Neuroscience and Biobehavioral Reviews

Download as pdf or txt
Download as pdf or txt
You are on page 1of 33

Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews


journal homepage: www.elsevier.com/locate/neubiorev

Review

A systematic review of fatigue in patients with traumatic brain injury:


The course, predictors and consequences
Tatyana Mollayeva a,b,c, , Tetyana Kendzerska d,1 , Shirin Mollayeva e,f,2 ,
Colin M. Shapiro g,h,3 , Angela Colantonio c,i,4 , J David Cassidy j,k,l,5
a

Graduate Department of Rehabilitation Science, Faculty of Medicine, University of Toronto, Canada


Collaborative Program in Neuroscience, University of Toronto, Canada
c
Toronto Rehab-University Health Network, Ontario, Canada
d
Department of Medicine, Division of Respirology, Sunnybrook Research Institute, University of Toronto, G1 06, 2075 Bayview Avenue, Toronto, Ontario
M4N 3M5, Canada
e
Faculty of Arts and Science, University of Toronto, Canada
f
Acquired Brain Injury Research Lab, University of Toronto, 160-500 University Avenue, Toronto, Ontario M5G 1V7, Canada
g
Toronto Western Hospital, University Health Network, 399 Bathurst Street, Rm 7MP421, Toronto, Ontario M5T 2S8, Canada
h
Youthdale Child & Adolescent Sleep Clinic, Ontario, Canada
i
Department of Occupational Science and Occupational Therapy, University of Toronto, Ontario, Canada
j
Faculty of Health, Institute of Sport Sciences and Clinical Biomechanics, University of Southern Denmark, Campusvej 55, Odense M 5230, Denmark
k
Division of Health Care and Outcomes Research, University Health Network, Canada
l
Division of Epidemiology, Dalla Lana School of Public Health, University of Toronto, Canada
b

a r t i c l e

i n f o

Article history:
Received 14 May 2014
Received in revised form
18 September 2014
Accepted 28 October 2014
Available online 6 November 2014
Keywords:
Post-traumatic fatigue
Traumatic brain injury
Rehabilitation
Systematic review

a b s t r a c t
Background: Fatigue is common after traumatic brain injury (TBI). Its risk factors, natural history and
consequences are uncertain. Best-evidence synthesis was used to address the gaps.
Methods: Five databases were searched for relevant peer-reviewed studies. Of the 33 articles appraised,
22 longitudinal studies were selected. Results were reported separately based on their timing of baseline
assessment.
Results: All studies document changes in fatigue frequency and severity with time, irrespective of setting
or TBI severity. There is limited evidence for certain clinical and psychosocial variables as predictors of
fatigue severity at follow-up. Early fatigue severity predicted persistent post-concussive symptoms and
Glasgow outcome score at follow-up.
Conclusions: Fatigue is present before and immediately following injury, and can persist long term. The
variation in ndings supports the idea of fatigue in TBI as a nonhomogeneous entity, with different factors
inuencing the course of new onset or chronic fatigue. To decrease the heterogeneity, we emphasize the
need for agreement on a core set of relevant fatigue predictors, denitions and outcome criteria.
PROSPERO registry number: CRD42013004262.
2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Abbreviations: APOE-4, apolipoprotein-4; BDI, Beck depression inventory; BFS, Barosso fatigue scale; CHART, Craig handicap assessment and reporting technique;
CNS, central nervous system; DRS, disability rating scale; FSS, fatigue severity scale; GCS, Glasgow coma scale; GOSE, Glasgow outcome scale-extended; GFI, global fatigue
inventory; HADS, hospital anxiety and depression scale; MFIS, modied fatigue impact scale; MFI, multidimentional fatigue inventory; mTBI, mild traumatic brain injury;
PRISMA, preferred reporting items for systematic reviews and meta-analyses; PCSC, post-concussion syndrome checklist; POMS, prole of moods scale; RCT, randomized
controlled trial; RPQ, Rivermead post-concussive questionnaire; SIGN, Scottish intercollegiate guidelines network; SF-36, 36-item short form health survey (from medical
outcomes study); TBI, traumatic brain injury; VAS, visual analog scale.
Corresponding author at: Toronto Rehabilitation Institute, 550 University Avenue, Rm 11207, Toronto, Ontario M5G 2A2, Canada. Tel.: +1 416 597 3422x7848;
fax: +416 946 8570.
E-mail addresses: [email protected] (T. Mollayeva), [email protected] (T. Kendzerska), [email protected] (S. Mollayeva),
[email protected] (C.M. Shapiro), [email protected] (A. Colantonio), [email protected] (J.D. Cassidy).
1
Tel.: +1 416 669 6759; fax: +1 416 946 8570.
2
Tel.: +1 416 978 1098; fax: +1 416 946 8570.
3
Tel.: +1 416 603 5800x5160; fax: +1 416 603 5292.
4
Address: Saunderson Family Chair in Acquired Brain Injury Research, Toronto Rehabilitation Institute, University of Toronto, 160-500 University Avenue, Toronto, Ontario
M5G 1V7, Canada. Tel.: +1 416 978 1098; fax: +1 416 946 8570.
5
Tel.: +45 6550 3471; fax: +45 2328 5051.
http://dx.doi.org/10.1016/j.neubiorev.2014.10.024
0149-7634/ 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

685

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods/design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
Data sources and searches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.
Inclusion criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.
Study design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.
Study review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.
Data extraction and quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.
Data synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.
Zero-time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.
Missing data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
Literature search and quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
Studies with baseline assessment up to one month post-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.
Studies with baseline assessment after one month post-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.
Assessment of TBI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.
Methods used for assessing fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.1.
Multi-item scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.2.
Single item assessment of fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.3.
Multiple measures of fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.7.
Overall predictors of fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.8.
The course of fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.9.
The course of fatigue, by injury severity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.10.
Fatigue severity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.11.
Impact of fatigue after TBI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.12.
Associations of fatigue with other clinically important variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.12.1.
Studies with baseline assessment up to one month post-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.12.2.
Studies with baseline assessment after one month post-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.13.
Medications, drugs and alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Factors associated with fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Frequency, severity and course of fatigue in TBI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.
Consequences of fatigue in TBI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.
Medication effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.
Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.
Pitfalls and controversies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Authors contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A.
Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

685
686
686
686
686
686
686
703
703
703
703
703
706
706
706
706
706
706
707
707
707
707
707
707
709
709
709
709
710
710
710
710
711
711
712
712
714
714
714
714
714

Contents
1.
2.

3.

4.

5.

1. Background
Traumatic brain injury (TBI), dened as an alteration in brain
function, or other evidence of brain pathology, caused by an external force (Brain Injury Association of America, 2013), is among
the most serious, disabling neurological disorders in all societies
and expected to rank as the major cause of death and disability
by the year 2020 (World Health Organization, 2002). Over the past
decades, evidence has emerged citing fatigue as a common, longlasting problem after TBI (Belmont et al., 2006; Ponsford et al., 2011;
Middleboe et al., 1992). It is burdensome to patients, and is associated with poor outcomes (Belmont et al., 2006; Ponsford et al.,
2011). In a number of studies, over half of the patients making up
the TBI samples reported fatigues negative effect on social, physical
and cognitive functioning (Ziino and Ponsford, 2006) and participation in everyday activities (Cantor et al., 2008), and role in increased
work-related and other disabilities (McCrimmon and Oddy, 2006).
Estimates of the incidence of fatigue after TBI vary from 21% to
73%, depending on the characteristics of the studied population
(e.g. severity of injury, time since injury, sampling of patients, etc.)
and the method used to identify fatigue (e.g. single item or fatigue
scales) (Belmont et al., 2006; Ponsford et al., 2011; Middleboe et al.,
1992; Borgaro et al., 2005; Lidvall et al., 1974).

The term fatigue has several meanings. It is recognized when


performance of an activity results in diminished capacity for carrying out a function (Chaudhuri and Behan, 2004). Within this,
physiological fatigue refers to the state of general tiredness due
to physical or mental exertion, which can be ameliorated by rest
(Schillings et al., 2007). A state that refers to a weariness unrelated
to previous exertion level, and not ameliorated by rest, is termed
pathological fatigue (Jason et al., 2010). Despite such characterization, fatigue in the TBI population is difcult to elucidate. This is
partly due to the numerous plausible biological causes of fatigue
(i.e. neuroanatomical, functional, psychological/psychiatric, biochemical, endocrine, sleep-related), independently or combined,
through which this symptom can evolve after brain injury (Fig. 1)
(Prins et al., 2006). To date, several narrative reviews have been
published to provide insight into the topic of post-traumatic fatigue
(PTF) (Belmont et al., 2006; Borgaro et al., 2005; Ponsford et al.,
2012; Levine and Greenwald, 2009). Nevertheless, there is still little
known about which specic clinical, behavioral and physiological factors are associated with its occurrence after brain injury;
nor whether fatigue remains the same in its frequency/intensity,
or changes over time. Finally, the overall health burden of this
symptom in the TBI population remains uncertain. Understanding the facets of fatigue in TBI can guide in differential diagnosis

686

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Fig. 1. Pathways to fatigue in traumatic brain injury.


Modied from Chaudhuri and Behan (2004), Finsterer and Mahjoub (2013) and Kluger et al. (2013).

and follow-up treatments. Moreover, identifying the most important contributors to PTF can change the view on the interventions
necessary to deal with this signicant symptom. This systematic
review was performed with the following goals, all with respect to
patients with TBI: (1) to determine the prognostic factors associated with fatigue onset; (2) to describe the course of fatigue; and
(3) to describe the health consequences of fatigue.
2. Methods/design

were excluded. Further, case reports, pediatric studies, dissertations, and articles with no primary data were excluded. For more
information, we refer the reader to the protocol (Mollayeva et al.,
2013a).
2.3. Study design
All experimental intervention and effectiveness studies of longitudinal design and observational cohort- and case control-designed
studies were considered for this review.

2.1. Data sources and searches


2.4. Study review
This review was conducted and reported in compliance with
the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines (Moher et al., 2009). The systematic
review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (Mollayeva et al.,
2013a) on April 25, 2013 (registration number CRD42013004262).
In collaboration with disease experts and a medical information specialist, we developed a comprehensive search strategy for
studying fatigue in TBI (Mollayeva et al., 2013a). All English language peer-reviewed studies with prospective or retrospective data
collection and a longitudinal design, found through PsycINFO, MEDLINE, EMBASE and CINAHL, published since 1806, 1946, 1974 and
1980, respectively, were eligible. Cochrane Database of Systematic
Reviews was also searched for studies published between 2005
and early April 2013. Publications identied from bibliographies
of identied articles and reviews were considered eligible. The
basic search can be found in Supplementary File 1. For the complete search strategy, we refer the reader to the published protocol
(Mollayeva et al., 2013a).
2.2. Inclusion criteria
Peer-reviewed, English language studies that investigated
fatigue in adult patients with a diagnosis of TBI and followed them
for any period were included. Studies that focused on a different
but parallel topic to fatigue (e.g. sleepiness, impaired alertness,
or vigilance) and studies about fatigue after brain injury due to
secondary pathological processes (e.g. edema, intracranial hemorrhages, ischemia/infarction, and systemic intracranial conditions)

In the rst stage of screening, two reviewers (TM and TK or TM


and SM) assessed study titles and abstracts for possible agreement
with the inclusion criteria. In the second stage, each reviewer individually assessed the full text of articles selected in the rst stage
to determine whether they met inclusion criteria. Differences of
opinion were resolved by discussion between reviewers, or by seeking advice from other experts (AC, CS, and JDC). Studies failing to
meet the inclusion criteria were excluded, with reasons listed in
Supplementary le 2.
2.5. Data extraction and quality assessment
The abstracted data included: (1) study characteristics (i.e.
author names, publication year, country, setting, design, sample
size, methods of measuring fatigue, and other variables [e.g. factors], number of participants assessed at each time point, time
between assessments, and time from injury to follow-up); (2)
participant characteristics (i.e. mean age, sex, denition of TBI,
localization of injury, and injury severity); (3) medications used
by or administered to participants; and (4) results (i.e. reported
frequencies of fatigue and other factors, and reported associations
between fatigue and other variables) (Tables 13).
For studies that fullled the inclusion criteria, two reviewers
(TM and TK) independently extracted data into data collection
forms grouped according to study design. The observational studies data were used to address the three research objectives (i.e.
prognostic factors, course of fatigue, and consequences). Randomized control trials (RCTs) were treated as cohorts, and the control

Table 1
Summary of study characteristics, including details on study sample, design, methods and results pertaining to fatigue.
Reference
Country
Sample by

Preliminary ndings from


study of fatigue in individuals
with mod-sev TBI 2 yrs
post-injury; document change
in nature of fatigue over time,
assess contributing factors
Prospective longitudinal
F/U: 6, 12, 1824 mos
post-injury
IC: inpatient rehabilitation;
TBI; 18 yrs at injury;
speak/write/read English;
informed consent
EC: conditions associated with
fatigue

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

n = 51
Attrition: 0
Age: 31 13
Sex: 76% M
TSI: discharge 6 mos
post-injury
IS: GCS motor; PTA length;
degree cranial midline
shift: mod-sev
AT: baseline
t1 : 2.6 1.8 mos
t2 : 12.6 1.2 mos
t3 : 23.2 3.4 mos

Statistical method

rmANOVA: change
over time
Post hoc pairwise t
tests: signicant
variables
p-value = .01

Medications

NR*
*Study #2, same
population: at year 1
11% illicit drug users,
19% classied as
problem substance
users; at year 2 29%
classied as problem
substance users

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

BFS: synthesis of items from


other scales: MAF, FSS, FAI, FIS,
GFS
BFS: higher scores = greater
fatigue
From BFS: scores for Global
Fatigue Index (GFI) of MAF and
FSS: higher scores = greater
fatigue
Scores: t1 , t2 , t3 :
BFS subscales (n = 46*):
Intensity: 36 17; 30 19;
33 18
Activities of daily living:
33 18; 28 18; 32 21.5
Socialization: 24 14; 20 13;
22 15
Mental functioning: 25 12;
20 12; 21 12
General impact: 11 7; 9 6;
10 6
Relieving factors: 16 8;
15 8; 18 6
Aggravating factors: 21 14;
18 11; 21 10
GFI (n = 43*): 23 10; 17 11;
20 11
FSS (n = 45*): 3.4 1.5;
2.9 1.6; 3.2 1.8
*Missing data, unanswered
questions

Score changes:
BFS subscales: NS
GFI: signicant
decrease t1 ct2
(t42 = 5.4; p = .0018;
effect size = .58)
FSS: NS
Notes:
Fatigue total scores had
same pattern of
change: highest at t1 ,
lowest at t2 , slight
increase at t3
GFI at t1 : only score
comparable to other
populations with
signicant fatigue
BFS subscales:
low-mod fatigue; avg
scores below 50% of
max score for each
subscale (exception:
relieving factors)
Associations:
Increased fatigue
t1 t2 , more sleep
problems (PSQI)
decreased and stable
fatigue scores,
decreased PSQI
Increased fatigue,
decreased cognitive
functioning; decreased
fatigue-increased
cognitive functioning;
similar for general
functioning, motor
symptoms

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Bushnik et al. (2008a)


US
Medical center
inpatient rehabilitation

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

687

688

Table 1 (Continued)
Reference
Country
Sample by

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

Statistical method

Medications

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

Compare fatigue reports of


participants 12 mos post-MHI
with those with other injury by
mild trauma; injury, BL
predictors of fatigue
Inception, cohort
F/U: 12 mos
IC: presented directly to ED w/I
24 h of injury; 18 yoa;
GCS 13; did not meet criteria
for activation of adult trauma
team; discharged directly from
ED; competency for informed
consent; mini-mental state
examination 18; able to
describe essential elements of
study
EC: transfer from other
hospital; non-English
speaking; being incarcerated;
medical evaluation resulting in
admission; state of PTA at
recruitment; LOC 30 min;
LOC not attributable to trauma

n = 359 (w/ 12 mos data)


3 groups: 1: HI w/ PTA
&/OR LOC; 2: HI only; 3:
other injury (n1 , n2 , n3 )
n1 = 58, n2 = 173, n3 = 128
* No medical info wrt
occurrence of brain injury
Attrition: 31.9%
Age: NR
Sex:
1: 44.8% M
2: 41% M
3: 43.7% M
TSI: 12 mos
IS: mild
AT: BL/n = 504
t1 : 1 mos (NR)
t2 : 3 mos (NR)
t3 : 12 mos/n = 359

2-tailed tests of
signicance, = .05
Chi-square: group
differences
(categorical)
Univariate ANOVA:
group differences
(continuous)
Hierarchical linear
regression: variable
associations

NR

MOS SF-36 Vitality subscale


Low scores on vitality subscale
indicate more fatigue
Mean SF-36 vitality subscale
scores at BL:
1: 52.8 9.5
2: 50.4 10.5
3: 53.4 8.6
*HI only group (2): greater
fatigue severity (p = .026)
t3 :
1: 52.3 12.2
2: 49.6 11.8
3: 53.0 10.4
*Signicant differences
between groups (F2,356 = 3.77,
p = .024, partial 2 = .02)

Pair-wise comparisons:
1: Lower mean score at
12 mos than other
injury group (p = .027)
2: Comparison NS

Driver and Ede (2009)


US
Community

Changes in mood in response


to 8 wk physical activity (PA)
intervention
Stratied random sampling
F/U: 8 wks
IC: >level 6 Ranchos Los
Amigos Scale of Cognitive
Functioning; TBI > 1 yr prior;
outpatients at rehabilitation
center
EC: NR

n = 18 TBI
2 groups: 1: PA; 2: control
Attrition: 0
Age: 37.7 2.3
Sex: NR
TSI:
1:40.8 14.7 mos
2:36.3 14.2 mos
IS:
Each group: 6 w/ left-sided
lesion proximity to frontal
pole; 1: 2 w/ damage to left
basal ganglia; 2:
2 w/damage to right
parietal occipital lobe
AT: BL
t1 : 8 wks

ANOVA: between,
within group
differences
Effect size: total
variance accounted for
by independent
variable

1: 5 taking SSRIs for


duration of program*
2: 6 taking SSRIs for
duration of program*
*Not as part of study

