepare in vivo glucose sensors using immobilized
redox hydrogels in which tethered mediator sites (e.g., ferrocene or Os(III)
redox species) are able to accept electrons from the reduced form of the
immobilized glucose oxidase enzyme after reaction with glucose.43 These
reduced mediators are then oxidized and hence the steady-state current is directly
proportional to glucose concentration. The advantage of the mediator-based
devices is that they partially reduce the need for outer membranes that carefully
control the ux of glucose relative to oxygen. Further, the applied potential to
the inner conductive electrode material can be less positive to reoxidize the
mediator (compared to hydrogen peroxide), reducing the need for exclusion
layers to block potential oxidizable species present in the in vivo milieu.
There have been a few optical-based glucose sensors reported with small size
requirements required for implantation.44 These are usually uorescent sensors
that utilize glucose-binding proteins and uorophore-labeled saccharides in a
competitive binding mode behind membranes that are permeable to glucose.
Recent developments in electrochemical glucose sensors have branched in
several directions from the second-generation sensors described above. Some
continue to utilize the glucose oxidase/hydrogen peroxide enzyme electrodes,

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

136

Chapter 5

but they feature novel outer layers to limit glucose diusion to extend the linear
range and to extend the stability and lifetime of the sensors. Others have
worked to eliminate electron-transfer mediators through the use of carbon
nanotubes. Amperometric glucose sensors developed by Tipnis et al.45 follow
the standard design of immobilized glucose oxidase to generate hydrogen
peroxide for oxidation. Glucose concentration is still determined from the
measured steady-state current, but these sensors incorporate a layer-by-layer
(LBL) assembled lm as an outer coat to both limit glucose diusion and
minimize calcication, which can clog sensor pores or crack the outer membrane. LBL coatings of humic acids (HAs) and ferric chloride (FeCl3) of
various thicknesses were compared to optimize the number of layers to tune
glucose diusion from the bulk to the sensor surface.45 Lin et al. constructed
sensors that utilized low density carbon nanotubes (CNT) for direct electron
transfer between the enzyme and the electrode surface so that sensors neither
required an electron mediator nor an outer glucose limiting layer.46,47 These
sensors had a large linear range (230 mM). Even more promising are sensors
fabricated with carbon nanotube bers, spun into nanoyarns.48 The bers are
unwound at the tip to create a brush onto which glucose oxidase is immobilized and an outer polyurethane outer membrane is applied. These sensors
exhibit limits of detection as low as 25 mM, although such low detection levels
are not required for blood glucose measurements. Using nickel or copper
nanoparticles to make a CNT array also lowered the limit of detection.48
Recent glucose sensors have also once again utilized optical detection to take
advantage of technological developments in optical materials and the high
degree of reversibility characteristic of optical sensors. Tierney et al. fabricated
glucose sensors that function by covalent binding of an acrylamide-based
hydrogel to the tip of an optical ber.49 The local glucose concentration swelled
and contracted the phenylboronic acid within the hydrogel, changing the optical length, which was measured by the optical ber. What especially sets these
optical sensors apart is their elimination of the need for a uorescent molecule
or tag. The sensors had high reversibility but could have interferences from
glycoproteins, polysaccharides, elevated blood levels of lactate, and the common anticoagulants heparin and citrate. Sensor function was also signicantly
temperature dependent, as the dierence between room and body temperature
signicantly aected the equilibrium swelling degree. Glucose sensors developed by Paek et al. utilize a reversible glucose binder, rather than a catalytic
enzyme reaction that consumes local glucose and optical detection, that would
not suer signal noise electromagnetic elds, and could use a detector that was
spatially separated from the sensor.50 The sensor functioned by competitive
binding between glucose and a ligand conjugate for the immobilized glucoseselective protein (concanavalin A), which caused a detectable wavelength shift
in the optical sensor. However, in order or preserve sensor lifetime for CGM
and recycle the ligand conjugate, the sensing area must be separated from the
sample by a semipermeable membrane selective for glucose diusion. All of
these sensors show promise in helping to extend the linear working range of
glucose sensors and replace the sensitive glucose oxidase enzyme; however, the

View Online

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

Table 5.1

137

The current commercially available in vivo electrochemical glucose

sensors, sensor lifetime, and the suggested frequency of recalibration.

Company

Sensor

Implantation time

Calibration frequency

Abbott

FreeStyle Navigator

5 days

Medtronic

Guardian

3 days

3 times on rst day of use

(1, 3, 24 h) not
re-calibrated until new
sensor implanted
34 nger pricks/day
(recommended), 1 per
12 hours required

Dexcom

Enlite
Seven Plus CGM

6 days
7 days

1 per 12 hours

biocompatibility of these sensors to thrombosis and immune response from

long-term in vivo implantation has yet to be assessed.
Despite tremendous advances in sensor design, however, none of these designs have yet achieved widespread use, and all existing commercial in vivo
glucose sensors are either based on electrochemical detection of hydrogen
peroxide or redox-mediator species. Indeed, Table 5.1 lists the current commercially available electrochemical glucose sensors, the lifetime of the sensor
in vivo, and the suggested frequency of recalibration. The only FDA approved
devices (by Medtronic-Minimed, Abbott Inc., Dexcom Inc.) require recalibration in vivo several hours after implantation (via repeated in vitro blood tests
by the diabetic patient).710 Even with such measures, the outputs of the devices
are still not reliable enough for the patient to exclusively use values for
monitoring glucose levels to decide on insulin doses.

5.2.3 Sensing Lactate in Blood

As mentioned above, continuous lactate measurements would be most useful
in intensive-care units within hospitals since elevated lactate levels in blood
are a negative predictor of a good outcome for such patients. Indeed, many
critical-care doctors believe that trends in blood-lactate levels provide the best
indicator of tissue oxygenation in patients, and when lactate levels continue to
rise, therapy is failing, and patient survival rate is greatly decreased.1517
Hence, an intravenous catheter that could provide reliable real-time lactate
concentrations would be highly desirable. Several indwelling lactate sensors
have been reported,51,52 and most are based on a design quite similar to the
aforementioned glucose sensors except that lactate oxidase is used as the
immobilized enzyme. The enzyme produces hydrogen peroxide, and thus
oxidation of this product species at a platinum, or other electrode material,
provides the transduction method. Since levels of lactate are generally lower
than glucose, controlling uxes of lactate into the immobilized enzyme, relative
to oxygen, is not quite as stringent, although still important.
Recent developments in lactate sensors have also utilized both electrochemical and optical detection methods. While glucose sensor development

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

138

Chapter 5

focused on applications for in vivo continuous monitoring for diabetic patients

and patients under tight glycemic control, several recent lactate sensors have
been designed for in vitro measurements of cell cultures, as increased lactate
production has been identied as a cancer marker. A potentiometric sensor
based on zinc-oxide nanorods was developed recently by Ibupoto et al. The
ZnO nanorods were grown on gold-coated glass and lactate oxidase was subsequently immobilized on the surface.53 The sensors showed excellent linear
response range, stability for more than three weeks, and high selectivity over
potential interfering species such as ascorbic acid, urea, glucose, galactose and
Mg21 and Ca21 ions. However, the sensors response decreased at pH48 and
with temperatures above room temperature, necessitating a stable operating
pH and temperature for the accurate and consistent measurements needed for a
continuous monitoring system.
Zheng, et al. developed an optical-ber nanosensor based on lactate dehydrogenase capable of measuring extracellular lactate concentration at the
single-cell level. Lactate dehydrogenase immobilized at the tip of an aluminumcoated, heated/pulled optical ber generates NADH as a uorescent byproduct, directly proportional to the local lactate concentration. When excited by
360-nm UV light, the NADH uoresces at 460 nm, allowing extremely sensitive
and selective detection. While this promises to be an excellent technique for
identifying cancerous cells in vitro, no plans have been made to use the sensors
for implanted in vivo measurements.54 Another very recent lactate sensor
designed by Mart n et al. utilizes genetically encoded Forster resonance energy
transfer (FRET) and uorescence-based detection to measure local lactate
extracellular concentration.55 The sensor design proved extremely selective and
very sensitive (single cell), but was pH sensitive and dicult to fabricate. Future
in vivo lactate sensors, designed for continuous monitoring of patients to
indicate health of critically ill patients may still utilize electrochemical
detection, but could also benet from the methods designed to improve the
biocompatibility of implanted in vivo glucose measurements.

5.3 Biocompatibility Issues that Inuence In Vivo Sensor

Performance and Reliability
Several research teams actually fabricated prototype blood-gas/pH-sensing
catheter products that were evaluated rst in animals and then in
humans.13,5659 Despite claims of success by some investigators, signicant
problems remain, both for intravascular and subcutaneous devices. A common
observation for some sensors during in vivo studies is the performance pattern
illustrated in Figure 5.6. Specically, after a given period of time, sensor output
signals often begin to deviate signicantly from the values obtained for discrete
blood samples analyzed in vitro on conventional laboratory instruments. The
pattern is always the same, with intravascular sensor values for pH and PO2
always lower than corresponding in vitro measurements and PCO2 levels always higher than the corresponding in vitro values. This behavior was initially

View Online

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

125
mm 100
Hg
75
50
mm 40
Hg 30
7.5
7.4
7.3

PO2
X XX

X XX X X
X XX

PCO2
X X X XX X X X X XX

X = in vitro blood gas

measurement
= in vivo, continuous blood gas
measurement

pH
X X XXX X X X X X X
2

Figure 5.6

139

4
6
Time (h)

Overall concept of in vivo blood gas/pH patient monitoring system based

on use of intravascular chemical sensors. Sensor inserted in to artery
would be able to continuously monitor oxygen, carbon dioxide and pH
levels in the atrial blood.

thought to arise exclusively from thrombus formation on the surface of implanted sensors, in which platelets and other entrapped metabolically active
cells create a local environment that diers in PO2, PCO2 and pH levels
compared to the bulk blood (due to cellular respiration). Although this eect
will certainly be accentuated by the formation of a thick thrombus layer, even
the adhesion of a single layer of metabolically active cells, e.g. platelets, could
cause a local change in analyte concentrations immediately adjacent to the
surface of the device. Similar errors can be observed for glucose measurements,
given the ability of cells to metabolize glucose, creating lower values near the
surface of cell coated sensors.

5.3.1 Details of Biological Response in Blood

Platelet adhesion and activation on foreign surfaces, such as the outer coatings
of intravascular sensors, is primarily governed by the ability of platelets to
recognize certain proteins that adsorb on the device surfaces once they are
inserted into the bloodstream. Two proteins, von Willebrands Factor (vWF)
and brinogen, play key roles in this process.60 vWF is a circulating protein
that readily adsorbs to collagen and foreign materials that do not an have intact
endothelial layer. Upon adsorption, vWF will bind with glycoprotein Ib (GPIb)
on the surface of platelets to form the GPIb-V-IX complex, which allows
platelet adhesion to the device surface.61 Adherent and activated platelets release a host of factors that promote more platelet adhesion and accelerate the
coagulation cascade including prostaglandins (PGH2 and PGG2), Ca21, and
thromboxane A2.62 Thomboxane A2 promotes the presentation of GP IIb/IIIa
on the surface of platelets that actively binds circulating brinogen, linking
platelets together and locally converting brinogen to insoluble brin.62 Fibrin

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

140

Figure 5.7

Chapter 5

Physiological response to the outer surface of a sensor placed into the

blood stream, ultimately causing activation of platelets and potentially
thrombus formation on all or part of the sensors surface. The coatings of
cells on sensor surface can cause signicant errors in the measurement of
the target analytes.

entangles blood cells and forms the physical barrier that covers devices placed
in blood. In addition to the role of platelets in coagulation, coagulation factor
XII and platelet factor 3 (which is released by adherent platelets) will bind
directly to the foreign surface and also convert prothombin to thrombin, which
will further promote the conversion of brinogen to brin (see Figure 5.7).62 As
mentioned above, even if signicant blood clots are not formed, the presence of
just a single layer of active platelets and other potential cells can alter the
analyte concentrations near the surface of the implanted sensor, yielding inaccurate values compared to the concentration measured in the bulk blood
phase. Within the realm of designing intravascular sensors, it is critical to be
aware that the dynamic nature of this surface interaction is a signicant challenge to obtaining accurate measurements of physiologically important
analytes.

5.3.2 Details of Biological Response in Subcutaneous Tissue

Sensors developed for real-time subcutaneous measurements, most notably
electrochemical glucose sensors,1221 do not suer from accuracy issues related
to platelet activation and thrombus formation. However, such sensors are
subject to their own unique biocompatibility problems that can inuence the
analytical performance of the implanted devices.63,64 Specically, a signicant
tissue inammatory response to a large foreign body (i.e., the sensor) can
lead to encapsulation of the devices within a sheath of leukocytes, macrophages, and broblasts (see Figure 5.8). This sheath can alter local analyte
levels (e.g., glucose) via metabolic reactions (e.g., inamed tissues consume
glucose at an accelerated rate12,32,65,66) and also greatly perturb mass transfer of
the analyte to the sensor surface as well as between the blood and the interstitial

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

Figure 5.8

141

Schematic of inammatory response process that takes place when a

sensor, such as a glucose sensor, is placed subcutaneously. The
increase in inammatory cells in the area and the ultimate encapsulation
of the device in a brous sheath creates a situation where levels of the
analyte near the sensor surface do not reect the concentrations in the
blood.

uid region. Such changes can eectively alter the calibration curve (e.g.,
change sensitivity) for the device and can also greatly aect the lag time of the
sensors response to varying glucose levels in the bulk blood.12,32,67
When foreign materials are placed in subcutaneous tissue, the inammatory
response poses signicant challenges to stable sensor performance. Upon implantation of a sensor, an acute response occurs that lasts from minutes up to
days.68 As is the case with intravascular sensors, nearly instantly, proteins
adsorb to the surface of the device. Due to the need to pass through vascularized tissue to place the sensors in subcutaneous tissue, both plasma and
tissue proteins such as albumin, brinogen, complement factors, globulins,
bronectin, and vitronectin adsorb to the device.68 Mast cells (MCs) present in

View Online

142

Chapter 5

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

69,70

subcutaneous tissue play a central role in regulating inammation.

Upon
injury, MCs will release a series of stored proinammatory agents including
vasoactive amine, chemotactic factors and proteases. As a result of sustained
stimulation, they will then synthesize and continue to release pro-inammatory
mediators that will recruit macrophages and lymphocytes. Additionally, they
release vascular endothelial cell growth factor (VEGF) that is important for
angiogenesis at the site of injury. MCs will also release cytokines that recruit
and activate broblasts that will eventually participate in brous
encapsulation.69
Following initial protein adsorption and release of proinammatory agents,
neutrophils are recruited to the site of implantation and begin attempting to
phagocytose and degrade the foreign device (in this case, the sensor).71 This
involves the release lysosomal proteases, reactive oxygen species, and chemokines that serve as potent signals to recruit cells to the wound site. Neutrophils
will persist at the wound for 12 days before they are replaced by monocytes/
macrophages (monocytes excite the vasculature and mature into macrophages
in tissue near the site of sensor implantation) and lymphocytes.71 The macrophages become activated and release IL-12 and IL-23, reactive nitrogen and
oxygen species (such as nitric oxide (NO), superoxides and H2O2), and proinammatory cytokines.72 Activated macrophages also continue to release chemoattractants to recruit more monocytes/macrophages. It has been shown that
there is heterogeneous activation of macrophages, dependent on many factors
that include the nature of the protein layer present and environmental signals
that recruit monocytes from circulation.73 The release of these active agents is
intended to kill bacteria at the implantation site and biodegrade the foreign
material that is in the subcutaneous tissue. These reactive species will additionally damage healthy tissue surrounding the implant and can lead to a state
of chronic inammation that ultimately leads to brous encapsulation of the
sensor.
As the macrophages adhere to the protein-coated sensor, they undergo a
cytoskeletal rearrangement in order to spread over the surface of the device.
Adherent macrophages then fuse to form foreign-body giant cells (FBGCs) that
further attempt to biodegrade the implanted device. After approximately 12
weeks, chronic inammation leads to the deposition of a collagen matrix by
broblasts to encapsulate the implant and isolate it from native tissue.74
The collagen layer thickness is partially dependent on factors such as the
implant texture, the protein layer adhered to the implant, and cells adhered to
the protein layer. Additionally, movement of the sensor over time can change
the thickness and level of inammation around the implant. It is important to
recognize that the material properties, topography, and placement in vivo aect
the biological response to the implant, but the biological response itself will
further inuence long-term tissue response.74 Again, this is a complicated and
dynamic response that changes over time. This continuous change creates a
highly challenging analytical environment in which reliable measurements must
be made in order to make the use of implanted subcutaneous sensors for in vivo
monitoring a clinical reality.

View Online

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

143

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

5.4 Strategies for Mediating Biological Response to

Implanted Sensors
The commercial development of implantable chemical sensors for routine
blood monitoring in humans has, at this point in time, reached an impasse of
sorts. With only three commercial subcutaneous glucose sensors approved for
limited use and no commercially available intravascular sensors, it appears that
the combination of problems related to cell adhesion, thrombus formation,
inammatory response, etc. yielding poor in vivo sensor performance is a formidable challenge to overcome. Key to ultimate success in this eld will be
developing coatings that can prevent/resist the in vivo biocompatibility issues
that play a dramatic role in the performance of implanted sensors. Fortunately,
there has been signicant fundamental research recently to address these issues.
In the following section, we discuss how various new coatings may provide a
solution to the dynamic biocompatibility problems that have plagued all in vivo
sensors developed to date.

5.4.1 Materials Development for Improved Biological Response

to Implanted Sensors
A great deal of research over the last two decades has focused on designing
materials that can be used to coat or fabricate sensors that are able to reduce
the in vivo biological response that is responsible for loss of sensor function over
time. An excellent review by Wisniewski and Reichert75 details approaches that
have been investigated to modify the materials surface properties to achieve
this goal. Briey, this review discusses use of hydrogels such as crosslinked
poly(hydroxyethyl methacrylate) (PHEMA) or poly(ethylene glycol) (PEG)
that allow water-soluble analytes to diuse to the sensing element, and the
hydrophilic nature of these materials is thought to improve biocompatibility.
Phospholipid surface modication has been used in an attempt to mimic
phospholipids present on cells in the body, thereby decreasing physiological
response toward the sensor. Another approach is to apply thin Naon
(a chemically inert anionic compound that contains both hydrophobic and
hydrophilic properties) coatings that show little adsorption of molecules from
solution or use surfactants applied to sensor surfaces that have shown reduced
protein adsorption. The use of natural materials such as modied cellulose has
been shown to reduce compliment activation, covalent attachment of antifouling agents such as phenylenediamine, diols, and glycols seem to enhance
biocompatibility. Additionally, using diamond-like carbons and controlling the
topography of sensor coats also enhances biocompatibility and stability
over time.
The development of biomimetic hydrogels for sensor coatings involves
PHEMA-based membranes with PEG incorporated and crosslinked with tetraethyleneglycol diacrylate. Such a material has been reported by Justin and
coworkers.76 This UV-cured material was applied to microdisc electrode
arrays, and it was shown that the diusion coecient of analyte was decreased,

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

144

Chapter 5

and pH response of the material from 6.1 to 8.8 caused a change in the impedance signal obtained. This material is expected to improve biocompatibility
of the coated sensor, but tuning of the porosity and crosslinking density is
needed to preserve the sensor response. Silica based sol-gels (SG) have also
been developed77 that include the incorporation of either PEG (SG-PEG),
heparin (SG-HEP), dextran sulfate (SG-DS), Naon (SG-NAF), or polystyrene sulfonate (SG-PSS). The composite materials have been tested for
toxicity in vitro with broblast cultures. The eect on cell proliferation was
dependent on which additive was included. SG-DS showed the most promising
results as glucose sensor coating when tested in BSA and serum, slowing the
growth of broblasts on the surface of the hybrid material, while also allowing
the best glucose response when coated on an amperometric sensor.
Work focused on controlling topography and stability of sensor coatings by
Ju and coworkers78 includes the development of a 3-dimensional porous type-I
collagen scaold crosslinked by nordihydroguaiaretic acid (NDGA) or glutaraldehyde (GA) that was applied to subcutaneous glucose sensors that were
implanted for up to 28 days. The NDGA crosslinked collagen coatings showed
a decrease in inammation compared to bare implanted sensors. Wang and
coworkers79 developed electrospun broporous polyurethane coatings for
implantable glucose sensors. The authors were able to control the dimensions
of the electrospun bers and their density on the sensors to secure certain advantages including a stronger mechanical stability, a greater ability to elongate
to accommodate the swelling enzyme layer on the sensor, a reduction in
mechanical weak spots along the length of the coating, bers that have excellent
interconnected porosity, and a ber structure that mimics natural extracellular
matrix structure. Although this material has not been tested in vivo yet, the
authors point out that in vitro sensor performance was not adversely aected
by the electrospun coating, and there is huge potential to apply biomimetic
electrospun bers to sensors once functioning of the ber coating is validated.
Ai, et al.80 developed a peruorocarboxylic acid ionomer (PFCI) that showed
reduced cracking compared to the commonly used peruorosulfonate ionomer
membrane (PFSI). The advantage was realized because PFCI possess higher
crystallinity and smaller ion clusters compared to PFSI, thereby reducing
mineralization in vivo and reducing cracking, while maintaining the inert
benets of peruoro materials.

5.4.2 Active Releasing Materials for Controlling Biological

Response
Although the development of static materials that are more cognizant of
potential biological response and biocompatibility problems is useful, it is
unlikely that changing the chemistry of the material itself will alleviate
biofouling and subsequent sensor failure when placed in the harsh, dynamic
physiological environment. Work has also been undertaken that relies on more
biomimetic approaches such as inclusion of specic surface active agents such
as CD47 and drug eluting materials, both targeted at controlling explicit

View Online

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

81

145

aspects of the inammatory response. Stachelek, et al. developed a material

that appended a recombinant CD47 to poly(vinyl chloride) (PVC) or
Tecothane, a polyether polyurethane (PU). CD 47 is a transmembrane protein
that binds to specic receptors on leukocytes and macrophages, which causes a
downregulation of inammatory cell attachment. The CD47-PU material was
tested in a rat subdermal implant model for 70 days and showed signicantly
reduced number of attached cells associated with both acute and chronic
inammation.
Several researchers have also proposed development of systems that will
release the anti-inammatory agent, dexamethasone (DX), and/or the angiogenic agent, vascular endothelial growth factor (VEGF). Norton et al.82 developed a hydrogel system based on 2-hydroxyethyl methacrylate that released
both DX and VEGF. The idea was to decrease inammation and encourage
angiogenesis around an implanted glucose sensor. Decreasing inammation
should decrease the thickness of the brous capsule and allow a stable background current to be reached more quickly, and increasing vasculature should
help increase glucose ux to the sensor. It was found that DX alone did decrease inammation after 2 weeks of implantation in rats. VEGF alone did
increase vascularity of the brous capsule, but when DX and VEGF are released at the same time from the hydrogel, while there is still a decrease in
inammation, there is no increase in vascularization. After six weeks, there is
no observable dierence between the drug-releasing hydrogels and control
implants, indicating that the positive eect is not maintained once the drugrelease reservoir is depleted. Sung, et al.83 then investigated a similar approach
by sequentially delivering DX and VEGF from a crosslinked PEG. The release
of DX is faster (kDX 0.5/day) and VEGF is slower (kVEGF 0.4/day). The
dierence in delivery rate is enough to see a marked decrease in inammation
and increase in vascular density in an ex ova chick embryo after 8 days. This is
signicant because it points to the temporal importance of delivering agents in
the appropriate sequence, thus more closely mimicking normal physiological
conditions.
Bhardwaj and coworkers84 tested the temporal eect of DX over much
longer term implants by fabricating dierent poly(lactic-co-glycolic acid)
(PLGA) microspheres that release DX for 30 days or PLGA/poly vinyl alcohol
(PVA) composite microspheres that release DX over a 3-month period. The
microspheres were implanted in rats to test for both acute and chronic inammation. Sustained release of DX was achieved by combining microspheres,
and the continued release did show suppression of inammation for 3 months,
while spheres that exhausted their drug after 30 days lost the ability to decrease
inammation. The same group then modied this approach and developed a
system that dispersed PLGA microspheres in a crosslinked PVA hydrogel.85
The motivation for this was that many dierent drugs in addition to the DX
that was tested could be encapsulated in PLGA microspheres to impart a wide
array of drug-release capability, and the PVA allows both analyte to diuse to
the sensor surface and drug to diuse into the tissue immediately surrounding
the sensor. The smart PLGA microsphere/PVA composite system showed

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

146

Chapter 5

good proof of concept tissue response when tested in both type-1 and type-2
diabetic rats for 1 month. Importantly, this system also recognizes the importance of tailoring the implanted material to the site of implantation and the
specic medical condition under which the sensors are being used and holds the
potential for incorporating additional biological mediators into the sensor
coating.
Another very promising active agent for release from implanted sensors is
nitric oxide (NO). Nitric oxide is a free radical-gas that is released endogenously in many physiological processes.86 Namely, within the context of indwelling sensors, NO is a potent antithrombotic agent87 and plays a role in
mediating the inammatory response in subcutaneous tissue.88 NO is an important signaling molecule in shifting inammation from acute stages to potentially helping in enhancing wound healing. If NO can be released from
intravascular or subcutaneous sensors at the appropriate levels and with good
temporal control, evidence suggests that NO may radically reduce the biological response to these indwelling sensors.89 A class of NO donor that has
been extensively used for improving the biological response toward implanted
sensors is diazeniumdiolates.8790 Figure 5.9 shows the structure of some examples of these molecules that have been used in polymeric materials to impart
NO release to materials for sensor fabrication. Each diazeniumdiolate molecule
will release 2 molecules of NO via thermal decomposition at 37 1C or a protonmediated decomposition mechanism. The rate of decomposition can be controlled by changing the molecular structure of the parent compound used
to form the diazeniumdiolate and/or by tuning the water uptake and pH of
the polymer matrix in which the molecule is embedded. Initial work dispersed
(Z)-1-[N-methyl-N-[6-(N-methylammoniohexyl)amino]]-diazen-1-ium-1,2-diolate
(DMHD/N2O2), evenly through a PDMS matrix coated on an amperometric
oxygen sensor.90 This material showed greatly reduced thromobogenicity,

Figure 5.9

Chemical structures of two diazeniumdiolate type NO donor species that

can be incorporated into polymers to coat implantable sensors, creating
NO release surfaces that can reduce activation and adhesion of platelets in
the blood stream and also decrease inammatory response for subcutaneous sensors. (A) diazeniumdiolated dimethylhexanediamine (B) diazeniumdiolated dibutyhexanediamine.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

147

but it was found that the hydrophilic DMHD/N2O2 leached out of the
polymer matrix, presenting potential problems by allowing the parent
diamine compound used to form the diazeniumdiolate (and present after NO
release) to circulate in the body. Some diamine compounds are known to be
carcinogenic.91
Because of the extremely promising results for inhibiting thrombus formation on the surface of the blood contacting device, work began to either
covalently link the NO donor to the backbone of the polymer matrix used or to
make the donor lipophilic enough that it would not exit the hydrophobic
polymer matrix to enter the aqueous environment of blood. To this end,
modied PDMS materials were developed that contained pendant secondary
amines that could be converted to diazeniumdiolates and result in materials
with dierent NO release properties. Zhang and coworkers92 developed
DACA/N2O2 and TACA/N2O2 PDMS that used diamine and triaminocontaining crosslinking agents to react with hydroxyl-terminated PDMS
chains. Once the crosslinked polymer was cured, it could be converted to the
diazeniumdiolate by reacting in 5 atm of NO. Polyurethanes capable of releasing NO were reported by Reynolds, et al.93 They developed two dierent
methods to synthesize polyurethanes, containing pendent amines in dierent
positions along the backbone of the PU that could be diazeniumdiolated, and
these polyurethans released NO for up to 6 days in aqueous solution at 37 1C.
These PDMSs and PUs are routinely used as membranes for sensors (O2, CO2,
glucose, lactate, etc.) and should be useful for improving the biological response to indwelling sensors.
Another approach to impart NO-release properties to polymers used for
sensor fabrication is to develop a more lipophilic analog of DMHD/N2O2 that
replaces the methyl side groups with butyl side groups (DBHD/N2O2)94 in an
eort to alter the partition coecient enough to keep the NO donor within the
PDMS matrix in which it is dispersed. There are two distinct advantages to
blending a discrete NO donor into the polymer matrix being used to fabricate a
sensor over modifying materials to covalently attach NO donors. First, it
provides the ability to alter the amounts of the NO donor blended into the
material, depending on the level and reservoir of NO donor desired for a given
application. Secondly, the base polymer matrix can be varied without changing
the fundamental synthetic chemistry of the polymer itself. A limitation to this
approach, however, is that the NO donor can still diuse out of the polymer,
albeit a greatly reduced risk compared to DMHD/N2O2. An alternative, which
allows similar advantages but reduces leaching, is the development of derivativized particle llers with NO-releasing surface chemistry, such that these
particles could be incorporated into dierent polymers, thereby imparting NO
release capability to the material without changing the fundamental chemistry
used to synthesize the matrix. These particles also have a much higher potential
loading capacity and release of NO owing to the fact that they have more
potential NO donating groups than just single molecules. Zhang and coworkers95 developed a series of silica particles modied by covalent attachment
of alkylamines to their surface that could be converted into diazeniumdiolates

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

148

Chapter 5

and subsequently blended into dierent polymers. This yielded mechanically

entrapped NO donors that were unable to leach from the polymer surface.
Frost and Meyerho96 also developed fumed silica particles derivatized with
dierent S-nitrosothiols, another class of NO donor (e.g., S-nitroso-cysteine, Snitroso-N-acetylcysteine, and S-nitroso-N-acetylpenicillamine) that could impart NO-release ability with dierent release proles compared to the
diazeniumdiolate-based particle/polymer composites in response to Cu1, or
visible light when blended into plasticized PVC lms.
There has also been extensive work to develop NO-releasing sol-gels (i.e.
xerogels) for potential implantable sensor applications.97100 Shin and
Schoensch97 note that these are attractive materials because they are
synthesized under mild conditions and are chemically exible and generally
porous, facilitating diusion of analytes through the matrix to the sensor.
Marxer and coworkers98,99 report the synthesis of sol-gels prepared by the
hydrolysis and condensation of amine-functionalized alkoxysilanes and alkyltrimethoxysilanes. The crosslinked materials were exposed to 5 atm of NO to
form diazeniumdiolate moieties covalently linked to the material. Four different aminosilanes (isobutyltrimethoxysilane, (aminoethylaminomethyl)phenethyltrimethoxysilane, (aminohexyl)aminopropyltrimethoxy-silane and
N-[3-(trimethoxysilyl)propyl]diethylenetriamine) all had dierent storage and
release properties, demonstrating the ability to achieve a wide range of NOrelease proles from this class of material. Shin et al.100 embedded NOreleasing sol-gel particles into a polyurethane matrix to use as an NO-release
coating for glucose sensors. Roughly 10200 mm NO-releasing particles were
obtained by grinding cured sol-gel monoliths synthesized by using (aminohexyl)aminopropyltrimethoxysilane reacted with methyltrimethoxysilane or
isobutyltrimethoxysilane. This further demonstrates the utility of these materials and the approach of incorporating NO release particles into a wide
variety of polymer matrices to obtain composite materials that are able to release physiologically relevant levels of NO.
These materials that release bioactive agents show immense promise in
controlling and mediating biological response around implanted sensors. One
limitation all these materials possess, however, is a nite reservoir of active
agent that is able to be released from materials preloaded with molecules such
as VEGF, DX or NO. Since these materials also release their agents continuously once release is initiated, this limits the lifetime of a device. NO is, consequently, released even when it may not be needed to limit thrombosis, or even
when a lower level of release may be adequate, and thus the total reservoir is
exhausted sooner than need be. This unnecessary NO release limits the eective
lifetime of the sensor.
In an eort to impart dynamic control over the level of NO released, RSNOderivatized fumed silica particles imbedded in a hydrophobic polymer matrix
showed dynamic NO release when illuminated with dierent intensities of
light.101 The hydrophobic polymer matrix excludes transition metal ions and
ascorbate ions such that NO release is only initiated by illumination, thereby
providing a dynamic on-o trigger for controlling NO release. Gierke et al.102

View Online

149

NO Flux 1010
(molcm2 min1)

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

NO Surface Flux 1010

(molcm2 min1)

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

y = 0.1637x + 0.6759
R2 = 0.9975

2
0

y = 0.1661x + 0.6005
R2 = 0.9991

5
10 15 20
Drive current (mA)

25

Figure 5.10

10

20

30
40
Time (min)

50

60

70

NO generated from S-nitroso-N-acetylpenicillamine-polydimethylsiloxane (SNAP-PDMS) coated on declad region of a 500 mm poly(methylmethacrylate) (PMMA) optical ber with drive current turned on and o
and increasing with each step. The pulses used in this example were in 4
min intervals, light o and then light on with increasing light applied in
each subsequent step from 2, 5, 10, 15 and 20 mA of Idrive applied and
then 15, 10, 5, 2 and 0 mA applied to the LED via the ED. The
corresponding surface uxes of NO generated was 0.95  0.04, 1.50
 0.13, 2.37  0.33, 3.19  0.45, 3.88  0.57, 3.13  0.13, 2.30  0.19,
1.42  0.12, and 0.91  0.04 (all1010 mol  cm2  min1), respectively.
The inset shows that the surface ux of NO generated is linearly related
to the drive current applied to the LED illuminating the coated optical
ber and is identical whether Idrive is applied form 020 mA or 200 mA.
Measurements were made at 22 1C with chemiluminescence detection.
(From ref. 103 with permission.)

then developed a modied polydimethylsiloxane with S-nitroso-N-acetylpenicillamine (SNAP-PDMS) covalently attached to the crosslinking agent
used to form the PDMS. This material, when coated onto declad optical bers,
enables the precise control over the level and duration of NO released in response to a programmed sequence of light generated by a wirelessly controlled
LED light source (see Figure 5.10).103 This system will permit control over NO
release and allow detailed information to be obtained regarding the level and
duration of NO needed to ensure controlled biological response in both blood
contacting and tissue contacting sensor applications. Riccio and coworkers104,105 developed RSNO-containing xerogels that also releases NO in
response to illumination using incandescent bulbs of dierent wattages. These
types of materials that are able to release NO in a dynamically controllable
manner may oer a means to more closely mimic the physiological release of
NO and have the potential to allow implanted sensors to reach a stable,

View Online

150

Chapter 5

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

interfacial blood/tissue response that may ultimately allow for widespread

clinical utility.

