Sie sind auf Seite 1von 16

Intensive Care Med (2007) 33:1876–1891

DOI 10.1007/s00134-007-0772-2 SYSTEMATIC REVIEW

Robert D. Stevens
David W. Dowdy
Neuromuscular dysfunction acquired in critical
Robert K. Michaels illness: a systematic review
Pedro A. Mendez-Tellez
Peter J. Pronovost
Dale M. Needham

P. J. Pronovost prolonged mechanical ventilation.


Received: 7 March 2007 Johns Hopkins Bloomberg School of Public
Accepted: 15 June 2007 Diagnostic criteria for CINMA were
Published online: 17 July 2007 Health, Department of Health Policy and
Management, not uniform, and few reports un-
© Springer-Verlag 2007 equivocally differentiated between
Baltimore MD, USA
polyneuropathy, myopathy, and mixed
D. M. Needham types of CINMA. The risk of CINMA
Johns Hopkins University School of was associated with hyperglycemia
Medicine, Department of
Pulmonary/Critical Care Medicine,
(and inversely associated with tight
R. D. Stevens (u) · R. K. Michaels ·
P. A. Mendez-Tellez · P. J. Pronovost Baltimore MD, USA glycemic control), the systemic
Johns Hopkins University School of inflammatory response syndrome,
Medicine, Department of sepsis, multiple organ dysfunction,
Anesthesiology/Critical Care Medicine, renal replacement therapy, and cat-
600 N Wolfe St, Meyer 8-140, Baltimore Abstract Objective: To determine echolamine administration. Across
21287, MD, USA the prevalence, risk factors, and studies, there was no consistent rela-
e-mail: rstevens@jhmi.edu outcomes of critical illness neuro- tionship between CINMA and patient
Tel.: +1-410-9557481
Fax: +1-410-6147903
muscular abnormalities (CINMA). age, gender, severity of illness, or use
Design: Systematic review. Data of glucocorticoids, neuromuscular
R. D. Stevens sources and study selection: MED- blockers, aminoglycosides, or mida-
Johns Hopkins University School of LINE, EMBASE, CINAHL, and the zolam. Unadjusted mortality was not
Medicine, Department of Neurology, Cochrane Library were searched for increased in the majority of patients
Baltimore MD, USA
reports on adult ICU patients who with CINMA, but mechanical ventila-
R. D. Stevens were evaluated for CINMA clinically tion and ICU and hospital stay were
Johns Hopkins University School of and electrophysiologically. Studies prolonged. Conclusions: The risk of
Medicine, Department of Neurosurgery, were included if they contained suffi- CINMA is nearly 50% in ICU patients
Baltimore MD, USA cient data to quantify the association with sepsis, multi-organ failure, or
D. W. Dowdy between CINMA and relevant ex- protracted mechanical ventilation.
Johns Hopkins Bloomberg School of Public posures and/or outcome variables. The association of CINMA with
Health, Department of Epidemiology, Measurements and results: CINMA frequently cited CINMA risk fac-
Baltimore MD, USA was diagnosed in 655 of 1421 [46% tors (glucocorticoids, neuromuscular
P. J. Pronovost
(95% confidence interval 43–49%)] blockers) and with short-term survival
Johns Hopkins University School of adult ICU patients enrolled in 24 is uncertain. Available data indi-
Medicine, Department of Surgery, studies, all with inclusion criteria cate glycemic control as a potential
Baltimore MD, USA of sepsis, multi-organ failure, or strategy to decrease CINMA risk.
1877

Introduction Methods
Many patients admitted to the intensive care unit (ICU) Data sources and searches
develop a syndrome of neuromuscular dysfunction
characterized by generalized muscle weakness and an Relevant publications were identified by searching Med-
inability to separate from mechanical ventilation [1]. line (1966–2006), EMBASE (1974–2006), CINAHL
Since this syndrome occurs in the absence of preex- (1982–2006) and the Cochrane Library (2006, Issue 2) on
isting neuromuscular disease, it is believed to reflect 29 April 2006, using the following terms: neuromuscular
illnesses or treatments occurring in the ICU. Early re- diseases, muscular diseases, paresis, polyneuropathy,
ports described two categories of acute neuromuscular myopathy, neuromyopathy, critical illness, critical care,
dysfunction, a polyneuropathy found in patients with critically ill, intensive care, ICU, sepsis, septic, systemic
sepsis and multiple organ failure [2, 3] and a myopathy inflammatory response syndrome, acute lung injury,
in patients with acute respiratory failure who received acute/adult respiratory distress syndrome, ARDS, mul-
glucocorticoids and/or neuromuscular blocking drugs [4, tiple organ dysfunction/failure, artificial ventilation,
5]. More recent work has delineated subtypes of critical mechanical ventilation. Terms were mapped to the ap-
illness polyneuropathy and myopathy, and has emphasized propriate “Medical Subject Heading” (MeSH) term and
that peripheral nerve and muscle involvement frequently “title/abstract” (tiab) text word in Medline, and to the
coexist [6–8]. The spectrum of neuromuscular disorders EMTREE subject headings in EMBASE, and “exploded”.
acquired in the ICU, which also includes neuromuscular Citations were not excluded on the basis of language. In
junction dysfunction, has been collectively referred to as addition to the electronic search, references from selected
“critical illness neuromyopathy” [9], “ICU-acquired pare- reports and review articles, as well as personal files,
sis” [7], “Critical illness myopathy and/or neuropathy” were hand searched. Only full-length reports published
(CRIMYNE) [10], or “critical illness neuromuscular in peer-reviewed journals were included in this review. If
abnormalities” (CINMA) [11]. data were missing or unclear in published reports, authors
The rising incidence and societal burden of critical ill- were contacted for additional information.
nesses such as sepsis and the acute respiratory distress syn-
drome [12–14], coupled with declining case fatality rates
and an aging population [14–16], suggest that the number Study selection
of patients with CINMA and its sequelae may be substan-
tial and likely to grow [17]. Many survivors of critical ill- We selected articles meeting the following inclusion
ness present with chronic morbidity linked to central and criteria. First, studies enrolled adults (> 16 years old)
peripheral neurological dysfunction [18–20], resulting in admitted to the ICU who were evaluated for CINMA
diminished functional status and quality of life months and using either diagnostic tests (nerve conduction velocities,
years after the acute period [21]. needle electromyography, direct muscle stimulation,
Knowledge of CINMA has increased through stud- histopathology of muscle or nerve tissue) or a combination
ies exploring the relationship between CINMA and of diagnostic tests and clinical findings (muscle weakness,
selected risk factors, as well as the impact of CINMA decreased or absent deep tendon reflexes, and/or failure
on short-term outcomes [7, 8, 22–25]. CINMA has to liberate from mechanical ventilation), with diagnostic
been incorporated as an outcome measure in recent tests obtained in at least all patients who had positive
large-scale randomized trials of ICU patients [26–28]. clinical findings. Second, studies compared patients with
Nevertheless, considerable uncertainty persists regarding CINMA to a group of ICU patients without clinical
the epidemiology, pathogenesis, and natural history of weakness and/or with negative diagnostic tests, such that
CINMA. We undertook a systematic review to syn- a quantitative estimate of the association between CINMA
thesize knowledge from published studies in this field and at least one relevant exposure and/or outcome variable
and to identify important areas for future research. was possible. Relevant exposure variables were patient
Specifically, our aims were threefold: first, to estimate demographics, admission diagnosis, severity of illness,
the prevalence of CINMA and its subsets using ex- physiological and metabolic status, organ dysfunction
plicit diagnostic criteria; second, to identify variables or failure, and the administration of pharmacological
that are associated with an increased (or decreased) agents. Relevant outcomes were the duration of mech-
risk of CINMA; and third, to evaluate the relationship anical ventilation, ICU and hospital mortality and length
between CINMA and patient outcomes, including mor- of stay, and any reporting of long-term outcomes such
tality, duration of mechanical ventilation, and length of as muscle weakness or functional status. Third, studies
hospitalization. contained original data with a sample size of ≥ 10 patients.
1878

There are no universally accepted or consensus-derived with decreased amplitudes of compound muscle action
definitions of CINMA and its subtypes. For the purposes potentials or sensory nerve action potentials. For critical
of this review, we accepted the classification used by illness myopathy (CIM), minimum criteria were the
study investigators provided that they explicitly met the presence of weakness acquired after admission to the
following minimum criteria. For critical illness polyneu- ICU and at least one of the following: demonstration
ropathy (CIP), minimum criteria were the presence of of short, low-amplitude motor unit potentials (requires
weakness acquired after admission to the ICU associated patient cooperation); decreased direct muscle stimula-

