Non-Pulmonary Critical Care

PREFACE
Non-pulmonary Critical Care: Managing Multisystem Critical Illness
C are of the critically ill patient is truly multi- gan dysfunction in the ICU setting. Clearly, a compre-
system management of highly complex patients who hensive review of the numerous organ system–based
typically have numerous acute physiological derange- medical conditions seen in our ICUs is well beyond
ments superimposed upon underlying medical ailments. the scope of a single issue of Seminars, more appropri-
Historically, the majority of patients admitted to inten- ately filling several thousand pages and hundreds of
sive care units (particularly medical ICUs) have respira- chapters typical of current major textbooks on critical
tory failure requiring mechanical ventilation, often along care medicine. Nevertheless, we have identified a cross-
with other acute and chronic pulmonary problems. section of key topics and have solicited state-of-the-art
Much ICU-related research and education has similarly reviews by highly qualified experts.
focused on pulmonary issues such as the acute respiratory Starting at the top, so to speak, Dr. Bleck presents
distress syndrome (ARDS), pneumonia, and mechanical a discussion of neurological complications of critical
ventilation. Additionally, a great majority of intensivists illness, with particular emphasis on metabolic encepha-
have received training in pulmonary and critical care lopathies, seizures, and neuromuscular conditions. Drs.
medicine. Thus it is not surprising that the phrase ‘‘non- Miller and Ely share their current understanding of the
pulmonary critical care’’ has arisen to address many rapidly evolving areas of delirium and cognitive dysfunc-
general critical care issues. However, this vast body of tion in the ICU. Drs. Tarditi and Hollenberg present a
knowledge might more appropriately be considered structured approach to the ICU patient who has cardiac
‘‘multisystem critical care’’ as an inclusive term that arrhythmias, emphasizing the challenges of managing
encompasses the many important organ system derange- tachyarrhythmias, whereas Dr. Axler discusses the eval-
ments that plague our ICU patients. uation and management of shock. Drs. Rinella and
This issue of Seminars provides scholarly and Sanyal provide a comprehensive review of acute hepatic
clinically relevant reviews of major non–pulmonary or- failure as well as acute complications of chronic advanced
1
Division of Pulmonary and Critical Care Medicine, Virginia Com- Non-pulmonary Critical Care: Managing Multisystem Critical Illness;
monwealth University Health System, Richmond, Virginia. Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Curtis N. Sessler, Semin Respir Crit Care Med 2006;27:199–200. Copyright # 2006
M.D., Division of Pulmonary and Critical Care Medicine, Box 980050, by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Virginia Commonwealth University Health System, Richmond, VA NY 10001, USA. Tel: +1(212) 584-4662.
23298-0050. E-mail: [email protected]. DOI 10.1055/s-2006-945524. ISSN 1069-3424.
199
200 SEMINARS IN RESPIRATORY AND CRITICAL CARE MEDICINE/VOLUME 27, NUMBER 3 2005
liver disease. Drs. Weisbord and Palevsky examine acute overview of infection control and prevention of nosoco-
renal failure, including sections on epidemiology, causes, mial infections in the ICU, with emphasis on ventilator-
prevention, and management. associated pneumonia and catheter-related bloodstream
Drs. Raghavan and Marik present an in-depth infection.
review of adrenal insufficiency and the many issues I would like to acknowledge the amazing exper-
related to hyperglycemia and glycemic control in the tise, attention to detail, and hard work of the authors of
ICU. Drs. Mercer, Macik, and Williams expertly ad- these articles, and convey my thanks for the steady hand
dress hematological disorders in the ICU, with particular of the Seminars Editor-in-Chief, Joseph P. Lynch, III,
emphasis on thrombocytopenia and bleeding disorders. M.D., and the understanding and support of my wife
Drs. Afessa and Peters provide a practical framework for and children, in creating this issue of Seminars.
evaluating and managing the many serious infectious
and non-infectious complications related to hemato-
poietic stem cell transplantation. Drs. Bearman, Munro, Curtis N. Sessler, M.D.
Sessler, and Wenzel conclude with a comprehensive Guest Editor 1
Neurological Disorders in the Intensive Care
Unit
Thomas P. Bleck, M.D., F.C.C.M.1,2,3
ABSTRACT
Neurological problems are common among critically ill patients; they often signal
that other organs are failing, but are themselves important causes of morbidity and
mortality. Cognitive function may suffer as a consequence of septic encephalopathy, the
pathophysiology of which is poorly understood; however, the affected patients usually
return to their baseline when sepsis resolves. Seizures and cerebrovascular disorders are also
common in the intensive care unit. Neuromuscular complications are important causes of
failure to wean from mechanical ventilation and lead to substantial long-term morbidity.
KEYWORDS: Acute quadriplegic myopathy, ARDS: acute respiratory distress

syndrome, critical illness myopathy, critical illness polyneuropathy, fulminant hepatic
failure, seizure, septic encephalopathy
Many conditions encountered in intensive care within 72 hours of admission.3 The study included all
affect the nervous system. The onset of an abrupt neuro- patients in need of intensive care except those with
logical complication is frequently obscured by the effects primarily cardiac disorders. Eighteen were admitted for
of the primary illness (e.g., metabolic encephalopathy acute neurological disease and five others for encephal-
may delay the recognition of an intracerebral hemor- opathy caused by drug overdose. Of the remaining 67,
rhage) or its treatment (such as sedation to allow greater 33% had a neurological complication of their medical
synchrony with a ventilator). Other neural problems, conditions (11 metabolic encephalopathy, four hypoxic-
such as critical illness polyneuropathy, may develop ischemic encephalopathy [HIE], and seven other neuro-
insidiously and become apparent only as the patient logical problems). Fifty-nine percent of the patients with
improves. At times the neurological problem has been neurological complications died, compared with 20% of
visible, but its manifestations may be inappropriately the nonneurological patients.
attributed to the presenting illness.1 The intensivist At the same time, we performed a 2-year pro-
should be perspicacious about changes in level of con- spective study of neurological complications among
sciousness or movement when investigating a fall in medical intensive care unit (MICU) patients to describe
oxygen saturation or a rising white blood cell count.2 the neurological complications encountered and to iden-
tify their effects on mortality and length of stay (LOS).4
Patients with a primarily neurological reason for admis-
EPIDEMIOLOGY sion to the MICU were excluded from analyses of
Isensee et al evaluated neurological problems in 100 mortality and LOS; more than half of this group had
consecutive medical intensive care unit (ICU) patients ischemic stroke or intracranial hemorrhage. The others
Departments of 1Neurology, 2Neurological Surgery, 3Internal Medicine, Non-pulmonary Critical Care: Managing Multisystem Critical Ill-
University of Virginia School of Medicine, Charlottesville, Virginia. ness; Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Thomas P. Bleck, Semin Respir Crit Care Med 2006;27:201–209. Copyright # 2006
M.D., University of Virginia School of Medicine, Neurology 800394, by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
McKim Hall 2025, Charlottesville, VA 22908-0394. E-mail: tbleck@ NY 10001, USA. Tel: +1(212) 584-4662.
virginia.edu. DOI 10.1055/s-2006-945531. ISSN 1069-3424.
201
Table 1 Neurological Complications Encountered in 217 Table 3 Etiologies of Metabolic Encephalopathy in a

Patients at Risk with Severe Medical Illnesses in the Medical Intensive Care Unit Population
Medical Intensive Care Unit
Etiology N
N (percent of patients
Complication with diagnosis)* Sepsis 19
Hepatic 18
Metabolic encephalopathy 62 (28.6) Renal 8
Seizures 61 (28.1) Hypertensive 7
Hypoxic-ischemic encephalopathy 51 (23.5) Hyperosmolar 4
Stroke 48 (22.1) Hypoglycemic 3
Other diagnoses 50 (23.0) Uncertain 3
*A single patient could have more than one complication; therefore, 12
Modified from Sprung et al.
the total number in this column exceeds the total number of patients.
Modified from Bleck et al.4
were classified as having a complication of a critical these, septic encephalopathy, without evidence for sig-
illness if they developed a neurological problem from a nificant hepatic or renal dysfunction or hypoxemia, was
medical disorder or its treatment. We collapsed the most common. The frequencies of different forms of
medical diagnoses into four categories: (1) sepsis, in- metabolic encephalopathy are detailed in Table 3. Seiz-
cluding bacteremia with shock, the sepsis syndrome, ures occurred in 61 patients, most often in patients with
and the acute respiratory distress syndrome (ARDS); vascular lesions. Hypoxic-ischemic encephalopathy oc-
(2) acute coronary artery disease; (3) other cardiac curred in 51 patients; in 27, the cause was primarily
disorders; and (4) all other patients (e.g., ventilatory cardiac, with pulmonary disease accounting for the re-
failure, gastrointestinal hemorrhage, hypotension not maining 24.
assigned to another category). Patients with neuro- Forty-eight patients suffered strokes while in the
logical complications were further divided into those MICU. Thirty-two of these were ischemic infarcts, 14
with metabolic encephalopathies, seizures, cerebrovas- were intracerebral hemorrhages, and two were subar-
cular disorders, hypoxic-ischemic encephalopathy, or achnoid hemorrhages. Thirteen stroke patients had an
other global brain disorders. These latter categories identified cause other than arteriosclerosis, including
were not mutually exclusive. Patients with clinically underlying autoimmune diseases and bacterial endocar-
apparent peripheral nervous system disorders were in- ditis. Stroke occurred in only 1% of patients with acute
cluded in the ‘‘other’’ group. myocardial infarction. This was less than the usually
We studied 1850 consecutively admitted patients; cited range of 1.7 to 2.4%.5,6
of these, 92 (4.9%) were admitted for a primary neuro-
logical reason. Of the remaining 1758 patients, 217
(12%) experienced neurological complications of their SEPSIS AND SEPTIC ENCEPHALOPATHY
underlying medical disease (Table 1). Table 2 details During the past 40 years, clinical analyses and investiga-
the neurological complication rates by admission cate- tions of cytokine mechanisms have markedly improved
gory. The mortality rate for all MICU patients was 32%, our understanding of the causes and pathogenesis of
but it was 55% for the 217 patients with neurological sepsis.7 Although bacteremia was previously considered
complications, compared with 29% for those without to be the sine qua non of systemic disease, occurring as a
neurological complications. Patients with neurological consequence of local infection, many patients suffer the
complications also had significantly longer MICU and same vasomotor disturbances and organ dysfunctions
hospital stays. without positive blood cultures. The foregoing epidemio-
Metabolic encephalopathy was the complication logical data indicate that septic encephalopathy is the
most frequently encountered, seen in 62 patients. Of most frequent neurological disorder encountered in
Table 2 Neurological Complication Rates by Primary Medical Intensive Care Unit

Admission Category
Percent of Patients with Complications
Category Seizure Vascular HIE Metabolic Other
Sepsis 11 6 10 21 11
Other medical condition 4 3 4 3 6
Coronary artery disease 1 1 1 1 1
Other cardiac condition 4 3 3 2 4
Modified from Bleck et al.4
NEUROLOGICAL DISORDERS IN THEICU/BLECK 203
medical intensive care; it is also one of the more poorly puric lesions. Eight of the patients had electroencepha-
recognized and understood. Septic encephalopathy was lograms (EEGs), three of which showed multifocal
described in 18278 but has only recently become a subject epileptiform activity. Pendlebury and associates identi-
of organized neurological interest. Young and coworkers fied 35 patients with multiple CNS microabscesses
provided a thorough prospective analysis of this disorder among 2107 consecutive autopsies.15 All these patients
in a large university ICU.9 This group required fever and had chronic, usually immunocompromising, diseases,
a positive blood culture for inclusion in their study, a very and were frequently septic before death. The most
restrictive definition of sepsis, which provided a homo- common organisms implicated were Staphylococcus aureus
geneous group for analysis. Patients were excluded for and Candida albicans. In contrast, we did not find
preexisting brain disease; frequent sedative or opiate microabscesses in four patients autopsied of our 14 fatal
administration; pulmonary, hepatic, or renal failure; en- septic encephalopathy cases; this may represent sampling
docarditis; or long bone fractures that might have pro- error or other differences in the populations studied.4
duced fat embolism. The pathophysiology of septic encephalopathy is
These workers identified 69 patients over 31 of great interest. The systemic mediators of inflamma-
months; by clinical examination, 20 of them were not tion are capable of damaging the blood–brain barrier.16
encephalopathic (NE), 17 were mildly encephalopathic Such disruption has been documented in an animal
(ME), and 32 were severely encephalopathic (SE). Pa- model early in sepsis.17 The behavioral effects of cyto-
tient age, blood pressure on entry into the study, and kines vary with the neuroanatomical structures affected
temperature did not vary significantly among the groups. but include thermogenic behaviors in the hypothala-
The lowest systolic and diastolic blood pressures were mus18 and somnolence in the locus caeruleus.19 Inter-
statistically significantly lower (but probably not biolog- ferons also alter individual cortical and hippocampal
ically significantly lower) in the ME and SE groups when neuronal functions, suggesting a myriad of effects on
compared with the NE group. Mortality depended on the memory and emotion.20 Brain catecholamine concen-
category of encephalopathy: none of the NE patients trations are decreased in experimental sepsis.21 Sepsis
died, whereas 35% of the ME and 53% of the SE patients leads to the release of S-100B, a protein predominantly
died. Several laboratory values showed a linear relation- expressed in CNS glial cells, into the systemic circula-
ship with the severity of encephalopathy, including white tion; this leakage was not affected by steroids in a human
blood cell count, PaO2, blood urea nitrogen (BUN), trial.22 Increased serum concentrations of S-100B are
creatinine, bilirubin, alkaline phosphatase, and potas- usually viewed as evidence of cellular damage, which may
sium. The serum albumin concentration was inversely not be reparable.
related to encephalopathy. cerebrospinal fluid (CSF) Cerebral blood flow (CBF) and cerebral oxygen
protein content was mildly elevated (60 to 85 mg/dL). extraction decrease in septic encephalopathy,23 along
Electroencephalographic abnormalities are more sensitive with the development of cerebral edema and disruption
indicators of central nervous system (CNS) dysfunction of the blood–brain barrier.24 Preliminary human studies
than the clinical examination, and also a powerful pre- suggest that these problems occur in several gray matter
dictor of survival.10 Evoked potential studies suggest that structures of the brain.25,26 Failure of cerebrovascular
brain dysfunction is even more prevalent in sepsis, being autoregulation is likely to compound these disorders,27
abnormal in 84%.11 A Veterans Administration (VA) producing cerebral ischemia. Both cerebral edema and
cooperative sepsis study also showed that ‘‘alterations in blood–brain barrier disruption appear to correlate with
mental status are common in septic patients, and are damage to astrocytic foot-processes.28,29 Aquaporin-4
associated with significantly higher mortality.’’12 expression increases in septic encephalopathy, but the
Eidelman and colleagues studied 50 patients with cellular events that trigger this change in sepsis require
severe sepsis and showed that encephalopathy was asso- further research.30
ciated with bacteremia and hepatic dysfunction.13 The The cause of the changes in CBF and oxygen
severity of encephalopathy correlated with mortality. extraction are less well understood. Focal elevations in
intracellular free calcium may cause neuronal dysfunc-
tion and also contribute to apoptotic31 or necrotic cell
PATHOLOGY AND PATHOPHYSIOLOGY loss.32 Activation of adenosine A1 receptors may be
The pathological basis of septic encephalopathy remains important in the development of a local CNS inflam-
obscure. Jackson et al autopsied 12 patients dying after matory response.33 Although cytokines have been sug-
severe, prolonged sepsis.14 They found cerebral micro- gested as mediators, a study of the effects of tumor
abscesses in eight patients and proliferation of astrocytes necrosis factor was unable to confirm its role in this
and microglia in three others; these findings suggested regard.34 However, a decline in CSF ascorbic acid
metastatic infection. Three of these patients also had concentration may reflect difficulty in safely handling
central pontine myelinolysis, and three had ischemic the oxygen-derived free radicals potentially resulting
strokes. The remaining patient demonstrated only pur- from both cytokine and nitric oxide excess.35 These
radicals may interfere with the mitochondrial electron Hyperventilation had previously been thought to
transport chain, leading to cellular energy deprivation.36 increase mortality, but a controlled trial demonstrated its
Magnesium administration may be able to attenuate utility.52 High-dose barbiturates may be used if mannitol
some of these problems,37 although this remains to be and hyperventilation fail to control ICP.53 The effect of
demonstrated in humans. posture is unpredictable,54 and computed tomographic
Abnormal systemic metabolism in sepsis may also (CT) scanning is debated as an indicator of the severity
contribute to CNS dysfunction. Mizock et al demon- of intracranial hypertension.55,56 Attempts to lower ICP
strated altered phenylalanine metabolism (elevated blood by reducing extracellular volume through hemofiltration
and CSF levels, and elevated phenylalanine metabolites) have not been effective.57 Dialysis against albumin may
in 11 patients with septic encephalopathy; in contrast, hold more promise.58 There is no current substitute for
patients with hepatic encephalopathy had elevated CSF invasive ICP monitoring in patients with FHF who are
concentrations of many other aromatic amino acids.38 in grade 3 (stuporous) or grade 4 (comatose). The need
Sprung and coworkers found elevated serum levels of for monitoring extends through the liver transplant
phenylalanine, ammonia, and tryptophan in encephalo- operation into at least the first postoperative day.59
pathic patients compared with infected patients with Recombinant factor VII administration immediately
normal sensoria, along with lower concentrations of before ICP monitor placement appears to reduce the
isoleucine.39 Other studies have confirmed this, and risk of intracerebral hemorrhage.60
find it linked to concentrations of calcitonin precursors Patients with FHF may also have elevated levels
and interleukin-6.40 The significance of these correlative of apparently endogenous 1,4-benzodiazepines,(3)
studies for the pathogenesis of encephalopathy is un- which may explain stupor or coma in patients who do
certain. Several authors have tried to implicate abnormal not have ICP elevations. An interaction between
hepatic and muscular metabolism of aromatic amino gamma-amino butyric acid (GABA) and elevated con-
acids in both hepatic and septic encephalopathies, but centrations of ammonia may be important in both FHF
this hypothesis has been challenged (see further discus- and more chronic forms of hepatic encephalopathy.61
sion in the next section). Antibiotic treatment of septic More recent work has concentrated on the putative role
patients may actually enhance these problems.41 of neurosteroid agonists of the benzodiazepine recep-
Septic patients are prey to a wide variety of other tor62 because researchers have been unable to identify a
metabolic disorders and intoxications that cause ence- compound that structurally resembles the pharmaceut-
phalopathy apart from the direct effects of sepsis on the ical benzodiazepine nucleus.
brain and on cerebral blood flow.42 Nonconvulsive status Chronic hepatic encephalopathy does not appear
epilepticus is an underrecognized problem in the pop- to cause ICP elevation unless intracranial bleeding
ulation at risk.43 There is neither a diagnostic test to supervenes. As in FHF, endogenous benzodiazepine-
discriminate sepsis from other causes of encephalopathy like compounds (GABAA agonists) likely play a major
nor a specific treatment for the CNS disturbance. From a role,(4) and the use of GABA antagonists is being
clinical standpoint, the diagnosis of septic encephalop- actively investigated but has not yet become standard
athy remains one of exclusion.44 practice. About 70% of chronic hepatic encephalopathy
patients awaken rapidly when given flumazenil (for the
duration of the drug’s effect).63,64 These patients have
Hepatic Failure and the Central Nervous System numerous other metabolic abnormalities that may con-
The current debate over the pathogenesis of hepatic tribute to their encephalopathy, including abnormalities
encephalopathy is of great importance to clinical neuro- in the Krebs cycle65 and of methionine metabolism.66
scientists. In fulminant hepatic failure (FHF), increased
intracranial pressure (ICP) has become a major cause of
death in patients awaiting transplantation.45 Patients OTHER CAUSES OF ENCEPHALOPATHY IN
whose ICP has been elevated may survive a transplant THE INTENSIVE CARE UNIT SETTING
but be left with CNS deficits.46 Renal and hypoxic encephalopathies are quite common
The mechanism of the cerebral edema that devel- in the ICU environment. Details of their diagnosis and
ops in FHF appears to be related to mitochondrial management are beyond the scope of this article, and
dysfunction47; it requires aggressive treatment.48 Ste- have been reviewed.67
roids are not effective, but mannitol has been useful.49 Iatrogenic causes of encephalopathy are also com-
As in many other causes of brain edema, abnormal mon in ICUs and should be actively excluded. Hypno-
function of matrix metalloproteinases probably contrib- sedative drugs and narcotic analgesics are the most
utes to the problem.50 As with most other causes of brain common agents in this category. Flumazenil and nalox-
edema, aquaporin-4 expression is involved; it may be one will reverse these drug-induced encephalopathies.
possible to decrease this by the administration of calci- The need to reverse the effects of these drugs in an
neurin antagonists.51 individual ICU patient should be carefully assessed
because the emergence of agitation or severe pain may be of anticonvulsant therapy; this was often justified by the
detrimental to the patient. If these drugs are used to argument that their underlying medical condition might
determine whether encephalopathy is due to medica- be adversely affected by subsequent seizures. We believe
tions, doses much smaller than those to treat respiratory that this constitutes excessively aggressive management
depression should be used initially. Flumazenil (0.1 to and that it postpones an appropriate search for etiology.
0.2 mg intravenously over 15 seconds), may be given For example, all three patients with recurrent seizures
every 60 seconds to a maximum of 1.0 mg. The risk of caused by nonketotic hyperglycemia were treated with at
seizures is always present when this drug is adminis- least two anticonvulsants; these drugs are known to be
tered.68 Naloxone, 0.04 to 0.08 mg intravenously, may ineffective in this condition.70
be given every 60 seconds to a total dose of 0.8 mg. When ICU patients do require treatment, phe-
Barbiturates are not antagonized by available agents. It nytoin remains a reasonable first choice. However, a
should be apparent from the preceding section on second agent, usually phenobarbital, was required in
hepatic encephalopathy that a response to flumazenil is most patients. Lorazepam was useful for the suppression
not specific for a benzodiazepine overdose. of breakthrough seizures, but the use of this drug was
Other encephalopathic conditions may arise dur- often continued without adequate attention to optimal
ing the course of ICU treatment. One of our poorly use of phenytoin or phenobarbital.
nourished patients developed Wernicke’s encephalop- Pentobarbital coma is often considered the treat-
athy from the dextrose administered as the vehicle for a ment of choice for refractory status epilepticus in ICU
lidocaine infusion to treat ventricular arrhythmias after patients. We have adopted the use of high-dose mid-
an acute myocardial infarction. The clinician must al- azolam in this circumstance.71 This method appears to
ways be aware that admission to the ICU does not be more rapidly effective and to have substantially fewer
prevent the development of intercurrent illnesses. and less severe adverse effects than pentobarbital, thio-
pental, or propofol.72
Seizures in the Intensive Care Unit

In our epidemiological study, 34 patients had simple NEUROMUSCULAR COMPLICATIONS
partial seizures (with or without secondary generaliza- OF SEPSIS
tion), and six had complex partial seizures (with or Although recognized by earlier authors such as Osler,
without secondary generalization).4 Twenty patients the modern era of interest in this problem began with the
had seizures that appeared to be generalized at onset, independent reports of three groups reported by Rivner
and six patients developed status epilepticus in the et al (four patients),73 Bolton et al (17 patients),74 and
MICU; all required at least two agents to terminate Roelofs (four patients).75 Bolton’s group followed their
their status (usually a benzodiazepine and phenytoin). initial description with a series of papers that character-
Two of these patients developed refractory status epi- ized the clinical,76 electrophysiological,77 and patholog-
lepticus and were treated with pentobarbital coma. ical aspects of critical illness polyneuropathy.78 Although
Three patients were admitted to the MICU for refrac- many other groups have made significant contributions
tory hypotension after receiving phenytoin infusions at to this area (e.g., Op de Coul et al),75 the work of Bolton
rates between 25 and 50 mg/min. The blood pressures of and his colleagues is in great measure responsible for the
these patients did not improve with fluid resuscitation recognition of this problem among intensivists. More
alone; all required dopamine infusions for several hours recent work continues to show a high incidence of this
to maintain blood pressure and systemic perfusion. condition.79
In contrast to our experience in other hospitalized These neuromuscular complications are presently
patients, the focal onset of partial seizures with secon- categorized anatomically. Critical illness polyneuropathy
dary generalization was usually noted and adequately is an axonal disorder affecting both sensory and motor
described. This was a useful guide to management nerves.78 Electrophysiological evidence of this condition,
because patients with partial seizures generally experi- as already noted, is present in 70% of septic patients,
enced seizure recurrence and thus probably benefited but the percentage with weakness sufficient to impede
from anticonvulsant treatment.69 ventilator weaning or ambulation is less. The phrenic
The diagnosis and management of seizures nerves are typically the most severely involved. Neuro-
should be pursued differently in ICU patients than in muscular junction (NMJ) dysfunction in the setting of
other patients. We had hoped to develop rules to predict critical illness is typically a consequence of a prolonged
which patients having seizures in the ICU might be effect of NMJ blocking agents, usually because of im-
evaluated without imaging procedures, but were unable paired clearance.80 Myopathy in critically ill patients is
to do so because most patients had vascular or infectious most commonly seen in those who have received NMJ
causes for their seizures. All patients experiencing a blocking agents and corticosteroids in the treatment of
single seizure in the ICU were treated with some form severe asthma.81 Although this has been reported most
commonly after vecuronium use, it also occurs after NMJ increased cerebrospinal fluid protein concentration. An-
blocking agents that do not depend on renal or hepatic tecedent infection with Campylobacter jejeuni is fre-
excretion (e.g., atracurium).82 Myopathy may also occur quently in patients with the axonal form of Guillain-
in the setting of some viral infections, such as influenza. Barré.94 Other differential diagnostic concerns are bot-
A syndrome of disseminated pyogenic myopathy has also ulism, which impairs presynaptic acetylcholine release,
been described,83 presumably as a consequence of bac- and myasthenia gravis, in which the motor end plate is
teremic seeding of muscles. damaged. These conditions have characteristic nerve
Critical illness polyneuropathy has also been conduction and electromyographic characteristics. No
noted after organ transplantation without sepsis.84 specific treatment for critical illness polyneuropathy is
Although most of the patients reported to have critical available. Anecdotal evidence does not favor the use of
illness polyneuropathy have been adults, this condition plasma exchange or intravenous immunoglobulin, but
has been recognized with increasing frequency in chil- there have been no definitive trials. Almost all patients
dren as well.85 recover eventually, but this may require 6 months or
In contrast to the body of experimental work on more of ventilatory support.
septic encephalopathy, very little is understood about the The prolonged effect of NMJ blocking agents can
pathogenesis of the neuromuscular complications of be suspected by lack of tendon reflexes and inability to
sepsis.86 Hyperglycemia correlates with the development produce muscle contraction with a bedside neuromus-
of critical illness polyneuropathy, but these may be cular stimulator. If the diagnosis is in question, it can be
independent markers of disease severity. However, the confirmed by formal nerve conduction studies and elec-
dramatic reduction in critical illness polyneuropathy tromyography. There is no specific treatment, but the
with intensive control of blood glucose in both medical87 condition will resolve when the agent in question has
and surgical88 ICU patients points to an important role cleared.
of hyperglycemia in the genesis of this condition, as well Critical illness myopathy is often accompanied by
as in many other neurological problems seen in critical substantial elevation of the serum creatine kinase (CK)
care units.89 The neuronal microenvironment is similar concentration, which serves to distinguish this condition
to that of the extracellular space of the brain; the same from steroid myopathy, in which the CK is usually
alterations that produce septic encephalopathy may be normal. Electromyography is usually an adequate diag-
responsible for critical illness polyneuropathy, but fur- nostic study, and muscle biopsy is only rarely required.95
ther investigation is clearly needed. Critical illness myo- Again, no specific treatment is available, but the con-
pathy may result from the functional denervation dition will resolve; whether it will resolve faster if
induced by NMJ blockade; it does not appear to be systemic steroids are reduced or discontinued is uncer-
simply a severe form of steroid myopathy. tain. A more recently described syndrome of acute
quadriplegic myopathy is also probably associated with
steroids; in this condition, muscle is electrically inexcit-
Diagnosis, Differential Diagnosis, and Prognosis able even with direct stimulation.96 Nerves are histolog-
Recognition of a neuromuscular complication of sepsis ically normal, but muscles show thick filament loss.97
generally occurs as the patient begins to recover from the Although no treatment is known, the prognosis for this
critical illness that first required ventilatory support. condition is for more rapid recovery than that for critical
Critical illness polyneuropathy is usually suspected illness polyneuropathy.
when the patient’s pulmonary mechanics (e.g., static There are many other neurological problems that
compliance) and gas exchange suggest that weaning may arise during the course of an ICU stay which may
should be possible, but the patient is too weak to tolerate impede weaning from mechanical ventilation. Kelly and
it.90 There is commonly some evidence of distal weak- Matthay prospectively studied 66 consecutive adult pa-
ness on examination. Tendon reflexes are often absent or tients requiring mechanical ventilation for more than
diminished, but critical illness polyneuropathy may be 48 hours to determine the reasons for their ventilatory
present without alteration in reflexes.91 Gorson provides problems.95 Neurological problems, primarily encepha-
an excellent framework for the evaluation of patients lopathies, were held responsible for the continuing need
with these problems.92 for ventilatory support in 32% of the patients, and
The diagnosis depends on electrophysiological contributed to this problem in another 41%. Although
studies, which demonstrate an axonal disorder. Electro- this study was not directed at patients who failed to wean
myographic studies of the diaphragm confirm the pres- after resolution of their presenting disease, it does high-
ence of denervation.93 The differential diagnosis is light the role of neurological problems early in critical
usually limited to consideration of the axonal form of illness. Spitzer and colleagues studied 21 patients who
the Guillain-Barré syndrome; this condition usually failed to wean after their presenting disease had im-
causes much more severe generalized weakness than proved to the point that their intensivists believed that
critical illness polyneuropathy, and typically causes an mechanical ventilation should no longer have been
necessary.98 Thirteen (62%) of these patients had a comparison with tumor necrosis factor. J Clin Invest 1991;87:
neuromuscular disorder that was either the major cause 1360–1366
of or contributed substantially to their ventilatory prob- 17. du Moulin GC, Paterson D, Hedley-Whyte J, Broitman SA.
E. Coli peritonitis and bacteremia cause increased blood–brain
lems. Only seven of these 13 patients had critical illness
barrier permeability. Brain Res 1985;340:261–268
polyneuropathy; other neuropathic conditions and un- 18. Dinarello CA, Cannon JG, Wolff SM. New concepts on the
suspected motor neuron disease were also uncovered. pathogenesis of fever. Rev Infect Dis 1988;10:168–189
Most intensivists will also be familiar with the less 19. De Sarro GB, Masuda Y, Ascioti C, Audino MG, Nistico G.
common acute myopathies which may complicate crit- Behavioural and ECoG spectrum changes induced by intra-
ical illnesses, especially in patients who have received cerebral infusion of interferons and interleukin 2 in rats are
neuromuscular junction blockade.99 antagonized by naloxone. Neuropharmacology 1990;29:167–
179
20. Reyes-Vazquez C, Prieto-Gomez B, Georgiades JA, Dafny
N. Alpha and gamma interferons’ effects on cortical and
REFERENCES hippocampal neurons: microiontophoretic application and
single cell recording. Int J Neurosci 1984;25:113–121
1. Bleck TP. Why isn’t this patient awake? J Intensive Care Med 21. Kadoi Y, Saito S, Kunimoto F, Imai T, Fujita T. Impairment
1993;8:155–156 of the brain beta-adrenergic system during experimental
2. Wijdicks EF. Neurologic complications in critically ill endotoxemia. J Surg Res 1996;61:496–502
patients. Anesth Analg 1996;83:411–419 22. Mussack T, Briegel J, Schelling G, Biberthaler P, Jochum M.
3. Isensee LM, Weiner LJ, Hart RG. Neurologic disorders in a Effect of stress doses of hydrocortisone on S-100B vs.
medical intensive care unit: a prospective survey. J Crit Care interleukin-8 and polymorphonuclear elastase levels in human
1989;4:208–210 septic shock. Clin Chem Lab Med 2005;43:259–268
4. Bleck TP, Smith MC, Pierre-Louis SJ, Jares JJ, Murray J, 23. Maekawa T, Fujii Y, Sadamitsu D, et al. Cerebral circulation
Hansen CA. Neurologic complications of critical medical and metabolism in patients with septic encephalopathy. Am J
illnesses. Crit Care Med 1993;21:98–103 Emerg Med 1991;9:139–143
5. Komrad MS, Coffey CE, Coffey KS, McKinnis R, Massey 24. Papadopoulos MC, Davies DC, Moss RF, Tighe D, Bennett
EW, Califf RM. Myocardial infarction and stroke. Neurology ED. Pathophysiology of septic encephalopathy: a review. Crit
1984;34:1403–1409 Care Med 2000;28:3019–3024
6. Thompson PL, Robinson JS. Stroke after acute myocardial 25. Finelli PF, Uphoff DF. Magnetic resonance imaging
infarction: relation to infarct size. BMJ 1978;2:457–459 abnormalities with septic encephalopathy. J Neurol Neurosurg
7. Nathan BR, Bleck TP. Central nervous system complications Psychiatry 2004;75:1189–1191
of critical medical illness. In: Miller DH, Raps EC, eds. 26. Sharshar T, Annane D, de la Grandmaison GL, Brouland JP,
Critical Care Neurology. Boston: Butterworth-Heineman; Hopkinson NS, Francoise G. The neuropathology of septic
1999:331–340 shock. Brain Pathol 2004;14:21–33
8. Bright R. Reports of Medical Cases Selected with a View of 27. Terborg C, Schummer W, Albrecht M, Reinhart K, Weiller
Illustrating the Symptoms and Cure of Diseases, with a C, Rother J. Dysfunction of vasomotor reactivity in severe
Reference to Morbid Anatomy. London: Longman, Rees, sepsis and septic shock. Intensive Care Med 2001;27:1231–
Orme, Brown and Green; 1827 1234
9. Young GB, Bolton CF, Austin TW, Archibald YM, Gonder 28. Papadopoulos MC, Lamb FJ, Moss RF, Davies DC, Tighe
J, Wells GA. The encephalopathy associated with septic D, Bennett ED. Faecal peritonitis causes oedema and
illness. Clin Invest Med 1990;13:297–304 neuronal injury in pig cerebral cortex. Clin Sci (Lond) 1999;
10. Young GB, Bolton CF, Archibald YM, Austin TW, Wells 96:461–466
GA. The electroencephalogram in sepsis-associated ence- 29. Ari I, Kafa IM, Kurt MA. Perimicrovascular edema in the
phalopathy. J Clin Neurophysiol 1992;9:145–152 frontal cortex in a rat model of intraperitoneal sepsis. Exp
11. Zauner C, Gendo A, Kramer L, et al. Impaired subcortical Neurol 2006;198:242–249
and cortical sensory evoked potential pathways in septic 30. Davies DC. Blood–brain barrier breakdown in septic
patients. Crit Care Med 2002;30:1136–1139 encephalopathy and brain tumours. J Anat 2002;200:639–646
12. Sprung CL, Peduzzi PN, Shatney CH, et al. Impact of 31. Semmler A, Okulla T, Sastre M, Dumitrescu-Ozimek L,
encephalopathy on mortality in the sepsis syndrome. The Heneka MT. Systemic inflammation induces apoptosis with
Veterans Administration Systemic Sepsis Cooperative Study variable vulnerability of different brain regions. J Chem
Group. Crit Care Med 1990;18:801–806 Neuroanat 2005;30:144–157
13. Eidelman LA, Putterman D, Putterman C, Sprung CL. The 32. Zhan RZ, Fujiwara N, Shimoji K. Regionally different
spectrum of septic encephalopathy: definitions, etiologies, and elevation of intracellular free calcium in hippocampus of septic
mortalities. JAMA 1996;275:470–473 rat brain. Shock 1996;6:293–297
14. Jackson AC, Gilbert JJ, Young GB, Bolton CF. The 33. Weeks SG, Silva C, Auer RN, Doig CJ, Gill MJ, Power C.
encephalopathy of sepsis. Can J Neurol Sci 1985;12:303–307 Encephalopathy with staphylococcal endocarditis: multiple
15. Pendlebury WW, Perl DP, Munoz DG. Multiple micro- neuropathological findings. Can J Neurol Sci 2001;28:260–264
abscesses in the central nervous system: a clinicopathologic 34. Moller K, Strauss GI, Qvist J, et al. Cerebral blood flow and
study. J Neuropathol Exp Neurol 1989;48:290–300 oxidative metabolism during human endotoxemia. J Cereb
16. Quagliarello VJ, Wispelwey B, Long WJ Jr, Scheld WM. Blood Flow Metab 2002;22:1262–1270
Recombinant human interleukin-1 induces meningitis and 35. Voigt K, Kontush A, Stuerenburg HJ, Muench-Harrach D,
blood-brain barrier injury in the rat: characterization and Hansen HC, Kunze K. Decreased plasma and cerebrospinal
fluid ascorbate levels in patients with septic encephalopathy. 55. Munoz SJ, Robinson M, Northrup B, et al. Elevated
Free Radic Res 2002;36:735–739 intracranial pressure and computed tomography of the brain
36. Barichello T, Fortunato JJ, Vitali AM, et al. Oxidative in fulminant hepatocellular failure. Hepatology 1991;13:209–
variables in the rat brain after sepsis induced by cecal ligation 212
and perforation. Crit Care Med 2006;34:886–889 56. Wijdicks EF, Plevak DJ, Rakela J, Wiesner RH. Clinical and
37. Esen F, Erdem T, Aktan D, et al. Effect of magnesium sulfate radiologic features of cerebral edema in fulminant hepatic
administration on blood–brain barrier in a rat model of failure. Mayo Clin Proc 1995;70:119–124
intraperitoneal sepsis: a randomized controlled experimental 57. Davenport A, Will EJ, Losowsky MS. Rebound surges of
study. Crit Care 2005;9:R18–R23 intracranial pressure as a consequence of forced ultrafiltration
38. Mizock BA, Sabelli HC, Dubin A, Javaid JI, Poulos A, used to control intracranial pressure in patients with severe
Rackow EC. Septic encephalopathy: evidence for altered hepatorenal failure. Am J Kidney Dis 1989;14:516–519
phenylalanine metabolism and comparison with hepatic 58. Sen S, Rose C, Ytrebo LM, et al. Effect of albumin dialysis on
encephalopathy. Arch Intern Med 1990;150:449–449 intracranial pressure increase in pigs with acute liver failure: a
39. Sprung CL, Cerra FB, Freund HR, et al. Amino acid randomized study. Crit Care Med 2006;34:158–164
alterations and encephalopathy in the sepsis syndrome. Crit 59. Keays R, Potter D, O’Grady J, Peachey T, Alexander G,
Care Med 1991;19:753–757 Williams R. Intracranial and cerebral perfusion pressure
40. Basler T, Meier-Hellmann A, Bredle D, Reinhart K. Amino changes before, during and immediately after orthotopic liver
acid imbalance early in septic encephalopathy. Intensive Care transplantation for fulminant hepatic failure. Q J Med
Med 2002;28:293–298 1991;79:425–433
41. Eggers V, Fugener K, Hein OV, et al. Antibiotic-mediated 60. Shami VM, Caldwell SH, Hespenheide EE, et al. Recombi-
release of tumour necrosis factor alpha and norharman in nant activated factor VII for coagulopathy in fulminant
patients with hospital-acquired pneumonia and septic ence- hepatic failure compared with conventional therapy. Liver
phalopathy. Intensive Care Med 2004;30:1544–1551 Transpl 2003;9:138–143
42. Kunze K. Metabolic encephalopathies. J Neurol 2002;249: 61. Jones EA, Basile AS. Does ammonia contribute to increased
1150–1159 GABA-ergic neurotransmission in liver failure? Metab Brain
43. Towne AR, Waterhouse EJ, Boggs JG, et al. Prevalence of Dis 1998;13:351–360
nonconvulsive status epilepticus in comatose patients. Neu- 62. Ahboucha S, Butterworth RF. Role of endogenous benzo-
rology 2000;54:340–345 diazepine ligands and their GABA-A–associated receptors in
44. Bleck TP. Sepsis on the brain. Crit Care Med 2002;30:1176– hepatic encephalopathy. Metab Brain Dis 2005;20:425–437
1177 63. Bansky G, Meier PJ, Rieder E, et al. Effects of the
45. Nora PM, Bleck TP. Increased intracranial pressure compli- benzodiazepine receptor antagonist flumazenil in hepatic
cating acute hepatic failure. J Crit Illn 1989;4:87–96 encephalopathy in humans. Gastroenterology 1989;97:744–
46. O’Brien CJ, Wise RJ, O’Grady JG, Williams R. Neurological 750
sequelae in patients recovered from fulminant hepatic failure. 64. Grimm G, Ferenci P, Katzenschlager R, et al. Improvement
Gut 1987;28:93–95 of hepatic encephalopathy treated with flumazenil. Lancet
47. Norenberg MD, Rao KV, Jayakumar AR. Mechanisms of 1988;2:1392–1394
ammonia-induced astrocyte swelling. Metab Brain Dis 2005; 65. Bessman SP, Wang W, Mohan C. Ammonia inhibits insulin
20:303–318 stimulation of the Krebs cycle: further insights into the
48. Bleck TP. Neurologic consequences of fulminant hepatic mechanism of hepatic coma. Neurochem Res 1991;16:805–
failure. Mayo Clin Proc 1995;70:195–196 811
49. Canalese J, Gimson AE, Davis C, Mellon PJ, Davis M, 66. Blom HJ, Ferenci P, Grimm G, et al. The role of
Williams R. Controlled trial of dexamethasone and mannitol methanethiol in the pathogenesis of hepatic coma. Hepatol-
for the cerebral oedema of fulminant hepatic failure. Gut ogy 1991;13:445–454
1982;23:625–629 67. Bleck TP. Metabolic encephalopathy. In: Weiner WJ,
50. Nguyen JH, Yamamoto S, Steers J, et al. Matrix metal- Shulman LM, eds. Emergent and Urgent Neurology. 2nd
loproteinase-9 contributes to brain extravasation and edema in ed. Philadelphia: Lippincott; 1999:223–253
fulminant hepatic failure mice. J Hepatol 2005 Nov 7; [Epub 68. Seger DL. Flumazenil–treatment or toxin. J Toxicol Clin
ahead of print] Toxicol 2004;42:209–216
51. Rama Rao KV, Chen M, Simard JM, Norenberg MD. 69. Frucht MM, Bleck TP. Seizures in the intensive care unit
Suppression of ammonia-induced astrocyte swelling by patient. In: Delanty M, ed. Seizures: Medical Causes and
cyclosporin A. J Neurosci Res 2003;74:891–897 Management. Totowa: Humana Press; 2002:309–318
52. Ede RJ, Gimson AE, Bihari D, Williams R. Controlled 70. Singh BM, Strobos RJ. Epilepsia partialis continua associated
hyperventilation in the prevention of cerebral oedema in with nonketotic hyperglycemia: clinical and biochemical
fulminant hepatic failure. J Hepatol 1986;2:43–51 profile of 21 patients. Ann Neurol 1980;8:155–160
53. Forbes A, Alexander GJ, O’Grady JG, et al. Thiopental 71. Kumar A, Bleck TP. Intravenous midazolam for the treat-
infusion in the treatment of intracranial hypertension ment of refractory status epilepticus. Crit Care Med 1992;20:
complicating fulminant hepatic failure. Hepatology 1989; 483–488
10:306–310 72. Prasad A, Worrall BB, Bertram EH, Bleck TP. Propofol and
54. Davenport A, Will EJ, Davison AM. Effect of posture on midazolam in the treatment of refractory status epilepticus.
intracranial pressure and cerebral perfusion pressure in Epilepsia 2001;42:380–386
patients with fulminant hepatic and renal failure after 73. Rivner MH, Kim S, Greenberg M, Swift TR. Reversible
acetaminophen self-poisoning. Crit Care Med 1990;18:286– generalized paresis following hypotension: a new neurologic
289 entity. Neurology 1983;33(Suppl 2):164
74. Bolton CF, Brown JD, Sibbald WJ. The electrophysiologic 87. Van den Berghe G, Wilmer A, Hermans G, et al. Intensive
investigation of respiratory paralysis in critically ill patients. insulin therapy in the medical ICU. N Engl J Med
Neurology 1983;33(Suppl 2):186 2006;354:449–461
75. Roelofs RJ, Cerra F, Bielka N, Rosenberg L, Delaney J. 88. van den Berghe G, Wouters P, Weekers F, et al. Intensive
Prolonged respiratory insufficiency due to acute motor neuro- insulin therapy in critically ill patients. N Engl J Med 2001;
pathy: a new syndrome? Neurology 1983;33(Suppl 2):240 345:1359–1367
76. Bolton CF, Gilbert JJ, Hahn AF, Sibbald WJ. Polyneurop- 89. Bleck TP. Intensive insulin therapy reduced the incidence of
athy in critically ill patients. J Neurol Neurosurg Psychiatry neurologic complications in critically ill patients. ACP J Club
1984;47:1223–1231 2006;144:20
77. Bolton CF, Laverty DA, Brown JD, Witt NJ, Hahn AF, 90. Bleck TP. The expanding spectrum of critical illness
Sibbald WJ. Critically ill polyneuropathy: electrophysiologic polyneuropathy. Crit Care Med 1996;24:1282–1283
studies and differentiation from the Guillain-Barré syndrome. 91. Hund EF, Fogel W, Krieger D, DeGeorgia M, Hacke W.
J Neurol Neurosurg Psychiatry 1986;49:563–573 Critical illness polyneuropathy: clinical findings and outcomes
78. Zochodne DW, Bolton CF, Wells GA, et al. Critical illness of a frequent cause of neuromuscular weaning failure. Crit
polyneuropathy: a complication of sepsis and multiple organ Care Med 1996;24:1328–1333
failure. Brain 1987;110:819–842 92. Gorson KC. Approach to neuromuscular disorders in the
79. Eikermann M, Koch G, Gerwig M, et al. Muscle force and intensive care unit. Neurocritical Care 2005;3:195–212
fatigue in patients with sepsis and multiorgan failure. 93. Bolton CF. AAEM minimonograph 340: clinical neuro-
Intensive Care Med 2006;32:251–259 physiology of the respiratory system. Muscle Nerve 1993;16:
80. Segredo V, Caldwell JE, Matthay MA, et al. Persistent 809–818
paralysis in critically ill patients after long-term administra- 94. Griffin JW, Li CY, Ho TW, et al. Pathology of the motor-
tion of vecuronium. N Engl J Med 1992;327:524–528 sensory axonal Guillain-Barré syndrome. Ann Neurol 1996;
81. Zochodne DW, Ramsay DA, Saly V, Shelley S, Moffatt S. 39:17–28
Acute necrotizing myopathy of intensive care: electrophysio- 95. Kelly BJ, Matthay MA. Prevalence and severity of neurologic
logical studies. Muscle Nerve 1994;17:285–292 dysfunction in critically ill patients: influence on need for
82. Davis NA, Rodgers JE, Gonzalez ER, Fowler AA III. continued mechanical ventilation. Chest 1993;104:1818–
Prolonged weakness after cisatracurium infusion: a case 1824
report. Crit Care Med 1998;26:1290–1292 96. Rich MM, Teener JW, Raps EC, Schotland DL, Bird SJ.
83. Young GB. Neurologic complications of systemic critical Muscle is electrically inexcitable in acute quadriplegic
illness. Neurol Clin 1995;13:645–658 myopathy. Neurology 1996;46:731–736
84. Rezaiguia-Delclaux S, Lefaucheur JP, Zakkouri M, Duvoux 97. Sander HW, Golden M, Danon MJ. Quadriplegic areflexic
C, Duvaldestin P, Stephan F. Severe acute polyneuropathy ICU illness: selective thick filament loss and normal nerve
complicating orthotopic liver allograft failure. Transplanta- histology. Muscle Nerve 2002;26:499–505
tion 2002;74:880–882 98. Spitzer AR, Giancarlo T, Maher L, Awerbuch G, Bowles A.
85. Sheth RD, Bolton CF. Neuromuscular complications of Neuromuscular causes of prolonged ventilator dependency.
sepsis in children. J Child Neurol 1995;10:346–352 Muscle Nerve 1992;15:682–686
86. Bolton CF. Sepsis and the systemic inflammatory response 99. Zochodne DW, Ramsay DA, Saly V, Shelley S, Moffatt S.
syndrome: neuromuscular manifestations. Crit Care Med Acute necrotizing myopathy of intensive care: electrophysio-
1996;24:1408–1416 logical studies. Muscle Nerve 1994;17:285–292
Delirium and Cognitive Dysfunction in the
Intensive Care Unit
Russell R. Miller III, M.D., M.P.H.1 and E. Wesley Ely, M.D., M.P.H.1,2,3
ABSTRACT
Delirium remains an underrecognized, but highly prevalent, form of organ

dysfunction in the intensive care unit (ICU). Intensivists have begun to benefit from
elucidation of risk factors for delirium in the ICU, some of which are modifiable, whereas
others are not. In the last 5 years, a new tool for use in detecting delirium among critically
ill patients has been adapted, validated, and found objectively reliable for use at the bedside
by nonpsychiatrists. Moreover, that tool—the Confusion Assessment Method for the
Intensive Care Unit (CAM-ICU)—has enabled determination of the serious sequelae of
delirium, including increased mortality, higher cost, longer length of hospital stay, failure
of extubation, and burdensome long-term cognitive impairment. Although prevention and
treatment options exist, little data guide current pharmacological approaches to delirium,
and nonpharmacological approaches have yet to be fully adopted by ICUs. Ongoing trials
will address some of these limitations, but large cohort studies within the ICU are needed
to further clarify risk factors and to identify targets to modify the occurrence and course of
delirium. Furthermore, consideration of a continuum may better elucidate the true
magnitude of acute brain dysfunction in the ICU.
KEYWORDS: Delirium, aged, lorazepam, cognitive impairment, mechanical ventilation,

sedatives, analgesics, critical care
In 2002, the Society of Critical Care Medicine issue among intensive care unit (ICU) patients. From
suggested that all critically ill patients should be moni- pharmacological to nonpharmacological interventions,
tored both for level of sedation and for delirium.1 With progress seems possible. It is likely as well that subsyn-
the advent of diagnostic tools to assess for sedation and dromal delirium—that is, the presence of some delirium
delirium since the last version of this topic in Seminars in features within a patient not otherwise diagnostic of full
2001,2 the Society’s guidelines have become possible. No delirium—may represent a form of brain dysfunction
longer should delirium be a ‘‘part of the scenery,’’ as along a spectrum heretofore not described among crit-
described in 2001, because the presence of delirium using ically ill patients. This article updates the knowledge
these tools has been shown independently to predict about how to predict and measure the frequency of
myriad untoward outcomes. delirium in the ICU, what its consequences are, how
An understanding of how to prevent and treat potentially to treat or prevent it, and how to think of
delirium is only in its infancy, though numerous med- delirium within the broader context of acute brain
ications and several trials have attempted to address this dysfunction in the ICU.
1
Division of Allergy, Pulmonary, and Critical Care Medicine, Van- Medicine, Center for Health Services Research, 6th Fl. Medical Center
derbilt University School of Medicine; 2Department of Medicine, East 6100, Vanderbilt University Medical Center, Nashville, TN
Center for Health Services Research, 3Veterans Affairs Tennessee 37232-8300. E-mail: [email protected].
Valley Geriatric Research, Education, and Clinical Center (GRECC), Semin Respir Crit Care Med 2006;27:210–220. Copyright # 2006
Nashville, Tennessee. by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Address for correspondence and reprint requests: Russell R. Miller NY 10001, USA. Tel: +1(212) 584-4662.
III, M.D., M.P.H., Division of Allergy, Pulmonary, and Critical Care DOI 10.1055/s-2006-945532. ISSN 1069-3424.
210
DELIRIUM AND COGNITIVE DYSFUNCTIONIN THEICU/MILLER, ELY 211
WHAT IS DELIRIUM? became the new reference standard for delirium among
Delirium is a disturbance of consciousness that develops non-ICU patients. However, the CAM’s adaptation and
acutely, tends to fluctuate, and is associated with in- applicability to a cohort of critically ill patients had not
attention, impaired cognition, and perceptual disturban- been tested as of the previous publication of this topic in
ces. Although delirium arguably has a clear definition, Seminars in 2001. Efforts to enhance the feasibility,
there may be significant heterogeneity of the disease, validity, and reliability of the CAM among ICU patients,
suggestive of complex neuropsychiatric pathophysiology. resulting in the CAM-ICU, have since revolutionized
Attempts to better characterize delirium have resulted in and standardized the detection and monitoring of delir-
the description of motoric subtypes: hyperactive, hypo- ium among the critically ill.
active, and mixed (both hyperactive and hypoactive
features).3,4 Hyperactive delirium is characterized by
increased psychomotor activity and hallucinations or DEMOGRAPHICS OF DELIRIUM
other active perceptual disturbances. Decreased psycho- Approximately 15 to 60% of hospitalized general med-
motor activity and inattention frequently constitute ical or surgical patients become delirious, with the
hypoactive delirium. Many patients demonstrate a com- occurrence highest among older patients.8–11 The prev-
bination of the two types. For instance, in a study by alence of delirium among mechanically ventilated ICU
Peterson et al of 614 consecutive medical ICU patients patients ranges from 7 to 80%, with the highest rates
who were monitored for delirium over 1 year, hypoactive likewise among patients > 65 years old.12–25
(44%) and mixed (55%) delirium were much more Prior supposition has been that the wide variation
common than pure hyperactive delirium ( 2%). The in prevalence of delirium among cohorts of critically ill
majority of patients exhibited hypoactive delirium at patients was due to differences among patient popula-
some point during their ICU stay, especially patients tions or by different severities of illness. Three caveats
> 65 years old.5 Accordingly, prior assumptions as to the exist. First, application of tools not intended for use
manifestations of delirium, and thus the frequency with among largely nonverbal cohorts of ICU patients (such
which it has been previously detected, may be fallacious. as the original CAM) may not be appropriate. McNicoll
Inadequate recognition of a complex disease and and colleagues compared the CAM to the CAM-ICU
reliance upon imperfect tools for detection limited rec- in a study of 22 alert, nonintubated ICU patients, noting
ognition of delirium into the 21st century. For many that the two instruments agreed 82% of the time but that
years, delirium has been described in the literature using the CAM identified delirium in four patients that the
such terms as ICU psychosis, encephalopathy of critical CAM-ICU did not.23 However, the frequent use of
illness, septic encephalopathy, acute confusion, neurological sedation and analgesia, even among nonventilated ICU
impairment, and toxic confusional state, among others. patients, calls into question the validity and reliability of
However, several, significant modifications in the termi- the CAM’s verbal assessment, and the authors did not
nology regarding delirium have occurred over the last 25 employ a reference standard (i.e., formal psychiatric
or more years, prompted in part by nosological stand- assessment). Second, the most recent revision of the
ardization from the American Psychiatric Association. DSM (DSM-IV) has been found to yield a higher
Since 1980 the reference standard for diagnosis of delir- prevalence of delirium than did the DSM-IIIR.26,27
ium by psychiatrists among verbal patients has been the Third, a recent, unpublished, prospective cohort study
Diagnostic and Statistical Manual of Mental Disorders of critically ill surgical and trauma patients suggests that
(DSM). Then, in 1990 Inouye et al6 adapted DSM-IIIR7 the prevalence of delirium among that cohort is 68%
criteria for delirium for use in noncritically ill patients. (pers. comm., P. Pandharipande, Vanderbilt University
The authors formally operationalized the resultant Con- Medical Center); therefore, the prevalence of delirium
fusion Assessment Method (CAM) as an algorithm using may not, as has been suggested, be any lower among
four of the features of delirium as described in the nonmedical ICU patients. Accounting for these caveats,
DSM-IIIR: acute onset, fluctuating course, altered level the prevalence of delirium among mechanically ventilated
of consciousness, inattention, disorganized thinking, al- ICU patients appears to vary from 50% to 80%, with a
tered level of consciousness, disorientation, memory range for nonventilated ICU patients from 21 to 68%.
impairment, perceptual disturbances (e.g., hallucinations,
delusions), psychomotor retardation or agitation, inap-
propriate speech or mood, and sleep/wake cycle disturb- RISK FACTORS FOR DELIRIUM
ance. The CAM required the presence of both (1) acute Because there have been relatively few studies of risk
onset of altered mental status and fluctuation thereof and factors for delirium, numerous potential risk factors
(2) inattention and either (3) disorganized thinking or exist. Moreover, there are likely multiple factors even
(4) altered level of consciousness. Following successful within a given patient that could lead to delirium. For
validation of the CAM and determination of its reli- many patients the ICU represents the intersection
ability even when used by nonpsychiatrists, the CAM of multiorgan dysfunction necessitating mechanical
Table 1 Risk Factors for Delirium

Baseline Characteristics Disease Factors Iatrogenic/Environmental Factors
Cognitive impairment Sepsis Sedative medications

Comorbidities Hypoxemia Analgesic medications
Age Metabolic derangements Use of bladder catheter
Severity of illness score Anticholinergic medications
Sleep quality/quantity
Adapted from Pandharipande et al.27a See American Psychiatric Association, Practice guidelines for the
treatment of delirium, for more details.63
ventilation, which begets sedatives and/or narcotics to medications. It is possible that delirium represents an
ensure patient comfort and patient–ventilator synchrony. idiosyncratic reaction to psychoactive (or other) med-
Those patients who do not require mechanical ventilation ications, accounting for some of the difference in the
frequently still evidence frequent disruptions of sleep (i.e., limited data to date. Further evaluation of the role of
from alarming monitors or intravenous pumps, blood psychoactive medications commonly used in the ICU
draws, bathing, or changes in positioning to prevent is warranted to elucidate mechanisms and the role of
decubitus ulcer formation); moreover, nonventilated these medications in the development of delirium.
ICU patients not uncommonly still require sedatives or Efforts to minimize the use of sedatives and analgesics
narcotics and suffer multiple organ dysfunctions. have been strongly encouraged for ventilated patients.32
For practicality and simplicity, we divide proven Additionally, alternative medications should be sought,
or potential risk factors among three categories: (1) host such as dexmedetomidine in place of benzodiazepines
characteristics, (2) features of the acute illness, and (3) for sedation.
environmental or iatrogenic factors (Table 1).27a Another important and common risk factor,
The accumulation of risk factors likely portends though unalterable, among hospitalized patients is older
development of delirium. In separate cohorts of non-ICU age. Increasing age has been shown to independently
patients, Francis et al28 and Inouye and Charpentier29 predict delirium among non-ICU patients.33,34 How-
noted at least a 60% risk of developing delirium among ever, this relationship has been unproven among ICU
patients with three or more risk factors. Critically ill patients > 65 years old, who account for nearly 60% of all
patients easily have at least three risk factors. In fact, in a ICU patient-days.35 Some have suggested that this
study of 53 consecutive medical ICU patients, Ely et al disconnect may be due to a generally faulty assumption
discovered that each patient had a mean ( standard that the risk factors outside the ICU are the same as
deviation) of 11 ( 4) delirium risk factors, with a range those within. For example, in a cohort of 216 critically ill
of three to 17.30 The accumulation of risk factors in medical and surgical patients in whom the overall
critically ill patients undoubtedly contributes to a high prevalence of delirium was low (19%), the authors
prevalence of delirium. were unable to show that common risk factors for
Perhaps the most universal and potentially mod- delirium (e.g., age) outside the ICU were significant
ifiable risk factor among critically ill patients is expo- among the critically ill.16 Likewise in a cohort of 818
sure to psychoactive medications such as sedatives and surgical ICU patients, age was not associated with
analgesics. Ely et al identified exposure to benzodiaze- delirium using logistic regression. However, this pop-
pines or narcotics in 98% of patients in an ICU ulation was young compared with typical ICU popula-
cohort.30 In a cohort of 216 medical and surgical ICU tions.13 Moreover, both of these studies suffered from
patients, Dubois et al found that morphine was the insufficient power or methodological problems, poten-
strongest predictor of delirium in a multivariable model, tially limiting their ability to detect differences among
with an increase in odds of at least sixfold over patients > 65 years old. Finally, in 2006 Pandharipande
5 months.16 Additionally, in a cohort of 198 mechanically et al demonstrated that age significantly increases risk
ventilated medical and cardiac ICU patients, Pandhar- for development of ICU delirium.31 The authors noted
ipande et al found that lorazepam had an independent a statistically significant, 2% increase in probability of
and dose-related temporal association with delirium.31 transition to delirium for each year beyond age 65, even
The risk for daily transition to delirium increased by 20% after controlling for relevant covariates.
per milligram of lorazepam [odds ratio (OR) ¼ 1.2 per
mg; 95% confidence interval (CI), 1.1 to 1.4; p ¼ .003]
after adjusting for 11 relevant covariates. Despite a trend HOW DO WE MEASURE DELIRIUM
toward significance, midazolam and fentanyl did not IN THE ICU?
significantly portend the same risk, likely because the At best, and despite its prevalence, medical doctors and
study was underpowered to detect a difference with these intensivists have markedly underdiagnosed delirium
Figure 1 Diagnosis of delirium with the Confusion Assessment Method for the intensive care unit (CAM-ICU). Adapted from Inouye
et al6 and copyright # 2002, E. Wesley Ely, M.D., M.P.H. and Vanderbilt University, all rights reserved. See http://www.icudelirium.org
for more information.
since its nosological inception in 1980.36–38 After the appropriate for use in screening for (rather than diag-
1987 revision of the DSM, Inouye et al6 developed and nosing) delirium.
validated a tool (the CAM) to assess hospitalized pa- Developed in 2001, the CAM-ICU is the only
tients for delirium. Other tools for use outside the ICU delirium assessment tool for use in critically ill patients
were subsequently developed and tested with varying that has been validated against a reference standard rater
degrees of success.39,40 However, they were not practical in multiple cohorts.14,15,21 Building upon the work of
for use in critically ill patients because they required Inouye et al,6 Hart et al,41,42 and others43 as well as the
verbal responses, took at least 5 to 10 minutes to DSM-IV,44 Ely et al14 created the CAM-ICU (Fig. 1)
complete, were not found valid and reliable among for use by nonpsychiatrists in mechanically ventilated
ICU patients, or were not demonstrated to be associated patients. When features 1 and 2 and either feature 3 or
with clinical outcomes. Hence, delirium continued to be feature 4 are present, a patient is said to be delirious, or
underrecognized in the ICU population. ‘‘CAM-ICU positive.’’ In the largest validation cohort of
Efforts to create an instrument to objectively 111 medical ICU patients using the CAM-ICU and as
evaluate for delirium among ICU patients began in compared with reference standard raters, two study
1996 and led to the development and validation of the nurses demonstrated high sensitivity (93 to 100%),
Cognitive Test for Delirium41,42 and the Intensive high specificity (98 to 100%), and high interrater reli-
Care Delirium Screening Checklist.17 The abbreviated ability (k ¼ 0.96; 95% CI, 0.92 to 0.99).14 Importantly,
version of the Cognitive Test for Delirium was pre- the CAM-ICU also distinguishes delirium from
sented in 1997 by Hart et al42 after evaluation in 19 coma—defined as the state of unarousability, unaware-
delirious patients (15 critically ill). Although it took but ness of the environment, and absence of spontaneous
a few minutes to complete and was shown to discrim- interaction or awareness of the interviewer—and de-
inate well among delirium, dementia, depression, and mentia. Requiring on average less than 2 minutes to
schizophrenia (p < .0001), the abbreviated Cognitive complete, the CAM-ICU has been validated, and its
Test for Delirium has not been prospectively validated reliability has been confirmed in other languages and at
in a cohort of ICU patients. In contrast, when the least one other region of the world.21 Its ease of use
Intensive Care Delirium Screening Checklist was first among nonpsychiatrists also makes the CAM-ICU
reported in 2001, it was heralded as valid in a general practical.45 Subsequent revisions of the CAM-ICU,
population of ICU patients. Using a consultant psy- including use of the validated Richmond Agitation
chiatrist as the reference standard rater and relying Sedation Scale to monitor level of consciousness46–48
largely upon chart review, the checklist of eight differ- and a greater reliance upon a form of the Vigilance A
ent delirium features derived from DSM criteria pre- random letter test43 rather than the picture recognition
dicted delirium with reported 99% sensitivity but only tool for the attention screening examination, are being
64% specificity, using a cutoff of four or more features incorporated in ongoing clinical trials. The development
as indicating delirium. Because of the low specificity, of a simple, objective, brief, valid, reliable, and widely
the authors concluded that the checklist was most accepted, bedside assessment tool for critically ill, often
Figure 2 Kaplan-Meier analysis of delirium in the intensive care unit and 6-month survival. Adapted from Ely et al.20
mechanically ventilated, patients has revolutionized discharge and noted a threefold increase in the risk of
the detection of delirium but also revealed its serious death by 6 months, even after adjustment for age,
sequelae. severity of illness, comorbid conditions, coma, and the
use of sedatives and analgesics (Fig. 2, adjusted hazard
ratio ¼ 3.2; 95% CI, 1.4 to 7.7; p ¼ 0.008).20
OUTCOMES AND PROGNOSTIC
SIGNIFICANCE OF DELIRIUM IN THE
INTENSIVE CARE UNIT Cost
With an effective tool to diagnose delirium, the next step Costs of care in the ICU are significantly higher among
was to determine the independent association of delir- those who develop delirium than in those who do not.
ium with multiple, clinically significant end points. In Milbrandt et al reported in 2004 that the costs to care for
the last 5 years, much has been learned about the impact a delirious, critically ill patient were significantly higher
of delirium both during hospitalization and for up to than for patients who never became delirious, even after
2 years later. On a daily basis, delirium and/or altered controlling for several covariates.51 The costs of caring
levels of consciousness contribute to an increased risk of for delirious, critically ill patients were a median $9014
complications related to the complex bedside monitor- and $14,730 higher for ICU and hospital stays, respec-
ing, drug delivery,16 and life-supporting therapies that tively. At least part of this increase would likely be
are commonplace in ICUs. Given that most patients accounted for by longer lengths of stay.
who develop delirium survive to hospital discharge, these
long-term sequelae will likely prove to be more burden-
some than currently imagined. Length of Stay
That ICU delirium independently prolongs hospital
length of stay has been demonstrated in two cohorts.
Mortality The first of these, a prospective cohort of 275 mechan-
Increased risk of death among delirious, noncritically ill ically ventilated ICU patients, demonstrated an ad-
patients for up to 2 years has been previously identi- justed hazard ratio of 2.0 ( p < .001) for longer
fied.49,50 In the last 5 years, additional data on critically hospital stay among delirious patients, even after
ill patients have shown a similar impact of delirium, controlling for covariates.20 Similarly, in a cohort of
though not always during the immediate hospitalization 261 consecutive, non-intubated medical ICU patients,
period.12,16,24 Lin et al demonstrated a 13-fold increased patients who had experienced delirium in the ICU had
risk for in-hospital death among 111 mechanically a 41% greater risk of remaining in the hospital relative
ventilated patients.21 Furthermore, Ely et al prospec- to non-delirious patients (p ¼ .023).25 Whether intu-
tively followed 275 mechanically ventilated, medical bated or not, it appears delirium may prolong length of
ICU patients for 6 months from the time of hospital hospital stay.
Failure of Extubation to do so frequently. However, for them to take their