5.4.3 Materials to Mimic Biological Form and Function

To more fully mimic biological tissue such as intact endothelium, Zhou
and Meyerho106 created a trilayer system that combined NO release with
surface-immobilized heparin. The idea is to more closely mimic the dual action of
antiplatelet NO and anticoagulant heparin. Wu and coworkers107 later combined
a continuously releasing NO PurSils layer top coated with a CarboSils with
surface-bound thrombomodulin and heparin that were shown to be biologically
active. As material development produces more matrices that exhibit better
biocompatibility, it is likely that combining these improved materials properties
with controlled active agent release and immobilized bioactive molecules will
lead to greatly enhanced performance of indwelling sensors.
An additional step toward creating longer-term implants would be to
eliminate the dependence on the reservoir of active agent present in the
implanted device. Klueh and coworkers108 developed a system in which they
genetically engineering cells to produce excess levels of VEGF. These cells were
entrapped in the tip of a glucose sensor. The goal was to increase vascularization around the implanted sensor to allow ample diusion of glucose to the
sensor. When implanted in ex ova chick embryos, they were able to demonstrate a dramatic increase in the density of blood vessels present around the
implant. This approach relies on the production of VEGF by cells in situ rather
than a preloading of the sensor with the active agent. Another approach to
avoid preloading donor molecules was demonstrated by Oh and Meyerho.109
They developed a lipophilic catalytic copper-containing species that was incorporated into poly(vinyl chloride) (PVC) and demonstrated that RSNOs
present in the bathing solution surrounding the material were decomposed to
release NO catalytically. Hwang and Meyerho110 then took a similar copper
cyclen structure and tethered it to the surface of biomedical-grade polyurethane
that was capable of spontaneously generating NO when in contact with blood.
This approach oers the advantage of using already existing medical grade
polymers and imparting them with NO-release properties. Cha and Meyerho111 then extended this by developing an immobilized organoselenium moiety
that was able to catalytically generate NO from RSNOs present in plasma.
These catalytic systems should be capable of nearly indenite production of NO
from their polymer surfaces because RSNOs are endogenously produced in the
body and circulate in blood and intestinal uid. As long as adequate levels of the
endogenous RSNOs and reducing equivalents are present, the material will
continue to generate physiological NO levels at the polymer/biological interface.

5.5 Summary
The goal of developing implantable blood gas, glucose, and lactate sensors
that can provide reliable, real-time clinically relevant results remains elusive.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

151

While great advances have been made in the development of miniaturized

electrochemical and optical sensors that have adequate sensitivity, selectivity
and stability when operated in vitro, the placement of such devices in the
in vivo environment creates a physiological response that greatly inuences
the quality of analytical results obtained. As summarized above, progress in
this area mandates the development of appropriate sensor coating materials
that can mitigate the normal activation of platelets and other cells when devices are placed intravascularly and the inhibition of the inammatory response when sensors are placed subcutaneously. These coatings must be
capable of diminishing the normal physiological responses to the implanted
sensors, yet they cannot adversely aect analytical response properties, including response times, sensitivity to the target analyte, and selectivity of the
sensor. A wide range of new, passive and active biomaterials coatings have
been devised in recent years specically to target the in vivo sensing challenge.
In the end, it is likely that a combination of immobilized agents and slow
chemical release strategies will be needed to resolve the physiological response
issues that have limited the analytical performance of implanted sensors to
date. Continued research in this area during the coming years will hopefully
achieve success and provide an array of new in vivo sensor tools that will
enhance the quality of health care and disease management for millions of
people worldwide.

References
1. V. Gubala, L. F. Harris, A. J. Ricco, M. X. Tan and D. E. Williams, Anal.
Chem., 2012, 84, 487.
2. P. St. Louis, Clin. Biochem., 2000, 33, 427.
3. F. B. Myers and L. P. Lee, Lab Chip, 2008, 8, 2015.
4. G. McGarraugh and R. Bergenstal, Diab. Technol. Ther., 2009, 11, 145.
5. S. Gelsomino, R. Lorusso, U. Livi, S. Romagnoli, S. M. Romano,
R. Carella, F. Luca`, G. Bille`, F. Matteucci, A. Renzulli, G. Bolotin, G. De
Cicco, P. Stefa`no, J. Maessen and G. F. Gensini, BMC Anesthesiol., 2011,
11, 1.
6. D. H. Bailey, E. J. da Silva and T. H. Cluton-Brock, Anesthesia, 2011,
66, 889.
7. V. R. Kondepati and H. M. Heise, Anal. Bioanal. Chem., 2007, 388, 545.
8. C. M. Girardin, C. Huot, M. Gonthier and E. Delvin, Clin. Biochem.,
2009, 42, 136.
9. E. Cheyne and D. Kerr, Diabet. Metab. Res. Rev., 2002, 18(Suppl 1), S43.
10. T. Kubiak, N. Hermanns, H. J. Schreckling, B Kulzer and T. Haak,
Diabetic Med., 2004, 21, 487.
11. M. C. Frost and M. E. Meyerho, Current Opin. Chem. Biol., 2002, 6, 633.
12. M. C. Frost and M. E. Meyerho, Anal. Chem., 2006, 78, 7370.
13. M. Ganter and A. Zollinger, Brit. J. Anaesth., 2003, 91, 397.
14. M. E. Meyerho, Trends Analyt. Chem., 1993, 12, 257.
15. P. Holloway, S. Benham and A. St. John, Clin. Chim. Acta., 2001, 307, 9.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

152

Chapter 5

16. Shapiro, N. Howell, M. Talmor, D. Nathanson, L. Lisbon, A. Wolfe and

R. Weiss, J. Ann. Emerg. Med., 2005, 45, 524.
17. F. Valenza, G. Aletti, T. Fossali, G. Chevallard, F. Sacconi, M. Irace and
L. Gattinoni, Crit. Care, 2005, 9, 588.
18. C. K. Mahutte, C. S. H. Sasson, J. R. Muro, D. R. Hansmann,
T. P. Maxwell, W. W. Miller and M. Yafuso, J. Clin. Monitor., 1990, 6, 147.
19. Y. Amao, Microchim. Acta, 2003, 143, 1.
20. T.-S. Yeh, C.-S. Chu and Y.-L. Lo, Sens, Actuators B, 2006, 119, 701.
21. C.-S. Chu and Y.-L. Lo, Sens, Actuators B, 2010, 151, 83.
22. J. W. Severinghaus, Ann. NY Acad. Sci., 1968, 148, 115.
23. W. Jin, J. Jiang, Y. Song and C. Bai, Res. Physiol. Neurobiol., 2012,
180, 141.
24. D. A. Nivens, M. V. Schiza and M. Angel, Talanta, 2002, 58, 543.
25. W. Jin, J. Jiang, X. Wang, X. Zhu, G. Wang, Y. Song and C. Bai, Res.
Physiol. Neurobiol., 2011, 177, 183.
26. H. Oenbacher, O. S. Wolfbeis and E. Furlinger, Sens. Actuators, 1986,
9, 73.
27. P. J. Kinlen, J. E. Heider and D. E. Hubbard, Sens. Actuators B, 1994,
22, 13.
28. M. Telting-Diaz, M. E. Collison and M. E. Meyerho, Anal. Chem., 1994,
66, 576.
29. R. K. Meruva and M. E. Meyerho, Biosens. Bioelectron., 1997, 13, 201.
30. M. A. Makos, D. M. Omiatek, A. G. Eweing and M. L. Heien, Langmuir,
2010, 26, 10386.
31. W.-D. Huang, H. Cao, S. Deb, M. Chiao and J. C. Chiao, Sens, Actuators
A, 2011, 169, 1.
32. G. S. Wilson and R. Giord, Biosens. Bioelectron., 2005, 20, 2388.
33. J. W. Mo and W. Smart, Front. Biosci., 2004, 9, 3384.
34. W. K. Ward, L. B. Jansen, E. Anderson, G. Reach, J. C. Klein and
G. S. Wilson, Biosens. Bioelectron., 2002, 17, 181.
35. J. Wang, Electroanalysis, 2001, 13, 983988.
36. T. Koschchinsky and L. Heinemann, Diabet. Metab. Res. Rev., 2001,
17, 113.
37. C. Choleau, J. C. Klein, G. Reach, B. Aussedat, V. Demaria-Pesce,
G. S. Wilson, R. Giord and W. K. Ward, Biosens. Bioelectron., 2002,
17, 647.
38. P. Waeger and M. Hummel, Diabet. Stowechsel und Herz, 2008,
17, 385.
39. N. Tubiana-Ru, J. P. Riveline and D. Dardari, Diabet. Metab, 2007,
33, 415.
40. I. B. Hirsch, D. Armstrong, R. M. Bergenstal, B. Buckingham,
B. P. Childs, W.L. Clarke, A. Peters and H. Wolpert, Diabet. Technol.
Ther., 2008, 10, 232.
41. R. J. McNichols and G. L. Cote, J. Biomed. Opt., 2000, 5, 5.
42. J. Wang, Talanta, 2008, 75, 636.
43. A. Heller, Current Opin. Chem. Biol., 2006, 6, 664.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

153

44. J. S. Marvin and H. W. Hellinga, J. Am. Chem. Soc., 1998, 120, 7.

45. R. Tipnis, M. Vaddiraju, F. Jain, D. Burgess and F.
Papadimitrakopoulos, J. Diabet. Sci. Technol., 2007, 2, 193.
46. Y. Lin, S. Taylor, H. Li, K. A. S. Fernando, L. Qu, W. Wang, L. Gu,
B. Zhou and Y. P. Sun, J. Mater. Chem., 2004, 14, 527.
47. Y. H. Lin, F. Lu, Y. Tu and Z. F. Ren, Nano Lett., 2004, 4, 191.
48. Z. Zhu, L. Garcia-Gancedo, A. Flewett, H. Xie, F. Moussy and W. Milne,
Sensors, 2012, 12, 5996.
49. S. Tierney, B. Falch, D. Hjelme and B. Stokke, Anal. Chem., 2009,
81, 3630.
50. S. Paek, I. Cho, D. Kim, J. Jeon, G. Lim and S. Paek, Biosens. Biolectron.,
2013, 40, 38.
51. Y. Hu, Y. Zhang and G. S. Wilson, Anal. Chim. Acta, 1993, 281, 503.
52. D. A. Baker and D. A. Gough, Anal. Chem., 1995, 67, 1536.
53. Z. Ibupoto, S. Shah, K. Kuhn and M. Willander, Sensors, 2012, 12, 2456.
54. X. Zheng, H. Yang and C. Li, Anal. Chem., 2010, 82, 5082.
55. A. Mart n, S. Ceballo, I. Ruminot, R. Lerchundi, W. Frommer and
L. Barros, Plos One, 2013, 8, 1.
56. B. Venkatesh, T. H. Clutton-Brock and S. P. Hendry, J. Med. Eng.
Technol., 1994, 18, 165.
57. B. Venkatesh, T. H. Clutton-Brock and S. P. Hendry, J. Cardiothorac.
Vasc. Anesth., 1995, 9, 412.
58. B. Venkatesh, T. H. Clutton-Brock and S. P. Hendry, Crit. Care Med.,
1994, 22, 588.
59. S. Divers, W. Marshall, and R. Foster-Smith, in Proceedings of the Conference on Electrolytes, Blood Gases and Other Critical Analytes Burritt, M.
S. Sena, S. F. DOrazio, P., Eds., AACC: Washington, D.C., 1992 pp. 19.
60. M. B. Gorbet and M. V. Sefton, Biomaterials, 2004, 25, 5681.
61. D. Dormann, K. J. Clemetson and B. E. Kehrel, Blood, 2000, 96, 2469.
62. K. C. Dee D. A. Puleo R. Bizios, An Introduction to Tissue-Biomaterial
Interactions, John Wiley & Sons, Inc., Hoboken, New Jersey, USA, 2002,
4, pp 6881.
63. S. P. Nichols, A. Koh, W. L. Storm, J. H. Shin and M. H. Schoensch,
Chem. Rev., 2013, 113, 2528.
64. Z. Zhu, L. Garcia-Gancedo, A. J. Flewitt, H. Xie, F. Moussy and W.
I. Milne, Sensors, 2012, 12, 5996.
65. J. M. Anderson, Cardiovasc. Path., 1993, 2, 33S.
66. J. M. Daley, J. D. Shearer, B. Mastrofrancesco and M. D. Caldwell,
Surgery, 1990, 107, 187.
67. D. A. Baker and D. A. Gough, Anal. Chem., 1996, 68, 1292.
68. J. M. Anderson, A. Rodriguez and D. T. Chang, Semin. Immunol., 2008,
20, 86.
69. U. Klueh, M. Kaur, Y. Qiao and D. L. Kreutzer, Biomaterials, 2010,
31, 4540.
70. P. T. Thevenot, D. W. Baker, H. Weng, M.-W. Sun and L. Tang, Biomaterials, 2011, 32, 8394.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

154

Chapter 5

71. J. D. Bryers, C. M. Giachelli and B. D. Ratner, Biotechnol. Bioeng., 2012,

109, 1898.
72. D. T. Chang, J. A. Jones, H. Meyerson, E. Colton, I. K. Kwon,
T. Matsuda and J. M. Anderson, J. Biomed. Mater. Res., 2008, 87A, 676.
73. J. M. Anderson, Ann. Rev. Mater. Res., 2001, 31, 81.
74. K. L. Helton, B. D. Ratner and N. A. Wisniewski, J. Diabet. Sci. Technol.,
2011, 5, 632.
75. N. Wisniewski and M. Reichert, Colloid Surf. B: Biointerfaces, 2000,
18, 197.
76. Justin, G. Finely, S. Rahman and A. R. A. Guiseppi-Elie, A. Biomed.
Microdevices, 2009, 11, 103.
77. A. Kros, M. Gerritsen, V. S. I. Sprakel, N. A. J. M. Sommerdijk,
J. A. Jansen and R. J. M. Nolte, Sens. Actuators B, 2001, 81, 68.
78. Y. M. Ju, B. Yu, L. West, Y. Moussy and F. Moussy, J. Biomed. Mater.
Res., 2009, 92A, 650.
79. N. Wang, K. Burugapalli, W. Song, J. Halls, F. Moussy, A. Ray and
Y. Zheng, Biomaterials, 2013, 34, 888.
80. F. Ai, Q. Wang, W. Z. Yuan, H. Li, X. Chen, L. Yang, Y. Zhang and
S. Pei, J. Mater. Sci., 2012, 47, 5181.
81. Stachelek, S. J. Finely, M. J. Alferiev, I. S. Wang, F. Tsai, R. K. Eckells,
E. C. Tomczyk, N. Connolly, J. M. Discher, D. E. Eckmann and
D. M. Levy, R.J. Biomaterials, 2011, 32, 4317.
82. L. W. Norton, H. E. Koschwanez, N. A. Wisniewski, B. Klitzman and
W. M. Reichert, J. Mater. Res., 2007, 81A, 858.
83. J. Sung, P. W. Barone, H. Kong and M. S. Strano, Biomaterials, 2009,
30, 622.
84. U. Bhardwaj, R. Sura, F. Papadimitrakopoulos and D. J. Burgess, Int.
J. Pharmaceut., 2010, 384, 78.
85. Wang, Y. Papadimitrakopoulos, F. Burgess J. Control. Release 2013, in
press. (http://dx.doi.org/10.1016/j.jconrel.2012.12.028).
86. S. Moncada, Ann. NY Acad. Sci., 1997, 811, 60.
87. M. W. Radomski, R. M. J. Palmer and S. Moncada, Biochem. Biophys.
Res. Commun., 1987, 148, 14821489.
88. S. P. Nichols, A. Koh, N. L. Brown, M. B. Rose, B. Sun, D. L. Slomnerg,
D. A. Riccio, B. Klitzman and M. H. Schoensch, Biomaterials, 2012,
33, 6305.
89. Y. Wu and M. E. Meyerho, Talanta, 2008, 75, 642.
90. K. A. Mowery, M. H. Schoensch, N. Baliga, J. A. Wahr and
M. E. Meyerho, Electroanalysis, 1999, 11, 681.
91. K. A. Mowery, M. H. Schoensch, J. E. Saavedra, L. K. Keefer and
M. E. Meyerho, Biomaterials, 2000, 21, 9.
92. H. Zhang, G. M. Annich, J. Miskulin, K. Osterholzer, S. I. Merz,
R. H. Bartlett and M. E. Meyerho, Biomaterials, 2002, 23, 1485.
93. M. M. Reynolds, J. A. Hrabie, B. K. Oh, J. K. Politis, M. L. Citro,
L. K. Keefer and M. E. Meyerho, Biomacromolecules, 2006, 7, 987.

View Online

14/10/2013 09:28:02.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00129

In Vivo Sensors for Continuous Monitoring of Blood Gases, Glucose, and Lactate

155

94. T. C. Major, D. O. Brant, M. M. Reynolds, R. H. Bartlett, M.

E. Meyerho, H. Handa and G. M. Annich, Biomaterials, 2010, 31, 2736.
95. H. Zhang, G. M. Annich, J. Miskulin, K. Staniewwicz, K. Osterholzer, S.
I. Merz, R. H. Bartlett and M. E. Meyerho, J. Am. Chem. Soc., 2003,
125, 5015.
96. M. C. Frost and M. E. Meyerho, J. Biomed. Mater., Res., 2005, 72A, 409.
97. J. H. Shin and M. H. Schoensch, Analyst, 2006, 131, 609.
98. S. M. Marxer, A. R. Rothrock, B. J. Nablo, M. E. Robbins and M.
H. Schoensch, Chem. Mater., 2003, 15, 4193.
99. S. M. Marxer, M. E. Robbins and M. Schoensch, H Analyst, 2005,
130, 206.
100. J. H. Shin, S. M. Marxer and M. H. Schoensch, Anal. Chem., 2004,
76, 4543.
101. M. C. Frost and M. E. Meyerho, J. Am. Chem. Soc., 2004, 126, 1348.
102. G. E. Gierke, M. Nielsen and M. C. Frost, Sci. Technol. Adv. Mater.,
2011, 12, 055007.
103. M. A. Starrett, M. Nielsen, D. M. Smeenge, G. E. Romanowicz and
M. C. Frost, Nitric Oxide, 2012, 27, 228.
104. D. A. Riccio, K. P. Dobmeier, E. M. Hetrick, B. J. Privett, H. S. Paul and
M. H. Schoensch, Biomaterials, 2009, 30, 4494.
105. D. A. Riccio, P. N. Coneski, S. P. Nichols, A. D. Broadnax and M.
H. Schoensch, Appl. Mater. Interfaces, 2012, 4, 796.
106. Z. Zhou and M. E. Meyerho, Biomaterials, 2005, 26, 6506.
107. B. Wu, B. Gerlitz, B. W. Grinnell and M. E. Meyerho, Biomaterials,
2007, 28, 4047.
108. U. Klueh, D. I. Dorsky and D. L. Kreutzer, Biomaterials, 2005, 26,
1155.
109. B. K. Oh and M. E. Meyerho, J. Am. Chem. Soc., 2003, 125, 9552.
110. S. Hwang and M. E. Meyerho, Biomaterials, 2008, 29, 2443.
111. W. Cha and M. E. Meyerho, Biomaterials, 2007, 28, 19.

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

CHAPTER 6

Peroxynitrite Electrochemical
Quantication: Recent Advances
and Challenges
SERBAN F. PETEU*a AND SABINE SZUNERITSb
a

National Institute for Research and Development in Chemistry and

Petrochemistry, 202 Splaiul Independentei, Sector 6, 060021 Bucuresti,
Romania; b Institut de Recherche Interdisciplinaire (IRI, USR 3078),
Universite Lille 1, Parc de la Haute Borne, 50 Avenue de Halley, BP 70478,
59658 Villeneuve dAscq, France
*Email: [email protected]

6.1 Introduction
The ageing population worldwide is aected by a high incidence of devastating
diseases, metabolic, cardiovascular and neurodegenerative to name a few,
caused by age-dependent formation of free radicals.13 Free radicals include
atoms, ions, molecules with unpaired valence electrons or open electron shell,
typically highly reactive and short lived.46
Clinical evidence shows the so-called reactive nitrogen and oxygen species
(RNOS), which include free radicals and peroxynitrite (ONOO) to play a
fundamental role in cell signaling, for example being produced to maintain
their integrity when challenged by environmentally unsafe exposures such as
mechanical stress, UV radiation, toxins in air or water, bacteria, or viruses.1
When these species are, however, in excess, the steady state maintained by
physiological processes (homeostasis) is disturbed. At this point, the nitro
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

156

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

157

oxidative metabolic stress develops and its actions over time leads to modications of bio-macromolecules and also interfere with the redox signaling
processes.48
Already established in the last decade as a powerful nitrating, nitrosating and
oxidative triple agent for cellular constituents, peroxynitrite (ONOO or
PON for short) is also being clinically established to exert a variety of detrimental, cytotoxic eects in cells and tissues, both in vitro and in living organisms.4,10,11 Here, ONOO is typically formed via the diusion-controlled
reaction between superoxide radical (O2 ) and nitric oxide (NO).48 The PON
anion is a short-lived species, a potent inducer of cell death, thus perhaps
rightfully hailed as the ugly side of nitric oxide and with a very short life time,
that is only about one second at physiological pH11. By contrast, NO is acting
as a good molecule, whose low-level production is important, including in
protecting organs, such as the liver, from ischemic damage. Consequently, the
trio ( NO; O2 ; ONOO) was notably dubbed5 as the Good, the Bad and the
Ugly. Its fate50 is illustrated in Scheme 6.1.
Starting from the 1990s with the seminal publications by J. S. Beckman,
H. Ischiropoulos, R. Radi, and their groups and others46 suggested the
biological formation of peroxynitrite by the near diusion-controlled reaction
in vivo between nitric oxide and superoxide radicals with their implications for
oxidative injury.4,5,7,1214,50
Ever since, the literature constantly provides solid clinical evidence showing
that endogenously generated peroxynitrite is correlated with acute cytotoxicity
and is thus incriminated in various pathologies and diseases.46 In fact, the
discovery of the peroxynitrite biological formation, role and fate, arguably
continues and accomplishes the nitric oxide crucial discovery as a signaling
molecule in the cardiovascular system15 pioneered by R.F. Furchgott, L.J.
Ignarro and F. Murad, the 1998 Physiology-Medicine Nobel laureates.
RO
ROO

O2

Reactive
Oxygen
Species

O2 +e+
H

Reactive
Nitrogen RNH2
Species

O2

O2
O2
NO2
N2O3

Scheme 6.1

RO
ROO

RH

Fe2+ Fe3+
+e
H+

H2O2

RH

OH

ONOOdecomp.
CO2

H+

ONOOH
decomp.

NO
NO2
N2O3

NO2
NO2+

NO2
N2O3

CO3

The principal RNOS derived from the biological conversion of oxygen

into superoxide ion and nitric oxide. Redrawn with the kind permission
of the American Chemical Society.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

158

Chapter 6

Before dealing with peroxynitrite detection methods and tough challenges, it

seems highly appropriate to briey review its role in biological systems.8,47,48
Herein, we table very succinctly the published clinical evidence for peroxynitrite
toxicity, especially to emphasize the vital need for the in vivo detection of this
analyte. The pathogenic eects of peroxynitrite have been aptly reviewed and
readers are referred to these articles.46
Furthermore, it is equally important to discuss eorts and strategies employed by others to attenuate the toxic eects of ONOO.7,10,3537,43 Some of
these potentially therapeutic methods are based on peroxynitrite isomerization
(A)

(B)

(D)

(C)

(E)
R=
R

O-

O-

TCPP

TPPS

N
N CH2 CH3

N CH3

N
R

(CH2 )n H

R
TE-2-PyP

TM-4-PyP

4-alkylpyridium

R1
N

2,2'-dialkylimidazolinium [R1= (CH2)nH]

2,2'-dimethoxyethylimidazolinium [R1=
(CH2)2OCH3]

N
R1

N
OCH3
O
O

FP15

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

159

to nitrate using metallo-macrocycle molecules and the catalytic reduction of

PON to nitrite and are examined extensively on models of several devastating
diseases models. An interesting fact is that same classes of metallo-macrocycles
decomposing or destroying peroxynitrite are arguably the ones that are also
electrochemically sensitive to PON, since in both cases catalytic processes are
involved. So, one can imagine perhaps designing PON sensitive lms, with detect-and-destroy double capabilities. Some of such PON decomposition
catalysts are shown in Scheme 6.2.

6.2 Challenges Confronted in the Accurate

Quantication of Peroxynitrite
Peroxynitrite is an extremely reactive molecule involved in numerous and
fast reaction pathways, thus the exogenous addition (from outside sources)
of ONOO into (multi)cellular model systems is less than ideal to study its
biological eects.6 There are a number of reviews that outline the advances
and challenges when confronted to PON detection.47,4750 Clinical and
pharmacological studies have discovered that PON is generated in vivo at steady,
low rates often quite the opposite of any analytical approach,4,5 where chemists
employ mostly short bolus exposures to characterize the PON-sensitive chemical
sensors.50 Thus, a gradual permeation of PON using an automatic variable-rate
delivery system (i.e. motor-driven syringe) seems better to model the in vivo
reality.6
The chemical instability of PON is an added diculty to be considered, with
special weight for in vivo studies where some investigators prefer the so-called

Scheme 6.2

Peroxynitrite decomposition catalysts: (A) FeTMPyP Fe (III)mesotetra(N-methyl-4-pyridyl) porphine chloride; (B) MnTDE-2-ImP51
Mn(III) meso-tetrakis(N,N 0 -diethylimidazolium-2-yl) porphyrin; and
(C) FeTPPS 5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato Fe
(III), chloride. (D) The general structure of several PON decomposition
catalysts including several Mn and Fe porphyrins.6 The metalloporphyrins group that have alkylpyridyl substituents are represented
by TE-2-PyP (tetrakis((N-ethyl) pyridynium-2-yl) l porphyrin) and
TMPyP (tetrakis((N-methyl) pyridynium-4-yl) l porphyrin). These readily decompose PON. By changing substituents position and alkyl chain
length, one can inuence both hydrophobicity and redox potential. (E)
Also, several specic R groups are specied.6 The metalloporphyrins
such as Mn(III) tetrakis [N-N 0 -diethylimida-zolium-2-yl]porphyrin
MnTDE-2-ImP5 contain dialkylimidazoline substituents as seen in
Scheme 2B where R1 is en ethyl group. Other acronyms are TCPP
(tetrakis 4-carboxylato-phenyl porphyrin); TPPS (5,10,15,20-tetrakis4 0 -sulfonatophenyl porphyrin). Finally, the FeCl tetrakis-2-(triethylene
glycol monomethyl ether) pyridyl porphyrin FP15 contains a triethylene
glycol monomethyl ether pyridynium moiety as the R-group linked to the
porphyrin ring. The acronyms in bold are referred to in Table 6.1.

View Online

160

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Table 6.1

Chapter 6

Peroxynitrite pathological eects and potential therapeutic

strategies.

Diseases

Pathogenic Eects of Peroxynitrite

Established Clinically in Disease
Models

PON Decomposition
Catalysts Potentially
Therapeutic

PON has been shown to trigger

Cardiac diseases
FP15 reduced myocardial
apoptosis in cardiomyocytes16,17
 Cardiovascular
infarct size in infarction
eects in vitro/
porcine model.38
endothelial18 and vascular smooth
19,20
in vivo
FeTPPS prevented
muscle cells.
 Myocardial
myocardial dysfunction in
Reperfusion (treatment to reduce
ischemia/
an inammatory heartmyocardial damage resulted from
reperfusion
failure model.7
infarction, organ transplantation)
injury
leads to additional tissue injury
 Myocarditis
facilitated by RNOS21,22 upon
 Cardiac
reperfusion, including by an apparent
allograft acute
PON formation increase.23,24
rejection
Myocarditis has been linked to
 Chronic heart
increased nitro-tyrosine immunefailure
reactivity, indicating a pathogenic
role of PON formation.25,26,64
Vascular diseases The pathogenic role of peroxynitrite Using FeTPPS, recent
in atherosclerosis27 and in arterial
 Atherosclerosis
studies have provided
hypertension is supported by recent additional data for role of
 Aging
ndings.28
 Hypertension
peroxynitrite.39
Increased oxidative stress in aging
FP15 improved vascular
also leads to functional inactivation function, cardiac function,
of NO by elevated concentrations of peripheral motor and
O2 favoring an enhanced PON
sensory nerve
conductance.38
formation.29
Chronic
Inammation
 Chronic
arthritis
 Inammation
of toxic origin
 Chronic
inammation
with cancer
potential

In arthritis, the PON could be the

FeTMPS exerted antinal nisher of previously
inammatory eects in
considered NO-dependent
paw-swelling models.40,41
inammation, as ascertained by a FP15 improved the eect of
series of ndings.7
autoimmune arthritis and
PON is suspected to play an important colitis in murine models.42
role in the development of organ
PON decomposition
damage and inammation prompted catalysts might develop
by various toxic chemicals.30
into potential tools for
chronic joint
A correlation between chronic
inammation and tumorigenesis has inammation.6,7
been thought. Long-term exposures
to PON nitro-oxidative stress could
lead to cancer, under inammatory
conditions.7

Neurodegenerative
diseases
 Multiple
sclerosis
 Parkinsons
 Alzheimers
 Huntingtons

The concurrently increased nitric oxide NOS inhibitors were helpful

and superoxide production
in neurodegeneration
associated with neurodegenerative
models, potentially by
disorders favors the in vivo generation averting peroxynitrite
of peroxynitrite.31
formation.7
Numerous ndings seem to strongly
Porphyrinic antioxidants
support the involvement of PON in
improved the outcome in
all neurodegenerative disorders.11
sclerosis murine models.44

View Online

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Table 6.1

161

(Continued)

Diseases
Diabetes and
complications
 Primary
diabetes
 Cardiovascular
dysfunction
 Neuropathy,
nephropathy,
retinopathy

Pathogenic Eects of Peroxynitrite

Established Clinically in Disease
Models

PON Decomposition
Catalysts Potentially
Therapeutic

Type 1 (insulin-dependent) diabetes

is an autoimmune destruction of
insulin producing beta cells,
thus resulting in long-term
hyperglycemia.
While the real trigger is still
unknown, the current ndings
converge for a crucial function of
nitric oxide, superoxide and
peroxynitrite in the pathogenesis
of diabetes and associated
complications32 with PON
potentially implicated in beta-cell
destruction.33,34

Mercaptoalkylguanidines
and FP15 reduced the
type 1 diabetes onset in
murine models.7,45
FP15 improves retinal
microcirculation in
diabetes rodent models.46

The acronyms utilized are (see also Scheme 6.2):

FeTMPS Fe (III)meso-tetra(2,4,6-trimethyl-3,5-disulfonato)porphine chloride
FeTPPS 5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato Fe(III), chloride
FP15 Fe(III)tetrakis-2-(N-triethylene glycol monomethyl ether) pyridyl porphyrin.

genuine ONOO. Several synthetic methods for the preparation of peroxynitrite have been reported and are being successfully used, including:
(a) the ozone reaction with azide ions in presence of low concentration
alkali;51
(b) the auto-oxidation of hydroxylamine in moderately alkaline NaOH
solution;52
(c) the nitrite reaction with acidied H2O2 followed by a base quench;53
(d) the ethoxyethyl nitrite reaction with H2O2 in a basic medium;54
(e) the UV photolysis of solid KNO3;55
(f) the two-phase displacement reaction by the hydroperoxide anion, in the
aqueous phase, on the isoamyl nitrite, the organic phase.56,57
Typically, the as-synthesized PON solution results with a double-digit
milliMolar nal concentration and could be stored at 50 1C being useable for
several months, with a typical loss, nonetheless, of about 2% PON per day.
Nevertheless, its real concentration can be assessed by absorbance in UV-Vis at
302 nm, (e 1700 mol1 L cm1). All these experiments are mostly carried out at
around 24 1C over an ice bath to minimize the spontaneous peroxynitrite
decay.57
Additionally, it has to be kept in mind that synthetic PON stock solutions
need a higher pH (such as pH 9.5) where the analyte is more stable and
characterization can be made adequately. By contrast, at physiological pH

View Online

162

Chapter 6

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

4,5,7

peroxynitrite is prone to fast decomposition.

This is one reason that other
routes for the PON formation have been looked at.
The use of donor solutions of SIN-1 (3-Morpholino-sydnonimine, also called
Linsidomine) has become widely accepted as one way to overcome this limitation. SIN-1 is being used as a PON producer at physiological pH, nowadays,
under ambient molecular oxygen conditions, it produces both nitric oxide and
superoxide as illustrated in Scheme 6.3 and therefore is a PON donor or
generator.5863
SIN-1 liberates superoxide anions and nitric oxide spontaneously in
solution with a 1 : 1 stoichiometry, thereby generating peroxynitrite continuously for a certain period of time. Indeed, in an aerobic aqueous solution SIN-1
decomposes readily to SIN-1A, which in the presence of an oxidant like
oxygen forms the unstable SIN-1A radical cation. The latter liberates NO
and eventually forms the stable end product 3-morpholino-iminoacetonitrile
SIN-1C, as shown in Scheme 6.3.
The quantitative determination of ONOO continues to be complicated, due
to a variety of intrinsic obstacles,4749 including
(i) inherent diculties to accurately reproduce the true in vivo kinetics of
PON in the model experiments in vivo;7,9,10
(ii) the potential of misinterpreting PON concentration as determined, if
the experimental conditions are not carefully optimized;8,47
(iii) the vast complexities of the in vivo real environment, as PON typically
interacts with more than one target per unit time, due to its high
reactivity.10

Scheme 6.3

The oxidative mechanism for ONOO release from SIN-1.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

163

The detection arsenal for peroxynitrite as well as its precursors nitric oxide
and superoxide, includes indirect biochemical assays, optical detection methods
with suitable dyes, electron (paramagnetic) spin resonance spectroscopy and
electrochemical sensing strategies.8,50,71
Biochemical assays measure PON concentrations indirectly, typically based
on up- or down- stream interactions, furthermore suering from additional
unwanted sample processing, such as freezing.6567 Other, more sophisticated
and extremely expensive techniques, such as electron spin resonance spectroscopy (ESR) used with spin-trapping agents, are aorded by only few select
laboratories, are considered too far outside of the scope of this chapter, and
plus these have been reviewed in detail elsewhere.8,50,71
By contrast, electrochemical detection using chemically modied electrodes89
is considered a great analytical technique when it comes to real-time, label-free
and direct measurements of these reactive species.11,47,50 One main drawback,
however, is that the selectivity sometimes implies the use of an additional
membrane or lm, which brings a trade-o between specicity or a slower
sensor response.68,69

6.3 Electrochemical Interfaces and Methods for

Peroxynitrite Detection
Several reviews on the PON electrochemical detection have outlined the
dierent materials and methods used.4,47 While the best perspectives for
label-free, in situ detection to date seem to be oered by electroanalytical
methods, surprisingly there are just a handful of papers dealing with
this issue.6873 Furthermore, most of the reports have not comprehensively
characterized the peroxynitrite-sensitive electrocatalytic matrices such as the
conductive polymer-manganese ion complex73 or the manganese (II)
tetraamino-phthalocyanine coated with poly(4-vinylpyridine).72
Thus, motivated by the challenge to nd better-performing interfaces for the
electrochemical detection of peroxynitrite, the state-of-the-art in PON sensors
will be discussed, including their sensitivity, response time, specicity against
other electroactive molecules with selectivity to peroxyitrite. Also, the advantages and drawbacks will emphasize their potential to be used in real biological
matrices. The discussion will be focused on two main aspects, the reaction
mechanism and the selectivity of the respective PON-sensitive lm, when
available. The other important performance factors such as limit of detection,
response time, sensitivity, are shown in Table 6.2.
Amatore and colleagues68,69 have pioneered a number of interesting methods
and dedicated micro- and nanosensors supported by theoretical and experimental models. From this body of work, herein will be outlined only some of
the eorts leading to peroxynitrite quantication. Simultaneous detection of
NO, O2  and ONOO at dierent potentials was possible with a new experimental method illustrated in Figure 6.1. They reported that a single-cell
oxidative burst could be in fact decomposed into each of the corresponding

View Online

164

Table 6.2

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Reference

Chapter 6

Electrochemical methods for peroxynitrite-sensitive interfaces.