Fatigue denition: NR
POMS fatigue-inertia subscale
Mean fatigue subscale scores at
BL:
1: 1.4 1.1
2: 1.2 .6
Mean fatigue subscale scores at
t1 :
1: .5 .6
2: 1.3 .6
Effect size:
1: 1.00
2: .08

Within group
differences (BL t1 ):
1: signicant (F = 4.7,
p < .05)
2: NS
Between group
differences:
Fatigue NR; signicant
wrt total POMS score
(F = 5.7, p < .05)

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

De Leon et al. (2009)


US
Level II community
hospital ED

Whether Tai Chi would have


immediate effect on mood
states in TBI group;
improvement of perceived
physical, emotional functions,
self-esteem, social functioning,
health over time
Within-group, between-group
with control
F/U: 6 wks
IC: mild, mod, severe TBI on
basis of
retrograde/anterograde
amnesia, PTA, and/or LOC with
associated outcomes
EC: NR

n = 18
2 groups: 1: Tai Chi (9); 2:
control, waiting list for Tai
Chi (9)
Attrition: 0
Age:
F: 40.2 12.5, M:
51.2 8.7, Sex: 50% M
TSI: mean = 8.7 yrs
IS: NR
AT: Before
t1 : After (6 wks)

t-tests: within,
between group
differences at time
points
ANOVA: within,
between group
differences over time
periods

NR

Fatigue denition: NR
MOS SF-36 Vitality subscale
VAMS Tired mood state scale
(Tai Chi group only)
SF-36 Vitality: Before:
1: 47.1 18.2
2: 47.5 20.2
After:
1: 40.7 22.3
2: 38.8 4.4
VAMS Tired: Before:
1: 54.4 6.0
After:
1: 52.5 5.8

SF-36 Vitality: Before:


1 vs. 2: NS (t = 0.04)
After:
1 vs. 2: NS (t = 0.25)
VAMS:
1: Fatigue NS (t = 1.10)

Hou et al. (2012)


UK
ED

Optimal early predictors for


post-concussional syndrome
(PCS) following mild TBI
(mTBI); cognitive, emotional,
behavioral, social perpetuating
factors in development of PCS
Prospective cohort
F/U: 3, 6 mos
IC: 1860 yoa, mTBI
EC: multi-trauma requiring
hospitalization, major
neurological/psychiatric
disorders

n = 126
Attrition: 25%
Age: 38.3 14.1
Sex: 63% M
TSI: 2 wks
IS: GCS: mild
AT: BL: 2 wks
post-injury/n = 126
t1 : 3 mos/n = 107
t2 : 6 mos/n = 107

t-tests, chi-square:
demographic, clinical
characteristics
Individual regression
analyses: cognitive,
emotional, behavioral
variables as covariates
with gender/age, PCS
outcome as dependent
variable
Logistic regressions
(LR): for signicant
variables from
individual regression
analyses
Stepwise backward LR:
derive models for 3, 6
mos
HosmerLemeshow
goodness of t
statistic: t of model
assessment

NR

Fatigue denition: NR
RPQ (including fatigue, sleep
disturbance items)
RPQ: fatigue, sleep disturbance
most commonly reported
symptoms are 3 and 6 mos
port-mTBI
Fatigue frequencies: NR
From bar graph:
Fatigue at 3 mos 33%, at 6
mos 28%
Sleep disturbance at 3 mos
27%, at 6 mos 24%

NR

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Gemmell and Leathem


(2006)
NZ
Psychology clinic, head
injury society

689

690

Table 1 (Continued)

Hutchinson et al.
(2009)
CA
University sports

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Determine whether athletes


with concussion and those
with musculoskeletal injuries
(MSI) differed in emotional
responses post-injury;
differences in BL emotional
status in 3 groups: mTBI, MSI,
active control (CTL) (control for
premorbid emotional
disturbance)
Prospective longitudinal cohort
F/U: 1,2,3 d
IC: university athletes in sport
with risk of concussion
EC: self-reported at
neuropsychological
assessment: >5 concussions;
learning disability; psychiatric
disorder

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

n = 53
3 groups: 1: mTBI (20); 2:
MSI (14); 3: CTL (19)
Attrition: 0
Age:
1: 20.1 1.8
2: 19.2 2.3
3: 21.6 1.6
Sex:
1: 60% M
2: 86% M
3: 47% M
TSI:
1: 96 hrs
IS:
1: concussion (mTBI) by
team physicians, therapists
AT: BL
t1 , t2 , t3 : 3 d*
*Nonconsecutive over 2
wks

Statistical method

Descriptive:
demographic variables,
mood scales
Cronbach alpha: scale
reliability
ANOVA: group
differences on POMS
subscales at BL;
physical characteristics
StudentNewmanKeuls
multiple-range test
(.05): F/U means
TukeyKramer
correction for type I
error

Medications

NR

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

Fatigue denition: NR
POMS fatigue subscale:
Reliability: .863
Scores:
Main effects: NS
Signicant interacting effect
for fatigue (F(6, 150), 10.11;
p < .001)
Difference at t1 : signicant for
1 (increase) vs. 2, 3

POMS fatigue subscale:


Difference BL t1 :
signicant for 1
Notes:
1: signicantly greater
fatigue, lack of energy
(POMS vigor subscale)
post-injury

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Reference
Country
Sample by

Jha et al. (2008)


US
Hospital

n = 51 (46*)
2 groups: 1: modanil rst
(27 (22*)); 2: placebo rst
(24)
*5 participants in group 1
withdrew
Demographic/clinical
characteristics reported for
n = 51**
**NS imbalances to affect
trial results
Attrition: 0
Age: 38.3 12.2
Sex: 69% M
TSI: 5.8 5.0 yrs
IS: GCS: mild (25.5%), mod
(23.5%), sev (51%)
AT (n = 46): BL
t1 : wk 4
t2 : wk 10

t-tests, chi-square:
continuous and
categorical,
respectively: BL
differences in
demographic/clinical
characteristics
between groups
Paired t-test: crude tx
effects
2-sample t-tests:
within group tx effects
Linear mixed-effects
regression: 4-wk
change in each of 2
periods for all
participants
Secondary analyses: tx
effects on secondary
end-points and at 10
wks

Modanil (400 mg)


Concurrent medication
use in exclusion criteria

MFIS
FSS
SF-12 (fatigue item)*
*Fatigue NR separately
FSS, MFIS high scores
Scores: BL; t1 ; t2 :
FSS:
1: 45.2 11.8; 39.4 15.6;
37.13 18.33
2: 44.46 12.17; 37.7 12.55;
36.91 14.08
MFIS:
1: 46.56 19.28;
38.65 16.09; 35.63 20
2: 47.17 15.53;
36.45 15.03; 33.55 18.16
Group medication switch
modanil to placebo, vice
versa:
FSS:
1 (placebo): 35.92 16.82;
33.74 16.16; 30.95 16.25
2 (modanil): 38.17 15.23;
31.38 10.66; 28.90 14.03
MFIS:
1: 36.27 17.67;
37.74 17.51; 31.20 19.44
2: 39.73 20.82;
28.91 19.06; 28.27 16.06

1 vs. 2:
modanil-placebo
scores:
Change wk 4-BL (p
value):
FSS: 2.33 12.96 (.54)
MFIS: 5.68 14.79 (.21)
Wk 10-BL:
FSS: .44 15.31 (.92)
MFIS: 4.03 16.93 (.43)
Group medication
switch modanil to
placebo, vice versa:
Wk 4-BL:
FSS: 2.55 11.07 (.45)
MFIS: 10.9 15.93
(.03)
Wk 10-BL:
FSS: 3.70 14.60 (.43)
MFIS: 8.07 16.61
(.14)
Notes:
Participants suggested
fatigue measures used
in study do not
accurately reect
fatigue experienced by
persons with TBI

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Efcacy of modanil for


treating fatigue, excessive
daytime sleepiness (EDS) in
persons with TBI; hypothesis:
modanil more efcacious
than placebo; outcomes wrt
cognitive function, health
related quality of life
Single-center randomized
blinded, placebo-controlled
cross-over
F/U: 4, 10 wks
IC: 1-year post-TBI; 1865 yoa;
received inpatient
rehabilitation at single model
system of care
EC: presence of
neurologic/neuropsychiatric
diagnosis; diagnosis by history
of other likely causes of EDS;
concurrent medication use
and/or clinically signicant
systemic disease that might
cause fatigue/diminished
arousal; epilepsy;
cardiovascular
disease/hypertension requiring
medical tx; history of severe
renal/hepatic impairment;
signicant
psychiatric/behavioral
disturbance; non-English
speaking; pregnant females/of
childbearing potential

691

692

Table 1 (Continued)
Reference
Country
Sample by

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

Statistical method

Medications

Results

Score differences over


time
Notes

Kaiser et al. (2010)


CH
Hospital neurology
department

Effect of modanil on
posttraumatic EDS and fatigue
Prospective, double-blind,
randomized,
placebo-controlled, pilot
F/U: 6 wks
IC*: presence of
fatigue/EDS/both since injury
* Patients from earlier study
(Baumann et al., 2007)
admitted for closed mild-sev
TBI to surgical intensive care
unit
EC: patients with neurologic,
psychiatric, other disorders,
medications that may cause
SWD; signicant SWD other
than posttraumatic vigilance
impairment at BL; chronic
sleep deprivation

n = 20
2 groups: 1: modanil
(10); 2: placebo (10)
Attrition: 0
Age:
1: 37 9
2: 43 19
Sex:
1: 80% M
2: 90% M
TSI:
1: 1.8 .9 yrs
2: 2 1.2 yrs
IS: GCS: mild-sev
AT: BL
t1 : 6 wks

Pearson, Spearman
correlation analyses;
two-tailed t-tests;
MannWhitney U
tests; multivariate
regression analyses

Modanil
(100200 mg)
Interfering medication
use as part of exclusion
criteria; caffeine, other
drugs not allowed
during course of study

FSS > 4
BL: frequency of fatigue
diagnosis:
1: .8
2: .8
BL: FSS
1: 5 1.4
2: 4.6 .8

t1 (p = 0.07):
1: .8 1
2: .0 .6
Notes:
Overall subjective
estimation of vigilance
impairment
amelioration:
1: much better (0%);
better (30%);
somewhat better
(30%); unchanged
(30%); worse (10%)
2: much better (10%);
better (10%);
somewhat better
(10%); unchanged
(70%); worse (0%)

Kempf et al. (2010)


CH
University neurology
department

Prevalence, characteristics of
post-traumatic sleep-wake
disorders (SWD)
Prospective, longitudinal,
clinical
F/U: 3 yrs
IC: acute, rst TBI; no SWD,
psychiatric/neurological
disorders prior; admitted
immediately after injury
EC: NR

n = 51
*Studied at 6 mos wrt SWD
(n = 65, Baumann et al.,
2007)
Attrition: 21.5%
Age: 40 16
Sex: 84% M
TSI: 3 yrs
IS: GCS: mild (42%), mod
(22%), sev (38%)
AT:
t1 : 6 mos/n = 65 (Baumann
et al., 2007)
t2 : 3 yrs/n = 51 (this study)

Correlation analyses
t-tests: parametric
MannWhitney U
tests: non-parametric
One-way ANOVA:
group differences
McNemar test: 
repeated dichotomous
measures

3 (antiepileptic drugs),
1 (zopidem for sleep)

Fatigue Severity Scale (FSS)


FSS 4
t1 : 17%
t2 : 35%
51%: fatigue associated
symptoms (daytime tiredness,
lack of energy, exhaustion)
since injury

NR

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Fatigue denition
Frequencies, scores

Whether patients with


cognitive/behavioral
impairment after brain injury
would benet from donepezil;
improvement would concern
one cognitive domain more
than others
Intervention
F/U: 3 mos
IC: outpatients of neurology
clinic; informed consent;
history of mod-sev TBI for 6
mos
EC: history of previous CNS
injury/disease; ongoing
drug/alcohol abuse; severe
speech/language disorders;
unstable psychiatric disorders;
compliance difculties; use of
AchE inhibitors

n = 10
Attrition: 33.3%
Age: 43 8
Sex: 60% M
TSI: 42 33 mos
IS: PTA: 8 10 d
AT: BL
t1 : 3 mos

Non-parametric
Wilcoxon: statistical
signicance of changes
by donepezil therapy

Donepezil (510 mg)


Ongoing alcohol/drug
abuse, current use of
acetylcholine AchE
inhibitors as part of
exclusion criteria

Fatigue disability measured by


29-item fatigue scale: severity,
specicity, psychological
consequences, effects of
sleep/rest on fatigue
Fatigue scale high score
severe symptoms
BL: mean score: 132.6 27.3

t1 : mean score:
126.1 32.3 (p = .92,
Z = .10)
Notes:
Subjects self-report
post-tx: 80% reported
medication-related
improvement in 1
cognitive/affectivebehavioral domain
40% wrt fatigue
dominating
improvement
Subjective fatigue
improvement did not
correlated with
decrease in fatigue
score only 2 patients
showed notable
decrease in score
2 patients that did not
report subjective
improvement had
notable decrease in
fatigue score
Discrepancies may be
result of varying
denitions of fatigue

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Khateb et al. (2005)


CH
University hospital
neurology clinic

693

Reference
Country
Sample by

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

Statistical method

Medications

694

Table 1 (Continued)
Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

Establish whether PCS were


experienced only after injury,
or whether the patient was
already suffering them before
the trauma; explore the
etiology of the PCS
Prospective longitudinal cohort
F/U: 2, 6, 14, 30, 90 d
post-injury
IC: cerebral concussion (CC);
PTA; hospital emergency
admission 24 and
examination on the second d
after injury; 15; able to
cooperate; not required
surgical tx; good knowledge of
Swedish
EC: with mental or comatic
illness; alcoholics and drug
addicts

n = 100 CC
Attrition: 0
Stats:
PTA (min):
<1: 17%; 15: 34%; 645:
29%; >45: 20%
Age: 33 y (av)
Sex: 69% M
TSI: <24 h
IS: PTA: mild-sev
AT: post-injury:
t1 : 2 d
t2 : 6 d
t3 : 14 d
t4 : 30 d
t5 : 90 d

Descriptive: summary
of all variables
Chi-square: discrete
variable differences
between PCS- and
C-groups
Phi coefcients for
symptoms clustering

NR

PCS-symptom questionnaire
Fatigue frequencies by PCS list:
t1 : 7%
t2 : 5%
t3 : 10%
t4 : 8%
t5 : 10%
Fatigue frequency at any given
time point 12%

Clusters unstable,
varying in extent and
character across time
points:
t1 :
1: headaches,
dizziness, fatigue,
concentration
impairment (largest)
2: memory
impairment, sensitivity
to light
t2 :
1: headaches, dizziness
2: headaches, fatigue
3: dizziness, fatigue
t3 :
1: headaches, fatigue,
dizziness
2: anxiety,
concentration
impairment
t4 :
1: anxiety, fatigue,
headaches
t5 :
1: headaches, anxiety
2: dizziness,
concentration
impairment
3: fatigue, anxiety
4: fatigue, headaches

Lundin et al. (2006)


SE
EDs

Report character, frequency,


course of persisting symptoms
and their relation to disability
through effects on daily
activities in mTBI patients
Prospective cohort
F/U: 1, 7, 14 d, 3 mos
post-injury
IC: blunt head trauma; LOC
and/or PTA; hospital admission
24 h post-injury; GCS 1415
at ED assessment; 1565 yrs
EC: LOC 30 min; PTA 24 h;
other signicant physical
injury/major neurological
disorder

n = 102 mTBI; 35 controls


Attrition: 16.4%
Stats for TBI:
Age: 37.3
Sex: 58% M
TSI: 24 h
IS: GCS: mild
AT: post-injury:
t1 : 1 d
t2 : 7 d
t3 : 14 d
t4 : 3 mos

MannWhitney U
tests: non-normal
continuous data
Chi-square/Fishers
exact tests: categorical
data
Bonferroni adjustment:
multiple comparisons
p = .05
Multilevel logistic
regression:
relationship between
stable patient
characteristics and
multiple
measurements

NR

Rivermead Post-Concussional
Questionnaire (RPQ);
Rivermead Head Injury F/U
Questionnaire (RHFUQ*): both
feature fatigue item
*Administered only at t4
RPQ score classication:
symptom resolution (1); mild
(2); moderate (3); severe (4)
RPQ fatigue frequencies:
t1: mTBI: 66.8%
(calculated from symptom load
bar graph)
t4 :
mTBI: 21%
Control: 11%

t1 t4 : signicant
decrease

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Lidvall et al. (1974)


SE
EDs

Investigate type of
psychosocial difculties
associated with head injury
(HI) at 1mos and 12 mos
post-injury; whether degree of
psychosocial impairment
relate to HI severity
Prospective cohort
F/U: 1 mo, 12 mos post-injury
IC: HI; LOC and PTA > 1 h, or obj
evidence of cerebral trauma;
trauma required hospital
admission; 1560 yrs
EC: previous CNS insult or
involvement (e.g. epilepsy, HI,
alcoholism, mental
retardation); psychiatric
disorder

n = 102 HI; 102 controls


Attrition: 0
Stats for HI:
Age: 26.33
Sex: *primarily single M
TSI: 24 h
IS: GCS: mild 58%; mod
11%; sev 28%
AT: post-injury:
t1 : 1 mos
t2 : 12 mos

2 by 2 Chi-square tests
MannWhitney U tests
comparison of change
with time
KruskalWallis
distribution-free
analysis of variance;
post hoc comparison
according to Tukeys
method for
unequal-sized groups

NR
Prior history of
epilepsy, alcoholism,
mental retardation as
part of exclusion
criteria

The Head Injury Symptom


Checklist
Frequencies:
t1 : 74%
t2 : 47%

A signicant reduction
in the number of
fatigue endorsed from
1 mo to 12 mos
post-injury (p < .001)
Signicant difference
between cases and
controls at 1 mo but
not at 12 mos
post-injury

Meares et al. (2011)