Fig. 1 Flow diagram of study se-


lection process
1879

tion amplitude or prolonged direct muscle stimulation (0–2), and the description of dropouts and withdrawals
latency; or muscle histopathology consistent with an (0–1).
acute myopathy. Mixed CIP/CIM was defined as a com-
bination of features consistent with both CIP and CIM.
We use the term “unclassified CINMA” for patients in Data synthesis and analysis
studies which did not fully differentiate between CINMA
subtypes. Since an explicit differentiation between CIP, CIM, and
A diagram of the study selection process is given in mixed CINMA was not available in most studies, these
Fig. 1. Using the study inclusion criteria above, article subtypes were not analyzed separately. In most of the
titles were screened for potentially relevant papers; these included studies, comparisons of exposure and outcome
papers were reviewed in abstract form, and relevant variables between patients with and without CINMA were
abstracts were chosen for full-text analysis. Each of made without adjustment for potential confounders. For
these steps was performed independently by two authors univariable comparisons, we recorded p-values provided
(R. D. S., D. W. D.); any title or abstract selected by either in the articles, and if these were not reported we calculated
author as potentially relevant was included in the next them from the original data; for multivariable compar-
round of review. Full-text review and final selection of isons, we recorded odds ratios (OR) or relative risks (RR)
articles was performed independently by two authors along with the 95% confidence interval (95% CI), and if
(D. W. D., R. K. M.), with discrepancies adjudicated by these data were not given, we calculated them using the
a third author (R. D. S.). Agreement between reviewers original data. In light of inter-study heterogeneity in both
was estimated using the kappa statistic. the number and type of covariates included in multivari-
able analyses, quantitative synthesis (i. e. meta-analysis)
of study results was not undertaken.
Data extraction and quality assessment

Using a standardized data collection form, two authors


(D. W. D., R. K. M.) independently abstracted data on Results
study design and quality, methods for diagnosing CINMA,
and patient demographic characteristics. The following Study search and selection
exposure variables were systematically recorded (when
available): severity of illness scores; organ dysfunction The initial search retrieved 9,299 citations. Sequential
scores; presence of the systemic inflammatory response review of titles, abstracts, and full-length articles ulti-
syndrome (SIRS), sepsis, or acute respiratory distress mately yielded 29 reports on 24 unique patient populations
syndrome; renal replacement therapy, parenteral nutrition (Fig. 1). Hereafter, unless otherwise noted, our findings
and hyperglycemia; glycemic control; and administration refer to these 24 patient populations. The included studies
of glucocorticoids, neuromuscular blockers, aminogly- had a mean sample size of 59 (range 10–405), and totaled
cosides, vasopressors or inotropic agents, metronidazole, 1,421 patients. Kappa statistics for inter-reviewer selection
furosemide, and midazolam. We also abstracted the of potentially relevant titles, abstracts, and full-text articles
following outcomes (when available): death in the ICU were 0.67, 0.72, and 0.79, respectively, demonstrating
and hospital, duration of mechanical ventilation, length of substantial agreement [33].
stay in the ICU and hospital, and long-term outcomes.
Methodological quality of individual observational
studies was assessed with the Newcastle–Ottawa Scale Study characteristics and quality
(NOS) [29], a validated instrument specifically designed to
evaluate the quality of observational studies in systematic Of the 24 studies, 19 were prospective cohorts [7, 8, 22,
reviews and meta-analyses [30, 31]. The NOS evaluates 24, 25, 34–46], 2 were retrospective cohorts [47, 48], 2
three domains of study methodology: the selection of were case–control studies [49, 50], and 1 was a random-
study groups (score range 0–4), the comparability of ized controlled trial [26] (Table 1). The criteria for enroll-
groups (score range 0–1), and the quality of ascertainment ment of ICU patients included mechanical ventilation in
of either the exposures (for case–control studies) or of 17 reports [median (range) duration of mechanical venti-
the outcomes of interest (for cohort studies) (score range lation 7 (2–10); in 14 reports, mechanical ventilation was
0–3). The composite NOS score ranges from 0 to 8, ≥ 7 days], sepsis or septic shock in 5 reports, and multi-
with a NOS > 5 indicating an acceptable methodological ple organ failure in 5. Patient exclusion criteria were stated
design. For the one randomized controlled trial included in in 20 studies and included presence or risk of prior neuro-
this review, methodological quality was assessed using the muscular disease in 19 reports and moribund status in 1.
Jadad scale [32]. This is a validated 3-item, 5-point scor- Patient follow-up extended beyond hospital discharge in 3
ing system evaluating trial randomization (0–2), blinding cohorts [7, 34, 43].
1880

Table 1 Study characteristics

Reference Study design No. of patients ICU Setting Patient enrollment criteria Follow up
enrolled
Amaya-Villar, 2005 [24] Prospective cohort 26 Mixed MV > 48 h, COPD, high-dose steroids In-hospital only
Bednarik, 2003 [8] Prospective cohort 46 Mixed (1 unit); > 2 organ failures with SOFA grade 3–5 28 days after enrollment
Neurological (1 unit)
Bednarik, 2005 [22] Prospective cohort 61 Mixed (1 unit); > 2 organ failures with SOFA grade 3–5 28 days after enrollment
Neurological (1 unit)
Bercker, 2005 [48] Retrospective cohort 45 Mixed ARDS and MV 29 days after enrollment
Berek, 1996 [34] Prospective cohort 22 Not reported SIRS or sepsis and MOF 60–90 days after enrollment
Campellone, 1998 [35] Prospective cohort 77 Surgical MV > 7 days or hospital LOS NR
> 14 days following OLT
Coakley, 1998 [36] Prospective cohort 44 Mixed MV and ICULOS > 7 days In-hospital only
De Jonghe, 2002 [7] Prospective cohort 95 Medical (3 units); MV > 7 days and evidence of wakefulness 45 days after enrollment
Surgical (2 units)
De Letter, 2001 [52] Prospective cohort 98 NR MV > 3 days 30 days after intubation
Druschky, 2001 [38] Prospective cohort 28 Neurological MV > 4 days 14 days of MV
Garnacho-Montero, 2001 [25] Prospective cohort 73 Mixed MV > 10 days, sepsis, MOF In-hospital only
Garnacho-Montero, 2005 [51] Prospective cohort 64 Mixed MV > 7 days, severe sepsis or septic shock In-hospital only
Hund, 1997 [50] Prospective cohort 28 Surgical “Prolonged” MV, sepsis ICU only
Kupfer, 1992 [40] Prospective cohort 10 Medical MV, vecuronium infusion ICU only
Lefaucheur, 2006 [41] Prospective cohort 30 Medical MV > 7 days, diffuse weakness NR
Leijten, 1995 [43] Prospective cohort 50 Mixed MV > 7 days, age < 75 12 months
Leijten, 1996 [42] Prospective cohort 38 Mixed MV > 7 days, age < 76 12 months
Mohr, 1997 [44] Prospective cohort 33 Mixed MOF (Goris score > 5 for > 5 days) In-hospital only
Rudis, 1996 [49] Case–control study 20 Mixed Cases: weakness after NMB In-hospital only
Tepper, 2000 [45] Prospective cohort 22 Mixed Septic shock ICU only
Thiele, 1997 [50] Case–control study 44 Surgical MV > 3 days, cardiac surgery ICU only
Thiele, 2000 [46] Prospective cohort 19 Surgical MV > 3 days, cardiac surgery ICU only
Van den Berghe, 2005 [54] Randomized 405 Mixed (mostly MV, EMG in subset with ICU LOS In-hospital only
controlled trial cardiac surgical) ICU LOS > or = 7 days
Witt, 1991 [47] Retrospective cohort 43 Mixed ICU LOS > 5 days, sepsis 6 months

MV, mechanical ventilation; COPD, chronic obstructive pulmonary disease; ARDS, acute respiratory distress syndrome; SIRS, systemic inflammatory response syndrome; MOF,
multiple organ failure; NR, not reported; LOS, length of stay; OLT, orthotopic liver transplantation; Goris score, see [66]; NMB, neuromuscular blockers; EMG, electromyogram
1881

Table 2 Quality of CINMA studies

Newcastle-Ottawa Scale [27]