As is the case with longer ICU and hospital stays among place alongside such stalwart, multidisciplinary inter-
delirious patients, the potential for delirium to contrib- ventions in the ICU as deep venous thrombosis preven-
ute to extubation failure necessarily subjects patients to tion or antibiotics, evidence that altering the occurrence
risk of complications. For example, failure of extubation or course of delirium decreases the frequency of unto-
poses risks of prolonged ventilation or of reintubation, ward outcomes is essential. Likely, this will require
including such complications as nosocomial pneumo- targeted prevention strategies and specialized, multi-
nia52 and death.53 That abnormal mental status signifi- disciplinary interventions, both nonpharmacological
cantly predicts failure of extubation has been previously and pharmacological.
studied.54,55 Nearly 10% of delirious, intubated patients
in one study versus 2.3% of nondelirious, intubated
patients self-extubated.16 Recently we have also demon- Nonpharmacological Approach
strated that the presence of delirium, specifically as Although several clinical trials of multicomponent, non-
determined by the CAM-ICU, is associated with a pharmacological interventions have been per-
threefold increased risk for reintubation within 48 hours formed,11,60,61 none targeted the ICU and all have met
of predominantly planned extubations.56 with modest, postdischarge success. The largest such
trials in the literature are those conducted by Inouye
et al11 and Lundström et al.61 The former, a non-
Long-Term Cognitive Impairment randomized trial of patients > 70 years old admitted
In addition to the predisposition of cognitive impair- either to a specialized ward or to a regular unit, featured
ment among critically ill patients for the development an intervention protocol. The protocol targeted such
of delirium, this condition may also independently features as cognitive stimulation, reorientation prompts,
cause or result in greater occurrence and severity of a sleep protocol, visual and hearing aids, reminders to
cognitive impairment for up to 3 years after ICU stay. prevent volume depletion, and walking/exercise. The
In a review of nine prospective studies evaluating a incidence of delirium among the intervention as com-
total of 1885 hospitalized medical and surgical (non- pared with those who received usual care was signifi-
ICU) patients, Jackson et al found that intrahospital cantly lower, 9.9% versus 15.0%, respectively (p ¼ .02).
delirium increased the likelihood of dementia as much Unfortunately, neither the severity of delirium nor the
as threefold up to 3 years from the time of hospital recurrence rates differed between the two groups, and
discharge.57 A smaller study that stringently examined subsequent follow-up of the patients 6 months after
34 previously mechanically ventilated patients demon- hospital discharge did not demonstrate sustained bene-
strated that, although ‘‘normal’’ at baseline almost one fits overall.62 However, the highest-risk delirious pa-
third were neuropsychologically impaired by standar- tients (arguably those most similar to critically ill
dized neuropsychological batteries at 6 months follow- patients) did report higher health and functional status
ing hospital discharge.58 More recently, Hopkins et al scores at 6 months than did high-risk, nondelirious
have confirmed the high prevalence of cognitive im- patients. The more recent trial by Lundström et al
pairment among previously critically ill patients 2 years incorporating a nurse-driven, multifactorial intervention
after discharge from the hospital.59 An area ripe for program among 400 hundred patients > 70 years old
future investigation, the likely role of delirium in resulted in significant reductions in duration of delirium
contributing to long-term cognitive impairment sug- (30% absolute risk reduction, p ¼ .001) and hospital
gests the need for larger, prospective cohort studies to length of stay (3 day reduction, p < .001).61 We hope
better identify risk factors for persistence of neuro- long-term follow-up will yield promising results and
cognitive dysfunction. As with mortality, cost, length anticipate that interventions to prevent or diminish the
of stay, and failure of extubation, the presence of long- consequences of delirium among both noncritically ill
term cognitive impairment among critically ill patients and ICU patients will remain fundamental to manage-
represents a significant, lasting burden for patients and ment of delirium in the future.
their families.
Drug Therapy
STRATEGIES FOR OPTIMAL That nonpharmacological interventions to prevent a
MANAGEMENT disease are prudent is no surprise. However, with a
With the development of effective diagnostic tools such widening array of medications at our disposal, physi-
as the CAM-ICU and the Richmond Agitation Seda- cians are tempted to act when faced with medical
tion Scale, the Society of Critical Care Medicine’s abnormalities. Accordingly, the temptation to seek
recommendations to monitor for delirium and sedation pharmacological interventions for diseases such as de-
level in all critically ill patients are easier to follow—and lirium in the ICU frequently outpaces the scientific
evidence to support their use. One need look no further plasma concentrations are typically reached within 5 to 8
than a 2004 survey of 912 intensivists, of whom 92% hours following enteral ingestion, though risperidone
considered delirium a substantial problem in the ICU reaches peak concentration within 1 hour, or within
despite only 40% routinely screening for it. Yet, 79% 1 hour for drugs administered intramuscularly. In con-
reported that delirium requires active intervention and trast to haloperidol, the atypical antipsychotics cause few
66% felt haloperidol should be the treatment of choice, side effects usually, though weight gain and hypotension
followed by lorazepam (12%) and atypical antipsy- are not uncommon, and there may be an increased risk of
chotics (4%). Despite this impulse to intervene phar- hyperglycemia or diabetes.75 The risk of extrapyramidal
macologically with haloperidol and the support of this symptoms or neuroleptic malignant syndrome is lower
position by both the Society of Critical Care Medicine1 than with haloperidol.
and the American Psychiatric Association,63 there are Data supporting the use of haloperidol either
no randomized, placebo-controlled trials to confirm alone or versus other medications is limited. Evidence
the efficacy of haloperidol in either the prevention or for the potential benefit of haloperidol, however, was
the treatment of delirium. Not surprisingly, no drugs recognized by Milbrandt et al. In a cohort of 989
have Food and Drug Administration approval for the mechanically ventilated ICU patients, those who re-
prevention or the treatment of delirium. Through 2005, ceived haloperidol within the first 48 hours were 16%
anecdotal evidence, uncontrolled trials, and a few less likely to die during the hospitalization.76 One
randomized trials comparing haloperidol to neurolep- explanation for this finding is the purported antiinflam-
tics or benzodiazepines constitute the basis for profes- matory effect of haloperidol.
sional societies’ recommendation of haloperidol. The only prospective trials to our knowledge to
Medications such as haloperidol and the so-called evaluate the efficacy of treatment of delirium in the ICU
atypical antipsychotics (e.g., olanzapine, risperidone, with antipsychotics are fraught with limitations but
ziprasidone, aripiprazole, quetiapine) are thought to provide reassurance as to the safety of the atypical
exert their antidelirium effect in at least two ways. First, antipsychotics. One unblinded study evaluated the effi-
the drugs are thought to ‘‘normalize’’ cerebral function by cacy and safety of enteral olanzapine as compared with
disinhibition of acetylcholine, blockade of dopamine enteral haloperidol in the treatment of 73 ICU patients
receptors, and activation of serotonin receptors. Second, who screened positive for delirium according to DSM-
some data suggest that haloperidol may exhibit antiin- IV criteria.22 Although there was no difference in the
flammatory effects upon the production of proinflam- development of delirium between the treatment groups,
matory cytokines.64,65 the study confirmed the safety of olanzapine because
The atypical antipsychotics do not have intra- none of the patients in the olanzapine arm but six
venous formulations. As such, those who have or are patients in the haloperidol arm experienced extrapyra-
investigating haloperidol as compared with an atypical midal symptoms. In addition to a simplistic random-
antipsychotic must rely upon either enteral or intra- ization scheme, a predominance of surgical ICU
muscular administration. Nonetheless, intravenous hal- patients, and a low severity of illness score (mean Acute
operidol is not only common but recommended,1,66,67 Physiology and Chronic Health II score of 12.7), the low
with a protocol of escalating dose (e.g., doubling every overall prevalence of delirium (21%) relative to other
30 minutes until desired effect) noted in professional ICU cohorts calls into question the generalizability of
guidelines.1 this study’s results. Although the trial did not intend to
Pharmacokinetics is important with antipsy- answer the question of whether olanzapine is equal or
chotics, as with all drugs, in portending risk of adverse superior to haloperidol, it did suggest that the former is
events. The half-life of haloperidol is 21 hours, with no more, and perhaps less, harmful. Moreover, similar
peak plasma concentrations noted within 2 to 6 hours results were found in a separate trial comparing haloper-
of dosing (enteral) or 20 minutes (intramuscular). idol to risperidone.77 Again, delirium scores decreased
Notable adverse effects can include hypotension that during drug administration in both groups (p < .05), but
antagonizes adrenaline (especially in parenteral there was no difference in the scores between the two
form), and dose-dependent QTc prolongation leading treatment groups (p ¼ .51). A placebo-controlled study
to cardiac tachyarrhythmias such as torsades de comparing placebo to haloperidol and to an atypical
pointes68–70 —particularly among patients with preex- antipsychotic would most appropriately address this
isting cardiac disease, those receiving other medica- question.
tions that prolong the QTc, those receiving > 35 mg As suggested by the first admonition of the
cumulative dose, and those with extrapyramidal symp- Hippocratic oath, it may be as important what medi-
toms,71 neuroleptic malignant syndrome,72 dyspho- cations we do not use as those we do use to alter the
ria,73 or laryngospasm.74 Meanwhile, the atypical occurrence and course of delirium. Almost all ICU
antipsychotics typically have half-lives of 20 þ hours, patients require and receive sedatives and analgesics,
with the exception of ziprasidone ( 7 hours). Peak particularly during the early part of their ICU stay, and
Figure 3 Proposed continuum of acute brain dysfunction in the ICU.
avoidance of harmful medications during that time at postacute skilled nursing facilities. They found that
seems prudent. Little evidence exists to date to guide patients with subsyndromal delirium had intermediate
this type of decision, but in a study by Breitbart et al,78 6-month mortality rates (25.0, 18.3, and 5.7% for
the authors randomized 30 hospitalized and subse- delirious, subsyndromal delirium, and ‘‘normal’’ patients,
quently delirious acquired immunodeficiency syn- respectively), intermediate rehospitalization rates, and
drome patients to receive either lorazepam (n ¼ 6), intermediate rates of complications, even after adjusting
chlorpromazine (n ¼ 13), or haloperidol (n ¼ 11).78 for age, preexisting dementia, and medical comorbidity.
They found that either haloperidol or chlorpromazine Identification of similar phenomenonology of acute
significantly reduced the delirium score as compared brain dysfunction in the ICU might help better target
with lorazepam (p < .001), leading the authors to therapeutic interventions.
suggest that lorazepam alone is ineffective in decreas-
ing delirium symptoms. That it is a known risk factor
further validates the assertion that outside well-known SUMMARY
drug withdrawal syndromes, benzodiazepines are not Since 2001 our knowledge of delirium in the ICU has
recommended for routine treatment of delirium. changed dramatically. There has been greater under-
Moreover, when used, they should be used in the standing, if persistent underappreciation, of the occur-
context of sedation protocols that employ intermittent rence of delirium among the critically ill. Also efforts to
bolus sedation, if at all possible.32,79,80 identify risk factors for delirium in the ICU have
Although minimal data beyond consensus opinion increased and are suggesting targets for refinements of
exist currently to guide prevention or treatment decisions ICU practice that will hopefully diminish the sequelae of
in ICU delirium, the CAM-ICU provides an effective delirium. With the additional development of the
way to evaluate the safety and efficacy of nonpharmaco- CAM-ICU, the independent association of delirium
logical and pharmacological interventions alike. Several, with untoward outcomes from mortality to long-term
ongoing, randomized, and/or placebo-controlled trials cognitive impairment has been possible to describe.
should begin to clarify the indications for use of the Now, further efforts to determine appropriate interven-
typical and atypical antipsychotics and the indications for tions and management of delirium must come to the
use of agents other than benzodiazepines for sedation. fore. Placebo-controlled trials for treatment of delirium
are on the horizon. In avoiding harmful medications or
pursuing beneficial ones, the prevention and treatment
SPECTRUM OF ACUTE BRAIN of delirium or any of the various subtypes along the
DYSFUNCTION continuum of acute brain dysfunction in the ICU will
In addition to delirium, other forms of brain dysfunction require a multidisciplinary approach. Diligent investiga-
such as coma, stupor, and the more recently described tion, including large cohort studies, will help identify
subsyndromal delirium may represent a continuum of potential new targets for intervention in the years ahead.
acute brain dysfunction (Fig. 3). Subsyndromal delirium
is perhaps best defined as the presence of some, but not
all, of the criteria for delirium.81–84 In the case of the GRANT SUPPORT
CAM-ICU, that could mean disorganized thinking Dr. Ely is a recipient of a K23 from the National
despite normal levels of consciousness and attention or Institute of Health (#AG01023–01A1).
a fluctuating level of consciousness despite preserved
attention and thinking. Distinction of these subtypes
of brain dysfunction may be pertinent in light of the
prevailing notion that increasing severity of delirium is REFERENCES
associated with worse outcomes outside the ICU.81–83
1. Jacobi J, Fraser GL, Coursin DB, et al. Clinical practice
For example, Marcantonio et al83 compared 504 patients guidelines for the sustained use of sedatives and analgesics in
who remained CAM negative (not delirious) or who the critically ill adult. Crit Care Med 2002;30:119–141
became CAM positive (delirious) or CAM intermediate 2. Ely EW, Siegel MD, Inouye SK. Delirium in the intensive
(subsyndromal delirium, those with one or more CAM care unit: an under-recognized syndrome of organ dysfunc-
features but not diagnostic of delirium) during their stay tion. Semin Respir Crit Care Med 2001;22:115–126
3. Meagher DJ, Hanlon DO, Mahony EO, Casey PR, 22. Skrobik YK, Bergeron N, Dumont M, Gottfried SB.
Trzepacz PT. Relationship between symptoms and motoric Olanzapine vs haloperidol: treating delirium in a critical care
subtype of delirium. J Neuropsychiatry Clin Neurosci 2000; setting. Intensive Care Med 2004;30:444–449
12:51–56 23. McNicoll L, Pisani MA, Ely EW, Gifford D, Inouye SK.
4. Lipowski ZJ. Delirium: Acute Confusional States. Rev. ed. Detection of delirium in the intensive care unit: comparison of
New York: Oxford University Press; 1990 confusion assessment method for the intensive care unit with
5. Peterson JF, Pun BT, Dittus RS, et al. Delirium and its confusion assessment method ratings. J Am Geriatr Soc 2005;
motoric subtypes: a study of 614 critically ill patients. J Am 53:495–500
Geriatr Soc 2006;54:479–484 24. Micek ST, Anand NJ, Laible BR, Shannon WD, Kollef MH.
6. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Delirium as detected by the CAM-ICU predicts restraint use
Horwitz RI. Clarifying confusion: the confusion assessment among mechanically ventilated medical patients. Crit Care
method: a new method for detection of delirium. Ann Intern Med 2005;33:1260–1265
Med 1990;113:941–948 25. Thomason JW, Shintani A, Peterson JF, Pun BT, Jackson JC,
7. American Psychiatric Association. Diagnostic and Statistical Ely EW. Intensive care unit delirium is an independent
Manual of Mental Disorders. 3rd rev. ed. Washington, DC: predictor of longer hospital stay: a prospective analysis of 261
American Psychiatric Association; 1987 non-ventilated patients. Crit Care 2005;9:R375–R381
8. O’Keeffe S, Lavan J. The prognostic significance of delirium 26. Cole MG, Dendukuri N, McCusker J, Han L. An empirical
in older hospital patients. J Am Geriatr Soc 1997;45:174– study of different diagnostic criteria for delirium among
178 elderly medical inpatients. J Neuropsychiatry Clin Neurosci
9. Levkoff SE, Evans DA, Liptzin B, et al. Delirium: the 2003;15:200–207
occurrence and persistence of symptoms among elderly 27. Laurila JV, Pitkala KH, Strandberg TE, Tilvis RS. Delirium
hospitalized patients. Arch Intern Med 1992;152:334–340 among patients with and without dementia: does the
10. Williams-Russo P, Urquhart BL, Sharrock NE, Charlson ME. diagnosis according to the DSM-IV differ from the previous
Postoperative delirium: predictors and prognosis in elderly classifications? Int J Geriatr Psychiatry 2004;19:271–277
orthopedic patients. J Am Geriatr Soc 1992;40:759–767 27a. Pandharipande P, Jackson J, Ely EW. Delirium: acute
11. Inouye SK, Bogardus ST Jr, Charpentier PA, et al. A cognitive dysfunction in the critically ill. Curr Opin Crit
multicomponent intervention to prevent delirium in hospi- Care 2005;11:360–368
talized older patients. N Engl J Med 1999;340:669–676 28. Francis J, Martin D, Kapoor WN. A prospective study of
12. Kishi Y, Iwasaki Y, Takezawa K, Kurosawa H, Endo S. delirium in hospitalized elderly. JAMA 1990;263:1097–
Delirium in critical care unit patients admitted through an 1101
emergency room. Gen Hosp Psychiatry 1995;17:371–379 29. Inouye SK, Charpentier PA. Precipitating factors for delirium
13. Aldemir M, Özen S, Kara IH, Sir A, Baç B. Predisposing in hospitalized elderly persons: predictive model and inter-
factors for delirium in the surgical intensive care unit. Crit relationship with baseline vulnerability. JAMA 1996;275:
Care 2001;5:265–270 852–857
14. Ely EW, Inouye SK, Bernard GR, et al. Delirium in 30. Ely EW, Gautam S, Margolin R, et al. The impact of
mechanically ventilated patients: validity and reliability of delirium in the intensive care unit on hospital length of stay.
the confusion assessment method for the intensive care unit Intensive Care Med 2001;27:1892–1900
(CAM-ICU). JAMA 2001;286:2703–2710 31. Pandharipande PP, Shintani A, Peterson J, et al. Lorazepam
15. Ely EW, Margolin R, Francis J, et al. Evaluation of delirium is an independent risk factor for transitioning to delirium
in critically ill patients: validation of the Confusion Assess- in intensive care unit patients. Anesthesiology 2006;104:21–
ment Method for the Intensive Care Unit (CAM-ICU). Crit 26
Care Med 2001;29:1370–1379 32. Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily
16. Dubois MJ, Bergeron N, Dumont M, Dial S, Skrobik Y. interruption of sedative infusions in critically ill patients
Delirium in an intensive care unit: a study of risk factors. undergoing mechanical ventilation. N Engl J Med 2000;342:
Intensive Care Med 2001;27:1297–1304 1471–1477
17. Bergeron N, Dubois MJ, Dumont M, Dial S, Skrobik Y. 33. Schor JD, Levkoff SE, Lipsitz LA, et al. Risk factors for
Intensive Care Delirium Screening Checklist: evaluation of delirium in hospitalized elderly. JAMA 1992;267:827–831
a new screening tool. Intensive Care Med 2001;27:859– 34. Marcantonio ER, Goldman L, Mangione CM, et al. A
864 clinical prediction rule for delirium after elective noncardiac
18. Rincon HG, Granados M, Unutzer J, et al. Prevalence, surgery. JAMA 1994;271:134–139
detection and treatment of anxiety, depression, and delirium 35. Angus DC, Kelley MA, Schmitz RJ, White A, Popovich J Jr.
in the adult critical care unit. Psychosomatics 2001;42:391– Caring for the critically ill patient: current and projected
396 workforce requirements for care of the critically ill and
19. McNicoll L, Pisani MA, Zhang Y, Ely EW, Siegel MD, patients with pulmonary disease: can we meet the require-
Inouye SK. Delirium in the intensive care unit: occurrence ments of an aging population? JAMA 2000;284:2762–
and clinical course in older patients. J Am Geriatr Soc 2003; 2770
51:591–598 36. Kakuma R, du Fort GG, Arsenault L, et al. Delirium in older
20. Ely EW, Shintani A, Truman B, et al. Delirium as a predictor emergency department patients discharged home: effect on
of mortality in mechanically ventilated patients in the survival. J Am Geriatr Soc 2003;51:443–450
intensive care unit. JAMA 2004;291:1753–1762 37. Inouye SK, Foreman MD, Mion LC, Katz KH, Cooney LM
21. Lin SM, Liu CY, Wang CH, et al. The impact of delirium on Jr. Nurses’ recognition of delirium and its symptoms: com-
the survival of mechanically ventilated patients. Crit Care parison of nurse and researcher ratings. Arch Intern Med 2001;
Med 2004;32:2254–2259 161:2467–2473
38. Ely EW, Stephens RK, Jackson JC, et al. Current opinions 56. Miller RR, Shintani AK, Girard TD, et al. Delirium predicts
regarding the importance, diagnosis, and management of extubation failure [abstract]. Proceedings of the American
delirium in the intensive care unit: a survey of 912 healthcare Thoracic Society 2006;3:A42
professionals. Crit Care Med 2004;32:106–112 57. Jackson JC, Gordon SM, Hart RP, Hopkins RO, Ely EW.
39. Rapp CG, Wakefield B, Kundrat M, et al. Acute confusion The association between delirium and cognitive decline: a
assessment instruments: clinical versus research usability. review of the empirical literature. Neuropsychol Rev 2004;14:
Appl Nurs Res 2000;13:37–45 87–98
40. Breitbart W, Rosenfeld B, Roth A, Smith MJ, Cohen K, 58. Jackson JC, Hart RP, Gordon SM, et al. Six-month
Passik S. The Memorial Delirium Assessment Scale. J Pain neuropsychological outcome of medical intensive care unit
Symptom Manage 1997;13:128–137 patients. Crit Care Med 2003;31:1226–1234
41. Hart RP, Levenson JL, Sessler CN, Best AM, Schwartz 59. Hopkins RO, Brett S. Chronic neurocognitive effects of
SM, Rutherford LE. Validation of a cognitive test for critical illness. Curr Opin Crit Care 2005;11:369–375
delirium in medical ICU patients. Psychosomatics 1996;37: 60. Marcantonio ER, Flacker JM, Wright RJ, Resnick NM.
533–546 Reducing delirium after hip fracture: a randomized trial. J Am
42. Hart RP, Best AM, Sessler CN, Levenson JL. Abbreviated Geriatr Soc 2001;49:516–522
cognitive test for delirium. J Psychosom Res 1997;43:417– 61. Lundström M, Edlund A, Karlsson S, Brannstrom B, Bucht
423 G, Gustafson Y. A multifactorial intervention program
43. Strub RL, Black FW. The Mental Status Examination in reduces the duration of delirium, length of hospitalization,
Neurology. 4th ed. Philadelphia: FA Davis; 2000 and mortality in delirious patients. J Am Geriatr Soc 2005;
44. American Psychiatric Association. Delirium, dementia and 53:622–628
amnestic and other cognitive disorders. In: Diagnostic and 62. Bogardus ST Jr, Desai MM, Williams CS, Leo-Summers L,
Statistical Manual of Mental Disorders. 4th ed. Washington, Acampora D, Inouye SK. The effects of a targeted multi-
DC: American Psychiatric Association. 1994 component delirium intervention on postdischarge out-
45. Pun BT, Gordon SM, Peterson JF, et al. Large-scale comes for hospitalized older adults. Am J Med 2003;114:
implementation of sedation and delirium monitoring in the 383–390
intensive care unit: a report from two medical centers. Crit 63. American Psychiatric Association. Practice guideline for the
Care Med 2005;33:1199–1205 treatment of patients with delirium. Am J Psychiatry 1999;
46. Ely EW, Gautam S, May L, et al. A comparison of different 156(Suppl 5):1–20
sedation scales in the ICU and validation of the Richmond 64. Moots RJ, Al Saffar Z, Hutchinson D, et al. Old drug, new
Agitation Sedation Scale (RASS) [abstract]. Am J Respir Crit tricks: haloperidol inhibits secretion of proinflammatory
Care Med 2001;163:A954 cytokines. Ann Rheum Dis 1999;58:585–587
47. Ely EW, Truman B, Shintani A, et al. Monitoring sedation 65. Song C, Lin A, Kenis G, Bosmans E, Maes M. Immuno-
status over time in ICU patients: reliability and validity of the suppressive effects of clozapine and haloperidol: enhanced
Richmond Agitation-Sedation Scale (RASS). JAMA 2003; production of the interleukin-1 receptor antagonist. Schiz-
289:2983–2991 ophr Res 2000;42:157–164
48. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond 66. Tesar GE, Stern TA. Rapid tranquilization of the agitated
Agitation-Sedation Scale: validity and reliability in adult intensive care unit patient. J Intensive Care Med 1988;3:195–
intensive care unit patients. Am J Respir Crit Care Med 201
2002;166:1338–1344 67. Tesar GE, Murray GB, Cassem NH. Use of high-dose
49. Francis J, Kapoor WN. Prognosis after hospital discharge of intravenous haloperidol in the treatment of agitated cardiac
older medical patients with delirium. J Am Geriatr Soc 1992; patients. J Clin Psychopharmacol 1985;5:344–347
40:601–606 68. Lawrence KR, Nasraway SA. Conduction disturbances
50. McCusker J, Cole M, Abrahamowicz M, Primeau F, Belzile associated with administration of butyrophenone antipsy-
E. Delirium predicts 12-month mortality. Arch Intern Med chotics in the critically ill: a review of the literature.
2002;162:457–463 Pharmacotherapy 1997;17:531–537
51. Milbrandt EB, Deppen S, Harrison PL, et al. Costs associated 69. Perrault LP, Denault AY, Carrier M, Cartier R, Belisle S.
with delirium in mechanically ventilated patients. Crit Care Torsades de pointes secondary to intravenous haloperidol after
Med 2004;32:955–962 coronary bypass grafting surgery. Can J Anaesth 2000;47:251–
52. Torres A, Gatell JM, Aznar E. Reintubation increases the 254
risk of nosocomial pneumonia in patients needing mechan- 70. Sharma ND, Rosman HS, Padhi ID, Tisdale JE. Torsades de
ical ventilation. Am J Respir Crit Care Med 1995;152: pointes associated with intravenous haloperidol in critically ill
137–141 patients. Am J Cardiol 1998;81:238–240
53. Esteban A, Alia I, Gordo F, et al. Extubation outcome 71. Bashford G, Bradd P. Drug-induced Parkinsonism associated
after spontaneous breathing trials with T-tube or pressure with dysphagia and aspiration: a brief report. J Geriatr Psy-
support ventilation. The Spanish Lung Failure Collabora- chiatry Neurol 1996;9:133–135
tive Group. Am J Respir Crit Care Med 1997;156(2 Pt 1): 72. Adnet P, Lestavel P, Krivosic-Horber R. Neuroleptic
459–465 malignant syndrome. Br J Anaesth 2000;85:129–135
54. Namen AM, Ely EW, Tatter SB, et al. Predictors of 73. King DJ, Burke M, Lucas RA. Antipsychotic drug-induced
successful extubation in neurosurgical patients. Am J Respir dysphoria. Br J Psychiatry 1995;167:480–482
Crit Care Med 2001;163(3 Pt 1):658–664 74. Ilchef R. Neuroleptic-induced laryngeal dystonia can mimic
55. Salam A, Tilluckdharry L, Amoateng-Adjepong Y, Man- anaphylaxis. Aust N Z J Psychiatry 1997;31:877–879
thous CA. Neurologic status, cough, secretions and extuba- 75. Lindenmayer JP, Czobor P, Volavka J, et al. Changes in
tion outcomes. Intensive Care Med 2004;30:1334–1339 glucose and cholesterol levels in patients with schizophrenia
treated with typical or atypical antipsychotics. Am J Psychiatry 80. Kollef MH, Levy NT, Ahrens TS, Schaiff R, Prentice D,
2003;160:290–296 Sherman G. The use of continuous I.V. sedation is associated
76. Milbrandt EB, Kersten A, Kong L, et al. Haloperidol use is with prolongation of mechanical ventilation. Chest 1998;114:
associated with lower hospital mortality in mechanically 541–548
ventilated patients. Crit Care Med 2005;33:226–229 81. Levkoff SE, Yang FM, Liptzin B. Delirium: the impor-
77. Han CS, Kim YK. A double-blind trial of risperidone and tance of subsyndromal states. Primary Psychiatry 2004;11:
haloperidol for the treatment of delirium. Psychosomatics 40–44
2004;45:297–301 82. Cole M, McCusker J, Dendukuri N, Han L. The prognostic
78. Breitbart W, Marotta R, Platt MM, et al. A double-blind significance of subsyndromal delirium in elderly medical
trial of haloperidol, chlorpromazine, and lorazepam in the inpatients. J Am Geriatr Soc 2003;51:754–760
treatment of delirium in hospitalized AIDS patients. Am J 83. Marcantonio ER, Kiely DK, Simon SE, et al. Outcomes of
Psychiatry 1996;153:231–237 older people admitted to postacute facilities with delirium.
79. Brook AD, Ahrens TS, Schaiff R, et al. Effect of a nursing- J Am Geriatr Soc 2005;53:963–969
implemented sedation protocol on the duration of mechanical 84. Liptzin B. What criteria should be used for the diagnosis of
ventilation. Crit Care Med 1999;27:2609–2615 delirium? Dement Geriatr Cogn Disord 1999;10:364–367
Cardiac Arrhythmias in the Intensive Care Unit
Daniel J. Tarditi, D.O.1 and Steven M. Hollenberg, M.D.1
ABSTRACT
Cardiac arrhythmias are a common problem encountered in the intensive care unit
(ICU) and represent a major source of morbidity. Arrhythmias are most likely to occur in
patients with structural heart disease. The inciting factor for an arrhythmia in a given
patient may be an insult such as hypoxia, infection, cardiac ischemia, catecholamine excess
(endogenous or exogenous), or an electrolyte abnormality. Management includes correc-
tion of these imbalances as well as medical therapy directed at the arrhythmia itself. The
physiological impact of arrhythmias depends on ventricular response rate and duration, and
the impact of a given arrhythmia in a given situation depends on the patient’s cardiac
physiology and function. Similarly, urgency and type of treatment are determined by the
physiological impact of the arrhythmia as well as by underlying cardiac status. The purpose
of this review is to provide an update regarding current concepts of diagnosis and acute
management of arrhythmias in the ICU. A systematic approach to diagnosis and evaluation
will be presented, followed by consideration of specific arrhythmias.
KEYWORDS: Arrhythmia, ICU, ventricular tachycardia, AV nodal reentrant tachycardia,

atrial fibrillation, atrial flutter, sinus tachycardia, Wolff-Parkinson-White syndrome,
electrical storm, bradycardia
A rrhythmias are a common dilemma confront- of an atrial kick may cause a dramatic increase in
ing the intensivist. They represent a major source of pulmonary pressures in patients with diastolic dysfunc-
morbidity, and they lengthen hospital stay. Arrhythmias tion. Similarly, tachyarrhythmias can decrease diastolic
are most likely to occur in patients with structural heart filling and reduce cardiac output, resulting in hypoten-
disease. The inciting factor for an arrhythmia in a given sion, in addition to producing myocardial ischemia.
patient may be an insult such as hypoxia, infection, Clearly, the impact of a given arrhythmia in a given
cardiac ischemia, catecholamine excess (endogenous or situation depends on the patient’s cardiac physiology and
exogenous), or an electrolyte abnormality. Management function. Similarly, urgency and type of treatment are
includes correction of these imbalances as well as medical determined by the physiological impact of the arrhyth-
therapy directed at the arrhythmia itself. mia as well as by underlying cardiac status.
The physiological impact of arrhythmias depends This review provides an update regarding current
on ventricular response rate and duration as well as on concepts of diagnosis and acute management of arrhyth-
the underlying cardiac function. Bradyarrhythmias may mias in the intensive care unit (ICU). A systematic
decrease cardiac output due to heart rate alone in approach to diagnosis and evaluation will be presented,
patients with relatively fixed stroke volumes, and loss followed by consideration of specific arrhythmias.
1
Divisions of Cardiovascular Disease and Critical Care Medicine, Non-pulmonary Critical Care: Managing Multisystem Critical Ill-
Cooper University Hospital, Camden, New Jersey. ness; Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Steven M. Semin Respir Crit Care Med 2006;27:221–229. Copyright # 2006
Hollenberg, M.D., Divisions of Cardiovascular Disease and Critical by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Care Medicine, Cooper University Hospital, One Cooper Plaza, NY 10001, USA. Tel: +1(212) 584-4662.
366 Dorrance, Camden, NJ 08103. E-mail: Hollenberg-Steven@ DOI 10.1055/s-2006-945525. ISSN 1069-3424.
cooperhealth.edu.
221
EVALUATION OF TACHYARRHYTHMIAS Adenosine is given as a rapid intravenous (IV) bolus

The first step in the evaluation of the critically ill patient of 6 mg, and a second dose of 12 mg can be given 1 to
with an arrhythmia is to assess hemodynamic stability. If 2 minutes later. The effects are more pronounced when
hemodynamics are compromised due to the arrhythmia, given through a central venous line, in which case
cardioversion should be performed unless pharmacolog- the dosage is then usually halved. The half-life of
ical treatment is immediately successful. However, be- adenosine is only 6 to 10 seconds. Severe broncho-
fore proceeding with cardioversion, one should consider spasm or wheezing can result from its use. Adenosine
whether the arrhythmia is in fact the basis for the can be proarrhythmic, most commonly the induction of
deterioration in hemodynamics. AF (2.7%),2 and there have been reports of asystole,
The next step in evaluation is to determine VT, and ventricular fibrillation (VF) following its
whether the arrhythmia is supraventricular or ventricular administration.3
in origin. First, one examines QRS width. A narrow
QRS complex (< 0.12 sec) indicates a supraventricular
tachycardia (SVT). Narrow complex tachycardias in- NARROW COMPLEX TACHYCARDIA
clude atrial fibrillation (AF), sinus tachycardia, atrioven- Regular narrow complex SVTs include sinus tachycar-
tricular nodal reentrant tachycardia (AVNRT), AV dia, AVNRT, AVRT, ectopic atrial tachycardia, and
reentry from the accessory pathway [Wolff-Parkinson- atrial flutter. Irregular narrow complex SVTs include
White syndrome (WPW)], atrial flutter, and atrial AF, multifocal atrial tachycardia, atrial flutter with
tachycardia. Wide QRS tachycardias include ventricular variable block, and sinus tachycardia with frequent
tachycardia (VT), SVT with preexisting bundle branch premature atrial complexes.
block, aberrant ventricular conduction, or SVT from The p wave morphology can suggest the origin of
AV reentry using an antegrade accessory pathway the atrial impulse. The p wave should be upright in lead
(WPW). II with a normal sinus mechanism. If inverted, this is
One should try not to rely solely on a rhythm strip suggestive of AVRT, AVNRT, or ectopic atrial tachy-
from one monitor lead for diagnosis; there can be cardia. P waves may be absent or difficult to discern in
variability in QRS width depending on which lead is the setting of tachycardia. The RP interval should be
examined. A 12-lead electrocardiogram (ECG) is more assessed on the 12-lead ECG, with a short RP interval
useful. Also, scrutiny of a previous ECG is often useful; (RP shorter than PR, and less than 70 msec) suggesting
for example, to identify preexisting bundle branch block AVNRT, and a long RP interval most likely indicating
or QTc interval prolongation. Marked left axis deviation AVRT via a slowly conducting accessory pathway. A
( 60 to 120 degrees) may indicate a ventricular origin heart rate of 150 beats per minute (bpm) should raise the
of the arrhythmia. It is noteworthy that ST segment suspicion of atrial flutter with 2:1 conduction.
depression during SVT lacks specificity in predicting
ischemia. In one series of 100 patients with SVT,
associated ST segment deviation was only 51% specific Regular Rhythms
(with a positive predictive value of only 6%) for signifi-
cant angiographic coronary artery disease or scinti- SINUS TACHYCARDIA
graphic evidence of ischemia.1 Sinus tachycardia often occurs as a response to a sympa-
Carotid sinus massage and other maneuvers that thetic stimulus (hypoxia, vasopressors, inotropes, pain,
increase vagal tone slow AV conduction time and in- dehydration, hyperthyroidism, etc.). The first step is to
crease refractoriness, and this can aid in the diagnosis review patient medications, including infusions, to ex-
through demonstration of p waves or interruption clude an iatrogenic etiology of the tachyarrhythmia.
of AVNRT or AV reentrant tachycardia (AVRT). Treatment focuses on identifying and trying to correct
Adenosine can also be used for this purpose. Responses the underlying cause. If ischemia is the cause and treat-
to vagal maneuvers or adenosine are listed in Table 1. ment is warranted, b-blockers are the first treatment
Table 1 Responses to Vagal Maneuvers or Adenosine

Arrhythmia Response to Vagal Maneuvers/Adenosine
Sinus tachycardia Gradual slowing with resumption of the tachycardia

Atrioventricular nodal reentrant Abrupt termination or only very transient slowing
tachycardia
Atrial fibrillation/flutter Increased atrioventricular block briefly with slowed ventricular response rate
Multifocal atrial tachycardia Increased atrioventricular block briefly with slowed ventricular response rate
Ventricular tachycardia Usually no response
CARDIAC ARRHYTHMIAS IN THEICU/TARDITI, HOLLENBERG 223
Figure 1 (A) Atrioventricular (AV) node demonstrating dual pathways: slow (a) pathway with short refractory period and fast (b)
pathway with long refractory period. (B) Premature impulse conducts down slow pathway while fast pathway is still refractory to
conduction. (C) As impulse conducts down slow pathway, the fast pathway recovers. (D) Impulse goes up fast pathway as it conducts to
the ventricle. (E) Impulse reenters cycle in AV node completing reentrant circuit.
option. However, it is worth considering that the sinus 250 to 350 bpm. Patients often present with 2:1 AV
tachycardia may be an appropriate hemodynamic response conduction with a ventricular rate of 150 bpm,
to hypotension, hypovolemia, or low cardiac output; if this although the AV conduction ratio can change abruptly.
is the case, overzealous use of b-blockers can reduce Acute treatment consists of AV-nodal-blocking drugs
cardiac output, with potentially disastrous consequences. for rate control. If the patient becomes clinically un-
stable, direct current–synchronized (DC-synchronized)
ATRIOVENTRICULAR NODAL REENTRANT TACHYCARDIA cardioversion with 50 J is usually sufficient, with success
AVNRT typically occurs at a heart rate of 140 to rate of 95 to 100%.5 IV ibutilide has an efficacy rate of
180 bpm. It is more prevalent in females and is not 76% for conversion to sinus rhythm in clinical trials
usually associated with structural heart disease. AVNRT but prolongs the QT interval and can provoke sustained
involves dual AV nodal pathways, usually with slow polymorphic VT in 1 to 2% of cases.6,7 Ibutilide should
conduction antegrade and retrograde conduction via a not be used in patients with a prolonged QTc interval
transiently refractory second pathway (Fig. 1). Therefore, (greater than 420 msec), or in those with underlying
the key to treatment is to block AV conduction. Acute sinus node disease. Other antiarrhythmics such as
treatment includes vagal maneuvers and IV adenosine. sotalol, procainamide, and flecainide have demon-
Long-term preventative therapy includes medications strated less efficacy for acute conversion.8–10 If a tem-
that suppress the initiating premature atrial contractions porary or permanent pacemaker is in place, atrial
(b-blockers) or slow AV conduction (nondihydropyridine overdrive (burst) pacing can sometimes restore sinus
calcium-channel blockers, b-blockers, and digoxin),4 or rhythm via overdrive suppression.
catheter ablation of one of the pathways. Long-term treatment of the ventricular rate in
atrial flutter usually consists of diltiazem, verapamil, b-
ATRIAL FLUTTER blockers, or digoxin. Class IC drugs (flecainide) are very
Atrial flutter is a macroreentrant arrhythmia identified effective in preventing atrial flutter, but by slowing the
by flutter waves, often best seen in the inferior leads, at atrial rate, they have the potential to cause 1:1 AV
Table 2 Intravenous Medications for Heart Rate Control in Atrial Fibrillation

Drug Loading Dose Onset Maintenance Dose
Diltiazem 0.25 mg/kg over 2 min 2–7 min 5–15 mg/h infusion
Esmolol 0.5 mg/kg over 1 min 5 min 0.05–0.2 mg/kg/min
Metoprolol 2.5–5.0 mg over 2 min up to three doses 5 min NA
Propanolol 0.15 mg/kg 5 min NA
Verapamil 0.075–0.15 mg/kg over 2 min 3–5 min NA
Digoxin 0.25 mg each 2 h up to 1.5 mg 2h 0.125–0.25 mg daily
NA, not applicable.
conduction, and should always be combined with AV- Digoxin controls ventricular response through a
nodal-blocking agents. centrally mediated vagal mechanism and by direct action
on the AV node. It controls resting heart rates in patients
who do not have increased catecholamine levels but is
Irregular Rhythms less effective in the ICU. IV digoxin begins to slow the
Irregular narrow complex SVT includes AF, multifocal heart rate in 30 minutes.18
atrial tachycardia, atrial flutter with variable block, and Cardioversion of a patient with AF carries a
sinus tachycardia with frequent premature atrial com- stroke risk from 1.1% if anticoagulated for 3 weeks to
plexes. 7% if not anticoagulated, even if AF duration is less than
1 week.19 Due to delay between resumption of organized
ATRIAL FIBRILLATION atrial electrical activity and of organized mechanical
AF is the most common narrow complex tachyarrhyth- contraction, there can be delay between cardioversion
mia in the ICU (second to VT overall).11 The preva- and embolic events ranging from 6 hours to 7 days.20
lence of AF in the general population increases Postcardiac surgery AF occurs in 25 to 40% of
exponentially with age, from 0.9% at age 40 to 5.9% patients, with peak incidence on day 2.21,22 Use of b-
in those over age 65.12 The most important risk factors blockers, amiodarone, sotalol and biatrial overdrive pac-
for development of AF in the general population are ing to prevent postoperative AF has been studied in
structural heart disease (70% in Framingham study clinical trials.23 Preoperative administration of sotalol
over 22-year follow-up), hypertension (50%),13 valvular and amiodarone is equally effective, but side effects of
heart disease (34%),14 and left ventricular hypertrophy. sotalol limit its use in comparison to amiodarone or b-
AF should be approached in the following manner: blockers. Standard treatment for postoperative AF is to
find the cause, fix the cause, control the rate, consider establish rate control, initially with IV (Table 2) and
rhythm control, and consider anticoagulation. Pharma- then with oral AV nodal blocking medications. There
cological agents for acute rate control include b-block- are numerous risk factors for postoperative AF, with
ers, nondihydropyridine calcium channel blockers, and advanced age being the most important. AF often runs a
digoxin. self-correcting course in this setting, with resumption of
Beta-blockers provide more effective rate control sinus rhythm in more than 90% of patients by 6 to 8
than calcium channel blockers at rest and during exer- weeks after surgery, and so cardioversion is not always
cise.15 Both oral and IV formulations are available. The necessary.24 Immediate cardioversion should be per-
most often used IV medication is metoprolol given at 2.5 formed in patients with recent onset AF accompanied
to 5.0 mg IV over 1 to 2 minutes every 5 to 10 minutes by symptoms or signs of hemodynamic instability result-
for a total of 15 mg as blood pressure tolerates. Esmolol, ing in angina, myocardial ischemia, shock, or pulmonary
0.5 mg/kg bolus, then 0.05 mg/kg/min infusion, is an edema without waiting for prior anticoagulation.
alternative with a more rapid onset and offset, which can Anticoagulation with IV heparin should be con-
be useful in unstable patients. sidered if AF persists for greater than 48 hours. The
Nondihydropyridine calcium channel blockers stroke risk in unanticoagulated patients taken as a whole
(diltiazem and verapamil) are also effective AV nodal is 2% per year (0.05% per day), but individual factors
blockers. Verapamil may have more negative inotropic modulate that risk. The risk factors for stroke are heart
properties than diltiazem and thus may induce hypo- failure, hypertension, age > 75 years, diabetes, prior
tension in patients with left ventricular dysfunction and history of transient ischemic attack (TIA) or stroke,
borderline blood pressure.16 Diltiazem is available in IV and female gender.25
form and is commonly used as a continuous infusion at a
rate of 5 to 15 mg per hour. Up to 93% of patients will MULTIFOCAL ATRIAL TACHYCARDIA
maintain a ventricular response rate < 100 bpm during a MAT is an irregular atrial tachycardia diagnosed by
24-hour infusion.17 identification of three or more p wave morphologies
and PR intervals. MAT is most often associated with 30% of cases of VT.32 Fusion beats, a hybrid of the
hypoxia in the setting of pulmonary disease but may supraventricular and ventricular complexes, occur when
occasionally be due to use of theophylline, metabolic two impulses, one supraventricular and one ventricular,
derangements, and end-stage cardiomyopathy. Treat- simultaneously activate the same territory of ventricular
ment consists of correcting hypoxia by either or both myocardium. The implication is that the wide complexes
treating underlying pulmonary disease and correcting are ventricular. Capture beats are occasional beats con-
electrolyte abnormalities.26 AV nodal blockers are some- ducted with a narrow complex, and such beats rule out
times useful to control the ventricular response in the fixed bundle branch block.
interim. It is better to err on the side of overdiagnosis of
VT. The potential consequences of misdiagnosis were
demonstrated in a study analyzing adverse events in-
WIDE COMPLEX TACHYCARDIA curred by patients with VT misdiagnosed as SVT and
The most frequently reported tachyarrhythmia in the given calcium channel blockers.33 Many of the patients
ICU setting is a wide complex tachycardia. The first promptly decompensated and some required resuscita-
step in treatment is establishing the diagnosis because VT tion. Interestingly, all of these patients were hemody-
is more ominous than SVT with aberrancy. VT is defined namically stable when first seen in VT.
by three or more consecutive ventricular beats. Sustained
VT is defined as more than 30 seconds of ventricular
beats at a rate of more than 100 bpm.27,28 Initial evalua- NONSUSTAINED VENTRICULAR TACHYCARDIA
tion should include obtaining a 12-lead ECG, and This common clinical problem, occurring equally in
measurement of serum potassium, calcium, and magne- women and men, is usually asymptomatic, with an
sium. The ECG should be examined and compared with incidence of 0 to 4% in the general population.34,35 A
prior ECGs with attention to QRS width in sinus major determinant of prognosis is the presence or
rhythm, prior Q waves that may indicate prior myocardial absence of underlying structural heart disease. The
infarction (MI), the presence of delta waves, as well as the Baltimore Longitudinal Study of Aging screened pa-
QT interval. A careful review of medications is para- tients aged 60 to 85 years old for cardiovascular disease
mount in excluding iatrogenic causes of VT. and followed them for 10 years; nonsustained ventricular
VT can be diagnosed using some clinical and tachycardia (NSVT) did not predict coronary events in
electrocardiographic clues, as outlined following here: this population.36 Therefore, in asymptomatic patients
with NSVT, a thorough history and physical examina-
1. Play the odds. VT is approximately four times more tion, echocardiography, and stress testing are usually
common than SVT with aberrancy. In one study of sufficient to exclude prognostically significant structural
200 consecutive patients with a wide QRS tachy- heart disease. Patients with symptoms of palpitations,
cardia, 164 were ventricular, 30 were SVT with syncope, or presyncope should undergo further evalua-
aberrancy, and six were SVT with antegrade con- tion to exclude episodes of sustained VT or other
duction.29 arrhythmias.
2. Ask the right questions. VT is much more common in Patients who have NSVT with structural heart
patients who have a history of MI or heart failure. disease (coronary heart disease, dilated cardiomyopathy,
3. Do not rely on hemodynamics alone. Circulatory col- or valvular heart disease) require more comprehensive
lapse is more common with VT than with SVT, but evaluation and management. As will be discussed here,
patients with VT may maintain a normal blood the prognosis of NSVT following a myocardial MI is
pressure. dependent upon the timing of onset of VT in relation to
4. Do not count on AV dissociation. This is present in less the incident MI. NSVT occurring in the first 48 hours of
than 50% of cases of VT and is difficult to identify at an MI is most likely related to reperfusion or ischemia
faster heart rates. and has no prognostic significance. However, NSVT
5. Do not count on irregularity. Regularity was identified occurring more than 1 week after MI doubles the risk of
in 90% of patients with SVT versus 78% with VT.30 sudden cardiac death (SCD) in patients with preserved
left ventricular function.37 The risk of SCD is increased
Other clues are useful in distinguishing VT from more than fivefold in patients with left ventricular
SVT. A QRS width of more than 0.14 seconds with dysfunction (ejection fraction less than 40%).38 The
right bundle branch block or 0.16 seconds during left risk of SCD is greatest in the first 6 months post-MI
bundle branch block favors VT.31 Comparison of QRS and persists for up to 2 years.
morphology during the tachycardia with the morphol- NSVT is present in up to 80% of patients with an
ogy of ventricular premature beats in sinus rhythm can be idiopathic dilated cardiomyopathy (ejection fraction
helpful. Other diagnostic clues suggestive of VT are [EF] < 40%).39 The current American College of
fusion and capture beats, but these are seen in only 20 to Cardiology/American Heart Association guidelines
recommend implantation of an internal cardiac defib- stopped if the patient becomes hypotensive or the QRS
rillator (ICD) for nonsustained VT in patients with widens by 50% above baseline. The most serious side
coronary disease, prior MI, LV (left ventricular) dys- effects of procainamide are hypotension and proarrhyth-
function, and inducible VF or sustained VT at electro- mia (most commonly torsades de pointes), both of which
physiological study that is not suppressible by a class I increase in frequency in patients with renal insufficiency
antiarrhythmic drug.40 Initial treatment of NSVT in the because of decreased excretion. If the QTc is longer than
setting of dilated cardiomyopathy should include cor- 500 msec the drug should be stopped immediately and
rection of electrolyte abnormalities, removal of exacer- the QTc followed closely. Cimetidine and amiodarone
bating factors (hypoxia, dehydration, medications, can increase levels of procainamide and its metabolite N-
vasopressors, etc.), and up titration of b-blockers. acetyl procainamide.45 Measurement of serum levels may
Mitral and aortic valve disease is associated with be useful, especially in patients with renal insufficiency.
NSVT, occurring in up to 20% of patients with mitral In patients with transvenous or epicardial pace-
valve prolapse (MVP) and 5% of patients with aortic makers, overdrive antitachycardia pacing is an option.
stenosis. In both severe mitral regurgitation and aortic The ventricular pacing rate should be 10 to 20 bpm
stenosis, NSVT does not appear to be associated with faster than the VT. Absent a reversible cause, an im-
increased risk of SCD.41–43 plantable cardioverter-defibrillator (ICD) should be
In patients at high risk as already described, considered in patients with recurrent monomorphic
further evaluation is warranted. This may include cardiac VT and an ejection fraction less than 40% or a history
catheterization, electrophysiological testing, and/or sig- of syncope.
nal-averaged ECG.
POLYMORPHIC VENTRICULAR TACHYCARDIA
MONOMORPHIC VENTRICULAR TACHYCARDIA Polymorphic VT with a normal QT interval is consid-
Monomorphic VT in the setting of a normal QT interval ered to be an ischemic rhythm that typically degenerates
usually occurs from a fixed substrate (i.e., scar) rather into VF. It is almost never asymptomatic and thus DC
than acute ischemia. The importance of monomorphic synchronized cardioversion is the initial recommended
VT depends on the clinical milieu in which it occurs and treatment. Polymorphic VT with a normal QTc is a
on the presence of underlying structural heart disease. more ominous sign than monomorphic VT in patients
Sustained monomorphic VT, either with or without with myocardial ischemia. Medications that might pre-
acute ischemia, portends a worse prognosis even after dispose to ischemia, such as inotropes or vasopressors,
hospital discharge.44 should be stopped or tapered, if possible, and b-blockers
The approach to treatment of sustained mono- started if blood pressure permits. Intraaortic balloon
morphic VT is based on the presence of hemodynamic pumping may be useful as a supportive measure, but
instability and/or other clinical factors (heart failure, revascularization is usually required. If withdrawal of
pulmonary congestion, shortness of breath, decreased vasopressors is contraindicated on a clinical basis, IV
level of consciousness, or myocardial ischemia). If any infusion of lidocaine or amiodarone should be initiated.
are present, then synchronized cardioversion is indi-
cated. Stable or recurrent monomorphic VT can be TORSADES DE POINTES
treated with lidocaine, procainamide, or amiodarone. Torsades de pointes is a French term translated as ‘‘twist-
The next step in evaluation and management of the ing of the points.’’ It is a syndrome composed of
patient is dependent on left ventricular function. If left polymorphic VT and a prolonged QTc interval (by
ventricular function is normal and the patient is not in definition 460 millisecondsec). This may be due to
heart failure, treatment with procainamide, amiodarone, various medications, including procainamide, disopyra-
lidocaine, or sotalol is recommended. The choices are mide, sotalol, phenothiazines, quinidine, some antibi-
limited to amiodarone or lidocaine in those with im- otics (erythromycin, pentamidine, ketoconazole), some
paired left ventricular function (EF < 40%). Amiodar- antihistamines (terfenadine, astemizole), and tricyclic
one can be given as a 150 mg IV bolus over 10 minutes antidepressants. Other etiologies include hypokalemia,
followed by an infusion of 360 mg (1 mg/min) over 6 hypocalcemia, subarachnoid hemorrhage, congenital
hours, and then 540 mg (0.5 mg/min) over the remain- prolongation of the QTc interval, and insecticide
ing 18 hours. The maximum total dose is 2.2 g over 24 poisoning.46 A key to treatment is correction of any
hours. Bradycardia and hypotension can result from IV exacerbating factors and normalization of electrolyte
amiodarone, in which case the rate of the infusion should disturbances, particularly hypomagnesemia, hypocalce-
be decreased. Lidocaine is administered by IV bolus of mia, and hypokalemia. Magnesium should be given 1 to
0.5 to 0.75 mg/kg, followed by continuous infusion at 1 2 g IV push over 30 to 60 minutes. Other potential
to 4 mg/min. Procainamide is administered at 20 mg/ treatments may include overdrive pacing or isoproterenol
min IV for a loading dose of 17 mg/kg, then continued to increase heart rate and thus shorten QTc. Admin-
as an infusion at 1 to 4 mg/min. The infusion should be istration of sodium bicarbonate IV can be useful to
antagonize the proarrhythmic effects of class I antiar- reduce sympathetic activation. Mechanical ventilatory
rhythmics.47 support and IV b-blockers can be used in conjunction,
but IV amiodarone is the pharmacological treatment of
WOLFF-PARKINSON-WHITE SYNDROME choice for this condition. If pharmacological therapy
(VENTRICULAR PREEXCITATION) and antitachycardia pacing are unsuccessful, electro-
AVRT using an accessory bypass tract, WPW, occurs in physiology mapping–guided catheter ablation can be
0.1 to 0.3% of the general population. An accessory considered, although this is often difficult in unstable
pathway bypass tract (bundle of Kent), bypasses the AV patients.55 The prognosis of patients with electrical
node and can activate the ventricles prematurely in sinus storm after ICD implantation is poor, with a 2.4-fold
rhythm, producing the characteristic delta wave. The increase in the risk of subsequent death, independent of
diagnosis of WPW is reserved for patients with both ejection fraction. The risk of SCD is greatest 3 months
preexcitation and tachyarrhythmias. In AVRT conduc- after an electrical storm.
tion can go down the bypass tract and back up the AV
node, producing a wide QRS complex (antidromic) or
down the AV node and back up the bypass tract, BRADYARRHYTHMIAS AND PACING
producing a narrow QRS complex (orthodromic). Asymptomatic bradyarrhythmias do not carry a poor
AVRT should be suspected in any patient whose heart prognosis and in general no therapy is indicated.56
rate exceeds 200 bpm. AF is a potentially life-threat- Recommended initial therapy for bradycardia inducing
ening arrhythmia in patients with WPW syndrome end organ perfusion problems is atropine IV 1.0 mg. The
because it can generate a rapid ventricular response presence of syncope, heart failure, or other symptoms
with subsequent degeneration into VF. This is impor- accompanying bradycardias is an indication for pace-
tant because one third of patients with WPW syndrome maker implantation. Third degree or advanced heart
have AF.48 block with either symptomatic bradycardia, pauses 3
Adenosine should be used with caution in any sec, or heart rate < 40 bpm is also an indication for
young patient suspected of having WPW because it pacemaker insertion. Class I indications (general agree-
may precipitate AF with a rapid ventricular response ment that a treatment is beneficial) for temporary trans-
rate down an antegrade accessory pathway. Procaina- venous pacing after an acute MI are listed here:
mide, ibutilide, and flecainide are preferred agents
because they slow conduction through the bypass tract. 1. Asystole
The long-term treatment of choice for symptomatic 2. Symptomatic bradycardia
patients is radiofrequency catheter ablation of the 3. Bilateral bundle branch block (BBB)
accessory pathway. a. Alternating BBB or right BBB (RBBB) with
alternating left anterior fascicular block (LAFB)/
ELECTRICAL STORM left posterior fascicular block (LPFB)
The definition of an electrical storm is more than three 4. New or indeterminate age bifascicular block with
distinct episodes of VT/VF within a 24-hour period.49 first-degree AV block
In patients with ventricular arrhythmias requiring ICD a. RBBB with LAFB or LPFB
implantation, the incidence of ventricular storm ranges b. Left BBB (LBBB)
from 10 to 30%.50,51 According to one study, the event 5. Mobitz type II second-degree AV block
occurred at an average of 133 135 days after ICD
implantation. Precipitating factors (hypokalemia, myo-
cardial ischemia, and heart failure) were identified in REFERENCES
only 26% of the patients in one study.
Evaluation should include measurement of se- 1. Imrie JR, Yee R, Klein GJ, Sharma AD. Incidence and clinical
significance of ST segment depression in supraventricular
rum electrolytes, obtaining an ECG, and further eval-
tachycardia. Can J Cardiol 1990;6:323–326
uation for ischemic heart disease, which may include 2. Tebbenjohanns J, Pfeiffer D, Schumacher B, Jung W, Manz
coronary angiography. Proarrhythmia secondary to M, Luderitz B. Intravenous adenosine during atrioventricular
antiarrhythmic drugs that prominently slow conduction reentrant tachycardia: induction of atrial fibrillation with
velocity, such as flecainide, propafenone, and morici- rapid conduction over an accessory pathway. Pacing Clin
zine, should be excluded.52,53 Treatment for proar- Electrophysiol 1995;18:743–746
rhythmia is hemodynamic support until the drug is 3. Pelleg A, Pennock RS, Kutalek SP. Proarrhythmic effects of
adenosine: one decade of clinical data. Am J Ther 2002;9:141–
excreted. If exacerbating factors (acute heart failure,
147
electrolyte abnormalities, proarrhythmia, myocardial 4. Winniford MD, Fulton KL, Hillis LD. Long-term therapy of
ischemia, and hypoxia) are corrected, repeated doses paroxysmal supraventricular tachycardia: a randomized, dou-
of IV amiodarone should be given, even if the patient is ble-blind comparison of digoxin, propranolol and verapamil.
already on oral amiodarone.54 Deep sedation can help Am J Cardiol 1984;54:1138–1139
5. Lown B. Electrical reversion of cardiac arrhythmias. Br Heart 21. Ommen SR, Odell JA, Stanton MS. Atrial arrhythmias
J 1967;29:469–489 after cardiothoracic surgery. N Engl J Med 1997;336:1429–
6. Ellenbogen KA, Stambler BS, Wood MA, et al. Efficacy of 1434
intravenous ibutilide for rapid termination of atrial fibrillation 22. Hashimoto K, Ilstrup DM, Schaff HV. Influence of clinical
and atrial flutter: a dose-response study. J Am Coll Cardiol and hemodynamic variables on risk of supraventricular
1996;28:130–136 tachycardia after coronary artery bypass. J Thorac Cardiovasc
7. Stambler BS, Wood MA, Ellenbogen KA, Perry KT, Surg 1991;101:56–65
Wakefield LK, VanderLugt JT. Efficacy and safety of 23. Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S.
repeated intravenous doses of ibutilide for rapid conversion Interventions on prevention of postoperative atrial fibrillation
of atrial flutter or fibrillation. Ibutilide Repeat Dose Study in patients undergoing heart surgery: a meta-analysis. Circu-
Investigators. Circulation 1996;94:1613–1621 lation 2002;106:75–80
8. Sung RJ, Tan HL, Karagounis L, et al. Intravenous sotalol for 24. Stebbins D, Igidbashian L, Goldman SM, et al. Clinical out-
the termination of supraventricular tachycardia and atrial come of patients who develop atrial fibrillation after coronary
fibrillation and flutter: a multicenter, randomized, double- artery bypass surgery [abstract]. PACE 1995;18:798
blind, placebo-controlled study. Sotalol Multicenter Study 25. Gage BF, Waterman AD, Shannon W, Boechler M, Rich
Group. Am Heart J 1995;129:739–748 MW, Radford MJ. Validation of clinical classification schemes
9. Kingma JH, Suttorp MJ. Acute pharmacologic conversion for predicting stroke: results from the National Registry of
of atrial fibrillation and flutter: the role of flecainide, Atrial Fibrillation. JAMA 2001;285:2864–2870
propafenone, and verapamil. Am J Cardiol 1992;70:56A– 26. Wang K, Goldfarb BL, Gobel FL, Richman HG. Multifocal
60A discussion A-1A atrial tachycardia. Arch Intern Med 1977;137:161–164
10. Suttorp MJ, Kingma JH, Jessurun ER, Lie AHL, van 27. Wagner G. Marriott’s Practical Electrocardiography. Phila-
Hemel NM, Lie KI. The value of class IC antiarrhythmic delphia, PA: Lippincott Williams and Wilkins; 2001
drugs for acute conversion of paroxysmal atrial fibrillation or 28. Shenasa M, Borggrefe M, Haverkamp W, Hindricks G,
flutter to sinus rhythm. J Am Coll Cardiol 1990;16:1722– Breithardt G. Ventricular tachycardia. Lancet 1993;341:
1727 1512–1519
11. Trappe HJ, Brandts B, Weismueller P. Arrhythmias in the 29. Wellens HJ, Brugada P. Diagnosis of ventricular tachycardia
intensive care patient. Curr Opin Crit Care 2003;9:345– from the 12-lead electrocardiogram. Cardiol Clin 1987;5:
355 511–525
12. Feinberg WM, Blackshear JL, Laupacis A, Kronmal R, Hart 30. Josephson ME, Wellens HJ. Differential diagnosis of supra-
RG. Prevalence, age distribution, and gender of patients with ventricular tachycardia. Cardiol Clin 1990;8:411–442
atrial fibrillation: analysis and implications. Arch Intern Med 31. Wellens HJ, Bar FW, Lie KI. The value of the electrocardio-
1995;155:469–473 gram in the differential diagnosis of a tachycardia with a
13. Kannel WB, Abbott RD, Savage DD, McNamara PM. widened QRS complex. Am J Med 1978;64:27–33
Epidemiologic features of chronic atrial fibrillation: the 32. Brugada P, Brugada J, Mont L, Smeets J, Andries EW. A new
Framingham study. N Engl J Med 1982;306:1018–1022 approach to the differential diagnosis of a regular tachycardia
14. Davidson E, Weinberger I, Rotenberg Z, Fuchs J, Agmon J. with a wide QRS complex. Circulation 1991;83:1649–1659
Atrial fibrillation: cause and time of onset. Arch Intern Med 33. Tchou P, Young P, Mahmud R, Denker S, Jazayeri M,
1989;149:457–459 Akhtar M. Useful clinical criteria for the diagnosis of
15. Koh KK, Song JH, Kwon KS, et al. Comparative study of ventricular tachycardia. Am J Med 1988;84:53–56
efficacy and safety of low-dose diltiazem or betaxolol in 34. Kostis JB, McCrone K, Moreyra AE, et al. Premature
combination with digoxin to control ventricular rate in chronic ventricular complexes in the absence of identifiable heart
atrial fibrillation: randomized crossover study. Int J Cardiol disease. Circulation 1981;63:1351–1356
1995;52:167–174 35. Raftery EB, Cashman PM. Long-term recording of the
16. Phillips BG, Gandhi AJ, Sanoski CA, Just VL, Bauman JL. electrocardiogram in a normal population. Postgrad Med J
Comparison of intravenous diltiazem and verapamil for the 1976;52(Suppl 7):32–38
acute treatment of atrial fibrillation and atrial flutter. 36. Fleg JL, Kennedy HL. Cardiac arrhythmias in a healthy
Pharmacotherapy 1997;17:1238–1245 elderly population: detection by 24-hour ambulatory electro-
17. Ellenbogen KA, Dias VC, Plumb VJ, Heywood JT, Mirvis cardiography. Chest 1982;81:302–307
DM. A placebo-controlled trial of continuous intravenous 37. Anderson KP, DeCamilla J, Moss AJ. Clinical significance of
diltiazem infusion for 24-hour heart rate control during atrial ventricular tachycardia (3 beats or longer) detected during
fibrillation and atrial flutter: a multicenter study. J Am Coll ambulatory monitoring after myocardial infarction. Circula-
Cardiol 1991;18:891–897 tion 1978;57:890–897
18. Jordaens L, Trouerbach J, Calle P, et al. Conversion of atrial 38. Buxton AE, Marchlinski FE, Waxman HL, Flores BT,
fibrillation to sinus rhythm and rate control by digoxin in Cassidy DM, Josephson ME. Prognostic factors in non-
comparison to placebo. Eur Heart J 1997;18:643–648 sustained ventricular tachycardia. Am J Cardiol 1984;53:1275–
19. Arnold AZ, Mick MJ, Mazurek RP, Loop FD, Trohman 1279
RG. Role of prophylactic anticoagulation for direct current 39. Larsen L, Markham J, Haffajee CI. Sudden death in
cardioversion in patients with atrial fibrillation or atrial flutter. idiopathic dilated cardiomyopathy: role of ventricular arrhyth-
J Am Coll Cardiol 1992;19:851–855 mias. Pacing Clin Electrophysiol 1993;16:1051–1059
20. Bjerkelund CJ, Orning OM. The efficacy of anticoagulant 40. Gregoratos G, Abrams J, Epstein AE, et al. ACC/AHA/
therapy in preventing embolism related to D.C. electrical NASPE 2002 guideline update for implantation of cardiac
conversion of atrial fibrillation. Am J Cardiol 1969;23:208– pacemakers and antiarrhythmia devices: summary article: a
216 report of the American College of Cardiology/American
Heart Association Task Force on Practice Guidelines (ACC/ 49. Exner DV, Pinski SL, Wyse DG, et al. Electrical storm
AHA/NASPE Committee to Update the 1998 Pacemaker presages nonsudden death: the antiarrhythmics versus im-
Guidelines). Circulation 2002;106:2145–2161 plantable defibrillators (AVID) trial. Circulation 2001;103:
41. Kligfield P, Levy D, Devereux RB, Savage DD. Arrhythmias 2066–2071
and sudden death in mitral valve prolapse. Am Heart J 1987; 50. Greene M, Geist M, Paquette M, et al. Long-term follow-up
113:1298–1307 of implantable defibrillator therapy in patients with electrical
42. Kligfield P, Hochreiter C, Niles N, Devereux RB, Borer JS. storm [abstract]. Pacing Clin Electrophysiol 1997;20:1207
Relation of sudden death in pure mitral regurgitation, with 51. O’Donoghue S, Patia EV, Waclawski S, et al. Transient
and without mitral valve prolapse, to repetitive ventricular electrical storm: prognostic significance of very numerous
arrhythmias and right and left ventricular ejection fractions. automatic defibrillator discharges [abstract]. J Am Coll
Am J Cardiol 1987;60:397–399 Cardiol 1997;17:352A
43. Wolfe RR, Driscoll DJ, Gersony WM, et al. Arrhythmias in 52. Passman R, Kadish A. Polymorphic ventricular tachycardia,
patients with valvar aortic stenosis, valvar pulmonary stenosis, long Q-T syndrome, and torsades de pointes. Med Clin
and ventricular septal defect: results of 24-hour ECG North Am 2001;85:321–341
monitoring. Circulation 1993;87:I89–101 53. Tschaidse O, Graboys TB, Lown B, Lampert S, Ravid S. The
44. Newby KH, Thompson T, Stebbins A, Topol EJ, Califf prevalence of proarrhythmic events during moricizine therapy
RM, Natale A. Sustained ventricular arrhythmias in patients and their relationship to ventricular function. Am Heart J
receiving thrombolytic therapy: incidence and outcomes. 1992;124:912–916
The GUSTO Investigators. Circulation 1998;98:2567– 54. Kowey PR. An overview of antiarrhythmic drug management
2573 of electrical storm. Can J Cardiol 1996;12(Suppl B):3B–8B;
45. Trujillo TC, Nolan PE. Antiarrhythmic agents: drug discussion 27B–28B
interactions of clinical significance. Drug Saf 2000;23:509– 55. Brugada J, Berruezo A, Cuesta A, et al. Nonsurgical
532 transthoracic epicardial radiofrequency ablation: an alternative
46. Kossmann CE. Torsade de pointes: an addition to the in incessant ventricular tachycardia. J Am Coll Cardiol 2003;
nosography of ventricular tachycardia. Am J Cardiol 1978;42: 41:2036–2043
1054–1056 56. Gregoratos G, Cheitlin MD, Conill A, et al. ACC/AHA
47. Banai S, Tzivoni D. Drug therapy for torsade de pointes. guidelines for implantation of cardiac pacemakers and
J Cardiovasc Electrophysiol 1993;4:206–210 antiarrhythmia devices: a report of the American College of
48. Campbell RW, Smith RA, Gallagher JJ, Pritchett EL, Cardiology/American Heart Association Task Force on
Wallace AG. Atrial fibrillation in the preexcitation syndrome. Practice Guidelines (Committee on Pacemaker Implanta-
Am J Cardiol 1977;40:514–520 tion). J Am Coll Cardiol 1998;31:1175–1209
Evaluation and Management of Shock
Olivier Axler, M.D., Ph.D., F.C.C.P.1
ABSTRACT
Shock is one of the most frequent situations encountered in the intensive care unit
(ICU). Important new concepts have emerged for shock management in recent years. The
concept of early goal-directed therapy has evolved from the basic management concepts for
septic shock delivered in a structured fashion. Numerous cardiovascular techniques,
methods, and strategies have been developed as novel alternatives to the use of the
pulmonary artery catheter. Among these techniques, echocardiography, esophageal Dop-
pler, and arterial pulse contour analysis show great promise. Prediction of responsiveness to
fluid administration is a key component of the management of shock, as is assessing
cardiovascular performance. The intensivist has several options to evaluate and treat shock.
Further research should yield additional important advances.
KEYWORDS: Shock, cardiovascular protocols, cardiovascular techniques; central

venous pressure; pulmonary arterial catheter
S hock is a common cause of admission in any OVERVIEW, DEFINITIONS, AND