AU
Level 1 trauma hospital

Investigate course of PCS,


PCS-like symptoms;
relationship pre-injury,
injury-related, post-injury
factors to PCS development
Prospective
F/U: 3 mos since hospitalized
IC: trauma center patient;
traumatic nonbrain
injury/mTBI; hospital
admission 24 h post-injury;
rst assessment 14 d
post-injury; 1865 yrs;
IQ 70; adequate
understanding of English
EC: mod-sev TBI/intracranial
lesion; self-harm physical
injury; psychotic; history of
cognitive impairment;
medically unstable;
interstate/overseas visitor;
pregnant
EC at 3 mos: >5 mos
post-injury at F/U; inadequate
effort on testing by failure on
memory test

n = 62 mTBI; 58 TC
Attrition: 0
Stats for TBI:
Age: 35.7 14.5
Sex: 67.7% M
TSI: 4.8 3.1 d
IS: GCS: mild
AT: post-injury:
t1 : 14 d
t2 : 3 mos

MannWhitney U
tests: non-normal
continuous data
Chi-square/Fishers
exact tests: categorical
data
Bonferroni adjustment:
multiple comparisons
p = .05
Multilevel logistic
regression:
relationship between
stable patient
characteristics and
multiple
measurements

Opiate administration
at t1 /t2 (n):
mTBI (37/62)
Control (37/58)
Marijuana use:
mTBI: 24.4%
Control: 19%
AUDIT alcohol screen
mTBI = 6.4 6.8; 8
hazardous alcohol use
indicator
At least 1 subs use
disorder: 12.5%

PCSC: adapted version


including fatigue symptom
PCSC symptoms: 5-point scale;
clinically signicant if scored
as 3, indicating often for
frequency
Fatigue symptom frequencies
(3 on PCSC):
None:
mTBI: 40.3%
Control: 20.7%
Present at t1 and t2 :
mTBI: 21%
Control: 32.8%
Present at t2 only:
mTBI: 14.5%
Control: 15.5%

Fatigue symptom
frequencies (3 on
PCSC):
Absent at t2 :
mTBI: 24.2%
Control: 31%
Difference in frequency
between mTBI and
controls in
presence/absence at
t1 /t2 of fatigue: NS

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

McLean et al. (1993)


US
Medical Center

695

696

Table 1 (Continued)
Reference
Country
Sample by

Investigate PCS symptoms,


effects of sociodemographic
factors and expectations on
symptoms
Prospective controlled cohort
F/U: 3 mos, 1 yr post-injury
IC: admitted for head trauma
(w/ LOC) evaluation/tx; LOC
15 min
EC: history of drug/alcohol
abuse, epilepsy, signicant
psychiatric/neurological
disorder; previous concussion;
concussion-related seizures;
focal neurological signs,
abnormal neurological status
at admission; other major
injury leading to
hospitalization; hospital stay
>1 wk

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

n = 217 concussion; 221


minor injury controls
Attrition: 7.8%; 4%
Stats for concussion group:
Age:
M: 33 13
F: 38 14
Sex: 66% M
TSI: 714 d
IS: NR
AT: post-injury
BL/n = 217
t1 : 3 mos/n = 200
t2 : 1 yr/n = 192

Statistical method

Chi-square w/ Yates
correction:
between-group
symptom comparison
2-Sided t test: VAS
score comparison
Multiple regression:
headache, cognitive
dysfunction VAS scores
as dependent,
demographic
characteristics as
independent

Medications

NR
Prior history of alcohol
abuse, drug abuse as
part of exclusion
criteria
Alcohol intolerance
31% at 3 mos
post-injury, 31% at 1 yr
post-injury

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

VAS fatigue item


Fatigue: score 50 on VAS
VAS fatigue scores:
BL: between-group differences
NS
t1 (p = .002):
Concussion: 50 28
Control: 41 29
t2 (p = .08):
Concussion: 50 30
Control: 44 28
VAS fatigue scores in
participants:
Married 1 yr (p = .17):
Concussion: 51 30
Control: 45 27
Unmarried 1 yr (p = .11):
Concussion: 49 30
Control: 42 29
With 1 yr high education
(p = .01):
Concussion: 52 30
Control: 42 27
With 1 yr low education
(p = .72):
Concussion: 47 30
Control: 45 29

NR

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Mickeviciene et al.
(2004)
LT
Hospital emergency
ward

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Norrie et al. (2010)


NZ
Hospital

n = 159
Attrition: 14.1%; 10.6%
Age: 35.9 15.6
Sex: 64% M
TSI: 1 wk; 10 d
IS: GCS: mild
AT: post-injury
BL/n = 263
t1 : 3 mos/n = 159
t2 : 6 mos/n = 159

Pearson correlation
coefcients: primary
dependent variables,
fatigue prevalence,
severity, energy,
depression, anxiety
associations
One-way ANOVA:
fatigue prevalence,
severity, energy change
Hierarchical
regression: t2 fatigue
severity with t1 fatigue
severity, depression,
anxiety
Receiver operating
characteristic curve:
sensitivity, specicity
of fatigue item on RPQ
in distinguishing those
w/ and w/o
pathological fatigue by
FSS at t2

NR
Regular intake of
psychoactive drugs,
history of drug abuse
as part of exclusion
criteria

FSS (severity)
RPQ fatigue item (prevalence)
SF-36v2 Vitality Subcale
FSS cut-off = 3.7
RPQ fatigue prevalence
frequency of item rating 2
(max 4)
SF-36v2 Vitality high
score = low fatigue (max 100)
Frequencies:
BL:
FSS: 54.1%
RPQ: 67.3%
t1 :
FSS: 35.8%
RPQ: 29.6%
t2 :
FSS: 34%
RPQ: 26.4%
Scores:
BL:
FSS: 3.99 1.53
RPQ: 2.09 1.24
SF-36v2 Vitality: 46.57 24.72
t1 :
FSS: 3.3 1.4
RPQ: 1.0 1.1
SF-36v2 Vitality: 60.2 19.7
t2 :
FSS: 3.2 1.4
RPQ: .96 1.1
SF-36v2 Vitality: 62.1 20.2
Correlations between
measures at time points
(p .000, unless otherwise
indicated):
FSS BL w/: FSS t1 : .53; FSS t2 :
.49; RPQ BL: .57; RPQ t1 : .30;
RPQ t2 : .38; SF-36v2 BL: NS;
SF-36v2 t1 : .42; SF-36v2 t2 : .4
FSS t1 w/: FSS t2 : .76; RPQ BL:
.16 (p .05); RPQ t1 : .45; RPQ
t2 : .5; SF-36v2 BL: NS; SF-36v2
t1 : .66; SF-36v2 t2 : .39
FSS t2 w/: RPQ BL: .2 (p .05);
RPQ t1 : .4; RPQ t2 : .62; SF-36v2
BL: NS; SF-36v2 t1 : .56;
SF-36v2 t2 : .59

Within-subject effects
BL-t2 (p < .0005):
FSS: F2,157 = 23.60;
2 = .23
RPQ: F2,157 = 60.556;
2 = .44
SF-36v2 Vitality:
F2,157 = 17.573; 2 = .18
Notes:
RPQ fatigue item
unsatisfactory for
prediction of
pathological fatigue
post-mTBI

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Prevalence, severity,
predictors, covariates of fatigue
in persons with mTBI
Longitudinal prospective
F/U: 3, 6 mos post-injury
IC: patients presenting to
hospital with mild closed head
injury; GCS 1315;
LOC < 20 min; PTA < 24 h
EC: abnormal CT scan; regular
admission of psychoactive
drugs/history drug abuse;
central neurological
disorder/psychiatric condition;
skull/facial fractures/multiple
trauma/other major trauma

697

Reference
Country
Sample by

698

Table 1 (Continued)
Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

Statistical method

Medications

Results

Fatigue denition
Frequencies, scores

Report on post-concussive
symptoms and associated
cognitive, psychological,
functional outcomes in persons
with uncomplicated mTBI
Prospective
F/U: 1 wk, 3 mos post-injury
IC: admitted to emergency and
trauma center;
trauma/accelerationdeceleration movement to
head w/ LOC < 30 min,
PTA < 24 h, GCS 1315 in last
24 h; 18 yrs; English-speaking
EC: general anesthesia after
injury; breath alcohol
>.05 mg/L at recruitment;
under inuence of illicit drug
at injury; focal neurological
signs/seizures and/or
intracerebral abrnormalities
based on CT; dominant upper
limb injury disabling from use
of computer mouse; spinal
precautions, cannot sit upright;
previous cognitive impairment,
neurological illness, major
alcohol/drug abuse, other
psychiatric impairment
affecting daily functioning; not
available for F/U

n = 123 (90 analyzed), 100


TC (TC) (80 analyzed)
Attrition: 9.8%; 18.9%
Stats for mTBI:
Age: 35.0 13.1
Sex: 74% M
TSI: 24 hrs at BL
IS: GCS: mild; PTA:
103 191 min (n = 118);
LOC: 61.4 110 s (n = 111)
AT:
BL: ED/n = 123
t1 : 1 wk/n = 90
t2 : 3 mos/n = 90

qq, box plots;


KolmogorovSmirnov;
ShapiroWilk:
distribution normality
Univariate;
multivariate;
repeated-measures:
between-group score
comparison
Chi-square: categorical
variables, not normally
distributed
MannWhitney U:
continuous variables,
not normally
distributed
Wilcoxon Signed
Ranks; Friedman:
within-subject changes
over time
p = .05

Narcotic analgesics by
self-report:
BL: 62.6% mTBI, 44% TC
t1 : 18.2% mTBI, 21.1%
TC
t2 : 2.2% mTBI, 2.5% TC

NA

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Ponsford et al. (2012)


AU
Emergency and trauma
center

RPQ BL w/: RPQ t1 : .24; RPQ t2 :


.34; SF-36v2 BL: NS; SF-36v2
t1 : .2 (p .05); SF-36v2 t2 :
.2 (p .05)
RPQ t1 w/: RPQ t2 : .6; SF-36v2
BL: NS; SF-36v2 t1 ; .57;
SF-36v2 t2 : .51
RPQ t2 w/: SF-36v2 BL: NS;
SF-36v2 t1 : .52; SF-36v2 t2 :
.7
SF-36v2 BL w/: SF-36 t1 /t2 : NS
SF-36v2 t1 w/: SFv2 t2 : .69
Sensitivity/specicity of RPQ
fatigue item (i.e. discriminating
symptom at t2 on FSS):
BL, t1 , t2 : sensitivity = .74, .42,
.55; specicity = .63, .24, .12
PCS post-concussive
symptoms checklist with
fatigue item
SF-36 Vitality (scores NR
separately)
Fatigue frequencies by PCS:
BL (p < .001):
mTBI: 73.3%
TC: 47.5%
t1 (p = .019):
mTBI: 61.1%
TC: 42.5%
t2 (p = .424):
mTBI: 37.1%
TC: 22.5%
SF-36 Vitality:
TC signicantly higher median
pre-injury, t1 score (z = 3.11,
p = .002; z = 2.96, p = .007)
t2 : mTBI had lower mean score
vs. TC (z = 2.33, p = .020)

Score differences over


time
Notes

Schoenberger et al.
(2001)
US
Tx seekers/patients of
neurologists/rehabilitation
clinics

n = 12; 2 groups: 1:
immediate tx; 2: waitlist
control
Attrition: 0
Age: 2153
Sex: 16.7% M
TSI: 36 mos21 yrs; mean:
7.7 yrs
IS: mild (75%); mod sev
(25%); PTA (self-report):
41.7%; LOC: 127 d
AT:
t1 : 1: pre-tx; 2: BL
t2 : 1: post-tx; 2: pre-tx
t3 : 1: 3 mos post-tx; 2:
post-tx
t4 : 1: NA; 2: 3 mos post-tx

ANCOVAs:
between-group
ANOVAs:
within-group; changes
over time
p = .05 (not adjusted for
multiple tests)

Change in medication
with tx: NA (33.3%);
eliminated (25%);
decreased (16.7%); no
change (25%)

MFI general fatigue, physical


fatigue, mental fatigue,
reduced motivation, reduced
activity
Fatigue by MFI: 5-point
severity scale/item; subscale
scores 420 higher score,
greater severity
Stratied MFI scores by group
and time point:
1: t1 , t2 : total (74.8 20.4,
48.5 20.9); general
(17.2 4.0, 9.8 4.8); physical
(16 6.2, 10 3.5); mental
(17.2 3.3, 10.3 6.3); reduced
activity (14.3 5.6, 11.3 5.4);
reduced motivation (10.2 4.8,
7 2.8)
2: t1 , t2 : total (61.5 18.9,
61.3 20.6); general
(14.8 4.2, 14 4.6); physical
(10.5 4.5, 10.8 5.3); mental
(15.5 3.8, 15.7 3.5); reduced
activity (10.7 4.7, 10.8 5.3);
reduced motivation (10 3.9,
10 4.9)
Stratied MFI scores by
time-point:
Pre-tx: total (68.1 20.8);
general (15.6 4.4); physical
(13.4 6.2); mental
(16.4 3.3); reduced activity
(12.6 5.4); reduced
motivation (10.1 4.6)
Post-tx: total (50.1 19.0);
general (11.2 4.8); physical
(9.8 3.3); mental (11 4.9);
reduced activity (10.1 4.8);
reduced motivation (8 3.6)
F/U: total (47.3 20.0); general
(10.5 5.1); physical
(9.3 5.0); mental (10.7 4.4);
reduced activity (8 4.2);
reduced motivation (7.7 4.4)

Between-group
comparison of MFI
scores: total (F = 3.68,
p < .1); general (F = 8.04,
p < .05); physical
(F = 2.88); mental
(F = 9.10, p < .05);
reduced activity
(F = .24); reduced
motivation (F = 1.99)
Changes in MFI scores:
total (F = 8.43, p < .01);
general (F = 6.5, p < .01);
physical (F = 4.02,
p < .05); mental
(F = 14.68, p < .001);
reduced activity
(F = 3.48, p < .1);
reduced motivation
(F = 2.72, p < .1)
Signicant differences
(p < .05):
Pre-post, F/U: total,
general, mental

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Preliminary study of efcacy of


FNS in persons with TBI
Preliminary experimental
randomized
F/U: post-tx: 3 mos
IC: mild-mod sev closed head
injury; informed consent
EC: penetrating head injury;
substance abuse/psychotic
diagnosis pre-injury; seizure
pre-/post-injury;
pregnant/trying to get
pregnant

699

700

Table 1 (Continued)
Reference
Country
Sample by

Study cognitive recovery 312


mos post-TBI; use of
neuropsychological tests to
predict functional outcome
Prospective
F/U: post-injury: 3, 12 mos
IC: admission to trauma center
with acute TBI; 1655 yrs;
24 h post-injury; uency in
Norwegian
EC: earlier neurological
disorder; spinal cord injuries
w/ current TBI; severe
psychiatric
disorders/substance abuse

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

n = 115; mild n = 40; mod


n = 34; sev n = 41
Attrition: 7.8%
Age (at injury): mild:
35.9 11.4; mod:
33.5 10.8; sev:
28.5 10.4
Sex:
mild: 63% M; mod: 74% M;
sev: 71% M
TSI: 24 h (at recruitment)
IS:
Mild: GCS (14.7 .6); PTA
(.08 (range 01) d);
AIShead 3 (23%); ISS 15
(15%); intracranial
pathology by CT/MRI (28%)
Mod: GCS (10.8 1.3); PTA
(5.25 (range 030) d);
AIShead 3 (79%); ISS 15
(68%); intracranial
pathology by CT/MRI (85%)
Sev: GCS (5.5 1.8); PTA
(35.83 (range 0128) d);
AIShead 3 (100%); ISS 15
(98%); intracranial
pathology by CT/MRI
(100%)
AT: post-injury
t1 : 3 mos/n = 115
t2 : 12 mos/n = 106

Statistical method

Parametric statistics
(Pearson) chi square:
categorical variables
ANOVA:
between-severity
group comparison
Multiple regression:
demographic, IS
predictors
Principal components
analysis w/ varimax
rotation:
neuropsychological
variable assessment at
t2 (to reduce predictors
in multiple regression)
Bonferroni corrections:
signicant tests w/
multiple comparisons
Post hoc: cognitive
functioning in mild TBI
p < .05 (2-tailed)

Medications

At t2 , drug/alcohol use
by Alcohol use
disorders identication
test
Alcohol >once/mo
(chi(8) = 24.1, p < .01):
mild, mod (48%); sev
(27%)
n = 106
Alcohol and/or drugs
23/wk: 13%;
4/wk: 7%

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

FSS: avg 17 across 9 items


total 07 (higher score, more
fatigue)
t2 , between-severity groups:
NS; mean score: 4 1.8
At t1 , fatigue (along w/ 2/3
cognitive components, PTA,
intracranial pathology) as a
predictor of Glasgow Outcome
Scale-Extended (GOSE) at t2 :
signicant (R2 = .61, p < .001)
Fatigue, by FSS, as predictor of
GOSE at t2 : B = .13; SE = .04;
= .25 (p < .001)
Signicant correlations at t2 :
fatigue w/ (n = 96):
Memory/speed: .38 (p < .001)
Verbal/reasoning: .2 (p < .01)
GOSE: .39 (p < .001)

NR
Notes:
Less fatigue predicts
better outcome
No effects of sex,
education, TBI severity
on fatigue at t2

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Sigurdardottir et al.
(2009)
NO
Level I trauma center

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

Study self-reported fatigue


incidence pre- and post-mTBI;
compare incidences at time
points to controls; examine
association between fatigue
and APOE 4 genotype
Longitudinal population-based
F/U: post-injury
IC: in subset of previous
longitudinal study (previously
tested); participated in 2
evaluations in previous study
EC: missing data; mini-mental
state examination score <23;
not genotyped; dementia at
1st/2nd/3rd evaluation in
previous study

n = 31 mTBI; 62 controls
Attrition: 0
Stats for mTBI:
Age (at entry): 55.2 13.6
Sex: 58.1% M
TSI (n = 18):
(injury-post-seccion):
19.7 14.5 mos
IS (self-report, conrmed
w/ criteria): mild
AT:
BL: pre-injury (previous
study)
F/U: post-injury
Controls: 1st, 2nd
assessments from previous
study

McNemars test:
differences pre- to
post-injury; for
controls, differences
1st to 2nd assessment
Fishers exact test:
between-group
comparison (i.e. mTBI
vs. controls)

NR

Yes/no questions about


presence of fatigue
Fatigue presence: Yes to Do
you often feel fatigued?
Fatigue frequencies:
Pre-injury/1st assessment:
mTBI: 16.1%
Control: 25.8%
Post-injury/2nd assessment
(p < .05):
mTBI: 41.9%
Control: 19.4%
Post-injury fatigue in mTBI w/
APOE 4: 58%; w/o APOE 4:
32%
2nd assessment fatigue in
controls w/ APOE 4: 17%; w/o
APOE 4: 21%
mTBI w/ APOE 4 vs. control w/
APOE 4: signicant (p = .02)
mTBI w/o APOE 4 vs. control
w/o APOE 4: NS (.52)

Fatigue frequency
within-group change:
mTBI: signicant
(p < .05)
Control: NS

van der Naalt et al.