Selection (0–4) Comparability (0–1) Outcome/Exposure (0–3) Total
Amaya-Villar, 2005 [24] 3 0 2 5
Bednarik, 2003 [8] 3 0 1 4
Bednarik, 2005 [22] 3 1 2 6
Bercker, 2005 [48] 4 0 2 6
Berek, 1996 [34] 3 0 2 5
Campellone, 1998 [35] 3 0 1 4
Coakley, 1998 [36] 3 0 2 5
De Jonghe, 2002 [7] 2 1 2 5
De Letter, 2000 [52] 3 1 2 6
Druschky, 2001 [38] 3 0 2 5
Garnacho-Montero, 2001 [25] 3 1 3 7
Garnacho-Montero, 2005 [51] 3 1 3 7
Hund, 1997 [50] 1 0 2 3
Kupfer, 1992 [40] 1 0 3 4
Lefaucheur, 2006 [41] 2 0 2 4
Leijten, 1995 [43] 3 0 3 6
Leijten, 1996 [42] 3 0 2 5
Mohr, 1997 [44] 1 0 1 2
Rudis, 1996 [49] 4 0 3 7
Tepper, 2000 [45] 3 0 1 4
Thiele, 1997 [50] 3 0 2 5
Thiele, 2000 [46] 2 0 2 4
Witt, 1991 [47] 1 0 1 2
Median (range) 3 (1–4) 0 (0–1) 2 (1–3) 5
Mean (SD) 2.6 (0.9) 0.2 (0.5) 2.0 (0.65) 4.80

Jadad scale [30] Randomization Blinding Follow up Total


Van den Berghe, 2001 [54] 2 0 1 3

For cohort studies “selection” refers to the representativeness of the exposed cohort (yes = 1, no = 0), selection of the non exposed cohort
(adequate = 1, not adequate = 0), ascertainment of exposure (clear = 1, unclear = 0), demonstration that outcome of interest was not present
at start of study (yes = 1, no = 0); “comparability” refers to adjustment for bias/confounding (yes = 1, no = 0); “outcome” refers to outcome
assessment, i.e. independent blind assessment (yes = 1, no = 0), appropriate duration of follow up (yes = 1, no = 0), adequacy of follow-up
(> 90% = 1, ≤ 90% = 0). For case control studies, “selection” refers to the case definition (adequate = 1, inadequate = 0), representative-
ness of the cases (yes = 1, no = 0), selection of controls (adequate = 1, inadequate = 0), definition of controls (adequate = 1, inadequate = 0);
“comparability” refers to adjustment for bias/confounding (yes = 1, no = 0); Exposure assessment refers to ascertainment of exposure (ad-
equate = 1, inadequate = 0); identical method of ascertainment for cases and controls (yes = 1, no = 0), non-response rate (same rate for both
groups = 1, other = 0). Maximum score = 8, minimum = 0. Adapted from [29]

The methodological quality of the studies was fair, with ease as a cause of critical illness [35, 46, 49, 50, 54], in-
7 studies achieving a NOS > 5 [22, 25, 37, 43, 48, 49, dicating possible enrollment of patients with non-CINMA
51] (Table 2). The quality of reports published after 2000 conditions. Six studies met minimum criteria to adequately
was not significantly higher than in the older studies [mean discriminate between cases of CIP, CIM, and CIP/CIM [8,
(SD) NOS 5.5 (1.0) versus 4.6 (1.3), p = 0.08]. Statistical 22, 34, 37, 41, 49]; these reports all contained either an-
comparisons in the majority of studies were made with- alysis of motor unit potentials, direct muscle stimulation,
out adjustment for potential confounders. Multivariable re- or muscle histopathology. In all other studies, data were
gression analysis was available in 8 reports (Table 8) [7, insufficient to make this distinction.
22–25, 34, 52, 53].

Prevalence of CINMA
Diagnosis of CINMA
Of the 1421 subjects included in this review, a total
Diagnosis of CINMA was confirmed with electrophysio- of 655 (46%; 95% CI 43–49%) were diagnosed with
logical tests in all studies, and clinical findings were docu- CINMA, with a median (range) prevalence of CINMA
mented in all but three reports [24, 25, 54] (Table 3). Five in the studies of 57% (9–87%) (Table 4). Of the 655
studies did not explicitly exclude patients with pre-existing patients diagnosed with CINMA, 51 (7.8%) had CIP,
neuromuscular disease or patients with neuromuscular dis- 49 (7.5%) had CIM, 44 (6.7%) had mixed CIM/CIP,
1882

Table 3 Diagnosis of CINMA

Exclusion
of prior NMD
Documentation
of weakness
Repetitive
stimulation
CMAP
SNAP
MUP
Fibrillation
potentials
DMS
No. of patients (%)
with muscle biopsy

Reference Timing of first CINMA evaluation


Amaya-Villar, 2005 [24] Median (range) 6 (2–12) days of MV + + – – – + – – 3 (12)
Bedoarik, 2003 [8] < 3 days after ICU admission + + + + + + + + 11 (24)
Bednarik, 2005 [22] < 3 days after ICU admission + + + + + + + + 11 (18)
Bercker, 2005 [48] NR + + – – – – – – –
Berek, 1996 [34] 14–28 days after ICU admission + + – + + + + – –
Campellone, 1998 [35] 2 weeks after liver transplantation – + + + + + + – 5 (7)
Coakley, 1998 [36] NR + + + + + – + – 24 (55)
De Jonghe, 2002 [7] Mean (SD) 12.4 (6.8) of MV + + + + + – + – 10 (11)
De Letter, 2000 [52] 4 days MV + + + + + + + – 32 (33)
Druschky, 2001 [38] 4 days of MV + + + + + – + – –
Garnacho-Montero, 2001 [25] 10 days of MV + – + + + – + – –
Garnacho-Montero, 2005 [51] Onset of weaning from MV + – + + + – + – –
Hund, 1997 [50] 5–7 days of MV + + – + – – + – 2 (7)
Kupfer, 1992 [40] > 24 h after discontinuation of vecuronium infusion + + + – + – – – –
Lefaucheur, 2006 [41] Mean (SD) 13.7 (7.8) days after awakening from coma + + + + + + + + –
Leijten, 1995 [43] 7–9 days of MV + + + + + – + – –
Leijten, 1996 [42] 7–9 days of MV + + + + + – + – –
Mohr, 1997 [44] 1–2 days before discharge from ICU + + – + + – + – –
Rudis, 1996 [49] Median 6.5 days after discontinuing NMB – + + + + + + – –
Tepper, 2000 [45] < 3 days of septic shock + + – + + – + – –
Thiele, 1997 [50] 5–7 days of MV – + – + + – + – –
Thiele, 2000 [46] 5–7 days of MV – + – + + – + – –
Van den Berghe, 2005 [54] 1 week after ICU admission – – – – – – + – –
Witt, 1991 [47] 20 days after ICU adnrission + + – + + – + – –

CINMA, critical illness neuromuscular abnormality; NMD, neuromuscular disease; CMAP, compound muscle action potential; SNAP, sensory nerve action potential; MUP, motor
unit potential; DMS, direct muscle stimulation; MV, mechanical ventilation; NMB, neuromuscular blockers; + = information available in report; – = information not available in
report
1883

Table 4 Prevalence of CINMA


Reference No. No. (%) No. (%) No. (%) No. (%) No. (%) No. (%) No. (%) with
of patients without with with with with mixed with NMJ unclassified
enrolled CINMA CINMA CIP CIM CIP/CIM defect CINMA
Amaya-Villar, 2005 [24] 26 17 (65%) 9 (35%) 9 (35%)
Bednarik, 2003 [8] 46 10 (43%) 26 (57%) 7 (15%) 8 (17%) 11 (24%)
Bednarik, 2005 [22] 61 26 (43%) 35 (57%) 12 (20%) 14 (23%) 9 (15%)
Bercker, 2005 [48] 45 18 (40%) 27 (60%) 27 (60%)
Berek, 1996 [34] 22 4 (18%) 18 (82%) 18 (82%)
Campellone, 1998 [35] 77 70 (91%) 7 (9%) 7 (9%)
Coakley, 1998 [36] 44 7 (16%) 37 (84%) 37 (84%)
De Jonghe, 2002 [7] 95 71 (75%) 24 (25%) 24 (25%)
De Letter, 2001 [52] 98 66 (67%) 32 (33%) 11 (11%) 12 (12%) 7 (7%) 2 (2%)
Druschky, 2001 [38] 28 12 (43%) 16 (57%) 16 (57%)
Garnacho-Montero, 2001 [25] 73 23 (32%) 50 (68%) 50 (68%)
Garnacho-Montero, 2005 [51] 64 30 (47%) 34 (53%) 34 (53%)
Hund, 1997 [50] 28 8 (29%) 20 (71%) 20 (71%)
Kupfer, 1992 [40] 10 3 (30%) 7 (70%) 2 (2%) 5 (50%)
Lefaucheur, 2006 [41] 30 4 (13%) 26 (87%) 1 (3%) 9 (30%) 16 (53%)
Leijten, 1995 [43] 50 21 (42%) 29 (58%) 29 (58%)
Leijten, 1996 [42] 38 20 (53%) 18 (47%) 18 (47%)
Mohr, 1997 [44] 33 26 (89%) 7 (21%) 7 (21%)
Rudis, 1996 [49] 20 10 (50%) 10 (50%) 2 (10%) 6 (30%) 1 (5%) 1 (5%)
Tepper, 2000 [45] 22 3 (14%) 19 (86%) 19 (86%)
Thiele, 1997 [50] 44 37 (84%) 7 (16%) 7 (16%)
Thiele, 2000 [46] 19 7 (37%) 12 (63%) 12 (63%)
Van den Berghe, 2005 [54] 405 250 (62%) 155 (38%) 155 (38%)
Witt, 1991 [47] 43 13 (30%) 30 (70%) 30 (70%)