intensive care unit (ICU) and also occurs frequently DIAGNOSIS OF SHOCK
during the course of critical illness. Shock is associated Shock is traditionally defined by multisystem organ
with significant morbidity and mortality and represents a hypoperfusion, whatever its specific cause, leading to
medical emergency. Early, targeted therapy is crucial; the common physical signs. It can also be defined as an
first hour of care may be key to a successful outcome.1,2 inability to assure adequate cellular and tissue oxygen
Therefore, it is important that physicians are aware of supply and removal of waste products of cellular metab-
updated concepts and management guidelines for treat- olism, thus overwhelming the compensatory mecha-
ing patients with shock. Although the principles of nisms of the organism.
shock management are well established, there is consid- The presentation of shock may be obvious but can
erable heterogeneity of bedside management. also be latent and incomplete, leading to a delayed
This heterogeneity is apparent not only with diagnosis, potentially worsening the prognosis and de-
accurate clinical identification of a shock state3 but also creasing chances of reversal. The clinician must be
in regard to evaluation and therapy. Critical care soci- familiar with different clinical patterns of shock and
eties and other experts have published evidence-based the pathophysiological aspects of shock, including car-
guidelines for diagnostic criteria and therapeutic strat- diovascular (ventricular pressure–volume curves, cardiac
egies4–8; however, these recommendations generally fo- function curves), biochemical (oxygenation cascades),
cus on severe sepsis and septic shock. This article reviews and immunological (mediators and cytokine cascades)
the traditional criteria and current guidelines for man- features.
agement of shock, the traditional and newer diagnostic The clinical signs and symptoms of shock have
and monitoring techniques, and therapeutic strategies. been known for years9 and have been presented in
1
Cardiology Department, Centre Hospitalier Territorial Gaston Non-pulmonary Critical Care: Managing Multisystem Critical Illness;
Bourret, Noumea, New Caledonia, France. Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Olivier Axler, Semin Respir Crit Care Med 2006;27:230–240. Copyright # 2006
M.D., Ph.D., F.C.C.P., Cardiology Department, CHT Gaston by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Bourret, 98800 Noumea, New Caledonia, France. E-mail: olivier. NY 10001, USA. Tel: +1(212) 584-4662.
[email protected]. DOI 10.1055/s-2006-945526. ISSN 1069-3424.
230
EVALUATION AND MANAGEMENT OF SHOCK/AXLER 231
comprehensive reviews.1,10 Several points are worthy of TRADITIONAL AND NEWER METHODS
emphasis. First, the diagnosis must be made quickly, AND TECHNIQUES TO ASSESS
followed by classification of type of shock. Second, the MECHANISMS OF SHOCK
sensitivity and specificity of each sign are highly variable. Several simple tools used in the initial management of
Third, the quantitative aspects of these signs are useful shock (e.g., electrocardiogram; chest radiographs; routine
but vary depending upon the clinical circumstances. The hospital chemistries; blood gases) are well known and will
first step is to begin the correct resuscitation measures, not be further discussed here. In this section, invasive and
not only to achieve therapeutic goals but also to confirm noninvasive techniques to assess hemodynamics are dis-
the diagnosis. Depending on the response to the ther- cussed. Measurement of central venous pressure (CVP)
apeutic intervention, the diagnosis can be confirmed or or pulmonary artery pressure (PAP) may be useful to
corrected, allowing adjustment of treatment. classify the mechanism of shock. Further, measurement
In its complete clinical presentation, shock clas- of SCVO2 [via a catheter in the superior vena cava
sically includes: tachypnea; tachycardia; low systolic, (SVC)] or mixed SVO2 (via a pulmonary arterial cathe-
diastolic, and mean blood pressures (BPs); diaphoresis; ter) may be useful to diagnose and monitor the impact of
poorly perfused skin and extremities; cyanosis; mottling therapeutic interventions in patients with shock.17
of cool and moist extremities; altered mental status Arterial pulse contour techniques are used to
(ranging from decreased state of consciousness to agi- measure cardiac output (CO), and requires arterial
tation); and decreased urine output. Joly and Weil access.18,19 This technique is incorporated in more
emphasized the role of the ‘‘cold great toe,’’11 whereas sophisticated devices, measuring other parameters as
we have noted that cold knees have an excellent diag- preload with an estimation of fluid responsiveness for
nostic specificity and sensitivity for shock (unpublished the LiDCO plus System (LiDCO Ltd., Cambridge,
data). UK) via a pulse power analysis.18–20 Another system,
However, these signs are not always present con- PiCCO (Pulsion Medical Systems AG, Munich,
comitantly, and some of these features may be absent or Germany) also measures intrathoracic volumes and
borderline. For instance, in classical shock, BP is de- extravascular lung water from transpulmonary indica-
creased, and a commonly accepted threshold for a tor dilution.21,22 This latter technique mandates fem-
resuscitation goal in septic shock is 65 mm Hg for oral arterial and central venous access, whereas the
mean arterial pressure (MAP).6 However, not all hypo- former requires only a radial artery and a peripheral
tensive states are associated with shock, and not all vein. Analysis of the systemic systolic pressure, pulse
shocks present with hypotension. In fact, some shock pressure and stroke volume (SV), and their respiratory
states present with high BP at the onset because of the variations, provides an excellent assessment of preload
adaptive adrenergic response. Early septic shock is clas- and estimation of fluid responsiveness.19
sically ‘‘hyperdynamic,’’ with increased pulse pressure Ultrasound techniques that are useful to assess
and warm extremities. In addition, low BP is relative fluid status and cardiac function include esophageal
to the baseline BP for a given patient. Invasively meas- Doppler23 and echocardiography (transthoracic and
ured BP using an intra-arterial catheter is more accurate transesophageal).24 Methods such as monitoring splanch-
than cuff measurements.12 Chronotropic medications, or nic blood flow or monitoring the microcirculation with
sinus or atrioventricular dysfunction, can blunt the videomicroscopy have been utilized in research investiga-
tachycardic response. Oliguria is often defined as urine tions but have no or limited clinical utility.
output less than 0.5 mL/kg/hr.6 From a laboratory
standpoint, blood lactate is a robust clue of shock arising
from cellular and tissue hypoxia,3 and precedes acidemia. Traditional Methods
This was recently confirmed,2 with a threshold of
4 mmol/L consistent with shock. However, the specific- CENTRAL VENOUS PRESSURE MONITORING
ity of blood lactate is imperfect because any condition Although CVP was recently shown to be somewhat
exceeding the aerobic threshold leads to increased lactate inaccurate to assess preload,25,26 CVP, when integrated
levels, and some metabolic conditions increase lactate into an algorithm, was a useful parameter to guide volume
levels without shock (i.e., any sympathetic activation).13 administration in a cohort of septic patients.2 However,
Tissue hypercapnia is known to correlate well CVP measurements should be interpreted with caution,
with decreased blood flow.14 The first organ to be even at low or high values, when the value is used in
studied was the gastric mucosa, but this technique has isolation.10,25–28 Some authors emphasize the need to
largely been abandoned. Sublingual PCO2 has emerged incorporate the cardiac and venous return curves for a
as a good predictor of shock (irrespective of cause), and correct interpretation of CVP.29,30 Indeed, the basic con-
in some studies was superior to blood lactate levels and cept espoused is to use CVP with simultaneous measure-
mixed venous oxygen saturation (SVO2) or central ment of cardiac output. In a study of 33 ICU patients, the
venous oxygen saturation (SCVO2).15,16 right atrial pressure (RAP) decreased at least 1 mm Hg
with inspiration [with a decreased pulmonary artery oc- ICU echocardiography24,44 and pulse contour analysis
clusion pressure (PAOP) of at least 2 mm Hg with techniques (PiCCO and LiDCO)18,19,21,22 have sup-
inspiration] and increases in CO of 250 mL/min.30 planted PACs in most European ICUs. Despite the
This analysis can differentiate ‘‘relative hypovolemia’’ (or limitations of PACs, recent guidelines (e.g., the Survival
more accurately a fluid responsive state) from euvolemia. Sepsis Campaign) continue to recommend PACs for
This concept was discussed in a recent excellent review.31 the assessment of severe sepsis and septic shock.6 More-
Despite the low reliability of CVP to assess preload, this over, a recent paper focusing on ‘‘practice parameters
parameter continued to be widely measured by 93% of a for hemodynamic support of sepsis in adult patients’’
cohort of European intensivists in 1998.32 favors the use of PACs, and states that ‘‘echocardiography
may also be useful to assess ventricular volumes and
CENTRAL VENOUS OXYGEN SATURATION (SCVO2) cardiac performance.’’7
The O2 saturation in a central vein (most often the SVC) Several measured and derived values are available
was recently shown to be a key component of early goal- from a PAC to determine the mechanism of shock:
directed therapy in the emergency department (ED)2; PAP, PAOP, right ventricular pressure (RVP) and
values < 70% are consistent with incomplete resuscita- RAP, CO by thermodilution, and its modified deriva-
tion. This parameter is less useful after several days of tives: (1) semicontinuous cardiac output (using a thermal
severe critical illness and severe tissue oxygenation defi- coil in the right ventricular portion of the PAC;
cit.17,33,34 A low SVO2 generally reflects low CO because (2) calculation of right ventricular end-diastolic volume
oxygen extraction by the tissues is greater in cardiac (RVEDV) from measurement of right ventricular ejec-
failure.34,35 Low SCVO2 is also associated with a poor tion fraction (RVEF); SVO2 and related oygenation
prognosis34 and often appears earlier than any other variables; oxygen consumption (VO2); oxygen delivery
clinical sign of shock.2 In a cohort of patients with septic (DO2); and O2 extraction ratio (O2ER).
shock, Rivers and colleagues demonstrated mortality Measurement of PAP is a very important param-
reduction of 15% by maintaining a therapeutic algorithm eter to diagnose pulmonary arterial hypertension (PAH)
that maintained the following parameters: SCVO2 as may be seen in ARDS, pulmonary embolism, right
> 70%; CVP > 8 to 12 mm Hg; MAP > 65 mm Hg; ventricular infarction, obstructive lung disease, left heart
urine output > 0.5 mL/kg/h.2 SCVO2 and SVO2 are diseases, and primary PAH. Its measurement is generally
usually similar but can diverge in some cases, particularly easy and its interpretation is the least problematic of all
in severe sepsis, due to greater O2 extraction in the PAC-derived data.
hepatosplanchnic circulation. However, SCVO2 is pre- Using the PAOP is one of the most controversial
ferred to SVO2 because it can be measured more simply issues related to PAC. Classically, hypovolemic shock has
from a central venous catheter rather than a pulmonary low right and left heart filling pressures, whereas left
artery catheter (PAC).36–38 SCVO2 can be continuously ventricular cardiogenic shock is associated with elevated
monitored using a special catheter or intermittently by PAOP and RAP. Historically, PAOP has been consid-
direct repeated samples. SCVO2 correlates with outcome ered to provide information regarding preload and the
in all kinds of shock, even during cardiopulmonary presence or absence of pulmonary edema. However,
resuscitation.17,33,34 The research by Rivers and col- measurement and interpretation of PAOP may be diffi-
leagues addressed very early shock,2 and studies have cult.37,38,45 The utility of PAOP to assess volume status
not demonstrated benefit for patients later in the course has recently been challenged 25,26,46; this is also true for
of shock when oxygen extraction may be impaired and RAP.25,26,46 This can be explained by a frequent absence
SCVO2 exceeds 80%.17,33,34 of linearity between left ventricular end-diastolic volume
(LVEDV) and left ventricular end-diastolic pressure
PULMONARY ARTERY CATHETERS (LVEDP); second, disparity between LVEDP and
The PAC has been used to differentiate various mecha- PAOP may exist. The LVEDV/LVEDP relationship
nisms of shock since the early 1970s, but utilization of data can be profoundly modified by LV compliance factors
from PACs has many pitfalls. First, hemodynamic values such as left ventricular hypertrophy (LVH), myocardial
are frequently misinterpreted, leading to incorrect treat- ischemia, positive end-expiratory pressure (PEEP), and
ment.37,38 Second, recent studies found that the use of active exhalation. Further, PAOP can overestimate
PAC did not confer any benefit compared with no PAC LVEDP if mitral stenosis or mitral regurgitation are
use36,39,39a,39b; further, some studies suggested deleterious present, and conversely underestimates LVEDP when
effects of PACs, particularly in patients presenting diastolic dysfunction or hypervolemia exist. These con-
with acute respiratory distress syndrome (ARDS) or ditions are frequent in patients presenting with shock but
shock.35,36,39 However, PAC-directed therapy was shown are often not appreciated. Thus PAOP should be inter-
to be cost effective in the preoperative period.40–43 In preted cautiously.45 Notwithstanding these pitfalls, 58%
North America, there continues to be relatively wide- of European intensivists continued to measure PAOP as
spread use of PACs,6 whereas newer techniques such as part of monitoring critically ill ICU patients in 1998.32
Measurement of CO is one of the most important thoracic echocardiography (TTE) is noninvasive and
issues in the management of shock.1,47–49 Thermodilu- relatively easy to perform after an adequate training.
tion is the gold standard method for measuring CO for Transesophageal echocardiography (TEE) is modestly
clinical use because measurement is relatively straight- invasive and requires some degree of sedation for patient
forward and does not present the same technical diffi- comfort but is safe and highly accurate. Echocardiog-
culties as PAOP. Additionally, the thermodilution raphy provides an excellent assessment of cardiac func-
technique was validated against electromagnetic flow, tion and estimates left and right heart filling pressures
Fick, and dye dilutions techniques. However, thermo- and can be useful to determine the cause of shock.56
dilution has important limitations. Specifically, thermo- Echocardiography provides acceptable estimates of most
dilution can overestimate CO in low output states, parameters gleaned from pulmonary artery catheters
whereas significant tricuspid regurgitation leads to (PACs) (i.e., CO; right arterial pressure (RAP) from
underestimation of CO. Other confounding issues in- inferior vena cava (IVC) size and ventilatory variations,
clude intracardiac shunts and temperature issues.47 In systolic pulmonary artery pressure (SPAP); left and right
this context, echocardiography is helpful. Recently, the ventricular filling pressures; ejection fraction; ventricular
left ventricular outflow tract (LVOT) pulsed Doppler interdependence; right heart function; diastolic dysfunc-
method has been employed as an alternative method to tion; left ventricular hypertrophy (LVH); ischemic heart
measure CO.49,50 Some authors believe that this should disease, valvular diseases, and so on). Recent publications
become the new gold standard49,50 unless significant emphasized the value of echocardiography to predict
aortic valve disease exists. fluid responsiveness using heart–lung interactions ba-
Cardiac output can be monitored with a modified sics.57–65 Many studies have shown that echocardiogra-
PAC. This PAC incorporates a thermal coil in the right phy (either TTE or TEE) may be invaluable to monitor
ventricular portion of the catheter, and continuous CO therapeutic interventions or hemodynamic changes in
measurement is based on the delivery of electrically critically ill patients.62–80 TEE is more useful than TTE
generated heat to the blood near the right atrium and for this purpose.66–79 In most of the studies, TEE
ventricle and the resulting temperature change in the provided clinically useful information in 60 to 90% of
PA. This technique avoids performing an intermittent cases. More importantly, TEE had a direct favorable
injection and yields a continuous CO display. The impact on the acute care management.79
accuracy is good compared with thermodilution, pro- Our recent experience with TTE has been favor-
vided regular calibration is performed.51 able (unpublished data). The value of TTE was recently
Recent refinements of the PAC allow calculation underscored in a study by Joseph et al, who noted
of RVEF. This catheter has a rapid-response thermo clinically useful and reliable information in 70 to 80%
‘‘slur’’ and intracardiac electrocardiogram electrode, al- of critically ill ICU patients who had TTE.80 Recent
lowing the calculation of RVEF. Combined with SV, improvement in imaging, software, and electronic sys-
the RVEF allows the calculation of right ventricular end tems have improved the quality and utility of images
diastolic volume (RVEDV). This parameter has been gleaned from TTE. Transthoracic echocardiography is
extensively studied in circulatory shock,52,53 but in the useful as a diagnostic tool for critically ill patients in
most important studies comparing RVEDV before and shock (or impending shock) but in some cases, TEE is
after a fluid challenge, significant difference was found in necessary for a more accurate assessment.79 Specific
only one of 15 studies.53 Intraindividual changes in indications for TEE include: aortic dissection (when
RDEDV with various treatments are more useful than computed tomographic angiography is inconclusive, or
absolute values.25,52,54 Continuous fiberoptic measure- to complete it if necessary); endocarditis (especially when
ment of SVO2, coupled with traditional PAC parame- a valvular prosthesis is present); complicated cardiac
ters, is available with some catheters. surgery; marked obesity; poor echogenicity with TTE;
Finally, the relationship of oxygen delivery/con- intracavitary thrombi; and cardiac sources of emboli.
sumption ratio (DO2/VO2) has been used for both TEE is the preferred method to assess the SVC size
research and clinical indications among critically ill and its ventilatory variations, a new powerful parameter
patients for more than 3 decades.55 However, awareness to predict fluid volume responsiveness.64
and incorporation of these variables into clinical proto- In a patient with shock, echocardiography (usu-
cols have not been shown to influence outcome.55 ally TTE) can provide prompt (within 15 minutes)
assessment of critical variables, including size of cardiac
chambers; left and right systolic function; cardiac output
Newer Methods (49); wall motion abnormalities; valvular pathology; LV
filling pressures from a combination of parameters ob-
ECHOCARDIOGRAPHY-DOPPLER tained from mitral flow, Doppler tissue imaging (DTI),
Cardiac ultrasound (echocardiography) has increasingly pulmonary venous flow (PVF), early diastolic mitral flow
been utilized in ICUs within the past few years. Trans- propagation velocity (Vp), PA pressures; RV filling
pressures; and pericardial imaging. Echocardiography Additionally, in a cohort of septic patients on MV,
can estimate fluid volume responsiveness using heart– measurement of the SVC by TEE, and ventilatory
lung interactions. This concept assumes that the meas- collapsibility of the SVC predicted the cardiac response
urement of some parameters before fluid loading can to fluid challenge.64 The 36% threshold of variability
predict a significant increase of cardiac ouput in hemo- (Dmax-Dmin:Dmax) could define responders and non-
dynamically unstable patients.54 This estimation uses responders in CO, with 90% sensitivity and 87% specif-
inferior vena cava (IVC)62,63 or SVC,64 size and ven- icity.64 Although measurement of IVC diameter
tilatory variations, as well as the respiratory variations of ventilatory variations was studied only in septic patients
the LVOT.57–59 These parameters correlated strongly in on MV, we use it in every patient in acute circulatory or
these 5 studies with the concept of ‘‘fluid responsive- acute respiratory failure, even among patients not requir-
ness.’’ The measurement of left heart filling pressures ing MV. However, outcomes data in these other patient
requires more sophisticated echo devices. These two new populations are lacking.
steps represent one of the major advances in the manage- The second important new parameter in assessing
ment of shock. shock is the ventilatory variation of left ventricular out-
Some echocardiographists prefer to use TTE first, flow tract (LVOT) (called also aortic) Doppler veloc-
and then TEE if necessary, and some always use directly ities. Two studies found that this parameter was a strong
TEE. This is an ongoing discussion, and this choice predictor of preload responsiveness: one in septic pa-
depends upon the ability of each echocardiographist. tients on MV (using TEE),57 and one with a rabbit
Published data regarding the assessment of pre- model.58 The first study defined the respiratory variation
load using echocardiography are disappointing. Preload as the ratio of the difference between maximal velocities
was initially assessed by measuring left ventricular end- to the mean of these two velocities. A ventilatory
diastolic diameters, areas, or volumes; close correlations variation of LVOT blood flow velocity > 12% was
were found with blood loss or expansion in normal associated with a 15% increase of CI with a 91% positive
subjects81 or perioperative conditions.82 However, in predictive value. A ventilatory variation < 12% had a
ICU settings, recent studies found that the left ventricle 100% negative predictive value. This could imply that no
end-diastolic area (LVEDA) failed to accurately predict volume expansion was necessary with a high degree of
fluid requirement or fluid overload status, particularly confidence. There was a high degree of correlation
when compared with newer methods to assess fluid between baseline ventilatory variation and degree of CI
responsiveness.57,83–85 Therefore, echocardiographic increase after volume expansion.57 This important study
measurements of LV and RV size are no longer used can be extended to TTE. We regularly use this method
to assess volume status. However, diameter of the IVC in all patients in shock to assess potential fluid respon-
and its ventilatory variations are invaluable to predict siveness. Patients must be on MV and well adapted to
fluid responsiveness. This measurement is simple to their ventilator, and must be free of arrhythmias. Slama
assess, with a short learning curve (1 hour). The IVC et al found similar results in a rabbit model, using TTE,
diameter and its ventilatory variations are measured with with progressive blood withdrawal.58 In these two stud-
TTE, on a subcostal view, in M mode. Measurement ies, this parameter was more powerful than all other
parameters include IVC diameter (D) at end-expiration parameters (CVP, PAOP, left ventricular end-diastolic
(Dmin) and at end-inspiration (Dmax); distensibility area) that had been used for the past several years. The
index of the IVC (dIVC) calculated as the ratio of most recent studies showed that ‘‘static’’ echocardio-
Dmax-Dmin:Dmin and expressed as a percentage, or graphic parameters failed to consistently predict re-
((Dmax-Dmin:Dmax þ Dmin):2 ).62,63,86 Measurement sponse to fluid loading.57–59,83–85
of IVC size and respiratory variation is useful to predict The second important advance is the ability of
response to a fluid challenge.62,63,65 The useful threshold echocardiography to estimate LV and RV filling pres-
was 12% of variability, with a positive and negative sures (LVFP and RVFP). These measurements were
predictive value of 93% and 92% in one study of septic extensively studied over the past 10 years in the cardio-
patients requiring mechanical ventilation (MV).62 An- logical arena86–90 but were only recently applied to the
other group studied 23 patients in septic shock requiring critically ill (noncardiac) patients in ICUs.91–93 These
MV.63 The size and ventilatory variation of the IVC measurements do not accurately measure preload, but
(IVCVV) predicted a positive response to a fluid chal- may predict fluid responsiveness. An algorithm is now
lenge [ 15% increase of the cardiac index (CI) follow- available to determine if LVFP are predictive of PAOP
ing a 7 mL/kg fluid challenge]. IVCVV was defined in as ‘‘high’’ (> 15 mm Hg) or ‘‘not high’’ ( 15 mm Hg).
this study by Dmax-Dmin:Dmin. There was an excellent This analysis is usually applied when the LVEF is
correlation (r ¼ 0.9) between an 18% IVCVV at baseline decreased (< 45%). This algorithm is determined by
and 15% increase in CI after a fluid loading, with 90% the analysis of the combination of pulsed Doppler of
specificity and 90% sensitivity.63 Importantly, baseline mitral flow, tissue Doppler imaging, color M-mode of
CVP did not accurately predict fluid responsiveness. mitral flow, pulmonary venous flow (PVF), left atrial
size, interatrial septum shape and movement, diastolic Ea, Em/E/a, Vp, Em/Vp, LA size. When LVEF is
PA pressure from pulmonary regurgitation. Mitral flow normal (except in cases of LVH), LVFP is rarely high.
pulsed Doppler shows an early wave (E wave) and a later This algorithm has been used for several years in cardiac
wave (A wave) if the rhythm is sinus. The peak velocities patients89,90 but has only recently been tested among
are measured (Em and Am). A small E wave, with critically ill noncardiac ICU patients.91–93
an E:A ratio < 1, and an E deceleration time (DTE) Right ventricular filling pressures can also be
> 150 msec are usually associated with diastolic assessed from systolic and mean PA pressures derived
LVFP < 15 mm Hg; when some discrepancies appear, from the tricuspid regurgitation jet and from the early
other parameters help to determine LVFP. These pa- diastolic pulmonary regurgitant jet, respectively. IVC
rameters are (1) the maximal early diastolic velocity (Ea) size and its collapsibility index, as were discussed earlier,
of the mitral annulus at the lateral- or septal most basal are markers of RAP. Hepatic vein pulsed Doppler also
point, measured from DTI; (2) early flow (LV flow provides information on right heart filling pressures.
propagation velocity) Vp using M-mode color Doppler; Despite these intriguing data, no study has exam-
(3) PVF, where the systolic fraction is measured: this ined the impact of echocardiography in critically ill
fraction is the ratio between the velocity time integral noncardiac ICU patients on mortality. Because echocar-
(VTI) of the systolic component of PVF, to the sum of diography is commonly employed in many ICUs, a
the VTI of the diastolic and systolic waves (S:S þ D); randomized trial to assess the clinical impact of this
this ratio is nevertheless less and less used and replaced technique may be difficult to accomplish.61
by the difference between duration of pulmonary A wave
and mitral A wave (Apd–Amd) duration; if this ratio is ESOPHAGEAL DOPPLER MONITORING
greater than 20 ms, this is in keeping with a high LVFP; Esophageal Doppler monitoring is an exciting method
(4) left atrial (LA) diameter and area, an indirect to manage shock because it couples ultrasonography with
indicator of LVFP; (5) diastolic PAP (PAPd) obtained continuous monitoring capabilities.23 This is a simple
from the end-diastolic velocity of the pulmonary regur- and quick technique, with a short period of training.
gitant flow because this parameter approximates PAOP; Analysis of the size and shapes of the waveform allows
(6) interatrial septal motion and curvature; (7) mitral therapeutic modifications regarding fluid therapy and
regurgitant flow from continuous Doppler allows to the use of inotropic, vasodilator, and vasopressor agents.
assess left ventricle diastolic pressure (LVDP) from the Outcome and hospital stay have improved with this
difference between systolic blood pressure SBP and technique in high-risk surgical patients.23
pressure gradient from mitral regurgitation maximal
velocity; (8) aortic regurgitation flow from continuous VENOUS AND TISSUE PCO2
Doppler allows to assess LVFP (difference between Venous and tissue PCO2 have been used for more than 3
diastolic blood pressure (DBP) and diastolic left ven- decades to manage shock, but sublingual PCO2 has
tricle-aorta ( LV–AO gradient). Systemic blood pres- emerged as the simplest and most reliable technique to
sures (BP) must be measured simultaneously. The assess venous and tissue PCO2. In patients in the ED
quantitative combination of mitral flow (Em, Am, E and in the ICU, sublingual PCO2 indicates the presence
wave deceleration time, DTEm), with DTI (Ea), early of shock but does not allow any classification of
flow Vp using M-mode color Doppler data, PVF, LA shock.3,15,16
size, and the other parameters just defined has led to an
algorithm to assess if LVFP are ‘‘high’’ (left atrial pressure ARTERIAL PRESSURE VARIATION
(LAP or PAOP > 15 mm Hg) or ‘‘not high’’ (normal or Arterial pressure variation yields important information
low). These values can be estimated for ‘‘not high’’ LVFP: for volume status in hemodynamically unstable patients
Em:Ea < 1, DTE > 150 to 200 ms, Em:Vp < 1.5, Em/ mechanically ventilated and perfectly sedated and may
Ea < 8, S/S þ D > 55%, LA size normal, Apd–Amd predict fluid responsiveness in these patients. Measure-
< 20 ms (Amd ¼ duration of atrial mitral wave, Apd ¼ ment of SV and pulse pressure respiratory variations
duration of atrial pulmonary wave), LVFP < 15 mmHg provides ancillary information.94–96 Values above 10 to
from aortic and mitral regurgitation, atrial septum shifted 13% indicate with a high sensitivity and specificity that
to the left. a volume expansion will increase CO.94–96 Perel and
LVFP can be estimated ‘‘high’’ (PAOP > 15 mm colleagues developed a respiratory systolic variation test
Hg), if Em/Am > 2, DTE < 150 ms, Em/Vp > 2.5, (RSVT) that uses three consecutive incremental pres-
Em/Ea > 15, S/S þ D < 55%, LA size increased, Apd– sure-controlled (10, 20, and 30 cm H2O) breaths to test
Amd > 20 ms, LVFP > 15 mmHg form aortic and the response of venous return, left ventricular stroke
mitral regurgitation, atrial septum shifted to the right. volume (LVSV), and systolic arterial pressure. A slope is
However, because areas of uncertainty exist, multiple then determined, and a prediction of an increased CO
parameters must be examined. The order of importance can be tested with volume expansion.94,95 In interpreting
of measurements is as follows: LVEF, Em, Am, DTE, these tests, it is important to note that a positive fluid
responsiveness test does not imply that there is an THORACIC ELECTRICAL BIOIMPEDANCE
indication for giving fluids, because a normal heart will Thoracic electrical bioimpedance has been used by some
increase CO with any volume expansion. By contrast, teams to assess CO continuously,3 but it has limited
even in a real hypovolemic status, CO may not increase if application.
one or both ventricles is not on its ascending limb of the
Starling curve. Thus RSVT is not 100% sensitive and
specific. Performing and interpreting these tests require MANAGEMENT OF SHOCK
a knowledge of heart–lung interactions.54,61 The classification of mechanisms of shock has been
elegantly analyzed elsewhere.1,3,10 The initial manage-
ARTERIAL PULSE CONTOUR ANALYSIS METHODS ment of shock requires an immediate clinical assessment,
Arterial pulse contour analysis permits continuous mon- basic vital signs, and clinical examination. Prompt as-
itoring of SV and CO.18,19,21,22 These methods calculate sessment and decision are imperative regarding fluid
the area under the systolic portion of the arterial pressure resuscitation, MV, and use of inotropic, vasoconstrictor,
waveform, which, divided by aortic impedance, allows or vasodilator agents.1 Echocardiography can be per-
estimation of LVSV. Generally, the CO obtained by formed at the bedside and may be invaluable in planning
these techniques is calibrated with an absolute CO value, resuscitative efforts. A standardized ‘‘ABC’’ resuscitation
as an indicator dilution technique. Preload and fluid protocol [management that incorporates components of
responsiveness can be estimated, in continuously meas- airway (A), breathing (B), and circulation (C)] such as
uring respiratory variations of systolic arterial pressure, advanced cardiac life support (ACLS) guidelines, can
pulse pressure, pulse pressure variation (PPV), and improve outcomes in the first few minutes of severe
stroke volume (SV). Two systems are commercially cardiorespiratory failure. Intubation and MV are corner-
available using arterial pulse contour analysis: the stones of therapy in this setting.1
PiCCO system and the PulseCo or LiDCO system. The circulation step has evolved recently because
Although the basic principle to measure CO is the such protocols have been shown to improve survival.1,2
same, the LiDCO system uses a very sophisticated algo- Current guidelines for hemodynamic support of sepsis
rithm that appears very accurate. This algorithm is set up advise following predetermined protocols in patients in
on arterial pulse power analysis rather than morphology.94 shock. The protocol espoused by Rivers and colleagues
This technology uses lithium (Li) as an indicator and has advocates aggressive volume resuscitation in septic shock
been extensively tested in the United States and Europe and aims to achieve threshold values of multiple param-
but is not available in some countries. We have used this eters (e.g., hemoglobin, MAP, CVP, SCVO2, urine
system extensively in our critically ill patients in New output, and lactate) within the first hour. Adoption of
Caledonia and have found it easy to use (unpublished this goal-directed protocol was associated with a clear
data). This technique generally uses a radial arterial reduction of mortality. However, protocols must be dis-
catheter and a peripheral or central venous catheter. cussed and accepted by whole ICU or ED teams. Further,
The PiCCO technology has been more adoption of such protocols should not preclude careful
widely employed in ICUs, and enables measurement assessment of the causes of shock because therapeutic
of volumetric parameters and extravascular lung interventions may differ depending upon the underlying
water.18,19,21,22 The analysis of the wave form is simpler cause.1,6,7 Protocolized cardiovascular management based
than with with the LiDCO technology and is a calcu- on ventricular–arterial coupling has been proposed by
lation of the area under the systolic portion of the Pinsky.97,97a,97b The Pinsky protocol involves both im-
arterial pressure waveform. Ideally, PiCCO requires a mediate and ‘‘second-step’’ procedures. The second step
femoral arterial catheter, but a peripheral vein may be relies on therapies based on specific mechanisms of shock.
usable. This technology allows measurement of volu- Specific drugs and therapeutics are tailored to the under-
metric parameters such as the global end-diastolic lying cause of shock (e.g., thrombolysis of a myocardial
volume index (GEDVI) and the intrathoracic blood infarction or pulmonary embolism; drainage for pericar-
volume index (ITBVI) by transcardiopulmonary ther- dial tamponade; appropriate cardiovascular drugs, antiar-
modilution. These parameters are more accurate than rhythmic drugs, inotropic, vasoconstrictor, or vasodilator
left and right cardiac filling pressures to estimate pre- agents; adequate fluid loading, etc.).1
load in critically ill patients but do not reliably assess
fluid responsiveness.19,25
The noninvasive Modelflow-Finapress (Finapress COST-EFFECTIVENESS OF EACH METHOD
Medical Systems, Arnhem, The Netherlands) method AND STRATEGY
also uses similar arterial pulse contour analysis of finger Data comparing morbidity, mortality, and cost of ther-
arterial pressure in critically ill patients.18 Expressed as apeutic interventions are critical.43
cardiac index (CI), the results are similar to studies using Regarding cost-effectiveness, only the PAC
more invasive methods.18 has been extensively studied. Esophageal Doppler and
echocardiography are increasingly being used in many using passive leg raising.96,102 This is an important step
ICUs, but outcomes data comparing these strategies to because previous studies assessing prediction of preload
conventional strategies are lacking. Randomized studies responsiveness were validated in patients requiring
to compare the clinical impact of echocardiography with continuous MV and sedation. In patients with cardio-
arterial pulse contour analysis would be of interest but to vascular or respiratory instability, we use in our ICU,
our knowledge have not been done. The appropriate use echocardiography to predict fluid responsiveness using
and optimal interpretation of data are at least as im- IVC and aortic Doppler respiratory variations, even in
portant as the choice of the technique itself. patients not requiring MV (unpublished data).
DISCUSSION CONCLUSION
Although the classification of shock has not changed, Within the last few years, protocol-driven management
several major changes have evolved within the past few of patients with shock has been associated with improved
years. First, the major scientific critical care societies outcomes. These sentinel studies focused on septic and
have defined quantitative parameters (primarily for sep- hypovolemic shock, but lessons learned may apply to
tic shock) to allow a faster and a more unified diagnosis other types of shock states. Monitoring techniques that
not only at the bedside but also for common inclusion couple measurement of CO with ventilatory variations
criteria in clinical studies.2,8,98,99 Second, early manage- of systolic arterial pressure, pulse pressure, and SV
ment of patients in shock must be aggressive and enhance the ability to predict fluid responsiveness in
orchestrated. This critical time has been called ‘‘the circulatory failure. With this strategy, inappropriate
golden first hour,’’ and is to be compared with the delays volume expansion can be avoided in some patients,
defined for thrombolysis of cerebral vascular accidents whereas patients who may benefit from fluid expansion
(CVAs) and ST elevation myocardial infarction are more easily and readily identified.
(STEMI).
Within the past decade, novel techniques have
supplanted the use of PACs [e.g., echocardiography, REFERENCES
esophageal Doppler monitoring, arterial pulse contour
analysis (PiCCO and LiDCO technologies), sublin- 1. Walley KR, Wood DH. Shock. In: Hall JB, Schmidt GA,
gual PCO2 monitoring, etc.]. Regarding echocardio- Wood DH, eds. Principles of Critical Care. 3rd ed. New
graphy, we believe that a multilevel training can be York: McGraw Hill; 2005:249–265
2. Rivers E, Nguyen B, Havstadt S, et al. Early goal therapy in
defined: a basic level with five to 20 diagnoses that can
the treatment of severe sepsis and septic shock. N Engl J Med
be learned in a few hours,100and a second level that 2001;345:1368–1377
would require traditional echocardiography training 3. Weil MH. Defining hemodynamic instability. In: Pinsky
(over 1 or more years). These two levels of training MR, Payen D, eds. Functional Hemodynamic Monitoring.
are being set up in France.20 We think that every ICU Update in Intensive Care and Emergency Medicine. No. 42.
should have within its team an intensivist, also board New York: Springer-Verlag; 2005:9–17
certified in cardiovascular diseases, and in echocardiog- 4. American College of Chest Physicians/Society of Critical
Care Medicine. Consensus Conference: definitions of sepsis
raphy. The choice of equipment is important. Within
and organ failure and guidelines for the use of innovative
the ICU, optimal equipment and technology are essen- therapies in sepsis. Crit Care Med 1992;20:864–874
tial. In this setting, handheld machines are not ad- 5. Levy MM, Fink MP, Marshal JC, et al. 2001 SCCM/
equate for the sickest patients.101We believe that ESICM/ACCP/ATS/SIS International Sepsis Definition
echocardiography is the most important tool to Conference. Crit Care Med 2003;31:1250–1256
promptly diagnosis and treat shock. A careful echocar- 6. Dellinger RP, Carlet JM, Masur H, et al. Surviving Sepsis
diographic examination can establish an accurate initial Campaign for management of severe sepsis and septic shock.
Crit Care Med 2004;32:858–873
diagnosis; serial studies 12 or 24 hours later may assist
7. Hollenberg SM, Ahrens TS, Annane D, et al. Practice
in follow-up. For unstable patients, continuous mon- parameters for hemodynamic support of sepsis in adult
itoring is required. The arterial pulse contour techni- patients: 2004 update. Crit Care Med 2004;32:1928–
ques are useful in this context and can replace the 1948
PAC. Importantly, CO, prediction of preload respon- 8. Annane D, Belissant E, Cavaillon JM. Septic shock. Lancet
siveness with respiratory variations of systolic arterial 2005;365:63–78
pressure, pulse pressure, SV, and other hemodynami- 9. Weil MH, Shubin H. Shock following acute myocardium
infarction: current understanding of hemodynamic mechan-
cally derived variables can be obtained with these
isms. Prog Cardiovasc Dis 1968;11:1–17
techniques.18,19,21,22 Esophageal Doppler may also be 10. Shoemaker WC. Diagnosis and treatment of shock and
useful in such circumstances.23,59,96 Esophageal Dop- circulatory dysfunction. In: Shoemaker WC, Ayres SM,
pler was beneficial in patients requiring MV but less Grenvik A, Holbrook PR, eds. Textbook of Critical Care.
in spontaneous breathing activity or with arrythmias, 4th ed. Philadelphia: WB Saunders; 2000:92–114
11. Joly HR, Weil MH. Temperature of the great toe as an pressure and cardiorespiratory changes. Surgery 1975;78:
indicator of the severity of shock. Circulation 1969;39:131– 304–315
139 28. Shippy CR, Appel PL, Shoemaker WC. Reliability of
12. Cohn JN. Blood pressure measurement in shock. Mechan- clinical monitoring to assess blood volume in critically ill
isms of inaccuracy in auscultatory and palpatory methods. patients. Crit Care Med 1984;12:107–112
JAMA 1967;199:118–122 29. Magder S. How to use central venous pressure measure-
13. Barnett VT, Schmidt GA. Acid–base disorders. In: Hall JB, ments. Curr Opin Crit Care 2005;11:264–270
Schmidt GA, Wood DH, eds. Principles of Critical Care. 30. Magder S. The use of central venous pressure in critically ill
2nd ed. New York: McGraw Hill; 1998:1169–1181 patients. In: Pinsky MR, Payen D, eds. Update in Intensive
14. Weil MH, Rackow EC, Trevino R, et al. Difference in Care and Emergency Medicine. No. 42. New York: Springer-
acid–base state between venous and arterial blood during Verlag; 2005:299–311
cardiopulmonary resuscitation. N Engl J Med 1986;315: 31. Smith T, Grounds RM, Rhodes A. Central venous pressure:
153–156 uses and limitations. In: Pinsky MR, Payen D, eds.
15. Weil MH, Nakagawa Y, Tang W, et al. Sublingual Functional Hemodynamic Monitoring. Update in Intensive
tonometry: a new noninvasive measurement for diagnosis Care and Emergency Medicine. No. 42. New York: Springer-
and quantitation of circulatory shock. Crit Care Med 1999; Verlag; 2005:99–110
27:1225–1229 32. Boldt J, Lenz M, Kumle B, et al. Volume replacement
16. Marik PE. Regional carbon dioxide monitoring to assess the strategies on ICU: results from a postal survey. Intensive
adequacy of tissue perfusion. Curr Opin Crit Care 2005;11: Care Med 1998;24:147–151
245–251 33. Gattinoni L, Brazzi L, Pelosi P, et al. A trial of goal-oriented
17. Reinhart K, Bloos F. The value of venous oxymetry. Curr hemodynamic therapy in critically ill patients: SVO2
Opin Crit Care 2005;11:259–263 Collaborative Group. N Engl J Med 1995;333:1025–1032
18. Jansen JRC, van den Berg PCM. Cardiac output by 34. Hall JB. Mixed oxygen saturation (SVO2). In: Pinsky MR,
thermodilution and arterial pulse contour techniques. In: Payen D, eds. Functional Hemodynamic Monitoring.
Pinsky MR, Payen D, eds. Functional Hemodynamic Update in Intensive Care and Emergency Medicine. No.
Monitoring. Update in Intensive Care and Emergency 42. New York: Springer-Verlag; 2005:233–240
Medicine. No. 42. New York: Springer-Verlag; 2005:135–152 35. Connors AF Jr, Speroff T, Dawson NV, et al. The effective-
19. Reuter DA, Goetz AE. Arterial pulse contour analysis: ness of right heart catheterization in the initial care of critically
applicability to clinical routine. In: Pinsky MR, Payen D, ill patients. JAMA 1996;276:889–897
eds. Functional Hemodynamic Monitoring. Update in 36. Richard C, Warszawski J, Anguel N, et al. Early use of the
Intensive Care and Emergency Medicine. No. 42. New pulmonary artery catheter and outcome in patients with
York: Springer-Verlag; 2005:175–182 shock and acute respiratory distress syndrome. JAMA 2003;
20. Vignon P, Guéret P. Formation des réanimateurs à 290:2713–2720
l’échographie cardiaque. Réanimation. 2004;13:126–130 37. Iberti TJ, Fisher EP, Leibowitz AB, et al. A multicenter
21. Groenenveld ABJ, Breukers RMBGE, Verheij J. Clinical study of physician’s knowledge of the pulmonary artery
value of intrathoracic volumes from transpulmonary indi- catheter. Pulmonary Artery Catheter Study Group. JAMA
cator dilution. In: Pinsky MR, Payen D, eds. Functional 1990;264:2928–2932
Hemodynamic Monitoring. Update in Intensive Care and 38. Gnaegi A, Feihl F, Perret C. Intensive care physicians’
Emergency Medicine. No. 42. New York: Springer-Verlag; insufficient knowledge of right heart catheterization at the
2005:153–163 bedside: time to act? Crit Care Med 1997;25:213–220
22. Groenenveld ABJ, Verheij J. Methodology and value of 39. Sandham JD, Hull RD, Brant RF, et al. A randomized
assessing extravascular lung water. In: Pinsky MR, Payen D, controlled trial of the use of pulmonary artery catheters in
eds. Functional Hemodynamic Monitoring. Update in high risk surgical patients. N Engl J Med 2003;348:5–14
Intensive Care and Emergency Medicine. No. 42. New 39a. The National Heart, Lung, and Blood Institute Acute
York: Springer-Verlag; 2005:175–182 Respiratory Distress Syndrome (ARDS) Clinical Trials
23. Singer M. Esophageal Doppler monitoring. In: Pinsky MR, Network. Pulmonary-artery versus central venous catheter
Payen D, eds. Functional Hemodynamic Monitoring. to guide treatment of acute lung injury. N Engl J Med
Update in Intensive Care and Emergency Medicine. No. 2006;354:2213–2224
42. New York: Springer-Verlag; 2005:193–204 39b. Shure D. Pulmonary-artery catheters—peace at last? N Engl
24. Vignon P. Hemodynamic assessment of critically ill patients J Med 2006;354:2273–2274
using echocardiography Doppler. Curr Opin Crit Care 2005; 40. Kern JW, Shoemaker WC. Meta-analysis of hemodynamic
11:227–234 optimization in high risk patients. Crit Care Med 2002;30:
25. Michard F, Teboul JL. Predicting fluid response in ICU 1686–1692
patients: a critical analysis of the evidence. Chest 2002;121: 41. Boyd O, Grounds RM, Bennett ED. A randomized clinical
2000–2008 trail of the effect of deliberate perioperative increase of
26. Kumar A, Anel R, Bunnell E, et al. Pulmonary artery oxygen delivery on mortality in high risk patients. JAMA
occlusion pressure and central venous pressure fail to predict 1993;270:2699–2707
ventricular filling volume, cardiac performance, or the 42. Wilson J, Woods I, Fawcett J, et al. Reducing the risk of
response to volume infusion in normal subjects. Crit Care major elective surgery: randomized controlled trial of
Med 2004;32:691–699 preoperative optimization of oxygen delivery. BMJ 1999;
27. Baek SM, Makabali G, Bryan-Brown C, et al. Plasma 318:1099–1103
expansion in surgical patients with high CVP; the relation- 43. Wendon J. Cost effectiveness of monitoring techniques. In:
ship of blood volume to hematocrit, CVP, pulmonary wedge Pinsky MR, Payen D, eds. Functional Hemodynamic
Monitoring. Update in Intensive Care and Emergency 60. Jardin F. Cyclic changes in arterial pressure during mechan-
Medicine. No. 42. New York: Springer-Verlag; 2005:397–413 ical ventilation. Physiological note. Intensive Care Med
44. Slama M. Cardiac preload evaluation using echocardio- 2004;30:1047–1050
graphic technique. In: Pinsky MR, Payen D, eds. Functional 61. Pinsky MR. Using ventilation-induced aortic pressure and
Hemodynamic Monitoring. Update in Intensive Care and flow variation to diagnose preload reponsiveness [editorial].
Emergency Medicine. No. 42. New York: Springer-Verlag; Intensive Care Med 2004;30:1008–1010
2005:259–268 62. Feissel M, Michard F, Faller JP, et al. The respiratory cycle
45. Marini JJ, Leatherman JW. Pulmonary artery occlusion in inferior vena cava diameter as a guide to fluid therapy.
pressure: measurement, significance, and clinical uses. In: Intensive Care Med 2004;30:1834–1837
Pinsky MR, Payen D, eds. Functional Hemodynamic 63. Barbier C, Loubieres Y, Schmitt C, et al. Respiratory
Monitoring. Update in Intensive Care and Emergency changes in inferior vena cava diameter are helpful in
Medicine. No. 42. New York: Springer-Verlag; 2005:111–134 predicting fluid responsiveness in ventilated patients. Inten-
46. Michard F, Alaya S, Zarka V, et al. Global end diastolic sive Care Med 2004;30:1740–1746
volume as an indicator of cardiac preload in patients with 64. Vieillard-Baron A, Chergui K, Rabiller A, et al. Superior
septic shock. Chest 2003;124:1900–1908 vena collapsibility as a gauge of volume status in ventilated
47. Axler O. Physiologie et Méthodes de Mesure du Débit septic patients. Intensive Care Med 2004;30:1734–1739
Cardiaque (Physiology and Methods of Cardiac Output 65. Vignon P. Evaluation of fluid responsiveness in ventilated
Measurement). 1996. Thèse de Sciences de la Vie et de la septic patients: back to venous return [editorial]. Intensive
Santé. (PhD Thesis). Université Paris VI. Faculté de Care Med 2004;30:1699–1701
Médecine Cochin. Paris 66. Pavlides GS, Hauser AM, Stewart JR, et al. Contribution of
48. Axler O, Tousignant C, Thompson CR, et al. Comparison transesophageal echocardiography to patient diagnosis and
of transesophageal echocardiographic, Fick and thermodilu- treatment: a prospect analysis. Am Heart J 1990;120:910–
tion cardiac output in critically ill patients. J Crit Care 914
1996;11:109–116 67. Pearson AC, Castello R, Labovitz AJ, et al. Safety and utility
49. Axler O, Megarbane B, Lentschener C, et al. Comparison of of transesophageal echocardiography in the critically ill
cardiac output methods measured by echocardiographic patient. Am Heart J 1990;119:1083–1089
volumes and aortic pulsed Doppler methods during mechan- 68. Oh JK, Seward JB, Khanderia BK, et al. Transesophageal
ical ventilation. Intensive Care Med 2003;29:208–217 echocardiography in critically ill patients. Am J Cardiol
50. Jardin F. Right heart catheterization at bedside: a critical 1990;66:1492–1499
view. Intensive Care Med 1995;21:291–295 69. Reichert C, Visser C, Koolen J, et al. Transesophageal
51. Janssen JRC, van den Berg PCM. Cardiac ouput by echocardiography in hypotensive patients after cardiac
thermodilution and arterial pulse contour techniques. In: operations: comparison with hemodynamic parameters.
Pinsky MR, Payen D, eds. Functional Hemodynamic J Thorac Cardiovasc Surg 1992;104:321–326
Monitoring. Update in Intensive Care and Emergency 70. Hwang JJ, Shyu KG, Chen JJ, et al. Usefulness of
Medicine. No. 42. New York: Springer-Verlag; 2005:135– transesophageal echocardiography in the treatment of
152 critically ill patients. Chest 1993;104:861–866
52. Boldt J. Right ventricular end-diastolic volume. In: Pinsky 71. Vignon P, Mentec H, Terré S, et al. Diagnostic accuracy and
MR, Payen D, eds. Functional Hemodynamic Monitoring. therapeutic impact of transthoracic and transesophageal
Update in Intensive Care and Emergency Medicine. No. 42. echocardiography in mechanically ventilated patients in the
New York: Springer-Verlag; 2005:269–281 ICU. Chest 1994;106:1829–1834
53. Michard F, Teboul JL. Using heart–lung interactions to 72. Poelaert J, Trouerbach J, De Buyzere M, et al. Evaluation of
assess fluid responsiveness during mechanical ventilation. transesophageal echocardiography as a diagnostic and
Crit Care Med 2000;4:282–289 therapeutic aid in a critical care setting. Chest 1995;107:
54. Pinsky MR. Assessment of indices of preload and preload 774–779
responsiveness. Curr Opin Crit Care 2005;11:235–239 73. Heidenreich PA, Stainback RF, Redberg RF, Schiller NB,
55. Vincent JL. DO2/VO2 relationships. In: Pinsky MR, Payen Cohen NH, Foster E. Transesophageal echocardiography
D, eds. Functional Hemodynamic Monitoring. Update in predicts mortality in critically ill patients with unexplained
Intensive Care and Emergency Medicine. No. 42. New hypotension. J Am Coll Cardiol 1995;26:152–158
York: Springer-Verlag; 2005:251–258 74. Sohn DW, Shin GJ, Oh JK, et al. Role of transesophageal
56. Lichtenstein D, Axler O. Intensive use of general ultrasound echocardiography in hemodynamically unstable patients.
in the intensive care unit, a prospective study of 150 patients. Mayo Clin Proc 1995;70:925–931
Intensive Care Med 1993;19:353–355 75. Slama MA, Novara A, Van de Putte P, et al. Diagnostic and
57. Feissel M, Michard F, Mangin I, et al. Respiratory changes therapeutic implications of transesophageal echocardiogra-
in aortic blood velocity as an indicator of fluid responsiveness phy in medical ICU patients with unexplained shock,
in ventilated patients with septic shock. Chest 2001;119: hypoxemia, or suspected endocarditis [review]. Intensive
867–873 Care Med 1996;22:916–922
58. Slama M, Masson H, Teboul JL, et al. Respiratory variations 76. Megarbane B, Axler O, Mazhigi M, et al. Assessment of
of aortic VTI: a new index of hypovolemia and fluid echocardiography indications and results in an intensive care
responsiveness. Am J Physiol Heart Circ Physiol 2002; unit [abstract]. Am J Respir Crit Care Med 1999;159:A520
283:H1729–H1733 77. Schmidlin D, Schuepbach R, Bernard E, et al. Indications
59. Slama M. Monitoring of respiratory variations of aortic and impact of postoperative transesophageal echocardiogra-
blood flow velocity, using esophageal Doppler. Intensive phy in cardiac surgical patients. Crit Care Med 2001;29:
Care Med 2004;30:1182–1187 2143–2148
78. Colreavy FB, Donovan K, Lee KY, et al. Transesophageal pressure in surgical patients with postoperative circulatory
echocardiography in critically ill patients. Crit Care Med shock and acute lung injury. Anesthesiology 2003;98:1091–
2002;30:989–996 1100
79. Hüttemann E, Schelenz C, Kara F, et al. The use and safety 93. Combes A, Arnoult F, Trouillet JL, et al. Tissue Doppler
of transoesophageal echocardiography in the general ICU a imaging estimation of pulmonary capillary wedge pressure in
minireview. Acta Anaesthesiol Scand 2004;48:827–836 ICU patients. Intensive Care Med 2004;30:75–81
80. Joseph MX, Disney PJS, Da Costa R, et al. Transthoracic 94. Rhodes A, Sunderland R. Arterial pulse power analysis: the
echocardiography to identify or exclude cardiac cause of LiDCO plus system. In: Pinsky MR, Payen D, eds.
shock. Chest 2004;126:1592–1597 Functional Hemodynamic Monitoring. Update in Intensive
81. Kumar A, Anel R, Bunnell E, et al. Pulmonary artery Care and Emergency Medicine. No. 42. New York: Springer-
occlusion pressure and central venous pressure fail to predict Verlag; 2005:183–192
filling volume, cardiac performance, or the reponse to volume 95. Perel A, Preisman S, Berkenstadt H. Arterial pressure
infusion in normal subjects. Crit Care Med 2004;32:691–699 variation during positive-pressure ventilation. In: Pinsky
82. Cheung AT, Savino JS, Weiss SJ, et al. Echocardiography MR, Payen D, eds. Functional Hemodynamic Monitoring.
and hemodynamic indexes of left ventricular preload in Update in Intensive Care and Emergency Medicine. No. 42.
patients with normal and abnormal ventricular function. New York: Springer-Verlag; 2005:313–329
Anesthesiology 1994;81:376–387 96. Teboul JL, Monnet X, Richard C. Arterial pulse pressure
83. Axler O, Tousignant C, Thompson CR, et al. Small variation during positive pressure ventilation and passive leg
hemodynamic effect of typical rapid volume infusions in raising. In: Pinsky MR, Payen D, eds. Functional Hemo-
critically ill patients. Crit Care Med 1997;25:965–970 dynamic Monitoring. Update in Intensive Care and
84. Tavernier B, Makhotine O, Lebuffe G, et al. Systolic pressure Emergency Medicine. No. 42. New York: Springer-Verlag;
variation as a guide to fluid therapy in patients with sepsis- 2005:332–343
induced hypotension. Anesthesiology 1998;89:1313–1321 97. Pinsky MR. Protocolized cardiovascular management based
85. Tousignant CP, Walsh F, Mazer CD. The use of transe- on ventricular-arterial coupling. In: Pinsky MR, PayenD,
sophageal echocardiography for preload assessment in crtically eds. Functional Hemodynamic Monitoring. Update in
ill patients. Anesth Analg 2000;90:351–355 Intensive Care and Emergency Medicine. No. 42. New
86. Nagueh S, Kopelen HA, Zoghbi WA, et al. Relation of York: Springer-Verlag; 2005:381–395
mean right atrial pressure to echocardiographic and Doppler 97a. The National Heart, Lung, and Blood Institute Acute
parameters of right atrial and ventricular function. Circula- Respiratory Distress Syndrome (ARDS) Clinical Trials
tion 1996;93:1160–1169 Network. Comparison of two fluid management strategies
87. Giannuzzi P, Imparato A, Temporelli PL. Doppler-derived in acute lung injury. N Engl J Med, June 2006;10.1056/
mitral deceleration time of early filling as a strong predictor NEJM oa062200. Available at www.nejm.org
of pulmonary capillary wedge pressure in post-infarction 97b. Rivers EP. Fluid-management strategies in acute lung
patients with left ventricular systolic dysfunction. J Am Coll injury—liberal, conservative, or both? N Engl J Med, June
Cardiol 1994;23:1630–1637 2006;10.1056/NEJM e068105. Available at www.nejm.org
88. Nagueh S, Middleton KJ, Kopelen HA, et al. Doppler tissue 98. Bernard GR, Vincent JL, Laterre PF, et al. Recombinant
imaging: a non invasive technique for evaluation of of left human protein C Worldwide Evaluation in Severe Sepsis
ventricular relaxation and estimation of filling pressures. (PROWESS) study group: efficacy and safety of recombi-
J Am Coll Cardiol 1997;30:1527–1533 nant human activated protein C for severe sepsis. N Engl J
89. Gonzalez-Vilchez F, Ares M, Ayuela J, et al. Combined use Med 2001;344:699–709
of pulsed and color M-Mode Doppler echocardiography for 99. van den Berghe G, Wouters P, Weekers F, et al. Intensive
the estimation of pulmonary capillary wedge pressure: an insulin therapy in the critically ill patients. N Engl J Med
empirical approach based on an analytical relation. J Am Coll 2001;345:1359–1367
Cardiol 1999;34:515–523 100. Liang DH, Thompson N. A program for the use of hand-
90. Ommen SR, Nishimura RA, Appleton CP, et al. Clinical held echocardiography by intensivists to augment physical
Utility of Doppler Echocardiography and Tissue Doppler exam assessment. 33rd Critical Care Medicine Congress.
Imaging in the Estimation of Left Ventricular Filling Orlando Feb 2004. Crit Care Med 2003:33(Suppl)–
Pressures. A Comparative Simultaneous Doppler-Catheter- A21
ization Study. Circulation 2000;102:1788–1794 101. Vignon P, Frank MBJ, Lesage J, et al. Handheld echocardio-
91. Boussuges A, Blanc P, Molenat F, et al. Evaluation in left graphy with Doppler capability for the assessment of critically
ventricular filling pressure by transthoracic Doppler echo- ill patients: is it reliable? Intensive Care Med 2004;30:718–
cardiography in the intensive care unit. Crit Care Med 2002; 723
30:362–367 102. Monnet X, Rienzo M, Osman D, et al. Passive leg raising
92. Bouhemad B, Nicolas-Robin A, Benois A, et al. Echocardio- predicts fluid responsiveness in the critically ill. Crit Care
graphic Doppler assessment of pulmonary capillary wedge Med 2006;34:1402–1407
Intensive Management of Hepatic Failure
Mary E. Rinella, M.D.1 and Arun Sanyal, M.D.2
ABSTRACT
A substantial number of patients with liver failure are admitted to the intensive
care unit; thus a thorough understanding of the prevention and treatment of complications
in such patients is imperative. The management of liver failure is demanding and often
involves the combined efforts of many specialists. Critically ill patients with hepatic failure
encompass a broad spectrum of disease, ranging from acute liver failure in a patient with no
prior history of liver disease, to acute on chronic liver failure. The initial assessment and
management of acute liver failure are reviewed with an emphasis on the prevention and
treatment of brain edema in the pretransplant setting. The current treatment of compli-
cations resulting from decompensated chronic liver disease such as portal hypertensive
bleeding; infection, renal failure, and hepatic encephalopathy are then discussed.
KEYWORDS: Liver failure, cerebral edema, portal hypertension, management
T he management of liver failure is demanding plantation. Although ALF remains one of the most acute
and often involves the combined efforts of many special- serious illnesses, thoughtful intensive management can
ists. Critically ill patients with hepatic failure encompass optimize the patient’s chances for spontaneous hepatic
a broad spectrum of disease, ranging from acute liver regeneration or a successful liver transplant.3 When
failure in a patient with no prior history of liver disease, possible, etiology-targeted therapy should be initiated
to end-stage decompensated cirrhosis. Both sides of this (Table 1). The goal of management should be focused
spectrum present clinical challenges that involve many on the prevention of systemic infection, multiorgan fail-
organ systems. Although both sides in acute and chronic ure, hepatic encephalopathy (HE), and ultimately the
liver failure can have a poor prognosis, careful and development of brain edema.4–6 At this time liver trans-
comprehensive intensive care can improve outcome and plantation is the only definitive therapy for those who
bridge eligible patients to liver transplantation. Because fulfill criteria for poor prognosis7–9 (Table 2). The
acute and chronic liver failure are very distinct clinical challenge to the clinician is selection of patients for
entities, they will be discussed separately. transplant that have low likelihood of spontaneous sur-
vival but are not too ill to benefit from transplantation.
The principles of management of ALF are reviewed here:
ACUTE LIVER FAILURE
Acute liver failure (ALF) is a rapidly progressive, often
fatal syndrome characterized by jaundice, encephalop-
athy, and coagulopathy leading to multiorgan failure in a INITIAL EVALUATION AND MANAGEMENT
patient with no prior history of liver disease.1,2 In recent Early diagnosis and identification of the subject that is
years, advancements in supportive care have improved unlikely to improve spontaneously constitute a critical
survival and provided a more effective bridge to trans- first step in the management of ALF. The initial triage
1
Division of Hepatology, Northwestern University, Chicago, Illinois; Non-pulmonary Critical Care: Managing Multisystem Critical Ill-
2
Division of Gastroenterology, Department of Internal Medicine, ness; Guest Editor, Curtis N. Sessler, M.D.
Virginia Commonwealth University, Richmond, Virginia. Semin Respir Crit Care Med 2006;27:241–261. Copyright # 2006
Address for correspondence and reprint requests: Arun Sanyal, by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
M.D., Division of Gastroenterology, Department of Internal Med- NY 10001, USA. Tel: +1(212) 584-4662.
icine, Virginia Commonwealth University, MCV Box 980341, DOI 10.1055/s-2006-945528. ISSN 1069-3424.
Richmond, VA 23298-0341. E-mail: [email protected].
241
Table 1 Etiology-Targeted Therapy and an accurate history cannot be overemphasized be-

Etiology Potential Therapies cause both treatment and prognosis are significantly
affected by the underlying etiology. A detailed account
TOXIC of the psychiatric history, including suicidal ideation
Acetaminophen N-acetyl cysteine and family support, is essential to assess suitability
Amanita poisoning Penicillin and silibinin for transplantation. The timing of the psychiatric evalua-
VIRAL tion is of particular importance, given the rapid deterio-
Herpes simplex virus Acyclovir ration in mental status that occurs in such patients.
Acute heptatitis B Antivirals?
METABOLIC
Wilson’s disease Transplant DISEASE-TARGETED THERAPY
Autoimmune hepatitis Corticosteroids A thorough discussion of the differential diagnosis of
VASCULAR ALF is beyond the scope of this review; however,
Acute Budd-Chiari syndrome Directed thrombolysis, Table 1 provides a summary of common etiologies of
transjugular intrahepatic ALF for which potential therapies exist. Only acetami-
portosystemic shunt nophen will be discussed in more detail because it is the
PREGNANCY most common etiology of liver failure in the United
Acute fatty liver of Urgent delivery States and has an effective antidote.
pregnancy/HELLP
HELLP, hemolysis elevated liver enzymes low platelets.
Acetaminophen
of a patient with acute liver injury to an intensive care Idiopathic and drug-related liver injuries are the most
unit (ICU) is based on the presence of altered mental common causes of ALF in the United States.10 Of the
status and the degree of coagulopathy. It is imperative drug-related causes, acetaminophen overdose is the most
to admit most subjects with acute liver injury with an common cause of ALF in the United Kingdom and
international normalized ratio (INR) > 1.5 and all United States. Overdose can be either intentional or
subjects with mental status changes. Rapid deteriora- unintentional.11–13 The patient, family, and close contacts
tion can occur and is often irreversible in the patient must be questioned about regular alcohol use, dieting, diet
with ALF. It is therefore imperative that decisions pills, medications, or recent illness that may have resulted
regarding prognosis and appropriateness for liver trans- in poor nutrition. These factors greatly affect toxicity
plant be made early, and potentially suitable patients either through upregulating cytochrome p450 (alcohol
should be referred to a liver transplant center early in and other drugs) promoting the formation of toxic inter-
the evaluation process. mediates, or through glutathione depletion. Such details
The management of patients in liver failure are important because as little as 2.6 to 4.0 g of acetami-
requires a multidisciplinary approach involving hepatol- nophen can lead to liver failure in this setting.14–17
ogists, transplant surgeons, intensivists, and other sub- It is worth noting that, at the time of presenta-
specialists. The importance of a thorough physical exam tion, a patient with acetaminophen-induced liver failure
may have undetectable blood levels of acetaminophen.
Table 2 King’s College Criteria for Acute Liver Failure This is particularly true when the toxicity manifests itself
several days after ingestion of acetaminophen for ther-
Acetaminophen induced
apeutic purposes in a susceptible subject. However, in
Arterial blood pH < 7.3 (regardless of degree of
the majority of cases, detectable acetaminophen levels
encephalopathy)
are present at the time of presentation. When acetami-
If no acidosis then all three of the below criteria:
nophen overdose is confirmed, N-acetylcysteine (NAC)
Prothrombin time > 100 seconds
must be initiated in a timely manner, ideally within
Serum Creatinine > 2.5 mg/dL
16 hours of ingestion, to have a significant impact on
Grade 3 or 4 encephalopathy
survival. NAC decreases injury through enhancement of
Nonacetaminophen induced
glutathione synthesis resulting in less formation of
Prothrombin time > 100 seconds
acetaminophen’s hepatotoxic intermediate.18,19 Even if
If prothrombin time < 100 seconds, then any of the below
the patient is delayed in reaching the hospital or the
criteria (regardless of degree of encephalopathy):
diagnosis is not forthcoming, there is evidence that late
Drug-induced, non-A, non-B, halothane hepatitis
administration of NAC can be beneficial.20 NAC may
Time from jaundice to encephalopathy > 7 days
also improve outcome through its effects on micro-
Age < 10 or > 40 years
circulatory function. A large multicenter study (the
Prothrombin > 50 seconds
ALF study group) is currently addressing the utility of
Bilirubin > 17.5 mg/dL
NAC in nonacetaminophen-induced ALF.
INTENSIVE MANAGEMENT OF HEPATIC FAILURE/RINELLA, SANYAL 243
Table 3 Acute Liver Failure: General Management Guidelines

On Admission Daily Tid Hourly If Indicated
Monitoring IV access, CVP and arterial Blood sugar Mental status Mechanical intubation,
line, Foley catheter ICP monitoring
Thorough history Interview family members

and physical
Laboratories Liver panel, renal panel, Basic laboratories, AFP, Blood gas Changes in
CBC, PT, Hep A,B,C arterial ammonia ICP monitor
serologies, HSV, CMV, (or more if mental
EBV, ceruloplasmin, ANA, status deteriorating
anti-sm Ab, SPEP, phosphate,
HIV, acetaminophen factor V level
level, toxicology screen,
cosyntropin stimulation
test, TSH, blood type,
blood cultures
Imaging US with Doppler Head CT for neurological

changes or suspected
edema
Directed therapy where indicated Drugs, cooling for

cerebral edema
AFP, alpha feta protein; ANA, antinuclear antibody; anti-sm Ab, antismooth muscle antibody; CBC, complete blood count; CMV, cytomegalo-
virus; CT, computed tomography; CVP, central venous pressure; EBV, Epstein-Barr virus; HSV, herpes simplex virus; HIV, human immunode-
ficiency virus; ICP, intracranial pressure; IV, intravenous; PT, prothrombin time; SPEP, serum protein electrophoresis; TSH, thyroid stimulating
hormone; US, ultrasound.
MONITORING AND GENERAL GUIDELINES overload and pulmonary edema, especially when renal
Acute hepatic dysfunction has profound effects on many function begins to deteriorate. Alternate approaches in-
organs. Therefore, one must remain cognizant of the clude the use of plasmapheresis where a volume of plasma
ramifications of specific therapies on other systems. The equal to the amount infused is removed to prevent
mental status must be documented several times daily, in volume overload. Recently, recombinant factor VII
addition to frequent assessment of hepatic synthetic (40 mg/kg) has been used in conjunction with FFP to
function and blood glucose (Table 3). Although a liver rapidly correct coagulopathy prior to either intracranial
biopsy may be helpful if the diagnosis is in question, it can pressure monitor or central line placement in patients
be unreliable in predicting outcome and is risky given the with ALF.25
presence of underlying coagulopathy.21 Low serum phos-
phate and elevated a fetoprotein can be encouraging signs
of hepatic regeneration.22,23 In a retrospective analysis of MECHANICAL VENTILATION
ALF patients, 74% of patients with phosphate levels Mechanical ventilation should be initiated once ence-
< 2.5 were alive at 1 week, in contrast to none in those phalopathy deteriorates to grade 3 (West Haven cri-
with a serum phosphate > 5.24 teria) to protect the airway.26,27 In addition to preventing
Coagulopathy in ALF does not usually require aspiration in the patient with compromised mental status,
correction unless an invasive procedure is planned or intubation and sedatives help control agitation, which can
overt bleeding is present because the use of fresh frozen lead to surges in intracranial pressure. Patients with
plasma (FFP) can mask deterioration of liver function. A encephalopathy beyond grade 3 are very difficult to
common indication for the correction of coagulopathy manage without intubation and sedation.28
is placement of a central line. Traditionally, FFP and Sedation is best achieved with a short-acting
cryoprecipitates have been used for the correction of sedative alone or in combination with a short-acting
coagulopathy in subjects with ALF. This is only partially narcotic. Recent evidence supports the use of propofol
effective in correcting coagulopathy and its effects are for this purpose. In a small study, propofol was given to
short-lived. It is also associated with a risk of transmitting seven patients with ALF and profound encephalopathy.
cytomegalovirus infection and may contribute to volume Intracranial pressure (ICP) remained normal in six of
seven patients with ALF given propofol at 50 mg/kg/

min, suggesting propofol may have independent benefi-
cial effects on ICP.29 Paralytics are usually avoided
because they can mask seizure activity. However, they
may be used in specific cases to facilitate management
when the subject does not respond appropriately to
sedation. In such cases, it is imperative to consider the
possibility of seizure activity if the clinical picture
continues to deteriorate. H2 antagonists or proton
pump inhibitors may decrease the incidence of ulcer
disease in mechanically ventilated patients30; however,
the theoretical risk of increasing the incidence of
pneumonia has not been studied in this population.
PREVENTION AND MANAGEMENT

OF COMPLICATIONS
Figure 1 The effect of antibiotic prophylaxis on the prevalence
of documented infection patients with acute liver failure.
Circulatory Dysfunction (Adapted from Salmeron et al.47)
Derangements in circulatory function manifest early in
ALF and are often progressive. They are characterized by
generalized vasodilation, increased cardiac output, de- respectively. These infections presented at a median of
creased systemic vascular resistance, and a low mean 5, 3, and 2 days after the onset of ALF.44
arterial pressure (MAP).31–33 It is challenging to distin- Given the frequency of both gram-negative and
guish this clinical picture from the hemodynamics of gram-positive infections in this population, broad spec-
sepsis, particularly given that infection is common and trum antibiotic coverage should be administered avoid-
often a fatal complication.34 Factors such as adrenal ing aminoglycosides due to their nephrotoxicity.34
insufficiency also complicate management by making Although no randomized controlled trials have demon-
the vasculature less responsive to vasopressive agents.35,36 strated improved survival with prophylactic antibiotics,
In general terms, fluids and vasopressors should parenteral antibiotics are associated with a lower inci-
be used to maintain adequate cerebral perfusion pressure dence of infection46 (Fig. 1).47 Given these data, pro-
(CPP) (50 mm Hg to 65 mm Hg) while avoiding phylactic broad-spectrum antibiotics seem justifiable
cerebral hyperemia from hyperperfusion.37,38 Because given that uncontrolled infection in such patients is
the circulatory disturbance in ALF is characterized by often catastrophic.48,49
vasodilation and increased cardiac output, norepinephr-
ine is frequently the vasopressor of choice.
Systemic Inflammatory Response Syndrome
Even in the absence of documented infection, systemic
Infection inflammatory response syndrome (SIRS) is common in
Patients with ALF are particularly susceptible to severe those with ALF and is likely due to a surge of cytokine
infection due to many immunological defects such as release.50 In a study from King’s College, 57% of 887
defective phagocytic function and decreased comple- patients with ALF developed SIRS. The presence of
ment levels.39–42 Bacterial or fungal sepsis is a frequent SIRS on admission was independently associated with
cause of death in this population. Much like other more severe illness, worsening of encephalopathy, and
immunocompromised hosts, their response to infection subsequent death. In those patients that were infected
is atypical in that signs such as fever or leukocytosis are (54%), mortality increased with each additional compo-
absent in 30% of cases.43 Thus sepsis is both frequent nent of SIRS. At this point it remains unclear how
and difficult to diagnose in subjects with ALF. In a additional infection contributes or which component of
prospective study of 887 patients with ALF, one or more the observed inflammatory response originates from
bacterial infections occurred in 37.8%; however, an humoral factors released by the necrotic liver.34,46,51
incidence of up to 80% has been reported.44 Of these,
gram-positive cocci were the most common organisms
isolated, although Escherichia coli and Klebsiella were also Adrenocortical Insufficiency
frequent pathogens.45 Overall, pneumonias make up Adrenocortical insufficiency can worsen hyperdynamic
50% of bacterial infections in ALF with bacteremia cardiovascular collapse typical of ALF or septic shock.52
and urinary tract infections occurring in 20 and 25%, This should be considered when the patient fails
Figure 2 Factors leading to the development of brain edema and potential therapeutic interventions.
to respond to volume resuscitation.35 In sepsis, supra- on ICP.55,56 Moreover, intermittent hemodialysis has
physiological doses of steroids in patients with adrenal been associated with increases in ICP and decreases in
insufficiency have been shown to reduce vasopressor CPP, whereas the opposite has been shown in patients
requirements and improve outcome.36,53 Adrenal receiving CVVH.55,57
dysfunction appears to also be prevalent in patients
with ALF; 62% of patients with ALF were found to
have an abnormal response to high-dose corticotrophin Hepatic Encephalopathy and the Development
stimulation. The patients with pronounced hemody- of Intracranial Hypertension
namic instability had more marked evidence of adrenal The development of HE and subsequent cerebral edema
insufficiency, suggesting that it may contribute to the and intracranial hypertension (ICH) define prognosis in
pattern of cardiovascular collapse seen in liver failure.54 patients with ALF.2,7,58 Treatment options for such
The benefit of stress-dose steroids in this population patients are limited. As a result, 30% of patients
needs to be tested in a randomized-controlled trial; with ALF and cerebral edema succumb to cerebral
however, given these data, it is reasonable to look for herniation while awaiting an organ.7,31 Without urgent
and consider treating adrenal insufficiency in patients transplantation, mortality can exceed 90% in those who
with ALF. have uncontrolled ICH.
The pathogenesis of cerebral edema is complex
(Fig. 2). ALF leads to many hemodynamic changes,
Renal Failure including impairment of cerebrovascular autoregulation
Renal failure is common in those with advanced liver and blood flow. This impairment makes the standard
failure and is multifactorial in etiology. Common causes assumption that CPP ¼ MAP ICP less reliable.38
of renal failure in this population include prerenal Other factors such as high arterial ammonia levels
azotemia, renal ischemia, acute tubular necrosis, and contribute to brain edema through the accumulation
hepatorenal syndrome. A majority of patients with of glutamine and alanine in astrocytes. In response to
ALF complicated by profound hypotension and cerebral swelling, a vasodilating factor is released that leads to
edema will require renal replacement therapy.31 Due to increased CBF and thus increased ICP.59
the marked vasodilation that characterizes such patients, Arterial ammonia levels > 200 mmol/L in the
continuous venovenous hemofiltration (CVVH) tends to setting of ALF have been shown to herald impending
be better tolerated and may have more beneficial effects cerebral herniation and poor outcome.60,61 Other
Figure 3 Proposed algorithm for the use of an intracranial pressure monitor.
markers of brain cell dysfunction and damage such as Although treatments aimed at reducing ICP can
s100-b and neuron-specific enolase (NSE) have also be used without an ICP monitor, an accurate ICP
been evaluated as potential predictors of impending reading permits targeted therapy to optimize CPP and
herniation in the setting of ALF and acute on chronic detect abrupt surges in pressure that necessitate addi-
liver failure with negative results.62 Currently, no serum tional therapy. Concomitant measurement of jugular
markers of brain cell dysfunction reliably demonstrate bulb oxygen saturation,32,66 which allows measurement
neurological injury and poor outcome. of brain oxygen utilization, can also be useful in the
Unfortunately, it can be difficult to predict which management of these patients.32 Jugular bulb oxygen
patients are likely to develop elevated ICP. Clinical signs saturation > 80% or < 60% predicts elevation in ICP
such as arterial hypertension, fever, and agitation can with good sensitivity and specificity.31 Jugulovenous O2
precede episodes of severe ICH; however, these are not saturations < 50% may herald an increase in anaerobic
reliable predictors because elevated ICP is often clin- cerebral glycolysis, increased lactate:pyruvate ratio, and
ically silent.63 Although a computed tomographic (CT) worsening cerebral edema.67
scan is usually used to look for cerebral edema, a normal
scan does not exclude the presence of edema because its
appearance on imaging may be delayed. When and in Whom to Insert an Intracranial
Pressure Monitor
To justify the risks of ICP monitor placement, the
INTRACRANIAL PRESSURE MONITORING monitor needs to be placed under controlled circum-
A significant clinical challenge in the management of stances, when increased ICP is likely to rise but before
ALF is the decision to place an ICP monitor. There are uncontrolled ICH and herniation occur. ICP monitor-
no strict guidelines related to the use of these monitors ing should be considered for mechanically ventilated
and experience across institutions is highly variable. patients with grade 3 or 4 encephalopathy with poor
Noninvasive techniques have not proven to be benefi- prognosis (Table 2) but who are otherwise good candi-
cial and direct ICP monitoring is the only reliable dates for liver transplant (Fig. 3). Other predictors of
modality for the measurement of ICP. The benefit increased ICP such as arterial ammonia > 150 mmol/L
that can be derived from ICP monitoring is twofold. could be used to time monitor placement. In those with
First, it allows for the early detection and treatment of poor prognosis without orthopedic liver transplant
ICH because it can be clinically silent.63 Second, it can (OLT), ICP monitoring can guide therapy and prevent
provide invaluable information about the likelihood of surges in ICH before and during OLT.
neurological recovery when deciding whether to pro-
ceed with liver transplantation, such as when CPP is
persistently low. Sustained CPP < 40 mm Hg predicts Risks of Intracranial Pressure Monitoring
a high likelihood of ischemic brain injury that typically As with all interventions, the risks of ICP monitor
results in a poor neurological outcome after transplan- placement need to be balanced against the accuracy and
tation.64,65 Figure 3 proposes an algorithm for the use usefulness of the information to be gained. No random-
of ICP monitors in ALF. ized, controlled trial is available to compare different
catheters in the setting of ALF. Types of catheters include induction and maintenance of hypernatremia (145 to
epidural, subdural, parenchymal, and intraventricular 155 mmol/L) in patients with grade 3 or 4 encephalop-
catheters. Blei et al performed a survey of transplant athy resulted in a decreased incidence and severity of
centers across the United States. They estimated that ICH.72 Other techniques to reduce brain water accu-
20% of ICP monitoring resulted in intracranial bleeding. mulation through the reduction of arterial ammonia
Epidural catheters had the lowest rate of bleeding com- remain under investigation.73–75
plications (3.8%) and subdural and parenchymal catheters
the highest; 20% and 22%, respectively.68 A recent multi- MANNITOL
center study from the ALF study group showed that ICP Mannitol administration leads to increased plasma os-
monitors were only used in 92/332 patients (28%) with molality in brain capillaries, resulting in movement of
ALF and severe encephalopathy; however, the frequency water out of the brain according to Starling’s law. It has
of monitoring differed between centers. Ten percent had been shown to decrease episodes of cerebral edema and
intracranial bleeding as a result of the ICP monitor. In result in improved survival in a cohort of patients with
two of these patients, ICP monitoring was directly ALF (47.1 and 5.9%, respectively, p ¼ .008).76 Its use,
associated with the patient’s death.69 Although the risk however, is limited in renal failure and can lead to a
of complications is greater,70 subdural catheters give a paradoxical increase in brain swelling if osmolality is not
more reliable estimate of ICP than epidural catheters.68 controlled. If more than two doses are to be used, plasma
Bleeding complications can be decreased signifi- osmolality must be checked to assure that it remains
cantly with the use of recombinant factor rVIIa given < 320 Osm/L.
immediately before the procedure.25 The frequency of
factor VII dosing was variable in this study; however, as a HYPERVENTILATION
group all patients that received factor VII normalized Hyperventilation is an effective technique to decrease
their prothrombin time (PT) and were able to have ICP cerebral blood flow (CBF) and ICP. It does so through
monitors placed (compared with 38% in the FFP alone precapillary hypocapnic vasoconstriction and helps re-
group). The ideal initial dose and subsequent doses of store CBF autoregulation.61,77–79 Although prophylactic
factor VII necessitates further study.71 The data show hyperventilation appears to be ineffective in preventing
that ICP monitoring can be an effective tool for manag- the development of ICH,77 it can be useful in controlling
ing elevated intracranial pressure; however, ICP mon- acute surges in ICP.
itors have not been shown to improve survival. Currently
there is no consensus about the use of ICP monitoring or INDOMETHACIN
whether the more accurate but higher-risk subdural Indomethacin leads to cerebral vasoconstriction effects
catheters or the less accurate but safer epidural catheter via altering cerebral temperature and extracellular pH
should be used. Individual centers will continue to use and inhibition of the endothelial cyclooxygenase path-
what they are comfortable with; however, their decision way.80 Its effectiveness has been proven in an animal
may be influenced by the decreased availability of epi- model81 and in a small cohort of patients with ALF.82
dural catheters. However, due to its multiple systemic side effects in
patients with ALF, its routine use cannot be supported.
Prevention and Treatment of Increased THIOPENTAL SODIUM

Intracranial Pressure In a small, uncontrolled study, thiopental sodium was
Routine measures such as elevation of the head of the effective in reducing ICP.83 Unfortunately, its use is
bed to 30 degrees,55 sedation, minimal stimulation, and associated with significant hemodynamic derangements
mechanical ventilation to minimize cerebral stimulation that may necessitate escalation of vasopressor or inotropic
should be adhered to whenever possible. The manage- support. Thus thiopental use should be reserved for
ment should be focused on maintaining an adequate surges of ICH unresponsive to standard medical therapy.
CPP (> 50 mm Hg) while minimizing elevations in ICP
(< 20 mm Hg). Blood pressure should be maintained to HYPOTHERMIA
achieve a CPP between 50 and 65 mm Hg. Prolonged Moderate hypothermia (32 to 33 C) in animal models of
CPP below 50 mm Hg in the setting of ICH or an ICP ALF has been effective in improving encephalopathy
greater than 40 mm Hg is associated with poor out- and reducing brain water.84,85 Clinical studies of hypo-
come.65 thermia have also shown significant reduction in ICP.
Jalan et al were able to demonstrate that cooling patients
HYPERTONIC SALINE with refractory ICH to 32 C decreased ICP to < 20 mm
The use of hypertonic saline is thought to help restore Hg. Subsequently they demonstrated a reciprocal in-
the osmotic gradient across the astrocyte membrane. A crease in ICP with rewarming.86 Since this study, the
randomized, controlled trial recently demonstrated that same and other groups have also shown that moderate
hypothermia is effective in the prevention of ICH in oped thrombocytopenia; thus its use should be restricted
ALF, as a bridge to transplant,87 and during liver trans- in patients with DIC.102
plantation88 to prevent surges in ICP. The mechanism
of action is presumed to be multifactorial but includes
reduction of CBF, cerebrospinal fluid (CSF) ammonia, LIVER TRANSPLANTATION FOR ACUTE
and extracellular glutamate concentrations. Although LIVER FAILURE
hypothermia may be beneficial to the brain in decreasing Because the mortality of patients with ALF is almost
ICP, it also impacts coagulation and may promote universal once they meet criteria for poor prognosis
infection and cardiac rhythm disturbances. It is difficult (Table 2), they receive the utmost priority. The Status
to know, however, if hypothermia exacerbates such 1 classification has remained from the former process of
factors because these are well-known complications of organ allocation. To receive this listing, the patient must
liver failure per se. Some have raised concern that have no prior history of liver disease and fulfill criteria for
hypothermia may impair hepatic regeneration and ‘‘com- ALF with poor prognosis.
mit’’ patients to transplant.89,90 The degree to which Intensive medical supportive care is geared toward
hypothermia influences such factors needs to be ad- the avoidance of complications that could preclude
dressed in a well-designed randomized, controlled trial transplant. The decision to proceed with transplantation
before its use in routine practice can be justified. is a difficult one not only from a medical management
perspective but also due to the limited time available to
assess the patient’s social support network, mental stabil-
LIVER ASSIST DEVICES ity, compliance, and wishes. Psychiatric stability and
Liver assist devices are of two general types: biological strong social support are essential to a successful trans-
devices and artificial devices. Biological devices use living plant. It is imperative to accurately document active
hepatocytes to replace liver function, whereas artificial drug or alcohol abuse, suicidal ideation, or history of a
devices such as the molecular absorbent recirculating previous suicide attempt because these are examples of
system (MARS) aim to remove injurious substances potential contraindications to transplant.
from the circulation. Most of the trials in ALF have
used either porcine hepatocytes [bioartificial liver (BAL)
device]91 or human HepG2 cells (ELAD [extracorporeal ACUTE OR CHRONIC LIVER
liver assist device]).92 A large multicenter study of BAL FAILURE—GENERAL CONSIDERATIONS
in ALF resulted in no survival advantage.93 In a small Critically ill patients with cirrhosis admitted to the ICU
randomized controlled trial, ELAD resulted in less have poor survival even if they survive the initial ICU
severity of encephalopathy without improvement in stay. Mortality can reach 100% in cases complicated by
survival.94 Overall, the data on bioartificial liver devices septic shock.103–106 Acute hepatic decompensation in a
are disappointing in that they are quite costly and have patient with chronic liver disease (AoCLF) is most
not resulted in improved synthetic function or survival.95 commonly precipitated by gastrointestinal bleeding or
MARS removes free and albumin-bound toxins infection. As a result of this, major functions of the liver
via a polysulfone membrane.96,97 In contrast to bioartifi- such as the synthesis of key proteins, detoxification, and
cial devices, MARS has mainly been tested in acute on metabolic regulation are impaired to various degrees.
chronic liver failure (AoCLF). Two small randomized The imbalance created by the physiological needs of the
clinical trials have shown improvement in encephalop- critically ill patient and the liver’s limited ability to
athy and increased survival.98,99 MARS also appears to be perform key functions leads to life-threatening compli-
effective in ameliorating renal function in hepatorenal cations such as renal failure, infection, HE, cholestasis,
syndrome and hemodynamics through increasing sys- ascites, and bleeding.
temic vascular resistance (SVR).100 Although not a An important step in the initial management of
primary end point, survival was improved in a prospec- patients with AoCLF is the identification and treatment
tive, randomized, controlled trial of MARS in liver fail- of the precipitating factor that led to acute decompen-
ure.99 The study was terminated prematurely due to sation. Patients admitted to the ICU with AoCLF
perceived ethical considerations. Unfortunately, had as should be managed with a simultaneous multifaceted
little as one or two deaths occurred in the MARS group approach that will bridge eligible patients to transplan-
the difference in survival would have no longer been tation and improve short-term survival in nontransplant
statistically significant. The results of this trial would candidates.
have had a more profound impact if the study had
achieved its targeted enrollment. A recent meta-analysis
found that MARS offered no significant survival benefit PORTAL HYPERTENSIVE BLEEDING
over standard medical therapy.101 MARS appears to be Portal hypertensive hemorrhage is a serious and frequent
well tolerated; however, a fair number of patients devel- complication of advanced cirrhosis.107–109 Bleeding can
Table 4 Child-Pugh Score

Score Bilirubin/mmol/L Albumin/g/L INR Ascites Encephalopathy
1 <2 > 3.5 < 1.3 Absent 0