(1999)
NL
Hospital

Report long-term outcomes in


persons w/ mild-mod head
injury, GCS 914, irrespective
of length of hospital stay and
CT abnormalities; examination
of complaints, return to work;
whether GCS scores at
admission and length of PTA is
predictive of outcome
Prospective
F/U: post-injury: 1, 3, 6, 12 mos
IC: 1565 yrs; GCS 914 at
hospital admission; PTA 1 h
EC: earlier admission for head
injury; history of drug/alcohol
abuse; psychiatric
disorder/mental retardation
diagnosis; severe aphasia
interfering w/ report of PTA;
PTA 28 d

n = 67
Attrition: 0
Age: 33.2 14.7
Sex: 64.2% M
TSI: 12 h post-injury at
hospital admission
IS: GCS mean: 12.6 (range
914); PTA mean:
7.8 7.3; range 130 d)
Mild: 64.2%
Mod: 35.8%
AT: post-injury:
t1 : 1 mos
t2 : 3 mos
t3 : 6 mos
t4 : 12 mos

Students
t/MannWhitney U
tests: where
appropriate
Pearsons correlation
coefcients:
independent measure
associations
 analysis:
interobserver scoring
Chi-square w/
correction for
continuity: frequencies
Multivariate regression
by stepwise backward
method
Outcome variables
distribution: normal
for scales w/ 4 points;
all else categorical

NR
History of addiction to
alcohol or drugs as part
of exclusion criteria
Alcohol intolerance at
1, 3, 6 mos and 1
year = 6, 11, 17 and
20%, respectively

Fatigue as complaint on
symptoms checklist
Fatigue frequencies at time
points:
t1 : 57%
t2 : 61%
t3 : 45%
t4 : 45%
At all time points, features as
1/6 most frequent complaints
No correlation between injury
characteristics and complaints
at all time points

NR

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Sundstrom et al. (2007)


SE
Longitudinal
prospective cohort
study sample

701

702

Table 1 (Continued)
Reference
Country
Sample by

Study early clinical predictors,


PCS in mTBI persons with
persistent symptoms 2 mos
post-injury; study w/i a study
Prospective
F/U: post-injury: 1, 2, 8 wks
IC: mTBI patients from other
prospective study
EC: could not be reached for
F/U

Sample size
Attrition
Age, sex (% M)
Time since injury (TSI)
Injury severity (IS)
Assessment time points/N
assessed (AT: t1 , t2 , etc.)

n = 180; 2 mTBI groups: 1:


persistent PCS (PPCS)
n = 17; non-PPCS n = 163;
control group n = 40
Attrition: 0
Age:
1: 37.2 14.2
2: 35.7 16.4
Sex:
1: 41% M
2: 48% M
TSI: recruitment at ED
IS: GCS mean:
1: 14.9 .3
2: 14.9 .3
Intracranial lesions:
1: 53%
2: 15%
AT: post-injury:
t1 : 1 wk
t2 : 2 wks
t3 : 8 wks

Statistical method

Chi-square:
associations of PCS
items in groups 1 and 2
Logistic regression:
predictors of mTBI w/
PPCS
p < .05

Medications

NR

Results

Fatigue denition
Frequencies, scores

Score differences over


time
Notes

Checklist of PCS (CPCS)


featuring fatigue item
Fatigue frequencies by CPCS:
mTBI:
t1 : 23%
t2 : 11%
t3 : 3%
Fatigue frequency within 2
mos (previous study):
Control (n = 40): 8%
Fatigue frequencies by CPCS by
group:
1: t1 : 24%; t2 : 35%
2: t1 : 23%; t2 : 9%
t1 : OR = .02
t2 : OR = 2.76
Fatigue as one of most
common PCS in group 1
Fatigue strongly associated
with PPCS incidence at t2 :
2 = 11.12, p < .01

NR

APOE 4 apolipoprotein-4; BDI, Beck depression inventory; BFS, Barroso fatigue scale; BL, baseline; CNS, central nervous system; CPCS, checklist for post-concussion syndrome; CT, computed tomography; d, day; ED, emergency
department; IC/EC, inclusion/exclusion criteria; IS, injury severity; ISS, injury severity score; FSS, fatigue severity scale; FNS, exyx neurotherapy system; F/U, follow-up; GCS, Glasgow coma scale; GOSE, Glasgow outcome
scale-extended; GFI, global fatigue inventory; LOC, loss of consciousness; MFIS, modied fatigue impact scale; MRI, magnetic resonance imaging; MHI, mild head injury; mTBI, mild traumatic brain injury; mos, months; NR, not
reported; NS, not signicant; PCS, post-concussion syndrome; PCSC, post-concussion syndrome checklist; POMS, prole of moods scale; PPCS, persistent post concussive syndrome; PTA, post traumatic amnesia; RPQ, Rivermead
post-concussive questionnaire; PSQI, Pittsburgh sleep quality index; SF-36, 36-item short form health survey (from medical outcomes study); TC, trauma controls; TBI, traumatic brain injury; tx, treatment; TSI, time since injury;
SSRI, selective serotonin reuptake inhibitor; VAS, visual analog scale; wks, weeks; yrs, years.

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Yang et al. (2009)


TW
Level I trauma center

Objective
Design
Follow-up (F/U)
Inclusion/exclusion criteria
(IC/EC)

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

703

Table 2
Summary of reported predictors of fatigue.
Study

Measure and purpose in study

Results: predictors of fatigue

De Leon et al. (2009)

MOS SF-36 Vitality subscale


Severity measure (i.e. low score indicates more fatigue)

Predictors of severity at 12 months


Baseline fatigue (p = .000)
Counseling for mental health (p = .016)
Medical disability (p = .012)
Marital status (p = .006)
Litigation involvement (p = .044)

Norrie et al. (2010)

Sundstrom et al. (2007)

FSS
Frequency measure (i.e. score 3.7 indicates fatigue)
RPQ fatigue item
Frequency measure (i.e. symptom rating 2 indicates
fatigue)
MOS SF-36v2 Vitality subscale
Severity measure (i.e. low score indicates more fatigue)
Fatigue question
Frequency measure (i.e. answer yes to question Do
you often feel fatigued? indicates fatigue)

NS
Sex (p = .762)
Age (p = .507)
Education (p = .77)
Prior drug/alcohol treatment (p = .382)
Motor vehicle crash (p = .444)
Injury type (p = .792, p = .427)
Predictor of severity at 3 mos
Severity at wk 1 (FSS) (R = 0.53; p < .000)
Predictor of severity at 6 mos
Severity at wk 1 (R = 0.49; p < .000)
Severity at 3 mos (R = 0.76; p < .000)
Depression at 3 mos (B = .12; SE = .04; = .25; p < .0000)
Anxiety at 3 mos (B = .01; SE = .03; = .04; p = .610)
Predictor of frequency post-injury
APOE 4 genotype (p = .02)

APOE 4, apolipoprotein-4; MOS SF-36v2 NS, not signicant 36-item short form health survey vitality subscale (from medical outcomes study); FSS, fatigue severity scale;
RPQ, Rivermead post-concussive questionnaire; wk, week.

time point). Fatigue resolution/exacerbation/no change designations were reported.


Prognostic factors associated with fatigue were extracted for all
cohorts and untreated/no effect RCTs. All factors inuencing the
course of fatigue, as reported by authors, were considered associated with fatigue and not necessarily causal factors. To address
our third research objective (i.e. health consequences of fatigue in
TBI), we evaluated reports of poor health outcomes associated with
fatigue after TBI.

(i.e. untreated group) data and no effect data (i.e. intervention has
not effect) were utilized to address the second research objective
(i.e. to determine the course of fatigue) in patients with TBI.
Study quality was independently assessed by two reviewers (TM
and TK), using guidelines developed by Hayden et al. (2006) for
assessment of prognostic studies (Table 4). The appraisal was performed in two steps. First, the items related to six potential sources
of bias (i.e. study participation and attrition, associated factors and
outcome measurements, confounding measurement, and analyses)
were assessed, then presence of potential biases was judged Yes,
Partly, No, or Unsure. To summarize the level of evidence,
we used a system similar to the Scottish Intercollegiate Guidelines
Network (SIGN) methodology (SIGNPG, 2013): (i) +++ when all or
most of the quality criteria proposed by Hayden et al. were fullled
(i.e. allowing one Partly while appraising all potential sources of
bias); (ii) ++ when the majority of criteria were fullled; (iii) +
when few criteria were fullled (i.e. at least one Yes). Additionally, as proposed by SIGN, studies with retrospective data collection
did not receive a ++ rating, as this design is weaker than prospective data collection. We refer to group (i) as high quality studies;
group (ii) as good quality studies; and group (iii) as fair quality
studies.

The nature of our research questions related to fatigue in the TBI


population (i.e. prognostic factors, course, and consequences) raises
the issue of zero-time bias. In prognostic studies, testing should
start at a dened point, called zero time (Giobbie-Hurder et al.,
2013; van Rein et al., 2014). Designated zero times (i.e. baseline or
rst assessment) varied between studies included in this review.
For this reason and to best address our research questions, studies
were grouped based on whether baseline assessments were conducted before or after the one-month post-injury mark. This point
was arbitrarily set.

2.6. Data synthesis

2.8. Missing data

A best-evidence synthesis approach was applied, synthesizing


ndings from studies with sufcient quality through tabulation and
qualitative description (Slavin, 1995; Carroll et al., 2004a).
Results were grouped into three main categories: prognostic
factors of fatigue, course of the fatigue, and consequences of fatigue
(Tables 13). For studies utilizing measures of fatigue prevalence,
sample size-weighted mean frequencies were reported. Fatigue
severity measures used in the studies were reported with their corresponding sample mean scores. To determine the course of fatigue,
matching assessment times (i.e. time post-injury that fatigue was
measured) were grouped, with their corresponding fatigue frequencies, and a sample size-weighted mean frequency value was
calculated for time points with more than one contributing frequency value (i.e. more than one study reporting fatigue at that

Primary authors were contacted in the case of missing data. In


the case of duplicate publications and companion papers of a primary study, we attempted to maximize the yield of information by
the simultaneous evaluation of all available data (i.e. all data necessary to address the three objectives of research, see Section 2.5).
Original publications took priority.

2.7. Zero-time

3. Results
3.1. Literature search and quality assessment
Of 2745 articles identied, 33 were selected for full-text review
(Lidvall et al., 1974; Ponsford et al., 2012; De Leon et al., 2009;
Lundin et al., 2006; Meares et al., 2011; Mickeviciene et al., 2004;

704

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Table 3
Summary of reported consequences of fatigue.
Study

Measure and purpose in study

Results: consequences of fatigue

Norrie et al. (2010)

FSS
Frequency measure (i.e. score 3.7 indicates fatigue)
RPCQ fatigue item
Frequency measure (i.e. symptom rating 2 indicates
fatigue)
RPSQ
Presence/problem status of 16 post-concussional
symptoms
FSS
Severity measure (i.e. higher score indicates more
fatigue)

Predictors of persistent PCS (RPSQ total) at 6 mos


Fatigue severity at wk 1(R = 0.40; p < .000)
Fatigue severity at 3 mos (R = 0.53; p < .000)

Sigurdardottir et al. (2009)

Full sample
Predictor of GOSE
Fatigue severity at 3 mos (R2 = .61, p < .001)
Predictor of GOSE at 12 mos
Fatigue severity by FSS (B = .13; SE = .04; = .25; p < .001)
mTBI
Predictor of GOSE at 12 mos
Fatigue severity by FSS (R2 = .47, p < .01)
Moderate/severe TBI
Fatigue severity by FSS (R2 = .58, p < .001)

FSS, fatigue severity scale; GOSE, Glasgow Outcome Scale-Extended; mTBI, mild traumatic brain injury; mos, months; PCS, post-concussion syndrome; RPCSQ, Rivermead
post-concussion symptom questionnaire; wk, week.

Norrie et al., 2010; van der Naalt et al., 1999; Yang et al., 2009;
McLean et al., 1993; Sundstrom et al., 2007; Hutchinson et al., 2009;
Sigurdardottir et al., 2009; Kempf et al., 2010; Driver and Ede, 2009;
Gemmell and Leathem, 2006; Bushnik et al., 2008a; Jha et al., 2008;
Kaiser et al., 2010; Khateb et al., 2005; Schoenberger et al., 2001;
Hou et al., 2012; Bhambhani et al., 2008; Bateman et al., 2001;
Bushnik et al., 2008b; Cooper et al., 2009; Hillier et al., 1997; Wiart
et al., 2012; Kim et al., 1999; Haboubi et al., 2001; Olver et al., 1996;
Rees and Bellon, 2007) and 22 were included in the nal review

(Fig. 2) (Lidvall et al., 1974; Ponsford et al., 2012; De Leon et al.,


2009; Lundin et al., 2006; Meares et al., 2011; Mickeviciene et al.,
2004; Norrie et al., 2010; van der Naalt et al., 1999; Yang et al., 2009;
McLean et al., 1993; Sundstrom et al., 2007; Hutchinson et al., 2009;
Sigurdardottir et al., 2009; Kempf et al., 2010; Driver and Ede, 2009;
Gemmell and Leathem, 2006; Bushnik et al., 2008a; Jha et al., 2008;
Kaiser et al., 2010; Khateb et al., 2005; Schoenberger et al., 2001;
Hou et al., 2012). Supplementary Table 2 reports reasons for exclusion of 11 studies (Bhambhani et al., 2008; Bateman et al., 2001;

Table 4
Quality assessment of studies using guidelines developed by Hayden et al. (2006).
Study

Study participation

Time-zero

Study
attrition

Prognostic
factor

Outcome

Confounding
measurement
and account

Analysis

Reason for
exclusion

Overall
assessment

Bushnik et al. (2008a)


De Leon et al. (2009)
Driver and Ede (2009)
Gemmell and Leathem (2006)
Hou et al. (2012)
Hutchinson et al. (2009)
Jha et al. (2008)
Kaiser et al. (2010)
Kempf et al. (2010)
Khateb et al. (2005)
Lidvall et al. (1974)
Lundin et al. (2006)
McLean et al. (1993)
Meares et al. (2011)
Mickeviciene et al. (2004)
Norrie et al. (2010)
Ponsford et al. (2012)
Schoenberger et al. (2001)
Sigurdardottir et al. (2009)
Sundstrom et al. (2007)
van der Naalt et al. (1999)
Yang et al. (2009)

No
No
Partlye
Partlye
No
Partlye
No
Partlye
No
Partlye
No
No
No
No
No
No
No
Partlye
No
No
No
No

Yesi
No
Yesi
Yesi
Yesi
No
Yesi
Yesi
Yesi
Yesi
No
No
No
No
No
No
No
Yesi
Yesi
No
No
No

Yesd
Partlya
No
Not sure
No
Not sure
No
No
No
Partlya
No
No
No
Partlya
No
Partlya
Partlya
Not sure
No
No sure
No
Partlya

NA
No
NA
NA
No
NA
Partlyg
Not sure
NA
NA
NA
NA
NA
No
NA
No
NA
NA
No
No
NA
No

No
No
No
No
No
No
No
No
No
Partlyf
Partlyf
Partlyf
Partlyf
No
No
No
No
No
No
Partlyf
Partlyf
Partlyf

No
No
Not sure
No
Partlyh
Yesb
No
No
No
Partlyh
Partlyh
Partlyh
No
Partlyh
No
Partlyh
Partlyh
No
Partly
No
Partlyh
Partlyh

No
No
No
Partlyc
No
No
No
No
Partlyc
Partlyc
Partlyc
No
Partlyc
No
Partlyc
No
Partlyc
Partlyc
No
No
No
No

+
+++
+
+
+
+
+
+
+
+
++
++
++
++
+++
++
+
+
+
++
++
+

Yes yes, sources of potential bias are presented; No no potential bias; Not sure not enough details were reported to make a decision (in some cases authors were
contacted); NA not applicable according to the study design or type of analyses used.
a
Not all required information about study attrition was provided.
b
A study does not address the possibility of confounding.
c
Some errors in analyses performed were observed: e.g. limited details about analyses.
d
Completeness of follow-up was not adequate.
e
Small sample size.
f
Detail information about measure used was not provided or used measure was not validated.
g
Analyses performed were not adequate.
h
Not all important covariates were included or exclusion criteria are not completed.
i
Baseline assessment performed after 1 month post-injury.

Identification

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Articles identified through database


search (N=2745): Embase (n=1128),
Medline (n=724), PsycINFO (n=453),
CINAHL (n=265), Cochrane database of
systematic reviews (n=175)

705

Other articles identified (i.e. through


scan of references of relevant articles)
(n=3)

Articles screened (n=2748)

Screening

Duplicates excluded (n=2388)


Excluded based on title (n=169)

Remaining articles (i.e. after exclusion based


on title, removal of duplicates) (n=191)
Excluded based on abstract/full text
(n=94)
Excluded based on format (i.e.
conference presentations, theses, etc.)
(n=8)

Eligibility

Articles assessed for eligibility (i.e. reading


full text) (n=89)
Excluded based on inclusion criteria
(N=56): not refereed (n=21), no measure
of fatigue (n=16), study design (i.e. not
longitudinal) (n=9), no separate analysis
of TBI patients (n=5); focus on different
construct (i.e. apathy, sleepiness, etc.)
(n=5)

Articles to next round (n=33)

Included

Excluded based on quality assessment


(n=6)
Inability to obtain information after
contacting author(s) (n=3)
Same population in multiple studies (i.e.
included oldest study) (n=2)

Articles included in review (n=22)


Fig. 2. Flow chart documenting process of article selection for review. Embase (1974-4/11/2013); Medline (1946-4/14/2013); PsycINFO (1806-4/7/2013); CINAHL (19804/18/2013); Cochrane (2005-3/2013).