Detection Limit (LoD), Response Time (RT),
Sensitivity (S) and Selectivity

A Amatore et al., Chem. Eur. J., 2001 Special methods.

Amatore et al., Chem. Bio. Chem., The dierent species including PON were
reported as being selectively quantied by
2006
amperometry at their respective potentials
Amatore et al., Anal. Chem., 2010
(see text)
B Xue et al. Anal. Chem., 2000
LoD 18 nM
S 2.4103 nA/nM
Selectivity was examined (see text)
C Wang and Chen Talanta, 2010
LoD 100 nM
S 0.22 nA/nM
Selectivity was examined (see text)
D Bedioui et al., Phys. Chem. Chem. LoD 5000 nM
S 14106 nA/nM
Phys., 2010
Bedioui et al., NO Biol. Chem., 2012 Qualitative indicator of ONOO- bolus pH 7.1
Cortes et al., Electroanalysis, 2007 Selectivity not examined (see text)
Quinton, Griveau, Bedioui
Electrochem. Commun., 2010
Quinton et al., Lab on Chip, 2011
E Koh et al., Anal. Chem., 2010
LoD 2 nM
Selectivity determined (see text)
F Kubant et al., Electroanalysis, 2006 LoD 1 nM
Selectivity not apparent (see text)
G Peteu et al., Biosens. Bioelectron.,
LoD 200 nM
2010
TR 5 s
Peteu et al., An. Chim. Acta, 2013 S 13103 nA/nM
LoD 2500 nM
S 75103 nA/nM
Selectivity examined (see text)
H Szunerits et al., Analyst, 2013
LoD 5 nM
S 7.5 nA/nM
Selectivity not yet examined (see text)

species, specically reduced oxygen (O2 , H2O2) and reduced nitrogen (NO,
NO2, ONOO). The distance between the platinized carbon microelectrode
(CFE) tip and the cell surface was minimized, thus facilitating a fast detection
of species despite their short life spans. Then, amperometric tests performed
on broblast cells at dierent oxidative potentials, have shown that the cell
response level increased with the potential and, accordingly these responses
appeared to integrate multiple electroactive species. The in vivo experimental
data consisted of several oxidative waves, subsequently conrmed by in vitro
experiments with same electrodes for stable solutions. Figure 6.1 shows the
species H2O2, ONOO, NO, NO2 to be directly oxidized at the platinized
CFEs for their respective potentials vs. SSCE. Each species, including the
ONOO, could selectively be quantied by amperometry at their respective
potentials. While this experimental algorithm allowed new investigations in
carefully controlled bioenvironments, its denitive merits for unknown complex congurations in vivo, still remain to be seen.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

Figure 6.1

165

Individual steady-state voltammograms obtained at the same electrodes

for the respective species O2, H2O2, ONOO, NO, NO2, including
peroxynitrite.
Reproduced with kind permission from the PCCP Owner Societies.

Metallo-phthalocyanine and metallo-porphyrin complexes have been tested

as peroxynitrite-sensitive electrochemical interfaces. From the family of the
tetrapyrrolic complexes, the metallo-phthalocyanines (MPC) are chemically
and thermally stable, have good electrocatalytic activity, with diverse coordination and substitutional possibilities conferring versatile physicochemical
properties to MPCs.75,76 MPCs are known to behave as shuttles/relays in a
series of oxidative and reductive electron-transfer processes, even when immobilized on surfaces.7780 Furthermore, the MPCs can be readily tailored by
substituent variations onto the phenyl ring, with their electroactive metal center
typically conferring good electrocatalytic performance.81 Catalytic applications
of MPCs include sensing analytes such as oxygen, peroxide, thiols, dopamine,
epinephrine, nitrite.8183 In aqueous phase, the MPC electrochemistry is illustrated by the multiple and often reversible redox processes localized on their PC
ring and metal center.84 On the other hand, the metallo-porphyrin complexes
have shown similar characteristics.99
The applications targeting PON detection follow. The manganese tetraaminophthalocyanine (MnTAPC) was employed by both the groups of Jin71
and of Bedioui70 for the PON detection on MnTAPC-modied CFEs and Pt
disc microelectrode, respectively.
Jin and colleagues71 electropolymerized the MnTAPC onto 7-mm
diameter CFEs. Based on quantum chemistry, they showed the O* atom from
peroxynitrite ONOO* providing the lone-pair electron to the center of
the Mn atom of the MnTAPC when these form an axial coordination
complex. The reaction in Scheme 6.4 shows that PON was oxidized to nitric
dioxygen and nitrite. The detection method employed was dierential pulse
amperometry (DPA), the current measured being the dierence between the
values at the potentials 0.20 V and 0.0 V vs. Ag/AgCl. The majority of the
electroactive compounds did not interfere with the signal, or were signicantly attenuated, since apparently these were easily oxidized at potentials

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

166

Scheme 6.4

Chapter 6

The detailed catalytic reaction between MnTAPC and peroxynitrite.

Reproduced with kind permission from the American Chemical
Society.

over 0.2 V vs. Ag/AgCl. Additionally, a poly(4-vynilpyridine) lm (PVP)

positively charged was added and it performed as a cation-repulsive barrier
and as a diusional screen for larger biomolecules. The selectivity was tested
against a list of electroactive, interfering agents from biological media including: nitric oxide, nitrite, nitrate, hydrogen peroxide, ascorbic acid, uric
acid, dopamine, epinephrine, glutathione, arginine. However, this improvement in selectivity typically comes at the expense of the response time, which
is a drawback when fast detection is needed.
Bedioui and colleagues70 reported a similar MnTAPC lm polymerized in
DMF at a Pt disc microelectrode, showing the determination of stable PON
in aerobic alkaline aqueous solution of 1.2 M NaOH at pH 10.2 based on the
electrocatalytic reduction current at 0.45 V vs. Ag/AgCl according to an
overall mechanism involving a two-electron reduction of a manganese
oxo intermediate [MnIV O], as previously suggested, although the exact
mechanism has not been identied. The main advantage of both of
these MnTAPC-based PON electrodes is the relatively low potential applied
between 0.0 V and 0.5 V vs. Ag/AgCl in the peroxynitrite detection. This
should solve the problem with most interfering oxidizable compounds.
For another PON-sensitive method developed in the same group,48 the detection was reported at a bare gold RDE (rotating disc electrode) at 7.1 pH in
PBS. Since PON is extremely unstable in neutral solution, the currentpotential
curve was reconstructed by RDE amperometry in the interval (0.5 to 0.7) V vs.
SCE. The detection was reported to occur via electroreducing the conjugated
peroxynitrous acid ONOOH. While the detection is claimed to be selective due
to the 0.1 V operating potential, the PON bolus detection seems qualitative at
best, plus the sensitivity cannot easily be calculated and that is a clear drawback
of this method.
The same group published a proof of concept for an on-chip electrochemical
sensor array (ESA) platform.50 The ESA has individually addressable gold
ultramicroelectrodes (UMEs) of 50 mm in diameter and counter with Ag/AgCl
reference) for simultaneous screening of nitric oxide and peroxynitrite. The
ESA was interfaced with cell-culture wells.
Other electrochemical approaches to detect peroxynitrite include Wang and
Chen72 who employed a lm of cyanocobalamin (vitamin B12) electropolymerized on a modied glassy carbon electrode, showing good electrocatalytic activity for PON oxidation.

View Online

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

poly(CbI)(III) + O=N-O-O
poly(CbI)(II) OONO*
14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

2ONO2*

Scheme 6.5

-e
O2 + 2NO2

167

poly(CbI)(II) OONO*
poly(CbI)(II) + ONO2*

Reaction mechanism showing the redox-active cobalt center in poly(cyanocobalamin) lm to catalyze peroxynitrite decomposition.

Earlier studies have shown that cobalt (II) tetraphenylporphyrins

(Co(II)TPP) can bind the O* atom in peroxynitrite (ONOO*) to form an
axial coordination compound (Co(II) OONOTPP). The reversible binding
reaction of peroxynitrite with Co(II)TPP has been conrmed by kinetic and
mechanistic experiments.72
Because cyanocob(II)alamin or Cbl(II) also contains cobalt(II) porphyrin, the thought was that Cbl(II) binds peroxynitrite to form a stable Cbl(II)
OONO coordination compound. The equations showing the postulated
electrocatalysis mechanism of poly(cyanocobalamin) lm toward peroxynitrite
oxidation are illustrated in Scheme 6.5.
The postulated mechanism demonstrated the redox-active cobalt center in
poly(cyano-cobalamin) lm can catalyze peroxynitrite decomposition through
the formation of intermediates, which is in agreement with the mechanism of
peroxynitrite decay in basic aqueous media proposed previously.72
The selectivity was claimed by studying the interference from: glucose,
dopamine, ascorbic acid, uric acid, xanthine, l-arginine, H2O2, NO and NO3,
NO2. As a drawback, however, there were no data shown, or theoretical
reasoning, in support for these specicity claims.
Koh with colleagues73 synthesized a manganesepolymer complex lm
(polydithienyl-pyrrole-benzoic acid, or PDB) and electropolymerized it onto a
Pt microelectrode. This (Mn-pDPB) lm also had gold nanoparticles electrodeposited on the modied surface and it reportedly enhanced the PON reduction according to the reaction mechanism from Scheme 6.6.
In this work SIN-1 was employed as a peroxynitrite donor solution with a
half-life of 1426 min, during which time PON was produced continuously.
This was timely conrmed with its 302-nm absorbance peak and was,
furthermore, ascertained in inhibitory tests with a peroxynitrite scavenger, such
as uric acid. During such tests, rst the response current rose sharply after
several additions of PON, then the current arrived at a plateau, only to decline
abruptly following the addition of the PON scavenger uric acid.
An increase in selectivity was reported by using an outer layer of polyethyleneimine (PEI), the interferents tested included serotonin, dopamine,
ascorbic acid, bilirubin, and the PON decomposition molecules. The selectivity
was reported as satisfactory, allowing these UMEs to be employed for the PON
determination in vitro on glioma cells. To fully validate this interesting work, it
would be perhaps interesting to repeat the same experiments with the genuine peroxynitrite freshly synthesized.
Malinski and colleagues74 utilized a manganese(III) [22]paracyclophenylporphyrin (MnPCP) that was electrodeposited on a 5-mm long, 250-mm

View Online

168

Chapter 6

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

O N O O
2+
Mn
O

O O N O

Mn2+
O

O
Mn
O

3+

2+

Mn3+
O

Mn
O

Mn2+
O

PON
N
S

Pt

N
S

Pt

N
S

Pt

Scheme 6.6

Reaction mechanism involving the manganese-polymer complex polymerized and the gold nanoparticles electrodeposited from solution.
Reproduced with kind permission from the American Chemical Society.

Scheme 6.7

Electrocatalytic PON oxidation scheme by the hemin-modied GC

electrode.

diameter CFE for the PON sensor. They reported the detection of PON in the
presence of NO and O2.
The chrono-amperometry was conducted with a module integrating 3
working electrodes poised at dierent potentials, namely 0.67 V (for NO),
0.35 V (for O2) and 0.35 V (ONOO), although the nature of the reference
electrode was not clear from the paper. Here, an interesting fact is that the ratio
between the concentration of NO and ONOO, that is [NO]/[ONOO-] was
reported as a potential diagnostic marker, e.g. in cardiovascular diseases. While
this idea is surely interesting, a disadvantage here is that some details seem
unclear and so should be claried in the future.
Elsewhere,57,127 the electropolymerized iron(III) chloroprotoporphyrin IX
(hemin) thin lms were developed to examine the electrocatalytic oxidation of
peroxynitrite on glassy carbon and carbon-ber microelectrodes, in both
standing solutions and in ow-through systems. The reaction scheme is illustrated in Scheme 6.7.57
The PON tests revealed a peak-shaped catalytic wave, indicating a fast
catalytic oxidation of peroxynitrite that was not observed with protoporphyrinonly (without the iron) lms. In reference to Scheme 6.7, this suggests the

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

169

essential role of the bound iron center/atom to allow the oxidative catalytic
turnover of peroxynitrite, also consistent with other reported direct interaction
with metallo-macrocycles. In one study, this interaction led to PON
decomposition, with a set of iron porphyrins in solution is based at the FeIII
center, and yields to the isomerization product, nitrate, through a sequence of
binding, in-cage dissociation mechanism, and rebound, utilizing an
Fe(IV) O intermediate.57
Additionally, a comparison was made between the properties of hemin alone
with hemin-PEDOT lms. Hemin lm exhibited a moderate response mediated
by a direct electrocatalytic oxidation of peroxynitrite. By contrast, after adding
PEDOT to the matrix, the sensitivity of the hemin-PEDOT modied CFEs is
strongly enhanced. This was due to the tortuous porous hybrid surface shaped
as a nanocauliower that is characteristic of the PEDOT grown lm
superimposed with the apparent synergistic eects between PEDOT and the
hemin macrocycle molecules.
The specicity was veried for the main decomposition products of peroxynitrite, in this case (pH 10.5) nitrite and nitrate compared with PON for same
concentration 400 mM. The peak current ratio of PON compared with NO2
and NO3 was 6.7 times higher. Certainly, more work is needed to address the
selectivity against other typical interferents from biological media including:
ascorbic acid, uric acid, dopamine, glutathione, arginine.71
The next section will examine more on the use of PEDOT and graphene to
increase the response sensitivity. In addition, some just-published results on
using a (reduced graphenehemin) PON-sensitive lm will be discussed in
detail.

6.4 Strategies to Increase Response Sensitivity:

Electroactive Polymers or Graphene
Electroactive Polymers (EAPs) are one way to increase the response/sensitivity
of the electrochemical systems, also known as Inherently Conductive Polymers
(ICPs) these two terms being used interchangeably in the literature.8587
The common sense that plastics, unlike metals, do not conduct electricity was
shattered in the 1970s. The discovery and development of conductive polymers
was later rewarded with the 2000 Chemistry Nobel prize to A.J. Heeger, A.G.
MacDiarmid and H. Shirakawa.88
The use of EAPs deposition will be examined next, with a focus on the
electropolymerization method to apply a lm or (multi)layer onto a working
electrode (WE). Herein, the main goal is to exploit the EAPs potential and thus
to impart its qualities to the WE and improving the electroanalytical performance and function adequately as electrode modiers, enhancing the WE
electrocatalytic activity, while also often oering antifouling properties.8587
With their intrinsic advantages, nowadays the polythiophenes may be considered the preferred choice for amperometric sensing, although the polyaniline
and polypyrrole should not be undervalued for electroanalysis in general.90,91

View Online

170

Chapter 6
O

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

N
H

polypyrrole

Scheme 6.8

O
NH

S
PEDOT

polyaniline

Typical conductive polymers are polypyrrole, PEDOT and polyaniline.

One reason for the attractiveness of thiophene is the arsenal of functionalization tools aorded by its monomer ring, in order to confer good conductivity
and high stability to the electrode and/or to create properties such as higher/
lower specic anity or binding towards a specic analytes, or by contrast
interferents.
From the thiophene family, the poly-3,4-ethylenedioxythiophene or
PEDOT (Scheme 6.8), has been especially preferred for several reasons. The
two oxygen atoms coupled to the thiophene rings permits this monomer to be
oxidized at especially low potentials. This aspect, added to the very low steric
hindrance of the condensed rings and the molecule symmetry, leads the
PEDOT to experience a high conductivity and a narrow bandgap, being easily
oxidized with a wide anodic potential window.9294 Furthermore, it was
determined that its condensed ring polarity allows polymerization in aqueous
solution.87
The advancement of nanostructured hybrid materials came recently to oer
new perspectives to increase the sensors response (sensitivity) including the use
of conductive nanohybrids with an EAP complementing the electrocatalytic
activity of a partner such as metal (oxide) nano or microparticles.9598
Metal complexes or ions can also act as (co-)electrocatalysts, while the
overall (PEDOT-metal ions/complex) hybrid lm was shown to improve its
stability and, furthermore, decrease its resistance to charge transfer.87 In one
such case, the iron protoporphyrin (hemin) has been employed for the detection
of peroxynitrite (ONOO) electroassembled on a glassy carbon electrode.8
When the hemin was included in a PEDOT coating, the sensitivity and linear
range of the calibration plots was strongly enhanced with respect to the two
single components, that is the PEDOT or the hemin, separately. Notably, the
heminPEDOT hybrid lm enjoys a fractal three-dimensional cauliower
porous matrix with a very large specic area/volume, similar with the one
shown in Figure 6.2 as ascertained by SEM imaging.100
Additionally, a comparison was made between the properties of hemin
alone with heminPEDOT lms. Hemin lm exhibited a moderate response
(mediated a direct electrocatalytic oxidation of peroxynitrite and. By
contrast, after adding PEDOT to the matrix, the sensitivity of the hemin
PEDOT-modied CFEs is being strongly enhanced. This was due to the
tortuous porous hybrid surface shaped as a nanocauliower that is
characteristic of the PEDOT grown lm superimposed with the apparent
synergistic eects between PEDOT and the hemin macrocycle molecules.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

Figure 6.2

171

Nanohybrid of a metallo-porphyrin and PEDOT. The dimension between

arrows is less than 300 nm.

Further testing has arguably ascertained synergy eects developed between

the PEDOT and the hemin macrocycle molecules: this conclusion was supported by a signicant lowering of the oxidation potential for the nanohybrid
material, as compared with the responses from either one of the two
components when employed separately.8,87
Whereas such hybrid nanomaterials always seem desirable due to the ability
to tailor their properties to the specically desired application, however, the ne
tuning of the properties is often the result of a trial and error approach.87
Additionally, there are few publications outlining the use of PEDOT for sensors, in real biological media.
To optimize a particular electropolymerization method, the experimental
matrix has tens of variables or combinations to be addressed, even when the
same monomer is used, including: the nature of solvent, supporting electrolyte,
monomer concentration, lm thickness, the temperature and each of those or
their combination can aect the usability (structural, morphological, chemical
features) of the electrodeposited lm.
To this aim, the interaction at the lm/electrode interface is important, as
it is considered that the very rst layers of adsorbed molecules on the
WE surface very strongly aect the properties of the whole WE system.101 And
nally, in situ measurements performed by electrochemical, microgravimetric, microscopic and spectroscopic techniques performed on the electropolymerized lm can accurately characterize the thin layer formed on this
conductive substrate.102105

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

172

Chapter 6

Graphene, or other carbon-based materials such as glassy carbon, diamond,

diamond-like carbon, carbon nanotubes, fullerenes are another way to increase the
WE response and sensitivity. They have attracted much interest, because of their
marked structural dierences and their related variety of electronic and electrochemical properties.106109 Their low cost, chemical stability, wide potential
window, rich surface chemistry and enhanced electrocatalytic activity for a
variety of redox reactions have made these materials of high importance for
analytical and industrial electrochemistry. Dierent forms of carbon nanomaterials, including carbon nanotubes, carbon nanobers and fullerenes have
been more recently employed for electroanalytical applications and have been
shown to outperform the classical carbon materials based on graphite, glassy
carbon, diamond and carbon black.110
In contrast to other carbon-based electrodes, carbon nanotubes are typically
grown from carbon-containing gases with the use of catalytic metal nanoparticles, which remain in the nanostructures even after extensive purication
procedures. As a consequence, the electrochemical behavior of carbon nanotubes is dominated by these metal nanoparticles.111 These metallic impurities
are not only causing problems for the construction of reliable sensors and
energy devices112,113 but are also responsible of toxicological hazards within
biological samples.113
The demonstration by Novoselov and Geim, the 2010 Nobel Laureates in
Physics, that single layers of graphene can be isolated from graphite and
identied by microscopy has pushed graphene to the forefront of research
in the design of electrochemical sensors.114,115 Most recently, heminfunctionalized graphene nanosheets have revealed PON activity.116,117
The potential of graphene to support organic molecules such as hemin, and
other porphyrin species, through p  p stacking interactions116122 makes this
material of high interest to study its resulting peroxynitrite activity.117 The
hemin-functionalized reduced graphene oxide (rGO/hemin) matrix can be
formed by reduction of graphene oxide to reduced graphene oxide using
hydrazine116 or dithiothreitol117 followed by immersion into hemin solution, or
by a simultaneous reduction and chemical modication process as illustrated in
Figure 6.3A.116
The incorporation of hemin into the RGO matrix can be revealed by the
presence of the Fe31/21 center of hemin119,122,123 at E 0.4 V (Figure 6.1B).,
where the amount of incorporated hemin was estimated as G was
(3.6  0.5)  108 mol cm2. This value is much larger than that reported for
heminpyrolytic graphite (0.74  108 mol cm2),124 heminmultiwalled carbon
nanotubes (0.27  108 mol cm2)125 and is comparable to hemin immobilized on
highly ordered mesoporous carbon (1.75  108 mol cm2).126
Addition of ONOO, generated from SIN-1 to a GCE/rGO/hemin electrode
shows an oxidative wave at Eox,1 1.17 V (Figure 6.4A). This wave is close to
the oxidation potential of hemin and is thus assigned to the electrochemical
oxidation of hemin on the rGO platform. In the presence of ONOO, an
electrocatalytic oxidations ONOO mediated by oxidized hemin centers seems
to occur (Figure 6.4B).127 The Fe31 center in the rGO/hemin lm is oxidized to

View Online

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

CH3

(A)

H3C
HOOC
OH

COOH

OH
HO

HOOC

H3C
O
HO

N N
Cl Fe
N N

CH3
O
OH

Hemin

HOOC
H3C
H3C
HO O
HOOC

sonication for 5 h
COOH

N
Cl F
N e N
N

CH3

COOH

CH3
O COOH
OH

GO

rGO/hemin

(B)

100

i / A

50
0
50
100
1,6

0,8

0,8

1,6

E / V vs. Ag/AgCl

Figure 6.3

(A) Schematic representation of the synthesis of rGO/hemin; (B) Cyclic

voltammograms of bare glassy carbon (black) and a rGO/hemin modied
GCE electrode prepared by drop casting 525 mL of rGO/hemin (0.5 mg/
mL) onto GCE (black): solution: N2-saturated PBS buer (0.1 M, pH 7.4),
scan rate: 50 mV s1.

(A) 400

(B)

300

200
i/A

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

173

Fe /Fe

100
*

100

0.2

Figure 6.4

0.4

0.6 0.8
1
1.2
E/V vs. Ag/AgCl

1.4

1.6

(A) Cyclic voltammograms in the absence (black) and presence of 200 mM

SIN-1 (pH 7.4) on GCE electrode modied with rGO/hemin (blue); (B)
Proposed mechanism for the electrocatalytic oxidation of ONOO by
rGO/hemin interface.

View Online

174

Chapter 6

(A)

(B)

350

160

E=+1.1 V vs. Ag/AgCl

5 nM
5 nM

200
150
5 nM

50
0

5 nM
0

Figure 6.5

100

43.8

5 nM

100

120

44.3

44.8

43.8

5 nM

140

44.1

i/nA

250

i/nA

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

5 nM
300

60
40

44.0

20

48.3
5

80

10
time/min

15

20

8
12
[ONOO]/nM

16

(A) Typical amperometric response curve obtained using a GCE/rGO/

hemin electrode polarized at 1.1 V vs. Ag/AgCl with subsequent addition of SIN-1 and, (B) Calibration curve.

a high-valent iron form (e.g. iron oxo intermediate, [Fe41 O]) electrochemically at the electrode interface, which, in the presence ONOO, is rereduced back to Fe31 for further turnovers. The performance of the rGO/
hemin modied GCE towards the detection of peroxynitrite was evaluated by
chronoamperometry and showed a PON sensitivity ofE7.5  1.5 nA nM1 with
a low nanomolar detection limit of E5  1 nM, as illustrated in Figure 6.5.

6.5 Conclusions and Perspectives

This chapter oered an overview of the state-of-the-art in electrochemical
quantication of peroxynitrite, a powerful nitrating, nitrosating and oxidative
agent4,10,11 and product of the nitric oxide reacting with superoxide radical.48
In addition, this chapter examined the real challenges to detect this analyte,
together with the most recent advances.
In order to show the vital necessity for in vivo ONOO detection, the medical
facts are purposely outlined, as a solid body of published ndings has clinically
ascertained peroxynitrite to be a potent cell death inducer, in several devastating diseases.46 Equally important are the strategies to develop PON decomposition catalysts as potential therapies to attenuate its adverse eects.
These therapeutic methods are based on peroxynitrite isomerization to nitrate
using metallo-macrocycle molecules and the catalytic reduction of PON to
nitrite.
Currently, a number of these PON decomposition compounds are examined
extensively on diseases models,710 including by few dedicated startups or
spinos. Not surprisingly, the interest from the angel investors to nance
such potentially successful endeavors PON decomposition catalysts could

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

175

be massive in the United States for example, where clearly the medical benets
of any candidates successfully vetted by the Food and Drug Administration
(FDA) would bring vast scientic and nancial paybacks.
Interestingly, some of the classes of metallo-macrocycles decomposing or
destroying peroxynitrite710 are arguably also electrochemically sensitive to
PON,57,7072 since in both cases catalytic redox processes are involved. So, one
can imagine perhaps the synthesis of peroxynitrite-sensitive lms with detectand-destroy double capabilities?
There is little doubt that PON detection poses multiple challenges, especially
in vivo, stemming from its very short lifetime, coupled with an extremely high
reactivity within multiple pathways concurrently, in real environments.710,4749
The optical detection has been available for some time now (e.g. oxidation
of uorescent probes, chemiluminescence, probe nitration), however, it
was not a focus in this chapter. As these optical techniques tend to be
indirect, rather elaborate and relatively dicult to apply real time, it is not
surprising that one turns to electrochemical quantication that is conceptually
simple, more convenient for real-time, label-free and direct in situ measurements especially in the microsensor format89,127130 and including for
peroxynitrite.8,57,7074
The state-of-the-art in the electrochemical quantication of peroxynitrite was
examined in some detail. There are just a handful of groups involved, and some
PON detection limits were reported in the low-single digits.7074,116 This would
seem encouraging when associated with a fast response time. However, a major
diculty for implementation continues to be the specicity and not all authors
investigated it for their PON-sensitive lms. This selectivity is perhaps the most
essential element to consider, particularly for in vivo usability. Equally important are other performance factors such as sensor calibration, the geometry
of the sensitive surface be it macro-, micro-, or nanoscale and the noise in the
response that often has multiple roots.
The eorts to develop specicity for peroxynitrite were reviewed herein and
these included: (i) using an outer membrane to reduce competing interference
by means of charge repulsion or size exclusion71 (ii) employing interrogation
techniques with specic pulse proles, such as dierential pulse amperometry
(DPA) or dierential pulse voltammetry (DPV) that can help discriminate
between competing species,71 (iii) considering the specic redox potential of
each species including PON and designing methods to separate their contribution within the sensor response signal.68,69 However, one needs to be mindful
of possible trade-os between specicity at a cost of slower response time, especially since PON is a short-lived species with about 1 s lifetime.
In particular, two interesting hybrid lms were examined: the
polythiophenehemin57 and the reduced graphene oxidehemin complex116 and
their apparent signicant increase in sensor response was further scrutinized.
In the case of the PEDOT-hemin hybrid lm utilized for the rst time in PON
detection,8 the Fe ions acted as (co-)electrocatalysts, while the overall hybrid
lm seemed to improve the overall stability and furthermore enhance charge
transfer.87 Notably, the hemin-PEDOT hybrid lm was imaged by SEM, as a

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

176

Chapter 6

porous matrix typically shaped as a nanocauliower. When the hemin was

included in a PEDOT coating, the sensitivity and linear range of the calibration
plots was strongly enhanced with respect to any of the two single components
(PEDOT or hemin) tested separately, seemingly due to the extremely high
(specic area per volume) ratio100 superimposed with apparent synergy eects
between PEDOT and the hemin macrocycle molecules. This conclusion
was supported by a signicant lowering of the oxidation potential for the
PEDOThemin nanostructured material, compared with responses from either
one of the two, separately used components.8,87
And nally, the rGO/hemin nanocomposite was successfully tested for the
rst time,116 as a sensitive platform for the detection of peroxynitrite generated
by SIN-1 in PBS. The sensitivity of the rGO/hemin-modied electrodes
for peroxynitrite was 7.5  1.5 nA nM1 with a low nanomolar detection limit
of 5  1 nM. Compared with the rGO formed by hydrazine reduction and
postmodied with hemin, the sensitivity was more than an order of magnitude
lower at 0.6 nA nM1 and a detection limit more than twice larger.
The fundamental causes for such an enhancement are diverse and need to be
further investigated, as several graphene features may contribute collectively to
the enhanced performance, as discussed elsewhere.116
While more eorts are ongoing to better understand the catalytic mechanism
at play, this study clearly highlights for the rst time, the importance of the use
of graphene-supported hemin as a general strategy for the fabrication of highly
sensitive peroxynitrite sensors.116

Acknowledgements
Sabine Szunerits gratefully acknowledges nancial support from the Centre
National de Recherche Scientique (CNRS), the University Lille 1, Nord Pas
de Calais region and the Institut Universitaire de France (IUF). Serban F.
Peteu is grateful for nancial support from the Agentia Nationala de Cercetare
Stiintica (ANCS - UEFISCDI) through PN-II Project 184/2011. The bilateral
Partenariat Hubert Curien Program Brancusi is also acknowledged.