2 2.1–3 2.8–3.5 1.3–1.5 Mild I/II
3 >3 < 2.7 > 1.5 > Moderate III/IV
The Child-Pugh score is derived from the sum of the assigned points in each category. Child-Pugh A,
B, and C are defined as < 6, 7–9, and > 10, respectively. INR, international normalized ratio.
originate from gastroesophageal varices, portal hyper- namic instability is often present and needs to be
tensive gastropathy, or less commonly from portal col- addressed with continuous cardiac monitoring and
opathy, duodenal or rectal varices. A successful outcome good intravenous access. Knowledge of the central
depends on accurate diagnosis, prompt resuscitation, venous pressure is helpful to gauge the adequacy of
hemodynamic support, management of complications, volume administration without overresuscitating the
and control of active bleeding. Clinical factors associated patient because this can provoke more bleeding from
with bleeding include deterioration of liver function, previously decompressed varices or lead to other com-
portal vein thrombosis, hepatocellular carcinoma, and plications such as acute pulmonary edema.121,122 Pro-
ongoing alcohol use.110 phylactic endotracheal intubation should be considered
In cirrhosis, portal hypertension is the result of in the setting of significant upper gastrointestinal
the combined effects of increased portal venous inflow hemorrhage prior to endoscopic intervention for airway
and an increased resistance to that flow.111 The corrected protection. Figure 4 proposes an algorithm for the
sinusoidal pressure gradient is inversely associated with management of AVH.
risk of bleeding as well as outcome after bleeding.112–114
A hepatic venous pressure gradient (HVPG) of 12 mm
Hg is a threshold value, with variceal hemorrhage Esophageal Variceal Hemorrhage
occurring above this level112 and therapeutic interven- Mortality associated with AVH has improved in the
tion aiming at a reduction below this value.115,116 How- past decade.123 Based on data comparing patients
ever, many studies have shown that if portal pressure is presenting with AVH from 1981–82 to 1988–91, the
reduced in a sustained manner by 20%, the risk of late cohort experienced a significant decline in mortal-
bleeding is low (10% at 2 years), even if the HVPG ity at 30 days (20.8% vs 29.6%, p ¼ .0001) and at 6 years
remains above 12 mm Hg.109,117 (69.7% vs 74.5%, p ¼ .0001). For patients who survived
Acute variceal hemorrhage (AVH) occurs in 30 to the first 30 days, survival was slightly better in the late
35% of patients with cirrhosis and is associated with cohort on multivariate analysis (p ¼ .01).124 Spontane-
significant mortality. A recent retrospective analysis re- ous cessation of bleeding occurs in 50% of patients;
ported in-hospital, 6-week, and overall mortality rates of however, 60% of these patients will rebleed if there is no
14.2%, 17.5%, and 33.5%, respectively, suggesting im- intervention.120 Therapeutic options for the manage-
proved therapy of AVH has impacted survival.118 One- ment of variceal hemorrhage include pharmacological
year survival following a variceal bleed greatly depends on therapy (somatostatin, octreotide, vasopressin, or terli-
the severity of liver disease assessed by the Child-Pugh pressin), endoscopic therapy (endoscopic band ligation
classification (Table 4). Mortality after a variceal hemor- or sclerotherapy), balloon tamponade, transjugular in-
rhage is 5% and 50% in Child-Pugh A and C cirrhotic trahepatic portosystemic shunt (TIPS) or surgical
patients, respectively,119 and 70% of those that survive shunts (Fig. 4). Pharmacological treatment is effective
will rebleed.110,120 A hepatic venous pressure gradient and should be initiated on admission to the ICU.
> 20 mm Hg after an acute bleed is an independent Although studies have shown some drugs to be as
predictor of poor outcome. Such measurements may be effective as endoscopic therapy,125,126 this is controver-
useful to assess prognosis and tailor therapy.114 sial; thus we recommend endoscopic intervention in
The management of patients with gastroesopha- addition to pharmacotherapy for the control of variceal
geal varices involves (1) prevention of the initial bleed hemorrhage.
(primary prophylaxis is beyond the scope of this review),
(2) management of the acute bleed, and (3) prevention of
rebleeding (secondary prophylaxis). Gastric Variceal Hemorrhage
Gastric varices represent a less frequent, but important
source of bleeding in patients with portal hypertension.
Variceal Hemorrhage—General Considerations They are present in up to 57% of patients with esoph-
Acute hemorrhage typically presents with hematemesis ageal varices secondary to portal hypertension.127 Less
with or without melena or hematochezia. Hemody- commonly, they are found in the absence of esophageal
Figure 4 Proposed algorithm for the management of acute variceal bleeding.
varices.128 If isolated gastric varices are noted imaging and cardiac and pulmonary vascular glue emboli.136–138
must be performed to exclude splenic vein thrombosis Recently, pilot studies have found thrombin injections
because this is managed with splenectomy. In 117 to be another promising approach.139 Yet another way
patients with fundal varices, the size of the vessels, the to control bleeding gastric varices is with special de-
presence of red spots, and the degree of liver decom- tachable endoscopically placed snares.140,141 Although
pensation were predictors of bleeding.129 Although there is limited experience with the technique in the
bleeding from gastric varices is less common than bleed- United States, the Japanese report success with balloon-
ing from esophageal varices, bleeding episodes are often occluded retrograde transvenous obliteration (B-RTO)
more severe and carry a higher mortality.128,130–132 of isolated gastric varices.142,143 Unfortunately, most of
The management of bleeding from gastric vari- these modalities are not routinely available in the
ces differs from that of esophageal varices. Although United States.
gastric varices in continuity with esophageal varices can The first-line approach to the control of actively
be managed by endoscopic band ligation, isolated bleeding fundic varices is balloon tamponade. The air-
varices in the fundus of the stomach are not amenable way must be protected when this approach is taken. This
to either sclerotherapy or band ligation.133 However, is, however, a temporizing measure and such varices have
numerous reports have documented the efficacy of a great propensity to rebleed over time. It is therefore
cyanoacrylate injections (tissue glue) in achieving he- reasonable to decompress the portal vein with a TIPS as
mostasis in bleeding fundic varices.134,135 Cyanoacry- the definitive procedure of choice in most patients before
late glue leads to more rapid variceal obliteration and discharge from the hospital.144,145 It is often necessary to
more effective hemostasis than alcohol injection.134 embolize prominent portosystemic collaterals at the time
Severe complications associated with intravariceal cya- of TIPS to reduce the risk of bleeding. When this is not
noacrylate glue injection include mediastinitis, CVA, done, nonselective b-blockers may be used to reduce the
risks of rebleeding, although the data to support such an variceal hemorrhage; however, its use is limited due to
approach are weak at best. In those with well-preserved systemic effects such as coronary and mesenteric ische-
liver function, a surgical shunt can also provide long- mia.156 If vasopressin is used in conjunction with nitro-
term relief from bleeding.146 glycerin, these effects can be minimized.157,158
Terlipressin is a synthetic analogue of vasopressin
with a longer half-life and fewer side effects. It is
Portal Hypertensive Gastropathy effective in the treatment of acute variceal bleeding
Portal hypertensive gastropathy (PHG) is a manifes- with or without endoscopic therapy and has been shown
tation of portal hypertension. Endoscopically it is to reduce mortality.125,159 It appears to be as effective as
characterized by a mosaic mucosal pattern with varying vasopressin or endoscopic treatment126,160 in controlling
degrees of submucosal hemorrhage. It is often asymp- acute variceal hemorrhage, but is not yet available in the
tomatic but may lead to chronic transfusion–requiring United States.
blood loss or acute bleeding (rarely). Although portal
gastropathy is more often seen in the setting of gastro-
esophageal varices, its severity does not correlate with Endoscopic Therapy
portal pressure. Nevertheless, bleeding is often amelio- Endoscopic variceal band ligation (EBL) is currently the
rated by reduction of portal pressure. Both octreotide preferred endoscopic technique for the management of
and nonselective b-blockade can be helpful in decreas- esophageal varices. It is at least as effective as endoscopic
ing acute bleeding and the degree of rebleeding from sclerotherapy (EST) but has a superior safety profile and
PHG via reduction of portal blood flow.147–149 In lower complication rate.133,161,162 EBL also decreases
refractory cases, TIPS can be effective at reducing the incidence of bleeding, when given as primary pro-
transfusion requirements.150 phylaxis,163,164 and death165 from variceal bleeding.
Band ligation leads to strangulation and subsequent
obliteration of the banded varix.
THERAPIES TO ACHIEVE HEMOSTASIS EST involves the injection of a sclerosant (sodium
morrhuate, ethanolamine, or polidocanol) into or around
Drug Therapy a varix. This leads to coagulative necrosis and obliter-
Drug therapy is an integral component of the manage- ation of varices in the vicinity of the injection. Compli-
ment of acute portal hypertensive hemorrhage and cations related to EST tend to occur more frequently and
should be started on presentation. To optimize the be more severe than with EBL and include esophageal
effectiveness of drug or endoscopic therapy, clotting ulceration, stricturing, esophageal perforation, pleural
abnormalities must be corrected. Target INR and effusion, and sepsis. Despite a higher complication
platelet count should be 1.5 and 75, respectively. In rate, a role still exists for EST. It can be a useful adjunct
the setting of renal failure platelets may be dysfunc- to EBL in the setting of massive hemorrhage with poor
tional and DDAVP (Desamino-D-Arginine Vasopres- visibility because the location of injection need not be as
sin) should be considered. FFP alone or in combination precise to achieve hemostasis.
with rFVIIa can be given to rapidly correct coagulop- Regardless of the choice of endoscopic manage-
athy. A recent randomized, clinical trial showed that ment for acute bleeding, follow-up endoscopy within
when given in addition to standard therapy, 100 mg/kg 1 to 2 weeks for further banding is essential to decrease
of rFVIIa may improve control of bleeding in patients the risk of rebleeding. EBL should then be continued in
with advanced cirrhosis.151 the future until variceal obliteration is achieved.
Somatostatin or octreotide, its synthetic analogue,
stops acute bleeding from varices in 80% of cases.152 It
does so through a reduction of portal pressure via effects Balloon Occlusion
on vasoactive peptides or through the prevention of Balloon tamponade is effective in the 5 to 10% of
postprandial hyperemia (blood meal). Octreotide has patients in which hemostasis cannot be achieved acutely
an excellent safety profile and can be given initially as a with medical or endoscopic therapy. It successfully stops
subcutaneous bolus of 50 to 100 mg followed by a bleeding from esophagogastric varices through external
continuous infusion of 50 mg/h for 3 to 5 days. Although compression of varices, but rebleeding occurs in 50%.166
it does decrease portal pressure acutely, these effects Many types of tubes exist (Sengsten-Blake Tube, Warne
appear to be short lived due to rapid tachyphylaxis.153 Surgical Products, Ltd., Armagh, Ireland, UK; and
Nevertheless, studies have shown a significant decrease Linton Tube, Bard Manufacturing, Covington, Geor-
in rebleeding after endoscopic therapy in those receiving gia, USA) with mild variations; however, in most cases;
octreotide infusion.154,155 inflation of the gastric balloon is adequate to stop
Vasopressin reduces splanchnic blood flow in variceal hemorrhage. It is quite effective as a bridge to
addition to portal pressure. It is effective in controlling more definitive therapy (TIPS or surgery) and cannot be
in place for more than 24 hours given the significant risk dures (Suguira or modification) or liver transplantation.
of esophageal necrosis and rupture.167 If surgery is necessary, it is best done at a center with
extensive experience.
Transjugular Intrahepatic Portosystemic

Shunt (TIPS) COMPLICATIONS ASSOCIATED WITH
TIPS involves placing a stent within the liver that VARICEAL BLEEDING
bridges a branch of the intrahepatic hepatic vein with
an intrahepatic branch of the portal vein, allowing Hepatic Encephalopathy
diversion of blood flow away from the cirrhotic liver, The initial approach to the patient with HE should focus
decompressing portal pressures, and reducing the im- on the identification and correction of any precipitant in
petus to bleed. In experienced hands TIPS is a very addition to treatment of the encephalopathy. Common
effective method for stopping acute hemorrhage from precipitants include gastrointestinal (GI) bleeding,
esophageal varices but can be less effective in gastric medications, infection, dehydration, electrolyte distur-
variceal bleeding.168 In the setting of variceal bleeding, bances, constipation, excessive protein load, portosyste-
TIPS should be reserved for cases that are refractory to mic shunting (TIPS or spontaneous), and worsening
endoscopic therapy,169 or in the case of gastric varices, liver function (Table 5). Potential precipitants such as
used in conjunction with embolization. If recurrent these must be individually considered and subsequently
bleeding occurs in a patient with a TIPS in place, the excluded. In the case of infection, spontaneous bacterial
most important intervention is interrogation of the peritonitis (SBP) is the most common infection in this
TIPS to document and treat thrombosis or stenosis. population and must be ruled out with a diagnostic
Until recently, TIPS was complicated by fre- paracentesis as already discussed. Often, HE is the
quent stenosis and thrombosis.168 Polytetrafluoroethy- only manifestation of SBP.
lene (PTFE)-coated stents have significantly improved
stent patency and the need for reintervention.170–172 TREATMENT OF HEPATIC ENCEPHALOPATHY
Although they have not been directly compared with Nonabsorbable disaccharides and antibiotics have been
surgical shunts, these data suggest they offer comparable shown to modify gut flora and decrease blood ammonia
results. levels, but these are not necessarily related (indicating
nonbacterial sources of ammonia, which may also be
decreased by these compounds). Lactulose is the first-
Surgery line therapy for HE. It is most effective if given orally
TIPS has significantly reduced the need for shunt and titrated to a dose that achieves three to four soft
surgery; however, surgery remains a good option in bowel movements a day. A common mistake in the ICU
selected cases when TIPS is not technically feasible or is continued lactulose despite diarrhea in the encepha-
fails or in patients with significant portal hypertension in lopathic patient. Not only will this not improve ence-
the face of preserved hepatic synthetic function. The phalopathy, it may worsen it through free water
long-term patency rate of shunt surgery is thought to be depletion. If the patient is having adequate bowel move-
superior to that of TIPS; however, newer coated stents ments on lactulose, but continues to be confused, several
may challenge this assumption.170–172 Surgical alterna- agents can be added. Nonabsorbable antibiotics such as
tives for acute portal hypertensive hemorrhage include neomycin or rifaxamin are effective for the treatment of
total or selective shunt surgery, devascularization proce- HE either alone or in conjunction with lactulose. Met-
ronidazole is also efficacious; however, side effects limit
Table 5 Precipitants of Hepatic Encephalopathy prolonged use. Zinc is a cofactor for the urea cycle and
can increase the clearance of ammonia. Zinc levels are
Infection
decreased in patients with cirrhosis and HE. Supple-
Gastrointestinal bleeding
mentation with zinc sulfate 600 mg/d normalizes zinc
Medical noncompliance
levels, decreases ammonia, and improves HE.173
Medication—sedatives, narcotics, other
Branched-chain amino acids have not convincingly
Electrolyte disturbances
shown improvement in HE; however, they may be
Portosystemic shunting
considered in patients that are not receiving protein in
Transjugular intrahepatic portosystemic shunt
any form.174,175
Spontaneous
When HE is refractory to medical treatment
Dehydration
other possibilities must be entertained. In those with a
Excessive protein load
TIPS, occlusion or narrowing of the stent lumen can
Constipation
improve mental status.176,177 If no TIPS or surgical
Worsening liver function
shunt is present, abdominal imaging should be obtained
Table 6 Diagnostic Criteria for Hepatorenal Syndrome location from the gut is thought to be the most common
(International Ascites Club) mode of ascitic fluid inoculation.189,190 Predisposing
MAJOR factors for the development of SBP include advanced
Chronic or acute liver disease with advanced hepatic failure liver disease, gastrointestinal bleeding, ascitic total pro-
and portal hypetension tein fluid content 1 g/dL and a previous history of
Creatinine > 1.5 mg/dL or24-hour creatinineclearance < 40mL/min SBP.181,191–193
Absence of shock, ongoing infection, use of nephrotoxic
drugs, gastrointestinal or renal fluid losses > 500 g/d or INTERPRETATION OF A DIAGNOSTIC PERITONEAL TAP
> 1000 g/d in the setting of edema It is crucial to have a very low threshold to perform a
Urine protein < 500 mg/dL diagnostic paracentesis in the patient with suspected
No ultrasonographic evidence of primary renal disease SBP. Ideally, this should be done prior to the initia-
No sustained improvement in renal function after hydration tion of antibiotics. Peritoneal fluid should be sent
MINOR for cell count, culture, albumin, total protein, LDH
Urine sodium < 10 mEq/L (lactate dehydrogenase) and glucose. Blood culture
Urine osmolality > plasma osmolality bottles should be inoculated at the bedside to improve
Urine red blood cells < 50 per high power field yield.
Urine output < 500 mL/d SBP is defined as an ascitic fluid polymorphonu-
Serum sodium < 130 mEq/L clear (PMN) 250 cells/mm3, in the setting of a
positive monomicrobial ascitic fluid culture.194,195 Cul-
For the diagnosis of hepatorenal syndrome, all major criteria must be
met. Minor criteria are supportive but not necessary for the diagnosis. ture positivity can fluctuate from one tap to the next;
thus culture-negative neutrocytic ascites should be
to look for a prominent portosystemic collateral that treated with antibiotics as previously outlined.196
could be radiographically embolized.178 Monomicrobial non-neutrocytic bacterascites is
another common variant of SBP. It is defined as a fluid
cell count < 250 cells/ mm3 with a positive culture.197
Infection Runyon and Hoefs and Chu et al found that 62 to 86%
Bacterial infections complicate 35 to 66% of cases of of cases of monomicrobial bacterial ascites resolve spon-
gastrointestinal bleeding in patients with AoCLF.179–184 taneously, and those that did not resolve were sympto-
Not only is infection common after bleeding, it may also matic on presentation.197,198 Thus, in a clinically stable
provoke rebleeding.179 Furthermore, bacterial infection asymptomatic patient, one could observe or consider
and the use of prophylactic antibiotics are independently repeat diagnostic paracentesis. However, in a critically
associated with failure to control variceal hemorrhage in ill patient with AoCLF in the ICU, antibiotic therapy
the first 5 days of admission.180 Many randomized, may be the best course of action.
controlled trials have shown that antibiotic prophylaxis
targeted at enteric organisms (such as quinolones or TREATMENT OF SPONTANEOUS BACTERIAL PERITONITIS
third-generation cephalosporin) is effective in the pre- Patients should be given a non-nephrotoxic antibiotic
vention of postbleeding infection in cirrhotic patients. In with good enteric coverage such as a third-generation
addition, a meta-analysis of randomized, controlled cephalosporin. Cefotaxime 2 g (q8h) is the best-
trials concluded that antibiotic prophylaxis resulted in studied antibiotic for the treatment of SBP.199,200
less infections and improved short-term survival in Other antibiotics of comparable spectrum can be
bleeding patients with cirrhosis.185 Given these data, used and can be tailored if the organism is identified.
prophylactic systemic antibiotics should be given to Once the patient has been on antibiotics for 48 hours,
cirrhotic patients with gastrointestinal hemorrhage. a diagnostic tap must be repeated to assess response to
treatment. If there is not a significant decrement in the
white blood cell (WBC) count, antibiotic coverage
Spontaneous Bacterial Peritonitis should be broadened. Once treatment efficacy is estab-
SBP is a frequent and severe complication in those with lished, antibiotics should be given for 5 days.201 Intra-
cirrhosis and ascites.186 It is associated with significant venous albumin is integral to the treatment of SBP
morbidity and mortality by precipitating renal failure in and should be used in conjunction with antibiotics. It
30%,187 worsening HE, and causing hemodynamic col- has been shown in a randomized, controlled trial to
lapse in an already critically ill patient. The deterioration decrease the incidence of renal failure and subsequent
of renal function is the most sensitive predictor of in- mortality when compared with antibiotics alone.202
hospital mortality.187,188 Renal failure often results from Based on these data, albumin should be given at a
a reduction in effective circulating blood volume, cyto- dose of 1.5 mg/kg on day 1 and 1 mg/kg on day 3. A
kine surges, and activation of the renin-angiotensin recent study compared albumin to plasma expansion
system precipitated by infection.187,188 Bacterial trans- with hydroxyethyl starch. Fernandez and colleagues
found that only albumin improved hemodynamics in Table 6 illustrates the International Ascites Club classi-
patients with SBP, suggesting that it may also have fication of HRS.
direct effects on the vascular endothelium.203 Lifelong The pathophysiology of HRS is complex.
secondary antibiotic prophylaxis is mandatory in the Splanchnic arteriolar vasodilation leads to central vaso-
patient with a history of SBP. Oral quinolones are dilation and compensatory activation of systemic and
most commonly used (norfloxacin); however, many renal vasoconstrictor systems.107,211 The resultant renal
antibiotics are effective for secondary prophylaxis of vasoconstriction leads to reduced glomerular filtration
SBP. rate and increased water and sodium retention.
ASCITES Treatment of Hepatorenal Syndrome

Ascites is the result of avid water and sodium retention Liver transplantation is the ultimate treatment for
characteristic of the altered hemodynamics of cirrhosis HRS. After transplantation renal function returns to
and portal hypertension. It is associated with a 50% baseline in most cases.212,213 Combination drug ther-
2-year survival. Uncomplicated ascites is managed apy that counteracts renal and systemic vasoconstriction
with sodium restriction (< 88 mmol/d) and diuretics; leading to arterial hypotension and central hypovolemia
potassium sparing (i.e., spironolactone) alone, or in with vasoconstrictors and plasma expanders, respec-
combination with a loop diuretic (i.e., furosemide). tively, is the most effective strategy. Terlipressin, a
Diuretics are advanced until therapeutic efficacy is long-acting vasopressin analogue that stimulates
achieved or limited by worsening renal function or splanchnic V1a vasopressin receptors increases blood
hyponatremia. In diuretic-refractory or resistant cases pressure, GFR (glomerular filtration rate), and urine
repeat large-volume paracentesis (LVP) with albumin volume in patients with HRS.214,215 Unfortunately,
infusion, TIPS, or peritonovenous shunts can be terlipressin is not yet available in the United States.
effective in improving the ascites but does not improve In a small study of patients with HRS I, the combina-
survival.204,205 tion of the a-agonist midodrine (7.5 mg tid), octreotide
Ascites can be a difficult problem to manage (100 g SQ tid), and albumin (25 g/day) was effective in
in the ICU. Copious colloid and crystalloid infusion improving renal function.216 In a more recent study,
inevitably worsens ascites and diuretic use is often these findings were confirmed and insertion of a TIPS
limited by hyponatremia, hypotension, or renal failure. in a subset of patients led to further improvement in
Massive ascites can also alter respiratory mechanics and renal function.217
make breathing more labored or mechanical ventilation
more challenging. Occasionally, massive ascites can
worsen renal failure through compression of the renal LIVER TRANSPLANTATION—CHRONIC
arteries. LVP should be reserved for patient discomfort LIVER DISEASE
and improvement of respiratory mechanics when possi- Allocation of organs in chronic liver disease changed on
ble to avoid large-volume shifting and activation of February 27, 2002. The model of end-stage liver disease
vasoactive neurohumoral systems after paracentesis that (MELD) system was adopted to objectify the way in
can worsen renal perfusion. Such changes can be mini- which livers were allocated in the United States. It is a
mized with the use of albumin (6 to 8 g/L removed). survival model based on a composite of three laboratory
Albumin administration helps maintain intravascular values: serum bilirubin, serum creatinine, and INR.
volume and minimize postparacentesis circulatory dys- The model was originally used to assess short-term
function.205,206 mortality in cirrhotic patients undergoing elective
TIPS placement.218 This model was subsequently va-
lidated as an independent predictor of survival in
RENAL FAILURE patients with cirrhosis.219,220 Thus priority on the liver
Twenty percent of cirrhotic patients with tense ascites transplant waiting list is based on the patient’s blood
develop renal failure characterized by the hepatorenal group and the MELD score without emphasis on
syndrome (HRS).207,208 HRS is defined as functional waiting time.
renal impairment in a patient with advanced liver disease
in the setting of normal tubular function and renal
histology209 (Table 6). Two types of HRS have been SUMMARY
described; HRS I and HRS II, based upon the rapidity Management of the patient with acute or chronic
and extent of renal failure.210 HRS I is characterized by a hepatic failure remains a challenging problem, despite
rapid and severe deterioration of renal function with advances in intensive care. Liver failure typically has
survival measured in days to weeks, and HRS II repre- profound effects on other organ systems and the effects
sents a more indolent and stable renal dysfunction. of therapeutic interventions on other organs must be
considered. A multidisciplinary approach is most effec- 19. Lee WM. Acetaminophen and the U.S. Acute Liver Failure
tive and urgent transfer to a transplant center is man- Study Group: lowering the risks of hepatic failure. Hepatol-
datory in potential transplant candidates. Ideally, good, ogy 2004;40:6–9
20. Harrison PM, Keays R, Bray GP, Alexander GJ, Williams R.
comprehensive intensive care can support the patient
Improved outcome of paracetamol-induced fulminant hepa-
into spontaneous hepatic recovery; however, often tic failure by late administration of acetylcysteine. Lancet
the goal of therapy is to bridge patients to definitive 1990;335:1572–1573
therapy—liver transplantation. 21. Donaldson BW, Gopinath R, Wanless IR, et al. The role of
transjugular liver biopsy in fulminant liver failure: relation to
other prognostic indicators. Hepatology 1993;18:1370–1376
REFERENCES 22. Chung PY, Sitrin MD, Te HS. Serum phosphorus levels
predict clinical outcome in fulminant hepatic failure. Liver
1. Trey C, Lipworth L, Davidson CS. Halothane and liver Transpl 2003;9:248–253
damage. N Engl J Med 1969;280:562–563 23. Schmidt LE, Dalhoff K. Alpha-fetoprotein is a predictor of
2. Trey C, Davidson CS. The management of fulminant outcome in acetaminophen-induced liver injury. Hepatology
hepatic failure. Prog Liver Dis 1970;3:282–298 2005;41:26–31
3. Bernal W, Wendon J. Liver transplantation in adults with 24. Baquerizo A, Anselmo D, Shackleton C, et al. Phosphorus as
acute liver failure. J Hepatol 2004;40:192–197 an early predictive factor in patients with acute liver failure.
4. Blei AT. Medical therapy of brain edema in fulminant Transplantation 2003;75:2007–2014
hepatic failure. Hepatology 2000;32:666–669 25. Shami VM, Caldwell SH, Hespenheide EE, Arseneau KO,
5. Mas A, Rodes J. Fulminant hepatic failure. Lancet 1997;349: Bickston SJ, Macik BG. Recombinant activated factor VII
1081–1085 for coagulopathy in fulminant hepatic failure compared with
6. Larsen FS, Wendon J. Brain edema in liver failure: basic conventional therapy. Liver Transpl 2003;9:138–143
physiologic principles and management. Liver Transpl 2002; 26. Bernal W. Intensive care support therapy. Liver Transpl
8:983–989 2003;9(suppl 9):S15–S17
7. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early 27. Vaquero J, Chung C, Cahill ME, Blei AT. Pathogenesis of
indicators of prognosis in fulminant hepatic failure. Gastro- hepatic encephalopathy in acute liver failure. Semin Liver
enterology 1989;97:439–446 Dis 2003;23:259–269
8. Izumi S, Langley PG, Wendon J, et al. Coagulation factor V 28. Vaquero J, Blei AT. Etiology and management of fulminant
levels as a prognostic indicator in fulminant hepatic failure. hepatic failure. Curr Gastroenterol Rep 2003;5:39–47
Hepatology 1996;23:1507–1511 29. Wijdicks EF, Nyberg SL. Propofol to control intracranial
9. Pereira LM, Langley PG, Hayllar KM, Tredger JM, pressure in fulminant hepatic failure. Transplant Proc 2002;
Williams R. Coagulation factor V and VIII/V ratio as 34:1220–1222
predictors of outcome in paracetamol induced fulminant 30. Steinberg KP. Stress-related mucosal disease in the critically
hepatic failure: relation to other prognostic indicators. Gut ill patient: risk factors and strategies to prevent stress-related
1992;33:98–102 bleeding in the intensive care unit. Crit Care Med 2002;
10. Lee WM. Acute liver failure in the United States. Semin 30(Suppl 6):S362–S364
Liver Dis 2003;23:217–226 31. Jalan R. Intracranial hypertension in acute liver failure:
11. Seeff LB, Cuccherini BA, Zimmerman HJ, Adler E, Benjamin pathophysiological basis of rational management. Semin
SB. Acetaminophen hepatotoxicity in alcoholics: a therapeutic Liver Dis 2003;23:271–282
misadventure. Ann Intern Med 1986;104:399–404 32. Ellis A, Wendon J. Circulatory, respiratory, cerebral, and
12. Ostapowicz G, Fontana RJ, Schiodt FV, et al. Results of a renal derangements in acute liver failure: pathophysiology
prospective study of acute liver failure at 17 tertiary care cen- and management. Semin Liver Dis 1996;16:379–388
ters in the United States. Ann Intern Med 2002;137:947– 33. Trewby PN, Williams R. Pathophysiology of hypotension in
954 patients with fulminant hepatic failure. Gut 1977;18:1021–
13. Makin AJ, Wendon J, Williams R. A 7-year experience of 1026
severe acetaminophen-induced hepatotoxicity (1987–1993). 34. Rolando N, Wade J, Davalos M, Wendon J, Philpott-
Gastroenterology 1995;109:1907–1916 Howard J, Williams R. The systemic inflammatory response
14. Zimmerman HJ, Maddrey WC. Acetaminophen (paraceta- syndrome in acute liver failure. Hepatology 2000;32(4 Pt 1):
mol) hepatotoxicity with regular intake of alcohol: analysis of 734–739
instances of therapeutic misadventure. Hepatology 1995;22: 35. Bouachour G, Tirot P, Varache N, Gouello JP, Harry P,
767–773 Alquier P. Hemodynamic changes in acute adrenal insuffi-
15. Schiodt FV, Rochling FA, Casey DL, Lee WM. Acetami- ciency. Intensive Care Med 1994;20:138–141
nophen toxicity in an urban county hospital. N Engl J Med 36. Bollaert PE, Charpentier C, Levy B, Debouverie M,
1997;337:1112–1117 Audibert G, Larcan A. Reversal of late septic shock with
16. McClain CJ, Kromhout JP, Peterson FJ, Holtzman JL. supraphysiologic doses of hydrocortisone. Crit Care Med
Potentiation of acetaminophen hepatotoxicity by alcohol. 1998;26:645–650
JAMA 1980;244:251–253 37. McCashland TM, Shaw BW Jr, Tape E. The American
17. Murphy R, Swartz R, Watkins PB. Severe acetaminophen experience with transplantation for acute liver failure. Semin
toxicity in a patient receiving isoniazid. Ann Intern Med Liver Dis 1996;16:427–433
1990;113:799–800 38. Davies MH, Mutimer D, Lowes J, Elias E, Neuberger J.
18. Prescott LF, Critchley JA. The treatment of acetaminophen Recovery despite impaired cerebral perfusion in fulminant
poisoning. Annu Rev Pharmacol Toxicol 1983;23:87–101 hepatic failure. Lancet 1994;343:1329–1330
39. Wyke RJ, Yousif-Kadaru AG, Rajkovic IA, Eddleston AL, 57. Davenport A, Will EJ, Davison AM. Continuous vs.
Williams R. Serum stimulatory activity and polymorpho- intermittent forms of haemofiltration and/or dialysis in the
nuclear leucocyte movement in patients with fulminant management of acute renal failure in patients with defective
hepatic failure. Clin Exp Immunol 1982;50:442–449 cerebral autoregulation at risk of cerebral oedema. Contrib
40. Wyke RJ, Rajkovic IA, Eddleston AL, Williams R. Nephrol 1991;93:225–233
Defective opsonisation and complement deficiency in serum 58. O’Grady JG. Paracetamol hepatotoxicity: how to prevent.
from patients with fulminant hepatic failure. Gut 1980;21: J R Soc Med 1997;90:368–370
643–649 59. Blei AT. Hypothermia for hypertension. Lancet 1999;
41. Altin M, Hughes RD, Williams R. Neutrophil adherence (9195):2082
during hemoperfusion in fulminant hepatic failure. Int J 60. Clemmesen JO, Larsen FS, Kondrup J, Hansen BA, Ott P.
Artif Organs 1982;5:315–317 Cerebral herniation in patients with acute liver failure is
42. Clapperton M, Rolando N, Sandoval L, Davies E, Williams correlated with arterial ammonia concentration. Hepatology
R. Neutrophil superoxide and hydrogen peroxide production 1999;29:648–653
in patients with acute liver failure. Eur J Clin Invest 1997; 61. Strauss GI, Knudsen GM, Kondrup J, Moller K, Larsen FS.
27:164–168 Cerebral metabolism of ammonia and amino acids in patients
43. Rolando N, Harvey F, Brahm J, et al. Prospective study of with fulminant hepatic failure. Gastroenterology 2001;121:
bacterial infection in acute liver failure: an analysis of fifty 1109–1119
patients. Hepatology 1990;11:49–53 62. Strauss GI, Christiansen M, Moller K, Clemmesen JO,
44. Rolando N, Philpott-Howard J, Williams R. Bacterial and Larsen FS, Knudsen GM. S-100b and neuron-specific
fungal infection in acute liver failure. Semin Liver Dis 1996; enolase in patients with fulminant hepatic failure. Liver
16:389–402 Transpl 2001;7:964–970
45. Wade J, Rolando N, Philpott-Howard J, Wendon J. Timing 63. Keays RT, Alexander GJ, Williams R. The safety and value
and aetiology of bacterial infections in a liver intensive care of extradural intracranial pressure monitors in fulminant
unit. J Hosp Infect 2003;53:144–146 hepatic failure. J Hepatol 1993;18:205–209
46. Rolando N, Gimson A, Wade J, Philpott-Howard J, 64. Lidofsky SD, Bass NM, Prager MC, et al. Intracranial
Casewell M, Williams R. Prospective controlled trial of pressure monitoring and liver transplantation for fulminant
selective parenteral and enteral antimicrobial regimen in hepatic failure. Hepatology 1992;16:1–7
fulminant liver failure. Hepatology 1993;17:196–201 65. Donovan JP, Shaw BW Jr, Langnas AN, Sorrell MF. Brain
47. Salmeron JM, Tito L, Rimola A, et al. Selective intestinal water and acute liver failure: the emerging role of intracranial
decontamination in the prevention of bacterial infection in pressure monitoring. Hepatology 1992;16:267–268
patients with acute liver failure. J Hepatol 1992;14:280–285 66. Larsen FS, Knudsen GM, Hansen BA. Pathophysiological
48. Rolando N, Wade JJ, Stangou A, et al. Prospective study changes in cerebral circulation, oxidative metabolism and
comparing the efficacy of prophylactic parenteral antimicro- blood-brain barrier in patients with acute liver failure: tailored
bials, with or without enteral decontamination, in patients cerebral oxygen utilization. J Hepatol 1997;27: 231–238
with acute liver failure. Liver Transpl Surg 1996;2:8–13 67. Toftengi F, Larsen FS. Management of patients with
49. Vaquero J, Polson J, Chung C, et al. Infection and the fulminant hepatic failure and brain edema. Metab Brain
progression of hepatic encephalopathy in acute liver failure. Dis 2004;19:207–214
Gastroenterology 2003;125:755–764 68. Blei AT, Olafsson S, Webster S, Levy R. Complications of
50. Billiau A, Vandekerckhove F. Cytokines and their interac- intracranial pressure monitoring in fulminant hepatic failure.
tions with other inflammatory mediators in the pathogenesis Lancet 1993;341:157–158
of sepsis and septic shock. Eur J Clin Invest 1991;21:559–573 69. Vaquero J, Blei A, Fontana RJ, et al. Complications and use
51. Jalan R, Pollok A, Shah SH, Madhavan K, Simpson KJ. of intracranial pressure monitoring in patients with acute
Liver derived pro-inflammatory cytokines may be important liver failure and severe encephalopathy. Liver Transpl 2005;
in producing intracranial hypertension in acute liver failure. J 12:1581–1589
Hepatol 2002;37:536–538 70. Cordoba J, Blei AT. Cerebral edema and intracranial
52. Bihari DJ, Gimson AE, Williams R. Cardiovascular, pressure monitoring. Liver Transpl Surg 1995;1:187–194
pulmonary and renal complications of fulminant hepatic 71. Caldwell SH, Chang C, Macik BG. Recombinant activated
failure. Semin Liver Dis 1986;6:119–128 factor VII (rFVIIa) as a hemostatic agent in liver disease:
53. Briegel J, Forst H, Haller M, et al. Stress doses of a break from convention in need of controlled trials.
hydrocortisone reverse hyperdynamic septic shock: a pro- Hepatology 2004;39:592–598
spective, randomized, double-blind, single-center study. Crit 72. Murphy N, Auzinger G, Bernel W, Wendon J. The effect of
Care Med 1999;27:723–732 hypertonic sodium chloride on intracranial pressure in
54. Harry R, Auzinger G, Wendon J. The clinical importance of patients with acute liver failure. Hepatology 2004;39:464–
adrenal insufficiency in acute hepatic dysfunction. Hepatol- 470
ogy 2002;36:395–402 73. Clemmesen JO, Kondrup J, Nielsen LB, Larsen FS, Ott P.
55. Davenport A, Will EJ, Davison AM, et al. Changes in Effects of high-volume plasmapheresis on ammonia, urea,
intracranial pressure during haemofiltration in oliguric and amino acids in patients with acute liver failure. Am J
patients with grade IV hepatic encephalopathy. Nephron Gastroenterol 2001;96:1217–1223
1989;53:142–146 74. Awad SS, Swaniker F, Magee J, Punch J, Bartlett RH.
56. Davenport A, Will EJ, Davidson AM. Improved cardiovas- Results of a phase I trial evaluating a liver support device
cular stability during continuous modes of renal replacement utilizing albumin dialysis. Surgery 2001;130:354–362
therapy in critically ill patients with acute hepatic and renal 75. Rose C, Michalak A, Rao KV, Quack G, Kircheis G,
failure. Crit Care Med 1993;21:328–338 Butterworth RF. L-ornithine-L-aspartate lowers plasma and
cerebrospinal fluid ammonia and prevents brain edema in rats module inoculated with liver cells. Hepatology 1993;17:258–
with acute liver failure. Hepatology 1999;30:636–640 265
76. Canalese J, Gimson AE, Davis C, Mellon PJ, Davis M, 92. Sussman NL, Kelly JH. Extracorporeal liver support: cell-
Williams R. Controlled trial of dexamethasone and mannitol based therapy for the failing liver. Am J Kidney Dis
for the cerebral oedema of fulminant hepatic failure. Gut 1997;30(5, Suppl 4):S66–S71
1982;23:625–629 93. Demetriou AA, Brown RS Jr, Busuttil RW, et al. Prospec-
77. Ede RJ, Gimson AE, Bihari D, Williams R. Controlled tive, randomized, multicenter, controlled trial of a bio-
hyperventilation in the prevention of cerebral oedema in artificial liver in treating acute liver failure. Ann Surg
fulminant hepatic failure. J Hepatol 1986;2:43–51 2004;239:660–667 discussion 667–670
78. Strauss GI, Moller K, Holm S, Sperling B, Knudsen GM, 94. Ellis AJ, Hughes RD, Wendon JA, et al. Pilot-controlled
Larsen FS. Transcranial Doppler sonography and internal trial of the extracorporeal liver assist device in acute liver
jugular bulb saturation during hyperventilation in patients failure. Hepatology 1996;24:1446–1451
with fulminant hepatic failure. Liver Transpl 2001;7:352– 95. Jalan R, Sen S, Williams R. Prospects for extracorporeal liver
358 support. Gut 2004;53:890–898
79. Strauss G, Hansen BA, Knudsen GM, Larsen FS. 96. Stange J, Mitzner S, Ramlow W, Gliesche T, Hickstein H,
Hyperventilation restores cerebral blood flow autoregulation Schmidt R. A new procedure for the removal of protein
in patients with acute liver failure. J Hepatol 1998;28:199– bound drugs and toxins. ASAIO J 1993;39:M621–M625
203 97. Stange J, Ramlow W, Mitzner S, Schmidt R, Klinkmann H.
80. Jensen K, Ohrstrom J, Cold GE, Astrup J. Indomethacin Dialysis against a recycled albumin solution enables the
(Confortid) in severe head injury and elevated intracranial removal of albumin-bound toxins. Artif Organs 1993;17:
pressure (ICP). Acta Neurochir Suppl (Wien) 1992;55:47– 809–813
48 98. Stange J, Mitzner SR, Klammt S, et al. Liver support by
81. Chung C, Gottstein J, Blei AT. Indomethacin prevents the extracorporeal blood purification: a clinical observation.
development of experimental ammonia-induced brain edema Liver Transpl 2000;6:603–613
in rats after portacaval anastomosis. Hepatology 2001;34: 99. Lutkes P, Lutz J, Loock J, et al. Continuous venovenous
249–254 hemodialysis treatment in critically ill patients after
82. Tofteng F, Larsen FS. The effect of indomethacin on liver transplantation. Kidney Int Suppl 1999;(72):S71–S74
intracranial pressure, cerebral perfusion and extracellular 100. Mitzner SR, Stange J, Klammt S, et al. Improvement of
lactate and glutamate concentrations in patients with hepatorenal syndrome with extracorporeal albumin dialysis
fulminant hepatic failure. J Cereb Blood Flow Metab 2004; MARS: results of a prospective, randomized, controlled
24:798–804 clinical trial. Liver Transpl 2000;6:277–286
83. Forbes A, Alexander GJ, O’Grady JG, et al. Thiopental 101. Khuroo MS, Farahat KL. Molecular adsorbent recirculating
infusion in the treatment of intracranial hypertension system for acute and acute-on-chronic liver failure: a meta-
complicating fulminant hepatic failure. Hepatology 1989; analysis. Liver Transpl 2004;10:1099–1106
10:306–310 102. Sen SSC, Williams R. Artificial liver support: overview of
84. Peignoux M, Bernuau J, Benhamou JP. Total hepatectomy registry and controlled trials. In: Arroyo VFX, Garcia-Pagan
and hepatic vascular exclusion in the rat: a comparison, with JC, eds. Progress in the Treatment of Liver Diseases.
special reference to the influence of body temperature. Clin Barcelona: Ars Medica; 2003:429–435
Sci (Lond) 1982;62:273–277 103. Singh N, Gayowski T, Wagener MM, Marino IR. Outcome
85. Eguchi S, Kamlot A, Ljubimova J, et al. Fulminant hepatic of patients with cirrhosis requiring intensive care unit
failure in rats: survival and effect on blood chemistry and liver support: prospective assessment of predictors of mortality.
regeneration. Hepatology 1996;24:1452–1459 J Gastroenterol 1998;33:73–79
86. Jalan R, Damink SW, Deutz NE, Lee A, Hayes PC. 104. Kress JP, Rubin A, Pohlman AS, Hall JB. Outcomes of
Moderate hypothermia for uncontrolled intracranial hyper- critically ill patients denied consideration for liver transplan-
tension in acute liver failure. Lancet 1999;354:1164–1168 tation. Am J Respir Crit Care Med 2000;162(2 Pt 1):418–
87. Jalan R, Olde Damink SW, Deutz NE, et al. Moderate 423
hypothermia prevents cerebral hyperemia and increase in 105. Zimmerman JE, Wagner DP, Seneff MG, Becker RB, Sun
intracranial pressure in patients undergoing liver transplan- X, Knaus WA. Intensive care unit admissions with cirrhosis:
tation for acute liver failure. Transplantation 2003;75:2034– risk-stratifying patient groups and predicting individual
2039 survival. Hepatology 1996;23:1393–1401
88. Jalan R, Olde Damink SW, Deutz NE, Hayes PC, Lee A. 106. Aggarwal A, Ong JP, Younossi ZM, Nelson DR, Hoffman-
Moderate hypothermia in patients with acute liver failure Hogg L, Arroliga AC. Predictors of mortality and resource
and uncontrolled intracranial hypertension. Gastroenterol- utilization in cirrhotic patients admitted to the medical ICU.
ogy 2004;127:1338–1346 Chest 2001;119:1489–1497
89. Bernuau J. Acute liver failure: avoidance of deleterious 107. Schrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen
cofactors and early specific medical therapy for the liver are JH, Rodes J. Peripheral arterial vasodilation hypothesis: a
better than late intensive care for the brain. J Hepatol proposal for the initiation of renal sodium and water
2004;41:152–155 retention in cirrhosis. Hepatology 1988;8:1151–1157
90. Munoz SJ. Hypothermia may impair hepatic regeneration 108. Gores GJ, Wiesner RH, Dickson ER, Zinsmeister AR,
in acute liver failure. Gastroenterology 2005;128:1143–1144 Jorgensen RA, Langworthy A. Prospective evaluation of
author reply 1144–1145 esophageal varices in primary biliary cirrhosis: development,
91. Rozga J, Williams F, Ro MS, et al. Development of a natural history, and influence on survival. Gastroenterology
bioartificial liver: properties and function of a hollow-fiber 1989;96:1552–1559
109. Groszmann RJ, Bosch J, Grace ND, et al. Hemodynamic 127. Watanabe K, Kimura K, Matsutani S, Ohto M, Okuda K.
events in a prospective randomized trial of propranolol versus Portal hemodynamics in patients with gastric varices: a study in
placebo in the prevention of a first variceal hemorrhage. 230 patients with esophageal and/or gastric varices using portal
Gastroenterology 1990;99:1401–1407 vein catheterization. Gastroenterology 1988;95:434–440
110. The North Italian Endoscopic Club for the Study and 128. Sarin SK, Lahoti D, Saxena SP, Murthy NS, Makwana UK.
Treatment of Esophageal Varices. Prediction of the first Prevalence, classification and natural history of gastric
variceal hemorrhage in patients with cirrhosis of the liver and varices: a long-term follow-up study in 568 portal hyperten-
esophageal varices: a prospective multicenter study. N Engl J sion patients. Hepatology 1992;16:1343–1349
Med 1988;319:983–989 129. Kim T, Shijo H, Kokawa H, et al. Risk factors for hem-
111. Mahl TC, Groszmann RJ. Pathophysiology of portal orrhage from gastric fundal varices. Hepatology 1997;25:307–
hypertension and variceal bleeding. Surg Clin North Am 312
1990;70:251–266 130. Ohnishi K, Sato S, Saito M, et al. Clinical and portal
112. Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn HO, hemodynamic features in cirrhotic patients having a large
Atterbury CE, Glickman M. Portal pressure, presence of spontaneous splenorenal and/or gastrorenal shunt. Am J
gastroesophageal varices and variceal bleeding. Hepatology Gastroenterol 1986;81:450–455
1985;5:419–424 131. Rinella ME, Shah D, Vogelzang RL, Blei AT, Flamm SL.
113. Vorobioff J, Groszmann RJ, Picabea E, et al. Prognostic Fundal variceal bleeding after correction of portal hyperten-
value of hepatic venous pressure gradient measurements in sion in patients with cirrhosis. Gastrointest Endosc 2003;58:
alcoholic cirrhosis: a 10-year prospective study. Gastroenter- 122–127
ology 1996;111:701–709 132. Thakeb F, Salem SA, Abdallah M, el Batanouny M.
114. Moitinho E, Escorsell A, Bandi JC, et al. Prognostic value of Endoscopic diagnosis of gastric varices. Endoscopy 1994;
early measurements of portal pressure in acute variceal 26:287–291
bleeding. Gastroenterology 1999;117:626–631 133. Sarin SK, Govil A, Jain AK, et al. Prospective randomized
115. Grace ND, Groszmann RJ, Garcia-Tsao G, et al. Portal trial of endoscopic sclerotherapy versus variceal band ligation
hypertension and variceal bleeding: an AASLD single topic for esophageal varices: influence on gastropathy, gastric
symposium. Hepatology 1998;28:868–880 varices and variceal recurrence. J Hepatol 1997;26:826–832
116. Rossle M. Prevention of rebleeding from oesophageal-gastric 134. Sarin SK, Jain AK, Jain M, Gupta R. A randomized
varices. Eur J Gastroenterol Hepatol 2001;13:343–348 controlled trial of cyanoacrylate versus alcohol injection in
117. Escorsell A, Bordas JM, Castaneda B, et al. Predictive value patients with isolated fundic varices. Am J Gastroenterol
of the variceal pressure response to continued pharmacolo- 2002;97:1010–1015
gical therapy in patients with cirrhosis and portal hyperten- 135. Lo GH, Lai KH, Cheng JS, Chen MH, Chiang HT. A
sion. Hepatology 2000;31:1061–1067 prospective, randomized trial of butyl cyanoacrylate injection
118. Chalasani N, Kahi C, Francois F, et al. Improved patient versus band ligation in the management of bleeding gastric
survival after acute variceal bleeding: a multicenter, cohort varices. Hepatology 2001;33:1060–1064
study. Am J Gastroenterol 2003;98:653–659 136. Gallet B, Zemour G, Saudemont JP, Renard P, Hillion ML,
119. Benoit JN, Womack WA, Korthuis RJ, Wilborn WH, Hiltgen M. Echocardiographic demonstration of intracar-
Granger DN. Chronic portal hypertension: effects on diac glue after endoscopic obturation of gastroesophageal
gastrointestinal blood flow distribution. Am J Physiol 1986; varices. J Am Soc Echocardiogr 1995;8(5 Pt 1):759–761
250(4 Pt 1):G535–G539 137. Rickman OB, Utz JP, Aughenbaugh GL, Gostout CJ.
120. Graham DY, Smith JL. The course of patients after variceal Pulmonary embolization of 2-octyl cyanoacrylate after
hemorrhage. Gastroenterology 1981;80:800–809 endoscopic injection therapy for gastric variceal bleeding.
121. Kravetz D, Bosch J, Arderiu M, Pilar Pizcueta M, Rodes J. Mayo Clin Proc 2004;79:1455–1458
Hemodynamic effects of blood volume restitution following 138. See A, Florent C, Lamy P, Levy VG, Bouvry M. Cerebro-
a hemorrhage in rats with portal hypertension due to vascular accidents after endoscopic obturation of esophageal
cirrhosis of the liver: influence of the extent of portal- varices with isobutyl-2-cyanoacrylate in 2 patients. Gastro-
systemic shunting. Hepatology 1989;9:808–814 enterol Clin Biol 1986;10:604–607
122. Kruskall MS, Mintz PD, Bergin JJ, et al. Transfusion 139. Yang WL, Tripathi D, Therapondos G, Todd A, Hayes PC.
therapy in emergency medicine. Ann Emerg Med 1988;17: Endoscopic use of human thrombin in bleeding gastric
327–335 varices. Am J Gastroenterol 2002;97:1381–1385
123. D’Amico G, De Franchis R. Upper digestive bleeding in 140. Yoshida T, Harada T, Shigemitsu T, Takeo Y, Miyazaki S,
cirrhosis: posttherapeutic outcome and prognostic indicators. Okita K. Endoscopic management of gastric varices using a
Hepatology 2003;38:599–612 detachable snare and simultaneous endoscopic sclerotherapy
124. El-Serag HB, Everhart JE. Improved survival after variceal and O-ring ligation. J Gastroenterol Hepatol 1999;14:730–
hemorrhage over an 11-year period in the Department 735
of Veterans Affairs. Am J Gastroenterol 2000;95:3566– 141. Yoshida T, Hayashi N, Suzumi N, et al. Endoscopic ligation
3573 of gastric varices using a detachable snare. Endoscopy
125. Ioannou GN, Doust J, Rockey DC. Systematic review: 1994;26:502–505
terlipressin in acute oesophageal variceal haemorrhage. 142. Kanagawa H, Mima S, Kouyama H, Gotoh K, Uchida T,
Aliment Pharmacol Ther 2003;17:53–64 Okuda K. Treatment of gastric fundal varices by balloon-
126. Escorsell A, Ruiz del Arbol L, Planas R, et al. Multicenter occluded retrograde transvenous obliteration. J Gastroenterol
randomized controlled trial of terlipressin versus sclerother- Hepatol 1996;11:51–58
apy in the treatment of acute variceal bleeding: the TEST 143. Fukuda T, Hirota S, Sugimura K. Long-term results of
study. Hepatology 2000;32:471–476 balloon-occluded retrograde transvenous obliteration for the
treatment of gastric varices and hepatic encephalopathy. hemostatic balloon in patients with cirrhosis: a randomized
J Vasc Interv Radiol 2001;12:327–336 controlled trial. Arch Med Res 1997;28:241–245
144. Sanyal AJ. The use and misuse of transjugular intrahepatic 161. Laine L, Cook D. Endoscopic ligation compared with
portasystemic shunts. Curr Gastroenterol Rep 2000;2:61–71 sclerotherapy for treatment of esophageal variceal bleeding:
145. Chau TN, Patch D, Chan YW, Nagral A, Dick R, a meta-analysis. Ann Intern Med 1995;123:280–287
Burroughs AK. ‘‘Salvage’’ transjugular intrahepatic portosys- 162. Stiegmann GV, Goff JS, Michaletz-Onody PA, et al.
temic shunts: gastric fundal compared with esophageal Endoscopic sclerotherapy as compared with endoscopic
variceal bleeding. Gastroenterology 1998;114:981–987 ligation for bleeding esophageal varices. N Engl J Med
146. Henderson JM. Role of distal splenorenal shunt for long- 1992;326:1527–1532
term management of variceal bleeding. World J Surg 1994; 163. Sarin SK, Lamba GS, Kumar M, Misra A, Murthy NS.
18:205–210 Comparison of endoscopic ligation and propranolol for the
147. Merli M, Salerno F, Riggio O, et al. Transjugular intrahepatic primary prevention of variceal bleeding. N Engl J Med
portosystemic shunt versus endoscopic sclerotherapy for the 1999;340:988–993
prevention of variceal bleeding in cirrhosis: a randomized 164. Sarin SK, Guptan RK, Jain AK, Sundaram KR. A
multicenter trial. Gruppo Italiano Studio Trends Pharmacol randomized controlled trial of endoscopic variceal band
Sci (G.I.S.T.). Hepatology 1998;27:48–53 ligation for primary prophylaxis of variceal bleeding. Eur J
148. Zhou JH, Liu CY, Zhang RH, Wang HR, Liu KJ. Effects of Gastroenterol Hepatol 1996;8:337–342
octreotide on gallbladder pressure and myoelectric activity of 165. Lay CS, Tsai YT, Teg CY, et al. Endoscopic variceal ligation
Oddi sphincter in rabbits. World J Gastroenterol 1998;4: in prophylaxis of first variceal bleeding in cirrhotic patients
238–241 with high-risk esophageal varices. Hepatology 1997;25:
149. Garcia N, Sanyal AJ. Portal hypertensive gastropathy and 1346–1350
gastric antral vascular ectasia. Curr Treat Options Gastro- 166. Harry R, Wendon J. Management of variceal bleeding. Curr
enterol 2001;4:163–171 Opin Crit Care 2002;8:164–170
150. Kamath PS, Lacerda M, Ahlquist DA, McKusick MA, 167. Chojkier M, Conn HO. Esophageal tamponade in the
Andrews JC, Nagorney DA. Gastric mucosal responses to treatment of bleeding varices: a decadel progress report. Dig
intrahepatic portosystemic shunting in patients with cirrho- Dis Sci 1980;25:267–272
sis. Gastroenterology 2000;118:905–911 168. Sanyal AJ, Freedman AM, Luketic VA, et al. The natural
151. Bosch J, Thabut D, Bendtsen F, et al. Recombinant factor history of portal hypertension after transjugular intrahepa-
VIIa for upper gastrointestinal bleeding in patients with tic portosystemic shunts. Gastroenterology 1997;112:889–
cirrhosis: a randomized, double-blind trial. Gastroenterology 898
2004;127:1123–1130 169. Sanyal AJ, Freedman AM, Luketic VA, et al. Transjugular
152. Jenkins SA, Shields R, Davies M, et al. A multicentre intrahepatic portosystemic shunts for patients with active
randomised trial comparing octreotide and injection scler- variceal hemorrhage unresponsive to sclerotherapy. Gastro-
otherapy in the management and outcome of acute variceal enterology 1996;111:138–146
haemorrhage. Gut 1997;41:526–533 170. Bureau C, Garcia-Pagan JC, Otal P, et al. Improved clinical
153. Escorsell A, Bandi JC, Andreu V, et al. Desensitization to outcome using polytetrafluoroethylene-coated stents for
the effects of intravenous octreotide in cirrhotic patients with Trends Pharmacol Sci: results of a randomized study.
portal hypertension. Gastroenterology 2001;120:161–169 Gastroenterology 2004;126:469–475
154. Besson I, Ingrand P, Person B, et al. Sclerotherapy with or 171. Rossi P, Salvatori FM, Fanelli F, et al. Polytetrafluoroethy-
without octreotide for acute variceal bleeding. N Engl J Med lene-covered nitinol stent-graft for transjugular intrahepatic
1995;333:555–560 portosystemic shunt creation: 3-year experience. Radiology
155. Sung JJ, Chung SC, Yung MY, et al. Prospective randomised 2004;231:820–830
study of effect of octreotide on rebleeding from oesophageal 172. Ockenga J, Kroencke TJ, Schuetz T, et al. Covered
varices after endoscopic ligation. Lancet 1995;346:1666– transjugular intrahepatic portosystemic stents maintain lower
1669 portal pressure and require fewer reinterventions than
156. Conn HO, Ramsby GR, Storer EH, et al. Intraarterial uncovered stents. Scand J Gastroenterol 2004;39:994–999
vasopressin in the treatment of upper gastrointestinal 173. Marchesini G, Fabbri A, Bianchi G, Brizi M, Zoli M. Zinc
hemorrhage: a prospective, controlled clinical trial. Gastro- supplementation and amino acid-nitrogen metabolism in
enterology 1975;68:211–221 patients with advanced cirrhosis. Hepatology 1996;23:1084–
157. Bosch J, Groszmann RJ, Garcia-Pagan JC, et al. Association 1092
of transdermal nitroglycerin to vasopressin infusion in the 174. Kjaergard LL, Liu J, Als-Nielsen B, Gluud C. Artificial and
treatment of variceal hemorrhage: a placebo-controlled bioartificial support systems for acute and acute-on-chronic
clinical trial. Hepatology 1989;10:962–968 liver failure: a systematic review. JAMA 2003;289:217–222
158. Gimson AE, Westaby D, Hegarty J, Watson A, Williams R. 175. Blei AT, Cordoba J. Hepatic encephalopathy. Am J
A randomized trial of vasopressin and vasopressin plus Gastroenterol 2001;96:1968–1976
nitroglycerin in the control of acute variceal hemorrhage. 176. Madoff DC, Perez-Young IV, Wallace MJ, Skolkin MD,
Hepatology 1986;6:410–413 Toombs BD. Management of Trends Pharmacol Sci–related
159. Soderlund C, Magnusson I, Torngren S, Lundell L. refractory hepatic encephalopathy with reduced Wallgraft
Terlipressin (triglycyl-lysine vasopressin) controls acute bleed- endoprostheses. J Vasc Interv Radiol 2003;14:369–374
ing oesophageal varices: a double-blind, randomized, placebo- 177. Kerlan RK Jr, LaBerge JM, Baker EL, et al. Successful
controlled trial. Scand J Gastroenterol 1990;25:622–630 reversal of hepatic encephalopathy with intentional occlusion
160. Garcia-Compean D, Blanc P, Bories JM, et al. Treatment of of transjugular intrahepatic portosystemic shunts. J Vasc
active gastroesophageal variceal bleeding with terlipressin or Interv Radiol 1995;6:917–921
178. Shioyama Y, Matsueda K, Horihata K, et al. Post-Trends common but rarely recognized syndrome. Ann Intern Med
Pharmacol Sci hepatic encephalopathy treated by occlusion 1964;60:568–580
balloon-assisted retrograde embolization of a coexisting 195. Guarner C, Soriano G. Spontaneous bacterial peritonitis.
spontaneous splenorenal shunt. Cardiovasc Intervent Radiol Semin Liver Dis 1997;17:203–217
1996;19:53–55 196. McHutchison J, Runyon BA. Spontaneous bacterial perito-
179. Bernard B, Cadranel JF, Valla D, Escolano S, Jarlier V, nitis. In: Surawicz CM, Owen R, eds. Gastrointestinal and
Opolon P. Prognostic significance of bacterial infection in Hepatic Infections. Philadelphia, PA: WB Saunders; 1994:
bleeding cirrhotic patients: a prospective study. Gastroenter- 455–475
ology 1995;108:1828–1834 197. Runyon BA, Hoefs JC. Culture-negative neutrocytic ascites:
180. Goulis J, Armonis A, Patch D, Sabin C, Greenslade L, a variant of spontaneous bacterial peritonitis. Hepatology
Burroughs AK. Bacterial infection is independently associated 1984;4:1209–1211
with failure to control bleeding in cirrhotic patients with 198. Chu CM, Chang KY, Liaw YF. Prevalence and prognostic
gastrointestinal hemorrhage. Hepatology 1998;27:1207–1212 significance of bacterascites in cirrhosis with ascites. Dig Dis
181. Soriano G, Guarner C, Tomas A, et al. Norfloxacin prevents Sci 1995;40:561–565
bacterial infection in cirrhotics with gastrointestinal hemor- 199. Cabrera J, Arroyo V, Ballesta AM, et al. Aminoglycoside
rhage. Gastroenterology 1992;103:1267–1272 nephrotoxicity in cirrhosis: value of urinary beta 2-
182. Blaise M, Pateron D, Trinchet JC, Levacher S, Beaugrand microglobulin to discriminate functional renal failure
M, Pourriat JL. Systemic antibiotic therapy prevents from acute tubular damage. Gastroenterology 1982;82:
bacterial infection in cirrhotic patients with gastrointestinal 97–105
hemorrhage. Hepatology 1994;20(1 Pt 1):34–38 200. Felisart J, Rimola A, Arroyo V, et al. Cefotaxime is more
183. Hsieh WJ, Lin HC, Hwang SJ, et al. The effect of effective than is ampicillin-tobramycin in cirrhotics with
ciprofloxacin in the prevention of bacterial infection in severe infections. Hepatology 1985;5:457–462
patients with cirrhosis after upper gastrointestinal bleeding. 201. Runyon BA, McHutchison JG, Antillon MR, Akriviadis
Am J Gastroenterol 1998;93:962–966 EA, Montano AA. Short-course versus long-course anti-
184. Pauwels A, Mostefa-Kara N, Debenes B, Degoutte E, Levy biotic treatment of spontaneous bacterial peritonitis: a
VG. Systemic antibiotic prophylaxis after gastrointestinal randomized controlled study of 100 patients. Gastroenter-
hemorrhage in cirrhotic patients with a high risk of infection. ology 1991;100:1737–1742
Hepatology 1996;24:802–806 202. Sort P, Navasa M, Arroyo V, et al. Effect of intravenous
185. Bernard B, Grange JD, Khac EN, Amiot X, Opolon P, albumin on renal impairment and mortality in patients with
Poynard T. Antibiotic prophylaxis for the prevention of cirrhosis and spontaneous bacterial peritonitis. N Engl J Med
bacterial infections in cirrhotic patients with gastrointest- 1999;341:403–409
inal bleeding: a meta-analysis. Hepatology 1999;29:1655– 203. Fernandez J, Monteagudo J, Bargallo X, et al. A randomized
1661 unblinded pilot study comparing albumin versus hydro-
186. Toledo C, Salmeron JM, Rimola A, et al. Spontaneous xyethyl starch in spontaneous bacterial peritonitis. Hepatol-
bacterial peritonitis in cirrhosis: predictive factors of infec- ogy 2005;42:627–634
tion resolution and survival in patients treated with 204. Ochs A, Rossle M, Haag K, et al. The transjugular
cefotaxime. Hepatology 1993;17:251–257 intrahepatic portosystemic stent-shunt procedure for refrac-
187. Follo A, Llovet JM, Navasa M, et al. Renal impairment tory ascites. N Engl J Med 1995;332:1192–1197
after spontaneous bacterial peritonitis in cirrhosis: incidence, 205. Choudhury J, Sanyal AJ. Treatment of ascites. Curr Treat
clinical course, predictive factors and prognosis. Hepatology Options Gastroenterol 2003;6:481–491
1994;20:1495–1501 206. Luca A, Garcia-Pagan JC, Bosch J, et al. Beneficial effects of
188. Navasa M, Follo A, Filella X, et al. Tumor necrosis intravenous albumin infusion on the hemodynamic and
factor and interleukin-6 in spontaneous bacterial peritonitis humoral changes after total paracentesis. Hepatology 1995;
in cirrhosis: relationship with the development of renal im- 22:753–758
pairment and mortality. Hepatology 1998;27:1227–1232 207. Dagher L, Moore K. The hepatorenal syndrome. Gut 2001;
189. Garcia-Tsao G, Albillos A, Barden GE, West AB. Bacterial 49:729–737
translocation in acute and chronic portal hypertension. 208. Gines P, Cardenas A, Arroyo V, Rodes J. Management of
Hepatology 1993;17:1081–1085 cirrhosis and ascites. N Engl J Med 2004;350:1646–1654
190. Llovet JM, Bartoli R, March F, et al. Translocated intestinal 209. Gines P, Guevara M, Arroyo V, Rodes J. Hepatorenal
bacteria cause spontaneous bacterial peritonitis in cirrhotic rats: syndrome. Lancet 2003;362:1819–1827
molecular epidemiologic evidence. J Hepatol 1998;28:307– 210. Arroyo V, Gines P, Gerbes AL, et al. Definition and
313 diagnostic criteria of refractory ascites and hepatorenal
191. Tito L, Rimola A, Gines P, Llach J, Arroyo V, Rodes J. syndrome in cirrhosis. International Ascites Club. Hepatol-
Recurrence of spontaneous bacterial peritonitis in cirrhosis: ogy 1996;23:164–176
frequency and predictive factors. Hepatology 1988;8:27–31 211. Moller S, Bendtsen F, Henriksen JH. Pathophysiological
192. Andreu M, Sola R, Sitges-Serra A, et al. Risk factors for basis of pharmacotherapy in the hepatorenal syndrome.
spontaneous bacterial peritonitis in cirrhotic patients with Scand J Gastroenterol 2005;40:491–500
ascites. Gastroenterology 1993;104:1133–1138 212. Laffi G, La Villa G, Gentilini P. Pathogenesis and manage-
193. Runyon BA. Low-protein-concentration ascitic fluid is ment of the hepatorenal syndrome. Semin Liver Dis 1994;14:
predisposed to spontaneous bacterial peritonitis. Gastro- 71–81
enterology 1986;91:1343–1346 213. Arroyo V, Guevara M, Gines P. Hepatorenal syndrome in
194. Conn HO. Spontaneous peritonitis and bacteremia in cirrhosis: pathogenesis and treatment. Gastroenterology
Laennec’s cirrhosis caused by enteric organisms. A relatively 2002;122:1658–1676
214. Maurizio R, Eugenio C, Roberto PR. Results of sclerother- cirrhosis and type 1 hepatorenal syndrome. Hepatology
apy for bleeding esophageal varices in patients with 2004;40:55–64
schistosomal liver disease: a retrospective study. Hepatogas- 218. Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J,
troenterology 2000;47:424–428 ter Borg PC. A model to predict poor survival in patients
215. Ortega R, Gines P, Uriz J, et al. Terlipressin therapy with undergoing transjugular intrahepatic portosystemic shunts.
and without albumin for patients with hepatorenal syn- Hepatology 2000;31:864–871
drome: results of a prospective, nonrandomized study. 219. Wiesner RH, McDiarmid SV, Kamath PS, et al. MELD
Hepatology 2002;36(4 Pt 1):941–948 and PELD: application of survival models to liver allocation.
216. Angeli P, Volpin R, Gerunda G, et al. Reversal of type 1 Liver Transpl 2001;7:567–580
hepatorenal syndrome with the administration of midodrine 220. Wiesner R, Edwards E, Freeman R, et al. Model for end-
and octreotide. Hepatology 1999;29:1690–1697 stage liver disease (MELD) and allocation of donor livers.
217. Wong F, Pantea L, Sniderman K. Midodrine, octreotide, Gastroenterology 2003;124:91–96
albumin, and Trends Pharmacol Sci in selected patients with
Acute Renal Failure in the Intensive Care Unit
Steven D. Weisbord, M.D., M.Sc.1,2,3 and Paul M. Palevsky, M.D.1,3
ABSTRACT
Acute renal failure (ARF) is a common complication in critically ill patients, with
ARF requiring renal replacement therapy (RRT) developing in 5 to 10% of intensive
care unit (ICU) patients. Epidemiological studies have demonstrated that ARF is an
independent risk factor for mortality. Interventions to prevent the development of ARF are
currently limited to a small number of settings, primarily radiocontrast nephropathy and
rhabdomyolysis. There are no effective pharmacological agents for the treatment of
established ARF. Renal replacement therapy remains the primary treatment for patients
with severe ARF; however, the data guiding selection of modality of RRT and the optimal
timing of initiation and dose of therapy are inconclusive. This review focuses on the
epidemiology and diagnostic approach to ARF in the ICU and summarizes our current
understanding of therapeutic approaches including RRT.
KEYWORDS: Acute renal failure, renal replacement therapy
EPIDEMIOLOGY OF ACUTE RENAL 209 cases per million persons.1,2 Other series have
FAILURE demonstrated community-acquired ARF in 0.4 to
Acute renal failure (ARF) is an abrupt decline in kidney 0.9% of hospital admissions,3 whereas hospital-acquired
function that develops over a period of hours to days. ARF developed during 4.9 to 7.2% of hospitalizations,
Although simple to define conceptually, there is no with an apparent increase in incidence over a 2 decade
consensus on an operational definition of ARF. Defi- interval.4,5
nitions that have been used in clinical studies have been With a well-recognized relationship between se-
highly variable. Some have been based on changes in the verity of illness and risk for ARF, it is not surprising
serum creatinine (Scr) concentration with wide variabil- that the incidence of ARF increases dramatically in the
ity in the relative or absolute magnitude of change, intensive care unit (ICU) setting. Over a 10-month
others on the presence of oliguria, and still others on period spanning 1991 to 1992, Liano and Pascual
the need for renal replacement therapy (RRT). This studied the epidemiology of ARF in 13 tertiary-care
variation in definitions has confounded efforts to char- hospitals in Madrid, Spain.2 Of the 747 episodes of ARF
acterize the epidemiology of ARF. Interpretation and that were identified, 253 (34%) occurred in ICU pa-
comparison of epidemiological studies must therefore tients. In other studies, the prevalence of ARF in
take into consideration not only the characteristics of the critically ill patients has ranged from 3 to 25%. In a
patient population studied but also the particular criteria large, multinational study published a decade ago, ARF,
used to define ARF. defined as an increase in Scr to more than 3.5 mg/dL
In two large European studies, the overall or the presence of oliguria, was observed in 348
population-based annual incidence of ARF was 140 to of 1411 ICU patients (24.7%).6 Nearly half of these
1
Renal Section, VA Pittsburgh Healthcare System, Pittsburgh, Penn- Non-pulmonary Critical Care: Managing Multisystem Critical Illness;
sylvania; 2Center for Health Equity Research and Promotion; Guest Editor, Curtis N. Sessler, M.D.
3
Renal-Electrolyte Division, Department of Medicine, University of Semin Respir Crit Care Med 2006;27:262–273. Published by
Pittsburgh School of Medicine, Pittsburgh, Pennsylvania. Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY
Address for correspondence and reprint requests: Paul M. Palevsky, 10001, USA. Tel: +1(212) 584-4662.
M.D., Rm. 7E123 (111F-U), VA Pittsburgh Healthcare System, Univer- DOI 10.1055/s-2006-945527. ISSN 1069-3424.
sity Drive Division, Pittsburgh, PA 15240. E-mail: [email protected].
262
ACUTERENAL FAILUREIN THEICU/WEISBORD, PALEVSKY 263
patients had primarily medical causes for ICU admis- and colleagues observed an independent mortality risk
sion; however, in 25% of cases, ARF developed in the in patients developing radiocontrast nephropathy
setting of trauma or during the postoperative period. In (RCN).15 After adjustment for comorbidities, there
contrast, in a study of 17,126 patients in 30 Austrian was a 5.5-fold increased mortality risk associated with
ICUs, the incidence of ARF was only 4.9%, with ARF the development of ARF. The increased mortality risk
defined by the need for RRT.7 Similarly, Uchino and was greatest in patients with the lowest levels of comor-
colleagues reported a prevalence of severe ARF of 5.7% bid illness, and declined as the burden of comorbid
in 29,269 patients treated at 54 centers in 23 countries, conditions increased. Although this study was not re-
using a definition of ARF based on the presence of stricted to critically ill patients, other studies restricted to
oliguria or azotemia severe enough to warrant RRT.8 critically ill patients have demonstrated a similar in-
Although the varying incidence of ARF in these and creased mortality risk. Using their database of patients
other studies almost certainly relates to the different treated in Austrian ICUs, Metnitz and colleagues com-
criteria employed to define this syndrome, it is quite clear pared survival in patients with and without ARF based
that ARF complicates the clinical course of a substantial on disease-specific observed to expected (O:E) mortality
proportion of patients requiring ICU care. ratios.7,16 Among patients with lower severity of illness,
ARF was associated with a sevenfold higher O:E mortal-
ity ratio.16 Differences in the more severely ill patients
OUTCOMES OF ACUTE RENAL FAILURE were less robust, suggesting that among the most crit-
Much like efforts to describe the epidemiology of ARF, ically ill, risk-adjustment for severity of illness becomes
attempts to characterize outcomes associated with ARF increasingly difficult, and extrarenal factors are progres-
are confounded by the different definitions that have sively more important predictors of outcome. Similarly,
been employed. A series of studies in demographically Chertow and colleagues demonstrated a nearly eightfold
diverse populations have demonstrated in-hospital mor- increased mortality risk associated with ARF following
tality rates in critically ill patients with ARF that range open heart surgery.17
from 35% to as high as 75%.7,9–13 A variety of factors As part of a study addressing the effect of ARF in
have been associated with mortality risk. In the recent ICU patients, Clermont and colleagues prospectively
multinational study by Uchino and colleagues, tracked outcomes of 1530 ICU admissions at a single
independent predictors of hospital mortality included institution over a 10-month period, categorizing patients
use of vasopressors and mechanical ventilation, septic into three groups: patients with ARF, patients with end-
shock, cardiogenic shock, and hepatorenal syndrome.8 stage renal disease (ESRD), and patients without evi-
Although the definition of ARF used in this study dence of renal failure.9 Mortality among patients who
precluded an assessment of the effect of RRT on hospital developed ARF was double that observed among pa-
mortality, this variable has emerged as one of the primary tients with ESRD and four times greater than that seen
predictors of hospital mortality in multiple other studies in patients without renal dysfunction (23% vs 11% vs
of ARF in critically ill patients. Metnitz and colleagues 5%). Although Acute Physiology and Chronic Health
found a fourfold higher mortality rate among patients Evaluation (APACHE) III scores were comparable in
requiring RRT for ARF than in patients without ARF the ARF and ESRD cohorts, hemodynamic instability
(62.8% vs 15.6%; p < .001).7 This excess mortality per- and leukocytosis were more common in patients with
sisted after adjustments for illness severity, age, and ARF and the study lacked a robust multivariable adjust-
treatment center. Similarly, in the Program to Improve ment for confounding factors. Despite these limitations,
Care in Acute Renal Disease (PICARD), a multicenter the results of this study imply that the pathophysiolog-
observational study of ARF in the United States, Mehta ical effects of ARF extend beyond the loss of organ
and colleagues also demonstrated markedly higher mor- function.
tality rates in patients with ARF who required RRT than
in those who did not (45% vs 24%).14 Thus it appears
that not only is ARF associated with increased mortality ETIOLOGIES OF ACUTE RENAL FAILURE
risk compared with other critically ill patients, but RRT Early recognition of ARF and prompt determination of
is one of the strongest negative prognostic factors, its etiology facilitate the institution of appropriate ther-
although this may merely reflect RRT serving as a apy and can potentially mitigate its severity. ARF can be
surrogate for severity of renal injury. broadly classified into prerenal, intrinsic, and postrenal
It had been widely believed that with the wide- etiologies. Prerenal azotemia, the most common form of
spread availability of RRT to manage the uremic con- hospital-acquired ARF, represents a decline in glomer-
sequences of renal failure, patients with ARF died ular filtration rate (GFR) due to hemodynamic factors
with, but not of, their renal failure. However, evidence that diminish effective renal perfusion. The defining
accrued over the past decade suggests that this paradigm pathological feature in prerenal azotemia is the absence
is incorrect. In the earliest of these studies, Levy of histological damage to the renal parenchyma. The
Table 1 Laboratory Findings in the Syndromes of Acute Renal Failure