Bushnik et al., 2008b; Cooper et al., 2009; Hillier et al., 1997; Wiart
et al., 2012; Kim et al., 1999; Haboubi et al., 2001; Olver et al., 1996;
Rees and Bellon, 2007; Meares et al., 2011). Main analyses featured
11 inception cohort studies with baseline assessment performed
within one-month post-injury (Lidvall et al., 1974; Ponsford et al.,
2012; De Leon et al., 2009; Lundin et al., 2006; Meares et al., 2011;
Mickeviciene et al., 2004; Norrie et al., 2010; van der Naalt et al.,
1999; Yang et al., 2009; McLean et al., 1993; Hutchinson et al.,
2009). Separate analyses of studies with baseline assessment after
one month included 11 studies (McLean et al., 1993; Sundstrom
et al., 2007; Sigurdardottir et al., 2009; Kempf et al., 2010; Driver
and Ede, 2009; Gemmell and Leathem, 2006; Bushnik et al., 2008a;
Jha et al., 2008; Kaiser et al., 2010; Khateb et al., 2005; Schoenberger
et al., 2001; Hou et al., 2012), among them six RCTs (Driver and Ede,

2009; Gemmell and Leathem, 2006; Jha et al., 2008; Kaiser et al.,
2010; Khateb et al., 2005; Schoenberger et al., 2001).
All 22 studies (Lidvall et al., 1974; Ponsford et al., 2012; De Leon
et al., 2009; Lundin et al., 2006; Meares et al., 2011; Mickeviciene
et al., 2004; Norrie et al., 2010; van der Naalt et al., 1999; Yang et al.,
2009; McLean et al., 1993; Sundstrom et al., 2007; Hutchinson et al.,
2009; Sigurdardottir et al., 2009; Kempf et al., 2010; Driver and Ede,
2009; Gemmell and Leathem, 2006; Bushnik et al., 2008a; Jha et al.,
2008; Kaiser et al., 2010; Khateb et al., 2005; Schoenberger et al.,
2001; Hou et al., 2012) were assessed as having Partly or No
on all bias criteria. Two studies (De Leon et al., 2009; Mickeviciene
et al., 2004) were of high quality (+++), six (Lidvall et al., 1974;
Lundin et al., 2006; Norrie et al., 2010; van der Naalt et al.,
1999; McLean et al., 1993; Sundstrom et al., 2007) were of good

706

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

quality (++) and the remaining 14 (Ponsford et al., 2012; Yang


et al., 2009; Hutchinson et al., 2009; Sigurdardottir et al., 2009;
Kempf et al., 2010; Driver and Ede, 2009; Gemmell and Leathem,
2006; Jha et al., 2008; Kaiser et al., 2010; Khateb et al., 2005;
Schoenberger et al., 2001; Bushnik et al., 2008b) were of fair quality
(+). The latter group were penalized by the SIGN criteria (SIGNPG,
2013) for incomplete statistical analysis, potential confounders,
selection bias due to study attrition, and/or zero-time bias (Table 4).
3.2. Study characteristics
Tables 24 summarize the study characteristics pertinent to our
research questions: population characteristics, denitions of TBI,
denitions of fatigue, follow-up time, statistical analysis methods,
and study results. Means were calculated for the reviewed studies
sample data.
3.3. Studies with baseline assessment up to one month
post-injury
Eleven studies featured a total of 1366 participants with TBI.
Nine studies performed recruitment at emergency departments,
hospitals, or trauma centers, making up 91% of the total group
(n = 1244) (Lidvall et al., 1974; Ponsford et al., 2012; De Leon et al.,
2009; Lundin et al., 2006; Meares et al., 2011; Mickeviciene et al.,
2004; Norrie et al., 2010; van der Naalt et al., 1999; Yang et al.,
2009). One study recruited participants from the community (7.5%
of the sample, or n = 102) (McLean et al., 1993). One study featured
university athletes within four days of concussion, making up 1.5%
of the sample (n = 20) (Hutchinson et al., 2009).
Nine studies featured strictly participants with mild TBI (Lidvall
et al., 1974; Ponsford et al., 2012; De Leon et al., 2009; Lundin et al.,
2006; Meares et al., 2011; Mickeviciene et al., 2004; Norrie et al.,
2010; Yang et al., 2009), one examined mild to moderate TBI (van
der Naalt et al., 1999), and one included all TBI severities (McLean
et al., 1993).
The study samples comprised males in a range between 42%
(De Leon et al., 2009) and 69% (Lidvall et al., 1974), with a mean
62.1 9.5% across studies. One study did not report a sex ratio, stating the sample consisted mainly of single males, and was included
in the mean calculation as 75% male (McLean et al., 1993). The mean
age ranged between 20.1 (Hutchinson et al., 2009) and 41.2 years
of age (De Leon et al., 2009), with a mean 33.5 5.6 years across
studies. Mean time since injury (TSI) to baseline assessment ranged
from one day (Lundin et al., 2006) to one month (Mickeviciene et al.,
2004), and the mean TSI across all studies was 0.34 0.38 months,
or 10.2 11.4 days.

2009; Gemmell and Leathem, 2006; Jha et al., 2008; Kaiser et al.,
2010; Schoenberger et al., 2001), two comprised patients with
moderate to severe TBI (Bushnik et al., 2008a; Khateb et al., 2005),
and two with mild TBI (Sundstrom et al., 2007; Hou et al., 2012).
The percentage of males ranged between 17%(Schoenberger
et al., 2001) and 85% (Kaiser et al., 2010), with a mean 62 19.2%
across samples. The mean age ranged from 31 years (Bushnik et al.,
2008a) to 55.2 years of age (Sundstrom et al., 2007), with a mean
39.9 6.6 years across studies. Mean TSI, across the ten studies, was
39.5 35.5 months (i.e. 1185 1065 days) and baseline assessment
times ranged from 2.6 months (Bushnik et al., 2008a) to 8.6 years
post-injury (Gemmell and Leathem, 2006).
3.5. Assessment of TBI
Considerable between-study variation was observed in TBI
diagnostic criteria and denitions, irrespective of TSI at baseline
assessment (Tables 13). Most studies (17/22) used a combinatorial
approach to conrm and assess TBI, using tools such as the Glasgow Coma Scale (GCS), duration of posttraumatic amnesia (PTA)
and loss of consciousness (LOC), and clinical evaluation (Lidvall
et al., 1974; Ponsford et al., 2012; De Leon et al., 2009; Lundin et al.,
2006; Meares et al., 2011; Mickeviciene et al., 2004; Norrie et al.,
2010; van der Naalt et al., 1999; Yang et al., 2009; McLean et al.,
1993; Hutchinson et al., 2009; Sigurdardottir et al., 2009; Gemmell
and Leathem, 2006; Bushnik et al., 2008a; Jha et al., 2008; Khateb
et al., 2005; Hou et al., 2012). Three studies (Sundstrom et al., 2007;
Driver and Ede, 2009; Schoenberger et al., 2001) used other methods, including patient report, description of damage and/or lesions
based on medical records, and diagnoses of referring professionals.
Two studies used GCS scores alone (Kempf et al., 2010; Kaiser et al.,
2010).
3.6. Methods used for assessing fatigue

3.4. Studies with baseline assessment after one month post-injury

Measures used to assess fatigue in the TBI population varied


depending on the study objectives. Studies where fatigue was not
a main focus most commonly used a single item for the symptom within a checklist with broad symptom coverage (9/22) (e.g.
Rivermead post-concussion questionnaire (RPQ) and the postconcussion syndrome checklist (PCSC)) (Tables 2 and 3). If fatigue
was studied more extensively, standardized measures looking at
different aspects of the symptom, such as momentary perception,
chronic characteristics, impact of fatigue on function, rating/rank of
fatigue intensity/severity and dimensions of fatigue (i.e. cognitive,
physical) were utilized. Four of the 22 studies used more than one
measure to assess fatigue. All fatigue scales were designed for other
populations, with some having psychometric properties described
in the TBI population (Supplementary le 3).

Eleven studies (Sundstrom et al., 2007; Sigurdardottir et al.,


2009; Kempf et al., 2010; Driver and Ede, 2009; Gemmell and
Leathem, 2006; Bushnik et al., 2008a; Jha et al., 2008; Kaiser et al.,
2010; Khateb et al., 2005; Schoenberger et al., 2001; Hou et al.,
2012) featured a total 482 participants. Eight of these studies
recruited from the community, making up 43.4% of the total group
(i.e. n = 209) (Sundstrom et al., 2007; Sigurdardottir et al., 2009;
Kempf et al., 2010; Driver and Ede, 2009; Gemmell and Leathem,
2006; Bushnik et al., 2008a; Kaiser et al., 2010; Schoenberger et al.,
2001). Bushnik et al. (2008a) (n = 51) and Sigurdardottir et al. (2009)
(n = 115) recruited from an inpatient rehabilitation center and a
level 1 trauma center, respectively. Hou et al. (2012) recruited
from the emergency department (n = 107), with baseline assessment performed at a laboratory at a later time.
Seven studies featured participants with all severities of TBI
(Sigurdardottir et al., 2009; Kempf et al., 2010; Driver and Ede,

3.6.1. Multi-item scales


Most studies (14/22) assessed fatigue based on standardized
self-report measures four used the fatigue severity scale (FSS)
(Sigurdardottir et al., 2009; Kempf et al., 2010; Jha et al., 2008;
Kaiser et al., 2010); four the short-form health survey-36 (SF36) vitality subscale (Ponsford et al., 2012; De Leon et al., 2009;
Norrie et al., 2010; Gemmell and Leathem, 2006); two a visual
analog scale (VAS) for fatigue (Mickeviciene et al., 2004; Gemmell
and Leathem, 2006), and two the prole of mood states (POMS)
fatigue-inertia scale (Hutchinson et al., 2009; Driver and Ede, 2009).
One study assessed fatigue with the Barroso fatigue scale (BFS), a
synthesis of ve independent scales, with additions (Bushnik et al.,
2008a). The BFS yields FSS and global fatigue index (GFI) scores. One
study utilized the modied fatigue impact scale (MFIS) (Jha et al.,
2008), one used the fatigue assessment inventory (Khateb et al.,
2005), and one again used the multidimensional fatigue inventory

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

(Schoenberger et al., 2001). We refer the reader to Supplementary


le 3 for descriptions of measures.
On the FSS, participants rate their level of agreement with
respect to nine statements about the severity of fatigue and its
impact on everyday activities (Krupp et al., 1989). The total score is
the mean, and higher scores indicate greater fatigue. In some cases,
studies reported frequency of fatigue by the FSS, dening presence
of the symptom as a total score 4 (Kempf et al., 2010; Kaiser et al.,
2010), or 3.7 (Norrie et al., 2010). Others used FSS scores as indicators of fatigue severity (Sigurdardottir et al., 2009; Bushnik et al.,
2008a; Jha et al., 2008).
On the four-item SF-36 vitality subscale, participants choose,
on a six-point scale, the frequency of events related to fatigue
and energy (Ware, 1992). Fatigue item means are combined with
reverse-scored energy means to yield a total score. Two studies
utilized two different versions of the measure (i.e. SF-36 and SF-36
version 2). SF-36 and the updated SF-36 version 2 are comparable
in terms of scores.

3.6.2. Single item assessment of fatigue


Nine papers used a single item or question to assess fatigue
(Lidvall et al., 1974; Ponsford et al., 2012; Lundin et al., 2006;
Meares et al., 2011; van der Naalt et al., 1999; Yang et al., 2009;
McLean et al., 1993; Sundstrom et al., 2007; Hou et al., 2012), presenting TBI patients with a list of symptoms (e.g. Rivermead PCSQ,
PCS checklist, etc.), including fatigue. One study determined presence of fatigue by participants Yes responses to the question, Do
you often feel fatigued? (Sundstrom et al., 2007).

3.6.3. Multiple measures of fatigue


Four studies used more than one measure to assess fatigue
(Ponsford et al., 2012; Norrie et al., 2010; Gemmell and Leathem,
2006; Jha et al., 2008). Gemmell and Leathem (2006) utilized the
SF-36 vitality subscale with the VAS for fatigue, for a measure of
fatigue severity. Jha et al. (2008) worked with the FSS and MFIS
and reported change in severity of fatigue with use of medications.
Norrie et al. (2010) utilized the SF-36 vitality subscale with the RPQ
for severity and frequency values. The SF-36 vitality subscale was
used again by Ponsford et al. (2012) together with a single item
from the PCS checklist, to report frequency and severity of fatigue
across time.

3.7. Overall predictors of fatigue


Only studies with baseline assessment prior to one month postinjury investigated predictors of fatigue. All statistically signicant
predictors of fatigue identied are reported in Table 2. Three studies, one of high quality and two of moderate quality (De Leon et al.,
2009; Norrie et al., 2010; Sundstrom et al., 2007) identied eight
factors signicantly associated with fatigue in TBI patients (Table 3).
The factors comprised earlier fatigue severity, signicant in two
studies (De Leon et al., 2009; Norrie et al., 2010), carriage of the
apolipoprotein E 4 allele, signicant in one study (Sundstrom et al.,
2007), having seen a counselor for a mental health issue, medical
disability, marital status (i.e. widowed, divorced, or separated) and
litigation involvement, all signicant in one study (De Leon et al.,
2009), and depression, also signicant in one study (Norrie et al.,
2010). One moderate quality study did not nd a signicant effect
of sex, education, or TBI type/severity on fatigue in a fully adjusted
model (Norrie et al., 2010). In that same study, anxiety at three
months was not a predictor of fatigue at six months (Norrie et al.,
2010).

707

3.8. The course of fatigue


Mean frequencies of fatigue at time points with more than one
reported value (i.e. two or more studies reported frequency at the
same time post-injury) were weighted based on sample size. For
studies where baseline assessment was conducted prior to or at
one month post-injury, mean weighted frequencies were 46.6%
(SD = 32.7, n = 206) (Lidvall et al., 1974; Ponsford et al., 2012), 45.9%
(SD = 24.8, n = 637) (Lidvall et al., 1974; Ponsford et al., 2012; Norrie
et al., 2010; Yang et al., 2009), 17.3% (SD = 13.6, n = 325) (Lidvall
et al., 1974; Meares et al., 2011; Yang et al., 2009), 45.2% (SD = 29,
n = 230) (Lidvall et al., 1974; van der Naalt et al., 1999; McLean
et al., 1993), 30.5% (SD = 11.7, n = 830) (Lidvall et al., 1974; Lundin
et al., 2006; Mickeviciene et al., 2004; Norrie et al., 2010), 32.4%
(SD = 7.3, n = 269) (Norrie et al., 2010; van der Naalt et al., 1999) and
37.4% (SD = 8.1, n = 354) (Mickeviciene et al., 2004; van der Naalt
et al., 1999; McLean et al., 1993) for two days, six days-one week,
two weeks, one month, three months, six months and one year
post-injury, respectively. The number of studies contributing to the
mean for a particular time point ranged from two studies for two
days and six months post-injury to seven studies for three months
post-injury (Fig. 3a).
For studies with baseline assessment after one-month postinjury, just one mean weighted frequency value was obtained for
one time point, 22.8% (SD = 5.4, n = 172) (Kempf et al., 2010; Hou
et al., 2012) at six months post-injury. Two studies contributed
values for calculation of this mean. The remaining time points comprised single studies and therefore one frequency value (Fig. 3b).
3.9. The course of fatigue, by injury severity
When fatigue frequency calculations were stratied by injury
severity, mean weighted frequencies could only be obtained for
the mild TBI group with baseline assessments conducted less than
or at one month after injury. The frequencies of fatigue were 46.6%
(SD = 32.7, n = 206) (Lidvall et al., 1974; Ponsford et al., 2012), 45.9%
(SD = 24.8, n = 637) (Lidvall et al., 1974; Ponsford et al., 2012; Norrie
et al., 2010; Yang et al., 2009), 17.3% (SD = 13.6, n = 325) (Lidvall
et al., 1974; Meares et al., 2011; Yang et al., 2009) and 27.8%
(SD = 7.8, n = 763) (Lidvall et al., 1974; Ponsford et al., 2012; Lundin
et al., 2006; Meares et al., 2011; Mickeviciene et al., 2004) for two
days, six days-one week, two weeks and three months post-injury,
respectively. The number of studies contributing to the means
ranged from two studies for two days post-injury to six studies
for three months. The mild TBI group with baseline assessments
performed after one-month post-injury and mild to moderate and
mixed groups all had one contributing study each. The studies featuring moderate to severe TBI did not report fatigue frequencies.
For the purpose of comparison of fatigue frequencies between
mild TBI and other severities, studies reporting frequencies in samples of mild to moderate and mixed severities of TBI were grouped.
Mean frequencies were 66.3% (SD = 8.5, n = 147) (van der Naalt et al.,
1999; McLean et al., 1993) and 46.2% (SD = 1, n = 162) (van der Naalt
et al., 1999; McLean et al., 1993) for one month and one-year post
TBI, respectively. The two contributing studies had baseline assessment performed within the rst month post-injury.
3.10. Fatigue severity
In two studies using the FSS, the sample mean scores at 12
months were 3.20 1.39 (Norrie et al., 2010) and 2.9 1.6 (Bushnik
et al., 2008a). One study, using an alternate FSS scoring system
(Jha et al., 2008), reported similar fatigue severity. In the two studies that utilized the SF-36 vitality subscale and the SF-36 vitality
subscale version 2, the mean scores at one year post-injury were
62.11 20.18 (Norrie et al., 2010) and 49.6 11.83 and 52.3 12.22

708

Table 5
Fatigue measures and their corresponding scores at assessment.
Study
Bushnik et al. (2008b)

Injury type/severity

Assessment

Measure

Sev(BL)

Sev(t1 )

Sev(t2 )

Sev(t3 )