CINMA, critical illness neuromuscular abnormality; CIP, critical illness polyneuropathy; CIM, critical illness myopathy; NMJ, neuromus-
cular junction

and 508 (77.6%) had unclassified CINMA. In the in one of these [7], renal replacement was associated with
six reports that differentiated between CIM, CIP and a 3.4-fold increase in the odds of developing CINMA
CIP/CIM [8, 22, 34, 41, 49, 52], the median (range) (95% CI 1.1–10.5), while the other study suggested
prevalence of these subsets within the study populations a reduction in risk [OR (95%CI) 0.27 (0.1–0.70)] [25].
were respectively 20% (0–30%), 13% (3–82%), and 9% In reports containing multivariable analyses, CINMA
(5–53%). was independently predicted by the following factors:
female gender [OR (95% CI) 4.6 (1.2–18.3)] [7]; renal
replacement therapy [OR (95% CI) 1.9 (1.0–3.8)] [26];
Exposure variables and CINMA APACHE III score [55] 30 days after institution of
mechanical ventilation [5.2 (2.8–9.8)] [52]; SOFA
Non-pharmacological exposures score [56] 1 week after ICU admission [RR (95% CI)
2.4 (1.02–5.53)] [22]; SIRS 1 week after admission [RR
Data on patient characteristics and commonly cited (95% CI) 3.74 (1.37–10.20)] [22]; SIRS 30 days after
CINMA risk factors are provided in Tables 5 and 6. Two institution of mechanical ventilation [OR (95% CI) 2.5
of 3 studies reported a univariable relationship between (1.2–4.8)] [52]; blood glucose levels [OR (95% CI)
SIRS and CINMA [22, 52], while sepsis was evaluated 1.24 (1.14–1.36) per 20 mg/dl increment] [53]; and total
in 12 reports, with a significant unadjusted association parenteral nutrition (OR [95% CI] 5.1 [1.1–23.0]) [25].
with CINMA noted in 6 [OR (95% CI) ranged from 2.4
(0.8–6.8) to 49 (4.7–519)] [24, 38, 46, 49, 50, 52]. Patients
with CINMA had significantly higher blood glucose levels Pharmacological exposures
in 5 of 6 studies [7, 26, 35, 48, 53], and in a subset of
405 patients enrolled in a randomized trial, a strategy of A number of studies investigated the relationship between
conventional (blood glucose target 180–200 mg/dl) versus pharmacological agents and CINMA. The majority of
intensive (blood glucose target 80–110 mg/dl) glycemic reports did not demonstrate a significant univariable
control resulted in an OR of CINMA of 2.6 (95% CI association between CINMA and glucocorticoids, amino-
1.6–4.2) [26]. Two of 3 studies found a univariable asso- glycosides, neuromuscular blocking agents, midazolam,
ciation between CINMA and renal replacement therapy; metronidazole, or furosemide. There was an unadjusted
1884

Table 5 Patient characteristics


Reference Age (years) Females (%) Severity of illness score Multiple organ failure score Sepsis (%)
CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA
Amaya-Villar, 2005 [24] 62 vs. 66, p = 0.2 NR AP2 23 vs. 14, p < 0.001 NR 67% vs. 6%, p = 0.002
Bednarik, 2003 [8] NR 54% vs. 25%, p = 0.09 NR NR NR
Bednarik, 2005 [22] NR NR NR SOFA score 8 vs. 6, p = 0.04* NR
Bercker, 2005 [48] 44 vs. 24, p = 0.01 NR SAPS2 35 vs. 32, SOFA score 5 vs. 5, p = 0.7 NR
p = 0.4
Berek, 1996 [34] 56 vs. 53, p = 0.44 17% vs. 50%, p = 0.41 SAPS1 15 vs. 15, p = NS NR NR
Campellone, 1998 [35] NR NR AP2 24 vs. 16, p = 0.005 NR NR
Coakley, 1998 [36] NR 46% vs. 42%, p = 0.88 NR NR NR
De Jonghe, 2002 [7] 68 vs. 59, p = 0.02** 50% vs. 20%, p < 0.004* SAPS2 53 vs. 47, p = 0.15 ODIN score 3 vs. 2.5, p = 0.08 38% vs. 20%, p = 0.08
No. of days with dysfunction in
≥ 2 organs 10 vs. 5, p < 0.001*
De Letter, 2001 [52] NR NR AP3 independently NR NR
predicted CINMA*
Druschky, 2001 [38] 70 vs. 64, p = 0.02 38% vs. 33%, p = 0.82 NR Goris score 1 vs. 0, p = 0.009 94% vs. 50%, p = 0.05
Garnacho-Montero, 2001 [25] 62 vs. 62, p = 0.92 NR AP2 17 vs. 18, p = 0.94 SOFA score 12 vs. 11, p = 0.29 100% vs. 100%, p = 0.33
Garnacho-Montero, 2005 [51] 61 vs. 62, p = 0.9 NR AP2 19 vs. 18, p = 0.3 SOFA score 7.2 vs. 7.2, p = 1 100% vs. 100%, p = NS
Hund, 1997 [50] NR NR NR NR NR
Kupfer, 1992 [40] 38 vs. 32, p = 0.51 29% vs. 33%, p = 0.88 NR NR NR
Lefaucheur, 2006 [41] 63 vs. 46, p = 0.143 33% vs. 44%, p = 0.81 NR NR 100% vs. 100%, p = NS
Leijten, 1995 [43] 59 vs. 54, p = 0.22 28% vs. 33%, p = 0.66 AP2 22 vs. 23. p = 0.74 22 vs. 11 patients 48% vs. 43%, p = 0.7
had MOF, p = 0.08
Leijten, 1996 [42] 58 vs. 55, p = 0.49 33 vs. 30%, p = 0.83 AP2 22 vs. 23, p = 0.66 MODS 5.3 vs. 3.6, p = 0.003 56% vs. 40%, p = 0.34
Mohr, 1997 [44] 48 vs. 52, p = 0.82 43% vs. 42%, p = 0.97 AP2 22 vs. 21, p = 0.34 NR 100% vs. 35%, p = 0.0027
Rudis, 1996 [49] 52 vs. 54, p = 0.97 70% vs. 70%, p = 1 AP2 61 vs. 62, p = 0.65 NR 40% vs. 0%, p = 0.001
Tepper, 2000 [45] NR NR NR NR NR
Thiele, 1997 [50] 68 vs. 74, p = NS NR NR NR 86% vs. 11%, p < 0.01
Thiele, 2000 [46] 66 vs. 68, p = NS 50% vs. 0%, p = 0.08 NR NR 58% vs. 0%, p < 0.05
Van den Borghe, 2005 [54] NR NR NR NR NR
Witt, 1991 [47] NR NR NR NR NR