BUN:Scr Ratio UNa (mEq/L) FENa Urinalysis Other Findings
Prerenal ARF > 20:1 < 20 < 1% Bland specific gravity > 1.015 FEurea < 35% hyperuricemia
Intrinsic ARF
ATN 10:1 > 40 > 2% Granular casts specific FEurea > 50%
gravity 1.010
AIN 10:1 Variable Variable RBCs, WBCs, WBC casts, Eosinophilia
eosinophiluria
GN Variable < 20 < 1% RBCs, RBC casts —
Intratubular Variable Variable Variable Crystalluria or immunoglobulin Urine or serum monoclonal
Obstruction light chains paraprotein
Vascular Variable Variable Variable Variable Hematuria
Postrenal ARF > 20:1 Variable Variable Variable Fluctuating urine output elevated
postvoid bladder
volume hydronephrosis
ATN, acute tubular necrosis; AIN, acute interstitial nephritis; GN, glomerulonephritis; BUN, blood urea nitrogen; Scr, serum creatinine; UNa, urine
sodium concentration; FENa, fractional excretion of sodium; FEurea, fractional excretion of urea; RBC, red blood cell; WBC, white blood cell.
physiological response to diminished renal perfusion is with upper tract obstruction, ureteral stenting or percu-
an increase in renal tubular sodium avidity and urinary taneous nephrostomies are required.
concentration. Clinically, the increase in sodium avid- Although pre- and postrenal ARF can cause or
ity is manifested by a urine sodium concentration less exacerbate a decline in renal function in critically ill
than 20 mEq/L and a fractional excretion of sodium patients, the most common etiology of ARF in the ICU
less than 1% (Table 1). In patients on diuretics, urinary is intrinsic ARF, most commonly due to ATN. In a
sodium may be increased as a result of pharmacological prospective analysis of patients hospitalized in 28 ICUs
blockade of reabsorptive pathways. In such patients, in France, 83% of recorded cases of ARF were attributed
a fractional excretion of urea of less than 35% is to ATN, with 54% associated with ischemic renal injury,
highly suggestive of a prerenal state. The hemody- 8% from nephrotoxic injury, and 21% of mixed etiol-
namically mediated secretion of vasopressin increases ogy.11 Among 253 episodes of ARF in critically ill
urinary concentration, resulting in a urine specific patients reported by Liano and colleagues, ATN ac-
gravity greater than 1.015 to 1.020 and urine osmo- counted for 75.9% of episodes compared with 61.4% in
lality greater than 300 mOsm/kg H2O. Increased non-ICU patients.18 Prerenal azotemia was the second
tubular reabsorption of urea in prerenal azotemia leading cause of ARF, with obstructive disease account-
may result in a disproportionate elevation in blood ing for less than 1% of cases.
urea nitrogen (BUN) concentration relative to Scr. Unlike prerenal azotemia, ATN is defined by the
Severe and protracted renal hypoperfusion can lead presence of tubular epithelial cell injury, apoptosis, and
to ischemic injury to the renal parenchyma and can necrosis, and is associated with the characteristic urinary
contribute to the development of acute tubular ne- finding of coarse granular ‘‘muddy brown’’ casts com-
crosis (ATN), making early recognition and treatment posed of sloughed epithelial cells and cellular debris. The
of prerenal azotemia essential. Prompt reversal of the resulting defects in renal tubular function generally result
abnormal hemodynamics will usually lead to full re- in impaired sodium reabsorption, a urine sodium con-
covery of renal function. centration greater than 40 mEq/L, and fractional ex-
Postrenal ARF results from anatomical obstruc- cretion of sodium greater than 2 to 3%. Defective urinary
tion to urine flow at the levels of the ureters or bladder concentration and dilution are manifested by isosthenu-
outlet. Although obstructive disease may be unilateral, ric urine, with a specific gravity of 1.010 and an
renal failure requires the presence of bilateral obstruction osmolality of 250 to 300 mOsm/kg H2O.
or unilateral obstruction of a solitary functional kidney. ATN may result from renal ischemia, endoge-
Postrenal ARF is usually readily diagnosed on the basis nous or exogenous nephrotoxins, and/or sepsis.
of bilateral hydronephrosis on renal ultrasound or the Although the pathophysiology of ATN remains in-
finding of an elevated postvoid residual bladder volume completely understood, robust conceptual models of its
(> 100 mL). Treatment of postrenal disease requires pathogenesis have been developed. In ischemic ATN,
relief of the obstruction. In patients in whom the the initial insult causes hypoxia of highly metabolically
obstruction is at the level of the urinary bladder, im- active tubular cells. Cells at the corticomedullary junc-
provement in renal function may be accomplished with tion are particularly susceptible to ischemic injury due
simple placement of a bladder catheter. Among patients to their high basal oxygen demand and relatively low
regional oxygen delivery. Although ATN associated Table 2 Prevention of Radiocontrast-Induced Acute
with sepsis has commonly been viewed as mediated Renal Failure
primarily by ischemic injury, recent data suggest pri- Effective interventions
macy of other pathophysiological processes.19 Recent Volume expansion with isotonic saline solutions
experimental animal models and limited human studies Avoidance of high-osmolar radiocontrast agents
suggest that overall renal perfusion is preserved or even Minimization of volume of radiocontrast
increased with sepsis, yet regional redistribution of Discontinuation of nonsteroidal antiinflammatory drugs
blood flow within the kidney may shunt blood away Potentially effective interventions
from the corticomedullary junction.20–22 Although spe- Sodium bicarbonate (as compared with sodium chloride)
cific mechanisms for deterioration in renal function in N-acetylcysteine
hyperdynamic sepsis remain unknown, recent work Iso-osmolar radiocontrast agents (as compared with
suggests that activation of cellular and humoral inflam- low-osmolar agents)
matory mediators and of coagulation pathways play a Theophylline
central role.23–27 In nephrotoxic ATN, direct cytotox- Ineffective or potentially harmful interventions
icity to tubular epithelial cells is the primary mechanism Dopamine
of renal injury. Epithelial cell injury leads to loss of Fenoldopam
polarity, activation of apoptotic pathways resulting in Atrial natriuretic peptide
both apoptotic and necrotic cell death, and sloughing of Diuretics
both viable and dying epithelial cells into the tubular Mannitol
lumen. The associated fall in GFR that defines ARF is Renal replacement therapy
mediated by a combination of intratubular obstruction
from the sloughed debris, backleak of glomerular fil-
trate across the denuded tubular basement membrane, that utilize intravascular radiocontrast, even in critically
and intrarenal vasoconstriction. Although the tubular ill patients, are planned sufficiently in advance to permit
epithelial cell injury has been the central focus of the implementation of preventative measures. Factors
understanding ATN, the role of endothelial cell injury, that predispose to the development of RCN include
generation of reactive oxygen species, and activation of preexisting renal insufficiency, diabetes mellitus, and
inflammatory pathways have been increasingly recog- effective intravascular volume depletion. The presence
nized as critical to the pathogenesis of ATN.28–31 of one or more of these risk factors should prompt the
A variety of other intrinsic forms of ARF such as institution of preventive interventions. The best vali-
acute interstitial nephritis, acute glomerular disease, dated of these strategies are the administration of intra-
atheromboembolic disease, and tumor lysis syndrome venous (IV) fluids and discontinuation of diuretics to
account for some cases of ARF in critically ill patients. expand the intravascular space. Although initial regi-
Nonetheless, the following discussion of the prevention mens for fluid administration in the prevention of RCN
and treatment of ARF will focus on patients with ATN employed hypotonic saline,32 isotonic saline, infused at
because it is the predominant form of ARF in the ICU 1 mL/kg/h for 12 hours prior to and 12 hours following
setting. the administration of radiocontrast, provides greater
protection than equal volumes of hypotonic saline.33
More recently, Merten and colleagues reported an in-
PREVENTION OF ACUTE RENAL FAILURE cidence of RCN of only 1.7% using isotonic sodium
The morbidity and mortality associated with ARF have bicarbonate administered at 3 mL/kg/h for 1 hour prior
energized multiple efforts to identify strategies to fore- to and at 1 mL/kg/h for 6 hours following radiocontrast
stall its development. Unfortunately, most episodes of administration compared with an incidence of 13.6%
ARF are unpredictable, and hence prophylactic inter- with equal volumes of isotonic sodium chloride.34
ventions are impractical. There are, however, specific Although these results are notable, the protocol for fluid
settings in which it is possible to identify high-risk administration was not comparable to that used in the
patients and institute preemptive strategies to decrease majority of other studies. It is therefore not possible to
the risk of ARF. establish the superiority of this regimen to more conven-
tional regimens. In addition, the mechanism for the
benefit associated with sodium bicarbonate is not well
Radiocontrast Nephropathy understood, although it is speculated that it may relate to
The administration of intravascular radiocontrast media decreased production of reactive oxygen species. Addi-
represents one of the most important settings for such tional studies involving multiple centers and with larger
strategies (Table 2). RCN is one of the most common numbers of patients are needed to validate the conclu-
forms of ATN, accounting for 10% of hospital- sions of this study. However, because the risks associated
acquired ARF.4 Many of the radiographic procedures with isotonic sodium bicarbonate administration are
minimal in the majority of patients, use of this agent is a decline in kidney function.50 The implications of this
reasonable alternative to infusions of isotonic saline. are significant because even small changes in Scr are
The role of the antioxidant N-acetylcysteine in associated with increased postoperative mortality,51
preventing RCN is controversial. Although the initial whereas postoperative ARF severe enough to require
study describing its use suggested a marked benefit,35 the use of RRT is associated with an approximately
subsequent studies have reached conflicting conclu- eightfold increased risk of death after adjusting for
sions.36,37 Because this agent is both inexpensive and comorbidities.17 Factors predisposing to the develop-
without significant side effects, its use is not unreason- ment of ARF following cardiac surgery include preop-
able while awaiting more definitive data. Theophylline, erative renal insufficiency, decreased left ventricular
an adenosine antagonist, has also been proposed as ejection fraction, and valvular surgery.52 Although
potentially beneficial in preventing RCN. In a meta- high-risk patients can be identified preoperatively, in-
analysis of six published studies, the effect size observed terventions to prevent postoperative ARF have not been
was similar to that seen with N-acetylcysteine; however, effective. In a randomized, controlled trial, prophylactic
this result did not reach statistical significance.38 Given administration of dopamine did not decrease the risk of
the potential for complications with this agent, partic- ARF, whereas furosemide increased its incidence.53
ularly in patients with cardiac disease, it is a less Although earlier studies of atrial natriuretic peptide
attractive agent than N-acetylcysteine for use in the did not demonstrate a benefit, a recent small study of
absence of a clear benefit. A variety of other pharmaco- low-dose recombinant human atrial natriuretic peptide
logical agents, including dopamine, mannitol, furose- reduced the need for dialysis in patients manifesting a
mide, atrial natriuretic peptide, and fenoldopam, have 50% increase in Scr following cardiac surgery.54 This
been shown to be ineffective, or even harmful, in suggests a potential benefit, but additional evaluation
preventing RCN and should not be used.39 will be necessary for confirmation. The potential benefit
The selection of radiocontrast agent is also an im- of off-pump coronary artery bypass graft (CABG) sur-
portant consideration in preventing RCN. Low-osmolar gery in decreasing the risk for renal injury compared with
radiocontrast agents are associated with a lower risk of on-pump CABG is controversial. Although case series
RCN compared with the older high-osmolar agents, have suggested a decreased risk of ARF,55,56 no benefit
particularly in patients with underlying kidney disin renal outcomes was reported in a more recent pro-
ease.40,41 In patients with both diabetes mellitus and spective observational study.57
renal insufficiency, iso-osmolar agents may be more
beneficial than low osmolar agents.42 Regardless of the
class of radiocontrast, the volume administered should be TREATMENT OF ESTABLISHED ACUTE
minimized because volumes in excess of 250 to 300 mL RENAL FAILURE
are associated with increased risk of RCN.43,44
Pharmacological Therapy
Multiple pharmacological agents including dopamine,
Rhabdomyolysis the dopamine receptor agonist fenoldopam, loop diu-
Rhabdomyolysis is an important cause of ARF, espe- retics, atrial natriuretic peptide, insulin-like growth
cially in trauma patients. The sine qua non of this factor-1 (IGF-1) and thyroxine are effective for the
condition is an elevation in the creatine phosphokinase treatment of ARF in animal models. However, similar
concentration. ARF results from the toxicity of myoglo- success has not been observed in human studies. No
bin and other intracellular constituents released from pharmacological agents have been clinically validated for
damaged myocytes. The early and aggressive adminis- the treatment of established ARF.
tration of IV fluids is well recognized as effective at
preventing ARF in this setting.45,46 In patients with DOPAMINE AND FENOLDOPAM
crush injuries, fluid administration with 1 L/h of normal When dopamine is administered in low doses (0.5–
saline should be initiated promptly, even before extrac- 2.0 mg/kg/min), renal plasma flow and GFR rise, and
tion of the patient, if possible. Although urinary alka- urinary sodium excretion increases.58,59 Based on these
linization with bicarbonate-containing fluids,47 and physiological effects, low-dose dopamine has been, and
forced mannitol diuresis48 have been advocated, the continues to be, used in critical care settings to attenu-
benefit of these agents is uncertain.49 ate the clinical impact of ARF and augment urine
output. However, its use is not supported by clinical
studies. In the largest prospective trial, Bellomo and
Postsurgical Acute Tubular Necrosis colleagues randomized 328 critically ill patients with
The incidence of ARF following cardiac and vascular early ARF to infusions of low dose dopamine or
surgery is significant, with some series demonstrating placebo.60 There was no benefit with regard to duration
that up to 40% of patients manifest a perioperative of ARF, need for RRT, or mortality associated with
dopamine. In a meta-analysis of over 60 published the use of diuretics to delay the initiation of RRT is
studies, Friedrich and colleagues also found no benefit associated with benefit or harm.67
of low dose dopamine on mortality or need for dialysis,
although there was a small benefit in terms of urine GROWTH FACTORS
volume excreted on the first day of therapy. However, Growth factors accelerate recovery of renal function in
even this benefit was not sustained beyond the first experimental models of ATN. In human trials, similar
day.61 Given the lack of benefit and the recognized benefit has not been observed. IGF-1 did not hasten
complications, most notably cardiac arrhythmias, asso- recovery of renal function, decrease the need for dial-
ciated with this agent, there is no role for low-dose ysis, or alter mortality.68 Not only did thyroxine not
dopamine in the management of ARF. improve renal recovery, it was associated with increased
The dopamine-receptor agonist fenoldopam has mortality.69,70
been evaluated in a small pilot study for the treatment of
ARF. In 155 critically ill patients with early ARF
randomized to fenoldopam or placebo, fenoldopam Renal Replacement Therapy
was associated with a trend toward decreased need for
dialysis and improved survival, although these findings TIMING OF INITIATION
did not reach statistical significance.62 Further evalua- RRT is the primary means for managing severe ARF,
tion using a larger patient sample will be necessary to especially when complicated by hyperkalemia, severe
adequately assess the potential role for this agent. metabolic acidosis, volume overload, or overt uremic
symptoms. Although the recognition of these clinical
LOOP DIURETICS indications for renal support is relatively straightfor-
Loop diuretics inhibit sodium transport in the loop of ward, there is uncertainty regarding the optimal time
Henle through inhibition of the Na-K-2Cl cotrans- to initiate RRT in the absence of these complications.
porter. It has been hypothesized that, by decreasing Advocates for the early initiation of RRT argue that
metabolic demand in this nephron segment, increasing RRT should be provided as soon as it is clear that the
urine flow, and washing out intratubular debris in more patient has sustained a significant and persistent re-
distal tubular segments, loop diuretics may be beneficial duction in GFR in order to maintain as normal a
in patients with ARF. Moreover, based on the observa- metabolic milieu as possible. The argument against
tion that patients with nonoliguric ARF have a better early initiation is that it will subject some patients to
prognosis than patients with oliguric ARF, loop diu- the risks of RRT who, if managed conservatively,
retics have been used in attempts to convert patients might recover renal function without requiring renal
from an oliguric to a nonoliguric state. Unfortunately, support. In addition, there is, at best, only limited data
clinical trials have failed to support the utility of loop suggesting an outcome benefit associated with early
diuretics in the treatment of ARF. In studies that are initiation of therapy.
over 2 decades old, the use of furosemide to increase The debate regarding timing of initiation of RRT
urine output had no impact the requirement for dialysis, extends back to the early 1960s when Teschan and
recovery of ARF, or mortality.63,64 colleagues, and Easterling and Forland reported benefits
A more recent observational study used propen- to ‘‘prophylactic’’ dialysis, initiated prior to uremic
sity scoring to adjust for factors leading to diuretic use.65 symptoms, in uncontrolled case series of patients with
Diuretic use was associated with an adjusted odds ratio ARF.71,72 During the next decade, a series of retrospec-
for mortality of 1.68 (CI, 1.96–2.64), for nonrecovery of tive studies supported the conclusion that earlier initia-
renal function of 1.79 (CI, 1.19–2.68), and for a com- tion of dialysis, prior to the development of advanced
posite end point of death or nonrecovery of renal azotemia, was associated with improved survival.73–75
function of 1.77 (CI, 1.14–2.76), suggesting potential The first prospective analysis of this issue was a study
harm with diuretic use. However, analysis of data from a of 18 consecutive patients assigned in an alternating
multinational study of 1743 critically ill patients with fashion to an early dialysis regimen (to maintain the
ARF did not reproduce these findings.66 Because all of BUN < 70 mg/dL and the Scr < 5 mg/dL), or late
the increased risk in the initial study was attributable to dialysis (BUN 150 mg/dL, Scr 10 mg/dL, or clin-
the diuretic-unresponsive patients, it has been suggested ical symptoms) published by Conger in 1975.76 Survival
that the adverse impact of diuretic therapy may result among patients receiving the intensive regimen was
from a delay in instituting RRT while using escalating superior to that observed in the nonintensive cohort
doses of diuretics. (64% vs 20% p < .01), and the frequency of gram-
In summary, diuretic therapy is not associated negative sepsis and gastrointestinal hemorrhage was
with an alteration in the clinical course of ARF. In diminished. In a subsequent study, Gillum and collea-
diuretic-responsive patients, their use may facilitate gues76a randomized 34 patients to receive either inten-
volume management; however, it is uncertain whether sive dialysis (maintaining the BUN < 60 mg/dL and
the Scr < 5 mg/dL) or nonintensive dialysis (BUN CRRT is an umbrella term used to describe a
100 mg/dL and Scr 9 mg/dL). Mortality was family of therapies that provide slow continuous
higher (59% vs 47%), although hemorrhagic and septic removal of solute and fluid. The variants of CRRT
complications were less frequent in the intensively differ with regard to the mode of vascular access for
treated patients; yet these differences did not achieve the extracorporeal circuit [arteriovenous (AV) or
statistical significance. These data form the basis for the venovenous (VV)] and the mechanism of solute re-
conventional teaching that dialysis should be initiated moval (hemodialysis, hemofiltration, or hemodiafiltra-
when the BUN approaches 100 mg/dL and that no tion).80,81 When CRRT was first introduced, arterial
further benefit is seen with earlier initiation of therapy. cannulation was utilized, with the gradient between
This dictum, however, is subject to the inherent limi- mean arterial pressure and central venous pressure
tations of retrospective analyses and underpowered pro- providing the driving force for blood flow through the
spective studies. extracorporeal circuit. Although the use of an AV
More recent investigation into the timing of RRT circuit provided for technological simplicity without
has focused on continuous RRT (CRRT). Gettings and the need for a blood pump or pressure monitors,
colleagues retrospectively compared outcomes among reliance on the AV pressure gradient limited the blood
100 adults with posttraumatic ARF who were initiated flow through the extracorporeal circuit. Moreover,
on CRRT when their BUN was < 60 mg/dL (early) or prolonged arterial cannulation was associated with un-
> 60 mg/dL (late).77 The early group was initiated on acceptably high complication rates.82 For these reasons,
CRRT an average of 9 days before the late group pump-driven VV modalities are now nearly universally
(10 15 days vs 19 27 days, p < .0001) and had a used.
substantially lower BUN at the time of initiation of Solute removal during CRRT can be provided by
therapy (43 13 mg/dL vs 94 28 mg/dL, p < .0001). either diffusion or convection. In continuous hemodial-
Survival was 39% in the early group compared with 20% ysis, diffusive solute transport predominates; in hemofil-
in the late initiation group (p ¼ .041). Although the two tration, convective transport predominates; and in
groups had similar levels of acuity of illness, the retro- hemodiafiltration there is a combination of both mech-
spective design of the study does not eliminate the anisms. Theoretically, convective clearance allows for
possibility that differences in outcomes were related to greater removal of middle and higher molecular weight
unrecognized differences in the clinical characteristics of solutes. It has been postulated that the potentially
the two groups. Similar findings have also been observed greater removal of inflammatory mediators using hemo-
in a recent retrospective study of CRRT following filtration favors this technique over purely diffusive
cardiac surgery.78 In the only prospective study evaluat- therapies. In one study, lower tumor necrosis factor-
ing timing of initiation of CRRT, Bouman and col- alpha (TNF-a) levels were achieved during continuous
leagues did not observe improved outcomes with early VV hemofiltration (CVVH) than during continuous
initiation of therapy, although the study is notable for its venovenous hemodialysis (CVVHD).83 However, no
small sample size and an overall patient survival that difference in clinical outcomes based on modality of
suggests a lower acuity of illness than most studies of CRRT has been reported.
ARF in critically ill patients.79 Thus, to date there are From a conceptual standpoint, it seems logical
inadequate data to permit consensus on the optimal that the use of CRRT with its gradual fluid and solute
timing of initiation of RRT. Resolution of this question removal would be superior to the rapid volume and
will require adequately powered randomized, controlled solute flux associated with IHD in the critically ill
trials because this question cannot be adequately an- patient with hemodynamic instability. However, clin-
swered using retrospective or observational data. ical trials have not demonstrated outcome benefits
associated with CRRT. The majority of studies com-
MODALITY OF RENAL REPLACEMENT THERAPY paring these modalities have been fraught with prob-
Over the past 2 decades there has been a rapid expansion lems related to disparities in disease severity because
of the modalities of RRT available for the management more seriously ill patients are more likely to receive
of ARF. Although the options were once limited to CRRT. Additionally, nonrandomized and/or retro-
intermittent hemodialysis (IHD) and peritoneal dialysis, spective study designs have confounded these compar-
the current armamentarium of therapies includes multi- isons. In a single-center retrospective comparison,
ple variants of CRRT and the more recently introduced Swartz and colleagues observed a twofold greater mor-
‘‘hybrid’’ therapies, such as sustained low-efficiency dial- tality in patients treated with CVVH compared with
ysis (SLED) and extended daily dialysis (EDD), which patients whose ARF was managed using IHD.84 After
combine the machine technology of conventional IHD adjusting for the greater burden of comorbid conditions
with the extended duration of CRRT. Unfortunately, in the patients managed using CVVH using two
despite the growing number of options, objective data to separate multivariate models, no difference in the
guide the selection of modality are limited. odds of death was observed between modalities. Similar
results were observed in a prospective, multicenter, The data comparing other modalities of RRT are
observational study by Guérin and colleagues.85 In limited. One randomized, controlled trial demonstrated
this series, mortality was 79% in 354 patients managed CVVH to be superior to peritoneal dialysis in infection-
with CRRT and 59% in patients managed with IHD. associated ARF.93 The generalizability of this study is
However, after performing logistic regression to adjust limited, however, by the predominance of malaria as the
for comorbidities, modality of RRT was not independ- underlying etiology of ARF. No studies have directly
ently associated with outcome. compared outcomes with ‘‘hybrid’’ therapies to either
In a randomized, controlled trial of 166 patients IHD or CRRT, although these therapies have been
with ARF conducted by Mehta and colleagues, ICU and shown to provide similar hemodynamic stability and
hospital mortality rates were 59.5% and 65.5%, respec- metabolic control to CRRT.94
tively, in patients randomized to CRRT and 41.5% In summary, current data are inadequate to guide
and 47.6%, respectively, in patients randomized to IHD selection of modality of RRT in ARF. Issues associated
(p < .02).86 Unfortunately, the randomization in this with study methodology, lack of comparability of treat-
study was imbalanced, resulting in higher APACHE III ment groups, and inadequate sample size limit the
scores and a higher prevalence of liver failure in the interpretation of studies that have attempted to address
patients randomized to CRRT. After adjusting for this question. The choice of modality of RRT should
the differences between groups using either logistic therefore be dictated primarily by local expertise and
regression or proportional hazards regression, there was availability of equipment and personnel.
no difference in mortality attributable to modality
of RRT. Two meta-analyses have attempted to compare DOSE OF RENAL REPLACEMENT THERAPY
outcomes between these modalities.87,88 One meta-anal- Guidelines for the dosing of dialysis for patients with
ysis that included both randomized and nonrandomized ESRD are well established. Unfortunately, similar
studies concluded that weakness in study quality signifi- guidelines for the dose of RRT for patients with ARF
cantly limited comparison between modalities, although do not exist. When determining the dose of IHD for
there was a suggestion that CRRT might be potentially patients with ARF, both the frequency and the dose of
superior when studies were weighted based on assessment each treatment session need to be considered. Only one
of study quality.87 The second meta-analysis restricted the study has evaluated the effect of IHD frequency on
included studies to six randomized trials, only one of outcomes among patients with ARF. In this study,
which, the study by Mehta and colleagues already dis- Schiffl and colleagues assigned 160 critically ill patients
cussed, was designed to evaluate mortality as an outcome with ATN in an alternating fashion to daily or alternate
and had been published as a peer-reviewed manuscript.88 day dialysis.95 Patients who received daily dialysis had
This meta-analysis found no difference in survival asso- decreased mortality 14 days after discontinuation of
ciated with modality of RRT. RRT (28% vs 46%, p ¼ .01), and shorter duration of
An additional randomized, controlled trial com- ARF (9 2 vs 16 6 days, p ¼ .001). Although these
paring IHD to CRRT was published subsequent to results are striking, concern has been raised that the dose
these two meta-analyses.89 In this study of 80 patients, of dialysis delivered to the alternate-day treatment group
acuity of illness was similar between the two treatment was exceptionally low, resulting in a markedly elevated
arms. Although CRRT was associated with greater net time-averaged BUN in these patients and a high in-
volume removal during the first 72 hours of therapy cidence of uremic complications, including gastrointes-
and greater hemodynamic stability than IHD, there was tinal bleeding, altered mental status, and infections.96
no observed difference in mortality between the two Thus, although demonstrating that increasing the dose
treatments. of dialysis from a very low level of alternate-day therapy
It has been suggested that, despite the absence of is associated with improved outcomes, this study does
a survival benefit with CRRT, recovery of renal function not provide convincing evidence that increasing the
may be more likely with this mode of therapy.86,90,91 The frequency of therapy provides added benefit to patients
clinical mechanism postulated for this benefit is the receiving an ‘‘adequate’’ delivered dose of therapy on an
lesser degree of hemodynamic instability with CRRT every other day or three-times per week schedule.
compared with IHD, leading to fewer episodes of intra- There are only limited data to establish what the
dialytic hypotension and associated renal ischemia. ‘‘adequate’’ dose of therapy should be. In a retrospective
Although greater recovery of renal function has been study, Paganini and colleagues evaluated survival as a
observed in surviving patients in these studies, limiting function of the delivered dose of dialysis in critically ill
the analysis to surviving patients fails to account for the patients with ARF.97 Although the dose of therapy
competing risk of mortality. When analyzed using the appeared to have no impact on outcome among patients
combined end point of death or nonrecovery of renal with either very high or very low acuity of illness, in
function, no difference in outcome can be ascribed to patients with intermediate severity of illness, doses of
modality.92 dialysis above the 50th percentile were associated with
improved survival compared with patients who received 2. Liano F, Pascual J. Epidemiology of acute renal failure: a
lower delivered doses of therapy. However, the median prospective, multicenter, community-based study. Madrid
dose of therapy was substantially lower than what would Acute Renal Failure Study Group. Kidney Int 1996;50:811–
818
be deemed appropriate in the chronic setting. In the
3. Kaufman J, Dhakal M, Patel B, Hamburger R. Community-
absence of other studies establishing a relationship acquired acute renal failure. Am J Kidney Dis 1991;17:191–
between dose and outcome, a consensus panel convened 198
by the multinational Acute Dialysis Quality Initiative 4. Nash K, Hafeez A, Hou S. Hospital-acquired renal
(ADQI) concluded that the patients with ARF should insufficiency. Am J Kidney Dis 2002;39:930–936
receive at least the same minimum dose of dialysis that is 5. Hou SH, Bushinsky DA, Wish JB, Cohen JJ, Harrington JT.
considered appropriate for patients with end stage kid- Hospital-acquired renal insufficiency: a prospective study. Am
J Med 1983;74:243–248
ney disease.98
6. de Mendonca A, Vincent JL, Suter PM, et al. Acute renal
The data regarding dosing of therapy in CRRT failure in the ICU: risk factors and outcome evaluated by the
are slightly more robust. Ronco and colleagues random- SOFA score. Intensive Care Med 2000;26:915–921
ized 435 patients to one of three doses of CVVH, 7. Metnitz PG, Krenn CG, Steltzer H, et al. Effect of acute
defined by ultrafiltration rates of 20 mL/kg/h, 35 mL/ renal failure requiring renal replacement therapy on outcome
kg/h, and 45 mL/kg/h.13 Survival was markedly higher in critically ill patients. Crit Care Med 2002;30:2051–
in the intermediate and high dose arms (57% and 58%, 2058
8. Uchino S, Kellum JA, Bellomo R, et al. Acute renal failure in
respectively) compared with the low dose arm (41%,
critically ill patients: a multinational, multicenter study.
p < .001). In a subsequent study, however, Bouman and JAMA 2005;294:813–818
colleagues observed no such advantage with higher doses 9. Clermont G, Acker CG, Angus DC, Sirio CA, Pinsky MR,
of CRRT.79 However, the overall survival of greater Johnson JP. Renal failure in the ICU: comparison of the
than 70% in this study suggests that the enrolled patients impact of acute renal failure and end-stage renal disease on
may not have been representative of the majority of ICU outcomes. Kidney Int 2002;62:986–996
critically ill patients with ARF. Therefore, although 10. Chertow GM, Christiansen CL, Cleary PD, Munro C,
Lazarus JM. Prognostic stratification in critically ill patients
definitive recommendations cannot be made, the data
with acute renal failure requiring dialysis. Arch Intern Med
suggest that CRRT should be dosed to provide an 1995;155:1505–1511
ultrafiltration rate of at least 35 mL/kg/h. Several large 11. Guérin C, Girard R, Selli JM, Perdrix JP, Ayzac L. Initial
randomized controlled trials are under way in the United versus delayed acute renal failure in the intensive care unit. A
States and elsewhere to better define the optimal dosing multicenter prospective epidemiological study. Rhone-Alpes
of RRT in ARF.99 Area Study Group on Acute Renal Failure. Am J Respir Crit
Care Med 2000;161:872–879
12. Maher ER, Robinson KN, Scoble JE, et al. Prognosis of
critically ill patients with acute renal failure: APACHE II
SUMMARY score and other predictive factors. Q J Med 1989;72:857–
ARF is common in the ICU. Preventive strategies 866
should be utilized in patients at high risk for RCN or 13. Ronco C, Bellomo R, Homel P, et al. Effects of different
rhabdomyolysis. RRT remains the mainstay of suppor- doses in continuous veno-venous haemofiltration on out-
tive care for the critically ill patient with established comes of acute renal failure: a prospective randomised trial.
ARF because no effective pharmacological therapy is Lancet 2000;356:26–30
14. Mehta RL, Pascual MT, Soroko S, et al. Spectrum of acute
available. The high prevalence of ARF in the ICU
renal failure in the intensive care unit: the PICARD
setting necessitates a firm understanding by critical experience. Kidney Int 2004;66:1613–1621
care providers of the salient issues related to timing of 15. Levy EM, Viscoli CM, Horwitz RI. The effect of acute renal
initiation of RRT, choice of modality, and optimal dose, failure on mortality: a cohort analysis. JAMA 1996;275:1489–
all of which remain subjects of substantial debate and 1494
active clinical investigation. 16. Joannidis M, Metnitz PG. Epidemiology and natural history
of acute renal failure in the ICU. Crit Care Clin 2005;21:239–
249
FUNDING
17. Chertow GM, Levy EM, Hammermeister KE, Grover F,
Dr. Weisbord is supported by a VA Health Services Daley J. Independent association between acute renal failure
Research and Development Career Development and mortality following cardiac surgery. Am J Med 1998;
Award. 104:343–348
18. Liano F, Junco E, Pascual J, Madero R, Verde E. The
spectrum of acute renal failure in the intensive care unit
REFERENCES compared with that seen in other settings. The Madrid Acute
Renal Failure Study Group. Kidney Int Suppl 1998;66:S16–
1. Feest TG, Round A, Hamad S. Incidence of severe acute renal S24
failure in adults: results of a community based study. BMJ 19. Klenzak J, Himmelfarb J. Sepsis and the kidney. Crit Care
1993;306:481–483 Clin 2005;21:211–222
20. Badr KF, Kelley VE, Rennke HG, Brenner BM. Roles for 39. Weisbord SD, Palevsky PM. Radiocontrast-induced acute
thromboxane A2 and leukotrienes in endotoxin-induced acute renal failure. J Intensive Care Med 2005;20:63–75
renal failure. Kidney Int 1986;30:474–480 40. Barrett BJ, Carlisle EJ. Meta-analysis of the relative
21. Kikeri D, Pennell JP, Hwang KH, Jacob AI, Richman AV, nephrotoxicity of high- and low-osmolality iodinated contrast
Bourgoignie JJ. Endotoxemic acute renal failure in awake rats. media. Radiology 1993;188:171–178
Am J Physiol 1986;250:F1098–F1106 41. Rudnick MR, Goldfarb S, Wexler L, et al. Nephrotoxicity of
22. Brenner M, Schaer GL, Mallory DL, Suffredini AF, Parrillo ionic and nonionic contrast media in 1196 patients: a
JE. Detection of renal blood flow abnormalities in septic and randomized trial. The Iohexol Cooperative Study. Kidney
critically ill patients using a newly designed indwelling Int 1995;47:254–261
thermodilution renal vein catheter. Chest 1990;98:170– 42. Aspelin P, Aubry P, Fransson SG, Strasser R, Willenbrock R,
179 Berg KJ. Nephrotoxic effects in high-risk patients undergoing
23. Knotek M, Rogachev B, Wang W, et al. Endotoxemic renal angiography. N Engl J Med 2003;348:491–499
failure in mice: role of tumor necrosis factor independent of 43. Bartholomew BA, Harjai KJ, Dukkipati S, et al. Impact of
inducible nitric oxide synthase. Kidney Int 2001;59:2243– nephropathy after percutaneous coronary intervention and a
2249 method for risk stratification. Am J Cardiol 2004;93:1515–
24. Cunningham PN, Dyanov HM, Park P, Wang J, Newell KA, 1519
Quigg RJ. Acute renal failure in endotoxemia is caused by 44. Marenzi G, Lauri G, Assanelli E, et al. Contrast-induced
TNF acting directly on TNF receptor-1 in kidney. J Immunol nephropathy in patients undergoing primary angioplasty for
2002;168:5817–5823 acute myocardial infarction. J Am Coll Cardiol 2004;44:
25. Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and safety 1780–1785
of recombinant human activated protein C for severe sepsis. 45. Ron D, Taitelman U, Michaelson M, Bar-Joseph G, Bursztein
N Engl J Med 2001;344:699–709 S, Better OS. Prevention of acute renal failure in traumatic
26. Allen DA, Harwood S, Varagunam M, Raftery MJ, Yaqoob rhabdomyolysis. Arch Intern Med 1984;144:277–280
MM. High glucose-induced oxidative stress causes apoptosis 46. Gunal AI, Celiker H, Dogukan A, et al. Early and vigorous
in proximal tubular epithelial cells and is mediated by multiple fluid resuscitation prevents acute renal failure in the crush
caspases. FASEB J 2003;17:908–910 victims of catastrophic earthquakes. J Am Soc Nephrol 2004;
27. Wan L, Bellomo R, Di Giantomasso D, Ronco C. The 15:1862–1867
pathogenesis of septic acute renal failure. Curr Opin Crit Care 47. Zager RA. Studies of mechanisms and protective maneuvers
2003;9:496–502 in myoglobinuric acute renal injury. Lab Invest 1989;60:619–
28. Molitoris BA, Sutton TA. Endothelial injury and dysfunc- 629
tion: role in the extension phase of acute renal failure. Kidney 48. Zager RA. Combined mannitol and deferoxamine therapy for
Int 2004;66:496–499 myohemoglobinuric renal injury and oxidant tubular stress:
29. Ysebaert DK, De Greef KE, De Beuf A, et al. T cells as mechanistic and therapeutic implications. J Clin Invest 1992;
mediators in renal ischemia/reperfusion injury. Kidney Int 90:711–719
2004;66:491–496 49. Brown CV, Rhee P, Chan L, Evans K, Demetriades D,
30. Friedewald JJ, Rabb H. Inflammatory cells in ischemic acute Velmahos GC. Preventing renal failure in patients with
renal failure. Kidney Int 2004;66:486–491 rhabdomyolysis: do bicarbonate and mannitol make a differ-
31. Bonventre JV, Zuk A. Ischemic acute renal failure: an ence? J Trauma 2004;56:1191–1196
inflammatory disease? Kidney Int 2004;66:480–485 50. Tuttle KR, Worrall NK, Dahlstrom LR, Nandagopal R,
32. Solomon R, Werner C, Mann D, D’Elia J, Silva P. Effects of Kausz AT, Davis CL. Predictors of ARF after cardiac surgical
saline, mannitol, and furosemide to prevent acute decreases in procedures. Am J Kidney Dis 2003;41:76–83
renal function induced by radiocontrast agents. N Engl J Med 51. Lassnigg A, Schmidlin D, Mouhieddine M, et al. Minimal
1994;331:1416–1420 changes of serum creatinine predict prognosis in patients after
33. Mueller C, Buerkle G, Buettner HJ, et al. Prevention of cardiothoracic surgery: a prospective cohort study. J Am Soc
contrast media-associated nephropathy: randomized compar- Nephrol 2004;15:1597–1605
ison of 2 hydration regimens in 1620 patients undergoing 52. Thakar CV, Arrigain S, Worley S, Yared JP, Paganini EP. A
coronary angioplasty. Arch Intern Med 2002;162:329–336 clinical score to predict acute renal failure after cardiac surgery.
34. Merten GJ, Burgess WP, Gray LV, et al. Prevention of J Am Soc Nephrol 2005;16:162–168
contrast-induced nephropathy with sodium bicarbonate: a 53. Lassnigg A, Donner E, Grubhofer G, Presterl E, Druml W,
randomized controlled trial. JAMA 2004;291:2328–2334 Hiesmayr M. Lack of renoprotective effects of dopamine and
35. Tepel M, van der Giet M, Schwarzfeld C, Laufer U, furosemide during cardiac surgery. J Am Soc Nephrol
Liermann D, Zidek W. Prevention of radiographic-con- 2000;11:97–104
trast-agent-induced reductions in renal function by acetylcys- 54. Sward K, Valsson F, Odencrants P, Samuelsson O, Ricksten
teine. N Engl J Med 2000;343:180–184 SE. Recombinant human atrial natriuretic peptide in ischemic
36. Pannu N, Manns B, Lee H, Tonelli M. Systematic review of acute renal failure: a randomized placebo-controlled trial. Crit
the impact of N-acetylcysteine on contrast nephropathy. Care Med 2004;32:1310–1315
Kidney Int 2004;65:1366–1374 55. Van Belleghem Y, Caes F, Maene L, Van Overbeke H,
37. Nallamothu BK, Shojania KG, Saint S, et al. Is acetylcysteine Moerman A, Van Nooten G. Off-pump coronary surgery:
effective in preventing contrast-related nephropathy? A meta- surgical strategy for the high-risk patient. Cardiovasc Surg
analysis. Am J Med 2004;117:938–947 2003;11:75–79
38. Bagshaw SM, Ghali WA. Theophylline for prevention of 56. Ascione R, Nason G, Al-Ruzzeh S, Ko C, Ciulli F,
contrast-induced nephropathy: a systematic review and meta- Angelini GD. Coronary revascularization with or without
analysis. Arch Intern Med 2005;165:1087–1093 cardiopulmonary bypass in patients with preoperative
nondialysis-dependent renal insufficiency. Ann Thorac 75. Kleinknecht D, Jungers P, Chanard J, Barbanel C, Ganeval
Surg 2001;72:2020–2025 D. Uremic and non-uremic complications in acute renal
57. Schwann NM, Horrow JC, Strong MD III, Chamchad D, failure: evaluation of early and frequent dialysis on prognosis.
Guerraty A, Wechsler AS. Does off-pump coronary artery Kidney Int 1972;1:190–196
bypass reduce the incidence of clinically evident renal 76. Conger JD. A controlled evaluation of prophylactic dialysis in
dysfunction after multivessel myocardial revascularization? post-traumatic acute renal failure. J Trauma 1975;15:1056–
Anesth Analg 2004;99:959–964 1063
58. Weisberg LS, Kurnik PB, Kurnik BR. Dopamine and renal 76a. Gillum DM, Dixon BM, Yanover MJ, et al. The role of
blood flow in radiocontrast-induced nephropathy in humans. intensive dialysis in acute renal failure. Clin Nephrol 1986;25:
Ren Fail 1993;15:61–68 249–255
59. Olsen NV, Olsen MH, Bonde J, et al. Dopamine natriuresis 77. Gettings LG, Reynolds HN, Scalea T. Outcome in post-
in salt-repleted, water-loaded humans: a dose-response study. traumatic acute renal failure when continuous renal replace-
Br J Clin Pharmacol 1997;43:509–520 ment therapy is applied early vs. late. Intensive Care Med
60. Bellomo R, Chapman M, Finfer S, Hickling K, Myburgh J. 1999;25:805–813
Low-dose dopamine in patients with early renal dysfunction: a 78. Elahi MM, Lim MY, Joseph RN, Dhannapuneni RR, Spyt
placebo-controlled randomised trial. Australian and New TJ. Early hemofiltration improves survival in post-cardiotomy
Zealand Intensive Care Society (ANZICS) Clinical Trials patients with acute renal failure. Eur J Cardiothorac Surg
Group. Lancet 2000;356:2139–2143 2004;26:1027–1031
61. Friedrich JO, Adhikari N, Herridge MS, Beyene J. Meta- 79. Bouman CS, Oudemans-Van Straaten HM, Tijssen JG,
analysis: low-dose dopamine increases urine output but does Zandstra DF, Kesecioglu J. Effects of early high-volume
not prevent renal dysfunction or death. Ann Intern Med continuous venovenous hemofiltration on survival and
2005;142:510–524 recovery of renal function in intensive care patients with
62. Tumlin JA, Finkel KW, Murray PT, Samuels J, Cotsonis G, acute renal failure: a prospective, randomized trial. Crit Care
Shaw AD. Fenoldopam mesylate in early acute tubular Med 2002;30:2205–2211
necrosis: a randomized, double-blind, placebo-controlled 80. Bellomo R, Ronco C, Mehta R. Nomenclature for continuous
clinical trial. Am J Kidney Dis 2005;46:26–34 renal replacement therapies. Am J Kidney Dis 1996;28(suppl 3):
63. Borirakchanyavat V, Vongsthongsri M, Sitprija V. Furose- S2–S7
mide and acute renal failure. Postgrad Med J 1978;54:30– 81. Ronco C, Bellomo R. Continuous renal replacement therapy:
32 evolution in technology and current nomenclature. Kidney Int
64. Brown CB, Ogg CS, Cameron JS. High dose frusemide in 1998;53(suppl 66):S160–S164
acute renal failure: a controlled trial. Clin Nephrol 1981;15: 82. Bellomo R, Parkin G, Love J, Boyce N. A prospective
90–96 comparative study of continuous arteriovenous hemodiafiltra-
65. Mehta RL, Pascual MT, Soroko S, Chertow GM. Diuretics, tion and continuous venovenous hemodiafiltration in critically
mortality, and nonrecovery of renal function in acute renal ill patients. Am J Kidney Dis 1993;21:400–404
failure. JAMA 2002;288:2547–2553 83. Kellum JA, Johnson JP, Kramer D, Palevsky P, Brady JJ,
66. Uchino S, Doig GS, Bellomo R, et al. Diuretics and mortality Pinsky MR. Diffusive vs. convective therapy: effects on
in acute renal failure. Crit Care Med 2004;32:1669–1677 mediators of inflammation in patient with severe systemic
67. Liangos O, Rao M, Balakrishnan VS, Pereira BJ, Jaber BL. inflammatory response syndrome. Crit Care Med 1998;26:
Relationship of urine output to dialysis initiation and mortality 1995–2000
in acute renal failure. Nephron Clin Pract 2005;99:c56–c60 84. Swartz RD, Messana JM, Orzol S, Port FK. Comparing
68. Hirschberg R, Kopple J, Lipsett P, et al. Multicenter clinical continuous hemofiltration with hemodialysis in patients with
trial of recombinant human insulin-like growth factor I in severe acute renal failure. Am J Kidney Dis 1999;34:424–432
patients with acute renal failure. Kidney Int 1999;55:2423– 85. Guérin C, Girard R, Selli JM, Ayzac L. Intermittent versus
2432 continuous renal replacement therapy for acute renal failure in
69. Acker CG, Flick R, Shapiro R, et al. Thyroid hormone in the intensive care units: results from a multicenter prospective
treatment of post-transplant acute tubular necrosis (ATN). epidemiological survey. Intensive Care Med 2002;28:1411–
Am J Transplant 2002;2:57–61 1418
70. Acker CG, Singh AR, Flick RP, Bernardini J, Greenberg A, 86. Mehta RL, McDonald B, Gabbai FB, et al. A randomized
Johnson JP. A trial of thyroxine in acute renal failure. Kidney clinical trial of continuous versus intermittent dialysis for
Int 2000;57:293–298 acute renal failure. Kidney Int 2001;60:1154–1163
71. Teschan PE, Baxter CR, O’Brien TF, Freyhof JN, Hall WH. 87. Kellum JA, Angus DC, Johnson JP, et al. Continuous versus
Prophylactic hemodialysis in the treatment of acute renal intermittent renal replacement therapy: a meta-analysis.
failure. Ann Intern Med 1960;53:992–1016 Intensive Care Med 2002;28:29–37
72. Easterling RE, Forland M. A five year experience with 88. Tonelli M, Manns B, Feller-Kopman D. Acute renal failure
prophylactic dialysis for acute renal failure. Trans Am Soc in the intensive care unit: a systematic review of the impact of
Artif Intern Organs 1964;10:200–208 dialytic modality on mortality and renal recovery. Am J
73. Parsons FM, Hobson SM, Blagg CR, McCracken CB. Kidney Dis 2002;40:875–885
Optimum time for dialysis in acute reversible renal failure: 89. Augustine JJ, Sandy D, Seifert TH, Paganini EP. A
description and value of an improved dialyser with large randomized controlled trial comparing intermittent with
surface area. Lancet 1961;1:129–134 continuous dialysis in patients with ARF. Am J Kidney Dis
74. Fischer RP, Griffen WO Jr, Reiser M, Clark DS. Early 2004;44:1000–1007
dialysis in the treatment of acute renal failure. Surg Gynecol 90. Manns B, Doig CJ, Lee H, et al. Cost of acute renal failure
Obstet 1966;123:1019–1023 requiring dialysis in the intensive care unit: clinical and
resource implications of renal recovery. Crit Care Med 2003; 95. Schiffl H, Lang SM, Fischer R. Daily hemodialysis and
31:449–455 the outcome of acute renal failure. N Engl J Med 2002;346:
91. Jacka MJ, Ivancinova X, Gibney RT. Continuous renal 305–310
replacement therapy improves renal recovery from acute renal 96. Bonventre JV. Daily hemodialysis: will treatment each day
failure. Can J Anaesth 2005;52:327–332 improve the outcome in patients with acute renal failure?
92. Palevsky P, Baldwin I, Davenport A, Goldstein S, Paganini E. N Engl J Med 2002;346:362–364
Renal replacement therapy and the kidney: minimizing the 97. Paganini EP, Taployai M, Gormastic M, et al. Establishing
impact of renal replacement therapy on recovery of acute renal a dialysis therapy/patient outcome link in intensive care unit
failure. Curr Opin Crit Care 2005;11, In press acute dialysis. Am J Kidney Dis 1996;28(suppl 3):S81–S89
93. Phu NH, Hien TT, Mai NT, et al. Hemofiltration and 98. Kellum JA, Mehta RL, Angus DC, Palevsky P, Ronco C. The
peritoneal dialysis in infection-associated acute renal failure in first international consensus conference on continuous renal
Vietnam. N Engl J Med 2002;347:895–902 replacement therapy. Kidney Int 2002;62:1855–1863
94. Kielstein JT, Kretschmer U, Ernst T, et al. Efficacy and 99. Palevsky PM, O’Connor T, Zhang JH, Star RA, Smith MW.
cardiovascular tolerability of extended dialysis in critically ill Design of the VA/NIH Acute Renal Failure Trial Network
patients: a randomized controlled study. Am J Kidney Dis (ATN) Study: Intensive versus conventional renal support in
2004;43:342–349 acute renal failure. Clin Trials 2005;2:423–435
Stress Hyperglycemia and Adrenal
Insufficiency in the Critically Ill
Murugan Raghavan, M.D., M.R.C.P. (UK)1 and
Paul E. Marik, M.D., F.R.C.P. (C), F.C.C.P.2
ABSTRACT
Critical illness evoked by trauma, extensive surgery, or severe medical illnesses is

the ultimate example of acute severe physical stress. The endocrine response in a critically
injured and stressed patient is varied and complex. Although the acute and chronic phases of
critical illness are characterized by distinct endocrine responses, the diagnosis of these
disorders is controversial. The inability to define the endocrine change as either adaptation
or pathology renders the issue of treatment even more controversial. In addition, patients
may have preexisting endocrine diseases, either previously diagnosed or unknown, and hence
endocrine evaluation in a critically ill patient poses a major challenge to the health care
provider. This review provides a novel insight into the dynamic endocrine alterations that
occur during evolution of stress hyperglycemia and adrenal insufficiency in the critically ill
patient and the available evidence for the therapy of these disorders.
KEYWORDS: Stress hyperglycemia, adrenal insufficiency, ICU, critically ill, therapy,

mortality
S tress hyperglycemia and adrenal insufficiency reported incidence of A-1 varies widely (0 to 77%) in the
are the most common endocrine disorders during critical ICU depending on the population of patients studied and
illness that impact survival. An increasingly robust body the diagnostic criteria used.9,10 However, the overall
of literature describes adverse clinical outcomes associ- incidence of A-1 in critically ill patients approximates
ated with hyperglycemia in critically ill patients.1–7 A 30%, with an incidence as high as 50 to 60% in patients
recent review of 1826 consecutive patients admitted to a with septic shock.10 It is important to recognize these
medical-surgical intensive care unit (ICU) showed that patients because this disorder has a high mortality rate if
hospital mortality was strongly associated with glycemic untreated.11 Therapy of patients with relative A-1 with
control during ICU stay. Patients with mean glucose corticosteroids has been shown to increase survival.10
levels between 80 and 99 mg/dL during ICU stay had a
9.6% hospital mortality; this increased to 12.5% among
patients with a mean glucose level of 100 to 119 mg/dL ENDOCRINOLOGY OF STRESS
and was as high as 42.5% in patients whose mean glucose Stress associated with critical illness is characterized by
level exceeded 300 mg/dL.8 activation of the hypothalamic-pituitary-adrenal (HPA)
The second most common endocrine disorder axis with the release of cortisol from the adrenal
during critical illness is adrenal insufficiency (A-1). The gland.9,12 This phenomenon is an essential component
1
Department of Critical Care Medicine, University of Pittsburgh 19107. E-mail: [email protected].
Medical Center, Pittsburgh, Pennsylvania; 2Division of Pulmonary Non-pulmonary Critical Care: Managing Multisystem Critical
and Critical Care Medicine, Thomas Jefferson University, Philadel- Illness; Guest Editor, Curtis N. Sessler, M.D.
phia, Pennsylvania. Semin Respir Crit Care Med 2006;27:274–285. Copyright # 2006
Address for correspondence and reprint requests: Paul E. Marik, by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
M.D., Division of Pulmonary and Critical Care Medicine, Thomas NY 10001, USA. Tel: +1(212) 584-4662.
Jefferson University, 834 Walnut St., Ste. 650, Philadelphia, PA DOI 10.1055/s-2006-945533. ISSN 1069-3424.
274
ENDOCRINE DISORDERS/RAGHAVAN, MARIK 275
of general adaptation to stress and contributes to the Stress hyperglycemia but not preexisting diabetes has
maintenance of cellular and organ homeostasis. In addi- been shown to be associated with a worse outcome
tion to increased cortisol secretion the stress response is following acute myocardial infarction and stroke.22–25
characterized by a marked increase in the release of The plasma glucose level on admission has been shown
norepinephrine, epinephrine, as well as glucagon and to be an independent predictor of prognosis after
growth hormone.13–15 Insulin levels are usually normal myocardial infarction.22,23 In diabetic patients with
or decreased, despite peripheral insulin resistance.16–18 It acute myocardial infarction, therapy to maintain blood
has been suggested that insulin release may be sup- glucose at a level below 215 mg/dL improves out-
pressed as the result of increased activation of the come.24,26,27 The presence of hyperglycemia following
pancreatic a receptors.17 In addition to causing insulin an ischemic or hemorrhagic stroke is associated with a
resistance, cytokines such as interleukin-1 (IL-1) and two to threefold increased mortality and significant
tumor necrosis factor (TNF) inhibit insulin release, an impairment in functional recovery.25,28
effect that appears to be concentration dependent.19 The Glucose has been shown to be a powerful proin-
low to normal insulin levels together with insulin resist- flammatory mediator,29,30 and tight glycemic control
ance in the presence of increased secretion of the below 110 mg/dL with insulin has been shown to exert
counterregulatory hormones results in stress hypergly- antiinflammatory effects in the critically ill patient.31
cemia. An inability to respond to stress by mounting an Glucose also has been shown to exert prothrombotic
appropriate cortisol response or cortisol resistance results effects and to increase oxidative stress due to increased
in clinical adrenal insufficiency syndrome (Fig. 1). lipid peroxidation.32 In critically ill surgical and burn
patients tight glycemic control has been demonstrated to
reduce the risk of sepsis-related morbidity.1,33–35 The in
EPIDEMIOLOGY OF STRESS vitro responsiveness of leukocytes stimulated by inflam-
HYPERGLYCEMIA matory mediators has been shown to inversely correlate
The prevalence of stress hyperglycemia during critical with glycemic control.36,37
illness is difficult to establish due to limited data and In critically ill patients, hyperglycemia has been
variations in the definition of hyperglycemia. Stress shown to cause deleterious effects on mitochondrial
hyperglycemia has been previously defined as plasma oxidation, whereas intensive insulin therapy (blood glu-
glucose above 200 mg/dL20; however, in view of the cose < 110 mg/dL) contributes to mitochondrial integ-
results of the Leuven Intensive Insulin Therapy Trial rity.38 Vanhorebeek et al showed that in seven of nine
(see later discussion), stress hyperglycemia should be patients who died in the ICU with hyperglycemia,
considered in any critically ill patient with a blood hepatic mitochondria exhibited hypertrophic abnormal
glucose in excess of 110 mg/dL.1 In a study of septic irregular cristae and reduced matrix electron density.
nondiabetic ICU patients 75% had a baseline blood However, only one of 11 patients given intensive insulin
glucose level above 110 mg/dL.21 In the Leuven Inten- treatment had these morphological abnormalities
sive Insulin Therapy Trial, 12% of patients had a base- (p ¼ .005). This effect on ultrastructure was associated
line blood glucose above 200 mg/dL; however, 74.5% of with higher activity of respiratory-chain complex I and
patients had a baseline blood glucose above 110 mg/dL, complex IV in the intensive group than in the conven-
with 97.5% having a recorded blood glucose level above tional group, suggesting a direct link between hyper-
110 mg/dL sometime during their ICU stay.1 glycemia and cellular energetic dysfunction. Strict
The metabolic milieu in which stress-induced glycemic control with intensive insulin therapy pre-
hyperglycemia develops in the critically ill in the absence vented or reversed ultrastructural and functional abnor-
of preexisting diabetes mellitus is complex. A combina- malities of hepatocyte mitochondria and maintained
tion of several factors, including the presence of excessive cellular integrity.38
counterregulatory hormones such as glucagon, growth
hormone, catecholamines, glucocorticoids, and cyto-
kines such as IL-1, IL-6, and TNF combined with BENEFICIAL IMMUNE-MODULATORY
exogenous administration of catecholamines, dextrose, ROLE OF INSULIN DURING CRITICAL
and nutritional support together with relative insulin ILLNESS
deficiency, play an important role.12 Besides control of hyperglycemia, insulin has potent
acute antiinflammatory effects. In a group of obese
subjects, Dandona and colleagues demonstrated that an
DELETERIOUS EFFECTS OF STRESS infusion of insulin was associated with a significant fall in
HYPERGLYCEMIA IN THE CRITICALLY ILL proinflammatory transcription factor nuclear factor
To some extent the deleterious effects of hyperglyce- kappa-B (NF-kB), and increase in inhibitory kappa-B
mia in the critically ill are similar to that of actual (IkB) in mononuclear cells.39 Intensive insulin therapy
diabetes, although the time scale obviously differs. has been shown to reduce endothelial activation and
Figure 1 Pathogenesis of stress hyperglycemia and adrenal insufficiency in critically ill. Stress induced by various stimuli results in
release of CRH (corticotropin releasing hormone) from hypothalamus and adrenal corticotropic hormone (ACTH) from pituitary. ACTH
stimulates cortisol release from the adrenal gland. Various cytokines inhibit cortisol synthesis and release through the hypothalamic-
pituitary-adrenal (HPA) axis modulation. In addition epinephrine, norepinephrine, and glucagon, along with cytokines and endotoxins,
inhibit insulin release, induce insulin resistance, and promote hepatic gluconeogenesis, thereby contributing to stress hyperglycemia.
Cytokine-induced inhibition of high density lipoprotein (HDL) cholesterol and CLA-1 (lysosomal integral membrane protein-II analogous
1) receptor synthesis results in decreased delivery of HDL to the adrenal gland, thereby causing substrate deficiency for cortisol
synthesis, resulting in adrenal insufficiency (A-1) (hepatoadrenal syndrome). Decreased release of cortisol coupled with increased tissue
resistance contributes to clinical A-1 syndrome. A-1, adrenal insufficiency; ACTH, adrenocorticotrophic hormone; Apo-1, apolipoprotein,
CRH, corticotropin; IL, interleukin; TNF-a, tumor necrosis factor a; TGF, tumor growth factor; HDL, high density lipoprotein cholesterol;
CLA-1, scavenger receptor for HDL in liver and adrenals.
end-organ dysfunction.40 Endothelial-derived adhesion Intensive insulin therapy lowers ICAM-1 and E-selectin
molecules such as intracellular adhesion molecule-1 levels in critically ill patients, thereby reducing endothe-
(ICAM-1) and E-selectin are expressed in patients lial dysfunction.40 Insulin therapy is also a powerful
with prolonged critical illness and in nonsurvivors.41,42 anabolic stimulus to promote protein synthesis during
critical illness thereby counteracting the protein catabo- of bacteremia, acute renal failure, critical illness poly-
lism that occurs during critical illness.43–45 neuropathy, and transfusion requirements, a blood
glucose level of less than 110 mg/dL was required.
Indeed, a blood glucose level of 110–150 mg/dL was
INTENSIVE INSULIN THERAPY IN THE not effective on these morbidity measures as compared
CRITICALLY ILL with > 150 mg/dL.46
Van Den Berghe et al, in a prospective, randomized, Krinsley and Grissler,47evaluated an intensive
controlled study involving 1548 patients, demonstrated glucose management protocol in 800 heterogeneous
that intensive insulin therapy reduced mortality and critically ill adult patients. The protocol involved in-
morbidity among patients admitted to a surgical critical tensive monitoring and treatment to maintain plasma
care unit (the Leuven Intensive Insulin Therapy glucose values lower than 140 mg/dL. Continuous
Trial).1,46 These authors compared an intensive insulin intravenous insulin was used if glucose values exceeded
therapy regimen aimed to maintain blood glucose be- 200 mg/dL. The mean glucose value decreased
tween 80 and 110 mg/dL with conventional treatment in from 152.3 to 130.7 mg/dL (p < .001), marked by a
which insulin infusion was only initiated when glucose 56.3% reduction in the percentage of glucose values of
level was greater than 215 mg/dL and maintenance of 200 mg/dL or higher, without a significant change in
glucose between 180 and 200 mg/dL. At 12 months the incidence of hypoglycemia. The development of new
mortality was 4.6% with the intensive insulin regimen renal insufficiency decreased by 75% (p ¼ 0.03), and the
compared with 8.0% in the control group. The benefit number of patients undergoing transfusion of packed
was most apparent in patients with greater than 5 days of red blood cells decreased by 18.7% (p ¼ .04). Hospital
stay in the intensive care unit. Tight and early glycemic mortality decreased by 29.3% (p ¼ .002), and length
control was associated with the more rapid improvement of stay in the ICU decreased by 10.8% (p ¼ .01). In
of insulin resistance.46 Intensive insulin therapy was addition, intensive insulin therapy was shown to cause a
associated with reduced bloodstream infections by significant reduction in the incidence of total nosoco-
46%, acute renal failure by 41%, and critical illness mial infections, including intravascular device, blood-
polyneuropathy by 44%. Using multivariate analysis the stream, intravascular device–related bloodstream, and
authors suggested that improved metabolic control, as surgical site infections.
reflected by normoglycemia, rather than the infused Grey and Perdrizet48 randomized 61 surgical
insulin dose per se, was responsible for the beneficial ICU patients requiring treatment of hyperglycemia
effects of intensive insulin therapy. However, achieving (glucose values > or ¼ 140 mg/dL) to receive either
normoglycemia and the administration of insulin are standard insulin therapy (target glucose range, 180 to
linked, and from the available evidence it appears likely 220 mg/dL) or strict insulin therapy (target glucose
that both factors played a key role in the improved range, 80 to 120 mg/dL) throughout their ICU stay. A
outcome. significant reduction (p < .001) in mean daily glucose
The outcome data from the Leuven Intensive level was achieved in the strict glycemic control group
Insulin Therapy Trial indicates that there is a direct (125 36 mg/dL) in comparison with the standard
relationship between the degree of glycemic control and glycemic control group (179 61 mg/dL). A significant
hospital mortality.46 In the long-stay patients (> 5 days reduction (p < .05) in the incidence of total nosocomial
in the ICU) the cumulative hospital mortality was 15% infections, including intravascular device, bloodstream,
in patients with a mean blood glucose less than 110 mg/ and surgical wound infections, was observed in the strict
dL, 25% in those with a blood glucose between 110 and glucose control group in comparison with the standard
150 mg/dL, and 40% in those with a mean blood glucose glucose control group.
of greater than 150 mg/dL. In diabetic patients with It is noteworthy that in the Leuven Intensive
acute myocardial infarction, therapy to maintain blood Insulin Therapy Trial, all patients received between
glucose at a level below 215 mg/dL improves out- 200 and 300 g of intravenous glucose on the day of
come.24,26,27 These data suggest that even ‘‘modest’’ admission followed by parenteral or enteral (or both)
glycemic control will have an impact on patient outcome. nutrition started on the second ICU day. However,
This is important because in the ‘‘real world’’ it may be although early enteral feeding has been reported to
difficult (if not somewhat risky) to attempt to maintain improve organ function and decrease the length of
blood glucose in the range of 80 to 110 mg/dL. This hospital stay,49 parenteral nutrition is associated with
often requires the use of a continuous insulin infusion adverse outcomes during critical illness.50 Furthermore,
protocol and frequent blood glucose monitoring. How- hypocaloric enteral nutrition administered with slowly
ever, this goal may only be achievable in ICUs with a absorbed carbohydrate induces less hyperglycemia than
high nursing to patient ratio and close physician super- parenteral nutrition among critically ill.51–53
vision. On the other hand, the Leuven study showed Based on the foregoing results we recommend the
that to improve morbidity by reducing the incidence initiation of early enteral nutrition in all ICU patients on
the day of ICU admission.49,50,54 Enteral nutrition Table 1 Etiology of Adrenal Insufficiency during
should be commenced at a rate of 33 to 66% of calculated Critical Illness
intake (15 to 20 kcal/kg/d) and advanced to full calorific Syndromes Mechanism of A-1
goal of 20 to 25 kcal/kg/d over 3 to 5 days.55 Insulin
COMMON
infusion should be commenced in patients with blood
Reversible dysfunction of the HPA axis
glucose above 150 mg/dL (a threshold of 110 mg/dL
Sepsis/SIRS Primarya and secondaryb
may be appropriate in select ICUs). Subcutaneous in-
Drugs
sulin ‘‘sliding scales’’ may control stress hyperglycemia.
Corticosteroids Secondary
However, an insulin infusion is recommended if the
Etomidate Primary
blood glucose remains above 150 mg/dL after 24 hours
Ketoconazole Primary
on a sliding scale.
Rifampin Increased cortisol metabolism
Phenytoin Increased cortisol metabolism
ACTH and cortisol resistance Primary and secondary
ADRENAL INSUFFICIENCY IN THE
Adrenal Exhaustion Secondary
CRITICALLY ILL
Hypothermia Primary
In critically ill patients there has been a great deal of
Liver disease (hepatoadrenal
interest regarding the assessment of adrenal function and
syndrome)
the indications for adrenal replacement therapy.9,11,56–58
HDL (apolipoprotein-1) Primary
A-1, although once considered a rare diagnosis in the
deficiency
ICU, is currently being reported with increased fre-
Fulminant hepatic failure Primary and secondary
quency in critically ill patients. Although the exact
Chronic liver failure (cirrhosis) Primary
incidence of A-1 varies with the diagnostic test and
Liver transplantation Primary
concentration of cortisol used to diagnose the disorder,
Anticoagulation Primary and secondary
in one series 61% of critically ill septic shock patients
Heparin-induced Primary and secondary
had A-1 when a baseline cortisol concentration of
thrombocytopenia
< 25 mg/dL was used as the diagnostic threshold.56
Brain dead organ donors Primary and secondary
Adrenal failure can be caused by structural damage to
RARE
the adrenal gland, pituitary gland, or hypothalamus;
Metastatic cancer Primary and secondary
however, many critically ill patients develop reversible
Pituitary diseases Secondary
failure of the HPA axis.9
HIV infection Primary and secondary
Granulomatous diseases Primary and secondary
HYPOTHALAMO-PITUITARY AXIS AND ACTH, adrenal corticotropic hormone; A-1, adrenal Insufficiency;
HDL, high density lipoprotein; HIV, human immunodeficiency virus;
CORTISOL DURING STRESS HPA, hypothalamic-pituitary-adrenal; SIRS, systemic inflammatory
Severe illness and stress activate the HPA axis and response syndrome.
a
Primary A-1 is defined as the failure of the adrenal gland to produce
stimulate the release of corticotropin [also known as cortisol.9
b
adrenal corticotropic hormone (ACTH)] from the pi- Secondary A-1 is defined as adrenal failure secondary to hypo-
thalmo-pituitary-axis dysfunction.9,61
tuitary, which in turn increases the release of cortisol
from the adrenal cortex. This activation is an essential
component of the general adaptation to illness and stress, patients.56,59–62 However, the most common cause of
and contributes to the maintenance of cellular and organ acute A-1 is sepsis and systemic inflammatory response
homeostasis. syndrome (SIRS)56,63,64 (Table 1).
CAUSES OF ADRENAL INSUFFICIENCY IN