References
1. J. S. Dawane and V. A. Pandit, J. Clin. Diagn. Res., 2012, 6, 1796.
2. E. E. Lomonosova, M. Kirsch, U. Rauen and H. de Groot, Free Rad.
Biol. Med., 1998, 24, 511.
3. M. Tecder-Unal and H. Tufan, Free Rad. Biol. Med., 2002, 33, S43.
4. W. H. Koppenol, J. J. Moreno, W. A. Pryor, H. Ischiropoulos and
J. S. Beckman, Chem. Res. Toxicol., 1992, 5, 834.
5. J. S. Beckman and W. H. Koppenol, Am. J. Physiol.-Cell Physiol., 1996,
271, C1424.
6. C. Szabo, H. Ischiropoulos and R. Radi, Nature Rev. Drug Discov., 2007,
6, 662.
7. P. Pacher, J. S. Beckman and L. Liaudet, Physiol. Rev., 2007, 87, 315.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

177

8. S. F. Peteu, S. Banihani, M. M. Gunesekera, P. Peiris, O. A. Sicuia, and

M. Bayachou, Oxidative Stress: Diagnostics, Prevention, and Therapy, S.
Andreescu and M. Hepel, Am. Chem. Soc., Washington DC, 1st edn,
2011, ACS Symposium Series 1083, 311339.
9. H. Cai and D. G. Harrison, Circulat. Res., 2000, 87, 840.
10. G. Ferrer-Sueta and R. Radi, ACS Chem. Biol., 2009, 4, 161.
11. C. Amatore, S. Arbault, C. Bouton, J. C. Drapier, H. Ghandour and A.
C. W. Koh, Chem. Biol. Chem., 2008, 9, 1472.
12. S. Amemiya, J. D. Guo, H. Xiong and D. A. Gross, Anal. Bioanal. Chem.,
2006, 386, 458.
13. T. Finkel, Curr. Opin. Cell Biol., 2003, 15, 247.
14. B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and
Medicine, 3rd edn, Oxford University Press, Oxford, 1999.
15. R. F. Furchgott, Biosci. Rep., 1999, 19, 235.
16. M. A. Arstall, D. B. Sawyer, R. Fukazawa and R. A. Kelly., Circ. Res.,
1999, 85, 829.
17. V. Grishko, V. Pastukh, V. Solodushko, M. Gillespie, J. Azuma and
S. Schaer, Am. J. Physiol. Heart. Circ. Physiol., 2003, 285, H2364.
18. J. G. Dickhout, G. S. Hossain, L. M. Pozza, J. Zhou, S. Lhotak and
R. C. Austin, Arterioscler. Thromb. Vasc. Biol., 2005, 25, 2623.
19. J. Li, W. Li, J. Su, W. Liu, B. T. Altura and B. M. Altura, Exp. Biol. Med.,
2004, 229, 264.
20. J. Li, W. Li, J. Su, W. Liu, B. T. Altura and B. M. Altura, Neurosci. Lett.,
2003, 350, 173.
21. P. Ferdinandy and R. Schulz, Br. J. Pharmacol., 2003, 138, 532.
22. P. Pacher, R. Schulz, L. Liaudet and C. Szabo, Trends Pharmacol. Sci.,
2005, 26, 302.
23. P. Wang and J. L. Zweier, J. Biol. Chem., 1996, 271, 29223.
24. W. Yasmin, K. D. Strynadka and R. Schulz, Cardiovasc. Res., 1997,
33, 422.
25. A. A. Chaves, M. J. Mihm, B. L. Schanbacher, A. Basuray, C. Liu,
L. W. Ayers and J. A. Bauer, Cardiovasc. Res., 2003, 60, 108.
26. N. W. Kooy, S. J. Lewis, J. A. Royall, Y. Z. Ye, D. R. Kelly and
J. S. Beckman, Crit. Care Med., 1997, 25, 812.
27. J. S. Beckman, Y. Z. Ye, P. G. Anderson, J. Chen, M. A. Accavitti,
M. M. Tarpey and C. R. White, Biol. Chem. Hoppe-Seyler, 1994,
375, 81.
28. N. Escobales and M. J. Crespo, Curr. Vasc. Pharmacol., 2005, 3, 231.
29. A. Csiszar, Z. Ungvari, J. G. Edwards, P. Kaminski, M. S. Wolin,
A. Koller and G. Kaley, Circ. Res., 2002, 90, 1159.
30. A. Denicola and R. Radi, Toxicol., 2005, 208, 273.
31. P. Sarchielli, F. Galli, A. Floridi and V. Gallai, Amino Acids, 2003, 25,
427436.
32. M. J. Stevens, Curr. Vasc. Pharmacol., 2005, 3, 253.
33. W. L. Suarez-Pinzon, C. Szabo and A. Rabinovitch, Diabetes, 1997,
46, 907.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

178

Chapter 6

34. C. Szabo, J. G. Mabley, S. M. Moeller, R. Shimanovich, P. Pacher,

L. Virag, F. G. Soriano, J. H. Van Duzer, W. Williams, A. L. Salzman
and J. T. Groves, Mol. Med., 2002, 8, 571.
35. M. Thiyagarajan, C. L. Kaul and S. S. Sharma, Br. J. Pharmacol., 2004,
142, 899.
36. S. S. Sharma, S. Munusamy, M. Thiyagarajan and C. L. Kaul, J. Neurosurg., 2004, 101, 669.
37. D. Liu, F. Bao, D. S. Prough and D. S. Dewitt, J. Neurotrauma, 2005,
22, 1123.
38. C. Bianchi, H. Wakiyama, R. Faro, T. Khan, J. D. McCully, S. Levitsky,
C. Szab and F. W. Sellke, Ann. Thorac. Surg., 74, 1201.
39. S. Cuzzocrea, E. Mazzon, R. Di Paola, E. Esposito, H. Macarthur,
G. M. Matuschak and D. Salvemini, J. Pharmacol. Exp. Ther., 2006, 319, 73.
40. D. Salvemini, Z. Q. Wang, D. M. Bourdon, M. K. Stern, M. G. Currie
and P. T. Manning, Eur. J. Pharmacol., 1996, 303, 217.
41. D. Salvemini, Z. Q. Wang, M. K. Stern, M. G. Currie and T. P. Misko,
Proc. Natl. Acad. Sci. USA, 1998, 95, 2659.
42. J. G. Mabley, J. G. L. Liaudet, P. Pacher, G. Southan, J. T. Groves,
A. L. Salzman and C. Szabo, Mol. Med., 2002, 8, 581590.
43. G. S. Scott, S. Spitsin, R. Kean, T. Mikheeva, H. Koprowski and
D. Hooper, Proc. Natl Acad. Sci. USA., 2002, 99, 1630316308.
44. J. P. Crow, N. Y. Calingasan, J. Chen, J. L. Hill and M. F. Beal, Ann.
Neurol., 2005, 58, 258.
45. A. H. Cross, M. San, M. K. Stern and T. P. Misko, J. Neuroimmunol.,
2001, 107, 21.
46. R. Sugawara, T. Hikichi, T. N. Kitaya, F. Mori, I. Nagaoka, A. Yoshida
and C. Szabo, Curr. Eye Res, 2004, 29, 1116.
47. S. Borgmann, Analyt. Bioanalyt. Chem., 2009, 394, 95.
48. F. Bedioui, D. Quinton, S. Griveau and T. Nyokong, Phys. Chem. Chem.
Phys., 2010, 12, 9976.
49. M. M. Tarpey and I. Fridovich, Circulation Res., 2001, 89, 224.
50. C. Amatore, S. Arbault, M. Guille and F. Lemaitre, Chem. Rev., 2008,
108, 2585.
51. R. M. Uppu, G. L. Squadrito, R. Cueto and W. A. Pryor, Nitric Oxide,
Pt. B, 1996, 269, 285.
52. M. N. Hughes, H. G. Nicklin and W. A. Sackrule, J. Chem. Soc.-Inorg.
Phys. Theor., 1971, 23, 3722.
53. J. W. H. Reed and W. L. Jolly, J. Am. Chem. Soc., 1974, 96, 1248.
54. J. R. Leis, M. E. Pena and A. Rios, J. Chem. Soc.-Chem. Commun., 1993,
16, 1298.
55. R. C. Plumb and J. O. Edwards, J. Phys. Chem., 1992, 96, 3245.
56. R. M. Uppu and W. A. Pryor, Analyt. Biochem., 1996, 236, 242.
57. S. F. Peteu, P. Peiris, E. Gebremichael and M. Bayachou, Biosens.
Bioelectr., 2010, 25, 1914.
58. E. de la Fuente, G. Villagra and S. Bollo, Electroanalysis, 2007, 19, 1518.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

179

59. A. Chatterjee, A. Smith, J. D. Catravas and S. M. Black, Free Rad. Biol.

Med., 2008, 45, S44.
60. L. K. Cuddy, A. C. Gordon, S. A. G. Black, E. Jaworski, S. S.
G. Ferguson and R. J. Rylett, J. Neurosci., 2012, 32, 5573.
61. J. B. Gramsbergen, T. R. Larsen, S. Rossen and P. Roepstor,
J. Neurochem., 2007, 101, 31.
62. M. J. Kohr, C. J. Traynham, S. R. Roof, J. P. Davis and M. T. Ziolo,
J. Mol. Cell. Cardiol., 2010, 48, 645.
63. J. L. Trackey, T. F. Uliasz and S. J. Hewett, J. Neurochem., 2001, 79, 445.
64. P.T. Liu, C.E. Hock, R. Nagele, and P.Y.K. Wong, Am. J. Physiol.-Heart
Circulat. Physiol., 1997, 272, H2327.
65. Z. H. Taha, Talanta, 2003, 61, 310.
66. A. Daiber, M. Mehl and V. Ullrich, Nitric Oxide, 1998, 2, 259.
67. F. L. Kiechle and T. Malinski, Ann. Clin. Lab. Sci., 1996, 26, 501.
68. C. Amatore, S. Arbault and D. Bruce, P. de Oliveira, M. Erard, and M.
Vuillaume, Chem. Eur. J., 2001, 7, 4171.
69. C. Amatore, S. Arbault and A. C. W. Koh, Analyt. Chem., 2010,
82, 1411.
70. J. S. Cortes, S. G. Granados, A. A. Ordaz, J. A. L. Jimenez, S. Griveau
and F. Bedioui, Electroanal., 2007, 19, 61.
71. J. Xue, X. Y. Ying, J. S. Chen, Y. H. Xian, L. T. Jin and J. Jin, Analyt.
Chem., 2000, 72, 5313.
72. Y. Wang and Z.-Z. Chen, Talanta, 2010, 82, 534.
73. W. C. A. Koh, J. I. Son, E. S. Choe and Y.-B. Shim, Analyt. Chem., 2010,
82, 10075.
74. R. Kubant, C. Malinski, A. Burewics and T. Malinski, Electroanal., 2006,
18, 410.
75. F. Dumoulin, M. Durmus, V. Ahsen and T. Nyokong, Coord. Chem.
Rev., 2010, 254, 2792.
76. N. Sehlotho, T. Nyokong, J. H. Zagal and F. Bedioui, Electrochim. Acta,
2006, 51, 5125.
77. X. H. Qi and R. P. Baldwin, Electroanal., 1994, 6, 353.
78. X. H. Qi and R. P. Baldwin, J. Electrochem. Soc., 1996, 143, 1283.
79. M. Gulppi, S. Griveau, F. Bedioui and J. H. Zagal, Electrochim. Acta,
2001, 46, 3397.
80. T. Mugadza and T. Nyokong, J. Colloid Interface Sci., 2011, 354, 437.
81. K. S. Lokesh, K. De Wael and A. Adriaens, Langmuir, 2010, 26, 17665.
82. J. Arguello, H. A. Magosso, R. Landers and Y. Gushikem, J. Electroan.
Chem., 2008, 617, 45.
83. J. H. Zagal, S. Griveau, K. I. Ozoemena, T. Nyokong and F. Bedioui,
J. Nanosci. Nanotechnol., 2009, 9, 2201.
84. A. B. P. Lever, J. Porphyrins Phthalocyanines, 1999, 3, 488499.
85. G. G. Wallace and L. A. P. Kane-Maguire, Adv. Mater., 2002, 14, 953.
86. G. G. Wallace, M. Smyth and H. Zhao, TRAC-Trends Analyt. Chem.,
1999, 18, 245.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

180

Chapter 6

87. C. Zanardi, F. Terzi and R. Seeber, Analyt. Bioanalyt. Chem., 2012,

405, 509.
88. H. Shirakawa, E. J. Louis, A. G. MacDiarmid, C. K. Chiang and A.
J. Heeger, J. Chem. Soc., Chem. Commun., 1977, 16, 578.
89. R. W. Murray, Acc. Chem. Res., 1980, 13, 135.
90. C. Dhand, M. Das, M. Datta and B. D. Malhotra, Biosens. Bioelectron.,
011, 26, 2811.
91. A. Ramanavcius, A. Ramanavicliene and A. Malinauskas, Electrochim.
Acta, 2006, 51, 6025.
92. H. J. Ahonen, J. Lukkari and J. Kankare, Macromolecules, 2000,
33, 6787.
93. M. Lapkowski and A. Pron, Synth. Met., 2000, 110, 79.
94. L. Groenendaal, G. Zotti, P. H. Aubert, M. S. Waybright and
J. R. Reynolds, Adv. Mater., 2003, 15, 855.
95. D. W. Hatchett and M. Josowicz, Chem. Rev., 2008, 108, 746.
96. C. Zanardi, F. Terzi, L. Pigani, and R. Seeber, Encyclopedia of Polymer
Composites: Properties, Performance and Applications, M. Lechkov and S.
Prandzheva, Nova, New York, 2011, pp 174.
97. A. Bhattacharya and A. De, Prog. Solid. State. Chem., 1996, 24, 141.
98. R. Gangopadhyay and A. De, Chem. Mater., 2000, 12, 608.
99. S. Bahaduri and D. Mukesh, Homogeneous Catalysis: Mechanisms and
Industrial Applications, S. Bahaduri and D. Mukesh, 2000, Wiley, New
York.
100. S. F. Peteu and M. Bayachou, unpublished results.
101. F. Terzi, L. Pasquali, M. Montecchi, S. Nannarone, A. Viinikanoja,
T. Aaritalo, M. Salomaki, J. Lukkari, B. P. Doyle and R. Seeber, J. Phys.
Chem. C, 2011, 115, 17836.
102. M. Innocenti, F. Loglio, L. Pigani, R. Seeber, F. Terzi and R. Udisti,
Electrochim. Acta, 2010, 50, 1497.
103. A. R. Goncalves, M. E. Ghica and C. M. A. Brett, Electrochim. Acta,
2011, 56, 3685.
104. V. Ruiz, A. Colina, A. Heras, J. Lopez-Palacios and R. Seeber, Helv.
Chim. Acta, 2001, 84, 3628.
105. L. Pigani, A. Heras, A. Colina, R. Seeber and J. Lopez-Palacios, Electrochem. Commun., 2004, 6, 1192.
106. R. L. McCrerry, Chem. Rev., 2008, 108, 2646.
107. D. Tsasis, N. Tagmatarchis, A. Bianco and M. Prato, Chem. Rev., 2002,
106, 1105.
108. A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov and
A. K. Geim, Rev. Mod. Phys., 2009, 81, 109.
109. A. Fujishima, Y. Einaga, T. N. Rao, and D. A. Tryk, Diamond Electrochemistry, Elsevier, co-publication with BKC Inc. Japan, 2005.
110. E. Mercey, R. Sadir, E. Maillart, A. Roget, F. Baleux, H. Lortat-Jacob
and T. Livache, Analyt. Chem., 2008, 80, 3476.
111. M. Pumera, Langmuir, 2007, 23, 6453.

View Online

14/10/2013 09:28:06.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00156

Peroxynitrite Electrochemical Quantication: Recent Advances and Challenges

181

112. C. E. Banks, A. Crossley, C. Salter, S. J. Wilkins and R. G. Compton,

Angew. Chem. Int. Ed., 2006, 45, 2533.
113. M. Pumera and Y. Miyahara, Nanoscale, 2009.
114. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,
S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306,
666669.
115. K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, S. V. Khotkevich,
S. V. Morozov and A. K. Geim, Proc. Natl. Acad. Sci. USA, 2005,
102, 10451.
116. R. Oprea, S. F. Peteu, P. Subramanian, W. Qi, E. Pichonat, H. Happy, M.
Bayachou, R. Boukherroub, and S. Szunerits, Analyst, 2013, 138, 4345.
117. A. A. Vernekar and G. Mugesh, Chem. Eur. J., 2012, 18, 15122.
118. T. Xue, S. Jiang, Y. Qu, Q. Su, R. Cheng, S. Dubin, C.-Y. Chiu,
R. B. Kaner, Y. Huang and X. Duan, Angew. Chem. Int. Ed., 2012,
51, 3822.
119. Y. Guo, J. Li and S. Dong, Sens. Actuators B Chem., 2011, 160, 295.
120. C. X. Guo, Y. Lei and C. M. Li, Electroanalysis, 2011, 23, 885.
121. C. Xu, J. Li, X. Wang, J. Wang, L. Wan, Y. Li, M. Zhang, X. Shang and
Y. Yang, Mater. Chem. Phys., 2012, 132, 858.
122. Y. Guo, L. Deng, J. Li, S. Guo, E. Wang and S. Dong, ACS Nano, 2011,
5, 1282.
123. J. Wei, J. Qiu, L. Li, L. Ren, X. Zhang, J. Chaudhuri and S. Wang,
Nanotechnol, 2012, 23, 335707.
124. J. Chen, U. Wollenberger, F. Lisdat, B. Ge and F. W. Scheller, Sens.
Actuators B, 2000, 70, 115.
125. J. S. Ye, Y. Wen, W. DeZhang, H. F. Cui, L. M. Gan, G. Q. Xu and
F. S. Sheu, J. Electroanal. Chem., 2004, 562, 241.
126. H. Cao, X. Sun, Y. Zhang, C. Hu and N. Jia, Analyt. Meth., 2012, 4, 2412.
127. S. F. Peteu and M. Bayachou, Analyt. Chim. Acta, 2013, 780, 81.
128. D. Emerson, S. F. Peteu and R. M. Woden, Biotechnol. Tech., 1996, 10,
673.
129. S. F. Peteu, M. T. Widman and R. M. Worden, Biosens. Bioelectr., 1998,
780, 81.
130. R. Gutierez-Osuna and A. Hierlmann, Ann. Rev. Bioanalyt. Chem., 2010,
3, 250.

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

CHAPTER 7

The Diagnosis of
Myelodysplastic Syndromes
ALISON S. THOMAS AND CHRISTOPHER MCNAMARA*
Department of Haematology, Royal Free Hospital, London NW3 2QG,
United Kingdom
*Email: [email protected]
Myelodysplastic syndromes (MDS) comprise a heterogeneous group of clonal
stem-cell disorders characterised by dysplasia, ineective haematopoiesis and a
risk of transformation to acute myeloid leukaemia. MDS may arise de novo or
secondary to exposure to mutagenic agents. MDS is usually rst suspected due
to the presence of peripheral blood cytopenias or evidence of dysplasia on a
blood lm and the diagnosis is conrmed by the presence of otherwise unexplained signicant dysplasia in the bone marrow, often in the presence of
cytogenetic abnormalities. Originally described as a preleukaemic condition in
the early 20th century, the cytopenias that form part of the clinical picture of
MDS can give rise to transfusion dependence, recurrent infections and a
bleeding diathesis and the complications of bone marrow failure are the main
cause of mortality in patients, rather than leukaemic transformation. The
denition of MDS has evolved since the initial proposal of a French
AmericanBritish classication in 19761 as our understanding of the pathophysiology and natural history of the syndrome has advanced. Alongside this
has been the development of subclassications, aimed at dening groups with
diering diagnostic and prognostic features within this highly heterogeneous
disorder. Despite many scientic advances there remains, however, no single
biological or genetic marker that reliably identies all cases of MDS and
morphology remains the cornerstone of diagnosis. Accurate diagnosis of MDS

RSC Detection Science Series No. 2

Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

182

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

183

relies on the ability to dierentiate MDS from other causes of myelodysplasia

and a willingness to revisit the diagnosis in cases of uncertainty. There is a
pressing need for increased engagement in research programs that seek to
further our understanding of the molecular and biological mechanisms
underpinning the pathogenesis of this disorder.

7.1 Morphological Features of Myelodysplasia

Dysplasia, derived from the Greek words dns (dys-, abnormal) and plasiB
(plasis, growth or formation) describes the aberrant maturation of cells.
Myeloid dysplasia manifests morphologically as an excess of immature myeloid
precursors with cytoplasmic and nuclear abnormalities. These precursors may
be abnormally located within the bone marrow compartment, as visible on
trephine biopsy, denoted as abnormal location of immature precursors (ALIP).
Although patients have peripheral blood cytopenias, the bone marrow is
commonly hypercellular with an increased proportion of immature cells (leftshifted).
Myeloid dysplasia can aect any of the myeloid lineages (erythroid, granulocytic or megakaryocytic) and commonly aects more than one lineage
even if there is only a cytopenia in one lineage. Detection of dysplasia requires
examination of blood lms and bone marrow aspirates that are stained to a
high standard. Inadequately stained slides may lead to erroneous identication of abnormalities of granularity. Diagnosis of MDS requires the presence of unequivocal dysplasia in Z10% of at least one of the myeloid series.
Accurate determination of the percentage of blast cells is required for subclassication and at least 200 cells should be counted in the peripheral blood
and 500 in the bone marrow. Common dysplastic features in each lineage are
summarised in Table 7.1 and shown in Figure 7.1.

7.1.1 Erythroid Dysplasia

In MDS, dysplasia is most commonly seen in the erythroid series. The most
reliable dysplastic features are nuclear abnormalities including: nuclear
budding,2,3 internuclear bridging (connections between the nuclei of two
dierent cells), multinuclearity and megaloblastic features. Cytoplasmic features include patchy haemoglobinisation, intercytoplasmic bridging and
vacuolisation. An iron stain (e.g. Perls stain) should be performed on all
bone-marrow aspirates of patients with suspected MDS, both to assess iron
stores, as iron deciency can cause dysplasia, and to look for the presence of
ringed sideroblasts, the percentage of which is important in the subclassication of MDS. Ringed sideroblasts are erythroid precursors which
have at least ve siderotic granules in a perinucelar distribution, covering at
least a third of the nuclear circumference.4 An increased number of ringed
sideroblasts is a feature of a specic subclass of MDS, refractory anaemia
with ringed sideroblasts (RARS).

View Online

184

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

Table 7.1

Chapter 7

Common morphological features of dysplasia.

Lineage

Peripheral blood

Bone-marrow aspirate Bone-marrow trephine

Erythroid

Poikilocytosis:

Megaloblastoid
Defective maturation
features
in erythroid islands
Nuclear budding and
erythroblasts at
multinucleation
same stage of
Internuclear and
maturation
intercytoplasmic
bridging
Cytoplasmic
vacuolisation
Patchy
haemoglobinisation
Nuclear-cytoplasmic
asynchrony
Ringed sideroblasts

Ovalomacrocytes
Dacrocytes
Elliptocytes
HowellJolly bodies
Nucleated RBC
Basophilic stippling
Acquired HbH
features, including
H inclusions
Granulocytic
Hyposegmentation
(pseudo-Pelger
Huet) or
hypersegmentation
Hypogranulation
Ring-shaped nuclei
Megakaryocytic Platelet anisocytosis

Increased blasts
Maturation arrest
Nuclear-cytoplasmic
asynchrony

Abnormal location of
precursors (ALIP):
central rather than
periosteal location of
blasts

Hypolobated or
Megakaryocyte
nonlobated nuclei
clusters
Multiple separated
Monolobated
nuclei
megakaryocytes
Micromegakaryocytes

7.1.2 Granulocytic Dysplasia

Dysgranulopoiesis is characterised by nuclear abnormalities with both hypolobation (pseudo-PelgerHuet) and hyperlobation.5 Cytoplasmic granulation
may be decreased. There is often an increase in immature myeloid precursors
(e.g. blasts, promyelocytes). Accurate enumeration of the blast percentage is
important for diagnosis of acute myeloid leukaemia (myeloid blasts Z20%) and
for subclassication and prognosis of MDS. Myeloid blasts comprise: myeloblasts, monoblasts and megakaryoblasts; erythroid precursors are not counted
as blasts except in cases of erythroleukaemia. To ensure consistency in identication of blast cells in MDS a consensus statement containing detailed descriptions and diagrams of blast cells has been published.4

7.1.3 Megakaryocytic Dysplasia

It is recommended that at least 30 megakaryocytes are evaluated; megakaryocytic
dysplasia is more easily identied on trephine sections. Micromegakaryocytes

View Online

185

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

Figure 7.1

Morphological features of myelodysplasia. (A) erythroid intercytoplasmic

bridging and red blood cell poikilocytosis, MGG100. (B) erythroid
dysplasia and hypogranular myelocyte, MGG100. (C) multinucleate
late erythroblast with megaloblastic features, erythroblast with patchy
haemoglobinisation and abnormally lobated neutrophil, MGG100. (D)
ringed sideroblasts, Perls stain,100. (E) monolobated megakaryocytes in
5q- syndrome, MGG10. (F) trephine showing dysplastic megakaryocytes, (H) and (E),10. (G) pseudo-Pelger-Huet abnormality, MGG100
(H) hypogranular myelocyte, MGG100.

and multinucleate megakaryocytes are considered most reliable for the diagnosis of MDS.6 Micromegakaryocytes are small megakaryocytes with hypolobated nuclei. Multinucleate megakaryocytes have multiple, widely separated
nuclei in contrast to the polylobated nucleus of normal megakaryocytes and
there may be increased variation in size.

View Online

186

Chapter 7

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

7.2 Dierential Diagnosis of Myelodysplasia

Bone marrow ndings should be interpreted in the context of the clinical presentation and full blood count abnormalities. Minor dysplastic features are a
common nding in bone marrow specimens from healthy volunteers:7 minor
dyserythropoietic features were present in 84% of specimens (patchy haemoglobinisation 62%, cytoplasmic bridging 42%, binuclearity 24%); the percentage of erythroid cells with dysplastic features was 17%; 38% of samples
had a small number of dysplastic megakaryocytes. Diagnosis of MDS requires
the presence of signicant (i.e. Z10% of cells in any one lineage) dysplasia and
the exclusion of causes of nonclonal dysplasia (see Table 7.2).

7.2.1 Metabolic Abnormalities

Nutritional deciencies are a common cause of cytopenias and dysplasia. B12
and folate deciency both cause dyserythropoiesis and, if severe, pancytopaenia
and multilineage dysplasia. Deciencies of other B vitamins may also result in
similar abnormalities.8 Copper deciency, whether primary or secondary
to zinc excess, is increasingly recognised as a cause of reversible dysplasia.9
Exposure to heavy metals, particularly arsenic, which can be found in contaminated water supplies in certain geographic regions, may result in bonemarrow failure and dysplasia.10,11

7.2.2 Drugs and Toxins

Cytotoxic drugs can result in signicant dysplasia in any myeloid lineage. Many
drugs have been reported as causing myelodysplastic changes including:

Table 7.2

Dierential diagnosis of myelodysplasia.

Vitamin/minerals

Infections
Drugs and toxins

Inherited disorders
Other myeloid neoplasms

B12 deciency
Folate deciency
Copper deciency (may be secondary to zinc excess)
Arsenic poisoning
HIV
Parvovirus
Any severe infection
Alcohol
Chemotherapeutic agents
Mycophenolate mofetil
Ganciclovir
Sodium valproate
G-CSF, erythropoietin
Congenital dyserythropoietic anaemia
Inherited bone-marrow-failure syndromes
Acute myeloid leukaemia
Myeloproliferative syndromes

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

187

 growth factor therapy: granulocyte colony stimulating factor results in

hypergranulation, abnormal granulocytic nuclear lobation and an increased proportion of myeloblasts, particularly in the peripheral blood;
 ganciclovir;12,13
 sodium valproate;
 mycophenolate mofetil;12,13
 co-trimoxazole;
 alemtuzumab.14

7.2.3 Infectious Diseases

Myelodysplasia in very common in patients with HIV15 as a consequence of the
direct eects of the virus on haematopoiesis as well as opportunistic infections
and eects of antiretrovirals. Acute episodes of sepsis can stimulate a
reactive bone marrow response that may exhibit dysplastic features. Less
commonly, parvovirus infection can cause transient red cell aplasia with
dyserythropoiesis.16

7.2.4 Congenital Disorders

In isolated erythroid dysplasia, the diagnosis of congenital dyserythropoietic
anaemia should be considered. In younger patients, congenital bone-marrow
failure syndromes, which often have a myelodysplastic phase prior to leukaemic
transformation, should be included in the dierential. These include: Fanconi
anaemia, dyskeratosis congenita, DiamondBlackfan syndrome and
SchwachmannDiamond syndrome. Although the dysplasia in these inherited
bone marrow failure syndromes is clonal, they are due to an inherited rather
than acquired genetic defect and as such represent distinct disorders which
require specialised management.

7.2.5 Acute Myeloid Leukaemia and Other Haematopoietic

Neoplasms
The presence of Z20% myeloid blasts in the peripheral blood or bone marrow
leads to a diagnosis of acute myeloid leukaemia (AML) regardless of the
presence or absence of dysplasia. Patients with a preceding history of cytotoxic
chemotherapy or radiotherapy who develop myelodysplastic changes as a late
complication are classied as having a therapy-related myeloid neoplasm, a
subclassication of AML, regardless of blast percentage and have a generally
poor prognosis. The most commonly implicated cytotoxic agents are alkylating
agents and topoisomerase II inhibitors. As discussed below, some patients may
present with features of a myeloproliferative disorder (e.g. thrombocytosis or
monocytosis) but also have evidence of myelodysplasia. The nal diagnosis of
many of these patients is within the category of myelodysplastic/myeloproliferative neoplasms, which includes chronic myelomonocytic leukaemia.

View Online

188

Chapter 7

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

7.3 Initial Assessment of a Patient with Possible MDS

Initial assessment of a patient with suspected MDS starts with a full clinical
history, including a detailed drug history and history of any malignancy.
Clinical examination should be unremarkable in MDS, except for occasional
bruising in thrombocytopenic patients. Lymphadenopathy and hepatosplenomegaly are rare in MDS and their presence should prompt consideration of
alternative diagnoses.
Initial investigations should include assessment of peripheral blood cytopenias and morphology and detection of causes of nonclonal dysplasia (see
Figure 7.2). Anaemia is usually normocytic or macrocytic. An increased red cell
distribution width (RDW) reects variability in red cell size, but a normal mean
cell haemoglobin concentration (MCHC), indicates a normal amount of
haemoglobin for the cell size. Neutropenia is common and circulating immature cells may be seen on the blood lm. Whilst thrombocytopenia is more
common, occasionally thrombocytosis may be present. The importance of the
haematologist examining the blood lm for themselves prior to deciding to
perform a bone-marrow biopsy cannot be overemphasised.
Any metabolic deciencies should be corrected prior to further assessment of
the patient. Acute illness results in bone marrow stress and apparent dysplasia
making the diagnosis of MDS challenging in these circumstances; further investigation should be deferred until recovery if possible. If cytopenias remain
unexplained, then a bone marrow aspirate and trephine should be performed,
with a bone marrow sample being taken for cytogenetic analysis. The results of
the aspirate, trephine and cytogenetic analysis should be reviewed together in
conjunction with the clinical history and a summative, integrated report agreed.
In some cases a concrete diagnosis may not be possible and a period of observation followed by repeat evaluation is required.

7.4 Additional Diagnostic Tools for MDS

7.4.1 Bone Marrow Aspirate Cytochemistry
An iron stain (e.g. Perls stain) should be performed on all aspirates to enable
enumeration of ringed sideroblasts. Cytochemical staining can be performed to
conrm myeloid lineage of blasts (e.g. Sudan Black B) or to dierentiate granulocytic from monocytic cells (e.g. dual esterase) but in most laboratories
these stains have been superseded by immunological techniques.

7.4.2 Bone Marrow Trephine Biopsy

A bone marrow trephine complements an aspirate, with dierent morphological information available from each.17 Whilst the aspirate provides detailed
cytological information, the trephine provides an overview of bone marrow
architecture and maturation. Dysplastic features may been evident as disruption of normal bone marrow architecture18,19 (see Table 7.1). Cellularity is

View Online

189

The Diagnosis of Myelodysplastic Syndromes

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

Unexplained cytopenia
or abnormal blood film

History and examination incl drug

history, previous malignancy, risk
factors for HIV

Initial investigations:
FBC, including reticulocyte count
Blood film review
Haematinics: B12, folate, ferritin
Consider: copper, zinc, HIV test

Potential cause of myelodysplasia

identified

YES

Treat and re-review

NO
Acute illness e.g. sepsis

YES

Re-review once recovered

NO
Receiving growth factor therapy
(e.g. G-CSF, erythropoietin)

Unable to diagnose MDS whilst

receiving stop if possible

YES

NO
Unexplained cytopenia(s): proceed to
bone marrow aspirate +/ trephine

10% cells in at least 1 myeloid lineage

with unequivocal dysplasia

YES

MDS

Sub-classify
(see Figure 7.3)

NO
Specific cytogenetic abnormality present
which is considered presumptive evidence
of MDS

YES

NO
Idiopathic cytopenia of
undetermined significance

Figure 7.2

No clear diagnosis, re-assess

and consider repeat bone
marrow in 6 months

Approach to the patient with suspected MDS.

normally increased and reactive changes secondary to the disease process may
also be seen, e.g. mastocytosis, brosis and increased histiocytes.20 Immunohistochemistry for lineage conrmation and enumeration of blast cells can be
performed on trephine biopsies and are of particular benet if the aspirate has
been suboptimal. Importantly, a trephine biopsy may also provide evidence of

View Online

190

Chapter 7

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

an alternative disease process as explanation of cytopenias or dysplasia, e.g.

hairy cell leukaemia, metastatic carcinoma.

7.4.3 Cytogenetics
Cytogenetic abnormalities are seen in 4070% patients with MDS21,22 and their
detection has an important role both in the diagnosis and prognosis of MDS.
A wide range of chromosomal abnormalities have been described in MDS including numerical abnormalities (monosomy or trisomy), inversions and interstitial deletions within one chromosome and balanced translocations between two
chromosomes. Cytogenetic studies should be performed on bone marrow samples
rather than peripheral blood. Karyotyping can be undertaken using conventional
metaphase analysis. This has the advantage of detecting a wide range of abnormalities but can be dicult in the presence of complex abnormalities and can
be hampered by poor sample quality. Fluorescent in situ hybridisation (FISH)
can be performed on metaphase chromosomes or interphase nuclei but will only
detect the abnormalities specic to the panel of probes utilised.23

7.4.3.1

Role of Cytogenetics in Diagnosis of MDS

In cases of suspected MDS with cytopenias but minimal dysplasia, the presence
of specic cytogenetic abnormalities is considered sucient for the diagnosis of
MDS to be made in the absence of signicant morphological dysplasia. These
abnormalities are listed in Table 7.3. MDS can precede AML; there are three
cytogenetic abnormalities which are considered diagnostic of AML, regardless
of blast percentage: t(8;21)(q22.q22), inv(16)(p13.1q22) or t(16;16)(p13.1;q22)
or t(15;17)(q22;q21.1)

7.4.3.2

Role of Cytogenetics in Subclassication and Prognosis

of MDS

The nding of 5q- as the sole cytogenetic abnormality comprises a specic

subclass of MDS, 5q- syndrome. This is associated with characteristic morphological abnormalities of hypolobated and monolobated megakaryocytes; it
Table 7.3

Cytogenetic abnormalities, which if present, are diagnostic

of MDS in the presence of peripheral blood cytopenias,
even in the absence of signicant dysplasia.

Deletions
Translocations and inversions

5 or del(5q) or 7/del(7q)
Del(9q) or del(11q) or del(12q) or del(13q)
t(1;3)(p36.3;q21)
t(2;11)(p21;q23)
inv(3)(q21;q26.2)
t(3;21)(q26.2;q22.1)
t(6;9)(p23;q34)
t(11;16)(q23;p13.3)
t(17p) or i(17q) with loss of 17p

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

191

is associated with a favourable prognosis. Currently, there are no other cytogenetic abnormalities associated with specic clinical and morphological features. Monosomy 7 or complex cytogenetic abnormalities (45 cytogenetic
abnormalities) are associated with a poorer prognosis21 and are more common
in RAEB than other subclasses of MDS.22

7.4.4 Immunophenotyping
As myeloid cells develop, dierent genes are transcribed and their proteins expressed intracellularly and on the cell surface membrane. Loss of coordination of
the processes governing normal cell maturation occurs in MDS and may result in
aberrant or asynchronous expression of maturation-associated antigens on
myeloid cells. The aberrant expression or absence of a cell surface antigen may
provide supporting evidence for the presence of an abnormal cell population.
The role of immunophenotyping in the diagnosis of MDS remains, however,
controversial. The specicity of abnormalities in antigen expression to MDS has
not been extensively studied.24 Aberrant expression has been described in reactive left-shift25 and in nonhaematological disorders that can mimic MDS.26
Further data is required to clarify the role of immunophenotyping in the diagnosis of MDS. This requires standardisation of sample processing, antibody
panels and analysis protocols to enable comparisons between dierent data
sets.27 The current guidance is that immunophenotyping reports should be descriptive; the presence of Z3 phenotypic abnormalities may be suggestive of
MDS, but is not currently considered diagnostic.2 Immunophenotyping should
not be used in place of morphology to enumerate blast cells. Samples are often
haemodilute, leading to underestimation and not all blast cells express CD34.