CINMA, critical illness neuromuscular abnormality; NR, not reported; AP2, Acute Physiology and Chronic Health Evaluation II score; SOFA, sequential organ function assess-
ment; SAPS2, simplified acute physiology score 2; ODIN, organ dysfunction and/or infection score; AP3, Acute Physiology and Chronic Health Evaluation III score; MODS,
multiple organ score; Goris score, see ref. [66]; MOF, multiple organ failure; * Significant association with CINMA after multivariable adjustment; ** Non-significant association
with CINMA after multivariable adjustment
Table 6 CINMA risk factors
Reference Serum glucose or hyperglycemia Glucocorticoids Neuromuscular blockers Aminoglycosides
(mg/dL or % patients) (MP total dose or % patients) (Total dose or % patients) (% patients or dose)
CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA
Amaya-Villar, 2005 [24] NR 1649 vs. 979 mg, p = 0.05 44% vs. 18%, p = 0.19 22% vs. 0%, p = 0.1
Vecuronium 13 vs. 11 mg, p = 0.19
Bednarik, 2003 [8] NR NR NR NR
Bednarik, 2005 [22] NR NS association with CINMA NS association with CINMA NR
Bercker, 2005 [48] Peak, 166 vs. 144, p < 0.001 74% vs. 66%, p = 0.7 NR NR
Berek, 1996 [34] NR NR NR
Campellone, 1998 [35] 312 vs. 221, p = 0.007 2438 vs. 1674 mg, p = 0.02 NR NR
Coakley, 1998 [36] NR 24% vs. 29%, p = 1 62% vs. 71%, p = 1 NR
De Jonghe, 2002 [7] Peak, 360 vs. 259, p = 0.001** 54% vs. 18%, p = 0.001* 63% vs. 41%, p = 0.07** 67% vs. 42%, p = 0.04**
290 vs. 460 mg, p = 0.32 Vecuronium 13 vs. 16 mg, p = 0.11
De Letter, 2001 [52] NR NS association with CINMA** NS association with CINMA** NS association with CINMA**
Druschky, 2001 [38] NR Daily dose, 0 vs. 32 mg, p = 0.47 NR NR
Garnacho-Montero, 2001 [25] 30% vs. 30%, p = NS 14% vs. 17%, p = 0.7 18% vs. 4%, p = 0.16* 42% vs. 44%, p = 0.91
Garnacho-Montero, 2005 [51] NR NR 29% vs. 10%, p = 0.06 NR
Hund, 1997 [50] NR NR NR NR
Kupfer, 1992 [40] NR NR Vecuronium 1352 vs. 528 mg, p = 0.05 NR
Lefaucheur, 2006 [41] NR 58% vs. 25%, p = 0.31 19% vs. 25%, p = 0.78 NR
Leijten, 1995 [43] NR NR NR NR
Leijten, 1996 [42] NR NR NR NR
Mohr, 1997 [44] NR NR 100% vs. 77%, p = 0.26 100% vs. 35%, p = 0.6
Rudis, 1996 [49] NR 70% vs. 50%, p = 0.65 70% vs. 90%, p = NS 70% vs. 80%, p = 1
Vecuronium or pancuronium
1080 vs. 1165 mg, p = 0.76
Tepper, 2000 [45] NR NR NR NR
Thiele, 1997 [50] NR 100% vs. 11%, p < 0.01 86% vs. 30%, p = NS 14% vs. 5%, p = NS
16 vs. 20 mg/kg, p < 0.01 Vecuronium or pancuronium Total netilmycin
0.37 vs. 0.52 mg/kg, p = NS 23 vs. 16 mg/kg
Thiele, 2000 [46] 221 vs. 215, p = NS NR NR NR
Van den Berghe, 2005 [54] OR (95% CI) of CINMA NR NR NR
1.26 (1.09; 1.46) per mmol/l
of AM blood glucose level*
Witt, 1991 [47] NR NR NR NR

CINMA, critical illness neuromuscular abnormality; MP, methylprednisolone; NR, not reported; NS, not significant or non significant; HT, intensive insulin therapy; * Significant
association with CINMA after multivariable adjustment; ** Non-significant association with CINMA after multivariable adjustment
1885
1886

Table 7 CINMA outcomes


References ICU mortality (%) Hospital mortality (%) ICU length of stay (days) Hospital length of stay (days) Duration of MV (days)
CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA CINMA vs. no CINMA
Amaya-Villar, 2005 [24] 33% vs. 18%, p = 0.6 56% vs. 24%, p = 0.2 23 vs. 11, p < 0.003 33 vs. 21, p < 0.03 15 vs. 6, p = 0.006
Bednarik, 2003 [8] NR NR NR NR NR
Bednarik, 2005 [22] NR NR NR NR 16 vs. 3, p < 0.001
Bercker, 2005 [48] NR NR 38 vs. 26, p = 0.001 NR 20 vs. 13, p = 0.018
Berek, 1996 [34] 39% vs. 0%, p = 0.26 NR NR NR NR
Campellone, 1998 [35] NR 29% vs. 0%, p = 0.007 49 vs. 14, p = 0.002 NR NR
Coakley, 1998 [36] NR 24% vs. 14%, p = 1 NR NR NR
De Jonghe, 2002 [7] 17% vs. 6%, p = 0.2 NR NR NR 38 vs. 18, p = 0.003*
De Letter, 2001 [52] NR NR NR NR NR
Druschky, 2001 [38] NR 6% vs. 0%, p = 1 NR NR 15 vs. 8, p = 0.005
Garnacho-Montero, 2001 [25] 66% vs. 52%, p = 0.26 84% vs. 57%, p = 0.01 * 53 vs. 32, p = 0.01 93 vs. 41, p = 0.01 32 vs. 19, p = 0.002*
Garnacho-Montero, 2005 [51] 21% vs. 10%, p = 0.31 47% vs. 20%, p = 0.03 46 vs. 23, p < 0.0001 85 vs. 33, p < 0.001 34 vs. 14, p < 0.0001 *
Hund, 1997 [50] 50% vs. 100%, p = 0.025 NR NR NR NR
Kupfer, 1992 [40] NR NR NR NR NR
Lefaucheur, 2006 [41] NR NR NR NR NR
Leijten, 1995 [43] 48% vs. 19%, p = 0.03 NR NR NR 25 vs. 20, p = 0.03
Leijten, 1996 [42] 44% vs. 20%, p = 0.11 NR NR NR 11 vs. 8, p = 0.03
Mohr, 1997 [44] NR NR 27 vs. 21, p = 0.19 NR 23 vs. 17, p = 0.19
Rudis, 1996 [49] NS association with ClNMA 50% vs. 60%, p = 1 42 vs. 8, p = 0.001 68 vs. 11, p = 0.001 41 vs. 3, p = 0.001
Tepper, 2000 [45] NR NR NR NR NR
Thiele, 1997 [50] 57% vs. 8%, p < 0.05 NR 62 vs. 14, p < 0.01 NR 50 vs. 7, p < 0.01
Thiele, 2000 [46] 33% vs. 0%, p = 0.24 NR 40 vs. 20, p < 0.05 NR 32 vs. 12, p < 0.05
Siroen, 2005 [67] NS association with ClNMA NR NR NR NR
Witt, 1991 [47] NR NR NR NR NR

CINMA, critical illness neuromuscular abnormality; MV, mechanical ventilation; NR, not reported; * Significant association with CINMA after multivariable adjustment
1887

Table 8 Exposure and outcome variables in CINMA studies

Univariable analysis Multivariable analysis


Comparisons of patients No. of studies No. of studies No. of studies No. of studies
with and without CINMA with comparison with association with comparison with association
Exposures
Non-pharmacological
Blood glucose 6 5 2 1
Renal replacement therapy 3 2 1 1
SIRS 3 2 2 2
Sepsis 12 6 0 0
Multiple organ failure score 9 4 2 2
Severity of illness scores 12 3 1 1
Age 15 3 1 0
Gender 12 1 1 1
Total parenteral nutrition 1 0 1 1
Pharmacological
Catecholamines 3 3 0 0
Morphine 1 1 0 0
Glucocorticoids 12 4 2 1
Aminoglycosides 7 1 2 0
Neuromuscular blockers 12 1 3 1
Metronidazole 1 0 0 0
Furosemide 1 0 0 0
Midazolam 6 0 0 0
Outcomes
Duration of MV 13 12 3 3
Hospital length of stay 4 4 0 0
ICU length of stay 9 8 0 0
Death in ICU 12 3* 0 0
Death in hospital 7 3 1 1