THE INTENSIVE CARE UNIT PATHOPHYSIOLOGY OF A-1 DURING
Acute A-1 occurs in patients who are unable to increase CRITICAL ILLNESS
their production of cortisol during acute stress. This
includes patients with hypothalamic and pituitary dis- Sepsis and Systemic Inflammatory Response
orders (secondary A-1) and patients with destructive Syndrome–Induced Acute Reversible Adrenal
diseases of the adrenal glands (primary A-1) (Table 1). Insufficiency
Secondary A-1 is common in patients who have been There is increasing evidence of HPA insufficiency in
treated with exogenous corticosteroids. Increasingly A-1 critically ill septic patients,56,65 which appears to result
is being reported in patients with sepsis, human immu- from circulating suppressive factors released during sys-
nodeficiency virus infection, acute and subacute liver temic inflammation.66 It is important to recognize these
failure, brain-dead organ donors, and cardiac surgery patients because this disorder has a high mortality rate if
untreated.67 In our series of 59 patients with septic hospitalization. The baseline serum cortisol was
shock, 15 patients (25%) had primary A-1, 10 patients 18.8 16.2 mg/dL in the nonsurvivors compared with
(17%) had HPA-axis failure, and 11 patients (19%) had 13.0 11.8 mg/dL in the survivors (p < .001). Of those
ACTH resistance.56 Surviving septic patients had return patients with adrenal failure who were treated with
of adrenal function and did not require long-term treat- glucocorticoids, the mortality rate was 26% compared
ment with corticosteroids. with 46% (p < .002) in those who were not treated. In
those patients receiving vasopressor agents at the time of
adrenal testing, the baseline cortisol was 10.0 4.8 mg/
Adrenocorticotropin and Cortisol Resistance dL in those with A-1 compared with 35.6 21.2 mg/dL
Patients with systemic infections [e.g., sepsis, human in those with normal adrenal function. Vasopressor-
immunodeficiency virus (HIV)] may acquire A-1 asso- dependent patients who did not have adrenal failure
ciated with resistance to ACTH. In two recent studies in had a mortality rate of 75%.72
critically ill patients, we found that 30% of patients with
septic shock and 25% of critically ill, HIV-infected
patients acquired A-1 associated with ACTH resist- High Density Lipoprotein Deficiency
ance.59,68 In these patients pharmacological doses of and Adrenal Insufficiency
exogenous corticotropin did not increase their serum A-1 is increasingly being reported in patients with acute
cortisol levels, but high doses of corticotropin were able and subacute liver failure.60,72–75 The finding of an
to increase the levels into the normal range suggesting association between low serum apolipoprotein A-1
corticotropin resistance. (Apo-1) in patients with hepatic failure and A-1 sup-
Ali and colleagues reported a 40% decline in the ports the notion that liver disease may lead to impaired
number of glucocorticoid receptors (GRs) in the liver of cortisol synthesis.74–76 Apo-1 is the major protein com-
septic rats.69 The decline in hormone-binding activity ponent of HDL cholesterol synthesized principally by
was associated with a fall in GR messenger ribonucleic the liver. Experimental studies suggest that HDL is the
acid (mRNA). Decreased affinity of the GR from preferred lipoprotein source of steroidogenic substrate in
mononuclear leukocytes of patients with sepsis has the adrenal gland.76 At rest and during stress, 80% of
also been reported.70 In addition, Norbiato et al recirculating cortisol is derived from plasma cholesterol,
ported resistance to glucocorticoids in patients with the remaining 20% being synthesized in situ from acetate
acquired immunodeficiency syndrome (AIDS).68 Cor- and other precursors.77
tisol-resistant patients had clinical evidence of A-1 Recently, mouse scavenger receptor, class B, type
associated with decreased affinity of GRs for glucocor- 1 (SR-B1) and its human homologue (CLA-1) were
ticoids and decreased GR function. We as well as others identified as the high affinity HDL receptor mediating
have found that cortisol clearance from the circulation selective cholesterol uptake.78 The receptor for HDL
is impaired in many critically ill patients.71 This de- (CLA-1) is expressed at high levels in the parenchymal
creased clearance reflects decreased tissue uptake and cells of the liver and the steroidogenic cells of the adrenal
metabolism of cortisol. glands, ovaries, and testes. CLA-1 mRNA is highly
expressed in human adrenal glands, and the accumula-
tion of CLA-1 messenger RNA is upregulated by
Liver Failure–Associated Adrenal Insufficiency adrenocorticotropin in primary cultures of normal hu-
(the ‘‘Hepatoadrenal’’ Syndrome) man adrenocortical cells.77 Low Apo-1/HDL levels in
We have found a high incidence of adrenal failure in the critically ill may be pathogenetically linked to the
critically ill patients with liver disease, an entity for high incidence of adrenal failure in this group of
which we have coined the term hepatoadrenal syndrome.72 patients. Van der Voort and colleagues79 demonstrated
In our study of 245 patients with hepatoadrenal syn- that in critically ill patients, low HDL levels were
drome, high density lipoprotein (HDL) level at the time associated with an attenuated response to Synacthen.
of adrenal testing was the only variable predictive of Indeed, an inverse relationship noted between proin-
adrenal insufficiency (p < .0001). In vasopressor-de- flammatory mediators and HDL/Apo-1 levels is asso-
pendent patients with A-1, treatment with hydrocorti- ciated with poor outcome in the critically ill.80 This may
sone was associated with a significant reduction (p < .02) be mediated by low serum cortisol level and A-1,
in the dose of norepinephrine at 24 hours, whereas the suggesting that further studies are required to define
dose of norepinephrine was significantly higher (p < .04) the pathogenetic role and mechanisms of altered HDL/
in those patients with adrenal failure not treated with Apo-1 metabolism in acute illness.74
hydrocortisone. In vasopressor-dependent patients with- In our series of patients with end-stage liver
out A-1, treatment with hydrocortisone did not affect failure, we noted an incidence of adrenal failure in
vasopressor dose at 24 hours. One hundred and forty- 15% of patients with fulminant liver failure, 40 to
one of a total 340 patients (41%) died during their 50% in patients with end-stage liver failure, and 90%
of patients undergoing liver transplantation.72 Because both upregulated or downregulated (tissue resistance)
HDL is synthesized primarily by the liver and plays a during stress.91,92 These data suggest that the total
fundamental role in transporting cholesterol to the circulating cortisol level may be a poor indicator of
adrenal gland, patients with the hepatoadrenal syndrome glucocorticoid activity at the nuclear level. Notwith-
have low HDL levels. In our study, the mean HDL level standing these caveats and the fact that we currently
was 8 mg/dL in patients with hepatoadrenal syndrome, do not have a test that measures glucocorticoid activity,
whereas it was 34 mg/dL (p ¼ .01) in patients with assessment of the HPA axis is usually made on the basis
normal adrenal function. Furthermore, a low HDL level of a random (stress) cortisol level or the corticotropin
at admission to the ICU was predictive of the develop- stimulation test. In a highly stressed patient such as with
ment of adrenal failure (adrenal exhaustion syndrome) in severe sepsis and other shock states a random cortisol
patients who had preserved adrenal function at admis- level assesses the integrity of the entire HPA axis.88
sion. Based on these findings, we suggest the routine Dysfunction at the hypothalamic, pituitary, or adrenal
measurement of a random cortisol and HDL level in all level will result in low circulating cortisol levels
patients with end-stage liver disease and in all critically (< 20 mg/dL). A stress cortisol level of < 20 mg/dL in
ill patients at risk of adrenal failure. a patient with refractory hypotension should be treated
with low-dose (stress dose) hydrocortisone.9 Because
this cutoff is rather arbitrary, a patient with a cortisol
Endotoxemia and Adrenal Insufficiency level greater than 20 mg/dL but less than 35 mg/dL, who
Apart from low HDL levels and the reduced delivery of has refractory hypotension may warrant a trial of low-
substrate for cortisol synthesis, other mechanisms may dose hydrocortisone. It should be emphasized that a
contribute to the pathophysiology of the hepatoadrenal cortisol level of > 20 mg/dl does not exclude A-1 due to
syndrome. Patients with acute and chronic liver disease tissue resistance.
have increased levels of circulating endotoxin [lipopoly- A random cortisol level of < 15 mg/dL in a
saccharide (LPS)] and proinflammatory mediators such moderately stressed (vasopressor-independent) ICU pa-
as TNF.81 It is postulated that intestinal bacterial over- tient is suggestive of HPA dysfunction.57 In moderately
growth with increased bacterial translocation, together stressed patients, ‘‘adrenal reserve’’ can be assessed by the
with reduced Kupffer cell activity and portosystemic low dose (LD; 1 mg) corticotropin (Synacthen) stimula-
shunting results in systemic endotoxemia with increased tion test. A serum cortisol level of < 20 mg/dL
transcription of proinflammatory mediators.82,83 In ad- 30 minutes after an LD corticotropin stimulation test
dition, serum endotoxin levels increase further during is suggestive of primary A-1. It is important to empha-
the anhepatic phase of liver transplantation and remain size that the random and stimulated cortisol levels must
high for several days following transplantation.82 LPS as be interpreted in conjunction with the severity of illness
well as TNF may inhibit cortisol synthesis. Endotoxin and the patient’s clinical features.89 A moderately
has been shown to bind with high affinity to the HDL stressed ICU patient with a random cortisol level of
receptor (CLA-1) with subsequent internalization of the < 15 mg/dL or a stimulated level of < 20 mg/dL who has
receptor.84,85 LPS may therefore limit the delivery of no clinical signs of A-1 (unexplained fever, confusion,
HDL cholesterol to the adrenal gland. Furthermore, hemodynamic instability, or eosinophilia) does not
TNF as well as interleukin-1b and interleukin-6 has warrant treatment with stress doses of hydrocortisone.
been demonstrated to decrease hepatocyte synthesis and Annane et al11 showed that a high baseline cortisol
secretion of Apo-186 (Fig. 1). as well as inability to increase cortisol by 9 mg/dL
(delta cortisol) after a high dose corticotropin (250 mg
tetracosactrin) stimulation test was associated with a
DIAGNOSIS OF HYPOTHALAMIC- worse likelihood of survival. Following this study the
PITUITARY-ADRENAL AXIS FAILURE delta cortisol has become the standard diagnostic test of
Because there are no clinically useful tests to assess the choice to diagnose A-1 in the ICU (see later discussion).
cellular actions of cortisol (i.e., end-organ effects), the The ‘‘delta 9’’ has become the magical number that
diagnosis of A-1 is based on the measurement of serum distinguishes normal adrenal function from ‘‘relative’’
cortisol levels; this has resulted in much confusion and adrenal failure. However, the delta cortisol is a measure
misunderstanding.58,87–90 Circulating cortisol is bound of adrenal reserve and adrenal responsiveness to cortico-
to corticosteroid-binding globulin with < 10% in the tropin; it does not assess the integrity of the HPA axis
free bioavailable form. During acute illness, there is an and is not a measure of adrenal function. In a study by
acute decline in the concentration of corticosteroid- Dimopoulou and coauthors who evaluated the HPA axis
binding globulin as well as decreased binding affinity dysfunction in critically ill patients with traumatic brain
for cortisol, resulting in an increase in the free bio- injury, > 50% of healthy volunteers had a delta cortisol of
logically active fraction of the hormone.65,90 In addition, < 9 mg/dL.93 Similarly, in a study of patients with
the number of intracellular GRs has been reported to be respiratory failure and no evidence of HPA disease,
50% had a delta cortisol of < 9 mg/dL after endotracheal if the pressor agent could be discontinued within
intubation with midazolam anesthesia.94 We believe that 24 hours of the first dose of hydrocortisone.
the standard corticotropin stimulation test lacks sensitiv- Sixty-one percent of patients met the criteria of
ity for the diagnosis of A-1.89 As already discussed, a A-1 when a baseline cortisol concentration of < 25 mg/dL
threshold cortisol level of < 20 g/dL is inappropriately was used. Ninety-five percent of steroid-responsive pa-
low in critically ill patients. ‘‘Normal’’ critically ill patients tients had a baseline cortisol concentration < 25 mg/dL.
should elevate their cortisol level 25 mg/dL. Further- Receiver operating characteristic curve analysis revealed
more, 250 mg of corticotropin is supraphysiological that a stress cortisol concentration of 23.7 mg/dL was the
( 100-fold higher than normal maximal-stress ACTH most accurate diagnostic threshold for determination of
levels).87,95 The very high levels of corticotropin obtained the hemodynamic response to glucocorticoid therapy.
with 250 mg can override adrenal resistance to ACTH The sensitivity of a baseline cortisol < 25 mg/dL in
and result in a normal cortisol response. Therefore the predicting steroid responsiveness was 96%, compared
decision to treat patients with glucocorticoids should be with 54% for the low-dose test and 22% for the high
based on serum cortisol levels in conjunction with the dose test. The specificities of the tests were 57, 97, and
patient’s clinical features and severity of illness. In pa- 100%, respectively. The area under the receiver operating
tients with subtle clinical signs or a borderline random characteristic curve of the stress (baseline) cortisol con-
cortisol level, a therapeutic trial of treatment with stress- centration was 0.84; a stress cortisol concentration of
level doses of glucocorticoids may be warranted. The 23.7 mg/dL had the best discriminating power, with a
potential benefit of treatment with hydrocortisone in sensitivity of 0.86, a specificity of 0.66, a likelihood ratio of
certain patient groups, including those with severe sepsis 2.6, a positive predictive value of 0.62, and a negative
(not septic shock), hemodynamic instability in nonseptic predictive value of 0.88.
patients, severe community acquired pneumonia; patients However, as discussed previously, it is unclear at
treated with etomidate; and patients being weaned from this time whether a threshold of 20 mg/dL or 25 mg/dL
mechanical ventilation, deserve further investigation. should be used to determine treatment with hydro-
cortisone. Due to poor sensitivity of low-dose and
high-dose corticotropin stimulation testing we recom-
THERAPY OF ADRENAL INSUFFICIENCY mend that these tests be avoided in severely stressed,
During septic shock, treatment with stress-level doses of vasopressor-dependant septic shock patients to diagnose
hydrocortisone has been demonstrated to improve he- A-1.56,89,99
modynamic status, downregulate the proinflammatory From a practical point of view it is reasonable to
response, and improve survival10,11,96–98 Annane and initiate treatment with low-dose steroids in any patient
colleagues in a landmark placebo-controlled, random- presenting with septic shock and refractory hypotension
ized, double-blind, multicenter study, enrolled 300 adult pending the results of random cortisol (Fig. 2).97 Glu-
patients with septic shock after undergoing a short high- cocorticoids should be continued in those patients with a
dose (250 mg) corticotropin test.11 Patients were ran- stress cortisol level of < 20 mg/dL and in those patients
domly assigned to receive either hydrocortisone (50 mg with a level > 20 mg/dL who have demonstrated a clear-
IV q6h) and fludrocortisone (50 mg tablet once daily) cut hemodynamic response (lesser vasopressor require-
(n ¼ 151) or matching placebos (n ¼ 149) for 7 days. The ment) to glucocorticoid replacement.56,64,89 We believe
mortality was 53% in the corticosteroid group and 63% this to be a useful (although not the only) approach to
in the placebo group. Vasopressor therapy was with- the management of adrenal failure in the critically ill
drawn in 40% of patients who received placebo and in patient until more specific diagnostic tests become avail-
57% in the corticosteroid group (hazard ratio, 1.91; 95% able that can quantitate glucocorticoid activity at the
confidence interval, 1.29 to 2.84; p ¼ .001). cellular or nuclear level. The ‘‘best’’ dosing schedule has
Marik and Zaloga compared whether a baseline yet to be determined; however, currently hydrocortisone
(random) cortisol concentration < 25 mg/dL was a better at a dose of 50 mg q6h or 100 mg q8h is recommended.
discriminator of adrenal insufficiency than the standard Alternatively hydrocortisone can be given as a 100 mg
(250 mg) and the low-dose (1 mg) corticotropin stim- bolus, followed by an infusion at 10 mg/h. This latter
ulation tests in 59 patients with septic shock.56 Follow- regimen may result in better glycemic control. Hydro-
ing baseline cortisol level, patients were given 1 mg of cortisone should be continued for 5 to 7 days at the above
corticotropin (low dose), followed 60 minutes later by an dose before tapering, assuming that there is no recur-
injection of 249 mg of corticotropin (high-dose test). rence of signs of sepsis or shock. The hydrocortisone
Cortisol concentrations were obtained 30 and 60 mi- dose should then be reduced every 2 to 3 days by 50%,
nutes after low- and high-dose corticotropin. All pa- unless there is clinical deterioration, which would re-
tients were administered hydrocortisone (100 mg q8h) quire an increase in hydrocortisone dose. Currently,
for the first 24 hours while awaiting results of cortisol there are no data available to suggest how long hydro-
assessment. Patients were considered steroid responsive cortisone should be continued and when and if ACTH
Figure 2 Recommended strategy for evaluation and treatment of adrenal insufficiency among the critically ill.
testing should be performed (to confirm recovery of flammation. If untreated, both are associated with a
adrenal function). Furthermore, although there are few higher mortality. Currently available evidence is robust
data, routine treatment with fludrocortisone is not rec- enough to suggest that a tight glycemic control with
ommended at this time. insulin and therapy of A-1 with steroid supplementation
will improve survival in critically ill patients.
CONCLUSION
Stress hyperglycemia and A-1 are common in critically AUTHORS’ NOTE
ill patients. Multiple pathogenetic mechanisms are re- The authors have no financial interest in any of the
sponsible for each of these distinct metabolic syndromes; products mentioned in this article.
however, increased release of proinflammatory mediators
and counterregulatory hormones may play a pivotal role.
REFERENCES
Hyperglycemia per se is proinflammatory, whereas in-
sulin has antiinflammatory properties. Similarly, A-1 is 1. Van den Berghe G, Wouters P, Van Weekers F, et al.
associated with a proinflammatory state, whereas steroid Intensive insulin therapy in the critically ill patients. N Engl J
supplementation attenuates cortisol deficiency and in- Med 2001;345:1359–1367
2. Van den Berghe G, Schoonheydt K, Becx P, Bruyninckx F, myocardial infarction even in nondiabetic patients. Diabetes
Wouters PJ. Insulin therapy protects the central and peripheral Care 1999;22:1827–1831
nervous system of intensive care patients. Neurology 2005;64: 23. Zindrou D, Taylor KM, Bagger JP. Admission plasma
1348–1353 glucose: an independent risk factor in nondiabetic women
3. Krinsley JS. Effect of an intensive glucose management after coronary artery bypass grafting. Diabetes Care 2001;24:
protocol on the mortality of critically ill adult patients. Mayo 1634–1639
Clin Proc 2004;79:992–1000 24. Malmberg K, Norhammar A, Wedel H, Ryden L. Glyco-
4. Bochicchio GV, Sung J, Joshi M, et al. Persistent hyper- metabolic state at admission: important risk marker of
glycemia is predictive of outcome in critically ill trauma mortality in conventionally treated patients with diabetes
patients. J Trauma 2005;58:921–924 mellitus and acute myocardial infarction: long-term results
5. Laird AM, Miller PR, Kilgo PD, Meredith JW, Chang MC. from the Diabetes and Insulin-Glucose Infusion in Acute
Relationship of early hyperglycemia to mortality in trauma Myocardial Infarction (DIGAMI) study. Circulation 1999;
patients. J Trauma 2004;56:1058–1062 99:2626–2632
6. Finney SJ, Zekveld C, Elia A, Evans TW. Glucose control and 25. Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC.
mortality in critically ill patients. JAMA 2003;290:2041–2047 Stress hyperglycemia and prognosis of stroke in nondiabetic
7. Pittas AG, Siegel RD, Lau J. Insulin therapy for critically ill and diabetic patients: a systematic overview. Stroke 2001;32:
hospitalized patients: a meta-analysis of randomized con- 2426–2432
trolled trials. Arch Intern Med 2004;164:2005–2011 26. Malmberg K. Prospective randomised study of intensive
8. Krinsley JS. Association between hyperglycemia and increased insulin treatment on long term survival after acute myocardial
hospital mortality in a heterogeneous population of critically infarction in patients with diabetes mellitus. DIGAMI
ill patients. Mayo Clin Proc 2003;78:1471–1478 (Diabetes Mellitus, Insulin Glucose Infusion in Acute
9. Marik PE, Zaloga GP. Adrenal insufficiency in the critically Myocardial Infarction) Study Group. BMJ 1997;314:1512–
ill: a new look at an old problem. Chest 2002;122:1784–1796 1515
10. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, 27. Malmberg K, Ryden L, Efendic S, et al. Randomized trial of
Kupfer Y. Corticosteroids for severe sepsis and septic shock: a insulin-glucose infusion followed by subcutaneous insulin
systematic review and meta-analysis. BMJ 2004;329:489 treatment in diabetic patients with acute myocardial infarction
11. Annane D, Sebille V, Charpentier C, et al. Effect of treatment (DIGAMI study): effects on mortality at 1 year. J Am Coll
with low doses of hydrocortisone and fludrocortisone on Cardiol 1995;26:57–65
mortality in patients with septic shock. JAMA 2002;288:862– 28. Woo J, Lam CW, Kay R, Wong AH, Teoh R, Nicholls MG.
871 The influence of hyperglycemia and diabetes mellitus on
12. Marik PE, Raghavan M. Stress-hyperglycemia, insulin and immediate and 3-month morbidity and mortality after acute
immunomodulation in sepsis. Intensive Care Med 2004;30: stroke. Arch Neurol 1990;47:1174–1177
748–756 29. Garg R, Tripathy D, Dandona P. Insulin resistance as a
13. Hart BB, Stanford GG, Ziegler MG, Lake CR, Chernow B. proinflammatory state: mechanisms, mediators, and thera-
Catecholamines: study of interspecies variation. Crit Care peutic interventions. Curr Drug Targets 2003;4:487–492
Med 1989;17:1203–1222 30. Dandona P, Aljada A, Bandyopadhyay A. The potential
14. Van den BG. Neuroendocrine pathobiology of chronic critical therapeutic role of insulin in acute myocardial infarction in
illness. Crit Care Clin 2002;18:509–528 patients admitted to intensive care and in those with
15. Siegel JH, Cerra FB, Coleman B, et al. Physiological and unspecified hyperglycemia. Diabetes Care 2003;26:516–519
metabolic correlations in human sepsis: invited commentary. 31. Hansen TK, Thiel S, Wouters PJ, Christiansen JS, Van den
Surgery 1979;86:163–193 BG. Intensive insulin therapy exerts antiinflammatory effects
16. Clowes GH Jr, Martin H, Walji S, Hirsch E, Gazitua R, in critically ill patients and counteracts the adverse effect of
Goodfellow R. Blood insulin responses to blood glucose levels low mannose-binding lectin levels. J Clin Endocrinol Metab
in high output sepsis and septic shock. Am J Surg 1978;135: 2003;88:1082–1088
577–583 32. Ceriello A, Bortolotti N, Motz E, et al. Meal-induced
17. Mizock BA. Alterations in fuel metabolism in critical illness: oxidative stress and low-density lipoprotein oxidation in dia-
hyperglycaemia. Best Pract Res Clin Endocrinol Metab betes: the possible role of hyperglycemia. Metabolism 1999;48:
2001;15:533–551 1503–1508
18. Dahn MS, Jacobs LA, Smith S, et al. The relationship of 33. Mowlavi A, Andrews K, Milner S, Herndon DN, Heggers JP.
insulin production to glucose metabolism in severe sepsis. The effects of hyperglycemia on skin graft survival in the burn
Arch Surg 1985;120:166–172 patient. Ann Plast Surg 2000;45:629–632
19. Mehta VK, Hao W, Brooks-Worrell BM, Palmer JP. Low- 34. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A. Continuous
dose interleukin 1 and tumor necrosis factor individually intravenous insulin infusion reduces the incidence of deep
stimulate insulin release but in combination cause suppres- sternal wound infection in diabetic patients after cardiac
sion. Eur J Endocrinol 1994;130:208–214 surgical procedures. Ann Thorac Surg 1999;67:352–360
20. McCowen KC, Malhotra A, Bistrian BR. Stress-induced 35. Zerr KJ, Furnary AP, Grunkemeier GL, Bookin S, Kanhere
hyperglycemia. Crit Care Clin 2001;17:107–124 V, Starr A. Glucose control lowers the risk of wound infection
21. Frankenfield DC, Omert LA, Badellino MM, et al. Corre- in diabetics after open heart operations. Ann Thorac Surg
lation between measured energy expenditure and clinically 1997;63:356–361
obtained variables in trauma and sepsis patients. JPEN J 36. McManus LM, Bloodworth RC, Prihoda TJ, Blodgett JL,
Parenter Enteral Nutr 1994;18:398–403 Pinckard RN. Agonist-dependent failure of neutrophil
22. Norhammar AM, Ryden L, Malmberg K. Admission plasma function in diabetes correlates with extent of hyperglycemia.
glucose: independent risk factor for long-term prognosis after J Leukoc Biol 2001;70:395–404
37. Evans TW. Hemodynamic and metabolic therapy in critically with guidelines and relationship to clinical outcomes. Chest
ill patients. N Engl J Med 2001;345:1417–1418 2003;124:297–305
38. Vanhorebeek I, De Vos R, Mesotten D, Wouters PJ, Wolf- 56. Marik PE, Zaloga GP. Adrenal insufficiency during septic
Peeters C, Van den BG. Protection of hepatocyte mitochon- shock. Crit Care Med 2003;31:141–145
drial ultrastructure and function by strict blood glucose 57. Cooper MS, Stewart PM. Corticosteroid insufficiency in
control with insulin in critically ill patients. Lancet 2005; acutely ill patients. N Engl J Med 2003;348:727–734
365:53–59 58. Annane D, Sebille V, Troche G, Raphael JC, Gajdos P,
39. Dandona P, Aljada A, Mohanty P, et al. Insulin inhibits Bellissant E. A 3-level prognostic classification in septic shock
intranuclear nuclear factor kappaB and stimulates IkappaB in based on cortisol levels and cortisol response to corticotropin.
mononuclear cells in obese subjects: evidence for an anti- JAMA 2000;283:1038–1045
inflammatory effect? J Clin Endocrinol Metab 2001;86:3257– 59. Marik PE, Kiminyo K, Zaloga GP. Adrenal insufficiency in
3265 critically ill patients with human immunodeficiency virus. Crit
40. Langouche L, Vanhorebeek I, Vlasselaers D, et al. Intensive Care Med 2002;30:1267–1273
insulin therapy protects the endothelium of critically ill 60. Harry R, Auzinger G, Wendon J. The clinical importance of
patients. J Clin Invest 2005;115:2277–2286 adrenal insufficiency in acute hepatic dysfunction. Hepatology
41. Whalen MJ, Doughty LA, Carlos TM, Wisniewski SR, 2002;36:395–402
Kochanek PM, Carcillo JA. Intercellular adhesion molecule-1 61. Dimopoulou I, Tsagarakis S, Anthi A, et al. High prevalence
and vascular cell adhesion molecule-1 are increased in the of decreased cortisol reserve in brain-dead potential organ
plasma of children with sepsis-induced multiple organ failure. donors. Crit Care Med 2003;31:1113–1117
Crit Care Med 2000;28:2600–2607 62. Kilger E, Weis F, Briegel J, et al. Stress doses of hydrocortisone
42. Boldt J, Wollbruck M, Kuhn D, Linke LC, Hempelmann G. reduce severe systemic inflammatory response syndrome and
Do plasma levels of circulating soluble adhesion molecules improve early outcome in a risk group of patients after cardiac
differ between surviving and nonsurviving critically ill surgery. Crit Care Med 2003;31:1068–1074
patients? Chest 1995;107:787–792 63. Marik PE, Zaloga GP. The central nervous system hypo-
43. Sakurai Y, Aarsland A, Herndon DN, et al. Stimulation of thalamic-pituitary-adrenal axis in sepsis. Crit Care Med 2002;
muscle protein synthesis by long-term insulin infusion in 30:490–491
severely burned patients. Ann Surg 1995;222:283–294 64. Raghavan M, Marik PE. Management of sepsis during the
44. Gore DC, Wolf SE, Sanford AP, Herndon DN, Wolfe RR. early ‘‘golden hours.’’ J Emerg Med 2006; In press
Extremity hyperinsulinemia stimulates muscle protein syn- 65. Beishuizen A, Thijs LG, Vermes I. Patterns of corticosteroid-
thesis in severely injured patients. Am J Physiol Endocrinol binding globulin and the free cortisol index during septic
Metab 2004;286:E529–E534 shock and multitrauma. Intensive Care Med 2001;27:1584–
45. Woolfson AM, Heatley RV, Allison SP. Insulin to inhibit 1591
protein catabolism after injury. N Engl J Med 1979;300:14– 66. Zaloga GP. Sepsis-induced adrenal deficiency syndrome. Crit
17 Care Med 2001;29:688–690
46. Van den BG, Wouters PJ, Bouillon R, et al. Outcome 67. Schroeder S, Wichers M, Klingmuller D, et al. The
benefit of intensive insulin therapy in the critically ill: insulin hypothalamic-pituitary-adrenal axis of patients with severe
dose versus glycemic control. Crit Care Med 2003;31:359– sepsis: altered response to corticotropin-releasing hormone.
366 Crit Care Med 2001;29:310–316
47. Krinsley J, Grissler B. Intensive glycemic management in 68. Norbiato G, Bevilacqua M, Vago T, et al. Cortisol resistance
critically ill patients. Jt Comm J Qual Patient Saf 2005;31: in acquired immunodeficiency syndrome. J Clin Endocrinol
308–312 Metab 1992;74:608–613
48. Grey NJ, Perdrizet GA. Reduction of nosocomial infections 69. Ali M, Allen HR, Vedeckis WV, Lang CH. Depletion of rat
in the surgical intensive-care unit by strict glycemic control. liver glucocorticoid receptor hormone-binding and its mRNA
Endocr Pract 2004;10(Suppl 2):46–52 in sepsis. Life Sci 1991;48:603–611
49. Marik PE, Zaloga GP. Early enteral nutrition in acutely ill 70. Molijn GJ, Koper JW, van Uffelen CJ, et al. Temperature-
patients: a systematic review. Crit Care Med 2001;29:2264– induced down-regulation of the glucocorticoid receptor in
2270 peripheral blood mononuclear leucocyte in patients with
50. Marik PE, Pinsky M. Death by parenteral nutrition. Intensive sepsis or septic shock. Clin Endocrinol (Oxf) 1995;43:197–
Care Med 2003;29:867–869 203
51. Dickerson RN. Hypocaloric feeding of obese patients in the 71. Melby JC, Spink WW. Comparative studies on adrenal
intensive care unit. Curr Opin Clin Nutr Metab Care cortical function and cortisol metabolism in healthy adults and
2005;8:189–196 in patients with shock due to infection. J Clin Invest 1958;
52. Moore FA, Feliciano DV, Andrassy RJ, et al. Early enteral 37:1791–1798
feeding, compared with parenteral, reduces postoperative 72. Marik PE, Gayowski T, Starzl TE. The hepatoadrenal
septic complications: the results of a meta-analysis. Ann Surg syndrome: a common yet unrecognized clinical condition.
1992;216:172–183 Crit Care Med 2005;33:1254–1259
53. Jenkins DJ, Ghafari H, Wolever TM, et al. Relationship 73. Singh N, Gayowski T, Marino IR, Schlichtig R. Acute
between rate of digestion of foods and post-prandial adrenal insufficiency in critically ill liver transplant recipients.
glycaemia. Diabetologia 1982;22:450–455 Implications for diagnosis. Transplantation 1995;59:1744–
54. Marik PE, Zaloga GP. Gastric versus post-pyloric feeding: a 1745
systematic review. Crit Care 2003;7:R46–R51 74. Marik PE. Adrenal insufficiency: the link between low
55. Krishnan JA, Parce PB, Martinez A, Diette GB, Brower RG. apolipoprotein A-1 levels and poor outcome in the critically
Caloric intake in medical ICU patients: consistency of care ill? Crit Care Med 2004;32:1977–1978
75. Yaguchi H, Tsutsumi K, Shimono K, Omura M, Sasano H, 86. Ettinger WH, Varma VK, Sorci-Thomas M, et al. Cytokines
Nishikawa T. Involvement of high density lipoprotein as decrease apolipoprotein accumulation in medium from Hep
substrate cholesterol for steroidogenesis by bovine adrenal G2 cells. Arterioscler Thromb 1994;14:8–13
fasciculo-reticularis cells. Life Sci 1998;62:1387–1395 87. Soni A, Pepper GM, Wyrwinski PM, et al. Adrenal
76. Borkowski AJ, Levin S, Delcroix C, Mahler A, Verhas V. insufficiency occurring during septic shock: incidence, out-
Blood cholesterol and hydrocortisone production in man: come, and relationship to peripheral cytokine levels. Am J
quantitative aspects of the utilization of circulating cholesterol Med 1995;98:266–271
by the adrenals at rest and under adrenocorticotropin 88. Marik P, Zaloga G. Prognostic value of cortisol response in
stimulation. J Clin Invest 1967;46:797–811 septic shock. JAMA 2000;284:308–309
77. Liu J, Heikkila P, Meng QH, Kahri AI, Tikkanen MJ, 89. Marik PE. Unraveling the mystery of adrenal failure in the
Voutilainen R. Expression of low and high density lipoprotein critically ill. Crit Care Med 2004;32:596–597
receptor genes in human adrenals. Eur J Endocrinol 2000; 90. Hamrahian AH, Oseni TS, Arafah BM. Measurements of
142:677–682 serum free cortisol in critically ill patients. N Engl J Med
78. Calvo D, Gomez-Coronado D, Lasuncion MA, Vega MA. 2004;350:1629–1638
CLA-1 is an 85-kD plasma membrane glycoprotein that acts 91. Sun X, Fischer DR, Pritts TA, Wray CJ, Hasselgren PO.
as a high-affinity receptor for both native (HDL, LDL, and Expression and binding activity of the glucocorticoid receptor
VLDL) and modified (OxLDL and AcLDL) lipoproteins. are upregulated in septic muscle. Am J Physiol Regul Integr
Arterioscler Thromb Vasc Biol 1997;17:2341–2349 Comp Physiol 2002;282:R509–R518
79. van der Voort PH, Gerritsen RT, Bakker AJ, Boerma EC, 92. Liu DH, Su YP, Zhang W, et al. Changes in glucocorticoid
Kuiper MA, de Heide L. HDL-cholesterol level and cortisol and mineralocorticoid receptors of liver and kidney cytosols
response to synacthen in critically ill patients. Intensive Care after pathologic stress and its regulation in rats. Crit Care
Med 2003;29:2199–2203 Med 2002;30:623–627
80. Chenaud C, Merlani PG, Roux-Lombard P, et al. Low 93. Dimopoulou I, Tsagarakis S, Kouyialis AT, et al. Hypothala-
apolipoprotein A-1 level at intensive care unit admission and mic-pituitary-adrenal axis dysfunction in critically ill patients
systemic inflammatory response syndrome exacerbation. Crit with traumatic brain injury: incidence, pathophysiology, and
Care Med 2004;32:632–637 relationship to vasopressor dependence and peripheral inter-
81. Mookerjee RP, Sen S, Davies NA, Hodges SJ, Williams R, leukin-6 levels. Crit Care Med 2004;32:404–408
Jalan R. Tumour necrosis factor alpha is an important 94. Schenarts CL, Burton JH, Riker RR. Adrenocortical
mediator of portal and systemic haemodynamic derangements dysfunction following etomidate induction in emergency
in alcoholic hepatitis. Gut 2003;52:1182–1187 department patients. Acad Emerg Med 2001;8:1–7
82. Yokoyama I, Gavaler JS, Todo S, Miyata T, Van Thiel 95. Drucker D, Shandling M. Variable adrenocortical function in
DH, Starzl TE. Endotoxemia is associated with renal acute medical illness. Crit Care Med 1985;13:477–479
dysfunction in liver transplantation recipients during the 96. Keh D, Boehnke T, Weber-Cartens S, et al. Immunologic and
first postoperative week. Hepatogastroenterology 1995;42: hemodynamic effects of ‘‘low-dose’’ hydrocortisone in septic
205–208 shock: a double-blind, randomized, placebo-controlled, cross-
83. Rasaratnam B, Kaye D, Jennings G, Dudley F, Chin-Dusting over study. Am J Respir Crit Care Med 2003;167:512–520
J. The effect of selective intestinal decontamination on the 97. Dellinger RP, Carlet JM, Masur H, et al. Surviving Sepsis
hyperdynamic circulatory state in cirrhosis: a randomized trial. Campaign guidelines for management of severe sepsis and
Ann Intern Med 2003;139:186–193 septic shock. Crit Care Med 2004;32:858–873
84. Vishnyakova TG, Bocharov AV, Baranova IN, et al. Binding 98. Briegel J, Forst H, Haller M, et al. Stress doses of
and internalization of lipopolysaccharide by CLA-1, a human hydrocortisone reverse hyperdynamic septic shock: a prospec-
orthologue of rodent scavenger receptor B1. J Biol Chem tive, randomized, double-blind, single-center study. Crit Care
2003;278:22771–22780 Med 1999;27:723–732
85. Baranova I, Vishnyakova T, Bocharov A, et al. Lipopoly- 99. Mayenknecht J, Diederich S, Bahr V, Plockinger U, Oelkers
saccharide down regulates both scavenger receptor B1 and W. Comparison of low and high dose corticotropin
ATP binding cassette transporter A1 in RAW cells. Infect stimulation tests in patients with pituitary disease. J Clin
Immun 2002;70:2995–3003 Endocrinol Metab 1998;83:1558–1562
Hematologic Disorders in Critically Ill
Patients
Kelly W. Mercer, M.D.,1 B. Gail Macik, M.D.,1 and Michael E. Williams, M.D.1
ABSTRACT
Hematologic disorders are frequently encountered in the intensive care unit.

Thrombocytopenia, often defined as a platelet count below 100,000/mL, is common in
critically ill patients and may be associated with adverse outcomes. A systematic evaluation
of clinical and laboratory findings is necessary to ascertain the cause of the thrombocyto-
penia and to determine the correct therapy. Recognition of heparin-induced thrombocy-
topenia (HIT) is particularly important, given the risk of thrombosis associated with this
condition. Prompt cessation of all heparin products is required, and anticoagulation with a
direct thrombin inhibitor is recommended if HIT is strongly suspected. Coagulopathies are
also common in the critically ill, and are often due to vitamin K deficiency or disseminated
intravascular coagulation (DIC). A careful history and interpretation of clotting studies are
useful in defining the coagulation defect. Advances in understanding the pathogenesis of
DIC have generated new treatment approaches, such as the use of recombinant activated
protein C. Recombinant factor VIIa (rFVIIa) is a novel drug approved for use in patients
with congenital hemophilias and inhibitors. Although its use as a hemostatic agent is
currently being evaluated in several off-label scenarios, including trauma, intracerebral
hemorrhage, and liver disease, there are limited data to guide therapy in these conditions.
KEYWORDS: Thrombocytopenia, coagulopathy, heparin-induced thrombocytopenia,

disseminated intravascular coagulation, recombinant factor VIIa
H ematologic disorders are common among crit- clinical scenarios. Thrombocytopenia may result in a
ically ill patients and frequently contribute to adverse bleeding diathesis necessitating transfusions; it may
outcomes. This review describes the most frequent non- also predict for increased morbidity and mortality.
neoplastic hematologic problems encountered in the Successful management of thrombocytopenia requires
intensive care unit and summarizes current approaches prompt and accurate recognition of its underlying
to diagnosis and management. cause. Drug-induced thrombocytopenia can be partic-
ularly challenging because many critically ill patients
receive multiple medications. Heparin-induced throm-
THROMBOCYTOPENIA IN THE INTENSIVE bocytopenia is of special concern, given the associated
CARE UNIT risk of thrombosis and the unique treatment of this
Thrombocytopenia is one of the most common labo- disorder.
ratory abnormalities in the intensive care unit (ICU). It Thrombocytopenia is frequently defined as a
may occur via several mechanisms and in a variety of platelet count below 100,000/mL. The incidence of
1
Division of Hematology/Oncology, University of Virginia School of Non-pulmonary Critical Care: Managing Multisystem Critical Illness;
Medicine, Charlottesville, Virginia. Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Michael E. Semin Respir Crit Care Med 2006;27:286–296. Copyright # 2006
Williams, M.D., Hematology/Oncology Division, Box 800716, Uni- by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
versity of Virginia School of Medicine, Jefferson Park Ave., Charlot- NY 10001, USA. Tel: +1(212) 584-4662.
tesville, VA 22908. E-mail: [email protected]. DOI 10.1055/s-2006-945529. ISSN 1069-3424.
286
HEMATOLOGIC DISORDERS IN CRITICALLY ILL PATIENTS/MERCER ET AL 287
thrombocytopenia in studies of medical and surgical necessary, particularly if a primary bone marrow disorder
ICU patients has ranged from 21 to 41%.1–7 Severe is suspected.
thrombocytopenia (platelets less than 20,000/mL) is
much less common, typically developing in less than
5% of ICU patients.1–4,7 Risk factors for the develop- HEPARIN-INDUCED THROMBOCYTOPENIA
ment of thrombocytopenia in the ICU have included Recognition of heparin-induced thrombocytopenia
sepsis, disseminated intravascular coagulation (DIC), (HIT) and HIT with thrombosis (HITT) is of particular
organ failure, central or arterial lines, episodes of bleed- importance given its paradoxical association with throm-
ing or transfusion, cardiopulmonary resuscitation, and bosis. HIT/HITT is an immune-mediated disorder that
use of chemotherapy.1,3–6,8,9 Drug-related thrombocy- is triggered by exposure to any form of heparin13,14; it is
topenia was seen in only a minority of cases.1–5 Develop- also known as type II HIT, to distinguish it from the
ment of a low platelet count has been associated with non-immune-mediated, mild thrombocytopenia associ-
several adverse outcomes, including longer ICU and ated with heparin termed type I HIT. Type II HIT is
hospital stays, increased incidence of major bleeding caused by the generation of heparin-induced, platelet-
and requirement for transfusions, and increased morta- activating immunoglobulin G (IgG) antibodies that
lity.3–9 Correction of thrombocytopenia has been asso- recognize heparin-platelet factor 4 complexes.15–18 The
ciated with a decreased risk of death.5,6 resulting platelet activation and thrombin generation lead
Four major mechanisms can explain the develop- to a significant risk of both arterial and venous throm-
ment of low platelet counts: spurious thrombocytopenia, bosis. Depending on the patient population studied, the
decreased platelet production, disorders of distribution risk of thrombosis has ranged from 29 to 89%.14,19–23
and dilution, and increased platelet destruction.10,11 A Even after cessation of heparin therapy, the threat of
low platelet count should always be repeated because thrombosis persists.21,22 In one study, the 30-day risk of
improper collection techniques or laboratory error may thrombosis after the diagnosis of HIT was 53%.22
be to blame. Examination of the peripheral blood film is The incidence of HIT varies depending on the
essential; this may reveal platelet clumping, which can be patient population and the type of heparin preparation
induced by ethylenediaminetetraacetic acid (EDTA)- used. Patients undergoing cardiac or orthopedic surgery
dependent antibodies.12 If this is suspected, blood are among those at highest risk of developing HIT.24,25
should be redrawn in a tube containing an alternative HIT is less common in medical patients, with studies
anticoagulant such as heparin or citrate. suggesting a frequency of 1% or less.19,20,26 Patients
Platelet production may be adversely affected by receiving unfractionated heparin (UFH) are at increased
several conditions. These include congenital or acquired risk of HIT compared with those given the low molec-
disorders of hematopoiesis; liver disease with decreased ular weight heparin (LMWH) preparations.14,27 How-
production of thrombopoietin; bone marrow suppression ever, there is a high risk of in vivo cross-reactivity
secondary to drugs, toxins, or infectious agents; and between UFH and LMWH, and one agent may not be
nutritional deficiencies such as folate and vitamin B12 substituted for the other if HIT is suspected.24,28
deficiency. A dilutional thrombocytopenia may occur Diagnosis of HIT requires consideration of both
following massive transfusion of blood and plasma prod- clinical and serologic findings.29,30 Because nonpatho-
ucts. Patients with marked splenomegaly may have dis- genic heparin-platelet factor 4 antibodies occur com-
tributional thrombocytopenia secondary to increased monly in patients treated with heparin,30 a positive test
splenic sequestration. Increased platelet destruction, by for HIT antibodies is not sufficient to make the diag-
both immune and nonimmune mechanisms, represents nosis. Additionally, given the risk of thrombosis in
the most common reason for thrombocytopenia in ICU patients with HIT, appropriate therapy should not be
patients.10 Sepsis syndrome, DIC, immune thrombocy- withheld while awaiting the results of serologic testing.
topenic purpura (ITP), drug-induced thrombocytopenia, HIT should be suspected in patients who develop
and thrombotic thrombocytopenic purpura-hemolytic thrombocytopenia or experience a relative drop in pla-
uremic syndrome (TTP-HUS) are all associated with telet count of 50%,29 typically occurring 4 to 10 days
increased platelet destruction. after initiation of heparin therapy.31 Thrombocytopenia
Assessment of the patient with thrombocytopenia may occur much more rapidly after exposure to heparin,
should involve a thorough history and physical examina- however, in patients who have received heparin within
tion, as well as a review of all recent and current the previous 100 days.31 New or recurrent venous or
medications. Coagulation parameters and the presence arterial thromboses in patients who are receiving, or have
or absence of other cytopenias should be considered in recently received, a heparin product should also raise
conjunction with the patient’s platelet count. Evaluation suspicion for HIT. Lastly, skin lesions at heparin in-
of a peripheral blood smear may provide important jection sites or acute systemic reactions after bolus
information regarding the cause of the patient’s throm- administration of intravenous heparin are other clues
bocytopenia. Bone marrow aspiration and biopsy may be to a diagnosis of HIT.29,32
Table 1 Comparison of Argatroban and Lepirudin for Heparin-Induced Thrombocytopenia

Argatroban Lepirudin
Approved indication Prophylaxis or treatment of HIT; Treatment of HIT and associated

PCI in patients with HIT thrombosis
Structure Small molecule Recombinant hirudin
Clearance Hepatic Renal
Half-life 40–50 minutes 80 minutes
Dosing in HIT with thrombosis No bolus; initial infusion rate for With or without bolus 0.4 mg/kg; initial
patients with HIT, 2 mg/kg/min infusion rate 0.15 mg/kg/h
Dosing in HIT without thrombosis No bolus; initial infusion rate for patients No bolus; initial infusion rate 0.1 mg/kg/h
with HIT, 2 mg/kg/min
Goal aPTT ratio 1.5–3.0 1.5–2.5 HIT with thrombosis; 1.5–2.0
HIT without thrombosis
Monitoring 2 hours after adjustments 4 hours after adjustments
Antidote None None
Special Issues - Dose reduction required in hepatic insufficiency: - Prolonged half-life in renal failure may
0.5 mg/kg/min require marked dose reduction
- Significantly prolongs PT/INR - Antihirudin antibody formation in 50%;
- Dose for PCI in patients may increase risk of bleeding; may
with HIT: bolus 350 mg/kg over cause anaphylaxis on reexposure
3–5 minutes; initial infusion rate 25 mg/kg/min
HIT, heparin-induced thrombocytopenia; PCI, percutaneous coronary intervention; aPTT, activated partial thromboplastin time; PT, prothrombin
time; INR, international normalized ratio.
Adapted from Warkentin, Greinacher29 and Rice.35
The diagnosis of HIT is necessarily based on sulfated heparan sulfate, is no longer available in the
clinical suspicion, although laboratory testing for HIT United States but is approved in some other countries for
antibodies may help confirm or exclude the diagnosis. management of HIT. Other agents under investigation
Both platelet activation and antigen-based assays are but not yet FDA-approved include bivalirudin, a paren-
available. The serotonin release assay and the heparin- teral DTI, and fondaparinux, a pentasaccharide with
induced platelet aggregation assay are functional tests of exclusive antifactor Xa activity.34,35 Vitamin K antago-
platelet activation. Although these assays are more nists such as warfarin may cause venous limb gangrene
specific for clinical HIT than antigen-based tests, they and skin necrosis in patients with HIT, and are therefore
are technically demanding to perform and are not widely contraindicated in the acute setting. It is recommended
available.16,33 The most commonly used test is an that vitamin K antagonists be given in HIT only once
enzyme-linked immunoassay (ELISA) that directly de- the platelet count has recovered to at least 100 109/L,
tects the presence of antibodies binding to heparin- and only when given during overlapping therapy with an
platelet factor 4 complexes. Although simpler to perform alternate anticoagulant.29
and highly sensitive for the detection of HIT antibodies,
ELISA testing is less specific for clinically relevant
heparin antibodies than the functional assays.16,33 A COAGULOPATHY IN THE INTENSIVE
positive ELISA test should be interpreted in the context CARE UNIT
of the patient’s clinical findings. Coagulation defects are common in critically ill pa-
If HIT is suspected, all heparin products must tients.3–7 An adequate assessment of the bleeding patient
immediately be discontinued, including heparin flushes requires not only a thorough history and physical exami-
and heparin-coated catheters.29 Given the high risk of nation but also a good understanding of coagulation
thrombosis with HIT, even after discontinuation of laboratory tests. Frequent causes of coagulopathies in-
heparin,21 it is currently recommended that an alternate, clude vitamin K deficiency, use of warfarin, liver dys-
nonheparin anticoagulant replace heparin in patients function, and DIC. Treatment strategies are dictated by
strongly suspected of having HIT.29 In the United the severity of the coagulation disorder as well as its
States, the direct thrombin inhibitors (DTIs) argatroban underlying cause. Coagulopathy due to DIC is of par-
and lepirudin are the only agents approved by the Food ticular interest, given recent advances regarding its
and Drug Administration (FDA) for the treatment of pathogenesis, diagnosis, and management. Recombinant
HIT. Characteristics and dosing information for these factor VIIa (rFVIIa) is a novel hemostatic agent being
drugs are reviewed in Table 1. Danaparoid, a heparinoid tested in several off-label clinical settings including
containing a mixture of dermatan sulfate and low- trauma, intracerebral hemorrhage, and liver disease.
Coagulopathic bleeding typically presents with S. Decreased production of these factors in vitamin K–
the development of large soft tissue hematomas and deficient patients causes prolongation of the PT and a
palpable ecchymoses. Bleeding into joints, or hemarth- rise in the INR. In cases of protracted and severe
rosis, is a hallmark of the inherited hemophilias. deficiency, the aPTT may also lengthen. Vitamin K
Generalized oozing around intravenous catheters deficiency is a common cause of acquired coagulopathy
or from mucocutaneous sites is also suggestive of a in critical care patients.7,36,37 Factors contributing to the
systemic coagulation problem. Delayed bleeding fol- development of vitamin K deficiency include inadequate
lowing trauma or surgical procedures commonly devel- oral intake, recent major surgery, gastrointestinal disor-
ops in patients with coagulation defects because intact ders, and the use of broad spectrum antibiotics.38–40
platelet function achieves primary hemostasis. Special Recent guidelines developed by the British Committee
attention should be paid to any personal history of for Standards in Hematology recommend that adult
bleeding, including the timing and nature of previous ICU patients receive routine vitamin K supplementa-
episodes of bleeding and their relation to surgical tion, 10 mg thrice weekly. Fresh frozen plasma (FFP) is
procedures, tooth extractions, or trauma. A family not the treatment of choice to replenish vitamin K
history of bleeding and a detailed medication history stores.41
should also be obtained. Treatment of vitamin K deficiency involves re-
The initial laboratory assessment in patients with placement of vitamin K. Most studies that examined
coagulopathy includes evaluation of routine clotting vitamin K administration have involved patients who
tests: the prothrombin time (PT) and international were overanticoagulated with warfarin, an oral agent that
normalized ratio (INR); the activated partial thrombo- inhibits carboxylation of the vitamin K-dependent clot-
plastin time (aPTT); and fibrinogen. Analysis of the ting factors, causing a functional deficiency of these
pattern of abnormalities in the PT-INR and aPTT can proteins and elevating the PT-INR. Excessive antico-
be useful in defining the coagulation defect. Additional agulation is common in patients receiving warfarin.42
testing may be required if the cause of the coagulopathy Randomized, controlled trials have demonstrated that
remains unclear. The additional testing may include both oral and intravenous vitamin K preparations are
assays that measure levels of specific clotting factors; more efficacious, and shorten the INR more rapidly,
mixing studies to distinguish factor deficiencies from than subcutaneous vitamin K.43,44 In most patients,
factor inhibitors; and fibrin-fibrinogen degradation oral vitamin K results in significant INR reversal by
products (FDPs) and D-dimers, both of which indicate 24 hours.44–46 Intravenous vitamin K acts more rapidly,
fibrinolysis and are elevated in consumptive coagulo- lowering the INR within 4 to 6 hours.46,47 However,
pathies such as DIC. Examination of the peripheral administration of intravenous vitamin K has been asso-
blood smear can help rule out a microangiopathic ciated with a small risk of anaphylaxis48 and should
process. therefore be used only if rapid and urgent INR reversal
Coagulopathy associated with clinically important is necessary. The American College of Chest Physician
bleeding is noted in a significant minority of critical care Consensus Conference has recently published specific
patients. Chakraverty et al studied 235 intensive care guidelines for the correction of excessive anticoagulation
patients; a coagulopathy was identified based on clinical with warfarin.49
findings in approximately one in seven patients.7 Labo-
ratory evidence of coagulopathy was even more common
in this population because two thirds of patients had an DISSEMINATED INTRAVASCULAR
INR greater than 1.5. These patients had a significantly COAGULATION
higher rate of mortality than those who did not have a DIC is a systemic disorder characterized by derange-
coagulopathy. In a retrospective analysis, vitamin K ments of the coagulation and fibrinolytic systems, leading
deficiency was noted in 20% of patients, representing to widespread thrombosis and bleeding. Proinflammatory
the most common explanation for a prolonged PT. stimuli such as endotoxin and cytokines trigger tissue
Other investigators have described similar rates of vita- factor expression on mononuclear cells and vascular
min K deficiency in ICU patients.36 Liver failure, DIC, endothelium.50 Exposure of the blood to procoagulants
and use of drugs such as warfarin and heparin were also (largely tissue factor) causes systemic activation of the
identified as causes of coagulopathy. coagulation cascade, with increased thrombin production
and fibrin formation.51,52 Thrombin generation is en-
hanced by impaired antithrombotic mechanisms, with
VITAMIN K DEFICIENCY diminished levels of antithrombin (formerly known as
Vitamin K is a fat-soluble vitamin that plays a critical antithrombin III) and protein C.53–56 Fibrinolytic path-
role in the clotting system, functioning as a coenzyme in ways are also downregulated during DIC, secondary to
the posttranslational carboxylation of clotting factors II, increased levels of plasminogen activator inhibitor, type
VII, IX, and X, as well as the regulatory proteins C and 1 (PAI-1).57–59 Extensive intravascular fibrin deposition
ensues, resulting in microvascular thrombosis, impaired Table 2 Laboratory Markers in Disseminated

blood supply to various organs, and ultimately multiorgan Intravascular Coagulation
failure. Consumption of platelets and clotting factors Test Result
including fibrinogen may result in diffuse hemorrhage.51
PT/INR "
DIC is an acquired syndrome that arises in the
aPTT " and biphasic waveform
setting of another underlying disorder. Disease states
Thrombin time "
known to cause DIC include sepsis, severe trauma,
Platelet count #
malignancy (both solid tumor and hematological malig-
Fibrinogen #
nancies, particularly acute promyelocytic leukemia), ob-
D-dimer, FDPs, soluble fibrin "
stetrical complications, vascular abnormalities such as
Levels of specific clotting #
giant hemangiomas, and severe liver failure.51,60–62 Di-
factors (e.g., VII)
agnosis of DIC requires assessment of the underlying
Antithrombin #
clinical scenario in conjunction with appropriate labo-
Protein C #
ratory tests. DIC should be considered in patients with
PAI-1 "
an appropriate clinical syndrome such as sepsis, malig-
nancy, or trauma. Laboratory evaluation in DIC typically PT, prothrombin time; INR, international normalized ratio; aPTT,
activated partial thromboplastin time; FDPs, fibrin degradation prod-
reveals a consumptive coagulopathy, as demonstrated by ucts; PAI-1, plasminogen activator inhibitor, type 1; " , increasing; # ,
thrombocytopenia and prolongation of global clotting decreasing.
times, including the PT-INR, aPTT, and thrombin time
(Table 2). Recent studies have shown that development Although there are some experimental data suggesting
of an abnormal, biphasic waveform in the automated an advantage to using heparin in patients with DIC,70,71
aPTT coagulation assay may be an early predictor of randomized controlled trials have failed to demonstrate a
DIC and may correlate with higher mortality.63,64 In- beneficial effect.72 Therapeutic doses of heparin are
creased fibrinolysis is suggested by elevated levels of typically limited to patients with clinically overt throm-
FDPs and D-dimer.51,65 Increased thrombin generation botic complications associated with DIC, such as acral
may produce diminished levels of fibrinogen, although ischemia and purpura fulminans.51
fibrinogen values may be normal or even elevated as an Recently, several novel therapeutic agents that
acute-phase reactant in some cases of DIC.66 Low target specific elements of the coagulation system have
plasma levels of inhibitors of coagulation such as antith- been examined. Antithrombin (AT) is an essential
rombin and protein C contribute to the diagnosis.53–56 inhibitor of coagulation that acts via neutralization
Plasma levels of soluble fibrin are highly sensitive in of several enzymes in the clotting cascade, including
diagnosing DIC. However, they lack specificity, and a thrombin and factor Xa. Based on the findings that AT
reliable assay is not widely available.67 levels are diminished in DIC53,56 and that lower
The subcommittee on DIC of the International AT levels are associated with poorer outcomes,53,56,73
Society on Thrombosis and Haemostasis has recently AT concentrate has been administered to septic pa-
published a scoring system for DIC that incorporates tients in a randomized, placebo-controlled fashion.
simple and readily available laboratory tests (Table 3).68 Although initial trials indicated a mortality benefit of
In patients with a condition known to be associated with AT in patients with severe sepsis or septic shock,74,75 a
DIC, a score of 5 or more is compatible with DIC. This recent large, randomized, placebo-controlled trial of
scoring system was prospectively validated in a study of 2314 patients with severe sepsis failed to demonstrate a
217 critically ill medical and surgical patients admitted to survival advantage.76 Additionally, those patients who
the ICU with a clinical suspicion of DIC.69 The DIC received AT in conjunction with heparin had an in-
score was calculated every 48 hours. The score was found creased risk of hemorrhage.
to be highly accurate in the diagnosis of DIC, with a Another important anticoagulant mediator is tis-
sensitivity of 91% and a specificity of 97%. Increasing sue factor pathway inhibitor (TFPI), an endogenous
DIC score also correlated strongly with 28-day mortality. inhibitor of the extrinsic, or tissue factor–based, coagu-
The fundamental approach to treatment of DIC lation pathway. Phase II trials with recombinant TFPI
is prompt identification and aggressive management of (rTFPI) showed some promise with respect to mortality
the underlying disorder. Transfusion of blood products in severe sepsis.77,78 However, a large randomized,
may be required, although there are no consensus guide- placebo-controlled, multicenter phase III trial by Abra-
lines regarding their appropriate use. Transfusion should ham et al demonstrated no benefit for rTFPI in patients
not be administered purely in response to abnormal with severe sepsis and high INR.79 Patients who received
laboratory results. A combination of platelets, FFP, rTFPI had an increased risk of bleeding.
and/or cryoprecipitate is indicated in the actively bleed- Activated protein C (APC) is a serine protease
ing patient, or if the patient requires an invasive proce- with both antithrombotic and profibrinolytic properties.
dure or is at high risk for bleeding problems.41,51,52 APC inhibits thrombin generation via inactivation of
Table 3 Scoring System for Overt Disseminated Abraham et al.84 Enrollment was terminated early, after
Intravascular Coagulation accrual of 2640 patients, due to the low likelihood of
Laboratory Test Points Values achieving a significant reduction in 28 day mortality with
rhAPC. Twenty-eight day mortality and in-hospital
Platelet count > 100 ¼ 0
mortality were the same in the rhAPC and placebo
< 100 ¼ 1
arms; serious bleeding was significantly greater in the
< 50 ¼ 2
rhAPC-treated patients. In light of these results, the use
Elevated fibrin-related markers No increase ¼ 0
of rhAPC should be considered only in those patients
(e.g., soluble fibrin, fibrin Moderate increase ¼ 2
with severe sepsis and a high risk of death.
degradation products) Strong increase ¼ 3
Prolonged prothrombin time < 3 second ¼ 0
> 3 but < 6 second ¼ 1
RECOMBINANT FACTOR VIIA
> 6 second ¼ 2
In 1999, the FDA approved recombinant factor VIIa
Fibrinogen > 1 g/L ¼ 0
(rFVIIa) for the treatment of patients with congenital
< 1 g/L ¼ 1
hemophilia A or B and circulating inhibitors to factors
The above scoring system is for use only in patients with an under- VIII or IX. Recombinant FVIIa has since been evaluated
lying condition known to be associated with disseminated intra-
vascular coagulation. as a hemostatic agent in a growing number of off-label
A score of 5 is compatible with overt disseminated intravascular conditions, albeit primarily in case reports or small
coagulation.
Adapted from Taylor et al.68 controlled trials. Areas of active investigation include
traumatic bleeding, intracerebral hemorrhage, and coa-
clotting factors Va and VIIIa; it enhances fibrinolysis by gulopathy of liver disease. Recombinant FVIIa use has
inactivating PAI-1. APC also modulates anti-inflam- also been reported in nonhemophiliacs with acquired
matory and antiapoptotic pathways.80 Low levels of inhibitors to various clotting factors, hereditary clotting
protein C are predictive of a poor clinical outcome in factor deficiencies, platelet disorders, reversal of anti-
septic patients.54–56 Based on these findings as well as coagulation, surgical bleeding, and perioperative bleeding
encouraging results from a phase II study,81 Bernard et al prophylaxis.85 The typical charge for a 40 mg/kg dose of
conducted a randomized, double-blind, placebo-con- rFVIIa is approximately $4,000. Important questions
trolled, multicenter trial to evaluate the impact of a remain regarding the efficacy, optimal dose, safety, and
recombinant human APC (rhAPC), or drotrecogin cost-effectiveness of rFVIIa in these populations.
alfa activated (DrotAA), on 28-day all-cause mortality Recombinant FVIIa is a genetically engineered
in patients with severe sepsis.82 Patients received either analogue of the naturally occurring FVII protein, a
placebo (840 patients) or rhAPC as a continuous serine protease that becomes activated upon binding to
infusion of 24 mg/kg/h for a total of 96 hours tissue factor exposed at areas of endovascular damage.
(850 patients). The trial was stopped prematurely after The rFVIIa-tissue factor complex leads to factor X
the second planned interim analysis because of a statisti- activation, which results in the conversion of prothrom-
cally significant reduction in mortality in the patients bin to thrombin and, subsequently, the activation of
who received rhAPC. In the rhAPC-treated patients, platelets and other clotting factors. Although its precise
28-day mortality was 24.7%, as compared with 30.8% in mechanism of action remains a matter of debate, it has
the placebo population; the reduction in relative risk of been proposed that rFVIIa at pharmacological doses is
death was 19.4%. There was an increased incidence of able to bind to activated platelets and stimulate factor X
serious bleeding in those patients treated with rhAPC directly, in a tissue factor–independent manner. Factor
compared with placebo (3.5% vs 2.0%). Subgroup anal- Xa, in the presence of factor Va, generates a thrombin
ysis determined that the largest reduction in mortality burst resulting in formation of fibrin clot localized to the
occurred among APC-treated patients with more severe site of injury.85–87
disease and higher risk of death, as indicated by Acute Recombinant FVIIa has been investigated as a
Physiology and Chronic Health Evaluation (APACHE universal hemostatic agent in patients with uncontrol-
II) scores in the third and fourth quartiles.83 Based on lable bleeding due to traumatic coagulopathy. The initial
these results, in November 2001 the United States FDA case report, in 1999, described cessation of life-threat-
approved rhAPC for the treatment of adult patients with ening bleeding in a gunshot victim following treatment
severe sepsis and a high risk of death. The FDA’s with two doses of 60 mg/kg of rFVIIa.88 Subsequent
approval required an additional trial evaluating the series of patients with uncontrolled traumatic bleeding
efficacy and safety of rhAPC in patients with severe have also reported improvements in bleeding and base-
sepsis and a low likelihood of death, as defined by line coagulation assays, as well as decreased transfusion
APACHE II scores of < 25 or single-organ failure. A requirements, after administration of rFVIIa.89–93 The
randomized, double-blind, placebo-controlled trial of doses of rFVIIa have varied widely in these studies,
rhAPC in this population was recently reported by ranging from 36 to 218 mg/kg.
Two parallel randomized, placebo-controlled, Coagulopathy is a common cause of morbidity

double-blind, multicenter trials evaluating the safety and mortality in patients with end-stage liver disease
and efficacy of rFVIIa as an adjunctive hemostatic agent (ESLD).98 Despite limited literature to guide therapy,
in severely bleeding trauma patients were recently re- rFVIIa has been used for both treatment and prophylaxis
ported by Boffard et al.94 One hundred forty-three blunt of bleeding in patients with ESLD. Recombinant FVIIa
trauma patients and 134 penetrating trauma patients has been shown to temporarily correct a prolonged PT in
with severe bleeding randomly received either rFVIIa or nonbleeding patients with advanced cirrhosis.99 Dura-
placebo. The first dose of rFVIIa, 200 mg/kg, was tion of PT correction was dose dependent; the mean PT
administered following transfusion of the eighth unit normalized for 2 hours, 6 hours, and 12 hours following
of red blood cells (RBCs), with additional doses of rFVIIa doses of 5 mg/kg, 20 mg/kg, and 80 mg/kg,
100 mg/kg delivered 1 and 3 hours later. The primary respectively. Recombinant FVIIa has also been shown
end point was RBC transfusion requirements within to improve coagulation parameters in patients with
48 hours of the first dose of rFVIIa. In the blunt trauma fulminant hepatic failure.100,101 In a retrospective study
study, RBC transfusions were reduced by 2.6 units in the by Shami et al, patients with fulminant hepatic failure
rFVIIa-treated patients as compared with placebo who were given rFVIIa and FFP (seven patients) were
(p ¼ .02); the need for massive transfusion (> 20 RBC compared with those who received FFP alone (eight
units) was also significantly reduced (14% vs 33% of patients).100 Those in the rFVIIa group were able to
patients, p ¼ .03). Similar trends were noted in the undergo placement of intracranial pressure monitors
penetrating trauma patients, although these did not more frequently, had less anasarca, and demonstrated
meet statistical significance. Of note, no significant improved survival compared with those patients given
differences were observed between treatment arms in only FFP.
either study with respect to transfusion of other blood The efficacy and safety of rFVIIa in patients with
products including platelets, FFP, and cryoprecipitate. cirrhosis and upper gastrointestinal bleeding (UGIB)
There was no difference in the incidence of adverse was recently evaluated in a randomized, double-blind,
events, including thromboembolic complications, be- placebo-controlled trial.102 Two hundred forty-five cir-
tween the treatment groups in either study. Although rhotic patients with active UGIB were randomized to
there was a trend toward improved clinical outcomes eight doses of rFVIIa at 100 mg/kg or placebo, in
with rFVIIa, such as 30-day mortality and the develop- addition to standard treatment measures. Although
ment of multiple organ failure, these differences were normalization of PT occurred in the majority of patients
not statistically significant. However, the study was not in the rFVIIa arm, the study observed no benefit to
powered to evaluate these end points. rFVIIa in terms of the composite primary end point,
Patients with acute intracerebral hemorrhage which included failure to control bleeding within
(ICH) are at significant risk of morbidity and mortality, 24 hours of the initial dose, failure to prevent rebleeding
due in part to hematoma expansion caused by continued between 24 hours and day 5, or death within 5 days. In
bleeding or rebleeding within the first few hours after addition, there was no treatment effect with respect to
symptom onset.95 Early intervention with rFVIIa has RBC transfusion requirement, the number of elective or
been evaluated in patients with ICH in an effort to arrest emergent procedures performed, the length of stay in the
hematoma growth and improve outcomes in this pop- ICU or hospital, or in 5 or 42 day mortality rates. The
ulation. Following a phase II dose-escalation safety incidence of adverse events, including thromboembolic
trial,96 Mayer et al randomly assigned 399 patients events, was the same in both groups.
with spontaneous ICH diagnosed within 3 hours of Lastly, rFVIIa has been investigated as a prophy-
symptom onset to placebo or rFVIIa at doses of 40, lactic measure in patients with ESLD undergoing specific
80, or 160 mg/kg, administered within 1 hour of diag- procedures such as liver biopsy as well as in liver trans-
nosis.97 The percent change in intracerebral hematoma plantation.103–106 In a multicenter, randomized, double-
volume at 24 hours and clinical outcomes at 90 days were blind study, a single dose of rFVIIa, ranging from 5 to
measured. Patients in the placebo arm experienced a 120 mg/kg, was administered to 71 cirrhotic patients prior
significantly greater increase in hematoma volume than to laparoscopic liver biopsy.103 The PT normalized tran-
those patients who received rFVIIa (29% in the placebo siently in most patients; a longer duration of correction
group vs 16, 14, and 11% in groups given 40, 80, and was observed with higher doses. There was no correla-
160 mg/kg of rFVIIa, respectively). Mortality at 90 days tion, however, between the time to bleeding cessation
was significantly improved in the three rFVIIa groups postprocedure and the dose of rFVIIa received. Two
combined as compared with placebo (18% vs 29%, patients experienced thrombotic events, although the
p ¼ .02). Serious thromboembolic adverse events, in- authors concluded that these were not clearly related
cluding myocardial infarction and cerebral infarction, to the administration of rFVIIa. Lodge et al conducted
occurred in 7% of the rFVIIa-treated patients and in a multicenter, randomized, double-blind, placebo-
2% of patients in the placebo arm (p ¼ .12). controlled trial to evaluate the effect of rFVIIa in
182 patients with cirrhosis undergoing orthotopic liver 13. Babcock RB, Dumper CW, Scharfman WB. Heparin-
transplantation.106 Although rFVIIa significantly re- induced immune thrombocytopenia. N Engl J Med 1976;
duced the number of patients needing RBC transfusion, 295:237–241
14. Warkentin TE, Levine MN, Hirsh J, et al. Heparin-induced
there was no improvement in comparison to placebo with
thrombocytopenia in patients treated with low-molecular-
regard to the number of units of RBCs transfused, weight heparin or unfractionated heparin. N Engl J Med
intraoperative blood loss, hospitalization rate, total sur- 1995;332:1330–1335
gery time, or the proportion of patients requiring retrans- 15. Warkentin TE. Heparin-induced thrombocytopenia: patho-
plantation. Additionally, there was no difference in the genesis and management. Br J Haematol 2003;121:535–555
rate of thromboembolic events between the two groups. 16. Francis JL. Detection and significance of heparin-platelet
In summary, emerging data suggest that rFVIIa factor 4 antibodies. Semin Hematol 2005;42(3, Suppl 3):
S9–14
may be an effective drug for control of bleeding in certain
17. Rauova L, Poncz M, McKenzie SE, et al. Ultralarge complexes
nonhemophilic populations. At present, there is insuffi- of PF4 and heparin are central to the pathogenesis of heparin-
cient information to recommend the use of rFVIIa in induced thrombocytopenia. Blood 2005;105:131–138
many off-label scenarios, particularly in light of concerns 18. Visentin GP, Ford SE, Scott JP, Aster RH. Antibodies from
regarding its cost and unclear risk:benefit ratio. Further patients with heparin-induced thrombocytopenia/thrombo-
randomized data are needed to clarify the efficacy, safety, sis are specific for platelet factor 4 complexed with heparin or
and cost-effectiveness of rFVIIa in a variety of clinical bound to endothelial cells. J Clin Invest 1994;93:81–88
19. Girolami B, Prandoni P, Stefani PM, et al. The incidence of
settings.
heparin-induced thrombocytopenia in hospitalized medical
patients treated with subcutaneous unfractionated heparin: a
prospective cohort study. Blood 2003;101:2955–2959
REFERENCES 20. Prandoni P, Siragusa S, Girolami B, Fabris F. The incidence
of heparin-induced thrombocytopenia in medical patients
1. Bonfiglio MF, Traeger SM, Kier KL, Martin BR, Hulisz treated with low molecular weight heparin. Blood 2005;
DT, Verbeck SR. Thrombocytopenia in intensive care 106(9):3049–3054
patients: a comprehensive analysis of risk factors in 314 21. Wallis DE, Workman DL, Lewis BE, Steen L, Pifarre R,
patients. Ann Pharmacother 1995;29:835–842 Moran JF. Failure of early heparin cessation as treatment for
2. Hanes SD, Quarles DA, Boucher BA. Incidence and risk heparin-induced thrombocytopenia. Am J Med 1999;106:
factors of thrombocytopenia in critically ill trauma patients. 629–635
Ann Pharmacother 1997;31:285–289 22. Warkentin TE, Kelton JG. A 14-year study of heparin-
3. Baughman RP, Lower EE, Flessa HC, Tollerud DJ. induced thrombocytopenia. Am J Med 1996;101:502–507
Thrombocytopenia in the intensive care unit. Chest 1993; 23. Wester JP, Haas FJ, Biesma DH, Leusink JA, Veth G.
104:1243–1247 Thrombosis and hemorrhage in heparin-induced thrombo-
4. Vanderschueren S, De WA, Malbrain M, et al. Thrombo- cytopenia in seriously ill patients. Intensive Care Med 2004;
cytopenia and prognosis in intensive care. Crit Care Med 30:1927–1934
2000;28:1871–1876 24. Bartholomew JR. The incidence and clinical features of
5. Stephan F, Hollande J, Richard O, Cheffi A, Maier- heparin-induced thrombocytopenia. Semin Hematol 2005;
Redelsperger M, Flahault A. Thrombocytopenia in a surgical 42(3, Suppl 3):S3–S8
ICU. Chest 1999;115:1363–1370 25. Warkentin TE, Sheppard JA, Horsewood P, Simpson PJ,
6. Strauss R, Wehler M, Mehler K, Kreutzer D, Koebnick C, Moore JC, Kelton JG. Impact of the patient population on
Hahn EG. Thrombocytopenia in patients in the medical the risk for heparin-induced thrombocytopenia. Blood 2000;
intensive care unit: bleeding prevalence, transfusion require- 96:1703–1708
ments, and outcome. Crit Care Med 2002;30:1765–1771 26. Verma AK, Levine M, Shalansky SJ, Carter CJ, Kelton JG.
7. Chakraverty R, Davidson S, Peggs K, Stross P, Garrard C, Frequency of heparin-induced thrombocytopenia in critical
Littlewood TJ. The incidence and cause of coagulopathies in care patients. Pharmacotherapy 2003;23:745–753
an intensive care population. Br J Haematol 1996;93:460– 27. Martel N, Lee J, Wells PS. Risk of heparin induced
463 thrombocytopenia with unfractionated and low molecular
8. Cawley MJ, Wittbrodt ET, Boyce EG, Skaar DJ. Potential weight heparin thromboprophylaxis: a meta-analysis. Blood
risk factors associated with thrombocytopenia in a surgical 2005;106(8):2710–2715
intensive care unit. Pharmacotherapy 1999;19:108–113 28. Warkentin TE. Heparin-induced thrombocytopenia: patho-
9. Shalansky SJ, Verma AK, Levine M, Spinelli JJ, Dodek PM. genesis and management. Br J Haematol 2003;121:535–555
Risk markers for thrombocytopenia in critically ill patients: a 29. Warkentin TE, Greinacher A. Heparin-induced thrombo-
prospective analysis. Pharmacotherapy 2002;22:803–813 cytopenia: recognition, treatment, and prevention: the
10. Bogdonoff DL, Williams ME, Stone DJ. Thrombocytope- Seventh ACCP Conference on Antithrombotic and Throm-
nia in the critically ill patient. J Crit Care 1990;5:186–205 bolytic Therapy. Chest 2004;126(Suppl 3):311S–337S
11. Drews RE, Weinberger SE. Thrombocytopenic disorders 30. Warkentin TE. New approaches to the diagnosis of heparin-
in critically ill patients. Am J Respir Crit Care Med 2000; induced thrombocytopenia. Chest 2005;127(Suppl 2):35S–
162(2 Pt 1):347–351 45S
12. Bizzaro N. EDTA-dependent pseudothrombocytopenia: a 31. Warkentin TE, Kelton JG. Temporal aspects of heparin-
clinical and epidemiological study of 112 cases, with 10-year induced thrombocytopenia. N Engl J Med 2001;344:1286–
follow-up. Am J Hematol 1995;50:103–109 1292
32. Warkentin TE, Roberts RS, Hirsh J, Kelton JG. Heparin- 49. Ansell J, Hirsh J, Poller L, Bussey H, Jacobson A, Hylek E.
induced skin lesions and other unusual sequelae of the The pharmacology and management of the vitamin K
heparin-induced thrombocytopenia syndrome: a nested antagonists: the Seventh ACCP Conference on Antithrom-
cohort study. Chest 2005;127:1857–1861 botic and Thrombolytic Therapy. Chest 2004;126(Suppl 3):
33. Warkentin TE. Platelet count monitoring and laboratory 204S–233S
testing for heparin-induced thrombocytopenia. Arch Pathol 50. Esmon CT. The impact of the inflammatory response on
Lab Med 2002;126:1415–1423 coagulation. Thromb Res 2004;114:321–327
34. Kuo KH, Kovacs MJ. Fondaparinux: a potential new therapy 51. Levi M. Current understanding of disseminated intravas-
for HIT. Hematology (Am Soc Hematol Educ Program) cular coagulation. Br J Haematol 2004;124:567–576
2005;10:271–275 52. Hambleton J, Leung LL, Levi M. Coagulation: consultative
35. Rice L. Evolving management strategies for heparin-induced hemostasis. Hematology (Am Soc Hematol Educ Program)
thrombocytopenia. Semin Hematol 2005;42(3, Suppl 3): 2002:335–352
S15–S21 53. Mesters RM, Mannucci PM, Coppola R, Keller T,
36. Crowther MA, McDonald E, Johnston M, Cook D. Ostermann H, Kienast J. Factor VIIa and antithrombin III
Vitamin K deficiency and D-dimer levels in the intensive activity during severe sepsis and septic shock in neutropenic
care unit: a prospective cohort study. Blood Coagul patients. Blood 1996;88:881–886
Fibrinolysis 2002;13:49–52 54. Mesters RM, Helterbrand J, Utterback BG, et al. Prognostic
37. Staudinger T, Locker GJ, Frass M. Management of acquired value of protein C concentrations in neutropenic patients at
coagulation disorders in emergency and intensive-care high risk of severe septic complications. Crit Care Med
medicine. Semin Thromb Hemost 1996;22:93–104 2000;28:2209–2216
38. Alperin JB. Coagulopathy caused by vitamin K deficiency in 55. Yan SB, Helterbrand JD, Hartman DL, Wright TJ, Bernard
critically ill, hospitalized patients. JAMA 1987;258:1916– GR. Low levels of protein C are associated with poor
1919 outcome in severe sepsis. Chest 2001;120:915–922
39. Cohen H, Scott SD, Mackie IJ, et al. The development of 56. Fourrier F, Chopin C, Goudemand J, et al. Septic shock,
hypoprothrombinaemia following antibiotic therapy in multiple organ failure, and disseminated intravascular
malnourished patients with low serum vitamin K1 levels. coagulation: compared patterns of antithrombin III, protein
Br J Haematol 1988;68:63–66 C, and protein S deficiencies. Chest 1992;101:816–823
40. Krasinski SD, Russell RM, Furie BC, Kruger SF, Jacques 57. Biemond BJ, Levi M, ten Cate H, et al. Plasminogen
PF, Furie B. The prevalence of vitamin K deficiency in activator and plasminogen activator inhibitor I release during
chronic gastrointestinal disorders. Am J Clin Nutr 1985; experimental endotoxaemia in chimpanzees: effect of inter-
41:639–643 ventions in the cytokine and coagulation cascades. Clin Sci
41. O’Shaughnessy DF, Atterbury C, Bolton MP, et al. Guide- (Lond) 1995;88:587–594
lines for the use of fresh-frozen plasma, cryoprecipitate and 58. Mesters RM, Florke N, Ostermann H, Kienast J. Increase of
cryosupernatant. Br J Haematol 2004;126:11–28 plasminogen activator inhibitor levels predicts outcome of
42. Penning-van Beest FJ, van Meegan E, Rosendaal FR, leukocytopenic patients with sepsis. Thromb Haemost 1996;
Stricker BH. Characteristics of anticoagulant therapy and 75:902–907
comorbidity related to overanticoagulation. Thromb Hae- 59. Watanabe R, Wada H, Miura Y, et al. Plasma levels of total
most 2001;86:569–574 plasminogen activator inhibitor-I (PAI-I) and tPA/PAI-1
43. Raj G, Kumar R, McKinney WP. Time course of reversal of complex in patients with disseminated intravascular coagula-
anticoagulant effect of warfarin by intravenous and sub- tion and thrombotic thrombocytopenic purpura. Clin Appl
cutaneous phytonadione. Arch Intern Med 1999;159:2721– Thromb Hemost 2001;7:229–233
2724 60. Kobayashi T, Terao T, Maki M, Ikenoue T. Diagnosis and
44. Crowther MA, Douketis JD, Schnurr T, et al. Oral vitamin management of acute obstetrical DIC. Semin Thromb
K lowers the international normalized ratio more rapidly Hemost 2001;27:161–167
than subcutaneous vitamin K in the treatment of warfarin- 61. Gando S. Disseminated intravascular coagulation in trauma
associated coagulopathy: a randomized, controlled trial. Ann patients. Semin Thromb Hemost 2001;27:585–592
Intern Med 2002;137:251–254 62. Barbui T, Falanga A. Disseminated intravascular coagulation
45. Wilson SE, Watson HG, Crowther MA. Low-dose oral in acute leukemia. Semin Thromb Hemost 2001;27:593–
vitamin K therapy for the management of asymptomatic 604
patients with elevated international normalized ratios: a brief 63. Toh CH, Samis J, Downey C, et al. Biphasic transmittance
review. CMAJ 2004;170:821–824 waveform in the APTT coagulation assay is due to the
46. Watson HG, Baglin T, Laidlaw SL, Makris M, Preston FE. formation of a Ca( þþ )-dependent complex of C-reactive
A comparison of the efficacy and rate of response to oral and protein with very-low-density lipoprotein and is a novel
intravenous vitamin K in reversal of over-anticoagulation marker of impending disseminated intravascular coagulation.
with warfarin. Br J Haematol 2001;115:145–149 Blood 2002;100:2522–2529
47. Hung A, Singh S, Tait RC. A prospective randomized study 64. Fernandes B, Giles A. An abnormal activated partial
to determine the optimal dose of intravenous vitamin K in thromboplastin time clotting waveform is associated with
reversal of over-warfarinization. Br J Haematol 2000;109: high mortality and a procoagulant state. Lab Hematol 2003;
537–539 9:138–142
48. Riegert-Johnson DL, Volcheck GW. The incidence of 65. Yu M, Nardella A, Pechet L. Screening tests of dissemi-
anaphylaxis following intravenous phytonadione (vitamin nated intravascular coagulation: guidelines for rapid and
K1): a 5-year retrospective review. Ann Allergy Asthma specific laboratory diagnosis. Crit Care Med 2000;28:
Immunol 2002;89:400–406 1777–1780
66. Levi M, de Jonge E, van der Poll T, ten Cate H. Disseminated 83. Ely EW, Laterre PF, Angus DC, et al. Drotrecogin alfa
intravascular coagulation. Thromb Haemost 1999;82:695–705 (activated) administration across clinically important sub-
67. Horan JT, Francis CW. Fibrin degradation products, fibrin groups of patients with severe sepsis. Crit Care Med 2003;
monomer and soluble fibrin in disseminated intravascular 31:12–19
coagulation. Semin Thromb Hemost 2001;27:657–666 84. Abraham E, Laterre PF, Garg R, et al. Drotrecogin alfa
68. Taylor FB Jr, Toh CH, Hoots WK, Wada H, Levi M. (activated) for adults with severe sepsis and a low risk of
Towards definition, clinical and laboratory criteria, and a death. N Engl J Med 2005;353:1332–1341
scoring system for disseminated intravascular coagulation. 85. Roberts HR, Monroe DM, White GC. The use of
Thromb Haemost 2001;86:1327–1330 recombinant factor VIIa in the treatment of bleeding
69. Bakhtiari K, Meijers JC, de Jonge E, Levi M. Prospective disorders. Blood 2004;104:3858–3864
validation of the International Society of Thrombosis and 86. Hoffman M, Monroe DM III. The action of high-dose
Haemostasis scoring system for disseminated intravascular factor VIIa (FVIIa) in a cell-based model of hemostasis.
coagulation. Crit Care Med 2004;32:2416–2421 Semin Hematol 2001;38(4, Suppl 12):6–9
70. Slofstra SH, van’t Veer C, Buurman WA, Reitsma PH, ten 87. Monroe DM, Hoffman M, Oliver JA, Roberts HR. Platelet
Cate H, Spek CA. Low molecular weight heparin attenuates activity of high-dose factor VIIa is independent of tissue
multiple organ failure in a murine model of disseminated factor. Br J Haematol 1997;99:542–547
intravascular coagulation. Crit Care Med 2005;33:1365–1370 88. Kenet G, Walden R, Eldad A, Martinowitz U. Treatment of
71. du Toit HJ, Coetzee AR, Chalton DO. Heparin treatment traumatic bleeding with recombinant factor VIIa. Lancet
in thrombin-induced disseminated intravascular coagulation 1999;354:1879
in the baboon. Crit Care Med 1991;19:1195–1200 89. Martinowitz U, Kenet G, Lubetski A, Luboshitz J, Segal E.
72. Feinstein DI. Diagnosis and management of disseminated Possible role of recombinant activated factor VII (rFVIIa) in
intravascular coagulation: the role of heparin therapy. Blood the control of hemorrhage associated with massive trauma.
1982;60:284–287 Can J Anaesth 2002;49:S15–S20
73. Okabayashi K, Wada H, Ohta S, Shiku H, Nobori T, 90. Martinowitz U, Kenet G, Segal E, et al. Recombinant
Maruyama K. Hemostatic markers and the sepsis-related activated factor VII for adjunctive hemorrhage control in
organ failure assessment score in patients with disseminated trauma. J Trauma 2001;51:431–438
intravascular coagulation in an intensive care unit. Am J 91. Dutton RP, Hess JR, Scalea TM. Recombinant factor VIIa
Hematol 2004;76:225–229 for control of hemorrhage: early experience in critically ill
74. Baudo F, Caimi TM, de CF, et al. Antithrombin III (ATIII) trauma patients. J Clin Anesth 2003;15:184–188
replacement therapy in patients with sepsis and/or post- 92. Dutton RP, McCunn M, Hyder M, et al. Factor VIIa for
surgical complications: a controlled double-blind, rando- correction of traumatic coagulopathy. J Trauma 2004;57:
mized, multicenter study. Intensive Care Med 1998;24:336– 709–718
342 93. Geeraedts LM Jr, Kamphuisen PW, Kaasjager HA, Verwiel
75. Eisele B, Lamy M, Thijs LG, et al. Antithrombin III in JM, van Vugt AB, Frolke JP. The role of recombinant factor
patients with severe sepsis: a randomized, placebo-con- VIIa in the treatment of life-threatening haemorrhage in
trolled, double-blind multicenter trial plus a meta-analysis blunt trauma. Injury 2005;36:495–500
on all randomized, placebo-controlled, double-blind trials 94. Boffard KD, Riou B, Warren B, et al. Recombinant factor
with antithrombin III in severe sepsis. Intensive Care Med VIIa as adjunctive therapy for bleeding control in severely
1998;24:663–672 injured trauma patients: two parallel randomized, placebo-
76. Warren BL, Eid A, Singer P, et al. Caring for the critically ill controlled, double-blind clinical trials. J Trauma 2005;59:8–
patient: high-dose antithrombin III in severe sepsis: a 15
randomized controlled trial. JAMA 2001;286:1869–1878 95. Qureshi AI, Tuhrim S, Broderick JP, Batjer HH, Hondo H,
77. Abraham E, Reinhart K, Svoboda P, et al. Assessment of the Hanley DF. Spontaneous intracerebral hemorrhage. N Engl
safety of recombinant tissue factor pathway inhibitor in J Med 2001;344:1450–1460
patients with severe sepsis: a multicenter, randomized, 96. Mayer SA, Brun NC, Broderick J, et al. Safety and feasibility
placebo-controlled, single-blind, dose escalation study. Crit of recombinant factor VIIa for acute intracerebral hemor-
Care Med 2001;29:2081–2089 rhage. Stroke 2005;36:74–79
78. Abraham E. Tissue factor inhibition and clinical trial results 97. Mayer SA, Brun NC, Begtrup K, et al. Recombinant
of tissue factor pathway inhibitor in sepsis. Crit Care Med activated factor VII for acute intracerebral hemorrhage. N
2000;28(Suppl 9):S31–S33 Engl J Med 2005;352:777–785
79. Abraham E, Reinhart K, Opal S, et al. Efficacy and safety of 98. Kujovich JL. Hemostatic defects in end stage liver disease.
tifacogin (recombinant tissue factor pathway inhibitor) in Crit Care Clin 2005;21:563–587
severe sepsis: a randomized controlled trial. JAMA 2003; 99. Bernstein DE, Jeffers L, Erhardtsen E, et al. Recombinant
290:238–247 factor VIIa corrects prothrombin time in cirrhotic patients: a
80. Esmon CT. The protein C pathway. Chest 2003;124(Suppl 3): preliminary study. Gastroenterology 1997;113:1930–1937
26S–32S 100. Shami VM, Caldwell SH, Hespenheide EE, Arseneau KO,
81. Bernard GR, Ely EW, Wright TJ, et al. Safety and dose Bickston SJ, Macik BG. Recombinant activated factor VII
relationship of recombinant human activated protein C for for coagulopathy in fulminant hepatic failure compared with
coagulopathy in severe sepsis. Crit Care Med 2001;29:2051– conventional therapy. Liver Transpl 2003;9:138–143
2059 101. Pavese P, Bonadona A, Beaubien J, et al. FVIIa corrects the
82. Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and coagulopathy of fulminant hepatic failure but may be
safety of recombinant human activated protein C for severe associated with thrombosis: a report of four cases. Can J
sepsis. N Engl J Med 2001;344:699–709 Anaesth 2005;52:26–29
102. Bosch J, Thabut D, Bendtsen F, et al. Recombinant factor for colonic polypectomies in patients with cirrhosis and
VIIa for upper gastrointestinal bleeding in patients with coagulopathy. Dig Dis Sci 2003;48:1414–1424
cirrhosis: a randomized, double-blind trial. Gastroenterology 105. Hendriks HG, Meijer K, de Wolf JT, et al. Reduced
2004;127:1123–1130 transfusion requirements by recombinant factor VIIa in
103. Jeffers L, Chalasani N, Balart L, Pyrsopoulos N, Erhardtsen orthotopic liver transplantation: a pilot study. Transplanta-
E. Safety and efficacy of recombinant factor VIIa in patients tion 2001;71:402–405
with liver disease undergoing laparoscopic liver biopsy. 106. Lodge JP, Jonas S, Jones RM, et al. Efficacy and safety
Gastroenterology 2002;123:118–126 of repeated perioperative doses of recombinant factor
104. Anantharaju A, Mehta K, Mindikoglu AL, Van Thiel DH. VIIa in liver transplantation. Liver Transpl 2005;11:973–
Use of activated recombinant human factor VII (rhFVIIa) 979
Major Complications following
Hematopoietic Stem Cell Transplantation
Bekele Afessa, M.D.1 and Steve G. Peters, M.D.1
ABSTRACT
Tens of thousands of patients undergo hematopoietic stem cell transplantation