7.5 Diagnostic Diculties in MDS

7.5.1 Unexplained Cytopenias in the Presence of Minimal
Dysplasia
In some cases, patients will have unexplained cytopenias but will not meet the
minimum morphological criteria for signicant dysplasia of at least 10% of
cells of at least one myeloid lineage showing unequivocal dysplasia. In these
cases, additional eort should be made to ensure that the diagnostic samples
are of adequate quality and that other causes of cytopenias and dysplasia have
been excluded. There are some specic cytogenetic abnormalities (see
Table 7.3), which if present, are considered sucient for a presumptive diagnosis of MDS to be made in the absence of signicant dysplasia.2 Diagnosis can
be dicult in the absence of cytogenetic abnormalities; virtually all
bone marrow specimens will have at least some mild dysplasia. In a study of
bone marrows of patients with dierent causes of anaemia,28 dysplasia was seen
in at least one lineage in 93% of iron-deciency anaemia, 92% of megaloblastic
anaemia and 84% of haemolytic anaemia. The percentage of dysplastic cells
was similar in these anaemias to patients given a diagnosis of refractory

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

192

Chapter 7

anaemia with normal cytogenetics. In patients with unexplained signicant

cytopenias (haemoglobin o10 g/dL, neutrophils o1.8109/L or platelets
o100109/L) the term idiopathic cytopenia of undetermined signicance has
been suggested, emphasising the lack of diagnostic clarity. These patients
should be monitored and a repeat bone marrow evaluation after six months
considered.

7.5.2 Myelioproliferative/Myelodysplastic Overlap Syndromes

Persistent monocytosis (41109/L) accompanied by the presence of dysplasia
in one or myeloid lineages but with a blast percentage of o20% is diagnostic
for chronic myelomonocytic leukaemia (CMML). Although originally considered part of MDS, CMML also shares some features of myeloproliferative
neoplasms (MPN) and in the WHO 3rd edition, a new category of myelodysplastic/myeloproliferative neoplasms was created, incorporating CMML.
Some patients have laboratory and morphological features consistent with a
diagnosis of MDS, but have additional myeloproliferative features (e.g.
platelets Z450109/L or WBC 413109/L) without a preceding history of
myeloproliferative disorders. These patients are also considered to have a
myelodysplastic/myeloproliferative neoplasm. Within this group are patients
who have anaemia accompanied by thrombocytosis, with ringed sideroblasts
present in excess on the bone marrow. Up to 2/3 of these patients will have a
JAK2 V617F mutation;29,30 in patients initially diagnosed with refractory anaemia with ringed sideroblasts who develop a thrombocytosis, acquisition of
the JAK2 V617F mutation at the time of development of thrombocytosis has
been demonstrated.30 These patients fall into a provisional entity of refractory
anaemia with ring sideroblasts associated with marked thrombocytosis.

7.5.3 Cytopenias in the Presence of a Hypocellular Bone

Marrow
In approximately 10% of cases of MDS the bone marrow is hypocellular and
this can raise a diagnostic challenge in dierentiating MDS from aplastic anaemia. In aplastic anaemia, cellular elements will be reduced but should be
morphologically normal. The peripheral blood lm should be carefully examined for evidence of dysplasia if leukocyte counts are signicantly reduced a
lm may need to be prepared from the buy coat. If the aspirate is paucicellular, blast percentage can be estimated by immunohistochemistry on the trephine specimen. Cytogenetic abnormalities are far commoner in MDS than
aplastic anaemia and the presence of clonal cytogenetic abnormalities would
favour a diagnosis of MDS.31

7.5.4 Dysplasia in the Presence of Bone Marrow Fibrosis

Bone marrow brosis occurs in approximately 10% MDS. Diculty in obtaining an aspirate, which is often haemodilute, reduces the quantity of

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

193

diagnostic material available and a good quality trephine specimen of adequate

length is essential. The dierential diagnosis includes: acute panmyelosis with
myelobrosis, acute megakaryoblastic leukaemia and chronic myeloproliferative neoplasms. Immunohistochemistry should be performed to enumerate
CD34 positive blast cells and to identify megakaryocytic cells. Evaluation for
the JAK2 V617F mutation may be helpful in distinguishing brotic MDS from
myelobrosis; the mutation is present in approximately 50% of myelobrosis
but is infrequent in MDS.32

7.6 Subclassication of MDS

Once the diagnosis of MDS has been established, classication into a specic
subtype of MDS is required. Classications are consensus denitions of specic
disease entities and currently the most widely used classication of MDS is the
WHO 4th Edition Classication.2 This incorporates clinical and laboratory
data to arrive at diagnostic criteria and subclassications of MDS. Subclassication is important given the heterogeneous features and variable
prognosis of MDS. The classication of MDS is designed to be used in routine
clinical practice, as well as to provide common denitions for clinical trials and
other studies. It incorporates data available at the time it was written; in this
rapidly advancing eld new data is constantly being published as our understanding of MDS evolves. This is recognised in the 4th edition of the WHO
classication where provisional entities are described based on new ndings
where further data is needed to clarify and conrm the clinical signicance.

7.6.1 Original FAB Classication

The original classication of MDS was the 1976 FrenchAmericanBritish (FAB)
classication,1 where the category of dysmyelopoietic syndromes was designed to
help distinguish acute leukaemia from less acute disorders that carried a risk of
progression to acute leukaemia. Two categories were described: CMML and refractory anaemia with excess blasts (RAEB). Diagnosis was based entirely on the
morphological appearances of peripheral blood and bone marrow. RAEB was
dened as bone marrow blasts 1030%, with 3050% representing probable
progression to AML. AML was diagnosed if blasts were 450%.
The FAB classication was revised in 198233 in recognition that there was a
wide range of dysplastic morphological appearances and that the risk of
transformation to AML was variable and, to a certain extent, correlated with
morphological features. The expansion to ve categories aimed to enable better
assessment of the prognostic signicance of these morphological features:
 Refractory anaemia (RA): o1% blasts in peripheral blood (PB), o5%
blasts in bone marrow (BM); granulocytic and megakaryocytic series
usually normal
 Refractory anaemia with ringed sideroblasts (RARS): o1% blasts in PB,
o5% blasts in BM, ringed sideroblasts 415% all nucleated cells in BM

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

194

Chapter 7

 RA with excess blasts (RAEB): 520% blasts in BM and o5% in PB

 CMML: monocytes 41109/L, o5% blasts in PB
 RAEB in transformation (RAEB-T): 45% blasts in PB, 2030% blasts in
BM or Auer rods.

7.6.2 WHO Classication 3rd Edition and 4th Edition

Whilst the FAB classications were based solely on morphological appearances, the World Health Organisation (WHO) classications incorporate
morphological, genetic, immunophenotypic and clinical features in an attempt
to identify distinct molecular pathways underlying neoplastic disorders.
Haematological malignancies, including MDS, were rst included in the 3rd
edition of the WHO Classication of tumours in 2001. These have been subsequently revised and a 4th edition published in 2008.24 The WHO classication
introduced a new category, MDS/MPN overlap syndromes, recognising the
dysplastic and proliferative features present in some patients such as those with
CMML. Building on the FAB classication, the WHO classications have
continued to rene subclassications of MDS, aiming to be prognostically
relevant. A diagnostic algorithm for subclassifying MDS is presented in
Figure 7.3. As the availability of therapies for MDS increases, it is vital to
correctly subclassify patients and all diagnostic material should be reviewed by
haematopathologists experienced in the diagnosis of MDS. A review of the
diagnosis of patients referred to a tertiary cancer centre in the USA revealed a
discordant diagnosis in 12% with over two-thirds of these having higher risk
disease (i.e. RAEB) than initially diagnosed.34 A number of prognostic scoring
systems are available (e.g. revised IPSS,35 WPSS36), aimed at stratifying risk of
overall survival and evolution to AML. These incorporate information on blast
percentages, cytopenias and cytogenetic abnormalities alongside other
clinical data (e.g. transfusion dependence); sucient information should be
provided on the integrated diagnostic report to enable clinicians to utilise these
scores.

7.7 Evolving Understanding of the Pathophysiology of

MDS
Much is still unknown about the molecular mechanisms underlying the
pathogenesis of MDS. There are many similarities in the gene abnormalities
identied in MDS and AML, including abnormalities in tumour suppressor
genes and oncogenes, but progression from MDS to AML is highly variable.
Whilst metaphase analysis and FISH-based approaches allow detection of
translocations and deletions, array-based approaches enable identication of
gene over/underexpression, loss of heterozygosity and copy-number variations.
In addition to abnormalities in genes involved in regulation of the cell cycle,
epigenetic mechanisms of gene expression regulation and mRNA splicing may
be important in the pathogenesis of MDS.

View Online

195

The Diagnosis of Myelodysplastic Syndromes

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

Dysplasia in 10% cells

in at least one myeloid
lineage

Cytogenetic
abnormality
considered
diagnostic of MDS

No

Yes

No

Yes

Blasts 20% in
PB or BM

Yes

Re-consider
diagnosis

MDS-U

Acute myeloid leukaemia

No
Previous
chemotherapy/radiotherapy

Yes

t-MDS

No
Auer rods OR 5-19% blasts in
PB or 10-19% blasts in BM

Yes

RAEB-2

No

2-4% blasts in PB or
5-9% blasts in BM

Yes

RAEB-1

No
Dysplasia of 10% cells in 2 myeloid
lineages

Yes

RCMD

No
Pancytopenia

Yes

MDS-U

No

Thrombocytopenia

Anaemia

Neutropenia

RT

Figure 7.3

RN

15% ringed
sideroblasts

Yes

No
RARS
RA

Subclassication of MDS.

7.7.1 Altered Methylation

Altered methylation is an important epigenetic mechanism of gene expression
regulation and mutations have been found in the TET2 (ten-eleven-translocation) gene family in up to 23% patients with MDS37 and isocitrate dehydrogenase (IDH) oncogenes in 3.6% patients,38 both resulting in
hypermethylation.

View Online

196

Chapter 7

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

7.7.2 Abnormalities in RNA Splicing Machinery

RNA splicing is accomplished by a complex of ve small nuclear ribonucleoproteins (snRNPs) and other proteins on pre-mRNAs that form the spliceosome. Mutations within any of these proteins can result in aberrantly spliced
mRNA. Whole exome sequencing of MDS specimens has revealed novel mutations involving components of the spliceosome, particularly those involved in
3 0 splice site recognition39 in up to 40% of MDS patients.39,40 Mutations have
also been found in patients with other haematological and nonhaematological
malignancies, but at lower levels (15%). Mutations in SRSF2 and ZRSR2 are
more common in those with RAEB and associated with poorer survival.40
Mutations in SF3B1 are found in up to 75% of patients with ringed sideroblasts;39 knockdown of SF3B1 in normal cells leads to development of ringed
sideroblasts.41
MDS represents a heterogeneous group of clonal disorders of haematopoietic stem cell maturation. First identied by the presence of bone marrow
dysplasia and a propensity to develop into acute leukaemia, it is clear that
current diagnostic criteria for MDS incorporate a number of dierent entities
with vastly dierent natural histories, ranging from RAEB-2 with a high risk of
leukaemia and short life expectancy to refractory anaemia, which may run a
more indolent course. Care should be given in arriving at the diagnosis, particularly in the absence of signicant dysplasia and cytogenetic abnormalities;
observation and re-evaluation are prudent in these circumstances. Common
nonclonal causes of MDS must be excluded prior to the initiation of any
specic (e.g. chemotherapeutic) therapy for MDS. As molecular diagnostic
techniques advance, coupled with large cohort analyses of patients, so too may
our understanding of the underlying pathogenesis of the myelodysplastic syndromes. In future it may be possible to diagnose and subclassify these syndromes on the basis of molecular and genetic, rather than primarily
morphological abnormalities.

References
1. J. M. Bennett, D. Catovsky, M. T. Daniel, G. Flandrin, D. A. Galton,
H. R. Gralnick and C. Sultan, Proposals for the classication of the acute
leukaemias. French-American-British (FAB) co-operative group, Br. J.
Haematol., 1976, 33(4), 451458.
2. S. H. Swerdlow, E. Campo, N. L. Harris, E. S. Jae, S. A. Pileri, H. Stein,
J. Thiele and J. W. Vardiman, WHO Classication of Tumours of Haematopoietic and Lymphoid Tissues, 4th edn. World Health Organisation,
2008.
3. J. Fradera, Erythroid karyorrhexis in myelodysplasia: bone marrow aspirate. Blood, 2008, 112 (3), 479.
4. G. J. Mufti, J. M. Bennett, J. Goasguen, B. J. Bain, I. Baumann,
R. Brunning, M. Cazzola, P. Fenaux, U. Germing, E. Hellstrom-Lindberg,
I. Jinnai, A. Manabe, A. Matsuda, C. M. Niemeyer, G. Sanz,

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

5.
6.

7.
8.

9.

10.

11.
12.

13.

14.

15.

16.

197

M. Tomonaga, T. Vallespi and A. Yoshimi, Diagnosis and classication of

myelodysplastic syndrome: International Working Group on Morphology
of myelodysplastic syndrome (IWGM-MDS) consensus proposals for the
denition and enumeration of myeloblasts and ring sideroblasts, Haematologica, 2008, 93(11), 17121717.
B. J. Bain, Dysplastic macropolycytes in myelodysplasia-related acute
myeloid leukemia, Am. J. Hematol., 2011, 86(9), 776.
E. Verburgh, R. Achten, V. J. Louw, C. Brusselmans, M. Delforge,
M. Boogaerts, A. Hagemeijer, P. Vandenberghe and G. Verhoef, A new
disease categorization of low-grade myelodysplastic syndromes based on
the expression of cytopenia and dysplasia in one versus more than one
lineage improves on the WHO classication, Leukemia, 2007, 21(4), 668
677.
B. J. Bain, The bone marrow aspirate of healthy subjects, Br. J. Haematol.,
1996, 94(1), 206209.
A. Bazarbachi, S. Muakkit, M. Ayas, A. Taher, Z. Salem, H. Solh and
J. H. Haidar, Thiamine-responsive myelodysplasia, Br. J. Haematol., 1998,
102(4), 10981100.
J. D. Hu, Y. K. Keung, M. Thakuri, M. W. Beaty, D. D. Hurd, J. Owen
and I. Molnar, Copper deciency causes reversible myelodysplasia, Am. J.
Hematol., 2007, 82(7), 625630.
W. N. Rezuke, C. Anderson, W. T. Pastuszak, S. R. Conway and
S. I. Firshein, Arsenic intoxication presenting as a myelodysplastic syndrome: a case report, Am. J. Hematol., 1991, 36(4), 291293.
D. D. Westho, R. J. Samaha and A. Barnes, Arsenic intoxication as a
cause of megaloblastic anemia, Blood, 1975, 45(2), 241246.
G. A. Kennedy, T. D. Kay, D. W. Johnson, C. M. Hawley, S. B. Campbell,
N. M. Isbel, P. Marlton, R. Cobcroft, D. Gill and G. Cull, Neutrophil
dysplasia characterised by a pseudo-Pelger-Huet anomaly occurring with
the use of mycophenolate mofetil and ganciclovir following renal
transplantation: a report of ve cases, Pathology (Phila.), 2002, 34(3),
263266.
A. B. Taegtmeyer, O. Halil, A. D. Bell, M. Carby, D. Cummins and
N. R. Banner, Neutrophil dysplasia (acquired pseudo-Pelger anomaly)
caused by ganciclovir, Transplantation, 2005, 80(1), 127130.
S. D. Gibbs, D. A. Westerman, C. McCormack, J. F. Seymour and
P. H. Miles, Severe and prolonged myeloid haematopoietic toxicity with
myelodysplastic features following alemtuzumab therapy in patients with
peripheral T-cell lymphoproliferative disorders, Br. J. Haematol., 2005,
130(1), 8791.
D. S. Karcher and A. R. Frost, The bone marrow in human immunodeciency virus (HIV)-related disease. Morphology and clinical correlation, Am. J. Clin. Pathol., 1991, 95(1), 6371.
S. L. Carpenter, S. A. Zimmerman and R. E. Ware, Acute parvovirus B19
infection mimicking congenital dyserythropoietic anemia, J. Pediatr.
Hematol. Oncol., 2004, 26(2), 133135.

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

198

Chapter 7

17. N. Ngo, I. A. Lampert and K. N. Naresh, Bone marrow trephine ndings

in acute myeloid leukaemia with multilineage dysplasia, Br. J. Haematol.,
2008, 140(3), 279286.
18. G. Tricot, C. De Wolf-Peeters, R. Vlietinck and R. L. Verwilghen, Bone
marrow histology in myelodysplastic syndromes. II. Prognostic value of
abnormal localization of immature precursors in MDS, Br. J. Haematol.,
1984, 58(2), 217225.
19. G. Tricot, C. De Wolf-Peeters, B. Hendrickx and R. L. Verwilghen, Bone
marrow histology in myelodysplastic syndromes. I. Histological ndings in
myelodysplastic syndromes and comparison with bone marrow smears, Br.
J. Haematol., 1984, 57(3), 423430.
20. A. J. Stewart and R. D. Jones, Pseudo-Gaucher cells in myelodysplasia,
J. Clin. Pathol., 1999, 52(12), 917918.
21. D. Haase, U. Germing, J. Schanz, M. Pfeilstocker, T. Nosslinger,
B. Hildebrandt, A. Kundgen, M. Lubbert, R. Kunzmann,
A. A. Giagounidis, C. Aul, L. Trumper, O. Krieger, R. Stauder,
T. H. Muller, F. Wimazal, P. Valent, C. Fonatsch and C. Steidl, New insights into the prognostic impact of the karyotype in MDS and correlation
with subtypes: evidence from a core dataset of 2124 patients, Blood, 2007,
110(13), 43854395.
22. O. Pozdnyakova, P. M. Miron, G. Tang, O. Walter, A. Raza, B. Woda and
S. A. Wang, Cytogenetic abnormalities in a series of 1,029 patients with
primary myelodysplastic syndromes: a report from the US with a focus on
some undened single chromosomal abnormalities, Cancer, 2008, 113(12),
33313340.
23. A. M. Cherry, M. L. Slovak, L. J. Campbell, K. Chun, V. Eclache,
D. Haase, C. Haferlach, B. Hildebrandt, A. M. Iqbal, S. C. Jhanwar,
K. Ohyashiki, F. Sole, P. Vandenberghe, D. L. VanDyke, Y. Zhang and
G. W. Dewald, Will a peripheral blood (PB) sample yield the same diagnostic and prognostic cytogenetic data as the concomitant bone marrow
(BM) in myelodysplasia?, Leuk. Res., 2012, 36(7), 832840.
24. J. W. Vardiman, J. Thiele, D. A. Arber, R. D. Brunning, M. J. Borowitz,
A. Porwit, N. L. Harris, M. M. Le Beau, E. Hellstrom-Lindberg, A. Teeri
and C. D. Bloomeld, The 2008 revision of the World Health Organization
(WHO) classication of myeloid neoplasms and acute leukemia: rationale
and important changes. Blood, 2009, 114 (5), 937951.
25. W. G. Finn, A. M. Harrington, K. M. Carter, R. Raich, S. H. Kroft and
A. O. Hero, Immunophenotypic signatures of benign and dysplastic
granulopoiesis by cytomic proling, Cytom. B Clin. Cytom., 2011, 80(5),
282290.
26. F. Steinbach, F. Henke, B. Krause, B. Thiele, G. R. Burmester and
F. Hiepe, Monocytes from systemic lupus erythematous patients are severely altered in phenotype and lineage exibility, Ann. Rheum. Dis., 2000,
59(4), 283288.
27. A. A. van de Loosdrecht, C. Alhan, M. C. Bene, M. G. Della Porta,
A. M. Drager, J. Feuillard, P. Font, U. Germing, D. Haase,

View Online

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

The Diagnosis of Myelodysplastic Syndromes

28.

29.

30.

31.

32.

33.

34.

35.

199

C. H. Homburg, R. Ireland, J. H. Jansen, W. Kern, L. Malcovati, J. G. Te

Marvelde, G. J. Mufti, K. Ogata, A. Orfao, G. J. Ossenkoppele, A. Porwit,
F. W. Preijers, S. J. Richards, G. J. Schuurhuis, D. Subira, P. Valent,
V. H. van der Velden, P. Vyas, A. H. Westra, T. M. de Witte, D. A. Wells,
M. R. Loken and T. M. Westers, Standardization of ow cytometry in
myelodysplastic syndromes: report from the rst European LeukemiaNet
working conference on ow cytometry in myelodysplastic syndromes,
Haematologica, 2009, 94(8), 11241134.
D. Liu, Z. Chen, Y. Xue, D. Lu, Y. Zhou, J. Gong, W. Wu, J. Liang,
Q. Ma, J. Pan, Y. Wu, Y. Wang, J. Zhang and J. Shen, The signicance of
bone marrow cell morphology and its correlation with cytogenetic features
in the diagnosis of MDS-RA patients, Leuk. Res., 2009, 33(8), 10291038.
H. Szpurka, R. Tiu, G. Murugesan, S. Aboudola, E. D. Hsi, K. S. Theil,
M. A. Sekeres and J. P. Maciejewski, Refractory anemia with ringed
sideroblasts associated with marked thrombocytosis (RARS-T), another
myeloproliferative condition characterized by JAK2 V617F mutation,
Blood, 2006, 108(7), 21732181.
L. Malcovati, M. G. Della Porta, D. Pietra, E. Boveri, A. Pellagatti,
A. Galli, E. Travaglino, A. Brisci, E. Rumi, F. Passamonti, R. Invernizzi,
L. Cremonesi, J. Boultwood, J. S. Wainscoat, E. Hellstrom-Lindberg and
M. Cazzola, Molecular and clinical features of refractory anemia with
ringed sideroblasts associated with marked thrombocytosis, Blood, 2009,
114(17), 35383545.
F. R. Appelbaum, J. Barrall, R. Storb, R. Ramberg, K. Doney, G. E. Sale
and E. D. Thomas, Clonal cytogenetic abnormalities in patients with
otherwise typical aplastic anemia, Exp. Hematol., 1987, 15(11), 11341139.
D. P. Steensma, G. W. Dewald, T. L. Lasho, H. L. Powell, R. F. McClure,
R. L. Levine, D. G. Gilliland and A. Teeri, The JAK2 V617F activating
tyrosine kinase mutation is an infrequent event in both atypical myeloproliferative disorders and myelodysplastic syndromes, Blood, 2005,
106(4), 12071209.
J. M. Bennett, D. Catovsky, M. T. Daniel, G. Flandrin, D. A. Galton,
H. R. Gralnick and C. Sultan, Proposals for the classication of the
myelodysplastic syndromes, Br. J. Haematol., 1982, 51(2), 189199.
K. Naqvi, E. Jabbour, C. Bueso-Ramos, S. Pierce, G. Borthakur,
Z. Estrov, F. Ravandi, S. Faderl, H. Kantarjian and G. Garcia-Manero,
Implications of discrepancy in morphologic diagnosis of myelodysplastic
syndrome between referral and tertiary care centers, Blood, 2011, 118(17),
46904693.
P. L. Greenberg, H. Tuechler, J. Schanz, G. Sanz, G. Garcia-Manero,
F. Sole, J. M. Bennett, D. Bowen, P. Fenaux, F. Dreyfus, H. Kantarjian,
A. Kuendgen, A. Levis, L. Malcovati, M. Cazzola, J. Cermak,
C. Fonatsch, M. M. Le Beau, M. L. Slovak, O. Krieger, M. Luebbert,
J. Maciejewski, S. M. Magalhaes, Y. Miyazaki, M. Pfeilstocker,
M. Sekeres, W. R. Sperr, R. Stauder, S. Tauro, P. Valent, T. Vallespi,
A. A. van de Loosdrecht, U. Germing and D. Haase, Revised international

View Online

200

14/10/2013 09:28:24.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00182

36.

37.

38.

39.

40.

41.

Chapter 7

prognostic scoring system for myelodysplastic syndromes, Blood, 2012,

120(12), 24542465.
L. Malcovati, M. G. Porta, C. Pascutto, R. Invernizzi, M. Boni,
E. Travaglino, F. Passamonti, L. Arcaini, M. Maoli, P. Bernasconi,
M. Lazzarino and M. Cazzola, Prognostic factors and life expectancy in
myelodysplastic syndromes classied according to WHO criteria: a basis
for clinical decision making, J. Clin. Oncol., 2005, 23(30), 75947603.
O. Kosmider, V. Gelsi-Boyer, M. Cheok, S. Grabar, V. la-Valle, F. Picard,
F. Viguie, B. Quesnel, O. Beyne-Rauzy, E. Solary, N. Vey, M. HunaultBerger, P. Fenaux, M. Mansat-De, E. V. Delabesse, P. Guardiola,
C. Lacombe, W. Vainchenker, C. Preudhomme, F. Dreyfus,
O. A. Bernard, D. Birnbaum and M. Fontenay, TET2 mutation is an independent favorable prognostic factor in myelodysplastic syndromes
(MDSs), Blood, 2009, 114(15), 32853291.
F. Thol, E. M. Weissinger, J. Krauter, K. Wagner, F. Damm,
M. Wichmann, G. Gohring, C. Schumann, G. Bug, O. Ottmann,
W. K. Hofmann, B. Schlegelberger, A. Ganser and M. Heuser, IDH1
mutations in patients with myelodysplastic syndromes are associated with
an unfavorable prognosis, Haematologica, 2010, 95(10), 16681674.
K. Yoshida, M. Sanada, Y. Shiraishi, D. Nowak, Y. Nagata,
R. Yamamoto, Y. Sato, A. Sato-Otsubo, A. Kon, M. Nagasaki,
G. Chalkidis, Y. Suzuki, M. Shiosaka, R. Kawahata, T. Yamaguchi,
M. Otsu, N. Obara, M. Sakata-Yanagimoto, K. Ishiyama, H. Mori,
F. Nolte, W. K. Hofmann, S. Miyawaki, S. Sugano, C. Haferlach,
H. P. Koeer, L. Y. Shih, T. Haferlach, S. Chiba, H. Nakauchi, S. Miyano
and S. Ogawa, Frequent pathway mutations of splicing machinery in
myelodysplasia, Nature, 2011, 478(7367), 6469.
F. Damm, O. Kosmider, V. Gelsi-Boyer, A. Renneville, N. Carbuccia,
C. Hidalgo-Curtis, V. Della, L. V. Couronne, L. Scourzic, V. Chesnais,
A. Guerci-Bresler, B. Slama, O. Beyne-Rauzy, A. Schmidt-Tanguy,
A. Stamatoullas-Bastard, F. Dreyfus, T. Prebet, B. S. de, N. Vey,
M. A. Morgan, N. C. Cross, C. Preudhomme, D. Birnbaum, O. A. Bernard
and M. Fontenay, Mutations aecting mRNA splicing dene distinct
clinical phenotypes and correlate with patient outcome in myelodysplastic
syndromes, Blood, 2012, 119(14), 32113218.
V. Visconte, H. J. Rogers, J. Singh, J. Barnard, M. Bupathi, F. Traina,
J. McMahon, H. Makishima, H. Szpurka, A. Jankowska, A. Jerez,
M. A. Sekeres, Y. Saunthararajah, A. S. Advani, E. Copelan, H. Koseki,
K. Isono, R. A. Padgett, S. Osman, K. Koide, C. OKeefe,
J. P. Maciejewski and R. V. Tiu, SF3B1 haploinsuciency leads to formation of ring sideroblasts in myelodysplastic syndromes, Blood, 2012,
120(16), 31733186.

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

CHAPTER 8

The Prospects for Real-Time

Raman Spectroscopy for
Oesophageal Neoplasia
MAX ALMOND,a GAVIN RHYS-LLOYD,a
JO HUTCHINGS,a GEETA SHETTY,a NEIL SHEPHERD,a,b,e
CATHERINE KENDALL,a,e NICHOLAS STONEa,c AND
HUGH BARR*a,d,e
a

Biophotonics Research Unit, Leadon House, Gloucestershire Royal

Hospital, Great Western Road, Gloucester GL13NN, United Kingdom;
b
Department of Histopathology, Gloucestershire Hospitals NHS Foundation
Trust, United Kingdom; c Biomedical Imaging and Bio Sensing, School of
Physics, University of Exeter, EX4 4QL, United Kingdom; d Three Counties
Cancer Resection Unit & Upper Gastrointestinal Surgical Unit,
Gloucestershire Hospitals NHS Trust, Gloucester GL13NN, United
Kingdom; e Faculty of Bioscience and Medicine, Craneld Health, Craneld
University, Bedfordshire, United Kingdom
*Email: [email protected]

8.1 The Clinical Problem

Over the past 40 years the incidence of oesophageal adenocarcinoma has increased more rapidly than any other solid tumour in the Western world and
now represents a major public health problem. Recently, the National United
Kingdom Oesophagogastric Cancer Audit reported a 5-year age-adjusted

RSC Detection Science Series No. 2

Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

201

View Online

202

Chapter 8

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

survival rate of 10%. There is a clear imperative to bring forward the diagnosis
of this cancer and to identify and treat precancerous changes.
Barretts oesophagus has been identied as a major risk factor. It is dened
as a change in the distal oesophageal epithelium of any length that can be
recognised as columnar type mucosa at endoscopy and is conrmed by biopsy
of the tubular oesophagus. This can be recognised endoscopically but often
the subtle molecular degeneration to cancer cannot be seen.2
The major healthcare problem is that the population prevalence estimates of
Barretts oesophagus in the UK and US populations is approximately 1 million
and 4 million individuals, respectively. There is also evidence that the incidence
of Barretts oesophagus is increasing by around 2% per year.36
Although Barretts oesophagus is accepted as a signicant risk factor for the
development of adenocarcinoma, the overall risk of progression and the risk of
disease specic mortality are low. Two large population-based cohort studies
reported annual rates of progression to be 0.120.13% per year.7,8
The natural history of preneoplastic/dysplastic Barretts oesophagus has
been dicult to ascertain. Studies have suered from inadequate power, insucient follow-up durations, sampling errors at endoscopy, and considerable
diagnostic variability due to histological subjectivity. Two phenotypes, lowgrade dysplasia (LGD) and high-grade dysplasia (HGD), indicate increased
risk of cancer development. A cohort study of 111 patients with LGD showed a
relative risk of 9.7 (95% condence intervals of 4.421.5) for progression to
HGD or adenocarcinoma.9 If detected at the baseline endoscopy the risk was
4.8 (95% CI 2.6-8.8). 7 LGD is hard to diagnose from biopsy samples and may
be overdiagnosed. Of 1198 Barretts patients undergoing surveillance 12.5%
were diagnosed as LGD. A review of the histology by two external expert
pathologists showed only 0.7% were considered to have true LGD; 42% of
these patients progressed to cancer or HGD, after 51 months.10 Patients with
HGD have a 59% change of invasive cancer after 5 years.11

8.2 Endoscopic Recognition of Dysplasia and Cancer

Barretts oesophagus is usually identied in patients who are undergoing an
endoscopy for investigation of symptoms of acid reux, see Figure 8.1. It can be
easily recognised by due to its salmon pink appearance. The detection of early
dysplasia and early cancer using standard white-light endoscopy is very dicult
and highly subjective, even for experts. Therefore guidelines recommend that
quadrantic biopsies are taken every 2 cm of Barretts oesophagus with additional targeted biopsies from any visible areas of mucosal abnormality. These
are poorly adhered to outside of clinical trials, sample less than 5% of the
mucosa and may miss up to 57% of dysplasia.12,13
Several advanced endoscopic imaging techniques have been developed that
may aid targeted biopsies and identify high-risk areas,1418 The latest endoscopes are high resolution (one million pixels) and do improve detection of
early Barretts neoplasia in expert hands.19 In addition, spraying dyes or acetic
acid onto the surface of the oesophagus can allow more accurate lesion

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

Figure 8.1

203

Endoscopic picture of Barretts oesophagus. The nodular area with white

areas in the corner is clearly abnormal and is an early cancer. The patchy
reddened areas may simply be inammation, Barretts oesophagus or
invisible early cancer or dysplasia. These are only claried by biopsy.
However, the biopsy may miss an area of dysplasia and full mapping is
necessary. Every biopsy will produce bleeding further obscuring the view.

recognition. There are several methods for this chromoendoscopy. Absorptive stains (such as methylene blue) cross epithelial membranes selectively,
whereas contrast stains (such as indigo carmine) permeate into mucosal crevices
highlighting surface topography and mucosal irregularities. Success is varied
and acetic acid is currently favoured.
Specic targeted biomarkers do hold promise. The Fitzgerald group have
demonstrated that alterations in cell-surface glycans occur in Barretts
oesophagus during progression to adenocarcinoma. Selective binding of a
candidate lectin (wheat germ agglutinin) sprayed into an ex vivo oesophagus
enabled visualisation of high-grade dysplastic lesions, which were not detectable by conventional endoscopy.20
To avoid staining the surface narrow-band imaging (NBI) is an alternative
method to provide contrast during endoscopic examination. The oesophageal
surface is illuminated with blue and green light. Longer wavelengths penetrate
deeper into tissue. By using shorter wavelengths in the blue part of the spectrum
supercial capillary networks can be seen, and green light displays the subepithelial vessels. In combination, they produce an extremely high-denition
image of the mucosal surface allowing visualisation of subtle mucosal irregularities and alterations in vascular patterns consistent with dysplasia and intramucosal cancer.21,22 A recent review of NBI with magnication demonstrated
high accuracy for the diagnosis of HGD in Barretts oesophagus based on recognition of irregular mucosal pit patterns and/or irregular microvascularisation.23
Overall, data on the accuracy of NBI in Barretts oesophagus are inconclusive
and results of multicentre randomised controlled trials are awaited.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

204

Chapter 8

As well as morphological changes there are uorophore concentration differences between normal, metaplastic and neoplastic tissue. Autouoresecence
endoscopy (AFI) endoscopy has been used for the detection of high grade
dysplasia in Barretts oesophagus. Curvers et al. demonstrated an increased
detection rate of HGD/IMC using AFI compared to WLE alone (53% vs.
90%) but this came at the expense of a high false-positive rate of 81%.24
Optical coherence tomography OCT produces cross-sectional images of the
oesophageal surface. The probe has to be passed through the instrument
channel of an endoscope. In 55 patients with Barretts oesophagus, OCT was
able to dierentiate between HGD and adenocarcinoma with a sensitivity of
83% and a specicity of 75%.25 The acquisition of images and interpretation
can take some time and further developments are awaited.
Endoscopic confocal microscopy magnies the mucosa 1000-fold enabling
real-time visualisation of cellular structures. It has shown considerable potential for use in patients with Barretts oesophagus with reported accuracy of up
to 97.4% for the detection of dysplasia.26 There is considerable interuser
variation, it is time consuming, requires administration of an exogenous agent;
and is expensive.27
Elastic scattering spectroscopy (ESS) was one of the rst methods to be
examined for the detection of dysplasia in Barretts using a probe passed down
the instrumentation channel of the endoscope. Lovat et al. measured spectra
from 181 tissue sites from 81 patients that were correlated with consensus
histopathology. ESS identied HGD with a sensitivity of 92% and a specicity of
60% and was able to dierentiate these sites from inammation with a sensitivity
and specicity of 79%.28 This current level of accuracy is probably not sucient
to support clinical uptake of ESS, unless it can be improved through renement
of the technology or the use of a concomitant imaging modality.