TBI/mod-sev

t1 : 2.6 1.8 mos


t2 : 12.6 1.2 mos
t3 : 23.2 3.4 mos

GFI

NR

23 10

17 11

20 11

3.4 1.5

2.9 1.6

3.2 1.8

FSS

1: HI w/ PTA and/or LOC/mild


2: HI only/mild

BL: ED
t1 : 1 mo
t2 : 3 mos
t3 : 12 mos

SF-36v

1: 52.8 9.53
2: 50.4 10.48

NR

NR

1: 52.3 12.22
2: 49.6 11.83

Driver and Ede (2009)

TBI control/NR

BL: before start


t1 : 8 wks later

POMS
fatigue-inertia
subscale

1.24 .61

1.29 .57

NA

NA

Gemmell and Leathem


(2006)

TBI control/NR

BL: before start


t1 : 6 wks later

SF-36v

47.50 20.18

38.75 4.43

NA

NA

Hutchinson et al. (2009)

TBI/mild

BL: 96 h post-injury
t1 , t2 , t3 : 3 non-consecutive d over 2
wks

POMS
fatigue-inertia
subscale

4.3

6.8

4.4

Jha et al. (2008)

1: TBI/mild-sev modanil rst


2: TBI/mild-sev placebo rst

BL: before start


t1 : 4 wks
t2 : 10 wks

MFIS

1:
46.56 19.28
2:
47.17 15.53
1:
45.22 11.82
2:
44.46 12.17

1:
38.65 16.09
2:
36.45 15.03
1:
39.36 15.61
2: 37.7 12.55

1: 35.63 20
2:
33.55 18.16

NA

1:
37.13 18.33
2:
36.91 14.08

NA

FSS

Khateb et al. (2005)

Brain injury/PTA 8 10 d

BL: before start


t1 : 3 mos

29-item fatigue
scale

132.6 27.3

126.1 32.3

NA

NA

Mickeviciene et al. (2004)

Concussion/NR

BL: ED
t1 : 3 mos
t2 : 1 y

VAS fatigue
item

NR

50 28

50 30

NA

Norrie et al. (2010)

TBI/mild

t1 : 1 mo
t2 : 3 mos
t3 : 12 mos

SF-36 Vitality

NR

46.57 24.72

60.21 19.68

62.11 20.18

FSS

NR

3.99 1.53

3.29 1.44

3.20 1.39

Ponsford et al. (2012)

TBI/mild

BL: ED post-injury
t1 : 1 wk
t2 : 3 mos

PCSC Checklist

NR

2.8

2.2

NA

Schoenberger et al. (2001)

Closed HI/mod-sev

BL: before start


t1 : 68 wks later

MFI: Gen
MFI: Phys
MFI: Men

14.83 4.17
10.50 4.51
15.50 3.83

14.00 4.56
10.83 5.34
15.67 3.50

NA

NA

Sigurdardottir et al. (2009)

TBI/mild-sev

BL: ED
t1 : 3 mos
t2 : 12 mos

FSS

NR

NR

4.0 1.8

NA

BL, baseline; ED, emergency department; FSS, fatigue severity scale; GFI, global fatigue inventory; LOC, loss of consciousness; MFIS, modied fatigue impact scale; MFI, multidimensional fatigue inventory; HI, head injury; TBI,
traumatic brain injury; mos, months; NR, not reported; NA, not applicable; PCSC, post-concussion syndrome checklist; POMS, prole of moods scale; PTA, post-traumatic amnesia; sev, severity; SF-36 V, 36-item short form health
survey vitality subscale (from medical outcomes study); TBI, traumatic brain injury; VAS, visual analogue scale; wks, weeks; y, year.

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

De Leon et al. (2009)

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

(a)

6d

14d
33

5%

7%10 73.3%15
2d

(b)

30 d
10

31

27

10% 11% 45.2%

10

709

90 d
30

32

8% 57% 74%

7092d
10

26

28

29

27

15

30

10% 21% 28% 31.4% 35.5% 37.1% 61%

23%31 61.1%15 68.1%29

8%31

28.2%29 45%30

30%28 45%30 47%32

7d

60 d

180d

360d

90 d

648d

33%44

80%41

42%33

16%36 27%44

60%41

35%36

180d

690d

1080d

Fig. 3. (a) Reported frequencies of fatigue for studies with time zero 1-month post-injury. (b) Reported frequencies of fatigue for studies with time zero >1-month post-injury.

for two TBI subgroups (De Leon et al., 2009). Other researchers who
utilized the same standardized scales could not be compared on
the basis of their measurement of fatigue at different time points,
or missing data on scores (Table 5).
3.11. Impact of fatigue after TBI
One study where baseline assessment took place prior to one
month post-injury and one with baseline assessment at three
months post-injury investigated consequences of fatigue.
Table 4 shows lists the consequences signicantly associated
with fatigue. One study of moderate quality looked at the relationship of fatigue with persistent post-concussive symptoms
(Norrie et al., 2010), and one of fair quality looked at its association with the Glasgow outcome scale-extended (GOSE) total score
(Sigurdardottir et al., 2009).
Persistent post-concussive symptoms: Fatigue severity at one
week and three months predicted persistent post-concussive
symptoms at three months and six months, respectively, controlling for litigation, psychological/neurological disorders, and
substance abuse (Table 4) (Norrie et al., 2010).
GOSE total score: An association with the GOSE score was
found in Sigurdardottir et al.s study controlling for education, PTA,
intracranial pathology and relevant psychological tests in multivariate regression analysis models. In the mild TBI group, the FSS
total score was a signicant predictor of GOSE score (R2 = 0.47;
p < .001), explaining 23% of the total variance in GOSE score at one
year post-injury. Similar results were obtained for the moderate to
severe TBI group (R2 = 0.58; p < .001) (Table 4).
3.12. Associations of fatigue with other clinically important
variables
3.12.1. Studies with baseline assessment up to one month
post-injury
Lundin et al. (2006) found poor memory, sleep disturbance
and fatigue to be most commonly reported within their sample,
with early symptom overlap correlated with later results. Similarly,
Meares et al. (2011) found symptom overlap of fatigue, insomnia,
and irritability at ve days and three months post-injury, with some

participants recovering from and others developing the symptoms


as time went on.
Norrie et al. found a signicant increase in the percentage of
those with fatigue reporting depression and/or anxiety, both symptoms over the cut off indicating mild severity, at six months after
injury, compared with reports at three months. This increase coincides with a leveling off of fatigue percentages. As fatigue becomes
persistent, psychological factors such as anxiety and depression
tend to worsen (Norrie et al., 2010).
Ponsford et al.s mTBI group reported signicantly poorer general health, vitality, and mental health, as demonstrated by their
scores in the corresponding subscales of the SF-36, compared to
trauma controls; however, a similar pattern was observed when
participants completed the same scales but with regard to their
pre-injury status. The authors highlighted the importance of documenting pre-injury status in TBI studies.
3.12.2. Studies with baseline assessment after one month
post-injury
Bushnik et al., investigating changes in fatigue from 6 to 12
months post-injury, reported a signicant change in the Pittsburgh
sleep quality index (PSQI) scores: where fatigue increased, PSQI
scores were higher compared to cases where there was no change
or decreased fatigue. There were no other signicant group differences on the pain VAS, disability rating scale, neurobehavioral
functioning inventory motor subscale, the Craig handicap assessment and reporting technique (CHART) cognitive independence,
and CHART occupation (Bushnik et al., 2008a). The authors suggested that, for TBI individuals who complain of fatigue, assessing
sleep quality would be a high-yield correlate and possibly treatable
with behavioral and/or medication interventions.
Kempf et al. reported no associations between fatigue parameters and TBI severity, alcohol intake at time of injury, nor with sleep
duration, education, age or gender. They did, however, nd a moderate correlation between FSS and depression symptoms assessed
with the Beck depression inventory (BDI) (r = 0.46, p = 0.001), and
with anxiety symptoms assessed with the Hospital anxiety and
depression scale (HADS) (r = 0.37, p = 0.007). They also reported
coincidence of fatigue and excessive daytime sleepiness (Kempf
et al., 2010).

710

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

3.13. Medications, drugs and alcohol


Seven of the 22 studies listed regular intake of psychoactive
drugs and/or history of drug or substance abuse prior to TBI in
study exclusion criteria (Mickeviciene et al., 2004; Norrie et al.,
2010; McLean et al., 1993; Jha et al., 2008; Kaiser et al., 2010; Khateb
et al., 2005; Schoenberger et al., 2001). One study excluded persons
taking medications that cause sleep/wake disturbances, however,
details were not provided (Kaiser et al., 2010). Eight of the 22 studies
did not report on use of medications/illicit drugs/alcohol by participants prior to or over the course of the study (Lidvall et al., 1974; De
Leon et al., 2009; Lundin et al., 2006; Yang et al., 2009; Sundstrom
et al., 2007; Hutchinson et al., 2009; Gemmell and Leathem, 2006;
Hou et al., 2012). Nine studies reported a variation of data. Bushnik
et al. (2008a) reported use of alcohol by 62% prior to the rst
assessment, with 19% identied as drug users and heavy and/or
binge drinkers. Sigurdardottir et al. (2009) described, at the second
assessment, use of alcohol more than once per month in 48% of mild
to moderate injured persons and in 27% of the severely injured. Use
of alcohol and/or drugs more than two or three times per week was
reported by 17%, and more than four times per week by 7% of participants (Sigurdardottir et al., 2009). Ponsford et al. (2012) included
breath alcohol levels at recruitment exceeding 0.05 mg/L of alcohol,
inuence of illicit drugs at injury, history of signicant drug/alcohol
abuse affecting daily functioning in their exclusion criteria. Those
reporting alcohol or cannabis use, and who did not have cognitive
difculties pre-injury were not excluded. The study reported frequency of lifetime substance abuse to be 31.3% in the mTBI sample
and frequency of substance abuse in the previous three months to
be 6.7% (Ponsford et al., 2012). Meares et al. (2011) reported use of
opioids/opiates by 59.7% across the rst and second assessments,
and use of marijuana was reported in 24.4% of mTBI cases. In Driver
and Edes study, 61.1% reported intake of selective serotonin reuptake inhibitors (SSRIs). One study indicated changes in medication
regimes of participants with exyx neurotherapy system treatment, but did not clarify the nature of the changes (Schoenberger
et al., 2001).

4. Discussion
4.1. Factors associated with fatigue
When we sought evidence of a temporal relationship between
clinically important factors and fatigue we focused on: (1) TBI population characteristics (e.g. time since injury, severity of injury,
comorbid conditions, etc.) and (2) our outcome of interest (i.e.
fatigue) its frequency, severity, and denition, with the goal of
obtaining a set of risk factors that can be used for prognosis. Table 4
presents a descriptive summary of the available evidence. In summary, several potential risk factors for fatigue in TBI have been
investigated, including those related to demographics and socioeconomic status, injury severity, medical comorbidities, baseline
fatigue levels, genetic makeup, and physical and cognitive independence.
Fatigue at baseline, occurring at any time from injury through
the acute care course, was found to be a primary predictor of symptom chronicity in TBI of varying severities (Norrie et al., 2010).
Baseline fatigue was found to be one of the most powerful predictors of fatigue at follow-up (De Leon et al., 2009; Norrie et al., 2010).
Other studies on chronic fatigue syndrome show similar associations between long-lasting fatigue and fatigue at baseline (Cairns
and Hotopf, 1997; Nisenbaum et al., 2003; Kato et al., 2006), concurrent with the results of another study, where pre-stroke fatigue
was reported to be related to fatigue in the acute phase after stroke
(Lerdal et al., 2011). Despite reports of an impact of baseline fatigue

on outcome at follow-up, the clinically important critical values


of fatigue severity and duration following brain injury are not established. Future research should record frequent and specic data in
investigation of the etiology of pre-morbid and baseline fatigue, and
control for multiple factor interactions, in addition to the magnitude of effect attributable to individual factors in analyses. More
attention needs to be paid to patients with intensive fatigue at
baseline, as it is related to prognosis.
Female sex, education, GCS score, and alcohol use at the time
of injury were reported to have no associative value for fatigue
severity in a study of mild TBI (Kempf et al., 2010). Differences in
severity and frequency of fatigue between men and women have
been observed after stroke (Lerdal et al., 2011), depression (Khan
et al., 2002), obstructive sleep apnea (Chervin, 2002), heart disease (Ekman and Ehrenberg, 2002), and cancer (Miaskowsky, 2004),
with females more often reporting fatigue than their male counterparts. Sex-related differences in fatigue were investigated in just
one reviewed study, and further research is warranted. Other factors associated with fatigue frequency and severity at follow-up,
reported reviewed studies, included carriage of the APOE 4 allele
(Sundstrom et al., 2007), counseling for mental health, medical
disability, specic marital status (i.e. widowed, divorced, or separated), and involvement in litigation (De Leon et al., 2009). The APOE
4 allele in persons with TBI was previously reported to be linked to
an increased risk of Alzheimers disease (Jellinger et al., 2001). In a
study of the general population (OHara et al., 2005), a relationship
between sleep apnea (SA) and dementia through the APOE 4 allele
was observed. Sleep apnea, highly prevalent in the TBI population
(Mollayeva et al., 2013b), may explain the link between the APOE
4 allele and fatigue and dementia. While studies to date have outlined the separate relationships between TBI, SA, fatigue, the APOE
4 allele, and dementia, their complex interaction requires rigorous
study.
4.2. Frequency, severity and course of fatigue in TBI
This systematic review underlines the variation in frequency of
fatigue that exists after TBI, regardless of studies set time zeros
(Fig. 3a and b). Changes in the proportion of participants reporting
fatigue from the start of the study to its completion also varied,
some reporting gradual or abrupt increases or decreases, and others reporting frequency uctuation over the course of study. A
steep drop in frequency was observed at two weeks post-injury,
with most of the contributing studies featuring mTBI patients. This
drop might be explained by the current clinical management of
mTBI, including the prescription of a rest period of at least two
weeks after injury. As fatigue is inuenced by the degree of physical and/or cognitive exertion, as well as the amount of rest one has
received, it is possible that a mildly injured person, after completing
a course of rest, would not perceive fatigue; however, their fatigue
could resume when they return to regular duties and responsibilities. Unchanged frequencies were commonly associated with RCTs
with close follow-up times (e.g. 13 months). The observed variation in the natural history of fatigue post TBI may be related to the
tools utilized by the different researchers, and the constructs those
tools measured. Respondents interpretations of the construct of
fatigue, as well as its complex underlying pathogenesis with different mechanisms inter-related at different time points, are expected
to inuence the results obtained.
The dimensions assessed in the studies utilizing fatigue scales or
single items included momentary perception, chronic perception,
the impact of fatigue on function, rating of tiredness, dimensions
of fatigue (i.e. mental, physical), or severity of the fatigue (Supplementary le 3). The various measures also attach different weights
to different aspects of fatigue, depending on the conceptualization
of fatigue by the developer (Chaudhuri and Behan, 2004). In some

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

studies, fatigue was conceptualized as a one-dimensional entity in


which persons are deemed either fatigued or not fatigued based
solely on their perception of the experience at the time of completion.
Interpretation of scale items by the respondent can be signicantly confounded by the association of fatigue with other
symptoms, particularly apathy, excessive sleepiness, depression,
lack of motivation, anxiety, litigation and cognitive dysfunction. In
TBI patients, fatigue was reported to be associated with depression;
moreover, in the regression analyses in one reviewed study (Norrie
et al., 2010), the severity of fatigue was predicted by depression.
Kempf et al. (2010) reported that fatigue and excessive daytime
sleepiness coincided in their sample. Excessive daytime sleepiness
may be an indicator of central nervous system (CNS) pathology
due to brain injury (i.e. hypocretin/orexin deciency), as well as
related to quantity and quality of sleep (Mollayeva et al., 2013b;
Baumann, 2012; Nardone et al., 2011; Mathias and Alvaro, 2012;
Siebern and Guilleminault, 2012). While Kempf et al. (2010) studied and did not uncover a relationship between fatigue and sleep
duration, quality of sleep was not investigated. A number of sleep
disorders (i.e. sleep-related breathing disorder, periodic leg movement disorder, etc.) highly prevalent post-TBI (Nardone et al., 2011;
Mollayeva et al., 2013c) are characterized by frequent arousals,
which generally result in fragmented sleep, which can produce daytime sleepiness (Stepanski, 2002). Bushnik et al. (2008a) suggests
assessment of sleep quality as a valuable measure when studying
TBI patients with fatigue complaints.
Symptoms of fatigue and cognitive dysfunction have been
reported to overlap in persons with TBI (Johansson et al., 2009;
Zaben et al., 2013). This can potentially inuence accuracy of selfreport, as a person with cognitive dysfunction may not be able
to fully grasp the changes in fatigue since their injury, as well
as its impact on daily functioning, as required in completion of
certain self-report measures. Bushnik et al. (2008a) reported that
a subset of individuals who experienced signicant increase in
fatigue over the rst two years post-injury demonstrated poorer
outcomes in cognition, motor symptoms, and general functioning
compared to those with decreased or stable fatigue (Bushnik et al.,
2008a). A separate study had similar ndings subjective mental fatigue following brain injury was correlated with objectively
measured information processing speed (Johansson et al., 2009). In
other literature again, post-traumatic conditions such as hypopituitarism have been reported to have a wide range of manifestations,
including fatigue, myopathy, cognitive difculties, depression, and
behavioral changes (Zaben et al., 2013). The same degree of fatigue,
therefore, will not be perceived with equal intensity by persons
with different comorbid conditions or fatigue etiology. Moreover,
fatigue manifestation is thought to be differentially modulated by
a variety of factors within and between TBI persons with time.
Distinguishing fatigue as a result of TBI from fatigue associated
with comorbid conditions (i.e. depression, pain, anxiety, apathy,
sleep dysfunction, medication effect, etc.) is a complicated task. As
such, future research should consider use of additional measures
for common comorbidities when assessing PTF.
Fatigue severity (i.e. mean FSS scores) was higher in persons
with TBI than previously reported for healthy adults (2.3 0.7)
(Krupp et al., 1989), but lower than those in patients with systemic
lupus erythematosus (SLE) (4.7 1.5) (LaChapelle and Finlayson,
1998), rheumatoid arthritis (4.2 1.2) (Krupp et al., 1989) and psoriatic arthritis (6.9 2.4) (Cella et al., 2005). Bushnik et al. (2008a)
reported FSS scores obtained at 6, 12, and 1824 months postinjury, all falling within the score range for non-fatigued control
subjects (Table 5).
Studies differed in reports of fatigue severity over time, with
some noting changes with and others stability. It is plausible
that time since injury is a determinant of effectiveness of coping