CINMA, critical illness neuromuscular abnormality; MV, mechanical ventilation; SIRS, systemic inflammatory response syndrome;
* Two studies showing higher hospital mortality in CINMA patients; one study showing higher hospital mortality in patients without
CINMA

association between inotropic/vasopressor drug adminis- in patients with CINMA. CINMA independently pre-
tration and CINMA in the three cohorts in which it was dicted duration of mechanical ventilation in two reports
reported (Table 8) [26, 46, 50]. Multivariable analysis sug- [OR (95% CI) for prolonged mechanical ventilation 2.4
gested a relationship between CINMA and glucocorticoids (1.4–4.2) [23]; and mean increase in duration of mechan-
in one of two studies [OR (95% CI) 14.9 (3.2–69.8)] [7], ical ventilation, 11.6 days (95% CI 2.6–20.5)] [25]. In
and between CINMA and neuromuscular blockers in one one report, CINMA was the only independent predictor
out of three studies [OR (95% CI) 16.32 (1.34–199)] [25]. of failure to wean from mechanical ventilation [OR (95%
CI) 15.4 (4.6–52.3)] [51].
Patient outcomes after hospital discharge were reported
Outcomes and CINMA in three studies [7, 34, 43]. In one, clinically relevant re-
covery of motor function was noted in 15 of 16 CINMA
Univariable and multivariable relationships between patients evaluated at 9 months (94%) [7]. In another ICU
CINMA and selected outcomes are given in Tables 7 cohort followed up at 1 year, 4 of 11 (36%) CINMA pa-
and 8. Nine of the 12 studies which reported on un- tients had severe functional disability [43]. The third study
adjusted ICU mortality found no difference between evaluated 15 CINMA survivors at 3 months, finding clin-
patients with and without CINMA. Hospital mortality was ical and electrophysiological signs of polyneuropathy in
higher in CINMA patients in three of seven reports [24, respectively 7 (47%) and 11 (73%) patients [34].
25, 35]. Only one study evaluated hospital mortality in
a multivariable model, identifying CINMA as an inde-
pendent predictor of hospital death [OR (95% CI) 7.1
Discussion
(1.5–32.6)] [25].
The influence of CINMA on mechanical ventilation There are four major findings of this systematic review.
was assessed in 13 studies, which with one exception [44] First, evidence of neuromuscular dysfunction is found in
all showed increased duration of mechanical ventilation approximately 50% of adult ICU patients who receive
1888

prolonged mechanical ventilation, have sepsis or multiple populations, including patients not receiving mechanical
organ failure. Second, five of six reports found an as- ventilation or without multiple organ failure, in order to
sociation between CINMA and higher serum glucose delineate the scope of CINMA and achieve generalizable
levels, yet existing studies do not consistently support results.
several other generally accepted risk factors for CINMA A second implication of this review is the charac-
such as exposure to glucocorticoids or neuromuscular terization of risk factors for CINMA. Existing studies
blocking drugs. Third, although CINMA does not reliably suggest a positive association between blood glucose
predict ICU mortality in unadjusted models, it consis- levels and the incidence of CINMA, and two large ran-
tently and significantly increased duration of mechanical domized trials enrolling surgical [26] and medical [27]
ventilation and hospitalization, and it may be linked with ICU patients demonstrated that intensive insulin therapy
long-term neuromuscular weakness. Last, in the published substantially lowered the incidence of neuromuscular
studies, there is considerable heterogeneity in the way complications [54, 58]. These data support glycemic
CINMA is diagnosed, and CINMA subtypes are not well control as one potential strategy to decrease the risk of
differentiated. CINMA. On the other hand, the relationship between
In a prior systematic review, De Jonghe et al. syn- several other risk factors (e. g., neuromuscular blockers,
thesized data from eight prospective cohorts of CINMA glucocorticoids) and CINMA is unclear and needs to be
patients published between 1980 and 1997 [57]. Of the critically reevaluated; this was an unexpected finding, as
242 patients in these studies, 145 (60%) were diagnosed these drugs are widely viewed as having a direct or causal
with CINMA (the total number of patients included in relationship to CINMA. Since many of the available
this review was uncertain since, as the authors acknow- studies enrolled small numbers of patients and did not
ledge, there was possible duplication of patient samples properly adjust for covariates, more robust inferences
between at least two of the publications). The authors regarding CINMA risk factors and causation will require
recommended additional prospective studies to delin- larger patient cohorts incorporating predetermined multi-
eate CINMA risk factors and examine the contribution variable models. These studies will complement basic and
of CINMA to clinically relevant endpoints, including translational research into the etiology and pathogenesis
long-term outcomes. Our systematic review builds on of CINMA.
this earlier work and includes studies published up to Third, available data suggest there is a relationship
2006 with a total of 1421 patients, 655 (46%) of whom between CINMA and long-term patient outcomes [20].
were diagnosed with CINMA. Over half (13 of 24) of the Persistent weakness and residual motor and sensory
studies in our review were published after the study by neurological deficits are common findings in survivors
De Jonghe et al., with eight of these reports containing of critical illness, and electrophysiological testing can
statistical modeling to adjust for confounding of CINMA demonstrate residual neuromuscular dysfunction years
risk factors and outcomes. In addition to summarizing after the initial presentation [59]. In a seminal report on
results from a larger number of patients, our review has 109 acute respiratory distress syndrome survivors evalu-
several important new implications. ated at 1 year, patient-reported muscle weakness, rather
While the overall prevalence of CINMA in ICU than pulmonary dysfunction, was the most important fac-
patients remains unknown, we found that that this tor contributing to functional impairment (this study did
complication is remarkably common in patients with not include formal neuromuscular testing) [19]. A separate
prolonged mechanical ventilation, sepsis or multi-organ clinical and electrophysiological study of 22 patients who
failure. There are no known therapies for CINMA, yet the had a prolonged (> 28 days) critical illness and were
contribution of CINMA must be weighed when making assessed a median of 43 months after ICU discharge found
predictions about durations of mechanical ventilation that 13 (59%) had motor or sensory deficits on clinical
and hospitalization, the need for a tracheostomy, and examination and 21 (95%) had chronic partial denervation
anticipated long-term functional status [17]. CINMA on EMG [60]. Building on these observations, new cohort
should be entertained in all patients who fail to wean from studies are needed that are powered to quantify the impact
mechanical ventilation, even when other factors have been of CINMA on functional status, quality of life, and social
implicated in weaning failure. The relationship between and work-related reintegration in the months and years
CINMA and mechanical ventilation is complex, since following the acute illness [61].
mechanical ventilation may be a surrogate marker for Fourth, there is a need to apply standardized diagnostic
factors that are causally related to CINMA, while CINMA criteria of CINMA and its subtypes. Existing studies
in turn prolongs mechanical ventilation. More studies are heterogeneous in the use of diagnostic criteria. For
are needed to unravel these relationships, and to assess instance, in the largest patient sample within this re-
the relative contribution of CINMA and other factors, view [26, 62], there was no documented exclusion of
such as coexisting cardiac and pulmonary dysfunction, prior neuromuscular disease, no reporting of clinical
in patients who fail to wean from mechanical ventilation. findings, and the diagnosis of CIP was based exclusively
Future studies should also be directed to broader ICU on EMG results that are non-specific in differentiating
1889

CIP and CIM [1]. Uniform use of diagnostic criteria outcome characteristics specific to CINMA, and therefore
will enhance communication between clinicians and help we disqualified studies that did not independently confirm
investigators design studies to understand the epidemiol- clinical findings. Another limitation is that the majority
ogy, risk factors, and impact of specific CINMA subsets. of reported CINMA risk factors and outcome analyses
These criteria should integrate physical examination were univariable and prone to confounding due to the
findings with valid and reliable bedside neurophysio- observational study designs. An additional limitation
logical tests (e. g., nerve conduction velocities, needle is that it was not possible to standardize the way that
electromyography), recognizing that CINMA preva- exposure variables were defined or modeled in individual
lence is likely to be underestimated when only physical studies, likely contributing to heterogeneity in our results
findings are relied upon [8], while clinically relevant (for instance, several studies included in this review
disease may be overestimated with electrophysiolog- provided either no definition or non-validated definitions
ical tests [7]. Clinicians should also understand that of terms such as sepsis or multiple organ failure). Study
conventional electrophysiological techniques may not heterogeneity precluded any quantitative pooling of
reliably identify CINMA subtypes [8, 63], and in selected risk estimates into a formal meta-analysis. Finally, the
circumstances specific methods such as direct muscle quality of included studies was only fair, potentially
stimulation [41, 64, 65] might be helpful, for example in influencing the direction and magnitude of the observed
unconscious patients, who are unable to voluntarily effects.
contract muscles and in whom the contribution of CIM is In conclusion, this systematic review reveals that
likely to be overlooked [6]. CINMA is a common complication in critically ill patients
This review has several limitations. By excluding with mechanical ventilation, sepsis or multiple organ
reports that lacked electrophysiological or histopatho- failure. Existing studies do not reliably confirm certain
logical confirmation of physical examination findings, widely cited CINMA risk factors, while new data indicate
we may have omitted data on patients with clinically that intensive glycemic control might decrease the risk
significant CINMA. It has been argued that a perti- of CINMA. Large prospective studies are needed to
nent history and detailed neurological examination may substantiate these findings and to identify new factors
be sufficient to establish a diagnosis of CINMA, and which might contribute to CINMA and its long-term
that additional tests are of little practical value or may consequences, with the goal of preventing and treating
even overestimate clinically relevant CINMA; others these disorders in vulnerable populations.
have countered that electrophysiological confirmation is
essential to accurately identify CINMA [1]. While elec- Acknowledgements. The authors wish to thank Ahmet Hoke, MD,
trophysiological results may have little impact on clinical Department of Neurology, Johns Hopkins University, for his atten-
decision-making, our intent was to identify exposure and tive manuscript review and helpful comments.