8.3 Raman Spectroscopy

The attraction of Raman spectroscopy is near-instantaneous acquisition of
molecular information. Raman scattering was rst described in 1928 by an
Indian physicist of the same name who was very shortly afterwards awarded the
Nobel Prize for Physics.29 The great attraction of the technique is the highly
specic biochemical/molecular information obtained. In the past decade, the
potential of the technique for dierentiating between normal, premalignant,
and cancerous tissue has been extensively explored. Technological developments have occurred, which have allowed the possibility of an endoscopic
Raman probe to provide accurate, noninvasive optical diagnosis in real-time,
without the need for biopsy.
Two types of Raman scattering have been described, dependent on the state
of the molecule when the interaction with light occurs. Stokes scattering results
when light interacts with a molecule in its ground (unexcited) state. Nuclear
motion is initiated and the scattered light has a lower energy than the incident
light. However, where the molecule exists in an already vibrationally excited
state, as is seen at high temperatures, the scattered light is of a higher energy

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

205

than the incident light. This is known as anti-Stokes scattering. At room

temperature, most molecules are at the ground vibrational state and therefore
Stokes Raman scattering is much more prevalent than anti-Stokes scattering. It
is possible to calculate the proportion of molecules in a substance that are in the
ground and excited states using the ratio of intensities of Stokes and anti-Stokes
scattering.
As a molecule vibrates the surrounding electron cloud changes shape. This can
induce a change of the dipole moment or a change in the polarisation of the
molecule. For the vibrational change in a molecule to be Raman active there
must be a change in the polarisation of the molecule. That is to say that in shifting
from one vibrational state to another, the electron density relative to the position
of the nucleus must change. However, not all changes in vibrational states can be
detected by Raman spectroscopy. In some cases the change in vibrational mode
represents a change in the dipole moment of the molecule and not its polarisation.
This change is not detected by Raman spectroscopy but can be detected by infrared (IR) spectroscopy. An example of the dierences in vibrational modes can
be appreciated when the properties of a phenyl ring are compared to that of OH
bond. The phenyl ring has a large electron cloud and is highly polarisable, giving
a strong Raman signal, whereas an OH bond is less polarisable (so gives a weak
Raman signal), but a dipole change is readily induced, causing a strong IR signal.
These properties of the OH bond have important consequences for the application of spectroscopy to living-tissue diagnostics. As a result of the strong IR
signal, all water must be excluded from tissue samples before using IR spectroscopy to prevent profound distortion of the spectra from the marked dipole
eect of the OH bonds. This eect is not seen with Raman spectroscopy and so
allows this technique to be used to analyse fresh tissue. Any molecule of sucient
complexity will usually exhibit a usable Raman signal.
A Raman spectrum is a plot of Raman shift measured in wave numbers
(on the x-axis) against intensity (on the y-axis) (see Figure 8.2). The zero point
on the x-axis corresponds to the energy of the incident light. Data to the right
of this point therefore represents Stokes scattering, whereas data to the left
represents anti-Stokes scattering. A Raman spectrum is always relative to the
wavelength of the incident light. As a result, a specic wavelength of incident
light can be selected in order to determine the wavelength region of the
measured Raman spectra. The x-axis of a Raman spectrum is therefore rarely
measured in wavelength as this would vary with the incident light chosen for
excitation. Instead, wave number is used to represent the Raman shift and
facilitate comparison between spectra. A larger Raman shift indicates that a
greater energy is required to create vibrational motion in a given molecule and
crucially, the value of the Raman shift for a given compound is independent
of the frequency of the excitational light. The excitational light frequency does
directly aect the intensity of the Raman signal.
Raman peaks are spectrally narrow and correspond precisely to the vibration
of specic bonds within molecules. In addition, the intensity of a peak is proportional to the concentration of the specic molecular constituent that gives
rise to it. This allows specic vibrations to be assigned to certain spectral peaks,

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

206

Figure 8.2

Chapter 8

Raman spectrum acquired from normal squamous oesophageal mucosa.

enabling a precise qualitative and quantitative biochemical ngerprint to be

obtained.30 However, in complex molecules peak assignment can be dicult
and unreliable as the various peaks merge to form bands. It is then the shape
and distribution of these bands that can then be used as a chemically sensitive
signature of the specimen.
Analysis of pure natural compounds such as glycogen and collagen has
enabled scientists to assign peaks that are characteristic of these molecules. Glycogen molecules have characteristic peaks at 490, 853, 937, 1083 and
1123 cm1. Although potentially obscured by signals from other molecular
constituents, modern multivariate analysis techniques can be used to extract
these subtle spectral features and enable conclusions to be drawn regarding the
nature of a tissue specimen. This highlights the potential of Raman spectroscopy to measure objective, quantitative information that can enable precise
tissue analysis and medical diagnosis.

8.4 Instrumentation
Raman spectrometers must be very sensitive in order to detect Raman signals
of extremely low intensity. Other signals, in particular elastically scattered light,
have to be excluded. Otherwise, the weak Raman signal would be completely
obscured. This is achieved using a series of lters within the instrument. As the
positions of Raman peaks are relative to the wavelength of incident light, a
monochromatic (single-wavelength) incident laser is used for excitation. When
assessing biological tissue, incident laser light in the near-infrared (NIR) region
is generally the preferred source for excitation. At this wavelength there is
minimal tissue absorption and the risks of thermal or mutagenic eects are

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

207

therefore avoided. In addition, the eect of uorescence seen with visible and
UV light is largely overcome.
A series of mirrors typically guide the laser onto the sample. Scattered light
passes back into the spectrometer and is ltered by a notch or edge lter to
remove elastically scattered light. A lens focuses the light onto a slit, which
removes any stray light. The light is then split and dispersed by the diraction
grating and then reected onto the charge-coupled device (CCD).
The CCD typically consists of a sectored length of silicon. This enables
dierent frequencies of scattered light to be collected separately, and this information can then be relayed to a computer in order to construct a spectrum.
The shape of the spectrum will depend on the number of photons striking each
sector of the silicon. The two main types of CCD used for Raman spectroscopy
are backthinned and deep-depletion devices. Backthinned CCDs suer from
etaloning and for this reason deep-depletion CCDs are commonly preferred for
NIR applications.

8.5 Raman Spectroscopy for Medical Diagnosis

The ability of Raman spectroscopy to accurately identify pathological changes
in tissues has been proven using ex vivo tissue specimens from breast, cervix,
larynx, bladder, prostate, lymph nodes and oesophagus.3148 These studies have
encouraged clinicians about the potential of Raman spectroscopy to bring
forward cancer and precancer diagnosis. In particular, to understand the subtle
molecular degeneration that occurs before an invasive cancer becomes morphologically visible. The aim is to provide real-time rapid diagnosis without the
need for tissue removal and biopsy and to detect the earliest molecular changes
that indicate a segment of Barretts oesophagus has the potential to become
bad. A Raman probe could identify early neoplastic changes within the
oesophagus, which are often invisible using standard endoscopy. In addition
Raman spectroscopy could remove the subjectivity from histological assessment
and even eliminate the need for a biopsy altogether (so called optical biopsy).
Objective endoscopic diagnosis of neoplastic change in the oesophagus would
enable immediate intervention to take place. HGD in Barretts oesophagus
could be diagnosed and then treated by immediate endoscopic resection during a
single endoscopy session. Similarly, tumour margins could be assessed intraoperatively to ensure an adequate surgical resection. In addition, when rescoping
a patient with biopsy-proven focal HGD or IMC, Raman spectroscopy could
accurately identify the lesion so that endoscopic resection (of the correct area)
could be performed. The challenge of developing in vivo Raman systems is
considerable as several important factors must be addressed and overcome:
Weak signal strength:
6
8
J Raman signals are inherently weak (only 1 in 10 10
photons are
inelastically scattered).
J Biochemical changes in cells that have become dysplastic are very difcult to detect as they may be present in minute concentrations.

View Online

208

Chapter 8

When used on living tissue, the intensity of the excitatory laser must be
limited to prevent tissue damage. This limits the ability to generate
strong scattering signals from the tissue.
Signal interference:
J Biological tissues are inhomogeneous in composition and thus highly
scattering.
J Tissue uorescence.
J Spectral contamination caused by the background Raman and uorescence signals generated in the bre-optic materials.

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

Previous studies have recognised that simple peak analysis comparing

the intensities and wave numbers of individual peaks, fails to provide discrimination between pathology. Instead, one must consider diagnostic information to be spread throughout the spectrum and complex chemometric
approaches such as principal component analysis (PCA) and linear discriminant analysis (LDA) have been utilised to tease out biochemical dierences between tissue types.
When aiming to optimise signal strength we must remember that the intensity of a Raman peak is dependent on the polarisability of the molecule being
excited, the concentration of that molecule in the tissue, and the intensity of the
incident laser. The power of Raman scattering increases with the fourth power
of the frequency of the excitation source. In addition, the signal-to-noise ratio
(SNR) is proportional to the square of the intensity of the incident light
multiplied by the acquisition time.
Therefore, increasing both the incident laser power and the acquisition time
would be expected to generate a stronger Raman signal. However, in practice
this could cause injury to human tissue and therefore does not oer a realistic
solution. Moreover, prolonged acquisition times of more than a few seconds
would be impractical for in vivo use.
Another problem when using Raman spectroscopy to examine biological
tissue is that of tissue uorescence that often obscures the Raman signal.

8.6 Raman Spectroscopy Assessment of Oesophageal

Pathology
Several groups have explored the potential of Raman spectroscopy for diagnosis of oesophageal pathology in vitro.
In 2003 Kendall et al. analysed snap-frozen oesophageal biopsy specimens
using Raman microspectroscopy.49 Their subset analysis included normal
squamous mucosa, Barretts mucosa and adenocarcinoma, and showed a high
level of agreement (kappa statistic, k 0.89) between consensus pathological
opinion (the gold standard) and Raman measurements. They also showed that
the level of agreement between Raman spectroscopy and consensus histological
diagnosis compared favourably to that between an independent pathologist
and the consensus pathology opinion for the same biopsy specimens (k 0.76).

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

209

This illustrated the importance of gaining a consensus pathology opinion in

order to reduce interobserver variability, a nding supported by others.11,49
Shetty et al. used Raman spectroscopy to characterise some of the early
biochemical changes that occur in oesophageal mucosal cells as they progress
to Barretts metaplasia and then to dysplasia and cancer.43 They found a
marked variation in the distribution of various biochemical constituents in
normal, dysplastic and malignant tissue. The changes occurred in sugars,
lipids, and amino acids. Raman spectroscopy identied increased levels of
DNA, actin and oleic acid in dysplastic glandular tissue compared to normal
Barretts or squamous mucosa. Conversely, glycogen was particularly prominent in normal squamous tissue and reduced in neoplastic tissue. This study
further supported the potential of Raman spectroscopy for early diagnosis of
malignancy, and raised the possibility that Raman spectroscopy could perhaps
enable detection of biochemical changes in tissue before morphological
changes manifest. The proven ability of Raman spectroscopy to analyse and
classify biopsied oesophageal tissue has demonstrated potential for the technique to be used to aid histological assessment of tissue samples. The ability
to objectively characterise tissue based on its biochemical prole could help
to overcome the inherent subjectivity associated with histopathological
classication.
Recently, Hutchings et al. demonstrated that rapid Raman mapping of
oesophageal tissue can identify histologically relevant biochemistry in order to
distinguish between pathological subgroups in a clinically practicable timescale.38 Hutchings et al. used a Renishaw StreamLine system coupled to a
microscope tted with a Leica 50 long working distance objective (NA 0.5) to
map 12 oesophageal biopsy samples that had been snap frozen and sectioned.
A 830-nm diode laser was used for excitation which had a power of 70 mW at
the sample. A variety of mapping parameters were evaluated (see Figure 8.3).
The continuous readout of the CCD minimised time between sequential data
measurements resulting in the acquisition of thousands of spectra in relatively
short timescales. Mapping of a 2-mm oesophageal biopsy in 3090 min (0.5 s
acquisition time per 25.3 mm Raman pixels) in conjunction with principal
component analysis was shown to be sucient to generate spectral maps with
clearly identiable pathological regions.34
Rapid Raman mapping has the potential to provide prompt objective assessment of biopsied tissue samples without the need for tissue xation or
staining. This low signal-to-noise mapping lends itself to automation. Furthermore, objective diagnosis can be obtained by projecting measured spectral
data onto pre-established multivariate algorithms in order to obtain a diagnosis
based on the biochemical prole of the tissue, thus removing the subjectivity of
histological reporting. Work is currently underway to further evaluate the
diagnostic potential of rapid Raman mapping as an aid to oesophageal histopathology evaluation, particularly with regard to the grading of Barretts-associated dysplasia. This ability to map out pathological regions within
oesophageal biopsy specimens, which can be correlated with histology appearance, provides further evidence of the ability of Raman spectroscopy to

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

210

Figure 8.3

Chapter 8

A high-quality Raman map is able to identify 4 distinct areas that have

been discriminated through dierences in their biochemical composition.
Each has a characteristic spectral ngerprint which can be distinguished
using PC-fed LDA.

objectively assess tissue and therefore gives additional support to the proposal
of developing diagnostic Raman probes for in vivo use in the oesophagus.

8.7 Fibre-Optic Raman Spectroscopy

In order to develop a clinically applicable in vivo Raman probe for upper
gastrointestinal endoscopy, there are several specic scientic and practical
considerations that must be addressed and overcome.
The probe must be narrow, long enough to t down an endoscope
channel modern instruments have a 2.8-mm diameter port and are
90 cm in length. The design of the probe tip must not interfere with
manipulation of the articulated distal end of the endoscope.
Signal-to-noise ratio must be maximised, although acquisition times must
remain clinically applicable (up to 2 s). The probe and spectrometer must
be coupled to a data-processing unit so that short acquisition times are
translated to meaningful clinical results in near real time.
Data measurement should also be close to real time (short acquisition
times) to enable rapid assessment of large areas of oesophagus and to
minimise inaccuracy caused by scope movement or oesophageal peristalsis.
The probe must be safe for use in vivo (laser wavelength and power must
not cause tissue destruction or long-term mutagenic eects).
The probe must function eectively in the presence of endoscopic light
laboratory Raman spectroscopy requires reduced ambient light to
minimise unwanted spectral noise.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

211

The sampling depth and volume must be optimised. Metaplasia and

dysplasia develop in the oesophageal epithelium the optimum depth
from which to attain Raman signal is therefore 10200 mm.43 Sampling
from large surface areas would provide more rapid assessment of clinical
pathology but may decrease sensitivity and specicity for focal
abnormalities.
The probe must work in direct contact with the mucosal surface in order
to facilitate endoscopic use.
The probe must work safely and accurately in the presence of blood, saliva
and gastric juice.
The probe must be packaged so that it can be reused for multiple
endoscopies and must be robust enough to withstand sterilisation procedures or must be simple and cheap enough to manufacture to enable
single use.
Measurements must be reproducible on dierent days, by dierent endoscopists and using dierent probes and spectrometers;
High sensitivity and specicity must be achievable in the clinical setting
(and this must be tested and reproduced in vivo).
As Raman scattering is weak and the power of the incident laser must be low
(to minimise thermal injury), the design of a bre-optic Raman probe must
centre on maximising collection eciency, whilst minimising intrinsic scattering
and uorescence produced by the probe.
One of the most potentially dicult problems to overcome is the propensity
of probe bres to generate their own Raman and uorescence signals.50,51 In
some cases, the Raman signals may be greater than those arising from the tissue
being studied. Fibre signal is proportional to bre length and to the intensity of
excitatory light. It is generated in the delivery bre due to excitation by the laser
light and in the collection bres by elastically scattered light transmitting back
along the collection bre. Holographic notch lters or metal-oxide edge lters
can be used to remove elastically scattered light from the collection bre, and
narrow-bandpass lters are commonly used to remove bre-optic signal that
has been generated in the delivery bre.
In 1998, Mahadevan-Jansen et al. published in vivo results using a breoptic Raman probe to analyse dysplastic change in cervical tissue.52 Silica
silica bres were shown to cause less signal disruption than sapphire and
liquid guide bres despite their tendency to generate signicant background
signal at low wave numbers (below 900 cm1). The probe consisted of a single
excitation bre and 50 collection bres and had a total diameter of 12 mm.
The excitation bre was positioned at the edge of the probe, and the laser
output (at 789 nm) was reected by a gold mirror at an angle to produce a
900-mm diameter excitation spot on the tissue surface. A bandpass lter on the
distal tip of the excitation bre blocked uorescence and Raman scattering
generated by the bre, but transmitted the excitation light. Although
demonstrating that Raman spectra could feasibly be measured in vivo using a
bre optic, the dimensions of this probe and the time taken for spectral

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

212

Chapter 8

acquisition (90 s) meant this design was impractical for use in the
gastrointestinal tract.
The next important challenge was to develop an eective miniature breoptic probe that employed state-of-the-art probe technology but that was
compatible for use with a medical endoscope. An endoscope is 90 cm in length
and has a 2.8-mm diameter instrument channel through which a Raman probe
could be delivered. This presents a considerable technical challenge; to package
a probe with all its constituent parts (lters, collimating lenses, illumination and
collection channels, and external casing) into a durable device with a diameter
smaller than 2.8 mm.
In 1999 Shim et al. reported on utilising an endoscope compatible, bre-optic
Raman probe consisting of one central delivery bre (400 mm diameter) surrounded by seven collection bres each with a 300-mm diameter (Visionex, Inc,
Sunnyvale, CA, USA, probe).53 Filters were placed over the bres approximately 2.5 cm from the probe tip and the collection bres were part bevelled so
that they predominantly collected signal from a zone in front of the delivery
bre and at a specied collection depth, determined by the bevelled angle. This
design provided a means of controlling the sampling volume whilst optimising
collection eciency.
In 2000, Shim et al. reported the rst in vivo Raman measurement of gastrointestinal tissue using the Visionex probe described above.54 In this feasibility study the probe was passed through an endoscope into the oesophagus,
stomach and colon. Twenty patients successfully underwent the endoscopic
procedure following routine sedation and no complications were reported. The
probe was positioned to touch an area of mucosa for 5 s during which time
readings were taken. The probed site was then biopsied to allow comparison of
Raman spectra with histopathology at that site. Subtle dierences were reported in spectra from normal and diseased areas but the data acquired was not
sucient to discriminate between pathological groups. Whilst demonstrating
feasibility, Shim et al. concluded that more work was needed to enhance the
signal-to-noise ratio (SNR), to minimise acquisition times, and to improve data
interpretation before reliable results could be generated. This study illustrated
that endoscopic Raman spectroscopy could be performed safely in human
patients, but was not large enough to enable comparison with results generated
using laboratory-based Raman spectrometers.
Motz et al. used a novel probe to study dysplasia in rat palate tissue using
100 mW of laser light (at 830 nm) for excitation.54,55 Multivariate data analysis
of Raman spectra demonstrated dierences in low- and high-grade dysplasia.
They concluded that further modications to the probe were necessary to optimise collection depth as a signicant spectral contribution was collected from
tissue that was deeper than 500 mm. More recently, a greater emphasis has been
placed on targeted illumination of the epithelial lining of the gastrointestinal
tract. Precancerous dysplasia develops in epithelial cells and, by denition, has
not yet invaded deeper structures. For optimum diagnostic accuracy, probes
must collect signal from the epithelial layer and minimise spectral contributions
from scattering occurring in deeper tissues.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

213

In 2005, Wilson and coworkers published results of an in vivo trial in the

upper gastrointestinal tract of humans. They used a probe consisting of a
central delivery bre surrounded by six 300-mm diameter bevelled collection
bres that achieved an estimated sampling depth of 500 mm.56 In conjunction
with a diode laser (emitting at 785 nm) and a modern CCD detector they reported an accuracy of 87% for delineating between normal oesophageal tissue
and dysplasia.
Huang and coworkers recently described the use of a ball lens at the distal tip
of a novel Raman probe to focus excitation light onto the epithelium and help
collimate scattered light into the collection bres.57 In 2010, following a
number of preliminary in vitro and in vivo studies in rats, Huang et al. published
a series of results from in vivo trials of endoscopic Raman spectroscopy for the
detection of gastric neoplasia in humans. They used a 1.8-mm diameter Raman
probe consisting of 32 collection bres surrounding a single central delivery
bre. The probe had two stages of optical lters, at its proximal and distal ends,
and used 785 nm excitation. Each spectrum was acquired in 0.5 s with a laser
light irradiance of 1.5 W cm2. In December 2010 they reported a diagnostic
sensitivity of 82.1% and a specicity of 95.3% for delineating malignant gastric
ulcers from benign tissue.58 In 2011, the same Raman probe system was used in
the oesophagus demonstrating a sensitivity of 97.0% and a specicity of 95.2%
for detecting oesophageal cancer.59

8.8 Novel Raman Probe

We have developed a novel probe for in vivo analysis of oesophageal tissue
using a novel, custom-built, confocal Raman probe that has been designed
specically for compatibility with a standard medical endoscope. The Raman
probe has been designed and built at the Interface Analysis Centre, at the
University of Bristol.
A full description of the probe has been published by Day et al.37 The laser is
transmitted through the excitation bre, collimated by a GRIN lens, and ltered to remove Raman and elastic scattering signals generated within the bre.
A monolith reects the light onto a dichroic lter that reects only light at the
specied excitation wavelength onto a focusing GRIN lens. After interaction
with the sample, scattered light is collimated by the GRIN lens and transmitted
back to the dichroic lter that reects the elastically scattered light but allows
inelastically scattered wavelengths to pass through to a second dichroic lter.
The signal is then collimated by a GRIN lens and transmitted back to the
spectrometer through the collection bre. The probe is confocal, that is to say,
the exit aperture of the excitation bre is equal to the entrance aperture of the
collection bre. It has been designed to focus to 150 mm in order to sample from
the oesophageal epithelium and minimise signal interference from deeper
tissues. In addition, unlike previous commercial bre-optic Raman probes it
has a short tip so that endoscopic manipulation will be possible when taking
probe measurements in vivo. It is also designed for direct tissue contact in
order to facilitate in vivo endoscopic measurements. Initial analysis was on

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

214

Figure 8.4

Chapter 8

Photograph taken prior to tissue measurements after placement of the

perspex mapping grid. The distal oesophagus and proximal stomach have
been opened and the grid has been pinned in place so that it overlies the
distal oesophagus and gastro-oesophageal junction.

oesophagectomy specimens using a xed grid as shown (see Figure 8.4). Following spectral measurement, a biopsy was taken using standard endoscopy
forceps at each of the specied grid positions and immediately xed in formalin.
All samples were paran embedded, microtomed and stained with Haematoxylin and Eosin prior to expert histological reporting. In addition, endoscopic resection (ER) and endoscopic biopsy samples were used to augment the
spectral dataset. These specimens were placed on acetate and snap frozen in
liquid nitrogen in the endoscopy suite. They were subsequently allowed to
defrost at room temperature for 2 min (point biopsies) or 5 min (ERs) and
mounted on calcium uoride substrate with their mucosal surface facing upwards before measurement. All tissue samples were independently reported by
two expert Consultant Pathologists based at dierent hospitals in order to gain
consensus expert diagnosis.
Figure 8.5 shows the results for classication of neoplasia in Barretts
oesophagus. The results presented have demonstrated that a custom-built
Raman probe designed for endoscopic use can accurately discriminate between
a variety of oesophageal tissue types. Despite unavoidable background signal,
an inherent problem when aiming to measure subtle spectral features via a breoptic, principal-component-fed linear discriminant analysis (PC-fed LDA)
demonstrated sucient ability to detect the specic biochemical information to
enable tissue classication. Background subtraction/correction algorithms did
not yield improvements in classication performance.
Results have been validated using independent test datasets and crossvalidation techniques (e.g. leave-one-out). The independent test datasets, which
were acquired over a dierent timescale to the training dataset, demonstrated

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

Figure 8.5

215

Linear discriminant analysis scatter plot of the three group classication

model. All spectra were acquired in 1 s. NSq-normal squamous (normal
oesophagus) HGD/cancer is high-grade dysplasia and cancer, LGD/ND8
is low-grade dysplasia and nondysplastic Barretts Oesophagus. PC stands
for principle component and ld (LD) is Linear Discriminant.

accurate classication performance. We have also demonstrated that reliable

spectra can be acquired using dierent identically designed probes used by
dierent operators and that these results can be accurately classied using an
independent training algorithm.

8.9 Discussion
The data demonstrates that there is the possibility of real-time bre-optic probe
analysis of the oesophageal mucosa in clinically applicable timescales. In
addition, spectra can feasibly be measured under direct vision using WLE and /
or NBI. Shim et al. in 2000, demonstrated that in vivo Raman measurements
were feasible and safe in the oesophagus, although diagnostically useful information was not obtained.54 In vivo work by several groups has supported the
conclusions that Raman spectroscopy is safe and feasible and has demonstrated
more encouraging diagnostic data, although detection of early lesions using
other probe systems has proved dicult.55,56
The results presented and the experience from previous pilot clinical trials
support the need for an in vivo trial using this custom-built Raman probe
system. There are several practical considerations that need to be addressed:
 Probe durability Probes must be suciently durable to tolerate multiple
passes down an endoscope without it aecting spectral recordings. Early
in this project a probe was damaged during passage through a scope
prompting a change to the design of the external rigid coating of the distal
tip of the probe.
 Single or multiuse Our data has shown that a multiuse probe could
feasibly collect reliable data over at least a 2-year period.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

216

Chapter 8

 Manufacturing There needs to be an established mechanism for reproducing identically built probes to high standards.
 CE marking this is the mandatory legal conformity mark necessary for
products placed on the European market. It conrms that the manufacturer has ensured that the product is not only safe but also functions in
a way so as to achieve its intended purpose.
 Commercial support There needs to be a mechanism for funding a future
clinical trial. This will most likely be through a commercial collaboration.
After addressing these factors a carefully planned clinical trial will need to be
undertaken. The aims of the trial would need to be clearly dened. Prior to a
large, randomised trial a signicant in vivo spectral dataset would be required to
generate a robust classication model. The technique for ensuring that probe
spectra can be correlated with the current gold standard (consensus expert
histology) must be established. Huang et al. have described the use of an ER
cap to ensure that probe and biopsy sites match. However, this requires that all
patients are scoped using an ER cap, despite not necessarily needing therapeutic intervention. Alternatively, a double-lumen scope could be used so that
the probe and the biopsy forceps are passed simultaneously through the scope.
This would enable a biopsy to immediately be followed by a Raman measurement. However, slight adjustment by the endoscopists to allow for the different positions of the instrument channels will be required.
A Raman probe system could potentially have a role as a surgical adjunct
during oesophagectomy. It could enable surgeons to assess resection margins
intraoperatively to ensure they are clear of microscopic tumour deposits in
order to minimise the risk of locally recurrent disease. Similarly, the probe
could even act as an adjunct to oesohagogastric cancer staging, through
laparoscopic assessment of suspicious peritoneal or lymphatic deposits, or
potentially endoscopic assessment of the depth of tumour invasion through the
oesophagus. Previously, the ability of RS to stage transitional cell carcinoma of
the bladder based on urothelial measurements has been demonstrated.44 There
is potential for tumour depth T-staging in the oesophagus, however, modication to the confocal design of the probe described would be necessary if this
were intended.
There are several methodological limitations that must be highlighted. These
relate to the technique of spectral acquisition, the nature and design of the
probe, the data preprocessing and analysis, and to assumptions made between
the ex vivo environment and the real clinical environment at endoscopy.
 Point measurements This is a recognised limitation of the technique. It is
hoped that further developments may enable real-time imaging that could
facilitate assessment of whole Barretts segments in clinically applicable
timescales. Thus, Raman point measurements may need to be combined
with a wide eld imaging system to identify areas of maximal concern.
 Fluorescence The endogenous uorophores in blood are well known to
induce uorescence that may mask weak Raman signals. Although water

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

217

lavage could be performed during endoscopy to remove excess blood, the

uorescence eects may interfere with a clinicians ability to take multiple
targeted biopsies. This will need further evaluation in vivo. As uorescence
was very rare ex vivo (following mucosal lavage) and these spectra were
easily identiable as visual outliers prior to data analysis, the problems
caused by uorescence were minimal.
Probe background The custom-built Raman probe has signicant
background signal due to Raman, uorescence, and elastic scattering
signals being generated within the bre-optic cables. However, the probe
has been shown to be suciently sensitive to detect the important spectral
signals that enable tissue classication. Although not all visible above the
probe background, PC-fed LDA was able to extract this data for tissue
classication. Background subtraction/correction algorithms were not
shown to improve classication performance.
Exclusion of ambient light All measurements presented and discussed
were acquired following exclusion of ambient light as visible wavelengths can cause spectral disruption. However, both endoscopic white
light and NBI have been shown to be compatible with Raman
measurements.
Pressure eects Inadequate contact of the probe with the tissue may have
aected the spectral signal. However, as explained previously, this was a
translational project using a clinically applicable tool that was pressed
against the tissue by hand, as would be the case in vivo. No attempt was
taken to regulate probe pressure as this would be impractical in vivo.
Inability to subclassify the grade of dysplasia Despite measuring in excess of a thousand ex vivo tissue spectra, the numbers of homogeneous
LGD and HGD samples that were conrmed by consensus histological
diagnosis were too low to enable distinction between grades of dysplasia.
Prior to widespread clinical use, an in vivo trial of sucient power will
need to generate a large spectral database in order to produce a robust
classication model for in vivo spectra.
Neoadjuvant chemotherapy The majority of patients were treated with
chemotherapy prior to surgical resection. This often led to shrinkage and
brosis of the tumour, which may have aected its spectral signal.
Previous ablation therapy Some point biopsies/endoscopic resections
were taken from patients who had undergone previous ablation therapy. It
is unclear what eect, if any, this may have had on the Raman signal
obtained from this tissue. However, in a clinical setting it will be essential
that the probe can detect recurrent Barretts/dysplasia in patients who
have undergone previous endoscopic treatment. The ability to detect
buried metaplasia has not been evaluated.
Ischaemia eects The eects of tissue ischemia have not been fully established. This may mean that the results cannot be generalised to in vivo
results with complete accuracy. A large spectral dataset will need to be
established following an in vivo trial and at that point comparisons with
ex vivo data can be made.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

218

Chapter 8

 No assessment has been made as to whether the probe can identify

macroscopically invisible areas of dysplasia/cancer. In order for RS to be a
clinically applicable tool, it will need to be shown to improve recognition
of early lesions and therefore demonstrate the ability to change patient
care. This is impractical to do at this early stage but should be considered
when designing future in vivo clinical trials in order to demonstrate
diagnostic benets.
 All spectra were acquired from a biased population as all patients had a
history of Barretts oesophagus and/or adenocarcinoma. However, these
patients represent the target population for this clinical tool.
 It is possible systematic errors may have aected the dataset, for example
systematic drift in the intensity or wave number axes over time. However,
careful calibration prior to all sets of measurements was used to obviate
these eects. In addition, the data presented showed no evidence of systematic drift and instrument response (green-glass) correction failed to
improve the classication performance suggesting that major systemic
drift was not a problem.

References
1. National Oesophago-Gastric Cancer Audit. Third Annual Report, (2010).
2. A. Watson, R. C. Heading and N. A. Shepherd, Report of the Working
Party of the British Society of Gastroenterology, Principal Authors:
Guidelines for the Diagnosis and Management of Barretts Columnar-lined
Oesophagus. 2005.
3. N. Vakil, S. V. van Zanten, P. Kahrilas, J. Dent and R. Jones, Global
Consensus Group, Am. J. Gastro., 2006, 101, 1900.
4. J. Ronkainen, P. Aro, T. Storskrubb, S. E. Johansson, T. Lind, E. BollingSternevald, M. Vieth, M. Stolte, N. J. Talley and L. Agreus, Gastroenterology, 2005, 1291825.
5. J. Jankowski, H. Barr, K. Wang and B. Delaney, Brit. Med. J., 2010,
341, 4551.
6. L. B. Gerson and S. Banerjee, Gastro. Endosc., 2009, 70, 867.
7. S. Bhat, H. G. Coleman, F. Yousef, B. T. Johnston, D. T. McManus,
A. T. Gavin and L. J. Murray, JNCI, 2011, 103, 1049.
8. F. Hvid-Jensen, L. Pedersen, A. M. Drewes, H. T. Sorensen and P. FunchJensen, NEJM, 2011, 365, 1375.
9. M. Sikkema, C. W. Looman, E. W. Steyerberg, M. Kerkhof, F. Kastelein,
H. van Dekken, A. J. van Vuuren, W. A. Bode, H. van der Valk,
R. J. Ouwendijk, R. Giard, W. Lesterhuis, R. Heinhuis, E. C. Klinkenberg,
G. A. Meijer, F. ter Borg, J. W. Arends, J. J. Kolkman, J. van Baarlen,
R. A. de Vries, A. H. Mulder, A. J. van Tilburg, G. J. Oerhaus, F. J. ten
Kate, J. G. Kusters, E. J. Kuipers and P. D. Siersema, Am. J. Gastro.,
2011, 106, 1231.
10. W. L. Curvers, F. J. ten Kate, K. K. Krishnadath, M. Visser, B. Elzer,
L. C. Baak, C. Bohmer, R. C. Mallant-Hent, A. van Oijen, A. H. Naber,

View Online

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

11.
12.
13.
14.
15.

16.
17.

18.
19.
20.

21.
22.
23.
24.

25.

26.
27.
28.

29.
30.
31.
32.