711

strategies and thereby perception of symptom severity. A study


of persons with chronic fatigue syndrome (Brown et al., 2010)
reported better adaptive coping strategies with longer disease
duration. Alternatively, spinal cord injury patients showed no
changes in coping styles over time (Craig et al., 1994). Future longitudinal studies of coping by persons who sustained a TBI may
provide greater insight. Age differences between samples should
also be considered. In a study of fatigue in the general population,
Cella et al. (2002) reported that people older than 50 years in the
described more severe fatigue than the younger population. The
mean age in samples of reviewed studies reporting fatigue severity ranged from 20.1 1.8 (Hutchinson et al., 2009) to 45.7 10.8
(Gemmell and Leathem, 2006). We did not observe relationship
between age and severity of reported fatigue, however (Table 5).
Other factors related to the discussion of fatigue severity have to
do with the impact of brain injury on a persons ability to perform
pre-morbid duties and manage responsibilities. Diminished activity due to changes in lifestyle, with subsequent loss of muscle tone
and weakness, or muscle weakness due to neurological impairment, can result in greater fatigue perception associated with mild
activity (Chaudhuri and Behan, 2004).
4.3. Consequences of fatigue in TBI
Possible consequences of fatigue emerged in the studies
reviewed. Fatigue severity one week post-injury was associated
with persistent post-concussive symptoms at three months (Norrie
et al., 2010), and the FSS total score was signicantly associated
with the GOSE score for all severities of TBI (Sigurdardottir et al.,
2009). Post-concussion syndrome (PCS) itself refers to a group of
symptoms, including headache, dizziness, fatigue, and affective
and cognitive changes, that may be reported by patients after TBI
(McAllister, 1994). Thus, it is possible that severe post-concussive
symptoms that are not resolved over a short period (i.e. three
months), inuence fatigue outcomes. The fact that the GOSE, the
gold standard for assessing patient outcomes after TBI (Shukka
and Devi, 2011), was affected by baseline fatigue severity across all
injury severities at one year post-injury is signicant, as it suggests
that fatigue can be long lasting, with a low likelihood of resolution.
Consequently, diagnostic efforts that consider diverse factors and
comorbid conditions (Figs. 1 and 2) should be implemented in the
very early stages post-injury.
4.4. Medication effects
CNS depressants can cause or increase fatigue (Liska, 2008). In
the reviewed studies, just seven included their use in the exclusion criteria. Nine studies provided some information on the use
of medications/illicit drugs/alcohol by participants prior to or over
the course of the study. None, however, considered the potential
confounders in this relationship. Reported use of alcohol by 62%
of participants in the period before the rst assessment, with 19%
identied as drug users, heavy and/or binge drinkers is striking
(Sigurdardottir et al., 2009). Ethyl alcohol is a CNS depressant, and
the injured brain is particularly sensitive to its effects at the highest
centers (i.e. speech, thought, cognition) and lower brain functions
(i.e. spinal cord reexes, respiration), as the dosage increases (Liska,
2008). Norrie et al. (2010) reported that alcohol intake prior to the
injury was not correlated with fatigue severity at three months
after injury, as measured by FSS, however, the researchers did not
report alcohol intake of participants throughout the course of the
study. This is signicant, as studies in the general population have
reported fatigue to be the most severe hangover symptom (Penning
et al., 2012; Rohsenow et al., 2007).
Intake of SSRIs was reported by 61.1% of the participants in
Driver and Edes study. While this class of medications is a rst

712

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

line of treatment for depression following TBI (World Health


Organization, 2002), some drugs within this class (i.e. uoxetine and paroxetine) may be problematic due to their adverse
effects, including those related to fatigue and cognitive function
(Schmitt et al., 2001). Another reviewed study reported use of opioids/opiates by 59.7% of participants (Meares et al., 2011). Opium
alkaloids are narcotic analgesics and narcosis is dened by depression of the CNS leading to analgesia, drowsiness, changes in mood,
mental clouding, lethargy, apathy and subsequent unconsciousness
(Shukka and Devi, 2011). While currently there is no strong evidence directly relating physical and mental fatigue in TBI to side
effects of opiates and opiods (Chapman, 2002; Leong and Royal,
2004), data on its safety for chronic use is also lacking (Rhodes,
2012). While our discussion of medication effects and fatigue in
TBI is limited, given the complexity of the fatigue symptom and
incomplete data available, future research should consider such
effects, as the potential of medications to cross the bloodbrain barrier and mimic neurological decits and cause or exacerbate PTF, is
real (Daneman, 2012; Maher et al., 2011).
To complete our discussion, we follow with recommendations
for future research in the eld of fatigue and TBI. As mentioned,
along with the confounding effects and selection bias, the method
by which fatigue was measured contributed signicantly to the
variation observed in results. The words that one uses to dene
fatigue can be vague, especially if the reporter (i.e. patient with TBI)
has additional complaints related to constructs such as excessive
sleepiness and impaired alertness. As such, separate assessment
of each construct is preferable. When featured as one item within
a self-report measure, even when spontaneously endorsed and
ranked as the most important symptom, patients may rank their
fatigue experience understanding it as being exhausted, tired,
weak, while others may feel physically exhausted but mentally
alert. As such, a single question hampers interpretation of the score.
While it is not always the case that multi-item instruments are
more valid than a single item, especially if the global opinion of
the patient is of interest, adding one global item about the construct to a multi-item symptom measure in the future can help in
the interpretation and validation of the instrument in the population of interest. This is particularly relevant to the study fatigue
in the TBI population as, despite the number of multi-item standardized measures that have been utilized, only the FSS, the MFIS
and the SF-36 have been partially validated against other fatigue
measures in a TBI sample. Moreover, there are no psychometric
data on the responsiveness of these measures, implying limited
understanding of how much error exists when measuring changes
in fatigue over time. Currently, the eld of TBI requires further testing of existing self-report measures whose psychometric properties
were described in other target populations, focusing on measures
pertaining to the multidimensional etiology and state of PTF.
None of the studies reviewed applied technologies (i.e.
electroencephalography, functional magnetic resonance imaging,
magnetic resonance spectroscopy, regional brain volumes, motor
evoked potential, etc.) or markers of physiological processes
(i.e. function of hypothalamicpituitaryadrenal axis, autonomic
nervous system response, metabolic processes, immune system
response, etc.) to study the fatigue experienced by individuals with
TBI. The latter is important, as research has shown that the resting
pulmonary and cardiorespiratory function in patients with TBI is
compromised (Jankowski and Sullivan, 1990). This can be related
to deconditioning as a result of a more sedentary lifestyle (Giordon
et al., 1998). In a study of maximal physiologic responses during exercise in patients with moderate to severe TBI at 17.2 17
months after injury, several weeks of an exercise training program reduced physiologic fatigue (Bhambhani et al., 2005). Others
reported that aerobic tness in individuals with TBI enhanced cognition and improved mood (Carroll et al., 2004b; Cassidy et al.,

2014). This is extremely important as fatigue perception ratings


were found to be higher in patients with depression (Norrie et al.,
2010). Again another study (Jankowski and Sullivan, 1990) reported
that a 16-week circuit training program of moderate intensity and
prolonged duration increased TBI patients oxidative capacity and
muscular endurance and the index of physiologic fatigability was
shown to be useful for the assessment and evaluation of individuals with TBI. Similarly, in a reviewed study by Driver and Ede
(2009) fatigue elimination was reported after an eight-week group
aquatic program, with no changes in fatigue in the control group.
Thus, further study that accounts for the physiologic, objective performance, and/or homeostatic changes with regard to increased
perception/manifestation of fatigue after brain injury is within the
top priorities for future research.
4.5. Limitations
We acknowledge heterogeneity in the primary studies with
respect to sample characteristics (i.e. age, injury/localization of
injury, time since injury) and fatigue denitions. Another concern
related to the reviewed studies, largely of moderate quality, is
that severe TBI is underrepresented in the inception cohorts and the
evidence for the second and third research questions of this review
was based largely on mild TBI cohorts. Additionally, the majority of
the patients in the studies were men, which limits the precision of
estimates of predictors and consequences for fatigue in severe TBI,
especially for women (Table 2).
Most studies focused largely on the fatigue symptom; the
strength and signicance of associations with other factors (e.g.
sleep, other medical conditions, medication use or clinically important symptoms such as alertness, sleepiness) were often not
reported. Thus, the roles of other factors could be underestimated
in this review.
The focus of this review was the natural history of fatigue in
patients with TBI. To be consistent with our protocol (Mollayeva
et al., 2013a), results from all selected longitudinal studies were
used to address the rst research question (i.e. natural history of
fatigue). Since baseline fatigue assessment was performed at different times since injury, we attempted to mitigate zero-time effect
by reporting results with baseline assessments up to one month
post-injury and after one month, separately. Nevertheless, generalizability of results remains unclear due to inadequate reporting of
selection criteria, poor control of confounding effects, and attrition.
There are limitations to the presented data on fatigue measures
used in the reviewed studies (Supplementary le 3). For conciseness, properties of the measures, specically those related to
psychometrics, were not reported in great detail. Despite attempts
to include all relevant articles for their use in the TBI population, it
is possible studies were missed.
All articles included in this review are peer-reviewed. As such,
there is possibility for publication bias. Finally, the inclusion of only
English language articles could affect the generalizability of our
ndings.
4.6. Pitfalls and controversies
Despite the existence of clinical criteria for the diagnosis of PCS,
the self-reported nature of nonspecic symptoms such as fatigue
can be confounded by other factors (i.e. psychological distress, pain,
depression, etc.). This may be particularly apparent in patients with
insurance claims that are being disputed. Their need to provide
proof of disability may magnify fatigue symptoms and result in controversy about whether symptoms are indicators of brain injury
or are of behavioral origin (Carroll et al., 2004b; Cassidy et al.,
2014). Also, fatigue as a symptom is nonspecic to TBI. Fatigue
appears with other diagnostic labels in other clinical specialities

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

713

Fig. 4. Model for studying fatigue symptoms in patients with traumatic brain injury.

for example bromyalgia, chronic fatigue syndrome, endocrine


disorders, and patients with psychiatric illness (Norrie et al., 2010;
Cairns and Hotopf, 1997; Nisenbaum et al., 2003; Kato et al., 2006;
Lerdal et al., 2011; Cassidy et al., 2014). Previous systematic reviews
on the epidemiology, diagnosis, prognosis, treatment and costs of
mTBI raised the issue of specicity of self-reported symptoms such
as headache, fatigue, cognitive decits to mild TBI (Carroll et al.,
2004b; Cassidy et al., 2014), with the recommendation to replace
the term post-concussion syndrome with the term post-traumatic
symptoms.
The reviewed studies allowed comparison of fatigue severity
only between mild TBI participants and controls, limiting our discussion to a single severity of injury. While concussed athletes
had more fatigue compared to healthy controls at one week postinjury, that was not the case at two weeks (Hutchinson et al.,
2009). Similarly, fatigue was greater in the mild TBI group compared to controls shortly after the injury, but not at three months

post-injury in Lundin et al.s and Meares et al.s samples. De Leon


et al. (2009) found that fatigue severity at the one year follow-up
was not associated with the type of injury (i.e. mild TBI vs. non-head
injury) in a fully adjusted model. Pair-wise comparisons showed
lower fatigue scores in the mild TBI group at 12 months compared
to the other injury group. Contrariwise, Sundstrom et al. (2007)
reported that their mild TBI group had less fatigue pre-injury and
more post-injury compared to age-, sex-, and education-matched
controls.
Given this lack of specicity of the fatigue symptom, this topic
is perplexing and time consuming. An accurate investigation of
fatigue in TBI must begin with a clear denition of the most common symptomatic descriptor, feeling fatigued. Next, the cause of
fatigue must be determined and a diagnosis established. Although
the pathophysiology of fatigue after TBI is still poorly understood, the goal is to determine whether the fatigue is caused
by a correctable factor (i.e. depression, endocrine dysfunction,

714

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

deconditioning, poor sleep, etc.) so that interventions are applied


appropriately. Fig. 4 illustrates the proposed algorithm for study of
PTF.
5. Conclusions
Fatigue is a common symptom post TBI. Its frequency may
change over time, but fatigue can persist years after the injury.
This may be related to pre-morbid/early fatigue, mental health
issues, other medical conditions, and ongoing societal stressors.
Clinicians seeing patients with TBI at the acute stages post-injury
with high early fatigue intensity, mental health issues, and litigation involvement should be aware that these may be associated
with the development of persistent post-concussive symptoms.
The available evidence on the associative value of these factors, as
well as the consequences of fatigue, is currently not very strong,
as we found just three cohort studies addressing these issues.
More research is needed to establish associations between fatigue
and other clinically important pre- and post-morbid variables (i.e.
sleep dysfunction, depression, physical and cognitive impairments,
other medical/neurological disorders), and their impact on outcomes post-injury. Medication effects, personal factors such as
coping ability, physical deconditioning, stress level, and time factors should also be investigated. This is particularly important for
translation of research into clinical practice, in order to address
risk factors and course of condition. An international consensus,
similar to the National Institutes of Health and developed for rehabilitation in TBI in 1999 (Consensus Conference, 1999) advising on
how best to study clinically important symptoms such as fatigue in
TBI, is of utmost importance. In particular, there needs to be a consensus on the denition of pathological PTF, set times for baseline
assessment, recognizing the challenges in studying the symptom
in moderate-severe brain injury at time zero, clinically relevant
period of follow-up, acceptable attrition rates to ensure representative samples, and validated measures of outcome, all of which can
reduce heterogeneity of results. What will be left to focus on then
is the variety of lesions from TBI (i.e. white or gray matter, specic tract damage, lesion volume, localization of injury, etc.) and
inter-individual variability in perception and multifactorial fatigue
etiology, which may nd study of individual patients best. A caveat
to this point is that case-reports of patients whose symptoms and
clinical course do not t the typical picture, may lead to scientic
progress in the understanding of and appreciation for the complexity of the fatigue symptom post TBI (Yennurajalingam and Bruera,
2007).
Authors contributions
Tatyana Mollayeva had full access to all of the data in the
study and takes responsibility for the integrity of the data and the
accuracy of the data analysis. Study concept and design: Tatyana
Mollayeva, Colin M Shapiro, Angela Colantonio, J David Cassidy.
Acquisition of data: Tatyana Mollayeva, Tetyana Kendzerska, Shirin
Mollayeva. Analysis and interpretation of data: All authors. Drafting of the manuscript: Tatyana Mollayeva. Critical revision of the
manuscript for important intellectual content: All authors. Statistical analysis: Tatyana Mollayeva, Shirin Mollayeva. Administrative,
technical, and material support: Shirin Mollayeva. Study supervision: Angela Colantonio, Colin M Shapiro, J David Cassidy.
Acknowledgements
Our study had no external funding source. The rst author
was supported by 2012/2013 Toronto Rehabilitation Institute
Scholarship, the Ontario Graduate Scholarship 2012/2013 and the

2013/2015 Frederick Banting and Charles Best Doctoral Research


Award from the Canadian Institutes of Health Research. We recognize the support of the Toronto Rehabilitation Institute Foundation
and a grant to the Ministry of Health and Long Term Care
to the Toronto Rehabilitation Institute. Support was also provided through the Ontario Work Study Program. We gratefully
acknowledge the involvement of Ms. Jessica Babineau, information
specialist at the Toronto Rehabilitation Institute for her help with
the literature search. The authors have no conict of interest to
declare pertaining to this review.

Appendix A. Supplementary data


Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.neubiorev.
2014.10.024.
References
Bateman, A., Culpan, F.J., Pickering, A.D., Powell, J.H., Scott, O.M., Greenwood, R.J.,
2001. The effect of aerobic training on rehabilitation outcomes after recent
severe brain injury: a randomized controlled evaluation. Arch. Phys. Med. Rehabil. 82, 174182.
Baumann, C.R., 2012. Traumatic brain injury and disturbed sleep and wakefulness.
Neuromol. Med. 14 (3), 205212.
Baumann, C.R., Werth, E., Stocker, R., Ludwig, S., Bassetti, C.L., 2007. Sleep-wake
disturbances 6 months after traumatic brain injury: a prospective study. Brain.
130 (Pt 7), 18731883.
Belmont, A., Agar, N., Hugerom, C., Gallais, B., Azouvi, P., 2006. Fatigue and traumatic
brain injury. Ann. Readapt. Med. Phys. 49 (6), 283288.
Bhambhani, Y., Rowland, G., Farag, M., 2005. Effects of circuit training on body
composition and peak cardiorespiratory responses in patients with moderate to severe traumatic brain injury. Arch. Phys. Med. Rehabil. 86 (2),
268276.
Bhambhani, Y., Rowland, G., Farag, M., 2008. Reliability of peak cardiorespiratory
responses in patients with moderate to severe traumatic brain injury. Arch. Phys.
Med. Rehabil. 84, 16291636.
Borgaro, S.R., Baker, J., Wethe, J.V., Prigatano, G., Kwasnica, C., 2005. Reports of fatigue
during early recovery from traumatic brain injury. J. Head Trauma Rehabil. 20
(5), 416425.
2013. Brain Injury Association of America About Brain Injury, Retrieved from:
http://www.biausa.org/about-brain-injury.htm (26.11.13).
Brown, M.M., Brown, A.A., Jason, L.A., 2010. Illness duration and coping style in
chronic fatigue syndrome. Psychol. Rep. 106, 383393.
Bushnik, T., Englander, J., Wright, J., 2008a. Patterns of fatigue and its correlates over
the rst 2 years after traumatic brain injury. J. Head Trauma Rehabil. 1, 2532.
Bushnik, T., Englander, J., Wright, J., 2008b. The experience of fatigue in the rst 2
years after moderate-to-severe traumatic brain injury: a preliminary report. J.
Head Trauma Rehabil. 23 (1), 1724.
Cairns, R., Hotopf, M., 1997. A systematic review describing the prognosis of chronic
fatigue syndrome: a systematic review. GJM 90, 223233.
Cantor, J.B., Ashman, T., Gordon, W., Ginsberg, A., Engmann, C., et al., 2008. Fatigue
after traumatic brain injury and its impact on participation and quality of life. J.
Head Trauma Rehabil. 23 (1), 4151.
Carroll, L.J., Cassidy, J.D., Peloso, P., Garritty, C., Giles-Smith, L., 2004a. Systematic
search and review procedures: results of the WHO Collaborating Centre Task
Force on Mild Traumatic Brain Injury. J. Rehabil. Med. 2004 (Suppl. 43), 1114.
Carroll, L.J., Cassidy, J.D., Peloso, P.M., Borg, J., von Holst, H., Holm, L., et al., WHO,
2004b. Prognosis for mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J. Rehabil. Med. 43
(Suppl.), 84105.
Cassidy, J.D., Cancelliere, C., Carroll, L.J., Ct, P., Hincapi, C.A., et al., 2014. Systematic review of self-reported prognosis in adults after mild traumatic brain
injury: results of the international collaboration on mild traumatic brain injury
prognosis. Arch. Phys. Med. Rehabil. 95 (3 Suppl.), S132S151.
Cella, I.D., Lai, J.S., Chang, C.H., Petermabn, A., Slavin, M., 2002. Fatigue in cancer
patients compared with fatigue in the general United States population. Cancer
94, 528538.
Cella, D., Yount, S., Sorensen, M., 2005. Validation of the functional assessments of
chronic illness therapy fatigue scale relative to other instrumentation in patients
with rheumatoid arthritis. J. Rheumatol. 32, 811819.
Chapman, S., 2002. Effects of intermediate and long term use of opioids on cognition
in patients with chronic pain. Clin. J. Pain 18, S83S90.
Chaudhuri, A., Behan, P.O., 2004. Fatigue in neurological disorders. Lancet 363
(9413), 978988.
Chervin, R., 2002. Sleepiness, fatigue, tiredness, and lack of energy in obstructive
sleep apnea. Chest 118, 372379.
Consensus Conference, 1999. Rehabilitation of persons with traumatic brain injury.
JAMA 282 (10), 974983.