(HSCT) annually, 15 to 40% of whom are admitted to the intensive care unit. Pulmonary
complications are the most life threatening conditions that develop in HSCT recipients.
Both infectious and noninfectious complications occur more frequently in allogeneic
HSCT. The management of HSCT recipients requires knowledge of their immune status,
appropriate diagnostic evaluation, and early treatment. During the preengraftment phase
(0 to 30 days after transplant), the most prevalent pathogens causing infection are bacteria
and Candida species and, if the neutropenia persists, Aspergillus species. The early
postengraftment phase (30 to 100 days) is characterized by cytomegalovirus (CMV),
Pneumocystis jiroveci, and Aspergillus infections. During the late posttransplant phase
(> 100 days), allogeneic HSCT recipients are at risk for CMV, community-acquired
respiratory virus, and encapsulated bacterial infections. Antigen and polymerase chain
reaction assays are important for the diagnosis of CMV and Aspergillus infections. Diffuse
alveolar hemorrhage (DAH) and periengraftment respiratory distress syndrome occur in
both allogeneic and autologous HSCT recipients, usually during the first 30 days.
Bronchiolitis obliterans occurs exclusively in allogeneic HSCT recipients with graft versus
host disease. Idiopathic pneumonia syndrome occurs at any time following transplant.
Bronchoscopy is usually helpful for the diagnosis of the infectious pulmonary complications
and DAH.
KEYWORDS: Aspergillosis, bone marrow transplantation, cytomegalovirus infection,

diffuse alveolar hemorrhage, idiopathic pneumonia syndrome, periengraftment
respiratory distress syndrome, pneumonia, respiratory insufficiency
Tens of thousands of patients undergo hemato- 1

disease.1 As a result of life-threatening multiple organ
poietic stem cell transplantation (HSCT) annually. In dysfunctions, 15 to 40% of HSCT recipients receive
allotransplantation, the 100 day mortality rate ranges intensive care unit support, the majority of whom require
between 10 and 40% and the main causes of death are mechanical ventilation.2–4 The mortality rate of HSCT
graft versus host disease (GVHD), interstitial pneumo- recipients receiving invasive ventilation used to exceed
nitis, and multiple organ failure.1 In autotransplantation, 90%.2,5 Although more recent studies have shown im-
the 100 day mortality ranges between 5 and 20% and the provement in outcome, the mortality rate of HSCT re-
main cause of death is recurrence of the underlying cipients receiving mechanical ventilation is still high.3,6,7
1
Division of Pulmonary and Critical Care Medicine, Department of Non-pulmonary Critical Care: Managing Multisystem Critical Illness;
Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota. Guest Editor, Curtis N. Sessler, M.D.
Address for correspondence and reprint requests: Bekele Afessa, Semin Respir Crit Care Med 2006;27:297–309. Copyright # 2006
M.D., Division of Pulmonary and Critical Care Medicine, Mayo by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Clinic College of Medicine, 200 First St., SW, Rochester, MN NY 10001, USA. Tel: +1(212) 584-4662.
55905. E-mail: [email protected]. DOI 10.1055/s-2006-945530. ISSN 1069-3424.
297
This review describes the major post-HSCT complica- Table 1 Major Pulmonary Complications in
tions pertinent to pulmonary and critical care physicians. Hematopoietic Stem Cell Transplant Recipients
The management of HSCT recipients requires Infectious
knowledge of their immune status, appropriate diagnostic Viral
evaluation, and early treatment. The conditioning regi- Cytomegalovirus
men severely depresses preexisting immunity, which Respiratory syncytial virus
recovers along predictable patterns after transplantation.8 Influenza B
Bacterial
Gram-positive
TIMING AND TYPES OF PULMONARY Staphylococcus aureus
COMPLICATIONS Streptococcus pneumoniae
Pulmonary complications, occurring in 30 to 60% of Gram-negative
recipients, are the most common life-threatening con- Pseudomonas aeruginosa
ditions that develop following HSCT. The complications Fungal
are more frequent in allogeneic recipients, especially Aspergillus spp.
those with GVHD. Both infectious and noninfectious Candida spp.
pulmonary complications occur frequently (Table 1) Pneumocystis jiroveci
(Fig. 1). Based on the immunosuppression status, the Noninfectious
posttransplant period is divided into three phases: preen- Acute pulmonary edema
graftment, early posttransplant, and late posttransplant.9 Diffuse alveolar hemorrhage
The preengraftment phase (0 to 30 days) is characterized Periengraftment respiratory distress syndrome
by neutropenia and breaks in the mucocutaneous barriers Bronchiolitis obliterans syndrome
as a result of conditioning regimens and frequent vascular Bronchiolitis obliterans organizing pneumonia
catheterization. During this phase, the most prevalent Idiopathic pulmonary syndrome
pathogens causing infection are bacteria and Candida Delayed pulmonary toxicity syndrome
species and, if the neutropenia persists, Aspergillus species. Pulmonary cytolytic thrombotic syndrome
During neutropenia, there is no significant difference in
the type of infection between allogeneic and autologous
HSCT recipients.10 The early postengraftment phase
Figure 1 Timing of the major infectious and noninfectious complications following hematopoietic stem cell transplantation. BO,
bronchiolitis obliterans; DAH, diffuse alveolar hemorrhage; GVHD, graft versus host disease; IPS, idiopathic pneumonia syndrome; P
edema, pulmonary edema; PERDS, periengraftment respiratory distress syndrome; PCP, Pneumocystis jiroveci pneumonia; RSV,
respiratory syncytial virus. Phase I, preengraftment period; Phase II, early postengraftment period; Phase III, late postengraftment
period.
COMPLICATIONS FOLLOWING HEMATOPOIETIC STEM CELL TRANSPLANTATION/AFESSA, PETERS 299
(30 to 100 days) is dominated by impaired cell-mediated endemic seasons, respiratory syncytial virus, adenovirus,
immunity. The impact of this cell-mediated defect is picornavirus, influenza virus, and parainfluenza virus
determined by the development of GVHD and the should be included in the differential diagnoses of
immunosuppressant medications used to treat it. Cyto- respiratory symptoms in the HSCT recipient.19,20 This
megalovirus (CMV), Pneumocystis jiroveci, and Aspergillus section will focus on CMV.
species are the predominant pathogens during this phase.
The late posttransplant phase (> 100 days) is character-
ized by defects in cell-mediated and humoral immunity as CYTOMEGALOVIRUS PNEUMONIA
well as function of the reticuloendothelial system in
allogeneic transplant recipients. During this phase, allo- Epidemiology Although the frequency of CMV
geneic HSCT recipients are at risk for CMV infection, pneumonia in the early posttransplantation period has
varicella-zoster infection, Epstein-Barr–related lympho- been substantially reduced by prophylaxis, it continues to
proliferative disease, community-acquired respiratory vi- be a major cause of morbidity and mortality in the late
rus infection, and infection by encapsulated bacteria such posttransplant period. With early detection and prophy-
as Haemophilus influenzae and Streptococcus pneumoniae. In lactic and preemptive treatment, the rate of CMV
certain parts of the world, pulmonary tuberculosis occurs pneumonia has declined to less than 5%.21–23 Risk
during the late posttransplant phase.11,12 factors for CMV pneumonia include older age, positive
Noninfectious pulmonary complications also fol- CMV serology, allogeneic graft, and GVHD.24–28
low a characteristic time pattern.13 Pulmonary edema,
diffuse alveolar hemorrhage (DAH), and periengraft- Clinical Findings and Diagnostic Evaluation CMV
ment respiratory distress syndrome (PERDS) usually infections result from primary infection or reactivation,
occur during the first 30 days following transplant and most occur between 6 and 12 weeks after trans-
(Fig. 1). Idiopathic pneumonia syndrome (IPS) occurs plant.29,30 With the wider use of prophylactic therapy,
at any time following transplant. CMV pneumonia may occur beyond 100 days after
transplant.12,31 Although rare, CMV pneumonia may
develop prior to engraftment, especially in recipients
APPROACH TO PULMONARY with positive CMV serology.31 The clinical manifesta-
COMPLICATIONS tions of CMV infection vary from completely asympto-
When HSCT recipients present with pulmonary infil- matic to multiple organ dysfunction. HSCT recipients
trates and symptoms and signs of infection, most clini- with CMV pneumonia typically present with fever, non-
cians initiate empirical antibacterial therapy, adding productive cough, dyspnea, and hypoxemia.31
antifungal therapy if risk factors are present and there Plain chest radiographs and HRCT usually show
is no response to initial treatment.14 Cultures of blood, subtle patchy or diffuse ground-glass opacities, often
urine, and respiratory secretions should be obtained. In with small concurrent pulmonary nodules.29,32,33 CMV
the appropriate clinical setting, antigen and polymerase antigen and PCR assays are used for the early detection
chain reaction (PCR) assays for Aspergillus and CMV of infection in urine, blood, and respiratory secre-
may be helpful. Pulmonary function testing (PFT) and tions.34,35 In the appropriate clinical setting, the diag-
high-resolution computed tomography (HRCT) of the nosis of CMV pneumonia relies on the identification of
chest play important roles in suggesting specific diag- CMV from the lower respiratory tract by BAL and
nosis. If tolerated, we advocate early bronchoalveolar transbronchial or surgical lung biopsy. The definitive
lavage (BAL) with or without transbronchial lung bi- diagnosis requires positive culture results from BAL
opsy, transthoracic fine needle aspiration, or video-as- fluid samples, and identification of the characteristic
sisted thoracoscopic lung biopsy because specific cytopathic feature of intranuclear inclusions in either
diagnoses may lead to appropriate treatment and avoid BAL fluid or biopsy tissue.36
unnecessary and potentially harmful therapy (Fig. 2).
Prevention and Treatment The transfusion of CMV
seronegative blood products and leukocyte-depleted pla-
INFECTIOUS COMPLICATIONS telets, and prophylaxis and preemptive use of acyclovir,
valacyclovir, valganciclovir, ganciclovir, or foscarnet have
Viral Pneumonia reduced the rate of CMV infection in high-risk patients.
Viral infections are a major cause of morbidity and Although ganciclovir and foscarnet are effective for
mortality in HSCT recipients. CMV is the most com- CMV prophylaxis, their use is limited by marrow and
mon viral pathogen causing lower respiratory tract in- renal toxicities, and they are usually reserved for pre-
fection.15–17 Other viruses, including herpes simplex, emptive therapy or treatment. For the treatment of
varicella-zoster, Epstein-Barr, and human herpesvirus CMV disease, intravenous immunoglobulin is usually
6 and 8 may also cause pulmonary infections.18 During used in combination with ganciclovir or foscarnet.37
Figure 2 Diagnostic approach to the hematopoietic stem cell transplant recipient with pulmonary complication. BAL, bronchoalveolar
lavage; HRCT, high-resolution computed tomography of the chest; PFT, pulmonary function test; TBB, transbronchial lung biopsy; VATS,
video-assisted thoracoscopic surgery.
Clinical Course/Prognosis The mortality rate associ- posed that the diagnoses of invasive fungal infections in
ated with untreated CMV pneumonia may exceed the immunocompromised host be labeled as ‘‘proven,’’
90%.25 Although the prognosis has improved as a result ‘‘probable,’’ and ‘‘possible,’’ based on clinical, radiolog-
of effective prophylactic and preemptive treatment in ical, microbiological, and histological findings.45 Despite
high-risk patients, mortality still exceeds 50%.25,38,39 the availability of newer antifungal agents, including the
azoles and echinocandins, there has been little success in
the prevention and treatment of Aspergillus infection in
Fungal Pneumonia HSCT recipients. The following is an overview of
With improved prevention and treatment of other pulmonary aspergillosis after HSCT.
infections, invasive fungi have become the leading
infectious cause of morbidity and mortality in HSCT
recipients.40 The main risk factors for the development PULMONARY ASPERGILLOSIS
of invasive fungal infections are neutropenia and the
development of GVHD. Candida spp. and Aspergillus Epidemiology The reported incidence of invasive as-
spp. are the most common fungal pathogens causing pergillosis in allogeneic recipients ranges between 3.6
infection in HSCT recipients.41 With the wider use and 28%. Because of its association with neutropenia and
of fluconazole prophylaxis, the incidence of Candida GVHD, the incidence has a bimodal distribution.41 In
infection has declined and Aspergillus has become the the absence of neutropenia, invasive aspergillosis is
dominant cause of fungal pneumonia.40,42–44 Endemic uncommon in autologous recipients.10,46
fungi rarely cause infection in HSCT recipients.41
There is substantial controversy regarding the Clinical Findings and Diagnostic Evaluation Asper-
optimal criteria for the diagnosis of invasive fungal gillus infection is usually acquired through inhalation. In
infections. An international consensus group has pro- the absence of adequate immune response, the organism
invades the sinuses or lung tissue and disseminates fungal prophylaxis has not been shown to be effective in
to other organs by hyphal invasion of blood vessels.41 preventing invasive aspergillosis.54 During the high-
The clinical presentations of pulmonary aspergillosis risk periods, granulocyte transfusion and oral itracona-
are nonspecific and include fever, cough, sputum pro- zole or intravenous amphotericin B has been used until
duction, and pleuritic chest pain.47,48 Hemoptysis is the neutropenia resolves or the Aspergillus colonization
uncommon. disappears.55–57
The radiographic findings of pulmonary asper-
gillosis include nodules, diffuse infiltrates, and cav- Clinical Course/Prognosis Despite therapy, the re-
itations.33,49 The ‘‘halo sign’’ consists of a nodule ported case fatality rate of invasive aspergillosis ranges
surrounded by ground-glass attenuation and is seen between 29 and 87%.46,47,58 One-year survival rate is
most often in neutropenic patients. Cavitation is a 20%.42
late finding. The ‘‘hypodense sign,’’ the presence of
low density within nodules or areas of consolidation, PNEUMOCYSTIS JIROVECI PNEUMONIA
is reported to have 30% sensitivity and 100% specif- The incidence of Pneumocystis jiroveci pneumonia (PCP)
icity for the diagnosis of invasive pulmonary aspergil- in HSCT recipients has decreased sharply with trime-
losis.49 thoprim-sulfamethoxazole (TMP-SMZ) prophylaxis.
In high-risk patients, Aspergillus antigen and PCR In three recent studies of 2356 HSCT recipients, PCP
assays can be used to screen for and monitor the response developed in 33 (1.4%), all in patients who were not
to treatment of Aspergillus infection. However, both the receiving TMP-SMZ prophylaxis.12,16,40 No PCP was
antigen and PCR assays may have high false-positive identified in two recent autopsy studies of 121 HSCT
rates.50 recipients.59,60
Because of the high mortality associated with The radiographic findings of PCP consist of a
invasive aspergillosis, many clinicians use antifungal diffuse, bilateral reticular or granular pattern that can
drugs empirically. The firm diagnosis of aspergillosis in progress to alveolar consolidation.16 The diagnosis usu-
HSCT recipients requires tissue for histology and cul- ally requires BAL with or without transbronchial lung
ture. If the histopathological examination of lung tissue biopsy.17,61,62
shows Aspergillus, the diagnosis of Aspergillus pneumonia The drug of choice for both the prevention
is considered definitive.45 In the absence of the typical and the treatment of PCP is TMP-SMZ. For patients
histopathological findings, the presence of one micro- who do not tolerate TMP-SMZ, dapsone, aerosolized
biological criterion and one major or two minor clinical pentamidine, and atovaquone are alternatives for
criteria are required for the diagnosis of probable, and prophylaxis, although they are not as effective as
one microbiological or one major or two minor criteria TMP-SMZ.52
for possible Aspergillus pneumonia.45 The microbiolog-
ical criteria include positive Aspergillus culture from PULMONARY CANDIDIASIS
sputum or BAL fluid or Aspergillus antigen from BAL Candida infections, especially of the bloodstream, are
fluid. Chest computed tomographic (CT) findings of common during the neutropenic phase.63,64 The portal
halo sign, air-crescent sign, or cavity within an area of of entry is usually the gastrointestinal tract or indwelling
consolidation are considered major criteria. The minor central venous catheter. Isolated Candida pneumonia is
criteria include cough, chest pain, hemoptysis, dyspnea, uncommon. In one study of 1359 HSCT recipients, only
pleural rub, any pulmonary infiltrates not fulfilling the one late-onset pulmonary candidiasis was seen.40 Autol-
major criteria, and pleural effusion. ogous HSCT recipients are at low risk for fungal
infection after neutrophil recovery. However, Candida
Prevention and Treatment Because pulmonary asper- infection remains a major threat to patients receiving
gillosis is associated with high mortality, early initiation allogeneic transplant. Risk factors for Candida infections
of therapy is crucial in HSCT recipients with prolonged include increased age, prolonged neutropenia, GVHD,
neutropenia and GVHD. Voriconazole is currently a and use of radiation in the preparative regimen.63,64 The
first-line therapy. Amphotericin B, caspofungin, mica- radiographic findings predominantly consist of air-space
fungin, and itraconazole are alternatives. The optimal disease.16 Candida fungemia is responsive to early anti-
duration of antifungal therapy is unknown and depends fungal therapy and removal of foreign objects such as
on the extent of the infection, response to treatment, and central lines. Fluconazole prophylaxis will prevent in-
improvement of the underlying immune deficit.51 vasive disease due to susceptible Candida spp.52
Because spore inhalation is the usual route of
Aspergillus infection, nursing of HSCT recipients in BACTERIAL PNEUMONIA
high-efficiency particulate air (HEPA)-filtered rooms
is recommended for protection.52,53 The addition of Epidemiology Bacterial infections are common dur-
aerosolized amphotericin B inhalation to standard anti- ing the neutropenic period. However, because most of
the HSCT recipients receive prophylactic and empirical organisms in allogeneic recipients with chronic
therapy during this period, the exact incidence has not GVHD. The 23-valent pneumococcal polysaccharide
been well defined. In a recent study of 1359 HSCT vaccine should be administered at 12 to 24 months after
recipients from the Fred Hutchinson Cancer center, 58% transplant.66
of late-onset pneumonias were diagnosed clinically with-
out tissue or microbial documentation.40
The etiology of bacterial infection in neutropenic NON-INFECTIOUS PULMONARY
patients is in flux.65 Gram-negative bacteria, especially COMPLICATIONS
Pseudomonas aeruginosa, used to be the predominant
bacterial pathogens.66 However, with the use of prophy- Pulmonary Edema
lactic antibiotics against gram-negative infection, and Although pulmonary edema has been described as part
the use of intravascular catheters, gram-positive organ- of the capillary leak syndrome, it usually results from
isms are becoming more frequent.11,67 Persistent deficits large volumes of fluid and blood products infused during
in cellular and humoral immunity predispose HSCT conditioning and the immediate posttransplant pe-
recipients to intracellular pathogens such as Listeria riod.69,70 Characteristics include weight gain and echo-
monocytogenes and Nocardia species, and to encapsulated cardiographic findings of left ventricular end diastolic
organisms such as Streptococcus pneumoniae and Haemo- dilation.70 This condition differs from cardiogenic pul-
philus influenzae. monary edema, which is characterized by elevated brain
natriuretic protein level.71 Diagnostic criteria include
Clinical Findings and Diagnostic Evaluation The acute-onset dyspnea, crackles on examination, and re-
initial clinical symptoms of bacterial pneumonia in the duced arterial oxygen tension, in the absence of infec-
HSCT recipient may be subtle, but they usually evolve tion.70 Chest radiographs show bilateral diffuse ground-
quickly. Although some HSCT recipients with bacte- glass pulmonary opacities often accompanied by Kerley
rial pneumonia demonstrate nodules and masses on B lines.32 Cardiomegaly is usually absent. Pulmonary
chest radiograph, most present with focal consolida- edema can be prevented by fluid restriction and diuretic
tion.16,33 Uncommonly, the halo sign and cavitations therapy.70
may be seen. Pseudomonas and Legionella tend to cause a
multifocal consolidation.16 Blood and sputum cultures
are usually obtained, although the sputum may be Idiopathic Pneumonia Syndrome
nonspecific. If there is uncertainty in the diagnosis or
lack of response to empirical treatment, BAL may be EPIDEMIOLOGY
helpful. Idiopathic pneumonia syndrome (IPS) is characterized
by clinical features of pneumonia and diffuse lung
Treatment and Prevention If the causative organism injury in the absence of an identified infection. The
is identified, antibiotic therapy should focus on the overall frequency of IPS is 10%, with a reported range
specific pathogen. However, no organism is identified of 2 to 17%.13,40,72 The median time of onset of IPS is
in most HSCT recipients with suspected bacterial between 21 and 87 days. Risk factors for IPS include
pneumonia. The initial empirical antibiotic regimen old age, transplant for malignancy other than leuke-
should include coverage of Pseudomonas aeruginosa mia, pretransplant chemotherapy, total body irradia-
and methicillin-resistant Staphylococcus aureus. Because tion, GVHD, and positive donor cytomegalovirus
bacteria are carried on the hands, appropriate hand serology.13
washing should be practiced by all individuals in con-
tact with HSCT recipients.52 The data are insufficient CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION
to recommend routine use of gut decontamination and IPS is defined by the presence of widespread alveolar
systemic antibiotics for afebrile, asymptomatic patients. injury in the absence of lower respiratory tract infec-
However, some advocate using fluoroquinolone pro- tion characterized by the presence of acute, bilateral
phylaxis to reduce infection by gram-negative rods in pulmonary infiltrates, associated symptoms of cough
allogeneic HSCT recipients.68 Although growth fac- and dyspnea, hypoxemia, and restrictive physiology, in
tors shorten the duration of neutropenia after HSCT, the absence of infection or heart failure.73 The patho-
they do not reduce the attack rate of invasive bacterial genesis is not well defined.74–76 Lung tissue injury,
disease. Some advise giving intravenous immunoglo- inflammation, and cytokine release are implicated. The
bulin (IVIG) to prevent bacterial infection in HSCT clinical spectrum of IPS is broad, ranging from acute
recipients with unrelated graft and severe hypogamma- respiratory failure to incidental radiographic abnor-
globulinemia within 100 days after transplant. During malities.73 The usual presentations include dyspnea,
the late posttransplant phase, antibiotic prophylaxis is dry cough, hypoxemia, and nonlobar infiltrates.73 Be-
recommended to prevent infection with encapsulated cause IPS mimics infectious pneumonia, the majority
of the patients are on antibiotics at the time of CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION
diagnosis.77 Infection is excluded by the absence of a Symptoms of DAH typically include dyspnea, fever, and
pathogen in BAL fluid and by the lack of clinical cough.82 Hemoptysis is reported in fewer than 20% of
response to antimicrobial therapy. DAH and perien- patients.78,83 The onset of DAH is usually within the
graftment respiratory distress syndrome (PERDS) also first 30 (median between 11 and 24) days following
fulfill the diagnostic criteria of IPS.78,79 Despite over- HSCT.
lap in the clinical features of IPS, DAH, and PERDS, Arterial blood gas studies show hypoxemia. Chest
their responses to treatment and clinical courses are radiographs usually show alveolar and interstitial infil-
different.78 In our diagnostic approach to patients with trates involving middle and lower lung zones.84 The
suspected IPS, we perform BAL and, if there are no early radiographic changes may be subtle, unilateral, or
contraindications, transbronchial lung biopsy. We pro- asymmetrical.84 CT of the chest has a limited role in the
ceed to video-assisted thoracoscopic lung biopsy if diagnosis.
transbronchial lung biopsy is contraindicated or if Criteria for the diagnosis DAH include (1) signs
the transbronchial specimen is inadequate. Lung biop- and symptoms of pneumonitis, (2) no evidence of
sies of patients with IPS may show diffuse alveolar infection, and (3) BAL showing progressively bloodier
damage, organizing or acute pneumonia, and intersti- return from separate subsegmental bronchi, or > 20%
tial lymphocytic inflammation.77,80 hemosiderin-laden alveolar macrophages.13,78 Lung tis-
sues in DAH show diffuse alveolar damage.59,82
TREATMENT
There have been no randomized clinical trials address- TREATMENT
ing the treatment of IPS. Despite case reports of Based on retrospective studies, HSCT recipients with
patients with IPS responding to treatment with corti- DAH are treated with systemic corticosteroids.78,83 We
costeroids, studies with larger sample sizes have not commonly use methylprednisolone, 1 g daily in four
shown any outcome benefit.77,80,81 Currently, manage- divided doses for 5 days, followed by 1 mg/kg for 3 days,
ment consists of supportive care and prevention and tapering off over 2 to 4 weeks.78 There are case reports
treatment of infection. There is a report of three cases of allogeneic HSCT recipients with DAH successfully
of HSCT recipients with IPS whose lung function treated with recombinant factor VIIa.85,86
improved following the administration of etanercept.76
This observation awaits further confirmation by clinical CLINICAL COURSE AND PROGNOSIS
trial. The majority of HSCT recipients with DAH require
mechanical ventilation.78,83 The reported mortality
CLINICAL COURSE AND PROGNOSIS rate of DAH ranges between 48 and 100%.78,83 The
The clinical course of HSCT recipients with IPS is two most common reported causes of death had been
usually complicated by viral and fungal infections as multiple organ failure and sepsis.87,88 However, in a
well as pneumothorax, pneumomediastinum, subcuta- recent study, respiratory failure was the immediate
neous emphysema, pulmonary fibrosis, and autoimmune cause of death in the majority of HSCT recipients
polyserositis.13 The overall mortality of IPS is 74%, with with DAH.83
a reported range between 60 and 86%.13 The 1 year
survival rate is less than 15%.77,81 For those who require
mechanical ventilation, the hospital mortality exceeds Periengraftment Respiratory Distress
95%.77 Syndrome
EPIDEMIOLOGY
Diffuse Alveolar Hemorrhage PERDS occurs in 5% of autologous HSCT recipi-
ents.79 About one third of DAH occurs during the
EPIDEMIOLOGY periengraftment period and about one third of patients
DAH represents an important subset of IPS. It occurs in with PERDS have DAH.79,83
5% of HSCT recipients, in both allo- and autotrans-
plants, with a range of 2 to 21%.13 Risk factors for DAH CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION
include age > 40 years, intensive chemotherapy, total PERDS is a subset of IPS characterized by acute lung
body irradiation, the presence of inflammatory cells in injury during the neutrophil periengraftment period. It
BAL fluid, mucositis, and acute GVHD.13 There are no represents the pulmonary component of the engraft-
associations between the development of DAH and ment syndrome, which may also present with diarrhea
prolonged prothrombin or partial thromboplastin time and skin rash.79,89,90 The diagnostic criteria of PERDS
or low platelets. DAH is not corrected with platelet include the presence of fever and evidence of pulmonary
transfusion.82 injury in the form of hypoxia (SaO2 < 90%) and/or
pulmonary infiltrates on chest radiograph, in the ab- exclusion of other causes of this functional abnormal-
sence of cardiac dysfunction or infection, within 5 days ity in allogeneic HSCT recipients with chronic
of neutrophil engraftment.79 The median time to onset GVHD.13,95
of PERDS is 11 days (range, 4 to 25 days) after
transplant.79 TREATMENT
BAL may show neutrophilic alveolitis.79 Trans- The treatment of BO usually consists of corticoste-
bronchial lung biopsy is usually contraindicated because roids and augmented immunosuppression, targeting
of thrombocytopenia. Surgical lung biopsy is rarely chronic GVHD. However, only a minority show
necessary but may show diffuse alveolar damage.79 clinical improvement.13 In a recent study, Khalid
and colleagues reported eight HSCT recipients with
TREATMENT/PROGNOSIS BO whose PFT improved after treatment with azi-
High-dose corticosteroid therapy often leads to rapid thromycin.96 Randomized clinical trials are warranted
clinical improvement.79 Unlike DAH and IPS, only to define the role of macrolide therapy for BO.
about one third of HSCT recipients with PERDS Prophylaxis for PCP and Streptococcus pneumoniae
require intensive care unit admission and mechanical should be provided. In selected HSCT recipients
ventilation.79 The reported mortality rate of PERDS is with respiratory failure secondary to BO, lung trans-
26%.79 plantation may be an option.97
CLINICAL COURSE AND PROGNOSIS

Bronchiolitis Obliterans The airflow limitation in HSCT recipients with BO
improves in only 8 to 20%.13 The overall case fatality rate
EPIDEMIOLOGY is 59%, with a reported range of 14 to 100%.13 In a recent
About 45% of long-term survivor allogeneic HSCT study, the 5 year survival rate of 47 HSCT recipients
recipients develop chronic GVHD.91 Bronchiolitis ob- with BO was 10%, compared with 40% for those without
literans (BO) is a severe manifestation of chronic BO.98
GVHD characterized by airflow limitation.92 The over-
all frequency of BO in allogeneic HSCT recipients is
3.9%.13 Risk factors for BO include GVHD, older Bronchiolitis Obliterans Organizing Pneumonia
donor and recipient age, myeloablative conditioning,
methotrexate use, antecedent respiratory infection, and EPIDEMIOLOGY
serum immunoglobulin deficiency.13 The published medical literature on bronchiolitis oblit-
erans organizing pneumonia (BOOP) in HSCT recip-
CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION ients is limited to case reports with a maximum number
The pathogenesis of BO in HSCT recipients is not well of five patients.13 Although BOOP has been reported
understood. BO usually occurs between 2 months and mostly in allogeneic HSCT recipients with GVHD, it
9 years after transplantation.13 The clinical presentation also occurs in autologous transplant.99
includes dry cough and dyspnea in most, wheezing in
40%, and antecedent cold symptoms in 20%.13 CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION
Twenty percent of the patients with BO had no respi- HSCT recipients with BOOP present with dry cough,
ratory symptoms at the time of the abnormal pulmonary dyspnea, and fever, usually at 1 month to 2 years after
function testing.93 The diagnostic evaluation should transplant.13 PFT shows a restrictive defect, decreased
include organs likely to be affected by GVHD, including diffusing capacity for carbon monoxide, normal airflow
the liver, sinuses, and esophagus.13 flow, and hypoxemia.13 Chest radiographs and HRCT
PFT shows irreversible airway obstruction. The show patchy air space consolidation, ground-glass
chest radiograph may be normal or show hyperinfla- attenuation, and nodular opacities. Although usually
tion.13 HRCT of the chest may show decreased lung bilateral, radiographic abnormalities can be unilateral.99
attenuation, bronchial dilatation, centrilobular nod- Exhaled nitric oxide concentration is increased in
ules, and expiratory air trapping.13 BAL may show HSCT recipients with BOOP and decline in response
either or both neutrophilic and lymphocytic inflam- to treatment.100
mation.94 Transbronchial lung biopsy is usually non- The diagnosis of BOOP in the HSCT recipient
diagnostic. Video-assisted thoracoscopic lung biopsy, requires lung biopsy, occasionally transbronchial but
showing fibrinous obliteration of the small airway usually surgical.101 The histologic hallmark is patchy
lumen, provides a definitive histologic diagnosis of intraluminal fibrosis consisting of polypoid plugs of
BO.13 However, surgical biopsies are rarely indicated immature fibroblasts resembling granulation tissue, ob-
because the diagnosis can usually be made clinically by literating the distal airways, alveolar ducts, and peribron-
the presence of irreversible airflow obstruction and the chial alveolar spaces.102
TREATMENT AND PROGNOSIS CONCLUSIONS

Approximately 80% of HSCT recipients with BOOP HSCT recipients are at risk for multiple complications.
respond favorably to treatment with corticosteroids.13 Their immune defect and recovery follow a predictable
Radiographic abnormalities usually clear within 1 to course. Knowledge of their immune status at the various
3 months of initiating corticosteroid therapy. A favor- phases following transplant is essential to focus on the
able outcome has been observed in one allogeneic HSCT most likely complications. Some of the infectious com-
recipient with BOOP following the use of erythromycin plications, especially Candida and CMV, can be pre-
in conjunction with corticosteroids.103 The case fatality vented by prophylactic and preemptive therapy. There is
rate of BOOP in HSCT recipients is 19%.13 no effective prophylactic therapy for invasive Aspergillus
infection, which usually affects HSCT recipients with
prolonged neutropenia and GVHD. Because of the high
Delayed Pulmonary Toxicity Syndrome mortality associated with delayed diagnoses and inap-
propriate therapy, we advocate aggressive diagnostic
EPIDEMIOLOGY evaluation and early treatment.
In the 1990s, many patients with breast cancer were
treated with autologous HSCT following chemotherapy
REFERENCES
with cyclophosphamide, cisplatinum, and bischloroethy-
linitrosurea (BCNU).104 Delayed pulmonary toxicity 1. Report on state of the art in blood and marrow transplanta-
syndrome (DPTS) developed in up to 72% of these tion. IBMTR/ABMTR Newsletter 2003;10:7–10
patients.105–107 Currently, DPTS is uncommon because 2. Afessa B, Tefferi A, Hoagland HC, Letendre L, Peters SG.
the use of HSCT for metastatic breast cancer has not Outcome of recipients of bone marrow transplants who
been shown to be of benefit.108 require intensive care unit support. Mayo Clin Proc 1992;67:
117–122
3. Afessa B, Tefferi A, Dunn WF, Litzow MR, Peters SG.
CLINICAL FINDINGS AND DIAGNOSTIC EVALUATION
Intensive care unit support and Acute Physiology and
Patients with DPTS present with cough, dyspnea, and Chronic Health Evaluation III performance in hemato-
fever, 2 weeks to 4 months after transplant.106,107,109 In poietic stem cell transplant recipients. Crit Care Med
the appropriate clinical setting, DPTS is diagnosed by 2003;31:1715–1721
demonstration of a decline in carbon monoxide diffu- 4. Jackson SR, Tweeddale MG, Barnett MJ, et al. Admission of
sion capacity and exclusion of infectious causes.106,109 bone marrow transplant recipients to the intensive care unit:
CT of the chest may be normal but usually shows outcome, survival and prognostic factors. Bone Marrow
Transplant 1998;21:697–704
ground-glass opacities, linear or nodular opacities,
5. Crawford SW, Petersen FB. Long-term survival from
and consolidation.109 respiratory failure after marrow transplantation for malig-
nancy. Am Rev Respir Dis 1992;145:510–514
TREATMENT AND PROGNOSIS 6. Khassawneh BY, White P Jr, Anaissie EJ, Barlogie B, Hiller
DPTS responds to corticosteroid therapy.106,107,109 One FC. Outcome from mechanical ventilation after autologous
case of DPTS refractory to steroid was treated success- peripheral blood stem cell transplantation. Chest 2002;121:
fully with interferon-gamma.110 Inhaled fluticasone may 185–188
7. Scott PH, Morgan TJ, Durrant S, Boots RJ. Survival
be used to prevent DPTS.111 No death has been attrib-
following mechanical ventilation of recipients of bone
uted to DPTS.105–107,109 marrow transplants and peripheral blood stem cell trans-
plants. Anaesth Intensive Care 2002;30:289–294
8. Matulis M, High KP. Immune reconstitution after hema-
Pulmonary Cytolytic Thrombi topoietic stem-cell transplantation and its influence on
Pulmonary cytolytic thrombi (PCT) occurs exclusively in respiratory infections. Semin Respir Infect 2002;17:130–
allogeneic HSCT recipients, typically in the setting of 139
9. Gosselin MV, Adams RH. Pulmonary complications in
GVHD. Most of the reported PCT cases were from a
bone marrow transplantation. J Thorac Imaging 2002;17:
single institution and were under age 18.112,113 The 132–144
median onset of PCT is 72 days after transplanta- 10. Ninin E, Milpied N, Moreau P, et al. Longitudinal study
tion.113–115 All patients are febrile and some have cough of bacterial, viral, and fungal infections in adult recipients
at presentation.113–115 Chest CT may show multiple of bone marrow transplants. Clin Infect Dis 2001;33:41–
peripheral pulmonary nodules.115 The diagnosis of 47
PCT requires surgical lung biopsy, showing occlusive 11. George B, Mathews V, Srivastava A, Chandy M. Infections
among allogeneic bone marrow transplant recipients in
vascular lesions and hemorrhagic infarcts due to thrombi
India. Bone Marrow Transplant 2004;33:311–315
consisting of intensely basophilic, amorphous material 12. Yoo JH, Lee DG, Choi SM, et al. Infectious complications
that may extend through the vascular wall.112 The treat- and outcomes after allogeneic hematopoietic stem cell
ment of PCT is not well defined. There have been no transplantation in Korea. Bone Marrow Transplant 2004;
reported deaths attributed to PCT. 34:497–504
13. Afessa B, Litzow MR, Tefferi A. Bronchiolitis obliterans 28. Morfin F, Boucher A, Najioullah F, et al. Cytomegalovirus
and other late onset noninfectious pulmonary complications and adenovirus infections and diseases among 75 paediatric
in hematopoietic stem cell transplantation. Bone Marrow unrelated allogeneic bone marrow transplant recipients.
Transplant 2001;28:425–434 J Med Virol 2004;72:257–262
14. Hughes WT, Armstrong D, Bodey GP, et al. 2002 29. Gasparetto EL, Ono SE, Escuissato D, et al. Cytomegalo-
guidelines for the use of antimicrobial agents in neutropenic virus pneumonia after bone marrow transplantation: high
patients with cancer. Clin Infect Dis 2002;34:730–751 resolution CT findings. Br J Radiol 2004;77:724–727
15. Feinstein MB, Mokhtari M, Ferreiro R, Stover DE, 30. Machado CM, Dulley FL, Boas LS, et al. CMV pneumonia
Jakubowski A. Fiberoptic bronchoscopy in allogeneic bone in allogeneic BMT recipients undergoing early treatment of
marrow transplantation: findings in the era of serum preemptive ganciclovir therapy. Bone Marrow Transplant
cytomegalovirus antigen surveillance. Chest 2001;120:1094– 2000;26:413–417
1100 31. Konoplev S, Champlin RE, Giralt S, et al. Cytomegalovirus
16. Leung AN, Gosselin MV, Napper CH, et al. Pulmonary pneumonia in adult autologous blood and marrow transplant
infections after bone marrow transplantation: clinical and recipients. Bone Marrow Transplant 2001;27:877–881
radiographic findings. Radiology 1999;210:699–710 32. Wah TM, Moss HA, Robertson RJ, Barnard DL.
17. Patel NR, Lee PS, Kim JH, Weinhouse GL, Koziel H. The Pulmonary complications following bone marrow transplan-
influence of diagnostic bronchoscopy on clinical outcomes tation. Br J Radiol 2003;76:373–379
comparing adult autologous and allogeneic bone marrow 33. Escuissato DL, Gasparetto EL, Marchiori E, et al. Pul-
transplant patients. Chest 2005;127:1388–1396 monary infections after bone marrow transplantation: high-
18. Taplitz RA, Jordan MC. Pneumonia caused by herpesviruses resolution CT findings in 111 patients. AJR Am J
in recipients of hematopoietic cell transplants. Semin Respir Roentgenol 2005;185:608–615
Infect 2002;17:121–129 34. Boeckh M, Boivin G. Quantitation of cytomegalovirus:
19. Ebbert JO, Limper AH. Respiratory syncytial virus pneu- methodologic aspects and clinical applications. Clin Micro-
monitis in immunocompromised adults: clinical features and biol Rev 1998;11:533–554
outcome. Respiration 2005;72:263–269 35. Ljungman P, De Bock R, Cordonnier C, et al. Practices for
20. Ghosh S, Champlin RE, Englund J, et al. Respiratory cytomegalovirus diagnosis, prophylaxis and treatment in
syncytial virus upper respiratory tract illnesses in adult blood allogeneic bone marrow transplant recipients: a report from
and marrow transplant recipients: combination therapy with the Working Party for Infectious Diseases of the EBMT.
aerosolized ribavirin and intravenous immunoglobulin. Bone Bone Marrow Transplant 1993;12:399–403
Marrow Transplant 2000;25:751–755 36. Shelhamer JH, Gill VJ, Quinn TC, et al. The laboratory
21. Atkinson K, Nivison-Smith I, Dodds A, Concannon A, evaluation of opportunistic pulmonary infections. Ann
Milliken S, Downs K. A comparison of the pattern of Intern Med 1996;124:585–599
interstitial pneumonitis following allogeneic bone marrow 37. Sokos DR, Berger M, Lazarus HM. Intravenous immuno-
transplantation before and after the introduction of prophy- globulin: appropriate indications and uses in hematopoietic
lactic ganciclovir therapy in 1989. Bone Marrow Transplant stem cell transplantation. Biol Blood Marrow Transplant
1998;21:691–695 2002;8:117–130
22. Ibrahim A, Gautier E, Roittmann S, et al. Should 38. Limaye AP, Bowden RA, Myerson D, Boeckh M.
cytomegalovirus be tested for in both blood and bronchoal- Cytomegalovirus disease occurring before engraftment in
veolar lavage fluid of patients at a high risk of CMV marrow transplant recipients. Clin Infect Dis 1997;24:830–
pneumonia after bone marrow transplantation? Br J Haematol 835
1997;98:222–227 39. Nguyen Q, Champlin R, Giralt S, et al. Late cytomegalo-
23. Sakuma H, Hosoya M, Kanno H, et al. Risk of cytomega- virus pneumonia in adult allogeneic blood and marrow
lovirus infection after peripheral blood stem cell transplanta- transplant recipients. Clin Infect Dis 1999;28:618–623
tion. Bone Marrow Transplant 1997;19:49–53 40. Chen CS, Boeckh M, Seidel K, et al. Incidence, risk factors,
24. Yanada M, Yamamoto K, Emi N, et al. Cytomegalovirus and mortality from pneumonia developing late after
antigenemia and outcome of patients treated with preemp- hematopoietic stem cell transplantation. Bone Marrow
tive ganciclovir: retrospective analysis of 241 consecutive Transplant 2003;32:515–522
patients undergoing allogeneic hematopoietic stem cell 41. Marr KA, Bowden RA. Fungal infections in patients
transplantation. Bone Marrow Transplant 2003;32:801– undergoing blood and marrow transplantation. Transpl
807 Infect Dis 1999;1:237–246
25. Enright H, Haake R, Weisdorf D, et al. Cytomegalovirus 42. Marr KA, Carter RA, Boeckh M, Martin P, Corey L.
pneumonia after bone marrow transplantation: risk factors Invasive aspergillosis in allogeneic stem cell transplant
and response to therapy. Transplantation 1993;55:1339– recipients: changes in epidemiology and risk factors. Blood
1346 2002;100:4358–4366
26. Zekri AR, Mohamed WS, Samra MA, Sherif GM, El 43. Oliveira JS, Kerbauy FR, Colombo AL, et al. Fungal
Shehaby AM, El Sayed MH. Risk factors for cytomegalo- infections in marrow transplant recipients under antifungal
virus, hepatitis B and C virus reactivation after bone marrow prophylaxis with fluconazole. Braz J Med Biol Res 2002;
transplantation. Transpl Immunol 2004;13:305–311 35:789–798
27. Kim DH, Kim JG, Lee NY, et al. Risk factors for late 44. Wakayama M, Shibuya K, Ando T, et al. Deep-seated
cytomegalovirus infection after allogeneic stem cell trans- mycosis as a complication in bone marrow transplantation
plantation using HLA-matched sibling donor: donor patients. Mycoses 2002;45:146–151
lymphocyte infusion and previous history of early CMV 45. Ascioglu S, Rex JH, de Pauw B, et al. Defining opportunistic
infection. Bone Marrow Transplant 2004;34:21–27 invasive fungal infections in immunocompromised patients
with cancer and hematopoietic stem cell transplants: an nias in immunocompromised marrow transplant recipients.
international consensus. Clin Infect Dis 2002;34:7–14 Am Rev Respir Dis 1982;126:763–765
46. Grow WB, Moreb JS, Roque D, et al. Late onset of invasive 63. Goodrich JM, Reed EC, Mori M, et al. Clinical features and
Aspergillus infection in bone marrow transplant patients at a analysis of risk factors for invasive candidal infection after
university hospital. Bone Marrow Transplant 2002;29:15–19 marrow transplantation. J Infect Dis 1991;164:731–740
47. Soubani AO, Qureshi MA. Invasive pulmonary aspergillosis 64. Verfaillie C, Weisdorf D, Haake R, Hostetter M, Ramsay
following bone marrow transplantation: risk factors and NK, McGlave P. Candida infections in bone marrow
diagnostic aspect. Haematologia (Budap) 2002;32:427–437 transplant recipients. Bone Marrow Transplant 1991;8:177–
48. Soubani AO, Chandrasekar PH. The clinical spectrum of 184
pulmonary aspergillosis. Chest 2002;121:1988–1999 65. Collin BA, Leather HL, Wingard JR, Ramphal R.
49. Horger M, Einsele H, Schumacher U, et al. Invasive Evolution, incidence, and susceptibility of bacterial blood-
pulmonary aspergillosis: frequency and meaning of the stream isolates from 519 bone marrow transplant patients.
‘‘hypodense sign’’ on unenhanced CT. Br J Radiol 2005; Clin Infect Dis 2001;33:947–953
78:697–703 66. Lossos IS, Breuer R, Or R, et al. Bacterial pneumonia in
50. Bialek R, Moshous D, Casanova JL, Blanche S, Hennequin recipients of bone marrow transplantation: a five-year
C. Aspergillus antigen and PCR assays in bone marrow prospective study. Transplantation 1995;60:672–678
transplanted children. Eur J Med Res 2002;7:177–180 67. Ramphal R. Changes in the etiology of bacteremia in febrile
51. Stevens DA, Kan VL, Judson MA, et al. Practice guidelines neutropenic patients and the susceptibilities of the currently
for diseases caused by Aspergillus. Infectious Diseases isolated pathogens. Clin Infect Dis 2004;39(Suppl 1):S25–
Society of America. Clin Infect Dis 2000;30:696–709 S31
52. Centers for Disease Control and Prevention. Guidelines for 68. Kruger WH, Bohlius J, Cornely OA, et al. Antimicrobial
preventing opportunistic infections among hematopoietic prophylaxis in allogeneic bone marrow transplantation.
stem cell transplant recipients. Recommendations of CDC, Guidelines of the infectious diseases working party
the Infectious Disease Society of America, and the American (AGIHO) of the German Society of Haematology and
Society of Blood and Marrow Transplantation. MMWR Oncology. Ann Oncol 2005;16:1381–1390
2000;49(No. RR-10):1–128 69. Cahill RA, Spitzer TR, Mazumder A. Marrow engraftment
53. Kruger WH, Zollner B, Kaulfers PM, Zander AR. Effective and clinical manifestations of capillary leak syndrome. Bone
protection of allogeneic stem cell recipients against aspergil- Marrow Transplant 1996;18:177–184
losis by HEPA air filtration during a period of construction: 70. Dickout WJ, Chan CK, Hyland RH, et al. Prevention of
a prospective survey. J Hematother Stem Cell Res 2003; acute pulmonary edema after bone marrow transplantation.
12:301–307 Chest 1987;92:303–309
54. Schwartz S, Behre G, Heinemann V, et al. Aerosolized 71. Snowden JA, Hill GR, Hunt P, et al. Assessment of
amphotericin B inhalations as prophylaxis of invasive cardiotoxicity during haemopoietic stem cell transplantation
aspergillus infections during prolonged neutropenia: results with plasma brain natriuretic peptide. Bone Marrow
of a prospective randomized multicenter trial. Blood 1999;93: Transplant 2000;26:309–313
3654–3661 72. Wong R, Rondon G, Saliba RM, et al. Idiopathic
55. Grigg A. Prophylaxis and treatment of patients with asper- pneumonia syndrome after high-dose chemotherapy and
gillosis: an overview, including the Royal Melbourne Hospital autologous hematopoietic stem cell transplantation for high-
experience. J Antimicrob Chemother 2002;49(Suppl 1):75–80 risk breast cancer. Bone Marrow Transplant 2003;31:1157–
56. Oren I, Haddad N, Finkelstein R, Rowe JM. Invasive 1163
pulmonary aspergillosis in neutropenic patients during 73. Clark JG, Hansen JA, Hertz MI, Parkman R, Jensen L,
hospital construction: before and after chemoprophylaxis Peavy HH. NHLBI workshop summary: idiopathic pneu-
and institution of HEPA filters. Am J Hematol 2001;66: monia syndrome after bone marrow transplantation. Am Rev
257–262 Respir Dis 1993;147:1601–1606
57. Wilkin A, Feinberg J. Prophylaxis against fungal infections 74. Hauber HP, Mikkila A, Erich JM, et al. TNF-alpha,
and cytomegalovirus disease after bone marrow transplanta- interleukin-10 and interleukin-18 expression in cells of the
tion. Oncology 2000;14:1701–1708 bronchoalveolar lavage in patients with pulmonary compli-
58. Lin SJ, Schranz J, Teutsch SM. Aspergillosis case-fatality cations following bone marrow or peripheral stem cell
rate: systematic review of the literature. Clin Infect Dis 2001; transplantation: a preliminary study. Bone Marrow Trans-
32:358–366 plant 2002;30:485–490
59. Roychowdhury M, Pambuccian SE, Aslan DL, et al. 75. Schots R, Kaufman L, Van RI, et al. Proinflammatory
Pulmonary complications after bone marrow transplantation: cytokines and their role in the development of major
an autopsy study from a large transplantation center. Arch transplant-related complications in the early phase after
Pathol Lab Med 2005;129:366–371 allogeneic bone marrow transplantation. Leukemia 2003;17:
60. Sharma S, Nadrous HF, Peters SG, et al. Pulmonary 1150–1156
complications in adult blood and marrow transplant recipi- 76. Yanik G, Hellerstedt B, Custer J, et al. Etanercept (Enbrel)
ents: autopsy findings. Chest 2005;128:1385–1392 administration for idiopathic pneumonia syndrome after
61. Soubani AO, Qureshi MA, Baynes RD. Flexible broncho- allogeneic hematopoietic stem cell transplantation. Biol
scopy in the diagnosis of pulmonary infiltrates following Blood Marrow Transplant 2002;8:395–400
autologous peripheral stem cell transplantation for advanced 77. Kantrow SP, Hackman RC, Boeckh M, Myerson D,
breast cancer. Bone Marrow Transplant 2001;28:981–985 Crawford SW. Idiopathic pneumonia syndrome: changing
62. Springmeyer SC, Silvestri RC, Sale GE, et al. The role of spectrum of lung injury after marrow transplantation.
transbronchial biopsy for the diagnosis of diffuse pneumo- Transplantation 1997;63:1079–1086
78. Afessa B, Tefferi A, Litzow MR, Krowka MJ, Wylam ME, 93. Clark JG, Crawford SW, Madtes DK, Sullivan KM.
Peters SG. Diffuse alveolar hemorrhage in hematopoietic Obstructive lung disease after allogeneic marrow transplan-
stem cell transplant recipients. Am J Respir Crit Care Med tation: clinical presentation and course. Ann Intern Med
2002;166:641–645 1989;111:368–376
79. Capizzi SA, Kumar S, Huneke NE, et al. Periengraftment 94. St. John RC, Gadek JE, Tutschka PJ, Kapoor N, Dorinsky
respiratory distress syndrome during autologous hemato- PM. Analysis of airflow obstruction by bronchoalveolar
poietic stem cell transplantation. Bone Marrow Transplant lavage following bone marrow transplantation. Implications
2001;27:1299–1303 for pathogenesis and treatment. Chest 1990;98:600–607
80. Griese M, Rampf U, Hofmann D, Fuhrer M, Reinhardt D, 95. Crawford SW, Clark JG. Bronchiolitis associated with bone
Bender-Gotze C. Pulmonary complications after bone marrow transplantation. Clin Chest Med 1993;14:741–749
marrow transplantation in children: twenty-four years of 96. Khalid M, Al Saghir A, Saleemi S, et al. Azithromycin in
experience in a single pediatric center. Pediatr Pulmonol bronchiolitis obliterans complicating bone marrow trans-
2000;30:393–401 plantation: a preliminary study. Eur Respir J 2005;25:490–
81. Crawford SW, Hackman RC. Clinical course of idiopathic 493
pneumonia after bone marrow transplantation. Am Rev 97. Favaloro R, Bertolotti A, Gomez C, et al. Lung transplant at
Respir Dis 1993;147:1393–1400 the Favaloro Foundation: a 13-year experience. Transplant
82. Robbins RA, Linder J, Stahl MG, et al. Diffuse alveolar Proc 2004;36:1689–1691
hemorrhage in autologous bone marrow transplant recipi- 98. Dudek AZ, Mahaseth H, DeFor TE, Weisdorf DJ.
ents. Am J Med 1989;87:511–518 Bronchiolitis obliterans in chronic graft-versus-host disease:
83. Afessa B, Tefferi A, Litzow MR, Peters SG. Outcome of analysis of risk factors and treatment outcomes. Biol Blood
diffuse alveolar hemorrhage in hematopoietic stem cell Marrow Transplant 2003;9:657–666
transplant recipients. Am J Respir Crit Care Med 99. Hayes-Jordan A, Benaim E, Richardson S, et al. Open lung
2002;166:1364–1368 biopsy in pediatric bone marrow transplant patients. J Pediatr
84. Witte RJ, Gurney JW, Robbins RA, et al. Diffuse pulmonary Surg 2002;37:446–452
alveolar hemorrhage after bone marrow transplantation: 100. Kanamori H, Fujisawa S, Tsuburai T, et al. Increased
radiographic findings in 39 patients. AJR Am J Roentgenol exhaled nitric oxide in bronchiolitis obliterans organizing
1991;157:461–464 pneumonia after allogeneic bone marrow transplantation.
85. Hicks K, Peng D, Gajewski JL. Treatment of diffuse alveolar Transplantation 2002;74:1356–1358
hemorrhage after allogeneic bone marrow transplant with 101. Alasaly K, Muller N, Ostrow DN, Champion P, FitzGerald
recombinant factor VIIa. Bone Marrow Transplant 2002; JM. Cryptogenic organizing pneumonia: a report of 25 cases
30:975–978 and a review of the literature. Medicine (Baltimore) 1995;
86. Pastores SM, Papadopoulos E, Voigt L, Halpern NA. 74:201–211
Diffuse alveolar hemorrhage after allogeneic hematopoietic 102. Myers JL, Colby TV. Pathologic manifestations of bronch-
stem-cell transplantation: treatment with recombinant factor iolitis, constrictive bronchiolitis, cryptogenic organizing
VIIa. Chest 2003;124:2400–2403 pneumonia, and diffuse panbronchiolitis. Clin Chest Med
87. Lewis ID, DeFor T, Weisdorf DJ. Increasing incidence of 1993;14:611–622
diffuse alveolar hemorrhage following allogeneic bone 103. Ishii T, Manabe A, Ebihara Y, et al. Improvement in
marrow transplantation: cryptic etiology and uncertain bronchiolitis obliterans organizing pneumonia in a child after
therapy. Bone Marrow Transplant 2000;26:539–543 allogeneic bone marrow transplantation by a combination of
88. Metcalf JP, Rennard SI, Reed EC, et al. Corticosteroids as oral prednisolone and low dose erythromycin. Bone Marrow
adjunctive therapy for diffuse alveolar hemorrhage associated Transplant 2000;26:907–910
with bone marrow transplantation. University of Nebraska 104. Pedrazzoli P, Ferrante P, Kulekci A, et al. Autologous
Medical Center Bone Marrow Transplant Group. Am J Med hematopoietic stem cell transplantation for breast cancer in
1994;96:327–334 Europe: critical evaluation of data from the European Group
89. Akasheh M, Eastwood D, Vesole DH. Engraftment for Blood and Marrow Transplantation (EBMT) Registry
syndrome after autologous hematopoietic stem cell trans- 1990–1999. Bone Marrow Transplant 2003;32:489–494
plant supported by granulocyte-colony-stimulating factor 105. Bhalla KS, Wilczynski SW, Abushamaa AM, et al. Pul-
(G-CSF) versus granulocyte-macrophage colony-stimulat- monary toxicity of induction chemotherapy prior to standard
ing factor (GM-CSF). Bone Marrow Transplant 2003; or high- dose chemotherapy with autologous hematopoietic
31:113–116 support. Am J Respir Crit Care Med 2000;161:17–25
90. Maiolino A, Biasoli I, Lima J, Portugal AC, Pulcheri W, 106. Cao TM, Negrin RS, Stockerl-Goldstein KE, et al.
Nucci M. Engraftment syndrome following autologous Pulmonary toxicity syndrome in breast cancer patients
hematopoietic stem cell transplantation: definition of undergoing BCNU-containing high-dose chemotherapy
diagnostic criteria. Bone Marrow Transplant 2003;31:393– and autologous hematopoietic cell transplantation. Biol
397 Blood Marrow Transplant 2000;6:387–394
91. Carlens S, Ringden O, Remberger M, et al. Risk factors for 107. Chap L, Shpiner R, Levine M, Norton L, Lill M, Glaspy J.
chronic graft-versus-host disease after bone marrow trans- Pulmonary toxicity of high-dose chemotherapy for breast
plantation: a retrospective single centre analysis. Bone cancer: a noninvasive approach to diagnosis and treatment.
Marrow Transplant 1998;22:755–761 Bone Marrow Transplant 1997;20:1063–1067
92. Holland HK, Wingard JR, Beschorner WE, Saral R, Santos 108. Stadtmauer EA, O’Neill A, Goldstein LJ, et al. Conven-
GW. Bronchiolitis obliterans in bone marrow transplanta- tional-dose chemotherapy compared with high-dose che-
tion and its relationship to chronic graft-v-host disease and motherapy plus autologous hematopoietic stem-cell
low serum IgG. Blood 1988;72:621–627 transplantation for metastatic breast cancer. Philadelphia
Bone Marrow Transplant Group. N Engl J Med 2000;342: 112. Gulbahce HE, Pambuccian SE, Jessurun J, et al. Pulmonary
1069–1076 nodular lesions in bone marrow transplant recipients: impact
109. Wilczynski SW, Erasmus JJ, Petros WP, Vredenburgh JJ, of histologic diagnosis on patient management and prog-
Folz RJ. Delayed pulmonary toxicity syndrome following nosis. Am J Clin Pathol 2004;121:205–210
high-dose chemotherapy and bone marrow transplantation 113. Morales IJ, Anderson PM, Tazelaar HD, Wylam ME.
for breast cancer. Am J Respir Crit Care Med 1998;157:565– Pulmonary cytolytic thrombi: unusual complication of
573 hematopoietic stem cell transplantation. J Pediatr Hematol
110. Suratt BT, Lynch DA, Cool CD, Jones RB, Brown KK. Oncol 2003;25:89–92
Interferon-gamma for delayed pulmonary toxicity syndrome 114. Gulbahce HE, Manivel JC, Jessurun J. Pulmonary cytolytic
resistant to steroids. Bone Marrow Transplant 2003;31:939– thrombi: a previously unrecognized complication of bone
941 marrow transplantation. Am J Surg Pathol 2000;24:1147–
111. McGaughey DS, Nikcevich DA, Long GD, et al. Inhaled 1152
steroids as prophylaxis for delayed pulmonary toxicity 115. Woodard JP, Gulbahce E, Shreve M, et al. Pulmonary
syndrome in breast cancer patients undergoing high-dose cytolytic thrombi: a newly recognized complication of stem
chemotherapy and autologous stem cell transplantation. Biol cell transplantation. Bone Marrow Transplant 2000;25:293–
Blood Marrow Transplant 2001;7:274–278 300
Infection Control and the Prevention of
Nosocomial Infections in the Intensive
Care Unit
Gonzalo M.L. Bearman, M.D., M.P.H.,1 Cindy Munro, Ph.D., R.N., A.N.P.,2
Curtis N. Sessler, M.D.,3 and Richard P. Wenzel, M.D., M.Sc.4
ABSTRACT
Nosocomial infections continue to be significant causes of morbidity, mortality,

and added costs in the health care setting. Half of all life-threatening nosocomial
bloodstream infections and pneumonias occur in intensive care units (ICUs), despite
ICUs representing only 15 to 20% of all hospital beds. Thus an efficient focus for
prevention and control of life-threatening health care–associated infections should be in
ICUs. Further, growing antibiotic resistance complicates the therapy of serious infections.
Meticulous infection control practice with continued attention to hand hygiene is of
paramount importance. Strict adherence to evidence-based catheter insertion and main-
tenance policies reduces nosocomial bloodstream infections. Evidence-based prevention
strategies for ventilator-associated pneumonia, including management of respiratory
equipment according to published guidelines and maintaining backrest elevation at 30 to
45 degrees, are effective. For greatest risk reduction, multifaceted programs ensuring
maximal adherence with evidence-based infection control guidelines are needed.
KEYWORDS: Nosocomial infections, bloodstream infections, ventilator-associated

pneumonia, infection control
N osocomial infections add significant morbid- Therefore, current data are compelling that lives can be
ity, mortality, and costs to those expected from the extended by preventing BSIs occurring in hospitals.
patient’s underlying conditions alone. The concept of The crude mortality (all cause) for nosocomial
attributable mortality has been championed1 to examine BSIs is 25 to 30%,3 and most studies of pathogen-
the direct contribution to mortality due to the infection, specific, nosocomial BSIs show that the attributable
after accounting for the impact of the patient’s baseline mortality is at least half of the crude mortality ( 15%
illnesses. When examining the attributable mortality on average conservatively).4–7 Furthermore, the crude
from nosocomial bloodstream infections (BSIs) alone, and attributable mortality figures for nosocomial pneu-
Wenzel and Edmond showed that they are equivalent to monias are 30% and 10%, respectively. The data imply
the eighth leading cause of death in the United States.2 that one third of all pneumonia-related deaths are due
Furthermore, with conservative assumptions nosocomial directly to the lung infection. Importantly, half of all
BSIs annually lead to over 250,000 years of life lost.2 life-threatening nosocomial BSIs and pneumonias occur
1
Divisions of Quality HealthCare and Infectious Diseases, 2School [email protected].
of Nursing, 3Division of Pulmonary and Critical Care Medicine, Non-pulmonary Critical Care: Managing Multisystem Critical
4
Department of Internal Medicine, Virginia Commonwealth Univer- Illness; Guest Editor, Curtis N. Sessler, M.D.
sity Medical Center, Richmond, Virginia. Semin Respir Crit Care Med 2006;27:310–324. Copyright # 2006
Address for correspondence and reprint requests: Gonzalo M.L. by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York,
Bearman, M.D., M.P.H., Divisions of Quality HealthCare and NY 10001, USA. Tel: +1(212) 584-4662.
Infectious Diseases, P.O. Box 980019, Virginia Commonwealth DOI 10.1055/s-2006-945534. ISSN 1069-3424.
University Medical Center, Richmond, VA 23298-0019. E-mail:
310
INFECTION CONTROL AND PREVENTIONIN THEICU/BEARMAN ET AL 311
in critical care units, usually representing 15 to 20% of all vascular disease, and chronic renal insufficiency have
hospital beds. Thus an efficient focus for prevention and been common. In three patients, an extremity had
control of life-threatening health care–associated infec- been the site of the soft tissue or bone infection. In all
tions should be in ICUs. patients, the substrate organism was methicillin-resist-
Complicating the therapy of serious ICU-related ant S. aureus (MRSA), and the mechanism of resistance
infections is the problem of antibiotic resistance, a global was plasmid transfer of a mec A gene from a co-coloniz-
issue with wide country to country variation.8,9 Consider ing Enterococcus (vancomycin resistant enterococci) to
the leading causes of nosocomial BSIs (proportion of all the MRSA strain. Earlier it had been shown that 30%
nosocomial BSIs) in the United States: coagulase neg- of identified patients colonized with vancomycin-resist-
ative Staphylococci (31%), S. aureus (20%), Enterococci ant enterococci are also colonized with MRSA.15 Thus
(9%), and Candida species (9%) in rank order. In the the opportunities to see the continual emergence of fully
United States, over 90% of coagulase negative Staph- vancomycin-resistant strains are increasing. For the time
ylococci and half of all S. aureus nosocomial bloodstream being, therefore, clinicians have another reason to con-
isolates are resistant to methicillin.3 Approximately 30% trol MRSA: the reduction of a risk of selecting for
of enterococcal isolates are resistant to vancomycin, and vancomycin-resistant S. aureus.
over 10% of Candida species are resistant to first-gen- With respect to gram-negative rod BSIs, the fifth
eration triazoles. The crude mortality of BSIs for each through ninth leading causes of nosocomial BSIs, in
species, respectively, is 21%, 25%, 34%, and 39%.3 Surely order, are Escherichia coli, Klebsiella species, Pseudomonas
the challenge for clinicians is to treat early with appro- aeruginosa, and Enterobacter species. Their respective
priate antibiotics in the ICU because inappropriate crude mortality rates are 22%, 28%, 38%, and 27%.3 It
therapy in the first 24 hours doubled the expected should be pointed out that among patients with hospi-
mortality (62% vs 29%) in one study.10 Furthermore, tal-acquired BSIs, P. aeruginosa independently predicts
in the same study, after accounting for the impact of the death,11 and in the United States, 15% of isolates
underlying diseases in a multivariable analysis, the au- reported in 2003 were resistant to imipenam.9 Thus a
thors showed that the adjusted odds ratio for death if the recognition of their frequency and the antibiogram in the
patients were given inappropriate therapy was 6.9.10 It is ICU is key to effective therapy and prevention of death.
clear that the modern clinician in the ICU must not only In summary, nosocomial BSIs alone are equiva-
anticipate the likely organisms and the antibiograms but lent to the eighth leading cause of death in the United
also treat early. States. Along with nosocomial pneumonias, these life-
In examining the reasons for prescribing inad- threatening infections can be found disproportionately
equate therapy for ICU BSIs, Ibrahim and colleagues in critical care units, where increasing levels of antibiotic
showed that failure to treat for Candida BSIs had an resistance are prevalent. Astute clinicians need to know
adjusted odds ratio of 52.10 This observation suggests the likely pathogens and treat early with appropriate
that clinicians still fail to recognize the high likelihood of antibiotics empirically to minimize mortality. The pre-
Candida as a cause of serious ICU-related BSIs and thus vention of these infections will have a huge impact on
do not give antifungal therapy empirically. This is an morbidity, mortality, and costs of health care.
especially important issue because BSIs with Candida
independently predict death in analyses that account for
the underlying diseases11 and have a very high attribut- NOSOCOMIAL BLOODSTREAM
able mortality.5,12 Additionally, the modern clinician INFECTIONS IN THE ICU
must be aware of species differences because all of BSIs are among the most common and most deadly
C. krusei, 10 to 20% of C. glabrata, and up to 10% of nosocomial infections encountered.16 Intravascular
C. albicans are resistant to first-generation triazoles.13 catheterization ranks as the strongest independent
Recently, an argument has been made to use easily risk factor for nosocomial bacteremia,17 and the vast
identified risk factors to target only high risk ICU patients majority of primary bacteremias (those not clearly
for anti-Candida prophylactic therapy. The model as- associated with another defined site of infection) are
sumed a necessary 33% threshold of risk of candidemia for related to vascular catheters. Catheter-related BSI
preventive treatment and a 65% efficacy for the drug to (CRBSI) is associated with prolonged ICU and hospi-
be given to prevent candidemia. With the assumptions, tal length of stay, associated added costs, and attribut-
the number needed to treat (NNT) to prevent a Candida- able mortality.7,18–21 Because an estimated 5 million
related death would be only seven patients.14 central venous catheters (CVCs) are inserted in the
A major concern since 2002 has been the emer- United States annually,22 and BSI accompanies 3 to
gence of fully vancomycin-resistant isolates of S. aureus 5% of these catheters, as many as 250,000 CRBSI
acquired in the health care setting. So far, four cases have might occur annually in the United States alone.23
been reported, and fortunately no transmission has Accordingly, strategies to prevent these common and
occurred. In two patients, diabetes mellitus, peripheral costly infections are critically important.22,24–27
The principal mechanism by which infections personal preference for years. There are meta-analyses of
develop is via colonization of the skin surrounding the nonrandomized trials,32 and risk factor analysis from
catheter insertion site followed by migration of patho- nonrandomized trials,33 that suggest the subclavian vein
gens along the subcutaneous catheter tract and external approach has fewer CRBSIs than the internal jugular
catheter surface. A second important mechanism that approach vein. Nonrandomized studies34 and a single
becomes more prevalent with longer catheter duration is randomized trial35 demonstrate higher rates of catheter-
the colonization of stopcocks and catheter hubs with related infections with the femoral vein approach,
subsequent contamination of luminal surfaces and cath- although some large, nonrandomized case series show
eter infection. An important intermediary step is accu- only trends for higher CRBSI rates with the femoral
mulation of biofilm on the luminal surface. Additional approach.36–38 Accordingly, insertion of a CVC at the
less common mechanisms for CRBSI include hematog- subclavian vein site is preferred and the femoral vein site
enous seeding from distant foci, and rarely, infusate avoided in comprehensive strategies to reduce CRBSI
contamination. rates.39–41 The availability of two-dimensional ultra-
Strategies for prevention of CRBSI focus prin- sound (2-D US) may also influence site selection because
cipally on various measures designed to prevent colo- use of 2-D US is associated with lower rates of both
nization of external and luminal catheter surfaces with failed CVC insertion and CVC complications, with the
microorganisms. An expert panel convened by the U.S. most convincing data for the internal jugular site.42,43
Centers for Disease Control and Prevention (CDC) Additional simple measures that have been asso-
identified the following as major areas of emphasis for ciated with reduced infection rates in prospective studies
prevention of catheter-related infection: (1) educating include use of proper hand hygiene and a structured
and training health care providers who insert and approach to achieve high compliance with sterile tech-
maintain catheters, (2) using maximal sterile barrier nique through checklists, consolidated bedside CVC
precautions during CVC insertion, (3) using a 2% supply carts, and nursing empowerment to stop proce-
chlorhexidine preparation for skin antisepsis, (4) avoid- dures in which violations have occurred.41 Although
ing routine replacement of CVCs as a strategy to these strategies focus primarily on optimizing technique
prevent infection, and (5) using antiseptic/antibiotic- during catheter insertion, equal emphasis must be placed
impregnated short-term CVCs if the rate of infection is on the timely removal of the catheter, including formal
high despite adherence to other strategies.24 These questioning of the continued need for the catheter on
recommendations are based largely upon clinical trial daily ICU rounds.41 Additionally, because hub coloni-
evidence. Raad and coworkers28 demonstrated maximum zation can occur during frequent manipulation, disin-
barrier precautions (large sterile drape on the patient and fection with chlorhexidine/alcohol should be performed
sterile gown and gloves, mask and cap worn by operators) when hubs are accessed.44
to be a cost-effective approach associated with reduced Organized strategies that focus on either or both
catheter-related infections. Two percent chlorhexidine is education and improving the process of catheter inser-
superior to povidone-iodine for skin disinfectant prepa- tion and care have been shown to produce sustained
ration prior to CVC insertion, based upon a meta- reductions in rates of CRBSI.39–41,45–48 Major compo-
analysis of eight randomized, controlled trials that dem- nents of these and other programs, and recommenda-
onstrated lower rates of CRBSI for all catheters (risk tions from recent reviews are outlined in Table 1. Most
ratio ¼ 0.49, 95% confidence interval 0.28 to 0.88) and approaches focus on educating and modifying the be-
for CVC (RR ¼ 0.51, 95% CI ¼ 0.27 to 0.97) specifi- havior of ICU caregivers, including physicians and
cally.29 Scheduled replacement of CVC or balloon flota- nurses; however, some educational programs also target
tion pulmonary artery catheters was not found to reduce medical students.49,50
infectious and mechanical complications when compared In the event that CRBSI rates remain unaccept-
with an as needed approach,30 nor supported in a meta- ably high despite implementing strategies such as those
analysis of eight studies31; thus scheduled replacement is already described, experts recommend considering the
not endorsed. Educational programs and the role of use of antiinfective catheters.24 Catheters coated with
antibacterial/antiseptic-impregnated catheters will be chlorhexidine and silver sulfadiazine (CSS) have been
discussed. most widely studied and were demonstrated in a 1999
The CDC working group24 made additional level meta-analysis to be associated with reduced rates of
I recommendations regarding CVC insertion that in- CRBSI compared with uncoated catheters.51 However,
clude practicing proper hand hygiene, selecting the this catheter was found to be inferior to a minocycline-
subclavian vein insertion site to avoid infection whenever rifampin (MR)-impregnated catheter for preventing
feasible upon considering the risk for mechanical com- CRBSI in a head to head multicenter, randomized,
plications, and using a sterile gauze or sterile transparent controlled trial (RCT).52 Both catheters have been
semipermeable dressing to cover the insertion site. The modified and current products have coatings of both
site of CVC insertion has been the subject of debate and external and luminal surfaces. The newer CSS catheter
Table 1 Preventing Infectious Complications Additional technology-based or invasive ap-