219

P. Scholten, O. R. Busch, H. G. Blaauwgeers, G. A. Meijer and

J. J. Bergman, Am. J. Gastro., 2010, 105, 15231530.
B. J. Reid, P. L. Blount, Z. Feng and D. S. Levine, Am. J. Gastro., 2000,
95, 3089.
C. Bennett, S. Green, H. Barr, P. Bhandari, J. DeCaestecker, K. Ragunath,
R. Singh, A. Tawil and J. Jankowski, Cochrane System. Rev., 2010, 5.
M. Vieth, C. Ell, L. Gossner, A. May and M. Stolte, Endosc., 2004,
36, 776.
K. B. Dunbar and M. I. Canto, Gastro. Endosc., 2010, 72, 668.
L. Qiu, D. K. Pleskow, R. Chuttani, E. Vitkin, J. Leyden, N. Ozden,
S. Itani, L. Guo, A. Sacks, J. D. Goldsmith, M. D. Modell, E. B. Hanlon,
I. Itzkan and L. T. Perelman, Nature Med., 2010, 16, 603.
M. A. Kara, M. E. Smits, W. D. Rosmolen, A. C. Bultje, F. J. Ten Kate,
P. Fockens, G. N. Tytgat and J. J. Bergman, Gastro. Endosc., 2005, 62, 671.
W. Curvers, L. Baak, R. Kiesslich, A. Van Oijen, T. Rabenstein,
K. Ragunath, J. F. Rey, P. Scholten, U. Seitz, F. Ten Kate, P. Fockens and
J. J. Bergman, Gastroenterol., 2008, 134, 670.
T. Thomas, R. Singh and K. Ragunath, Surg. Endosc., 2009, 23, 1609.
M. A. Kara, F. P. Peters, W. D. Rosmolen, K. K. Krishnadath, F. J. ten
Kate, P. Fockens and J. J. Bergman, Endosc., 2005, 37, 929.
E. L. Bird-Lieberman, A. A. Neves, P. Lao-Sirieix, M. ODonovan,
M. Novelli, L. B. Lovat, W. S. Eng, L. K. Mahal, K. M. Brindle and
R. C. Fitzgerald, Nature Med., 2012, 18, 315.
M. A. Kara, F. P. Peters, W. D. Rosmolen, K. K. Krishnadath, F. J. ten
Kate, P. Fockens and J. J. Bergman, Endoscopy, 2005, 37, 929.
M. A. Kara, F. P. Peters, P. Fockens, F. J. ten Kate and J. J. Bergman,
Gastro. Endosc., 2006, 64, 176.
M. Filip, S. Iordache, A. Saftoiu and T. Ciurea, WJG, 2011, 17, 9.
W. L. Curvers, R. Singh, L. M. Song, H. C. Wolfsen, K. Ragunath,
K. Wang, M. B. Wallace, P. Fockens and J. J. Bergman, Gut, 2008, 57(2),
167172.
J. A. Evans, J. M. Poneros, B. E. Bouma, J. Bressner, E. F. Halpern,
M. Shishkov, G. Y. Lauwers, M. Mino-Kenudson, N. S. Nishioka and
G. J. Tearney, Clin. Gastro. Hepatico., 2006, 4, 3843.
M. Goetz and R. Kiesslich, Anticancer Res., 2008, 28, 353.
H. Pohl, T. Rosch, M. Vieth, M. Koch, V. Becker, M. Anders,
A. C. Khalifa and A. Meining, Gut, 2008, 57, 1648.
L. B. Lovat, K. Johnson, G. D. Mackenzie, B. R. Clark, M. R. Novelli,
S. Davies, M. ODonovan, C. Selvasekar, S. M. Thorpe, D. Pickard,
R. Fitzgerald, T. Fearn, I. Bigio and S. G. Bown, Gut, 2006, 55, 1078.
C. V. Raman and K. S. Krishnan, Nature, 1928, 122, 501.
N. Stone, C. Kendall and H. Barr, J. Biomed. Opt., 2008, 24, 203.
C. Kendall, J. Hutchings, H. Barr, N. Shepherd and N. Stone, Faraday
Discuss, 2011, 149, 279.
M. C. Grimbergen, C. F. van Swol, C. Kendall, R. M. Verdaasdonk,
N. Stone and J. L. Bosch, Appl. Spectrosc., 2010, 64, 8.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

220

Chapter 8

33. J. Horsnell, P. Stonelake, J. Christie-Brown, G. Shetty, J. Hutchings,

C. Kendall and N. Stone, The Analyst, 2010, 135, 3042.
34. J. Hutchings, C. Kendall, N. Shepherd, H. Barr and N. Stone, J. Biomed.
Opt., 2010, 15, 660.
35. C. Kendall, J. Day, J. Hutchings, B. Smith, N. Shepherd, H. Barr and
N. Stone, Analyst, 2010, 135, 3038.
36. M. Sattlecker, C. Bessant, J. Smith and N. Stone, Analyst, 2010, 135, 895.
37. J. C. Day, R. Bennett, B. Smith, C. Kendall, J. Hutchings, G. M. Meaden,
C. Born, S. Yu and N. Stone, Phys. Med. Biol., 2009, 54, 7077.
38. J. Hutchings, C. Kendall, B. Smith, N. Shepherd, H. Barr and N. Stone,
J. Biopoton., 2009, 2, 91.
39. C. Kendall, M. Isabelle, F. Bazant-Hegemark, J. Hutchings, L. Orr,
J. Babrah, R. Baker and N. Stone, Analyst, 2009, 134, 1029.
40. F. Bazant-Hegemark, K. Edey, G. R. Swingler, M. D. Read and N. Stone,
TCRT, 2008, 7, 483.
41. N. Stone, R. Baker, K. Rogers, A. W. Parker and P. Matousek, Analyst,
2007, 132, 899.
42. N. Stone, M. C. Hart Prieto, P. Crow, J. U and A. W. Ritchie, Anal.
Bioanal. Chem., 2007, 387, 1657.
43. G. Shetty, C. Kendall, N. Shepherd, N. Stone and H. Barr, Brit. J. Cancer,
2006, 94, 1460.
44. P. Crow, B. Barrass, C. Kendall, M. Hart-Prieto, M. Wright, R. Persad
and N. Stone, Brit. J. Cancer, 2005, 92, 2166.
45. P. Crow, A. Molckovsky, N. Stone, J. U, B. Wilson and
L. M. WongKeeSong, Urology, 2005, 65, 1126.
46. M. C. Prieto, P. Matousek, M. Towrie, A. W. Parker, M. Wright,
A. W. Ritchie and N. Stone, J. Biomed. Opt., 2005, 10(4), 440446.
47. N. Stone, C. Kendall, J. Smith, P. Crow and H. Barr, Faraday Discuss.,
2004, 126, 141.
48. P. Crow, N. Stone, C. A. Kendall, J. S. U, J. A. Farmer, H. Barr and
M. P. Wright, Brit. J. Cancer, 2003, 89, 106.
49. C. Kendall, N. Stone, N. Shepherd, K. Geboes, B. Warren, R. Bennett and
H. Barr, J. Path., 2003, 200, 602.
50. S. M. Angel and M. L. Myrick, Appl. Opt., 1990, 29, 1350.
51. M. L. Myrick, S. M. Angel and R. Desiderio, Appl. Opt., 1990, 29,
1333.
52. A. Mahadevan-Jansen, M. F. Mitchell, N. Ramanujam, U. Utzinger and
R. Richards-Kortum, Photochem. Photobiol., 1998, 68, 427.
53. M. G. Shim, B. C. Wilson, E. Marple and M. Wach, Appl. Spectrosc., 1999,
53, 619.
54. M. G. Shim, L. M. Song, N. E. Marcon and B. C. Wilson, Photochem.
Photobiol., 2000, 72, 146.
55. J. T. Motz, S. J. Gandhi, O. R. Scepanovic, A. S. Haka, J. R. Kramer,
R. R. Dasari and M. S. Feld, J Biomed. Opt., 2005, 10, 97.
56. L. M. Wong Kee Song, A. Molckovsky, K. K. Wang, L. J. Burgart,
B. Dolenko, R. J. Somorjai and B. C. Wilson, Proc. SPIE, 2005, 5692, 140.

View Online

14/10/2013 09:28:25.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00201

The Prospects for Real-Time Raman Spectroscopy for Oesophageal Neoplasia

221

57. Z. Huang, S. K. The, W. Zheng, J. Mo, K. Lin, X. Shao, K. Y. Ho, M. The

and K. G. Yeoh, Opt. Lett., 2009, 34, 758.
58. M. S. Bergholt, W. Zheng, K. Lin, K. Y. Ho, M. Teh, K. G. Yeoh, J. B. So
and Z. Huang, Analyst, 2010, 135, 3162.
59. M. S. Bergholt, W. Zheng, K. Lin, K. Y. Ho, M. Teh, K. G. Yeoh, J. B. So
and Z. Huang, TCRT, 2011, 10, 103.

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

CHAPTER 9

Volatile Analysis for Clinical

Diagnostics
CATHERINE KENDALL,*a,b HUGH BARRa,b,c AND
NARESH MAGANb
a

Biophotonics Research Unit, Leadon House, GHNHSFT, Gloucester,

GL1 3NN, United Kingdom; b Craneld Health, Craneld University,
MK43 0AL, United Kingdom; c Gloucestershire Royal Hospital,
Gloucestershire NHS Foundation Trust, United Kingdom
*Email: [email protected]

9.1 Volatile Diagnostics

The human nose is used as a qualitative analytical tool for discrimination/
detection of odours from food (sh, meat, cheese, wine), perfumes (cosmetics,
soaps) and avours. Since the time of the ancient Greeks, physicians have been
aware that it is possible to smell disease on the breath of those aicted. An
obvious example of this is the pear drop like smell of acetone in the breath of a
diabetic or bodily uids such as urine can also have distinctive smells, which are
specic to disease, such as the honey-like smell of urine in phenylketonuria.
People can be highly skilled at detecting odours, but there may be problems
because of a lack of sensitivity over time and the variation between individuals.
It may also sometimes be unsafe or unhealthy and variable depending on the
individual.1 To avoid these diculties, attempts have been made to build an
articial system that can mimic the olfactory sense of mammals and that can
detect and discriminate the production of volatile compounds from various
sources. Since the 1970s the potential of breath analysis as a way of detecting
RSC Detection Science Series No. 2
Detection Challenges in Clinical Diagnostics
Edited by Pankaj Vadgama and Serban Peteu
r The Royal Society of Chemistry 2013
Published by the Royal Society of Chemistry, www.rsc.org

222

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Volatile Analysis for Clinical Diagnostics

223

disease in the human body has been recognised. These early studies were primarily performed using gas chromatography-mass spectrometry (GC-MS). In
1971 Pauling et al.2 demonstrated that several hundred compounds were present in human breath and that some of these could be associated with abnormal
physiological states or disease. When considering breath analysis, diseases of
the lung and respiratory tract are the obvious choice of pathologies to be investigated as the vast majority of expired breath gas has been produced in the
lung. Breath analysis is attractive because it oers the potential for rapid,
noninvasive near-patient diagnosis.
All the methods for breath analysis rely on the detection of volatile organic
compounds or VOCs. Organic compounds are those that contain carbon and
are found in all living things. VOCs are organic compounds that easily become
gases or vapours; as such they can be detected in breath and bodily uids.
Many studies, including those using GC-MS have shown that for certain
pathological conditions (infection, malignancy, liver and cardiopulmonary
disease) specic patterns of VOCs can be detected. These patterns are known as
volatile ngerprints. It is hoped that these ngerprints will provide the basis
for accurate breath screening for disease in the future. These same technologies
can also be employed to smell the headspace above bodily uids such as blood
and urine, which adds other potential medical applications to the uses of these
technologies.

9.1.1 GC-MS and SIFT-MS Techniques

Techniques such as gas chromatography (GC), gas chromatography linked
with mass spectrometry (GC-MS) and selected ion ow tube-mass spectrometry (SIFT-MS) have been used to analyse volatile compounds. SIFT-MS
was developed in 1976 for real-time quantication of gas analysis of ambient
air, exhaled breath and the headspace above liquids.3,4 Headspace analysis of
liquid samples such as urine to detect nephrotoxicity,5 blood to detect bacterial
growth6 and human sweat to investigate stimulant compounds for mosquitoes
causing malaria7 have also been examined.
A large number of studies have been carried out to identify Microbial VOCs
(MVOCs) growing on a diversity of matrices using chromatographic techniques
but only a few have investigated these compounds from micro-organisms involved in ventilator associated pneumonia (VAP). Humphreys8 used GC-MS
for patients breath analysis in order to identify patients with VAP. Scotter
et al.9 used SIFT-MS to nd the ngerprints of volatiles from medically important fungi (Aspergillus avus, A.fumigatus, Candida albicans, Mucor racemosus, Fusarium solani and Cryptococcus neoformans) grown on a variety of
laboratory media (SDA, BHIB, Columbia agar, blood agar). Others10 have
identied MVOCs present in diverse building materials and from isolated and
subcultured pure mould cultures (Stachybotrys chartarum, Aspergillus penicillioides and Chaetomium globosum) related to allergic status using GC-MS
and solid-phase microextraction (SPME). Finally, there has been interest in
monitoring mycotoxigenic fungal strains of Penicillium roqueforti using SPME

View Online

224

Chapter 9

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

with GC-MS by the detection of some metabolites formed in the biosynthesis of

the mycotoxins.11

9.1.2 Electronic Noses

In the 1960s the development of instruments to detect odours started using
mechanical noses based on redox reactions. It was not until the 1980s when
integrated e-nose (electronic-nose) systems began appearing as research tools
and commercial devices. The rst electronic nose was described by Persaud
and Dodd in 1981.12 They attempted to replicate the various stages of the
human olfactory process using biochemical sensors. Volatile compounds react
on the surface of the sensors, neural networks can be used to analyse the sensor
responses and evaluate the main useful components of the data, enabling in
volatile odour recognition. The currently accepted denition of an e-nose, regarding the chemical sensor arrays used to detect odorant molecules, was
proposed by Gardner and Bartlett in 1994:13
An electronic nose is an instrument which comprises an array of electronic
chemical sensors with partial sensitivity and an appropriate pattern recognition
system, capable of recognising simple or complex odours.
The odorant molecules that interact with the surface of the sensors in an
e-nose (equivalent to the stimuli in the olfactory system) are typically hydrophobic and volatile due to their molecular weight (30300 Da), strength of
the interactions between molecules and molecular shape.14 Many of the
odorant molecules are produced by micro-organisms. The ability of the e-nose
to detect volatile ngerprints combined with pattern recognition systems may
permit the early detection of specic micro-organisms or groups of microorganisms.
A wide range of MVOCs are released from both primary and secondary
metabolites and obtained from the breakdown of protein and carbohydrates
during the metabolic pathways. These compounds can be alcohols, ketones,
esters, lactones, aldehydes, sulfur and nitrogen-based products, as well as
mycotoxins such as trichothecenes. These MVOCs can be characteristic and
dierent for each species.15,16 They also can vary depending on temperature,
the age of cultures, the nutritional status17 and the dierent water availability
of the substrates.16
Micro-organisms are known to produce a wide range of dierent organic
compounds that are released, with some of them being volatile. These MVOCs
may be used as ngerprints of microbial growth and perhaps to discriminate
between dierent species.
Analytical approaches involving more complex devices such as gas chromatography associated with mass spectrometry (GC-MS) and selected ion ow
tube-mass spectrometry (SIFT) have also been employed to detect volatile
biomarkers in headspace samples for similar applications. However, such approaches are more expensive and require a high level of expertise for analyses
and interpretation of data sets.

View Online

225

Volatile Analysis for Clinical Diagnostics

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

9.1.2.1

E-Nose Technology

Typically, an e-nose consists of three parts: a sensor array, where the sensors
measure electrical changes generated by adsorption of volatiles to the odoursensitive chemical surface; a transducer that converts signals in readable values
and the software for the analysis of the data. This structure has been copied
from the olfactory organ of mammals in the sense that the active material of the
sensors in an e-nose device (olfactory receptors) interacts with the volatiles
creating a sensor response (electric signal or pattern) which is transmitted to the
preprocessor (olfactory bulb) where the output is amplied and the noise is
reduced. In the nal stage the simplied signals (nerve impulses), as patterns of
responses, are processed into the data-analysis system (hypothalamus and olfactory cortex in the brain).18

9.1.2.2

Types of Odour Sensors

Since the rst model of an e-nose reported by Persaud and Dodd,12 many
advances have been made in the development of sensor technology. Some of the
most important aspects of sensors are that they must be sensitive enough to the
chemicals even at low concentrations and each sensor in the array must have
partial sensitivity. Thus, response to a range of dierent gases rather than to a
specic one is an advantage. This variable sensitivity leads to the production of
distinct qualitative volatile ngerprints, enabling the identication of dierent
patterns.
The e-nose sensor technologies can be classied according to the sensor
material and its transduction or conversion principle. Table 9.1 provides a
summary of the main sensor transducer systems and their characteristics.

Table 9.1

Summary of the sensor technologies for e-noses. (Adapted

from refs. 67, 68.)

Category

Measures

Material

Electrochemical
MOS

Resistance

Oxides

MOSFET
CP

Capacitance
Resistance

Sensitivity Principle

5500
ppm
Catalytic metals ppm
Polymers
0.1100
ppm

Optical
Optical bres
Intensity/
Fluorescent
Fluorescence
Spectrum
dyes
Chemoluminescence
Piezoelectric
SAW
Piezoelectricity Polymer-coated
resonating
QCM
quartz disc

Low ppb

Changes in
current or in
potential

Changes in light
intensity

1 pg mass Changes in mass,

change
density, viscosity
1 ng mass
and acoustic
change
phenomena

View Online

226

Chapter 9

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Electrochemical sensors can be:

a) Conductivity sensors: these exhibit a change in resistance when they are
exposed to volatile organic compounds. There are two types of conductivity sensors:
1. Conducting polymer sensors (CP sensors): This kind of sensor has
reversible physiochemical properties in terms of change in conductivity under ambient temperature conditions when the active material
(such as polypyrroles, thiophenes) is exposed to the chemicals, which
bond with the polymer backbone. They are extremely sensitive and the
standardisation of sample presentation is crucial for avoiding problems with humidity and drift over time.
2. Metal oxide semiconductor sensors (MOS sensors): In this case, the
doped materials with which the volatile organic compound (VOC)
interacts are oxides that can be made of tin, zinc, and tungsten. Those
materials are deposited between two metal contacts over a heating
element that operates at high temperatures, 200650 1C. When the
sample interacts with the oxide material, oxidation occurs. The output
signal is produced by a change in resistance caused by the reaction between the adsorbed oxygen and the gas molecules. They are quite sensitive (5500 ppm) and this can be improved by using noble metals or
changing the operating temperature. They are also resistant to humidity,
but they may have drift problems over time and they may be poisoned by
irreversible binding by dierent compounds such as sulfur or weak acids.
b) Metal oxide silicon eld eect sensors (MOSFET sensors): They consist
of three layers: a semiconductor, which can be made of silicon (Si) or
silicon carbide (SiC); a thick oxide layer (SiO2) as insulator; and on top a
thin catalytic metal layer made of platinum (Pt), iridium (Ir) or palladium
(Pd). In these sensors, the contact between the gas molecules and the
catalytic gate metal aects the voltage and thus change the capacitance
through the transistor. The gas detection is recorded as a voltage change
in the sensor signal. The sensitivity and selectivity can be determined by
the choice of the operation temperature, gate metal and structure of the
gate metal. The sensitivity of the sensors is normally high for low concentrations of the gases, while it becomes saturated for high concentrations of gases. MOSFET and MOS sensors operate at high
temperatures and are considered to be less sensitive to humidity with less
carryover from one measurement to another.19
Piezoelectric sensors can measure temperature, mass changes, pressure, force
and acceleration. The e-nose with piezoelectric sensors is designed to detect
mass changes. There are two types of piezoelectric sensors:
a) Quartz crystal microbalance (QCM sensors): They consist of a resonating
quartz disc that vibrates at a characteristic frequency (1030 MHz). When
the volatiles are absorbed there is an increase in the mass of the disc

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Volatile Analysis for Clinical Diagnostics

227

reducing the resonance frequency. This reduction is inversely proportional to the mass of the odorant absorbed.
b) Surface acoustic-wave devices (SAW sensors): These sensors measure
mass changes too but dier from those above in that they operate at much
higher frequencies and require waves to travel over the surface of the
device. They can be less sensitive than other piezoelectric sensors.
Optical sensors include, amongst other materials, glass bres coated with a
thin chemically active material that contains a uorescent dye. When the
volatiles interact with the dye under a light source of a determinate frequency,
or narrow range of frequencies, the polarity light of the uorescent dye is
altered and then a shift is produced in the emission spectrum. This change of
colour (or of uorescence, chemoluminescence, absorbance or reectance) can
be measured. Optical sensors have wide biological applications though they
present bleaching problems with sunlight.

9.2 Applications of Volatile Fingerprinting

These sensor array systems, so called e-noses, have been examined for potential
applications in many elds such as food quality control and spoilage, cosmetics
and beverages industries, environmental monitoring and more recently in
medical diagnostics.20

9.2.1 Food
E-nose technology has been applied in many elds, but the most frequently
considered has been the analysis of raw or manufactured food products;
freshness and maturity monitoring; shelf-life investigations; microbial detection
and control of food packaging material. In this eld, many studies have been
performed in order to nd the quality of raw materials like grains;21,22 to
evaluate the olive oil oxidation during storage;23 for freshness evaluation of
meats or sh; to classify a wide range of beverages such as coee, beer, wine;24
to recognise spoilage bacteria and yeasts in milk;25 to detect spoilage fungi in
bread26 and to monitor the ripening of Danish blue cheese.27
Today, there is signicant interest in detecting the levels of mycotoxins
produced by fungi in many food products. Thus, diverse studies using the
e-nose technology have been carried out to dierentiate between mycotoxigenic
and nonmycotoxigenic strains of Fusarium, Aspergillus, Penicillium in food raw
materials28,29 and in ham-based medium.30 Discrimination between mycotoxin
producers and nonproducers strains is based, as some studies have reported, on
the dierent volatile ngerprints produced because of the dierent biosynthetic
pathways for mycotoxin production used.
Food applications have received signicant attention because of the ease of
consistent sample presentation, sample size that may be required, it is noninvasive technique and the lack of sample preparation. Thus, potential exists in
using this approach as part of a quality-assurance system.31

View Online

228

Chapter 9

If the e-nose is able to recognise volatiles from bacterial and fungal growth on
food substrates, it may also be able to detect and diagnose certain illnesses.
There is a need for rapid, inexpensive and objective systems for clinical diagnosis. E-nose technology could provide a tool for early detection and reduce the
high costs and the time consuming nature of using molecular, microbiological
cultures and serological tests in the detection of diseases like Helicobacter
pylori,32 and Mycobacterium tuberculosis.18,33 The use of e-nose systems in
medicine for the early diagnosis of noninfectious diseases such as breast and
lung cancers34,35 kidney disorders and infectious diseases has been recognised33,3638 and will be discussed further. Clinical studies for rapid real-time
diagnosis have proved challenging. We conducted a prospective study to detect
the premalignant phenotype Barretts oesophagus that can progress to
oesophageal cancer. Patients attending for upper GI endoscopy who have given
written informed consent to participate were examined using the sning
endoscope aspirating the insuated gas through an e-nose device. The results
(Figure 9.1) revealed to complexity of obtaining a real-time in vivo diagnosis.
Although dierences are apparent the results are inconsistent and very individual. We will require technology improvement and very large studies. When
moving onto real patients, this very complex project encountered some
challenges.

Arbitrary Units

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

9.2.2 Medical Applications

Arbitrary Units
annotations ascribed to peaks for recognition

Figure 9.1

This gure shows that using an articial stomach it is possible to sni

the stomach and detect helicobacter infection. There is complete discrimination between H.pylori and sterile patterns.32 The project to develop
real-time sning endoscopy using volatile ngerprints in association
with standard visualization is complex.

View Online

Volatile Analysis for Clinical Diagnostics

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

9.2.2.1

229

Lung Cancer

In 1999 Phillips et al.39 found a combination of 22 VOCs in the breath of lungcancer patients that could be used to identify those with the disease. The
compounds were identied using GC-MS and were predominantly alkanes,
alkane derivatives and benzene derivatives. This was a cross-sectional study
that looked at breath samples from 108 patients with abnormal chest X-rays.
Patients then underwent bronchoscopy and histological or cytological sampling.
60 patients were conrmed to have lung cancer. Patients underwent breath
sampling within 24 h prior to bronchoscopy and the breath samples were analysed using GC-MS techniques. 67 VOCs were common to the breath samples of
62 patients, of these VOCs 22 were selected by discriminate analysis. Using this
approach for the detection of stage 1 lung cancer the technique had 100%
sensitivity and 81.3% specicity. Crossvalidation of the technique correctly
predicted the diagnosis in 71.7% with lung cancer and 66.7% of those without.
Di Natale et al.40 used an electronic nose to identify lung cancer patients.
A total of 62 breath samples were taken from 60 individuals by the bagcollection method. 35 patients had lung cancer, 18 were used as reference
subjects and 7 patients were postsurgery for lung cancer. 100% of the lung
cancer patients were correctly classied, 94% of the controls were correctly
classied and 44% of the postsurgical patients were also correctly classied.
The misclassied postsurgical patients were classied as healthy controls.
Machado et al.41 studied the exhaled breath from individuals with lung
cancer and compared this with subjects with noncancer diseases and a healthy
group as noncancer control volunteers. They demonstrated that the e-nose
could identify the dierent characteristics of exhaled breath in patients with
lung cancer in a noninvasive way. A total of 330 of 388 breath samples collected
from this group were correctly classied; an overall accuracy of 85%. The
electronic nose had 71.4% sensitivity and 91.9% specicity. More recently Peng
et al.35 were able to dierentiate between distinct cancer types by means of
patient breath analysis using a nanosensor array.

9.2.2.2

Breast Cancer

In 2003 Phillips et al.42 used 22 C4C20 alkanes found in breath to identify

women suering with breast cancer from those without breast cancer. Two
hundred breath samples were obtained from women with abnormal mammograms who subsequently underwent breast biopsy. Fifty one cases of breast
cancer were detected in 198 consecutive biopsies. The breath test distinguished
between women with breast cancer and healthy volunteers with a sensitivity of
94.1% (48/51) and a specicity of 73.8% (31/42) (crossvalidated sensitivity
88.2% (45/51), specicity 73.8% (31/42)). Compared to women with abnormal
mammograms and no cancer on biopsy, the breath test identied breast cancer
with a sensitivity of 62.7% (32/51) and a specicity of 84.0% (42/50) (crossvalidated sensitivity of 60.8% (31/51), specicity of 82.0% (41/50)). The
negative predictive value (NPV) of a screening breath test for breast cancer was

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

230

Chapter 9

superior to a screening mammogram (99.93% versus 99.89%); the positive

predictive value (PPV) of a screening mammogram was superior to a screening
breath test (4.63% versus 1.29%).
Previous in vitro studies suggest it could be possible to diagnose microbial
pathogens of medical importance such as anaerobic bacteria Clostridium spp
and Bacteroides fragilis, and even to discriminate between dierent Clostridia
based on volatiles produced.43

9.2.2.3

Microbial Infections

Furthermore, there has been an increasing need for early detection of microbial
infections in the hospital environment. Dutta et al.44 distinguished between
dierent types of Staphylococcus aureus species: MRSA, MSSA and coagulasenegative staphylococci (C-NS) from swab samples of infected patients using a
sensor array. Other acquired diseases of great importance in hospitals are those
related to prolonged mechanical ventilation like sinusitis, bronchitis and
pneumonia.38,45
In 2004 Hockstein et al.46 used e-nose analysis of the breath of ventilated patients in an intensive-care unit to diagnose ventilator associated pneumonia (VAP).
He compared breath analysis to computed tomography (CT) of the chest. They
report an accuracy of at least 80% when comparing e nose analysis to chest CT.
Adrie et al.47 have reported increased levels of nitric oxide in the breath of
patients with pneumonia compared to those without pneumonia.
Humphreys et al.38 investigated the diagnosis of VAP by detecting microorganisms in bronchoalveolar lavage (BAL) uid in a prospective comparative
study of e-nose and microbiology using LDA for the data analysis. 77% of the
samples were correctly classied and 68% when patients on antibiotics were
included.
Thaler et al. (2006)48 was able to identify correctly biolms versus nonbiolm
producing Pseudomonas and Staphylococcus species with accuracy ranging
from 72.2% to 100% in patients suering with rhinosinusitis.

9.2.2.4

Bacterial Infections

Pavlou et al.36 used an e-nose to identify samples of urine that were infected
with bacteria. They were further able to identify the species of bacteria causing
the infection. Shyknon et al. used e-nose technology to dierentiate between
dierent bacterial causes of otits media.49
Some recent studies have demonstrated the capacity of a hybrid metal-oxide/
metal-ion-based sensor systems to discriminate between dierent species of
Trichophyton that cause dermatophytosis.50 At present, the diagnosis of these
skin diseases takes up to 12 weeks. They were able to dierentiate between
species and distinct concentrations within 4 days of growth. This would permit
a rapid diagnosis, the appropriate use of antifungal drugs and also monitoring
their activity during the treatment. Naraghi et al.51 studied the potential of
using an e-nose for discriminating between ve antifungal treatments against

View Online

Volatile Analysis for Clinical Diagnostics

231

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

two dermatophytes species within 96120 h. This approach is important due to

the lack of accurate susceptibility testing, especially for dermatophytic fungi.

9.2.3 Environmental
Pollution has become one of the main concerns among public opinion and for
local governments. The water supply can be contaminated with faecal microorganisms and with chemical products discharged from industrial, agricultural
and animal farm activities. The traditional methods for detecting indicator
micro-organisms of pollution, such as microbiological cultures, can underestimate their presence in water. Sensor arrays have been used to monitor wastewater in treatment plants, sewage treatment works and to discriminate between
dierent products from sources as diverse as leather and automotive industries.
A wide spectrum of organisms, both bacteria and fungi, and chemical
components such as pesticides and insecticides can contaminate water supplies.52 E-noses have also been used to monitor the quality of potable water for
testing cyanobacterial species53 or detecting and dierentiating between
Streptomyces spp spores in reverse osmosis and tap water54 (Bastos and
Magan,54). Recently, they showed that soil quality and microbial status could be
dierentiated using volatile production patterns under dierent environmental
regimes.55 Other interesting applications such as control of health and safety air
quality in the aircraft-cabin environment have gained special interest.56

9.3 Ventilator-Associated Pneumonia (VAP)

Ventilator-associated pneumonia (VAP) is a signicant infection and has become of great medical importance because it is associated with antibioticresistant pathogens and might result in high rates of morbidity, mortality and
increased medical care costs. The diagnosis of VAP is still dicult since there is
no gold standard diagnostic test and the specicity and sensitivity of all existing
diagnostics are not very high. Clinical signs of VAP are not specic and the
scoring systems do not give high accuracy for VAP diagnosis.
VAP is the acquired pneumonia that occurs in patients who require mechanical ventilator support by endotracheal tube or tracheostomy forZ48 h.57
Therefore, VAP is considered a nosocomial disease because it is acquired in the
health care environment with no evidence that it is present or incubated at the
time of hospital admission.
VAP can be caused by a range of individual or mixtures of micro-organisms.
Gram-negative enteric rods, Staphylococcus aureus and Pseudomonas aeruginosa are the predominant organisms responsible for this infection. However,
many other micro-organisms can cause VAP including Streptococcus pneumoniae, Haemophilus inuenzae and Acinetobacter baumanii depending on the
onset of the illness.58 There are other opportunistic micro-organisms such as
Candida albicans and Aspergillus fumigatus that can play a role in immunecompromised patients. Polymicrobial infections appear to be especially high in
patients with acute respiratory distress syndrome (ARDS).

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

232

Chapter 9

Identication of the micro-organisms that may have the most clinically impact on a patient, depends on the individual patient, the type of intensive-care
unit (ICU), duration of stay and on any previous antibiotic treatment.58 Several
studies suggest that the time of onset determines that pathogens will infect and
the prognosis of the disease.59,60 For the latter authors H.inuenzae, S.pneumoniae, methicillin sensitive S.aureus (MSSA) and some Enterobacteriaceae
were common in early-onset VAP, while P.aeruginosa, Acinetobacter spp and
methicillin resistant S.aureus (MRSA) were typically more common in lateonset infections. The latter species belong to the group of micro-organisms that
potentially can be multidrug resistant (MDR). These pathogens have a high
and very variable range of mortality, between 050%, due to their specic risk
factors, patterns of clinical resolution and resistance.61
The incidence of VAP is estimated at between 828%59,60 and in all cases
there has been a close relationship between VAP and the length of mechanical
ventilation. Mortality rates due to VAP may reach values of up to 50%, especially in elderly populations, in severely ill patients, and those who may have
received inappropriate previous drug therapy. However, some studies have not
found any relationship between the high rates of mortality and VAP, but are
mostly due to the underlying diseases.
The diagnosis of VAP is dicult because there is no specic and single
clinical criterion. A large number of diseases, even noninfectious, may present
similar clinical signs such as ARDS, nosocomial tracheobronchitis, congestive
heart failure, atelactasis, pulmonary thromboembolism and pulmonary
haemorrhage. Likewise, the lack of a gold standard in diagnosis may result in a
nonspecic and sometimes unsuccessful treatment, which exposes the patients
to high levels of toxicity, longer stays in hospital, and elevates the costs and
risks of having resistant micro-organisms.62
There have been various clinical strategies for VAP diagnosis such as CPIS
(clinical pulmonary infection score) developed and modied by Pugin et al.63
although some consider CPIS of little usefulness for daily evaluation due to the
need of a waiting period for culture results and for its low sensitivity and
specicity.64 Also, CPIS has not been validated in immunosuppressed patients.
Regarding VAP treatment, the decisive points to be considered once the
therapy for VAP is chosen are the antibiotic selection and its duration. The
empiric antibiotic therapy has to be based on knowledge of the most common
micro-organisms and its local resistance pattern within the ICU,65 on the
duration of mechanical ventilation and on the previous treatment, if there was
one. The duration of the therapy has been lately reduced due to the observations that prolonged therapies lead to colonization with antibiotic resistant
bacteria such as P.aeruginosa, S.aureus or Enterobacteriaceae. This can result in
a relapse of VAP, mainly in the second week of treatment.66 An early eective
diagnosis and therapy are crucial for good outcomes with less antibiotic use
and therefore for reducing mortality.
Work has been undertaken to examine the potential for using qualitative
volatile ngerprints of specic micro-organisms involved in VAP infections
using the hybrid sensor e-nose to discriminate between dierent species; to

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Volatile Analysis for Clinical Diagnostics

233

build an in vitro model based on these data sets to interpret samples from VAP
patients to enable better and more accurate treatments to be applied.
Humphreys et al.38 investigated the diagnosis of VAP by detecting microorganisms in bronchoalveolar lavage (BAL) uid in a prospective comparative
study of e-nose and microbiology using LDA for the data analysis. 77% of the
samples were correctly classied and 68% when patients on antibiotics were
included.

Acknowledgements
The following people are kindly acknowledged for their help: Neus Planas
Pont, Napoleon Charaklias, Martyn Lee Humphries, Natasha Sahgal, Kamran
Naraghi.