References
1. Bolton CF (2005) Neuromuscular 7. De Jonghe B, Sharshar T, 11. Lorin S, Nierman DM (2002) Critical
manifestations of critical illness. Lefaucheur JP, Authier FJ, illness neuromuscular abnormalities.
Muscle Nerve 32:140–163 Durand-Zaleski I, Boussarsar M, Crit Care Clin 18:553–568
2. Bolton CF, Gilbert JJ, Hahn AF, Cerf C, Renaud E, Mesrati F, Carlet J, 12. Angus DC, Linde-Zwirble WT,
Sibbald WJ (1984) Polyneuropathy Raphael JC, Outin H, Bastuji-Garin S Lidicker J, Clermont G, Carcillo J,
in critically ill patients. J Neurol (2002) Paresis acquired in the intensive Pinsky MR (2001) Epidemiol-
Neurosurg Psychiatry 47:1223–1231 care unit: a prospective multicenter ogy of severe sepsis in the United
3. Bolton CF (1996) Sepsis and the study. JAMA 288:2859–2867 States: analysis of incidence, out-
systemic inflammatory response syn- 8. Bednarik J, Lukas Z, Vondracek P come, and associated costs of care.
drome: neuromuscular manifestations. (2003) Critical illness polyneuromy- Crit Care Med 29:1303–1310
Crit Care Med 24:1408–1416 opathy: the electrophysiological 13. Rubenfeld GD, Caldwell E, Peabody E,
4. MacFarlane IA, Rosenthal FD (1977) components of a complex entity. Weaver J, Martin DP, Neff M, Stern EJ,
Severe myopathy after status asthmati- Intensive Care Med 29:1505–1514 Hudson LD (2005) Incidence and
cus. Lancet 2:615 9. Lacomis D, Campellone JV (2002) outcomes of acute lung injury.
5. Lacomis D, Giuliani MJ, Van Cott A, Critical illness neuromyopathies. N Engl J Med 353:1685–1693
Kramer DJ (1996) Acute myopathy of Adv Neurol 88:325–335 14. Martin GS, Mannino DM, Eaton S,
intensive care: clinical, electromyo- 10. Latronico N, Bertolini G, Guarneri B, Moss M (2003) The epidemi-
graphic, and pathological aspects. Botteri M, Peli E, Andreoletti S, Bera P, ology of sepsis in the United
Ann Neurol 40:645–654 Luciani D, Nardella A, Vittorielli E, States from 1979 through 2000.
6. Latronico N, Fenzi F, Recupero D, Simini B, Candiani A (2007) Simplified N Engl J Med 348:1546–1554
Guarneri B, Tomelleri G, Tonin P, De electrophysiological evaluation of pe-
Maria G, Antonini L, Rizzuto N, Candi- ripheral nerves in critically ill patients:
ani A (1996) Critical illness myopathy the Italian multi-centre CRIMYNE
and neuropathy. Lancet 347:1579–1582 study. Crit Care 11:R11
1890

15. Abel SJ, Finney SJ, Brett SJ, Keogh BF, 25. Garnacho-Montero J, Madrazo- 33. Landis JR, Koch GG (1977)
Morgan CJ, Evans TW (1998) Reduced Osuna J, Garcia-Garmendia JL, The measurement of observer
mortality in association with the acute Ortiz-Leyba C, Jimenez-Jimenez FJ, agreement for categorical data.
respiratory distress syndrome (ARDS). Barrero-Almodovar A, Garnacho- Biometrics 33:159–174
Thorax 53:292–294 Montero MC, Moyano-Del-Estad MR 34. Berek K, Margreiter J, Willeit J,
16. Jardin F, Fellahi JL, Beauchet A, (2001) Critical illness polyneuropathy: Berek A, Schmutzhard E, Mutz NJ
Vieillard-Baron A, Loubieres Y, Page B risk factors and clinical consequences. (1996) Polyneuropathies in critically
(1999) Improved prognosis of acute A cohort study in septic patients. ill patients: a prospective evaluation.
respiratory distress syndrome 15 years Intensive Care Med 27:1288–1296 Intensive Care Med 22:849–855
on. Intensive Care Med 25:936–941 26. Van den Berghe G, Wouters P, Week- 35. Campellone JV, Lacomis D, Kramer DJ,
17. Hudson LD, Lee CM (2003) Neuro- ers F, Verwaest C, Bruyninckx F, Van Cott AC, Giuliani MJ (1998) Acute
muscular sequelae of critical illness. Schetz M, Vlasselaers D, Ferdinande P, myopathy after liver transplantation.
N Engl J Med 348:745–747 Lauwers P, Bouillon R (2001) Intensive Neurology 50:46–53
18. Hopkins RO, Weaver LK, insulin therapy in the critically ill 36. Coakley JH, Nagendran K,
Collingridge D, Parkinson RB, patients. N Engl J Med 345:1359–1367 Yarwood GD, Honavar M, Hinds CJ
Chan KJ, Orme JF Jr (2005) Two-year 27. Van den Berghe G, Wilmer A, Her- (1998) Patterns of neurophysiological
cognitive, emotional, and quality-of-life mans G, Meersseman W, Wouters PJ, abnormality in prolonged critical
outcomes in acute respiratory Milants I, Van Wijngaerden E, Bob- illness. Intensive Care Med 24:801–807
distress syndrome. Am J Respir baers H, Bouillon R (2006) Intensive 37. De Letter MA, van Doorn PA, Savelk-
Crit Care Med 171:340–347 insulin therapy in the medical ICU. oul HF, Laman JD, Schmitz PI, Op
19. Herridge MS, Cheung AM, Tansey CM, N Engl J Med 354:449–461 de Coul AA, Visser LH, Kros JM,
Matte-Martyn A, Diaz-Granados N, 28. Steinberg KP, Hudson LD, Good- Teepen JL, van der Meche FG (2000)
Al-Saidi F, Cooper AB, Guest CB, man RB, Hough CL, Lanken PN, Critical illness polyneuropathy and
Mazer CD, Mehta S, Stewart TE, Hyzy R, Thompson BT, An- myopathy (CIPNM): evidence for
Barr A, Cook D, Slutsky AS (2003) cukiewicz M (2006) Efficacy and local immune activation by cytokine-
One-year outcomes in survivors of the safety of corticosteroids for persistent expression in the muscle tissue.
acute respiratory distress syndrome. acute respiratory distress syndrome. J Neuroimmunol 106:206–213
N Engl J Med 348:683–693 N Engl J Med 354:1671–1684 38. Druschky A, Herkert M, Radespiel-
20. Latronico N, Shehu I, Seghelini E 29. Wells GA SB, O’Connell D, Peter- Troger M, Druschky K, Hund E,
(2005) Neuromuscular seque- son J, Welch V, Tugwell P (2000) The Becker CM, Hilz MJ, Erbguth F,
lae of critical illness. Curr Opin Newcastle–Ottawa Scale (NOS) for Neundorfer B (2001) Critical illness
Crit Care 11:381–390 assessing the quality of nonrandomised polyneuropathy: clinical findings and
21. Williams TA, Dobb GJ, Finn JC, studies in meta-analyses. Available cell culture assay of neurotoxicity as-
Webb SA (2005) Long-term sur- at: http://www.ohri.ca/programs/ sessed by a prospective study. Intensive
vival from intensive care: a review. clinical_epidemiology/oxford.htm Care Med 27:686–693
Intensive Care Med 31:1306–1315 (accessed 18 November 2006) 39. Hund E (1999) Myopathy in critically ill
22. Bednarik J, Vondracek P, Dusek L, 30. Jefferson T, Smith S, Demicheli V, patients. Crit Care Med 27:2544–2547
Moravcova E, Cundrle I (2005) Risk Harnden A, Rivetti A, Di Pietrantonj C 40. Kupfer Y, Namba T, Kaldawi E,
factors for critical illness polyneuromy- (2005) Assessment of the efficacy and Tessler S (1992) Prolonged weakness
opathy. J Neurol 252:343–351 effectiveness of influenza vaccines in after long-term infusion of vecuronium
23. De Jonghe B, Bastuji-Garin S, Shar- healthy children: systematic review. bromide. Ann Intern Med 117:484–486
shar T, Outin H, Brochard L (2004) Lancet 365:773–780 41. Lefaucheur JP, Nordine T, Rodriguez P,
Does ICU-acquired paresis lengthen 31. Tleyjeh IM, Tlaygeh HM, Hejal R, Brochard L (2006) Origin of ICU
weaning from mechanical ventilation? Montori VM, Baddour LM (2006) acquired paresis determined by direct
Intensive Care Med 30:1117–1121 The impact of penicillin resistance on muscle stimulation. J Neurol Neurosurg
24. Amaya-Villar R, Garnacho-Montero J, short-term mortality in hospitalized Psychiatry 77:500–506
Garcia-Garmendia JL, Madrazo- adults with pneumococcal pneumonia: 42. Leijten FS, De Weerd AW,
Osuna J, Garnacho-Montero MC, a systematic review and meta-analysis. Poortvliet DC, De Ridder VA, Ulrich C,
Luque R, Ortiz-Leyba C (2005) Clin Infect Dis 42:788–797 Harink-De Weerd JE (1996) Critical
Steroid-induced myopathy in patients 32. Jadad AR, Moore RA, Carroll D, Jenk- illness polyneuropathy in multiple
intubated due to exacerbation of inson C, Reynolds DJ, Gavaghan DJ, organ dysfunction syndrome and
chronic obstructive pulmonary disease. McQuay HJ (1996) Assessing the weaning from the ventilator. Intensive
Intensive Care Med 31:157–161 quality of reports of randomized Care Med 22:856–861
clinical trials: is blinding necessary?
Control Clin Trials 17:1–12
1891