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716


Cooper, J., Reynolds, F., Bateman, A., 2009. An evaluation of a fatigue management
intervention for people with acquired brain injury: an exploratory study. Br. J.
Occup. Ther. 72 (4), 174179.
Craig, A.R., Hancock, K., Chang, E., 1994. The inuence of spinal cord injury on coping
styles and self-perceptions two years after the injury. Aust. N. Z. J. Psychiatry 28,
307312.
Daneman, R., 2012. The bloodbrain barrier in health and disease. Ann. Neurol. 72
(5), 648672.
De Leon, M.B., Kirsch, N.L., Maio, R.T., Tan-Schriner, C.U., Millis, S.R., et al., 2009.
Baseline predictors of fatigue 1 year after mild head injury. Arch. Phys. Med.
Rehabil. 90, 956995.
Driver, S., Ede, A., 2009. Impact of physical activity on mood after TBI. Brain Inj. 23
(3), 203212.
Ekman, I., Ehrenberg, A., 2002. Fatigue in chronic heart failure does gender make
a difference? Eur. J. Cardiovasc. Nurs. 1, 7782.
Finsterer, J., Mahjoub, S.Z., 2013. Fatigue in healthy and diseased individuals. Am. J.
Hosp. Palliat. Care. 31 (5), 562575.
Gemmell, C., Leathem, J.M., 2006. A study investigating the effects of Tai Chi Chuan:
individuals with traumatic brain injury compared to controls. Brain Inj. 20 (2),
151156.
Giobbie-Hurder, A., Gelber, R.D., Regan, M.M., 2013. Challenges on guarantee-time
bias. J. Clin. Oncol. 31 (23), 29632969.
Giordon, W.A., Sliwinski, M., Echo, J., McLoughlin, M., Sheerer, M., et al., 1998. The
benet of exercise in individuals with traumatic brain injury: a retrospective
study. J. Head Trauma Rehabil. 13 (4), 5867.
Haboubi, N.J., Long, J., Koshy, M., Ward, A.B., 2001. Short-term sequelae of minor
head injury: 5 years experience of minor head injury clinic. Disability Rehabil.
23 (14), 635638.
Hayden, J.A., Cote, P., Bombardier, C., 2006. Evaluation of the quality of prognosis
studies in systematic reviews. Ann. Intern. Med. 144, 427437.
Hillier, S.L., Sharpe, M.H., Merzer, J., 1997. Outcomes 5 years post-traumatic brain
injury with further reference to neurophysical impairment and disability. Brain
Inj. 11 (9), 661675.
Hou, R., Moss-Morris, R., Peveler, R., Mogg, K., Bradley, B.P., et al., 2012. When a
minor head injury results in enduring symptoms: a prospective investigation of
risk factors for postconcussional syndrome after mild traumatic brain injury. J.
Neurol. Neurosurg. Psychiatry 83, 217223.
Hutchinson, M., Mainwaring, L.M., Comper, P., Pichards, D.W., Bisschop, S.M., 2009.
Differential emotional responses of varsity athletes to concussion and musculoskeletal injuries. Clin. J. Sport Med. 19 (1), 1319.
Jankowski, L.W., Sullivan, S.J., 1990. Aerobic and neuromuscular training effect on
the capacity, efciency, and fatigability of patients with traumatic brain injury.
Arch. Phys. Med. Rehabil. 71 (7), 500504.
Jason, L.A., Evans, M., Brown, M., Porter, N., 2010. What is fatigue? Pathological and
nonpathological fatigue. PM&R 2 (5), 327331.
Jellinger, K.A., Paulus, W., Wrocklage, C., Litvan, I., 2001. Effects of closed traumatic
brain injury and genetic factors on the development of Alzheimers disease. Eur.
J. Neurol. 8, 707710.
Jha, A., Weintraub, A., Allschouse, A., Morey, C., Cusick, C., et al., 2008. A randomized
trial of modanil for the treatment of fatigue and excessive daytime sleepiness
in individuals with chronic traumatic brain injury. J. Head Trauma Rehabil. 23
(1), 5263.
Johansson, B., Berglund, P., Ronnback, L., 2009. Mental fatigue and impaired information processing after mild and moderate traumatic brain injury. Brain Inj. 23,
1314.
Kaiser, P.R., Valko, P.O., Werth, E., Thomann, J., Meier, J., et al., 2010. Modanil
ameliorates excessive sleepiness after traumatic brain injury. Neurology 75,
17801785.
Kato, K., Sullivan, P.F., Evengard, B., Pedersen, N.L., 2006. Premorbid predictors of
chronic fatigue. Arch. Gen. Psychiatry 63, 12671272.
Kempf, J., Werth, E., Kaiser, P.R., Bassetti, C.L., 2010. Sleep-wake disturbances 3 years
after traumatic brain injury. J. Neurol. Neurosurg. Psychiatry 81, 14021405.
Khan, A.A., Gardner, C.O., Prescott, A.C., Kendler, S.K., 2002. Gender differences in
the symptoms of major depression in opposite-sex dizygotic twin pairs. Am. J.
Psychiatry 159, 14271429.
Khateb, A., Atmann, J., Annoni, J.-M., Diserens, K., 2005. Cognition-enhancing effects
of donepezil in traumatic brain injury. Eur. Neurol. 54, 3945.
Kim, S.H., Manes, F., Kosier, T., Baruan, S., Robinson, R., 1999. Irritability following
traumatic brain injury. J. Nervous Mental Dis. 187, 327335.
Kluger, B.M., Krupp, L.B., Enoka, R.M., 2013. Fatigue and fatigability in neurologic
illnesses: proposal for a unied taxonomy. Neurology 80 (4), 409416.
Krupp, L.B., laRossa, N.G., Muir-Nash, J., Steinberg, A.D., 1989. The fatigue severity
scale application to patients with multiple sclerosis and systemic lupus erythematosus. Arch. Neurol. 46 (10), 11211123.
LaChapelle, D.L., Finlayson, M.A.J., 1998. An evaluation of subjective and objective
measures of fatigue in patients with brain injury and healthy controls. Brain Inj.
12, 649659.
Leong, M., Royal, M., 2004. Opioid therapy in chronic non-cancer pain management.
Pract. Pain Manag., 4347.
Lerdal, A., Bakken, L.N., Rasmussen, E.F., Bejermann, C., Ryen, S., et al., 2011. Physical
impairment, depressive symptoms and pre-stroke fatigue are related to fatigue
in the acute phase after stroke. Disabil. Rehabil. 33 (4), 334342.
Levine, L., Greenwald, B.D., 2009. Fatigue in Parkinson Disease, stroke, and traumatic
brain injury. Phys. Med. Rehabil. Clin. N. Am., 20.
Lidvall, H., Linderoth, B., Norlin, B., 1974. Causes of post-concussional syndrome.
Acta Neurol. Scand. 50 (Suppl. 56), 3144.

715

Liska, K., 2008. Drugs and the Human Body. Prentice Hall.
Lundin, A., De Boussard, C., Edman, G., Borg, J., 2006. Symptoms and disability until
3 months after mild TBI. Brain Inj. 20 (8), 799806.
Maher, A.R., Maglione, M., Bagley, S., Suttorp, M., Hu, J.H., Ewing, B., Wang, Z., Timmer, M., Sultzer, D., Shekelle, P.G., 2011. Efcacy and comparative effectiveness
of atypical antipsychotic medications for off-label uses in adults: a systematic
review and meta-analysis. JAMA 306 (12), 13591369.
Mathias, J.L., Alvaro, P.K., 2012. Prevalence of sleep disturbances, disorders, and
problems following traumatic brain injury: a meta-analysis. Sleep Med. 13 (7),
898905.
McAllister, T.W., 1994. Mild traumatic brain injury and the postconcussive syndrome. In: Silver, J.S., Yudofsky, S.C., Hahles, R.E. (Eds.), The Neuropsychiatry
of Traumatic Brain Injury. American Psychiatric Press, Washington, DC, pp.
357392.
McCrimmon, S., Oddy, M., 2006. Return to work following moderate-to-severe traumatic brain injury. Brain Inj. 20 (10), 10371046.
McLean, A., Dikmen, S.S., Temkin, N.R., 1993. Psychosocial recovery after head injury.
Arch. Phys. Med. Rehabil. 74, 10411046.
Meares, S., Shores, E.A., Taylor, A.J., Batchelor, J., 2011. The prospective course of postconcussion syndrome: the role of mild traumatic brain injury. Neuropsychology
25 (4), 454465.
Miaskowsky, C., 2004. Gender differences in pain, fatigue, and depression in patients
with cancer. J. Natl. Cancer Inst. Monogr. 32, 132143.
Mickeviciene, D., Schrader, H., Obelieniene, D., Surkiene, D., Kuniskas, R., et al., 2004.
A controlled prospective inception cohort study on the post-concussion syndrome outside the medicolegal context. Eur. J. Neurol. 11, 411419.
Middleboe, T., Andersen, H.S., Birket-Smith, M., Friis, M.L., 1992. Minor head injury:
impact on general health after 1 year, a prospective follow-up study. Acta Neurol.
Scand. 85 (1), 59.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., 2009. Preferable reporting items for
systematic reviews and meta-analyses for PRISMA statement. PLoS Med., 6.
Mollayeva, T., Kendzerska, T., Mollayeva, S., Shapiro, C.M., Colantonio, A., Cassidy,
J.D., 2013a. Fatigue in adults with traumatic brain injury: predictors and consequences. A systematic review of longitudinal study protocols. Syst. Rev. 11 (2),
25.
Mollayeva, T., Colantonio, A., Mollayeva, S., Shapiro, C.M., 2013b. Screening
for sleep dysfunction after traumatic brain injury. Sleep Med. 14 (12),
12351246.
Mollayeva, T., Kendzerska, T., Mollayeva, S., Colantonio, A., 2013c. Sleep apnea in
traumatic brain injury: understanding its impact on executive function. J. Sleep
Disorders Ther. 2, 129.
Nardone, R., Bergmann, J., Kunz, A., Caleri, F., Seidl, M., et al., 2011. Cortical excitability changes in patients with sleep-wake disturbances after traumatic brain
injury. J. Neurotrauma 28 (7), 11651671.
Nisenbaum, R., Jones, J., Unger, E., Reyes, M., Reyes, W.C., 2003. A population based
study of the chronic fatigue syndrome. Health Quality Life Outcomes 1, 49.
Norrie, J., Heitger, M., Leathem, J., Anderson, T., Jones, R., et al., 2010. Mild traumatic
brain injury and fatigue: a prospective longitudinal study. Brain Inj. 24 (1314),
15281538.
OHara, R., Schrodek, C., Kraemer, H., Kryla, N., Cao, C., 2005. Nocturnal sleep
apneahypopnea is associated with lower memory performance in APOE
epsilon4 carriers. Neurology 65, 642644.
Olver, J.H., Ponsford, J.L., Curran, C.A., 1996. Outcome following traumatic brain
injury: a comparison between 2 and 5 years after injury. Brain Inj. 10 (11),
841848.
Penning, R., McKinney, A., Verster, J.C., 2012. Alcohol hangover symptoms and their
contribution to the overall hangover severity. Alcohol Alcohol. 47 (3), 248252.
Ponsford, J., Cameron, P., Fitzgerald, M., Grant, M., Mickocka-Walus, A., 2011. Longterm outcomes after uncomplicated mild traumatic brain injury: a comparison
with trauma controls. J. Neurotrauma 28, 937946.
Ponsford, J.L., Ziino, C., Parcell, D.L., Shekleton, J.A., Roper, M., et al., 2012. Fatigue
and sleep disturbance following traumatic brain injury their nature, causes,
and potential treatments. J. Head Trauma Rehabil. 27 (3), 224233.
Prins, J.B., van der Meer, J.W., Bleijenberg, G., 2006. Chronic fatigue syndrome. Lancet
367, 346.
Rees, R.J., Bellon, M.L., 2007. Postconcussion syndrome ebb and ow: longitudinal
effects and management. Neurorehabilitation 22, 229242.
Rhodes, C., 2012. Opioids: is it time to change? Pain Pract. 22 (1), 4346.
Rohsenow, D.J., Howland, J., Minsky, S.J., Greese, J., Almeida, A., et al., 2007. The
acute hangover scale: a new measure of immediate hangover symptoms. Addict.
Behav. 32 (6), 13141320.
Schillings, M.L., Kalkman, J.S., Lannsen, H.H., van Engelen, B.M., Bleijenberg, G., et al.,
2007. Experienced and physiological fatigue in neuromuscular disorders. Clin.
Neurophysiol., 292300.
Schmitt, J.A., Kruizinga, M.J., Riedel, W.J., 2001. Non-serotonergic pharmacological proles and associated cognitive effects of serotonin reuptake inhibitors.
J. Psychopharmacol. 15 (3), 173179.
Schoenberger, N.E., Shieett, S.C., Esty, M.L., Ochs, R.J., Matheis, R.J., 2001. Flexyx
neurotherapy system in the treatment of traumatic brain injury: an initial evaluation. J. Head Trauma Rehabil. 16 (3), 260274.
Shukka, D., Devi, B.I., 2011. Outcome measures for traumatic brain injury. Clin. Neurol. Neurosurg. 113, 435441.
Siebern, A.T., Guilleminault, C., 2012. Sleepiness and fatigue following traumatic
brain injury: a clear relationship? Sleep Med. 13 (6), 559560.
2013. Scottish Intercollegiate Guidelines Network Published Guidelines, Retrieved
from: http://www.sign.ac.uk.guidelines/ (24.09.13).

716

T. Mollayeva et al. / Neuroscience and Biobehavioral Reviews 47 (2014) 684716

Sigurdardottir, S., Andelic, N., Roe, C., Schanke, A.-K., 2009. Cognitive recovery and
predictors of functional outcome 1 year after traumatic brain injury. J. Intern.
Neuropsychol. Soc. 15, 740750.
Slavin, R.E., 1995. Best evidence synthesis: an intelligent alternative to metaanalysis. J. Clin. Epidemiol. 48, 918.
Stepanski, E., 2002. The effect of sleep fragmentation on daytime function. Sleep 25,
268276.
Sundstrom, A., Nilsson, L.-G., Cruts, M., Adolfsson, R., van Broeckhaun, C., Nyberg,
L., 2007. Fatigue before and after mild traumatic brain injury: prepostinjury comparisons in relation to apolipoprotein E. Brain Inj. 21 (10),
10491054.
van der Naalt, J., van Zomeren, A.H., Sluiter, W.J., Minderhoud, J.M., 1999. One year
outcome in mild to moderate head injury: the predictive value of acute injury
characteristics related to complaints and return to work. J. Neurol. Neurosurg.
Psychiatry 66, 207213.
van Rein, N., Cannegieter, S.C., Rosendaal, F.R., Reitsma, P.H., Lijfering, W.M., 2014.
Suspected survivor bias in case-control studies: stratify on survival time and use
a negative control. J. Clin. Epidemiol. 67 (2), 232235.

Ware, J.E., 1992. The MOS 36-item short form health survey (SF-36): conceptual
framework and item selection. Med. Care, 473483.
Wiart, L., Richer, E., Destaillats, J.M., Joseph, P.A., Dehail, P., Mazaux, J.M., 2012.
Psychotherapeutic follow up of our patients with traumatic brain injury: preliminary results of an individual neurosystemic approach. Ann. Phys. Rehabil.
Med. 55, 375387.
World Health Organization, 2002. Projection of Mortality and Burden of Disease to
2030. Death by Income Group, Geneva.
Yang, C.C., Hua, M.S., Tu, Y.K., Huang, S.-J., 2009. Early clinical characteristics of
patients with persistent post-concussion symptoms: a prospective study. Brain
Inj. 23 (4), 299306.
Yennurajalingam, S., Bruera, E., 2007. Palliative management of fatigue at the close
of life: it feels like my body is just worn out. JAMA 297 (3), 295304 (future
readings).
Zaben, M., El Ghoul, W., Belli, A., 2013. Post-traumatic head injury pituitary dysfunction. Disabil. Rehabil. 35 (6), 522525.
Ziino, C., Ponsford, J., 2006. Selective attention decits and subjective fatigue following traumatic brain injury. Neuropsychology 20 (3), 383390.

You might also like