References
1. H. T. Nagle, R. Gutierrez-Osuna and S. S. Schiman, The how and why of
electronic noses, Spectrum, IEEE, 1998, 35, 2231.
2. L. Pauling, et al., Quantitative analysis of urine vapor and breath by gasliquid partition chromatography, Proc. Natl. Acad. Sci. USA, 1971, 68(10),
23746.
3. P. Spanel and D. Smith, Selected ion ow tube mass spectrometry for online trace gas analysis in biology and medicine, Eur. J. Mass Spectrom.,
2007, 13, 7782.
4. P. Spanel and D. Smith, The challenge of breath analysis for clinical
diagnosis and therapeutic monitoring, Analyst, 2007, 132, 390396.
5. K. J. Boudonck, M. W. Mitchell, L. Nemet, L. Keresztes, A. Nyska,
D. Shinar and M. Rosenstock, Discovery of metabolomics biomarkers for
early detection of nephrotoxicity, Toxicol Pathol., 2009, 37, 280292.
6. J. M. Scotter, R. A. Allardyce, V. S. Langford, A. Hill and D. R. Murdoch,
The rapid evaluation of bacterial growth in blood cultures by selected ion
ow tube-mass spectrometry (SIFT-MS) and comparison with the BacT/
ALERT automated blood culture system, J. Microbiol. Methods, 2006, 65,
628631.
7. J. Meijerink, M. A. H. Braks, A. A. Brack, W. Adam, T. Dekker,
M. A. Posthumus, T. A. Van Beek and J. J. A. Van Loon, Identication of
olfactory stimulants for Anopheles gambiae from human sweat samples,
J. Chem. Ecol., 2000, 26, 13671382.
8. M. L. Humphreys, Volatile diagnostic techniques for ventilator associated
pneumonia. DM Thesis, 2010, Craneld University.
9. J. M. Scotter, V. S. Langford, P. F. Wilson, M. J. McEwan and
S. T. Chambers, Real-time detection of common microbial volatile
organic compounds from medically important fungi by Selected Ion Flow
Tube-Mass Spectrometry (SIFT-MS), J. Microbiol. Methods, 2005, 63,
127134.

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

234

Chapter 9

10. L. Wady, A. Bunte, C. Pehrson and L. Larsson, Use of gas

chromatography-mass spectrometry/solid phase microextraction for the
identication of MVOCs from moldy building materials, J. Microbiol.
Methods, 2003, 52, 325332.
11. J. C. R. Demyttenaere, R. M. Morina and P. Sandra, Monitoring and fast
detection of mycotoxin-producing fungi based on headspace solid-phase
microextraction and headspace sorptive extraction of the volatile metabolites, J. Chromat. A, 2003, 985, 127135.
12. K. Persaud and G. Dodd, Analysis of discrimination mechanisms in the
mammalian olfactory system using a model nose, Nature, 1982, 299(5881),
3525.
13. J. W. Gardner and P. N. Bartlett, A brief history of electronic nose, Sens.
Actuators B, 1994, 18, 210211.
14. J. W. Gardner and P. N. Bartlett, Electronic Noses: Principles and Applications. Oxford University Press, Oxford, 1999.
15. T. O. Larsen, Identication of cheese-associated fungi using selected ion
monitoring of volatile terpens, Lett. Appl. Microbiol., 1997, 24, 463466.
16. N. Magan and P. Evans, Volatiles as an indicator of fungal activity and
dierentiation between species, and the potential use of electronic nose
technology for early detection of grain spoilage, J. Stored Products Res.,
2000, 36, 319340.
17. K. Fiedler, E. Schutz and S. Geh, Detection of microbial volatile organic
compounds (MVOCs) produced by moulds on various materials, Int. J.
Hyg. Environ. Health, 2001, 204, 111121.
18. A. P. F. Turner and N. Magan, Electronic noses and disease diagnostics,
Nature Rev.: Microbiol., 2004, 2, 16.
19. E. Schaller, J. O. Bosset and F. Escher, Electronic nose and their application
to food, Lebensmittel-Wissenschaft und-Technologie, 1998, 31, 305316.
20. J. W. Gardner, H. W. Shin and E. L. Hines, An electronic nose system to
diagnose illness, Sens. Actuators B, 2000, 70, 1924.
21. A. Jonsson, F. Winquist, J. Schnurer, H. Sundgren and I. Lundstrom,
Electronic nose for microbial quality classication of grains, Int. J. Food
Microbiol., 1997, 35, 187193.
22. P. Evans, K. C. Persaud, A. S. McNeish, R. W. Sneath, N. Hobson and
N. Magan, Evaluation of a radial basis function neural network for the
determination of wheat quality from electronic nose data, Sens. Actuators
B, 2000, 69, 348358.
23. S. Buratti, S. Benedetti and M. S. Cosio, An electronic nose to
evaluate olive oil oxidation during storage, Ital. J. Food Sci., 2005, 17,
203210.
24. C. Di Natale, A. Macagnano, F. Davide, A. DAmico, R. Paolesse,
T. Boschi, M. Faccio and G. Ferri, An electronic nose for food analysis,
Sens. Actuators B, 1997, 44, 521526.
25. N. Magan, A. Pavlou and I. Chrysanthakis, Milk-sense: a volatile sensing
system recognises spoilage bacteria and yeasts in milk, Sens. Actuators B,
2001, 72, 2834.

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Volatile Analysis for Clinical Diagnostics

235

26. G. Keshri, P. Voysey and N. Magan, Early detection of spoilage moulds in

bread using volatile production patterns and quantitative enzyme assays,
J. Appl. Mycol., 2002, 92, 165172.
27. J. Trihaas, L. Vognsen and P. V. Nielsen, Electronic nose: New tool in
modelling the ripening of Danish blue cheese, Int. Dairy J., 2005, 15, 679691.
28. G. Keshri and N. Magan, Detection and dierentiation between mycotoxigenic and non-mycotoxigenic strains of two Fusaria spp. using volatile production proles and hydrolytic enzymes, J. Appl. Mycol., 2000, 89, 825833.
29. N. Sahgal, R. Needham, F. J. Cabanes and N. Magan, Potential for detection and discrimination between mycotoxigenic and non-toxigenic
moulds using volatile production patterns: A review, Food Additives Contamin., 2007, 24, 11611168.
30. M. Camardo Leggieri, N. Planas Pont, P. Battilani and N. Magan, Detection and discrimination between ochratoxin producer and non-producer
strains of Penicillium nordicum on a ham-based medium using an electronic
nose, Mycotox Res., 2010, DOI: 10.1007/s12550-010-0072-5.
31. N. Magan and N. Sahgal, Electronic noses for quality and safety
control, in Advances in Food Diagnostics. Eds. L. M. L. Nollet, F. Toldra
and Y. H. Hui, Blackwell, Iowa, USA, 2007.
32. A. K. Pavlou, N. Magan, D. Sharp, J. Brown, H. Barr and A. P. F. Turner,
An intelligent rapid odour recognition model in discrimination of Helicobacter pylori and other gastroesophageal isolates in vitro, Biosens. Biolectron., 2000, 15, 333342.
33. A. K. Pavlou, N. Magan, J. Jones, J. Brown, P. Klatser and A. P.
F. Turner, Detection of Mycobacterium tuberculosis (TB) in vitro and
in situ using an electronic nose in combination with a neural network
system, Biosens. Bioelectron., 2004, 20, 538544.
34. A. DAmico, G. Pennazza, M. Santonico, E. Martinelli, C. Rocioni,
G. Galluccio, R. Paolesse and C. Di Natale, An investigation on electronic
nose diagnosis of lung cancer, Lung Cancer, 2010, 68, 170176.
35. G. Peng, M. Hakim, Y. Y. Broza, S. Billan, R. Abdah-Bortnyak, A. Kuten,
U. Tisch and H. Haick, Detection of lung, breast, colorectal, and prostate
cancers from exhaled breath using a single array of nanosensors, Brit. J.
Cancer, 2010, 103, 542551.
36. A. K. Pavlou, Magan, C. McNulty, J. M. Jones, D. Sharp, J. Brown and
A. P. F. Turner, Use of an electronic nose system for diagnoses of urinary
tract infections, Biosens. Biolectron., 2002, 17, 893899.
37. P. Boillot, E. L. Hines, J. W. Gardner, R. Pitt, S. John, J. Mitchell and
D. W. Morgan, Classication of bacteria responsible for ENT and eye
infection using the Cyranose System, IEEE Sens. J., 2002, 3, 247253.
38. L. Humphreys, R. M. L. E. Orme, P. Oore, N. Charaklias, N. Sahgal,
N. Planas Pont, N. Magan, N. Stone and C. A. Kendall, Electronic nose
analysis of bronchoalveolar lavage uid, Eur. J. Clin. Invest., 2011, 41,
5258.
39. M. Phillips, et al., Volatile organic compounds in breath as markers of lung
cancer: a cross-sectional study, Lancet, 1999, 353(9168), 19303.

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

236

Chapter 9

40. C. Di Natale, et al., Lung cancer identication by the analysis of breath by

means of an array of non-selective gas sensors, Biosens Bioelectron., 2003,
18(10), 120918.
41. R. F. Machado, O. Deenderfer, T. Burch, S. Zheng, P. J. Mazzone,
T. Mekhail, C. Jennings, J. K. Stoller, J. Pyle, J. Duncan, R. A. Dweik and
S. C. Erzurum, Detection of lung cancer by sensor array analyses of exhaled breath., Am. J. Resp Crit. Care Med., 2005, 171, 12861291.
42. M. Phillips, R. N. Cataneo, C. Saunders, P. Hope, P. Schmitt and J. Wai,
Volatile biomarkers in the breath of women with breast cancer, J. Breath
Res., 2010, 4, DOI: 10.1088/1752-7155/4/2/026003.
43. A. K. Pavlou, A. P. F. Turner and N. Magan, Recognition of anaerobic
bacterial isolates in vitro using electronic nose technology, Lett. Appl.
Microbiol., 2002, 35, 366369.
44. R. Dutta, D. Morgan, N. Baker, J. W. Gardner and E. L. Hines, Sens.
Actuators B, 2005, 109, 355362.
45. M. J. Rumbak, Pneumonia in patients who require prolonged mechanical
ventilation, Microbes Infect., 2005, 7, 275278.
46. N. G. Hockstein, et al., Diagnosis of pneumonia with an electronic nose:
correlation of vapor signature with chest computed tomography scan
ndings, Laryngoscope, 2004, 114(10), 17015.
47. C. Adrie, et al., Exhaled and nasal nitric oxide as a marker of pneumonia
in ventilated patients, Am. J. Respir. Crit. Care Med., 2001, 163(5),
11439.
48. E. R. Thalen and C. W. Hansen, Use of an electronic nose to diagnose
bacterial sinusitis, Am. J. Rhinol., 2006, 20, 170172.
49. M. E. Shyknon, D. W. Morgan, R. Dutta, E. L. Hines and J. W. Gardner,
Clinical evaluation of the electronic nose in the diagnosis of ear, nose
and throat infection: a preliminary study, J. Laryngol. Otol., 2004, 118,
706709.
50. N. Sahgal, B. Monk, M. Wasil and N. Magan, Tricophyton species: use of
volatile ngerprints for rapid identication and discrimination, Brit. J.
Dermatol., 2006, 155, 12091216.
51. K. Naraghi, N. Sahgal, B. Adriaans, H. Barr and N. Magan, Use of
volatile ngerprints for rapid screening of anti-fungal agents for ecacy
against dermatophyte Trichophyton species, Sens. Actuators B, 2010, 146,
521526.
52. O. Canhoto and N. Magan, Electronic nose technology for the detection of
microbial and chemical contamination of potable water, Sens. Actuators B,
2005, 106, 36.
53. J. W. Gardner, H. W. Shin, E. L. Hines and C. S. Dow, An electronic nose
system for monitoring the quality of potable water, Sens. Actuators B,
2000, 69, 336341.
54. A. C. Bastos and N. Magan, Potential of an electronic nose for the early
detection and dierentiation between Streptomyces in potable water, Sens.
Actuators B, 2006, 116, 151155.

View Online

14/10/2013 09:28:34.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00222

Volatile Analysis for Clinical Diagnostics

237

55. A. C. Bastos and N. Magan, Soil volatile ngerprints: Use for discrimination between soil types under dierent environmental conditions, Sens.
Actuators B, 2007, 125, 556562.
56. J. D. Spengler and D. G. Wilson, Air quality in aircraft, J. Process. Mech.
Eng., 2003, 217, 323335.
57. J. D. Hunter, Ventilator associated pneumonia, Postgrad. Med. J., 2006,
82, 172178.
58. E. Bouza, C. Brun-Buisson, J. Chastre, S. Ewig, J. Y. Fagon,
C. H. Marquette, P. Munoz, M. S. Niederman, L. Papazian, J. Rello,
J. J. Rouby, H. Van Saene and T. Welte, Ventilator-associated pneumonia,
Eur. Respir. J., 2001, 17, 10341045.
59. J. Chastre and J. Y. Fagon, Ventilator-associated pneumonia, Am. J.
Respir. Crit. Care Med., 2002, 165, 867903.
60. M. H. Kollef, What is ventilator-associated pneumonia and why is it important?, Respir. Care, 2005, 50, 714724.
61. J. Rello, E. D az and A. Rodr guez, Etiology of Ventilator-Associated
Pneumonia, Clin. Chest Med., 2005, 26, 8795.
62. J. S. Solomkin, Ventilator-associated pulmonary infection: the germ theory
of disease remains viable, Microbes Infect., 2005, 7, 279291.
63. J. Pugin, R. Auckenthaler, N. Mili, J. P. Janssens, P. D. Lew and
P. M. Suter, Diagnosis of ventilator-associated pneumonia by bacteriologic
analysis of bronchoscopic and nonbronchoscopic blind bronchoalveolar
lavage uid, Am. Rev. Respir. Dis., 1991, 143(1), 11211129.
64. R. P. Baughman, Microbiologic diagnosis of ventilator-associated pneumonia, Clin. Chest Med., 2005, 26, 8186.
65. S. M. Koenig and J. D. Truwit, Ventilator-associated pneumonia: Diagnosis, treatment and prevention, Clin. Microbiol. Rev., 2006, 19, 637657.
66. P. J. W. Dennesen, J. A. M. Van der Ven, A. G. H. Kessels, G. Ramsay and
M. J. M. Bonten, Resolution of infectious parameters after antimicrobial
therapy in patients with ventilator-associated pneumonia, Am. J. Respir.
Crit. Care Med., 2001, 163, 13711375.
67. T. C. Pearce, S. S. Schiman H. T. Nagle and J. W. Gardner, Handbook of
Machine Olfaction. Wiley-VCH, Weinheim, 2003.
68. E. Mahmoudi, Electronic nose technology and its applications, Sens.
Transduc. J., 2009, 107, 1725.

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

Subject Index
ABL800 FLEX analyser 58
abnormal location of immature
precursors (ALIP) in
myelodysplasia 183
acoustic network analysis 16
acoustic wave-based biosensor
technology 1418
acquired immunodeciency syndrome
see AIDS
acute myeloid leukaemia (AML) 182,
193, 194
acute respiratory distress syndrome
(ARDS) 2312
AIDS (acquired immunodeciency
syndrome) 91, 1079
Alere Triages System 37
American Diabetes Association
(ADA) 77
American College of
Rheumatology 115
Ames Reectance Meter
(ARM) 6677
p-aminobenzoic acid (PABA) 40
aminopropyltriethoxysilane
(APTES) 10
amperometric enzyme
biosensors 5254
amperometry 13, 116
angel investors 1745
anodic stripping voltammetry
(ASV) 97
anticyclic citrullinated peptide (CCP)
antibody 115
antimicrobial peptides (AMPs) 43
Aspergillus species 227
autouorescence endoscopy (AFI) 204

bacterial infections 2301

Barretts oesophagus 202, 2034, 207,
209, 218, 228
benzocaine 40
biocompatibility (in vivo sensors
performance/reliability)
biological response in
subcutaneous tissue 1402
biological response in
blood 13940
performance patterns 1389
bioelectric recognition assay
(BERA) 15
biological response mediation
(materials)
active releasing
materials 14450
biological form/function 150
introduction 1434
biomodication approach (enzyme
properties) 44
biosensor technology and clinical
biochemistry laboratory signal
interference
biosensor technology
biosensor architecture
67
devices and
transduction 1121
probe attachment to
device surfaces 711
biosensors and measurement of
clinical targets 213
conclusions 301
laboratory clinical biochemistry
assays 16

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

Subject Index

signal interference and NSA

problem 239
surface chemistries and NSA
issue 2930
biosensors and measurement of
clinical targets
cost per analysis 22
functionality in complex
media 22
generic methodology 21
high throughput 21
speed of analysis 21
validation of doseresponse
and limit of detection
(LOD) 22
biotin 45
blood glucose monitoring system
(BGMS)
electrodes
uid-detection
electrode 72
reference/counter
electrode 72
substrate 701
working electrode 712
performance validation
accuracy 78
haematocrit
independence 79
interference 79
ISO 77, 78
precision 78
reliability 78
repeatability 778
sensors
calibration 767
chemistry 734
detection method 746
electrode 702
manufacturing 7980
performance
validation 779
reaction chamber 72
blood-glucose biosensors:
development and challenges
clinical utility and practical
problems 801

239

Continuous Blood Glucose

Monitoring System 812
history
1st generation 678
2nd generation 689
3rd generation 69
history of blood glucose
biosensors 669
introduction 656
sensors for BGMS 6980
Body Mass Index (BMI) 65
brain natriuretic peptide (BNP) 55
breast cancer 99, 22930
breath analysis 2223
bronchoalveolar lavage (BAL) 230,
233
C-reactive protein (inammation
marker) immunoassay 39
cancer 91, 92103
cancer antigen 15-3 (CA 15-3) 97
cancer antigen 19-9 (CA 19-9) 97
cancer antigen 125 (CA 125) 97
cancer biomarkers (electrochemical
sensors)
cancer cells 1003
carcinoembryonic antigen 923
genetic markers 989
other proteins 946
prostate-specic antigen
934, 97
simultaneous detection of
proteins 967
cancer cells 1003
carbon microelectrodes (CFEs) 164
carbon nanotubes (CNTs) 58, 136
Carbosils 150
carcinoembryonic antigen
(CEA) 923
cardiac disease 91, 1037
CE marking 216
celiac disease (gluten enteropathy)
biomarkers 91, 11618
cellulose 48
Center for Devices and Radiological
Health (CDRH) in US 77
central symmetry (crystals) 14

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

240

charge-coupled devices (CCDs) 21,

40, 207, 209, 213
chromoendoscopy 203
chronic myelomonocytic leukaemia
(CMML) 192, 193
CPIS (clinical pulmonary infection
score) 232
CoagucheckXS 41
cobalt (II) tetraphenylporphyrins
(Co/IITPP) 167
computed tomography (CT) 230
conducting polymer (CP)
sensors 2256
Continuous Glucose Monitoring
System (CGMS) 812
continuous glucose monitoring
(CGM) 134, 136
correction algorithm for BGMS
haematocrit 75
temperature 756
Coxiella burnetti 43
creatinine measurement 54
crosslinked enzyme aggregate
(CLEA) 47
cyanocob(II)alamin 167
cytogenetics (MDS)
diagnosis 190
subclassication and
prognosis 1901
deamidated gliadin peptide
(DGP) 117
devices and transduction
acoustic wave-based
biosensor technology
1418
electrochemical devices
1214
electromagnetic radiation and
optical biosensor
technology 1821
dexamethasone (DX) 145, 148
Dextrostix 67
type 2 diabetes 65
Diabetes Control and Complications
Trial (DCCT) 66

Subject Index

diabetes mellitus
statistics 65
see also blood-glucose
biosensors
diazeniumdiolates 1467
DIDNTB (tetra
bromosulfonphthalein dye) 41
dierential pulse amperometry
(DPA) 165, 175
dierential pulse voltammetry
(DPV) 100, 107, 111, 175
DNA
biosensors 112
sensors 56
sequences 98, 100
dysplasia, denition 183
E-nose technology see electronic
noses
elastic scattering spectroscopy
(ESS) 204
Elecsyss analysis systems 115
electroactive polymers (EAPs) 16970
electrochemical detection of
disease-related diagnostic
biomarkers
AIDS 91, 1079
biorecognition 122
cancer 91, 92103
cardiac disease 91, 1037
celiac disease 91, 11618
challenges
mass production 121
real clinical samples 121
sample throughput 121
stability 121
conclusions 122
hepatitis 91, 11015
introduction 8991
rheumatoid arthritis 91, 11516
urinary tract infection 91,
11721
electrochemical impedance
spectroscopy (EIS) 14, 103
electrochemical sensor array
(ESA) 166

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

Subject Index

electromagnetic piezoelectric acoustic

sensor (EMPAS) 16, 30
electromagnetic radiation and optical
biosensor technology 1820
electron spin resonance spectroscopy
(ESR) 163
electronic noses
description 224, 226
E-nose technology 225, 227
food 227
medical applications 228
odour sensors 2257
endoscopic resection (ER) 214, 216
a-enolase (ENOA) 956
enzyme-linked immunosorbent assay
(ELISA)
AIDS 107
cardiac biomarkers 1056
celiac disease 117
electrochemical detection of
diagnostic biomarkers 90,
121, 122
hepatitis B 110, 11213
hepatitis C 11314
immunoassay 2
murine double minute 2 96
PSA 97
rheumatoid arthritis 11516
erythroid dysplasia 183
Escherichia coli 119
Europe in vitro diagnostics market 89
European League against
Rheumatism 115
European Manufacturers Diagnostics
Association 89
ferrocene 14, 108, 135
brinogen 139
eld-eect transistors (FETs) 12, 17
ow-injection analysis (FIA) 22, 31
Food and Drug Administration
(FDA)
angel investors 175
blood glucose monitoring 77
continuous glucose
monitoring 82

241

in vivo electrochemical glucose

sensors 137
OraQuick HCV rapid antibody
test 114
food and E-nose technology 227
foreign-body giant cells (FBGCs) 142
Forster Resonance Energy transfer
(FRET) 138
free radicals 156
French-American-British (FAB)
classication of MDS 182, 193,
194
Fusarium species 223, 227
GC-MS (gas chromatography linked
with mass spectrometry) 223, 224,
229
genetic markers (cancer) 989
glutaraldehyde (GA) 47, 144
gold ultramicroelectrodes 166
granulocytic dysplasia 184
graphene 58, 1724
GRIN lens (Raman spectroscopy) 213
H-Filters system 38
Helicobacter pylori 228
hemin-PEDOT lms 169, 170, 1756
hepatitis biomarkers 91, 11015
B biomarkers 11013
C biomarkers 11315
types 110
hepatitis B surface antigen
(HBsAg) 11011, 113
high-grade dysplasia (HGD) 202,
203, 207, 217
HIV (human immune deciency virus)
HIV-1 1089
HIV-2 109
DNA biosensors 109
electrochemical sensors and
AIDS 1079
nanoparticle labels 42
PIMA CD4+ analyser 37
HIV-1 integrase 107
HIV-1 protease 1078
HIV-1 reverse transcriptase 107

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

242

human chorionic gonadotrophin

(hCG) for pregnancy testing 41, 42
human immune deciency virus
see HIV
human nose 222
humic acids (HAs) 136
i-STAT meter for blood gases 37, 58
ideal biosensor 223
IGFET (insulated gate FET) 12
IMMUNOFET 13
immunosensors 546
in vitro diagnostics market in
Europe 89
in vivo sensors for continuous
monitoring of blood gases, glucose
and lactate
biocompatibility 13842
biological response
mediation 14350
design 1308
introduction 12930
summary 1501
in vivo sensors design
glucose in subcutaneous
tissue 1347
lactate in blood 1378
PO2/PCO2/pH in blood 1304
indium-tin oxide 11
inammatory response process
1401
inherently conductive polymers
(IMPs) 169
inkjet printing of enzymes 45
integral membrane proteins (IMPs) 10
integrated chemistries for analytical
simplication and POCT
aptamers as biorecognition
elements 4850
biological recognition
aptamers 4850
enzyme-based
systems 478
immobilisation 436
commercialisation 578
conclusions 589

Subject Index

electrochemical sensors
amperometric enzyme
biosensors 524
DNA sensors 56
immunosensors 546
ion-selective
electrodes 502
uidics 369
introduction 356
lateral ow techniques from
simple colorimetric strips to
3D ow 4045
micromechanical
transduction 567
interdigital transducer (IDT) 17
interferometry 201
ion-selective electrodes (ISEs) 12,
502
ISFET (ion-selective eld eect
transducer) 13, 52
ISO 15197 (2003) for BGMS 77
isocitrate dehydrogenase (IDH)
oncogenes 195
lab-on-a-chip concept 5, 30, 58, 122
lactate sensors 534, 1378
lactoferrin 120
lateral ow techniques from simple
colorimetric strips to 3D ow
colorimetric strips 401
nanoparticle labels 413
new opportunities 43
LDA data analysis 230, 233
leukocyte esterase (LE) 120
Li+ selective uoroionophore 512
light-emitting diodes (LEDs) 40
linear discriminant analysis
(LDA) 208, 214
linosdomine see SIN-1
(1,3-morpholino-sydnomimine)
low-grade dysplasia 202, 217
lung cancer 229
MachZehnder device
(interferometry) 201
macrophages 140, 142

View Online

243

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

Subject Index

manganese tetraaminophthalocyanine
(MnTAPC) 1656
manganese(III) [2-2]paracyclo-phenyl
porphyrin (MnPCP) 1678
mast cells (MCs) 1412
MCF7 cancer cells 100
mean cell haemoglobin concentration
(MCHC) 188
medical applications (volatile
ngerprinting)
bacterial infections 2301
breast cancer 22930
lung cancer 229
microbial infections 230
megakaryocytic dysplasia 1845
mercaptohexanol 9
metal oxide semiconductor (MOS)
sensors 2256
metal oxide silicon eld eect
(MOSFET) sensors 2256
metallo-phthalocyanines (MPCs) 165
microbial infections 230
microbial VOCs (MVOCs) 223, 224
microelectrical mechanical systems
(MEMS) 36
microuidic paper analytical devices
(mPADS) 43
micromechanical transduction 567
migration inhibitory factor
(MIF) 11516
MMP (matrix metalloproteinase) 37
multi-walled carbon nanotube
polystyrene (MWCNT-PS)
composite 116
multidrug resistance (MDR)
organisms 232
multiple-tube fermentation
(MTF) 119
murine double minute 2 (MDM2) 96
myelodysplastic (MDS) syndromes
cytogenetics 1901
diagnostic diculties
cytopenias and
hypocellular bone
marrow 192

cytopenias and minimal

dysplasia 1912
dysplasia and bone
marrow brosis 1923
myelioproliferative/
myelodysplastic overlap
syndromes 192
diagnostic tools
bone marrow aspirate
cytochemistry 188
bone marrow trephine
biopsy 18890
immunophenotyping 191
dierential diagnosis
AML and other
haematopoietic
neoplasms 187
congenital disorders 187
drugs and toxins 1867
infectious diseases 187
metabolic abnormalities
186
initial assessment of patients 188
introduction 1823
morphology
description 1834
erythroid dysplasia 183
granulocytic dysplasia 184
megakaryocytic
dysplasia 1845
pathophysiology
altered methylation
1945
RNA splicing
machinery 196
subclassication
original FAB
classication 1934
other FAB
classication 1934
WHO classication 3rd/
4th edition 194
Naon coatings 143
nanocauliower shape 176
narrow-band imaging 203

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

244

National Health and Nutrition

Examination Survey III (US) 65
negative predictive value
(NPV) 22930
neutravidin 8
nitric oxide (NO)
in vivo sensors 142, 1469, 150
peroxynitrite 157, 162, 168
nitro oxidative metabolic stress 1567
S-nitroso-N-acetyl pencillamine
(SNAP-PDMS) 149
nonspecic adsorption (NSA) 239,
2930, 31
nordihydroguaiaretic acid
(NDGA) 144
odour sensors 2257
oesophageal neoplasia see Raman
spectroscopy
oestrogen receptor-a (ESR1) 99
one-test disposable system 6
ONOO see peroxynitrite
optical-ber oxygen sensors 132
optical biopsy 207
optical coherence tomography
(OCT) 204
optical pH sensors 13233
optical sensors 132, 227
optical-ber oxygen sensors 132
optrode term 18
OraQuick HCV rapid antibody
test 114
oxidised O-dianisidine (ODS) 55
p-aminobenzoic acid (PABA) 40
PEDOT-hemin lms 169, 170, 175
PEG alkanethiol 99
PEG eect 30
PEG (poly(ethylene glycol)) 30, 46,
99, 1434
Penicillium roqueforti 2234
Penicillium species 227
peruorocarboxylic acid ionomer
(PFCI) 144
peruorosulfonate ionomer
membrane (PFSI) 144

Subject Index

peroxynitrite (PON, ONOO )

electrochemical quantication
accuracy 15963
biochemical assays 163
conclusions and
perspectives 1746
decomposition catalysts 159,
160
interfaces and detection 1639
introduction 1569
pathological eects and
potential therapeutic
strategies 1601
radicals 1568
sensitivity eletroactive
polymers of graphene 16974
piezoelectric sensors 2267
Pima CD4+ analyser for HIV
testing 37
Plasmodium falciparum 423
PO2/PCO2/pH in blood 12930,
1304, 1389
point of care testing (POCT) see
integrated chemistries
polyaniline 170
poly-3, 4-ethylenedioxythiophene
(PEDOT) 1701, 1756
poly(4-vinylpyridine) lm 166
poly(cyano-cobalamin) lm 167
polydimethylsiloxane (PDMS) 104,
1467, 149
polydithenyl-pyrrole-benzoic acid
(PDB) 167
polyethersulfone (PES) 71
polyethyl terephthalate (PET) 71
polyethyleneimine (PEI) 167
polyethylphenylnaphthalate (PEN) 71
poly(hydroxyethyl methacrylate)
(PHEMA) 143
poly(lactic-co-glycolic acid)
(PLGA) 1456
polypyrrole 49, 170
poly(vinyl chloride) (PVC) 147, 150
poly (4-vynilpyridine) (PVP) lm 166
positive predictive value (NPV) 230
potentiometry 12

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

Subject Index

PPQ (pyrroloquinline quinone) 53

principal component analysis
(PCA) 208
principal-component-led linear
discriminant analysis (PC-fed
LSA) 214, 217
probe attachment to device surfaces
(biosensor technology)
covalent bonding 810
non-covalent attachment 78
self-assembled monolayer
chemistry 1011
procaine 40
prostate-specic antigen (PSA)
934, 97
prostate-specic membrane antigen
(PSMA) 97
Pursils 150
quartz crystal microbalance (QCM)
sensors 15, 116, 2257
Raman shift 205
Raman spectroscopy (real time) for
oesophageal neoplasia
clinical problem 2012
description 2046
discussion 21518
endoscopic recognition of
dysplasia and cancer
2024
bre-optic Raman
spectroscopy 21013
instrumentation 2067
medical diagnosis 2078
novel Raman probe 21315
oesophageal pathology 2068
Rayleigh wave see surface acoustic
wave
reactive nitrogen and oxygen species
(RNOS) 1567
recombinant immunoblot assay
(RIBA) 11314
red cell distribution width (RDW) 188
reduced graphene oxide (RgO)/hemin
nanocomposite 1723, 176

245

refractory anaemia with excess blasts

(RAEB) 191, 1934, 196
refractory anaemia with ringed
sideroblasts (RARS) 1934
Renishaw StreamLine system (Raman
mapping) 209
reverse transcriptase-linked
polymerase chain reaction
(RT-PCR) 114
rheumatoid arthritis (RA)
biomarkers 91, 11516
rheumatoid factor (RF) 115
RNA sequences 98
rotating disc electrode (RTE) 166
RSNO 1489, 150
salivary MMP 37
scanning electrochemical microscopy
(SECM) 68
self-assembled monolayer (SAM) 10
sensitivity: eletroactive polymers/
graphene
electroactive polymers 16971
other carbon-based
materials 1724
sensor technologies for e-noses 225
sensorgram 20
SIFT-MS (selected ion-ow tubemass spectrometry) 223, 224
signal-to-noise ratio (SNR) 213
silicon dioxide 11
simultaneous detection of proteins
(cancer markers) 967
SIN-1 (1,3-morpholinosydnonimine) 162, 167, 172, 176
single-walled carbon nanotube array
(SWCNT) 111
small nuclear ribonucleoproteins
(snRNPs) 196
sol-gels (SGs) 48, 144
solid-phase microextraction
(SPME) 223
squamous cell carcinoma antigen
(SCCA) 95
square-wave voltammetry (SWV) 90,
102, 105, 109

View Online

14/10/2013 09:28:44.
Published on 23 September 2013 on http://pubs.rsc.org | doi:10.1039/9781849737302-00238

246

Staphylococccus aureus 8, 56, 230

surface acoustic-wave (SAW) 16,
567, 225, 227
surface plasmon resonance
(SPR) 1820, 31, 11516
surface transfer wave (STW) 18
surface-blocking agents 44
target functional groups for biosensor
immobilization of proteins 89
Tecothane 147
TET2 (ten-eleven translocation) gene
family 195
tetraethylorthosilicate (TEOS) 8
the sning endoscope 228
thickness-shear mode (TSM)
technology 1417
thrombocytopenia 188
thyroid stimulating hormone (TSH)
test 2
tropinin I 103, 105
tropinin T 1036
ultramicroelectrodes (UMEs) 166
United Kingdom Oesophagogastric
Cancer Audit 2012
United Kingdom Prospective
Diabetes Study (UKPDS) 66
urinary tract infection (UTI),
biomarkers 91, 11721
Sensor Array 11920
uropathogens 11920
vaccinia virus 43
valinomycin 51
Vantixt sandwich 58

Subject Index

vascular endothelial growth factor

(VEGF) 97, 142, 145, 150
vascular endothelial growth factor-C
(VEGF-C) 97
Venezuelan equine encephalitis
virus 43
ventilator-assisted pneumonia
(VAP) 223, 2313
volatile analysis for clinical diagnosis
electronic noses 2247
GC-MS and SIFT-MS
technique 2234
ventilator-assisted
pneumonia 2313
volatile diagnostics 2223
volatile ngerprinting 22731
volatile ngerprinting
environmental 231
food 227
medical applications 2289
volatile organic compounds
(VOCs) 223, 226, 229
von Willebrands factor (vWF) 139
working electrodes (WEs) 169, 171,
172
World Health Organization (WHO)
diabetes 65
MDS classication 194
MDS subclassication 193
Yellow Spring Instrument (YSI) for
BGMS 76
Young device (interferometry) 201
zinc-oxide nanorods 138