43. Leijten FS, Harinck-de Weerd JE, 52. de Letter MA, Schmitz PI, Visser LH, 59. Zifko UA (2000) Long-term outcome of
Poortvliet DC, de Weerd AW (1995) Verheul FA, Schellens RL, Op de critical illness polyneuropathy. Muscle
The role of polyneuropathy in motor Coul DA, van der Meche FG Nerve [Suppl] 9:S49–S52
convalescence after prolonged mechan- (2001) Risk factors for the de- 60. Fletcher SN, Kennedy DD, Ghosh IR,
ical ventilation. JAMA 274:1221–1225 velopment of polyneuropathy and Misra VP, Kiff K, Coakley JH, Hinds CJ
44. Mohr M, Englisch L, Roth A, Bur- myopathy in critically ill patients. (2003) Persistent neuromuscular and
chardi H, Zielmann S (1997) Effects of Crit Care Med 29:2281–2286 neurophysiologic abnormalities in long-
early treatment with immunoglobulin 53. Van den Berghe G, Wouters PJ, term survivors of prolonged critical
on critical illness polyneuropathy Bouillon R, Weekers F, Verwaest C, illness. Crit Care Med 31:1012–1016
following multiple organ failure Schetz M, Vlasselaers D, Ferdinande P, 61. Angus DC, Carlet J (2003) Surviving
and gram-negative sepsis. Intensive Lauwers P (2003) Outcome benefit of intensive care: a report from the 2002
Care Med 23:1144–1149 intensive insulin therapy in the crit- Brussels Roundtable. Intensive Care
45. Tepper M, Rakic S, Haas JA, Woit- ically ill: insulin dose versus glycemic Med 29:368–377
tiez AJ (2000) Incidence and onset control. Crit Care Med 31:359–366 62. Van den Berghe G, Schoonheydt K,
of critical illness polyneuropa- 54. Van den Berghe G, Schoonheydt K, Becx P, Bruyninckx F, Wouters PJ
thy in patients with septic shock. Becx P, Bruyninckx F, Wouters PJ (2005) Insulin therapy protects
Neth J Med 56:211–214 (2005) Insulin therapy protects the central and peripheral nervous
46. Thiele RI, Jakob H, Hund E, Tantzky S, the central and peripheral nervous system of intensive care patients.
Keller S, Kamler M, Herold U, Hagl S system of intensive care patients. Neurology 64:1330–1331
(2000) Sepsis and catecholamine Neurology 64:1348–1353 63. Latronico N, Peli E, Botteri M (2005)
support are the major risk factors for 55. Knaus WA, Wagner DP, Draper EA, Critical illness myopathy and neuropa-
critical illness polyneuropathy after Zimmerman JE, Bergner M, Bastos PG, thy. Curr Opin Crit Care 11:126–132
open heart surgery’. Thorac Cardiovasc Sirio CA, Murphy DJ, Lotring T, Dami- 64. Rich MM, Bird SJ, Raps EC,
Surg 48:145–150 ano A, et al. (1991) The APACHE III McCluskey LF, Teener JW (1997)
47. Witt NJ, Zochodne DW, Bolton CF, prognostic system. Risk prediction of Direct muscle stimulation in
Grand’Maison F, Wells G, Young GB, hospital mortality for critically ill hos- acute quadriplegic myopathy.
Sibbald WJ (1991) Peripheral nerve pitalized adults. Chest 100:1619–1636 Muscle Nerve 20:665–673
function in sepsis and multiple organ 56. Vincent JL, Moreno R, Takala J, 65. Rich MM, Teener JW, Raps EC,
failure. Chest 99:176–184 Willatts S, De Mendonca A, Bruin- Schotland DL, Bird SJ (1996)
48. Bercker S, Weber-Carstens S, Deja M, ing H, Reinhart CK, Suter PM, Thijs LG Muscle is electrically inexcitable
Grimm C, Wolf S, Behse F, Busch T, (1996) The SOFA (Sepsis-related Organ in acute quadriplegic myopathy.
Falke KJ, Kaisers U (2005) Critical Failure Assessment) score to describe Neurology 46:731–736
illness polyneuropathy and myopathy in organ dysfunction/failure. On behalf 66. Goris RJ, te Boekhorst TP, Nuytinck JK,
patients with acute respiratory distress of the Working Group on Sepsis- Gimbrere JS (1985) Multiple-organ
syndrome. Crit Care Med 33:711–715 Related Problems of the European failure. Generalized autodestructive in-
49. Rudis MI, Guslits BJ, Peterson EL, Society of Intensive Care Medicine. flammation? Arch Surg 120:1109–1115
Hathaway SJ, Angus E, Beis S, Intensive Care Med 22:707–710 67. Siroen MP, van Leeuwen PA, Nij-
Zarowitz BJ (1996) Economic 57. De Jonghe B, Cook D, Sharshar T, veldt RJ, Teerlink T, Wouters PJ, Van
impact of prolonged motor weak- Lefaucheur JP, Carlet J, Outin H den Berghe G (2005) Modulation
ness complicating neuromuscular (1998) Acquired neuromuscular of asymmetric dimethylarginine in
blockade in the intensive care unit. disorders in critically ill patients: critically ill patients receiving intensive
Crit Care Med 24:1749–1756 a systematic review. Groupe de insulin treatment: a possible explanation
50. Thiele RI, Jakob H, Hund E, Genz- Reflexion et d’Etude sur les Neu- of reduced morbidity and mortality?
wuerker H, Herold U, Schweiger P, romyopathies En Reanimation. Crit Care Med 33:504–510
Hagl S (1997) Critical illness polyneu- Intensive Care Med 24:1242–1250
ropathy: a new iatrogenically induced 58. Hermans G, Wilmer A, Meersseman W,
syndrome after cardiac surgery? Milants I, Wouters PJ, Bobbaers H,
Eur J Cardiothorac Surg 12:826–835 Bruyninckx F, Van den Berghe G (2007)
51. Garnacho-Montero J, Amaya-Villar R, Impact of intensive insulin therapy on
Garcia-Garmendia JL, Madrazo- neuromuscular complications and
Osuna J, Ortiz-Leyba C (2005) Effect ventilator dependency in the medical
of critical illness polyneuropathy on the intensive care unit. Am J Respir Crit
withdrawal from mechanical ventilation Care Med 175:480–489
and the length of stay in septic patients.
Crit Care Med 33:349–354

Das könnte Ihnen auch gefallen