Associated with Central Venous Catheterization proaches to insertion of short-term CVCs in ICU
Program development patients include use of a chlorhexidine-impregnated
1. Establish a standardized interdisciplinary approach with sponge dressing,59 catheter hub ‘‘lock’’ with solutions
emphasis on education, training, and process containing alcohol, heparin, and/or vancomycin,60,61 use
management for infection control of heparin-coated CVCs,62 and subcutaneous tunneling
2. Develop and implement interdisciplinary structured of short-term CVCs.63,64 None of the techniques have
process tools, including been definitively demonstrated to reduce CRBSI in this
a. CVC insertion check list setting thus far.
b. CVC insertion cart that contains all supplies needed Peripherally inserted central catheters (PICC) are
for procedure; cart to be positioned at bedside increasingly utilized for central venous access in ICU
c. Real-time monitoring of insertion process with patients, in part because CRBSI rates for PICCs have
empowerment of observer (typically the bedside nurse) been widely regarded as being substantially lower than
to stop procedure if violation of sterile technique for traditional CVCs. A recent prospective study and a
observed systematic review of published literature indicates, how-
d. Process for daily explicit reminders for continued ever, that PICCs inserted exclusively in hospitalized
necessity of CVC adult patients have a CRBSI rate [systematic review:
3. Track catheter-related infection rates, particularly CRBSI, 1.9 (95% CI: 1.4–2.6)/1000 PICC days for hospitalized
and provide institution-wide feedback about infection rates adults; prospective study of 251 PICCs: 2.4/1000 PICC
and process measures days] that is in excess of rates for PICCs used in
4. Consider expanded use of PICCs outpatients (0.4/1000 PICC days) and similar to that
5. Consider use of anti-infective catheter if institutional CRBSI of CVCs inserted at jugular, subclavian, or femoral sites
rates remain excessive despite careful attention to (2.3/1000 CVC days).65
conventional infection control issues
Catheter insertion and care
1. Confirm that CVC is really needed CATHETER-ASSOCIATED URINARY TRACT
2. Select insertion site based upon CVC indication, patient INFECTIONS
risk factors for complications, patient anatomy, and Urinary catheters are inserted into more than 5 million
availability of two-dimensional ultrasound. Use subclavian patients in both acute care hospitals and extended care
vein site and avoid femoral vein site when possible facilities. Nosocomial bacteriuria predictably occurs in
3. Use precautions to prevent infection during insertion, 25% of patients requiring urinary catheterization for
including greater than 7 days.66,67 Several studies suggest that
a. Hand hygiene catheter-associated urinary tract infections, although
b. Maximum barrier precautions (sterile gown, sterile rarely fatal, can add $500 to $1000 to the direct cost of
gloves, cap, and mask for all operators; large sterile sheet) medical care, trigger the use of unnecessary antimicro-
c. Chlorhexidine skin disinfection bials, and provide a nosocomial reservoir of drug-resist-
d. Aseptic catheter handling. ant pathogens.68 As such, nosocomial urinary tract
4. Properly secure CVC and dress CVC site with dry sterile infections are of epidemiological importance.
gauze or semipermeable transparent dressing The majority of catheter-associated urinary tract
5. Practice good infection control technique during infections derive from the patient’s own periurethral and
management of CVC, including disinfecting catheter hubs perineal flora or from cross-transmission via the hands of
when accessing hubs health care workers during insertion and manipulation of
6. Remove CVC at earliest possible time; use structured the urinary catheter system. Risk factors for catheter-
process reminder associated urinary tract infections have been studied in
7. Do not perform routine scheduled CVC exchange well-designed, prospective studies with multivariate
statistical analysis. Most notably, significant risk factors
CRBSI, catheter-related bloodstream infection; CVC, central venous
catheter; PICCs, peripherally inserted central catheters. include prolonged catheterization (> 6 days), female
gender, urinary catheter insertion outside of the operat-
has been demonstrated to have lower rates of coloniza- ing room, diabetes, malnutrition, and other distant sites
tion, with trends for fewer CRBSI compared with of infection.
uncoated catheters in RCTs.53,54 Similarly, catheters Simple and effective risk reduction strategies
impregnated with silver, platinum, and carbon55–57 re- include the avoidance of unnecessary catheterizations.
duce colonization and may reduce CRBSI. Despite A recent paper by Maki and Tambyah69 suggested that
having a higher acquisition cost, use of antiinfective urinary catheter use should be limited to patients
catheters can be cost-effect if baseline CRBSI rate is requiring relief of urinary obstruction, patients under-
sufficiently high.58 going surgical repair of the genitourinary tract,
critically ill or post-operative patients requiring adequate Table 2 Prevention of Ventilator-Associated

measurement of urine output, and debilitated, paralyzed Pneumonia
or comatose patients. When no longer needed, the Surveillance
urinary catheter should be promptly removed. 1. Monitor hospital rates and trends using National
Nosocomial Infection Surveillance definition of pneumonia
2. Communicate data and goals to physicians and appropriate
VENTILATOR-ASSOCIATED PNEUMONIA staff on regular basis
Ventilator-associated pneumonia (VAP) is defined as Reduce nosocomial transmission of organisms
pneumonia occurring in a mechanically ventilated pa- 1. Adhere to Centers for Disease Control guidelines for
tient that was neither present nor developing at the time care of respiratory equipment
of intubation; thus clinical evidence of VAP occurs 48 2. Practice good hand hygiene
hours or more after intubation. VAP is a significant Patient care
cause of morbidity and mortality in critically ill adults. A 1. Maintain backrest elevation at 30–45 degrees
recent meta-analysis examining 89 VAP research studies 2. Avoid nasotracheal intubation and reintubation
published since 1990 concluded that VAP occurs in 10 3. Remove subglottic secretions by suctioning
to 20% of patients receiving more than 48 hours of 4. Remove secretions prior to endotracheal tube removal
mechanical ventilation.70 VAP doubles the mortality 5. Chlorhexidine gluconate (0.12%) oral rinse for adult
risk of critically ill patients, lengthens ICU and hospital cardiac surgery patients
stay, and increases hospital costs.
Early-onset VAP (diagnosed on or before day 4 of
intubation) has been associated with oropharyngeal lead to impairment of normal swallowing and respiratory
bacterial species (Haemophilus influenzae, Streptococcus clearance mechanisms. Diaphragmatic dysfunction asso-
pneumoniae, Staphylococcus aureus), whereas late-onset ciated with thoracic surgery can result in decreased
VAP (diagnosed after day 5 of intubation) is more functional residual lung capacity, closure of airways,
frequently associated with aerobic gram-negative bacte- and atelectasis.
ria (Pseudomonas aeruginosa or Acinetobacter species) and The CDC identifies all mechanically ventilated
MRSA.71–75 Additional risk factors for VAP with anti- patients as being at high risk for VAP and recommends
biotic-resistant organisms include recent antibiotic use monitoring hospital rates and trends in VAP using the
and recent admission to a health care facility.75 Patients National Nosocomial Infection Surveillance (NNIS)
with pneumonia caused by Pseudomonas or Acinetobacter system’s surveillance definition of pneumonia. However,
have the highest mortality rates and longer length routine surveillance culturing of patients or respiratory
of stay than patients with other pneumonia patho- equipment is not recommended. Availability of data to
gens.76–78 Nosocomial transmission of organisms appropriate hospital personnel provides feedback regard-
(through respiratory equipment or lack of provider ing efficacy of local VAP prevention efforts.
infection control behaviors) and introduction of gastric
or oropharyngeal organisms into the respiratory tract
(through microaspiration or related to care) are impli- Risk Related to Nosocomial Transmission
cated in VAP. Risk factors for VAP have been identi- of Microorganisms
fied, and several of these factors are attractive targets for The earliest CDC guidelines addressing nosocomial
prevention strategies.75 The most recent CDC recom- pneumonia, published in 1981, placed great emphasis
mendations for prevention of VAP address several of on standardization of practices related to care of respi-
these risk factors (Table 2).79 ratory equipment, and this area has been a continued
focus in subsequent reports. Recommendations related
to procedures for cleaning, sterilization or disinfection,
Risk Related to Patient Factors and maintenance of respiratory equipment now have a
Although all mechanically ventilated patients are at risk strong evidence base and are presented in detail in
for VAP, certain groups of patients are at especially high current the CDC report.79 Attention to compliance
risk and may merit additional clinical vigilance. The with these procedures is primarily the responsibility of
highest VAP rates occur in burn, trauma, and neuro- respiratory therapy but requires the cooperation and
surgical patients, followed by surgical, cardiothoracic, support of physicians and nurses.
and medical populations, whereas the pediatric popula- Hand hygiene is an essential component of VAP
tion has the lowest rate.80 Patients who undergo surgery reduction. In addition to hand decontamination as
and/or trauma of the head, neck, thorax, or abdomen previously discussed, gloves should be worn for han-
have been shown to be at increased risk.81–84 In such dling respiratory secretions or any objects contaminated
patients, respiratory tract instrumentation, anesthesia, with respiratory secretions. If soiling with respiratory
increased use of narcotics and sedatives, and pain may secretions is anticipated, a gown should be worn. Hand
decontamination and glove changes are required decontamination of the gastrointestinal tract (SDD), but
between contacts with different patients and in an these remain unresolved issues in practice recommenda-
encounter with a single patient between contacts with tions. Although stress-bleeding prophylaxis is necessary
a contaminated body site and the respiratory tract or for critically ill patients, gastric alkalinization may con-
respiratory equipment. tribute to VAP by increasing the pool of gastric organ-
isms that serve as a source of VAP pathogens. Despite
much research (including seven meta-analyses since
Risk Related to Positioning and Microaspiration 1991),110 a consensus regarding recommendations for
Backrest elevation to 30 to 45 degrees is an important prophylaxis with histamine-2 receptor antagonists versus
intervention in VAP prevention recommended by the sucralfate, balancing risk of stress-bleeding with risk of
CDC and others.85–87 The supine position has been VAP, has not yet emerged. Acidification of gastric
shown to be an independent risk factor for mortality in feedings has also been proposed as a method to reduce
mechanically ventilated patients77,88 and in all ICU gastric colonization, but there are insufficient data to
patients.89 Backrest elevation reduces risk of gastroeso- recommend routine acidification of feedings. Although
phageal reflux and aspiration90,91 and risk of VAP.92,93 studies have demonstrated reduction in VAP with se-
However, backrest elevation has been underutilized.94–97 lective digestive tract decontamination, this practice
Although patient positioning is a nursing activity, a increases the risks of antibiotic resistance to agents
written physician order for backrest elevation may in- used, overgrowth of nonsusceptible flora, and induction
crease time spent at 30 to 45 degrees.98 In 2005, the Joint of specific resistance mechanisms.111 The concerns about
Commission on the Accreditation of Healthcare Organ- emerging antibiotic resistance preclude recommenda-
izations (JCAHO) added an ICU core measure related tions for SDD at this time.
to backrest elevation (number of ventilator days where In summary, there are several evidence-based
the head of bed is elevated equal to or greater than 30 VAP prevention strategies, including management of
degrees) as part of documenting VAP prevention efforts. respiratory equipment according to published guidelines,
hand hygiene, and maintaining backrest elevation at
30–45 degrees. Many unresolved issues remain. Among
Risk Related to Endotracheal Tube Factors these are consensus on recommendations for stress-
The endotracheal tube enhances bacterial entry into the bleeding prophylaxis that balance risk of bleeding and
lung both by providing a conduit and by serving as a risk of VAP, and recommendations to reduce orophar-
reservoir in which bacteria remain inaccessible to host yngeal and gastric colonization.
defenses.99 VAP risk increases with nasal (versus oral)
intubation100–102 and endotracheal reintubation103,104;
these practices should be avoided if possible. Removal of HAND HYGIENE
subglottic secretions by suctioning through a dorsal lumen Hand hygiene, either by conventional hand washing or
above the endotracheal tube cuff, and removal of secre- disinfection, remains the single most effective method to
tions prior to tube removal, are also recommended.105–107 limit the spread of drug-resistant pathogens in the
hospital setting.112 Conceptually, the cross-transmission
of nosocomial pathogens is summarized as follows113:
Risk Related to Oropharyngeal and Gastric
Colonization *
Organisms are transferred to the hands of the health
Because the oropharyngeal and gastric secretions may care worker either from a patient or from the inani-
provide a nidus for growth of potential VAP pathogens mate environment.
that can be aspirated into the respiratory tract, reducing *
Nosocomial pathogens must then survive on the hand
oropharyngeal and gastric colonization has been a focus of the health care worker.
of VAP prevention efforts. While implementation of a *
Hand hygiene must be either inadequate or omitted.
comprehensive oral-hygiene program is recommended *
The contaminated hands of the health care worker
for those at risk for VAP, evidence-based oral care must then come into contact with another patient or
protocols have not been published. Oral chlorhexidine into contact with an inanimate surface that will later
gluconate (0.12%) rinse begun prior to cardiac surgery come into contact with the patient.
in adults and continued until extubation decreased
VAP108,109 and is now recommended, but because only Conceptually, the microorganisms of the hand
limited data are available for ICU patient populations, can be divided into transient flora and resident flora.114
the use of oral chlorhexidine rinse for prevention of VAP The resident flora is typically of low virulence pathogens
in all critically ill patients is an unresolved issue. such as Micrococcus, coagulase-negative Staphylococcus,
The incidence of VAP has been shown to be and Corynebacterium. Resident flora is difficult to remove
affected by gastric alkalinization, and selective digestive by hand washing. Although these organisms are rarely
pathogenic, they are of nosocomial significance when increase in hand hygiene compliance. Although the
introduced to the patient by invasive procedures and microbicidal effect of chlorhexidine may have resulted
indwelling devices. Transient flora are acquired by pain fewer infections, the difference in nosocomial infec-
tient contact or from the inanimate environment, are tions was also likely due to increased compliance with
loosely attached to the skin, and are more amenable to hand hygiene practices. Regardless, owing to their bac-
removal by hand washing.114 These organisms are of tericidal properties, medicated hand hygiene agents,
nosocomial significance and include MRSA, vancomy- including chlorhexidine, alcohol, and triclosan, should
cin-resistant Enterococcus (VRE), and multiple drug- be considered products of choice, especially in environ-
resistant gram-negative rods. ments with elevated rates of drug-resistant pathogens.
Nosocomial pathogens can be recovered from Sadly, data on health care worker hand hygiene
multiple hospital scenarios. Patient contact, including practice remain discouraging. The reasons for poor
contact with wounds and intact skin, can result in health compliance are multiple and have been studied by
care worker hand contamination.115–118 Areas of high numerous investigators. Observational studies of hand
nosocomial pathogen concentration on patient skin in- hygiene compliance report compliance rates of 5 to
clude the axillae, trunk, perineum, inguinal region, and 81%.123–125 Factors commonly cited that may influence
hands.115,117,118 poor adherence with hand hygiene include insufficient
Hand hygiene should be practiced by health care time, understaffing, patient overcrowding, lack of
workers before and after all patient contact. Several knowledge of hand hygiene guidelines, skepticism about
methods of hand hygiene exist and include washing hand washing efficacy, inconvenient location of sinks
with plain soap and water or using an antibacterial agent and hand disinfectants, and lack of hand hygiene pro-
such as alcohol, chlorhexidine gluconate, or triclosan as motion by the institution.113
either detergent washes or waterless hand rubs. Soap and Even in the ICU setting, hand hygiene remains
water can remove loosely adherent transient skin; how- notoriously poor. A British study performed both an
ever, these agents have minimal antimicrobial activity.113 observation and detailed survey of hand hygiene practices
For effective reduction of bacterial count, a 30-second in 16 ICUs.114 Compliance with hand hygiene and proper
hand rub is recommended. Several factors should be glove use, observed in 381 (non-nurse) health care pro-
considered when choosing a hand hygiene agent for the fessionals, ranged from 9 to 25%. Survey data suggested
ICU, including microbicidal effect, skin irritability, ease that poor compliance with hand hygiene in the ICU was
of use, and staff acceptance. secondary to multiple issues, including ineffective com-
With respect to skin irritability, several studies munication of infection control recommendations, insuf-
have demonstrated that hand washing with both plain ficient promotion of hand antisepsis, and a deficiency of
soap and water can result in skin irritation, dryness, and a infection control education.114 Poor compliance with
paradoxical increase in microbial counts on the hand hygiene was similarly observed by Kaplan and
skin.119,120 Medicated hand washing agents are bacter- McGuckin in a tertiary-care American hospital.126 Physi-
icidal (alcohol, chlorhexidine gluconate, triclosan) and cian compliance with hand hygiene was 19%, whereas
effectively reduced bacterial counts on the hands. More- compliance by the nursing staff was 63%. Greater com-
over, chlorhexidine has the advantage of producing a pliance with hand hygiene was observed among the
residual antibacterial effect, thereby limiting hand re- nursing staff with a 1:1 bed to sink ratio than those
contamination until the time of the next hand hygiene with a greater bed to sink ratio (76% vs 51%).126
episode.121 Efforts to improve hand hygiene in the ICUs will
At least one study supports the effectiveness of likely require multiple, simultaneous interventions, in-
chlorhexidine as a hand antiseptic agent with regard to cluding increased access to hand hygiene products. In a
infection control end points. Doebbeling et al compared study by Bischoff et al where alcohol-based hand sani-
different hand hygiene agents with the end result of tizers were introduced to an ICU, the greatest increment
hand hygiene compliance observation and the reduction in hand hygiene compliance was observed when the hand
of nosocomial infections in an ICU setting.122 During an sanitizer to health care worker ratio went from 1:4 to 1:1,
8-month period, a prospective, multiple crossover trial thereby underscoring the importance of accessibility.123
was conducted in three ICUs. The trial involved 1894 The CDC now suggests promoting alcohol-based hand
adult patients exposed to alternate months of either sanitizer access both by bedside dispensers and by health
chlorhexidine or 60% alcohol solution with the optional care worker pocket-sized dispensers.113 Similarly, Pittet
use of a nonmedicated soap. A greater frequency of and colleagues127 improved overall compliance with
nosocomial infections was seen with the combination hand hygiene by implementing a hospital-wide program
of alcohol and soap compared with the chlorhexidine with emphasis on education, promotion, and bedside,
hand hygiene agent (202 vs 152). However, during alcohol-based hand disinfection. The 3-year campaign
periods of chlorhexidine use, there was a corresponding consisted primarily of hand hygiene promotion through
decrease in the rate of nosocomial infections and an large, conspicuous posters promoting hand hygiene
throughout patient care areas. The project was supported Gowns

and heavily promoted by senior hospital management. Gowns have been used as part of contact precaution
Additionally, alcohol-based hand rub solutions were protocols to limit the spread of nosocomial pathogens.
distributed in large amounts, mounted on beds/walls, Several studies have documented colonization of health
and given to health care workers to encourage packet care worker apparel and instruments during patient care
carriage for convenience of use. During the study, seven activities without the use of gowns.130,131 One study by
institution-wide hand hygiene observational surveys were Boyce et al demonstrated the efficacy of disposable
performed twice yearly. Compliance with hand hygiene gowns in the prevention of health care worker clothing
improved from a baseline of 44% in 1994 to 66% in 1997. contamination.132 Srinivasen et al prospectively meas-
Of note, hand hygiene improved markedly among nurs- ured the effect of gown and glove use in a 16-bed
ing staff but remained poor for physicians. Additionally, medical ICU of a tertiary-care medical center.133 Over
over the study period, the overall prevalence of nosoco- a 3-month period, all admissions to a medical ICU were
mial infections decreased from 16.9 to 9.9%, MRSA screened for VRE by perirectal swab. Patients who were
transmission rates decreased from 2.16 to 0.93 episodes culture positive for VRE were isolated by hospital policy,
per 10,000 patient days, and the consumption of alcohol- requiring the use of gown and gloves for patient care. For
based hand rub increased from 3.5 to 15.4 L per 1000 the following 3 months, precautions were changed to
patient days. Unfortunately, because multiple interven- glove use alone. The VRE acquisition rate was 1.8 cases
tions were employed simultaneously, the relative effect of per 100 patient days at risk in the gown/glove group and
each component was difficult to properly assess. Thus, 3.78 per 100 patient days during glove use alone
although the most efficient and effective means for (p ¼ .04).
sustained improvements in hand hygiene compliance Nevertheless, with regard to the end point of
have yet to be defined, measures should at least include colonization and cross-transmission, there may be little
efforts that stress increased use of accessible, easy to use, incremental benefit to gown use over proper glove use
medicated hand hygiene products, coupled with a hospi- and hand hygiene alone. Pelke et al studied the effect of
tal-wide, administration-supported, high priority hand gowning in a neonatal ICU over an 8-month time frame
hygiene educational and promotional campaign. employing alternating 2-month gowning and nongown-
ing cycles.134 The outcomes of interest were colonization
patterns, necrotizing enterocolitis, respiratory syncytial
The Use of Gloves and Gowns to Limit virus, other nosocomial infections, mortality and hand
Cross-Transmission of Nosocomial Pathogens washing. The investigators failed to document any sig-
Gloves should be worn to prevent health care worker nificant difference between the gowning and nongown-
exposure to bloodborne pathogens and to prevent con- ing cohorts with respect to the rates of bacterial
tamination of hands with drug-resistant pathogens dur- colonization, infection type, or mortality. In addition,
ing patient care activities. Nevertheless, even with proper no significant difference in hand hygiene practice was
glove use, hands may become contaminated during the observed.134
removal of the glove or with microtears that allow for Slaughter et al prospectively compared universal
microorganism transmission.128 Nevertheless, glove use gloving versus universal gown and glove use on the
should not be a substitute for hand hygiene. The acquisition of VRE in a medical ICU.135 Half of the
promotion of glove use may increase compliance with 16 bed ICU was designated for universal gown and glove
hand hygiene protocols. A recent study by Kim and use during patient care activities, the other half was
colleagues observed the rate of hand disinfection with universal gloving for patient care activities. Rectal sur-
glove use and patient isolation.129 In this prospective, veillance cultures were taken daily from patients along
observational study, hand hygiene and glove use com- with monthly environmental cultures of bed rails, bed-
pliance were observed and measured in two ICUs of a side tables, and other common objects in patient rooms.
tertiary-care hospital. Over a 40-hour period of obser- The investigators found no superiority in the universal
vation, 589 opportunities for hand disinfection were use of gowns and gloves versus use of gloves alone in
noted. Overall hand hygiene compliance was 22%. The preventing the rectal colonization of VRE in a medical
investigators found a statistically significant, positive ICU cohort.135 Thus, although the use of gloves and
association between glove use and subsequent hand gowns is the convention for limiting the cross-trans-
disinfection (RR 3.9, 95% CI 2.5 to 6.0). Isolation mission of nosocomial pathogens, the incremental ben-
precautions did not significantly increase hand hygiene efit of gown use, in endemic settings, may be minimal.
compliance. For infection control purposes, glove use
should be promoted as a means of limiting hand con-
tamination with drug-resistant pathogens such as TRANSMISSION-BASED PRECAUTIONS
MRSA and VRE. Additionally, glove use and hand Transmission Based Precautions are for selected patients
hygiene should be promoted concurrently. who are known or suspected to harbor certain infections.
These precautions are divided into three categories, the hospital away from the isolation room should wear a
reflecting differences in disease transmission. Some dis- mask. Examples of diseases requiring droplet precau-
eases may require more than one isolation category. The tions are meningococcal meningitis, Haemophilus influ-
essential elements of transmission-based precautions are enza, influenza, mumps, and German measles (rubella).
summarized in the following sections.
Contact Precautions
Airborne Precautions Contact precautions prevent spread of organisms from
Airborne precautions are designed to prevent diseases an infected patient through direct (touching the patient)
that are transmitted by droplet nuclei or contaminated or indirect (touching surfaces or objects that have been in
dust particles. Droplet nuclei, because of their size, can contact with the patient) contact. This type of precaution
remain suspended in the air for prolonged periods, even requires the patient either be placed in a private room or
after the infected patient has left the room. Agents be cohorted with a roommate with the same infection.
requiring airborne precautions include Mycobacterium Health care workers should don gloves upon entering
tuberculosis, varicella-zoster virus, influenza, and measles the room. After patient care or environmental contact,
virus. All patients needing airborne precautions should the gloves should be removed and hand hygiene should
be assigned to a private room with special engineering be performed prior to leaving the room. In addition, the
and ventilation considerations. The door to this room use protective gowns has been advocated to decrease the
must be closed at all possible times. The isolation room risk of health care worker garment contamination. Pa-
must be maintained at negative pressure in comparison tient care items used for a patient in contact precautions,
to the surroundings. As such, droplet nuclei are pre- such as a stethoscopes and blood pressure cuffs, should
vented from traveling into the environment. In addition, not be shared with other patients unless they are properly
the air within the isolation room should either be vented cleaned and disinfected before reuse. Patients should be
to the outside or passed through high-efficiency particle restricted to the isolation room.
filters.136 Contact precautions are indicated for patients
All personnel entering the isolation room are with drug-resistant pathogens such as MRSA, VRE,
required by federal regulations to don masks for respi- and multidrug-resistant gram-negative rods. In addition,
ratory protection. If a patient must move from the contact isolation is recommended for diarrheal illnesses
isolation room to another area of the hospital, the patient of infectious origin and for infections with Clostridia
should be wearing a mask during the transport. Anyone difficile.
entering the isolation room to provide care to the patient
must wear a special mask called a respirator. These
respirator masks are approved by the National Institute Potential Adverse Effects of Isolation Practices
for Occupational Safety and Health and are capable of The use of strict isolation practices may have a detri-
filtering 1 mm particles with an efficiency of 95% (N-95 mental impact on the process and quality of patient care.
mask). By regulation, all health care workers must be fit Evans et al prospectively observed surgical patients both
tested for N-95 masks and must be taught to check for in the ICU and on a general surgical floor. Both in the
proper fit each time prior to use.136 Rapid airborne ICU and on the surgical floor, surgical patients in
isolation of patients with known or suspected multi- contact isolation had fewer health care worker visits
drug-resistant M. tuberculosis, along with proper N-95 and less contact time overall despite a higher severity
mask use by health care workers, is essential to limit the of illness as measured by APACHE (acute physiology
spread of this pathogen. and chronic health assessment) II score.137 Stelfox
et al138 studied the quality of medical care received by
patients isolated for MRSA-related infection control
Droplet Precautions precautions using a case control study design. Although
Droplet precautions prevent the transmission of organ- isolated and control patients had similar baseline char-
isms that travel via droplets generated during phonation, acteristics, isolated patients were twice as likely as non-
sneezing, coughing, or invasive respiratory tract proce- isolated patients to experience adverse events during
dures. These particles are not suspended in the air for their hospitalization. These adverse events included
extended periods and typically do not travel beyond supportive care measures and process of care measures
several feet from the patient. Patients who require such as days with incomplete or absent vitals signs, and
droplet precautions should be placed in a private room days without documented nursing and physician prog-
or should be cohorted with a roommate who is infected ress notes. Additionally, patients on MRSA contact
with the same organism. The door to the room may isolation expressed greater dissatisfaction with the qual-
remain open. Health care workers should wear a mask ity of their treatment.138 Similarly, Saint and colleagues
when within 3 ft of the patient. Patients moving about observed in a prospective cohort study of two in-patient
medical services, that patients in contact isolation were minimize the steps of catheter insertion, and the stand-
half as likely to be examined by an attending physician as ardized questioning of daily providers whether catheters
nonisolated patients.139 could be removed.41 Most importantly, a checklist to
Contact isolation may have a detrimental psycho- ensure adherence to evidence-based guidelines for pre-
logical impact on patients. One cross-sectional matched venting CRBSIs was implemented along with the em-
case control study compared contact-isolated versus powerment of nurses to stop the catheter insertion
nonisolated elderly patients.140 The level of depressive procedure if a violation of the guidelines was observed.
and anxiety symptoms exhibited by the contact isolation Evidence-based catheter insertion process-of-care meas-
group exceeded that of the noncontact isolation group. ures included a nurse’s checklist for observed hand
Catalano et al prospectively studied the impact of con- hygiene, patient skin antisepsis, and proper use of sterile
tact isolation on anxiety and depression in noncritically gloves, gown, and drape. Additionally, the checklist
ill hospitalized patients.141 Patients in contact isolation included confirmation that all personnel complied with
for either MRSA or VRE were compared with other infection control precautions and that a sterile field was
hospitalized patients with infectious diseases not requir- maintained during the procedure. The investigators
ing isolation. All patients were evaluated with the observed that physicians followed infection control
Hamilton Anxiety and Depression Rating scale at base- guidelines during 62% of the procedures. During the
line and then later during the hospital course. Although intervention time period, a nursing intervention was
no significant differences in baseline anxiety and depres- required in 32% of all CVC insertions. Following the
sion scores were noted, for patients in contact isolation, intervention, the CRBSI rate in the study ICU de-
statistically significant higher scores on both scales were creased from 11.3/1000 catheter days in the first quarter
reported later during the course of hospitalization. of 1998 to 0/1000 catheter days in the fourth quarter of
Thus the optimal strategy for control of endemic, 2002. The CRBSI rate in the control ICU was 5.7/1000
resistant pathogens such as MRSA or VRE has yet to be catheter days in the first quarter of 1998 and 1.6/1000
defined. ICU directors will have to weigh the infection- catheter days in the fourth quarter of 2002 (p ¼ .56). As
control benefits of patient isolation against the potential per the estimates of Berenholtz et al, the initiative
risk of adverse events. may have prevented 43 CRBSIs, eight deaths, and
$1,945,922 in additional costs per year in the study ICU.
Wall et al similarly studied the impact of using
INFECTION CONTROL PROCESS-OF-CARE real-time process measures to reduce catheter-related
MEASURES BSIs in the ICU.39 An interdisciplinary team developed
Traditional infection control programs involve surveil- a standardized, nursing checklist for CVC insertion.
lance and feedback of outcome measures, such as BSI Infection control practitioners scanned the completed
and VAP rates.142 Outcome measures are, however, checklists into a computerized database, thereby gener-
uncommon events that take longer to observe and also ating real-time measurements for the process of CVC
may not directly relate to individual or group perform- insertion. These infection control process measures al-
ance because they are frequently affected by factors lowed the ICU team to directly monitor adherence to
related to the patient and the healthcare system.143,144 evidence-based guidelines. After 2 years, the investiga-
However, process measures are attractive for several tors reported a persistent and historically low CRBSI
reasons. They provide operational and measurable rep- rate. Thus it appears that multifaceted programs ensur-
resentation of performance, relate to individual and/or ing maximal adherence with evidence-based infection
group performance, and are easier to measure than control guidelines are effective in reducing the incidence
outcomes. By potentially increasing overall accountabil- of catheter-related BSIs in the intensive care setting.
ity, they create opportunities to monitor and improve
performance. Provided a process measure and outcome
are linked, interventions that improves the process CONCLUSION
measure should theoretically improve the outcome.145 The prevalence of hospital-acquired, antibiotic-resistant
In addition, several recent studies have shown that pathogens has increased significantly over the past
performance feedback and accountability can have a 20 years. Hospital infection control programs are seen
positive influence on hand hygiene compliance and on as increasingly important for the control of antibiotic-
the reduction of CRBSI and catheter-related urinary resistant organisms. Strategies to control the spread of
tract infections.146,147 hospital-acquired infections by drug-resistant pathogens
Berenholtz et al studied the effect of a multi- are multiple. The patient, the health care worker, and the
faceted systems intervention on catheter-related BSIs in environment are reservoirs for drug-resistant pathogens.
the ICU.41 The strategy included a quality improvement For high-risk patients colonized with MRSA, such as
team that implemented five interventions. These were surgical candidates and those in ICU, decolonization
staff education, the creation of a catheter insertion cart to with nasal mupirocin should be considered. Patients
colonized with resistant pathogens such as MRSA, VRE, 6. Martin MA, Pfaller MA, Wenzel RP. Coagulase-negative
and drug-resistant gram-negative rods can contaminate staphylococcal bacteremia: mortality and hospital stay. Ann
the environment. As such, all health care facilities should Intern Med 1989;110:9–16
7. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream
develop policies for the terminal and periodic disinfection
infection in critically ill patients: excess length of stay, extra
of patient care areas and environmental services. Cross- costs, and attributable mortality. JAMA 1994;271:1598–
transmission of nosocomial pathogens by the hands of 1601
health care workers has been well documented. Metic- 8. Jones ME, Draghi DC, Thornsberry C, Karlowsky JA, Sahm
ulous hand hygiene should be practiced with medicated DF, Wenzel RP. Emerging resistance among bacterial
hand washing agents (alcohol, chlorhexidine gluconate, pathogens in the intensive care unit: a European and North
triclosan) that are bactericidal and effectively reduce American Surveillance study (2000–2002). Ann Clin Micro-
biol Antimicrob 2004;3:14
bacterial counts on the hands. Measures to promote
9. Wenzel RP, Sahm DF, Thornsberry C, Draghi DC, Jones
hand hygiene compliance should include efforts that ME, Karlowsky JA. In vitro susceptibilities of gram-negative
stress increased use of accessible, easy to use, medicated bacteria isolated from hospitalized patients in four European
hand hygiene products, coupled with a hospital-wide, countries, Canada, and the United States in 2000–2001 to
administration-backed, high priority hand hygiene cam- expanded-spectrum cephalosporins and comparator antimi-
paign. Glove use is beneficial in limiting the contami- crobials: implications for therapy. Antimicrob Agents Che-
nation of health-care worker hands but is not a substitute mother 2003;47:3089–3098
10. Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH.
for hand hygiene. Concerns about the contamination of
The influence of inadequate antimicrobial treatment of
personnel clothing with nosocomial pathogens has led to bloodstream infections on patient outcomes in the ICU
the use of gowns for patients in contact isolation. The setting. [see comments]. Chest 2000;118:146–155
incremental benefit of gowns and gloves use may be 11. Miller PJ, Searcy MA, Kaiser DL, Wenzel RP. The
minimal. Transmission-based precautions are useful for relationship between surgeon experience and endometritis
the control of nosocomial infections and include contact, after cesarean section. Surg Gynecol Obstet 1987;165:535–
airborne, and droplet precautions. Aggressive surveillance 539
12. Gudlaugsson O, Gillespie S, Lee K, et al. Attributable
for asymptomatic reservoirs may be of value but is not
mortality of nosocomial candidemia, revisited. Clin Infect
without controversy, including questions about efficacy Dis 2003;37:1172–1177
and effect on quality of care. Other considerations for an 13. Pfaller MA, Messer SA, Boyken L, et al. In vitro activities of
infection control program include antibiotic control pro- voriconazole, posaconazole, and fluconazole against 4,169
grams and surveillance systems for infections with noso- clinical isolates of Candida spp. and Cryptococcus neoformans
comial pathogens. This type of surveillance is essential for collected during 2001 and 2002 in the ARTEMIS global
establishing endemic rates, defining outbreaks, and de- antifungal surveillance program. Diagn Microbiol Infect Dis
2004;48:201–205
veloping institution-specific antibiograms. In the end,
14. Wenzel RP, Gennings C. Bloodstream infections due to
the purpose of a hospital infections surveillance program Candida species in the intensive care unit: identifying
is to define endemic rates, recognize outbreaks, and especially high-risk patients to determine prevention strate-
obtain data of value in recognizing the extent and gies. Clin Infect Dis 2005;41(Suppl 6):S389–S393
causation of the infections. These data are later applied 15. Franchi D, Climo MW, Wong AH, Edmond MB, Wenzel
for the planning and implementation of risk reduction RP. Seeking vancomycin resistant Staphylococcus aureus
policies and interventions. among patients with vancomycin-resistant enterococci. Clin
Infect Dis 1999;29:1566–1568
16. Richards MJ, Edwards JR, Culver DH, Gaynes RP.
REFERENCES Nosocomial infections in medical intensive care units in
the United States. National Nosocomial Infections Surveil-
1. Wenzel RP. Perspective: attributable mortality—the pro- lance System. Crit Care Med 1999;27:887–892
mise of better antimicrobial therapy. J Infect Dis 1998;178: 17. Rojo D, Pinedo A, Clavijo E, Garcia-Rodriguez A, Garcia
917–919 V. Analysis of risk factors associated with nosocomial
2. Wenzel RP, Edmond MB. The impact of hospital-acquired bacteraemias. J Hosp Infect 1999;42:135–141
bloodstream infections. Emerg Infect Dis 2001;7:174–177 18. Digiovine B, Chenoweth C, Watts C, Higgins M. The
3. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel attributable mortality and costs of primary nosocomial
RP, Edmond MB. Nosocomial bloodstream infections in US bloodstream infections in the intensive care unit. Am J
hospitals: analysis of 24,179 cases from a prospective nation- Respir Crit Care Med 1999;160:976–981
wide surveillance study. Clin Infect Dis 2004;39:309–317 19. Rello J, Ochagavia A, Sabanes E, et al. Evaluation of outcome
4. Landry SL, Kaiser DL, Wenzel RP. Hospital stay and of intravenous catheter-related infections in critically ill
mortality attributed to nosocomial enterococcal bacteremia: a patients. Am J Respir Crit Care Med 2000;162:1027–1030
controlled study. Am J Infect Control 1989;17:323–329 20. Soufir L, Timsit JF, Mahe C, Carlet J, Regnier B, Chevret S.
5. Wey SB, Mori M, Pfaller MA, Woolson RF, Wenzel RP. Attributable morbidity and mortality of catheter-related
Hospital-acquired candidemia: the attributable mortality and septicemia in critically ill patients: a matched, risk-adjusted,
excess length of stay. Arch Intern Med 1988;148:2642– cohort study. Infect Control Hosp Epidemiol 1999;20:
2645 396–401
21. Dimick JB, Pelz RK, Consunji R, Swoboda SM, Hendrix 39. Wall RJ, Ely EW, Elasy TA, et al. Using real-time process
CW, Lipsett PA. Increased resource use associated with measurements to reduce catheter related bloodstream infec-
catheter-related bloodstream infection in the surgical tions in the intensive care unit. Qual Saf Health Care 2005;
intensive care unit. Arch Surg 2001;136:229–234 14:295–302
22. Mermel LA. Prevention of intravascular catheter-related 40. Reduction in central line-associated bloodstream infections
infections. Ann Intern Med 2000;132:391–402 among patients in intensive care units: Pennsylvania, April
23. Saint S, Veenstra DL, Lipsky BA. The clinical and economic 2001–March 2005. MMWR Morb Mortal Wkly Rep 2005;
consequences of nosocomial central venous catheter-related 54:1013–1016
infection: are antimicrobial catheters useful? Infect Control 41. Berenholtz SM, Pronovost PJ, Lipsett PA, et al. Eliminating
Hosp Epidemiol 2000;21:375–380 catheter-related bloodstream infections in the intensive care
24. O’Grady NP, Alexander M, Dellinger EP, et al. Guidelines unit. Crit Care Med 2004;32:2014–2020
for the prevention of intravascular catheter-related infec- 42. Hind D, Calvert N, McWilliams R, et al. Ultrasonic locating
tions. Centers for Disease Control and Prevention. MMWR devices for central venous cannulation: meta-analysis. BMJ
Recomm Rep 2002;51:1–29 2003;327:361–364
25. Sessler CN. Preventing mechanical and infectious complica- 43. Milling TJ Jr, Rose J, Briggs WM, et al. Randomized,
tions of central venous catheterization. In: Vincent JL, ed. controlled clinical trial of point-of-care limited ultrasono-
Yearbook of Intensive Care and Emergency Medicine. graphy assistance of central venous cannulation: The Third
Berlin: Springer; 2004:544–557 Sonography Outcomes Assessment Program (SOAP-3)
26. Saint S. Prevention of intravascular catheter-associated infec- Trial. Crit Care Med 2005;33:1764–1769
tions. http:/ /www.ahrq.gov/clinic/ptsafety/chap16a.htm. 2001 44. Casey AL, Worthington T, Lambert PA, Quinn D, Faroqui
27. Eggimann P, Pittet D. Infection control in the ICU. Chest MH, Elliott TS. A randomized, prospective clinical trial to
2001;120:2059–2093 assess the potential infection risk associated with the
28. Raad II, Hohn DC, Gilbreath BJ, et al. Prevention of central PosiFlow needleless connector. J Hosp Infect 2003;54:288–
venous catheter–related infections by using maximal sterile 293
barrier precautions during insertion. Infect Control Hosp 45. Eggimann P, Harbarth S, Constantin MN, Touveneau S,
Epidemiol 1994;15:231–238 Chevrolet JC, Pittet D. Impact of a prevention strategy
29. Chaiyakunapruk N, Veenstra DL, Lipsky BA, Saint S. targeted at vascular-access care on incidence of infections
Chlorhexidine compared with povidone-iodine solution for acquired in intensive care. Lancet 2000;355:1864–1868
vascular catheter-site care: a meta-analysis. Ann Intern Med 46. Sherertz RJ, Ely EW, Westbrook DM, et al. Education of
2002;136:792–801 physicians-in-training can decrease the risk for vascular
30. Cobb DK, High KP, Sawyer RG, et al. A controlled trial of catheter infection. Ann Intern Med 2000;132:641–648
scheduled replacement of central venous and pulmonary- 47. Coopersmith CM, Rebmann TL, Zack JE, et al. Effect of an
artery catheters. N Engl J Med 1992;327:1062–1068 education program on decreasing catheter-related blood-
31. Cook D, Randolph A, Kernerman P, et al. Central venous stream infections in the surgical intensive care unit. Crit Care
catheter replacement strategies: a systematic review of the Med 2002;30:59–64
literature. Crit Care Med 1997;25:1417–1424 48. Higuera F, Rosenthal VD, Duarte P, Ruiz J, Franco G,
32. Ruesch S, Walder B, Tramer MR. Complications of central Safdar N. The effect of process control on the incidence of
venous catheters: internal jugular versus subclavian access– central venous catheter-associated bloodstream infections
a systematic review. Crit Care Med 2002;30:454–460 and mortality in intensive care units in Mexico. Crit Care
33. Heard SO, Wagle M, Vijayakumar E, et al. Influence of Med 2005;33:2022–2027
triple-lumen central venous catheters coated with chlorhex- 49. Martin M, Scalabrini B, Rioux A, Xhignesse MA. Training
idine and silver sulfadiazine on the incidence of catheter- fourth-year medical students in critical invasive skills
related bacteremia. Arch Intern Med 1998;158:81–87 improves subsequent patient safety. Am Surg 2003;69:437–
34. Goetz AM, Wagener MM, Miller JM, Muder RR. Risk of 440
infection due to central venous catheters: effect of site of 50. Sessler CN, Perry J, Sampson R, Edmond MB. CD-ROM-
placement and catheter type. Infect Control Hosp Epidemiol based program for medical student and housestaff education
1998;19:842–845 regarding central venous catheter procedure and prevention
35. Merrer J, De Jonghe B, Golliot F, et al. Complications of of complications [abstract]. Crit Care Med 2005;32:A67
femoral and subclavian venous catheterization in critically ill 51. Veenstra DL, Saint S, Saha S, Lumley T, Sullivan SD.
patients: a randomized controlled trial. JAMA 2001;286: Efficacy of antiseptic-impregnated central venous catheters
700–707 in preventing catheter-related bloodstream infection: a meta-
36. Deshpande KS, Hatem C, Ulrich HL, et al. The incidence of analysis. JAMA 1999;281:261–267
infectious complications of central venous catheters at the 52. Darouiche RO, Raad II, Heard SO, et al. A comparison of
subclavian, internal jugular, and femoral sites in an intensive two antimicrobial-impregnated central venous catheters.
care unit population. Crit Care Med 2005;33:13–20discus- Catheter Study Group. [see comments]. N Engl J Med
sion 234–235 1999;340:1–8
37. Durbec O, Viviand X, Potie F, Vialet R, Albanese J, Martin 53. Rupp ME, Lisco SJ, Lipsett PA, et al. Effect of a second-
C. A prospective evaluation of the use of femoral venous generation venous catheter impregnated with chlorhexidine
catheters in critically ill adults. Crit Care Med 1997;25: and silver sulfadiazine on central catheter-related infections:
1986–1989 a randomized, controlled trial. Ann Intern Med 2005;
38. Lorente L, Villegas J, Martin MM, Jimenez A, Mora ML. 143:570–580
Catheter-related infection in critically ill patients. Intensive 54. Brun-Buisson C, Doyon F, Sollet JP, Cochard JF, Cohen Y,
Care Med 2004;30:1681–1684 Nitenberg G. Prevention of intravascular catheter-related
infection with newer chlorhexidine-silver sulfadiazine- 71. Langer M, Cigada M, Mandelli M, Mosconi P, Tognoni G.
coated catheters: a randomized controlled trial. Intensive Early onset pneumonia: a multicenter study in intensive care
Care Med 2004;30:837–843 units. Intensive Care Med 1987;13:342–346
55. Corral L, Nolla-Salas M, Ibanez-Nolla J, et al. A 72. Pugin J, Auckenthaler R, Mili N, Janssens JP, Lew PD,
prospective, randomized study in critically ill patients using Suter PM. Diagnosis of ventilator-associated pneumonia by
the Oligon Vantex catheter. J Hosp Infect 2003;55:212– bacteriologic analysis of bronchoscopic and nonbroncho-
219 scopic ‘‘blind’’ bronchoalveolar lavage fluid. Am Rev Respir
56. Moretti EW, Ofstead CL, Kristy RM, Wetzler HP. Impact Dis 1991;143:1121–1129
of central venous catheter type and methods on catheter- 73. Driks MR, Craven DE, Celli BR, et al. Nosocomial
related colonization and bacteraemia. J Hosp Infect 2005; pneumonia in intubated patients given sucralfate as com-
61:139–145 pared with antacids or histamine type 2 blockers: the role of
57. Ranucci M, Isgro G, Giomarelli PP, et al. Impact of oligon gastric colonization. N Engl J Med 1987;317:1376–1382
central venous catheters on catheter colonization and 74. Prod’hom G, Leuenberger P, Koerfer J, et al. Nosocomial
catheter-related bloodstream infection. Crit Care Med pneumonia in mechanically ventilated patients receiving
2003;31:52–59 antacid, ranitidine, or sucralfate as prophylaxis for stress
58. Shorr AF, Humphreys CW, Helman DL. New choices for ulcer: a randomized controlled trial. Ann Intern Med 1994;
central venous catheters: potential financial implications. 120:653–662
Chest 2003;124:275–284 75. American Thoracic Society. Guidelines for the management
59. Levy I, Katz J, Solter E, et al. Chlorhexidine-impregnated of adults with hospital-acquired, ventilator-associated, and
dressing for prevention of colonization of central venous healthcare-associated pneumonia. Am J Respir Crit Care
catheters in infants and children: a randomized controlled Med 2005;171:388–416
study. Pediatr Infect Dis J 2005;24:676–679 76. Fagon JY, Chastre J, Hance AJ, Montravers P, Novara A,
60. Carratala J, Niubo J, Fernandez-Sevilla A, et al. Rando- Gibert C. Nosocomial pneumonia in ventilated patients: a
mized, double-blind trial of an antibiotic-lock technique for cohort study evaluating attributable mortality and hospital
prevention of gram-positive central venous catheter-related stay. Am J Med 1993;94:281–288
infection in neutropenic patients with cancer. Antimicrob 77. Kollef MH. Ventilator-associated pneumonia: a multivariate
Agents Chemother 1999;43:2200–2204 analysis. JAMA 1993;270:1965–1970
61. Leon C, Alvarez-Lerma F, Ruiz-Santana S, et al. Antiseptic 78. Kollef MH, Silver P, Murphy DM, Trovillion E. The effect
chamber-containing hub reduces central venous catheter- of late-onset ventilator-associated pneumonia in determin-
related infection: a prospective, randomized study. Crit Care ing patient mortality. Chest 1995;108:1655–1662
Med 2003;31:1318–1324 79. Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R.
62. Appelgren P, Ransjo U, Bindslev L, Espersen F, Larm O. Guidelines for preventing health-care-associated pneumo-
Surface heparinization of central venous catheters reduces nia, 2003: recommendations of CDC and the Healthcare
microbial colonization in vitro and in vivo: results from a Infection Control Practices Advisory Committee. MMWR
prospective, randomized trial. Crit Care Med 1996;24:1482– Recomm Rep 2004;53:1–36
1489 80. National Nosocomial Infections Surveillance (NNIS) Sys-
63. Randolph AG, Cook DJ, Gonzales CA, Brun-Buisson C. tem Report. Data summary from January 1992 through June
Tunneling short-term central venous catheters to prevent 2004, issued October 2004. Am J Infect Control 2004;32:
catheter- related infection: a meta-analysis of randomized, 470–485
controlled trials. Crit Care Med 1998;26:1452–1457 81. Chevret S, Hemmer M, Carlet J, Langer M. Incidence and
64. Timsit JF, Bruneel F, Cheval C, et al. Use of tunneled risk factors of pneumonia acquired in intensive care units:
femoral catheters to prevent catheter-related infection: a results from a multicenter prospective study on 996 patients.
randomized, controlled trial. Ann Intern Med 1999;130: European Cooperative Group on Nosocomial Pneumonia.
729–735 Intensive Care Med 1993;19:256–264
65. Safdar N, Maki DG. Risk of catheter-related bloodstream 82. Haley RW. Nosocomial infections in surgical patients:
infection with peripherally inserted central venous catheters developing valid measures of intrinsic patient risk. Am J
used in hospitalized patients. Chest 2005;128:489–495 Med 1991;91:145S–151S
66. Garibaldi RA, Mooney BR, Epstein BJ, Britt MR. An 83. Celis R, Torres A, Gatell JM, Almela M, Rodriguez-Roisin
evaluation of daily bacteriologic monitoring to identify R, Agusti-Vidal A. Nosocomial pneumonia: a multivariate
preventable episodes of catheter-associated urinary tract analysis of risk and prognosis. Chest 1988;93:318–324
infection. Infect Control 1982;3:466–470 84. George DL. Epidemiology of nosocomial pneumonia in
67. Stark RP, Maki DG. Bacteriuria in the catheterized patient. intensive care unit patients. Clin Chest Med 1995;16:29–44
What quantitative level of bacteriuria is relevant? N Engl J 85. Kollef MH. The prevention of ventilator-associated pneu-
Med 1984;311:560–564 monia. N Engl J Med 1999;340:627–634
68. Maki DG. Nosocomial bacteremia: an epidemiologic over- 86. Hixson S, Sole ML, King T. Nursing strategies to prevent
view. Am J Med 1981;70:719–732 ventilator-associated pneumonia. AACN Clin Issues 1998;
69. Maki DG, Tambyah PA. Engineering out the risk for 9:76–90quiz 145–146
infection with urinary catheters. Emerg Infect Dis 2001; 87. Tasota FJ, Fisher EM, Coulson CF, Hoffman LA.
7:342–347 Protecting ICU patients from nosocomial infections: prac-
70. Safdar N, Dezfulian C, Collard HR, Saint S. Clinical and tical measures for favorable outcomes. Crit Care Nurse 1998;
economic consequences of ventilator-associated pneumo- 18:54–65; quiz 66–57
nia: a systematic review. Crit Care Med 2005;33:2184– 88. Craven DE, Kunches LM, Kilinsky V, Lichtenberg DA,
2193 Make BJ, McCabe WR. Risk factors for pneumonia and
fatality in patients receiving continuous mechanical ventila- ventilator-associated pneumonia: a meta-analysis. Am J Med
tion. Am Rev Respir Dis 1986;133:792–796 2005;118:11–18
89. Kollef MH. The identification of ICU-specific outcome 106. Kollef MH, Skubas NJ, Sundt TM. A randomized clinical
predictors: a comparison of medical, surgical, and cardi- trial of continuous aspiration of subglottic secretions in
othoracic ICUs from a single institution. Heart Lung 1995; cardiac surgery patients. Chest 1999;116:1339–1346
24:60–66 107. Valles J, Artigas A, Rello J, et al. Continuous aspiration of
90. Torres A, Serra-Batlles J, Ros E, et al. Pulmonary aspiration subglottic secretions in preventing ventilator-associated
of gastric contents in patients receiving mechanical ventila- pneumonia. Ann Intern Med 1995;122:179–186
tion: the effect of body position. Ann Intern Med 1992;116: 108. Fourrier F, Duvivier B, Boutigny H, Roussel-Delvallez M,
540–543 Chopin C. Colonization of dental plaque: a source of
91. Orozco-Levi M, Torres A, Ferrer M, et al. Semirecumbent nosocomial infections in intensive care unit patients. Crit
position protects from pulmonary aspiration but not Care Med 1998;26:301–308
completely from gastroesophageal reflux in mechanically 109. Fourrier F, Cau-Pottier E, Boutigny H, Roussel-Delvallez
ventilated patients. Am J Respir Crit Care Med 1995; M, Jourdain M, Chopin C. Effects of dental plaque
152:1387–1390 antiseptic decontamination on bacterial colonization and
92. Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogue S, nosocomial infections in critically ill patients. Intensive Care
Ferrer M. Supine body position as a risk factor for Med 2000;26:1239–1247
nosocomial pneumonia in mechanically ventilated patients: 110. Collard HR, Saint S, Matthay MA. Prevention of ventilator-
a randomised trial. Lancet 1999;354:1851–1858 associated pneumonia: an evidence-based systematic review.
93. Grap MJ, Munro CL, Hummel RS 3rd, Elswick RK Jr, Ann Intern Med 2003;138:494–501
McKinney JL, Sessler CN. Effect of backrest elevation on 111. Duncan RA, Steger KA, Craven DE. Selective decontami-
the development of ventilator-associated pneumonia. Am J nation of the digestive tract: risks outweigh benefits for
Crit Care 2005;14:325–332; quiz 333 intensive care unit patients. Semin Respir Infect 1993;8:308–
94. Evans D. The use of position during critical illness: current 324
practice and review of the literature. Aust Crit Care 112. Larson EL. APIC guideline for handwashing and hand
1994;7:16–21 antisepsis in health care settings. Am J Infect Control 1995;
95. Grap MJ, Cantley M, Munro CL, Corley MC. Use of 23:251–269
backrest elevation in critical care: a pilot study. Am J Crit 113. Boyce JM, Pittet D. Guideline for Hand Hygiene in Health-
Care 1999;8:475–480 Care Settings. Recommendations of the Healthcare Infec-
96. Grap MJ, Munro CL, Bryant S, Ashtiani B. Predictors of tion Control Practices Advisory Committee and the
backrest elevation in critical care. Intensive Crit Care Nurs HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force.
2003;19:68–74 Society for Healthcare Epidemiology of America/Associa-
97. Grap MJ, Munro CL. Quality improvement in backrest tion for Professionals in Infection Control/Infectious Dis-
elevation: improving outcomes in critical care. AACN Clin eases Society of America. MMWR Recomm Rep 2002;51:1–
Issues 2005;16:133–139 45; quiz CE41–44
98. Helman DL Jr, Sherner JH III, Fitzpatrick TM, Callender 114. Sproat LJ, Inglis TJ. A multicentre survey of hand hygiene
ME, Shorr AF. Effect of standardized orders and provider practice in intensive care units. J Hosp Infect 1994;26:137–
education on head-of-bed positioning in mechanically 148
ventilated patients. Crit Care Med 2003;31:2285–2290 115. Sanford MD, Widmer AF, Bale MJ, Jones RN, Wenzel RP.
99. Levine SA, Niederman MS. The impact of tracheal Efficient detection and long-term persistence of the carriage
intubation on host defenses and risks for nosocomial of methicillin-resistant Staphylococcus aureus. Clin Infect Dis
pneumonia. Clin Chest Med 1991;12:523–543 1994;19:1123–1128
100. Aebert H, Hunefeld G, Regel G. Paranasal sinusitis and 116. Bertone SA, Fisher MC, Mortensen JE. Quantitative skin
sepsis in ICU patients with nasotracheal intubation. Inten- cultures at potential catheter sites in neonates. Infect Control
sive Care Med 1988;15:27–30 Hosp Epidemiol 1994;15:315–318
101. Grindlinger GA, Niehoff J, Hughes SL, Humphrey MA, 117. Bonten MJ, Hayden MK, Nathan C, et al. Epidemiology of
Simpson G. Acute paranasal sinusitis related to nasotracheal colonisation of patients and environment with vancomycin-
intubation of head-injured patients. Crit Care Med 1987; resistant enterococci. Lancet 1996;348:1615–1619
15:214–217 118. Larson EL, Cronquist AB, Whittier S, Lai L, Lyle CT,
102. Salord F, Gaussorgues P, Marti-Flich J, et al. Nosocomial Della Latta P. Differences in skin flora between inpatients
maxillary sinusitis during mechanical ventilation: a prospec- and chronically ill outpatients. Heart Lung 2000;29:298–305
tive comparison of orotracheal versus the nasotracheal route 119. Winnefeld M, Richard MA, Drancourt M, Grob JJ. Skin
for intubation. Intensive Care Med 1990;16:390–393 tolerance and effectiveness of two hand decontamination
103. Torres A, Aznar R, Gatell JM, et al. Incidence, risk, and procedures in everyday hospital use. Br J Dermatol 2000;
prognosis factors of nosocomial pneumonia in mechanically 143:546–550
ventilated patients. Am Rev Respir Dis 1990;142:523– 120. Boyce JM, Kelliher S, Vallande N. Skin irritation and
528 dryness associated with two hand-hygiene regimens: soap-
104. Torres A, Gatell JM, Aznar E, et al. Re-intubation increases and-water hand washing versus hand antisepsis with an
the risk of nosocomial pneumonia in patients needing alcoholic hand gel. Infect Control Hosp Epidemiol 2000;
mechanical ventilation. Am J Respir Crit Care Med 1995; 21:442–448
152:137–141 121. Wade JJ, Casewell MW. The evaluation of residual
105. Dezfulian C, Shojania K, Collard HR, Kim HM, Matthay antimicrobial activity on hands and its clinical relevance.
MA, Saint S. Subglottic secretion drainage for preventing J Hosp Infect 1991;18(Suppl B):23–28
122. Doebbeling BN, Stanley GL, Sheetz CT, et al. Comparative care unit. Arch Pediatr Adolesc Med 1994;148:1016–
efficacy of alternative hand-washing agents in reducing 1020
nosocomial infections in intensive care units. N Engl J Med 135. Slaughter S, Hayden MK, Nathan C, et al. A comparison of
1992;327:88–93 the effect of universal use of gloves and gowns with that of
123. Bischoff WE, Reynolds TM, Sessler CN, Edmond MB, glove use alone on acquisition of vancomycin-resistant
Wenzel RP. Handwashing compliance by health care enterococci in a medical intensive care unit. Ann Intern
workers: the impact of introducing an accessible, alcohol- Med 1996;125:448–456
based hand antiseptic. Arch Intern Med 2000;160:1017– 136. Guidelines for preventing the transmission of Mycobacterium
1021 tuberculosis in health-care facilities, 1994. Centers for Disease
124. Maury E, Alzieu M, Baudel JL, et al. Availability of an Control and Prevention. MMWR Morb Mortal Wkly Rep
alcohol solution can improve hand disinfection compliance 1994;43:1–132
in an intensive care unit. Am J Respir Crit Care Med 2000; 137. Evans HL, Shaffer MM, Hughes MG, et al. Contact
162:324–327 isolation in surgical patients: a barrier to care? Surgery 2003;
125. Muto CA, Sistrom MG, Farr BM. Hand hygiene rates 134:180–188
unaffected by installation of dispensers of a rapidly acting 138. Stelfox HT, Bates DW, Redelmeier DA. Safety of patients
hand antiseptic. Am J Infect Control 2000;28:273–276 isolated for infection control. JAMA 2003;290:1899–1905
126. Kaplan LM, McGuckin M. Increasing handwashing com- 139. Saint S, Higgins LA, Nallamothu BK, Chenoweth C. Do
pliance with more accessible sinks. Infect Control 1986;7: physicians examine patients in contact isolation less fre-
408–410 quently? A brief report. Am J Infect Control 2003;31:354–
127. Pittet D, Hugonnet S, Harbarth S, et al. Effectiveness of a 356
hospital-wide programme to improve compliance with hand 140. Tarzi S, Kennedy P, Stone S, Evans M. Methicillin-resistant
hygiene: Infection Control Programme. Lancet 2000;356: Staphylococcus aureus: psychological impact of hospitalization
1307–1312 and isolation in an older adult population. J Hosp Infect
128. Doebbeling BN, Pfaller MA, Houston AK, Wenzel RP. 2001;49:250–254
Removal of nosocomial pathogens from the contaminated 141. Catalano G, Houston SH, Catalano MC, et al. Anxiety and
glove: implications for glove reuse and handwashing. Ann depression in hospitalized patients in resistant organism
Intern Med 1988;109:394–398 isolation. South Med J 2003;96:141–145
129. Kim PW, Roghmann MC, Perencevich EN, Harris AD. 142. Delgado-Rodriguez M, Gomez-Ortega A, Sillero-Arenas
Rates of hand disinfection associated with glove use, patient M, Martinez-Gallego G, Medina-Cuadros M, Llorca J.
isolation, and changes between exposure to various body Efficacy of surveillance in nosocomial infection control in a
sites. Am J Infect Control 2003;31:97–103 surgical service. Am J Infect Control 2001;29:289–294
130. Zachary KC, Bayne PS, Morrison VJ, Ford DS, Silver LC, 143. Rubin HR, Pronovost P, Diette GB. The advantages and
Hooper DC. Contamination of gowns, gloves, and stetho- disadvantages of process-based measures of health care
scopes with vancomycin-resistant enterococci. Infect Control quality. Int J Qual Health Care 2001;13:469–474
Hosp Epidemiol 2001;22:560–564 144. Evans DB, Edejer TT, Lauer J, Frenk J, Murray CJ.
131. Boyce JM, Potter-Bynoe G, Chenevert C, King T. Measuring quality: from the system to the provider. Int J
Environmental contamination due to methicillin-resistant Qual Health Care 2001;13:439–446
Staphylococcus aureus: possible infection control implications. 145. Kritchevsky SB, Braun BI, Gross PA, Newcomb CS,
Infect Control Hosp Epidemiol 1997;18:622–627 Kelleher CA, Simmons BP. Definition and adjustment of
132. Boyce JM, Mermel LA, Zervos MJ, et al. Controlling Cesarean section rates and assessments of hospital perfor-
vancomycin-resistant enterococci. Infect Control Hosp Epi- mance. Int J Qual Health Care 1999;11:283–291
demiol 1995;16:634–637 146. Rosenthal VD, Guzman S, Safdar N. Reduction in
133. Srinivasan A, Song X, Ross T, Merz W, Brower R, Perl TM. nosocomial infection with improved hand hygiene in
A prospective study to determine whether cover gowns in intensive care units of a tertiary care hospital in Argentina.
addition to gloves decrease nosocomial transmission of Am J Infect Control 2005;33:392–397
vancomycin-resistant enterococci in an intensive care unit. 147. Rosenthal VD, Guzman S, Safdar N. Effect of education and
Infect Control Hosp Epidemiol 2002;23:424–428 performance feedback on rates of catheter-associated urinary
134. Pelke S, Ching D, Easa D, Melish ME. Gowning does not tract infection in intensive care units in Argentina. Infect
affect colonization or infection rates in a neonatal intensive Control Hosp Epidemiol 2004;25:47–50

Non-Pulmonary Critical Care

Uploaded by

Copyright:

Available Formats

Non-Pulmonary Critical Care

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Non-Pulmonary Critical Care

Uploaded by

Copyright:

Available Formats

PREFACE

Non-pulmonary Critical Care: Managing Multisystem Critical Illness

KEYWORDS: Acute quadriplegic myopathy, ARDS: acute respiratory distress

Table 1 Neurological Complications Encountered in 217 Table 3 Etiologies of Metabolic Encephalopathy in a

Table 2 Neurological Complication Rates by Primary Medical Intensive Care Unit

Seizures in the Intensive Care Unit

Delirium remains an underrecognized, but highly prevalent, form of organ

KEYWORDS: Delirium, aged, lorazepam, cognitive impairment, mechanical ventilation,

Table 1 Risk Factors for Delirium

Cognitive impairment Sepsis Sedative medications

Failure of Extubation to do so frequently. However, for them to take their

Figure 3 Proposed continuum of acute brain dysfunction in the ICU.

KEYWORDS: Arrhythmia, ICU, ventricular tachycardia, AV nodal reentrant tachycardia,

EVALUATION OF TACHYARRHYTHMIAS Adenosine is given as a rapid intravenous (IV) bolus

Table 1 Responses to Vagal Maneuvers or Adenosine

Sinus tachycardia Gradual slowing with resumption of the tachycardia

Table 2 Intravenous Medications for Heart Rate Control in Atrial Fibrillation

NA, not applicable.

KEYWORDS: Shock, cardiovascular protocols, cardiovascular techniques; central

S hock is a common cause of admission in any OVERVIEW, DEFINITIONS, AND

KEYWORDS: Liver failure, cerebral edema, portal hypertension, management

Table 1 Etiology-Targeted Therapy and an accurate history cannot be overemphasized be-

Table 3 Acute Liver Failure: General Management Guidelines

Thorough history Interview family members

Imaging US with Doppler Head CT for neurological

Directed therapy where indicated Drugs, cooling for

seven patients with ALF given propofol at 50 mg/kg/

PREVENTION AND MANAGEMENT

Figure 3 Proposed algorithm for the use of an intracranial pressure monitor.

Prevention and Treatment of Increased THIOPENTAL SODIUM

Table 4 Child-Pugh Score

1 <2 > 3.5 < 1.3 Absent 0

Figure 4 Proposed algorithm for the management of acute variceal bleeding.

Transjugular Intrahepatic Portosystemic

ASCITES Treatment of Hepatorenal Syndrome

KEYWORDS: Acute renal failure, renal replacement therapy

Table 1 Laboratory Findings in the Syndromes of Acute Renal Failure

Critical illness evoked by trauma, extensive surgery, or severe medical illnesses is

KEYWORDS: Stress hyperglycemia, adrenal insufﬁciency, ICU, critically ill, therapy,

CAUSES OF ADRENAL INSUFFICIENCY IN

Hematologic disorders are frequently encountered in the intensive care unit.

KEYWORDS: Thrombocytopenia, coagulopathy, heparin-induced thrombocytopenia,

Table 1 Comparison of Argatroban and Lepirudin for Heparin-Induced Thrombocytopenia

Approved indication Prophylaxis or treatment of HIT; Treatment of HIT and associated

ensues, resulting in microvascular thrombosis, impaired Table 2 Laboratory Markers in Disseminated

Two parallel randomized, placebo-controlled, Coagulopathy is a common cause of morbidity

Tens of thousands of patients undergo hematopoietic stem cell transplantation

KEYWORDS: Aspergillosis, bone marrow transplantation, cytomegalovirus infection,

Tens of thousands of patients undergo hemato- 1

CLINICAL COURSE AND PROGNOSIS

TREATMENT AND PROGNOSIS CONCLUSIONS

Nosocomial infections continue to be signiﬁcant causes of morbidity, mortality,

KEYWORDS: Nosocomial infections, bloodstream infections, ventilator-associated

Table 1 Preventing Infectious Complications Additional technology-based or invasive ap-

critically ill or post-operative patients requiring adequate Table 2 Prevention of Ventilator-Associated

throughout patient care areas. The project was supported Gowns

You might also like