Yang et al.

Critical Care (2018) 22:187

https://doi.org/10.1186/s13054-018-2111-0

RESEARCH Open Access

Corticosteroid use and intensive care unit-

acquired weakness: a systematic review
and meta-analysis
Tao Yang1†, Zhiqiang Li2†, Li Jiang1 and Xiuming Xi1*

Abstract
Background: The association between corticosteroid use and intensive care unit (ICU)-acquired weakness remains
unclear. We evaluated the relationship between corticosteroid use and ICU-acquired weakness in critically ill adult
patients.
Methods: The PubMed, Embase, Web of Science, Cochrane Central Register of Controlled Trials, and Cumulative Index
of Nursing and Allied Health Literature databases were searched from database inception until October 10, 2017. Two
authors independently screened the titles/abstracts and reviewed full-text articles. Randomized controlled trials and
prospective cohort studies evaluating the association between corticosteroids and ICU-acquired weakness in adult ICU
patients were selected. Data extraction from the included studies was accomplished by two independent reviewers.
Meta-analysis was performed using Stata version 12.0. The results were analyzed using odds ratios (ORs) and 95%
confidence intervals (CIs). Data were pooled using a random effects model, and heterogeneity was evaluated using the
χ2 and I2 statistics. Publication bias was qualitatively analyzed with funnel plots, and quantitatively analyzed with Begg’s
test and Egger’s test.
Results: One randomized controlled trial and 17 prospective cohort studies were included in this review. After a
meta-analysis, the effect sizes of the included studies indicated a statistically significant association between
corticosteroid use and ICU-acquired weakness (OR 1.84; 95% CI 1.26–2.67; I2 = 67.2%). Subgroup analyses
suggested a significant association between corticosteroid use and studies limited to patients with clinical
weakness (OR 2.06; 95% CI 1.27–3.33; I2 = 60.6%), patients with mechanical ventilation (OR 2.00; 95% CI 1.23–3.27;
I2 = 66.0%), and a large sample size (OR 1.61; 95% CI 1.02–2.53; I2 = 74.9%), and not studies limited to patients with
abnormal electrophysiology (OR 1.65; 95% CI 0.92–2.95; I2 = 70.6%) or patients with sepsis (OR 1.96; 95% CI 0.61–6.30;
I2 = 80.8%); however, statistical heterogeneity was obvious. No significant publication biases were found in the review.
The overall quality of the evidence was high for the randomized controlled trial and very low for the included
prospective cohort studies.
Conclusions: The review suggested a significant association between corticosteroid use and ICU-acquired weakness.
Thus, exposure to corticosteroids should be limited, or the administration time should be shortened in clinical practice
to reduce the risk of ICU-acquired weakness.
Keywords: Intensive care unit, ICU-acquired weakness, Corticosteroids, Corticosteroid use, Corticosteroid therapy,
Systematic review

* Correspondence: xixiuming2937@163.com
†
Tao Yang and Zhiqiang Li contributed equally to this work.
1
Department of Critical Care Medicine, Fu Xing Hospital, Capital Medical
University, 20A Fu Xing Men Wai Da Jie, Xicheng District, Beijing 100038,
China
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Yang et al. Critical Care (2018) 22:187 Page 2 of 10

Background databases from inception through October 10, 2017:

Intensive care unit-acquired weakness (ICUAW) is a com- PubMed, Embase, Cochrane Central Register of Con-
mon neuromuscular complication of critical illness. trolled Trials, Web of Science, and Cumulative Index of
ICUAW is associated with delayed weaning, longer inten- Nursing and Allied Health Literature. The search terms
sive care unit (ICU) and hospital stays, increased were used for PubMed (Additional file 1) and the other
healthcare-related costs, and higher ICU-related and databases. In addition, a manual search of references
hospitalization-related mortality [1–3]. Corticosteroid ther- cited by the selected articles and relevant review articles
apy is still the key treatment and recommendation for spe- was performed to identify other eligible studies.
cific critically ill patients [4, 5] because of its strong
anti-inflammatory and anti-fibrotic effects. Corticosteroid Selection criteria
therapy results in a shorter duration of mechanical ventila- All studies satisfying the following criteria were included:
tion, a faster resolution of shock [6], more vasopressor-free age > 18; RCTs and prospective cohort studies; diagno-
and organ-failure-free days [7], and lower mortality [7–9] ses of ICUAW confirmed using manual muscle testing
in patients with refractory septic shock. For patients with (MRC weakness scale) or diagnostic tests (electrophysio-
acute respiratory distress syndrome (ARDS), corticosteroid logical studies, histopathology of muscle or nerve tissue);
therapy may also improve hypoxemia [10] and reduce the and studies that evaluated the use of corticosteroids and
duration of mechanical ventilation [11, 12] and the ICU incidence of ICUAW. The exclusion criteria were as fol-
hospitalization period [13]. ICUAW occurs commonly in lows: patients with primary myopathies (e.g., idiopathic
critically ill patients, but the role of corticosteroid therapy inflammatory myopathies) or polyneuropathies (e.g., my-
in ICUAW remains controversial. Researchers and authors asthenia gravis, Guillain-Barré syndrome); and studies
have raised significant concerns regarding the side effects of with insufficient data reported.
corticosteroids in terms of ICUAW development and have
attempted to examine the relationship. Some clinical stud- Study selection and data abstraction
ies have indicated that corticosteroids may contribute to de- Two reviewers (TY and ZqL) independently reviewed
veloping ICUAW, yet others have demonstrated decreasing and selected studies based on the inclusion criteria. Data
odds of developing ICUAW. However, other studies could were extracted independently by each reviewer using a
not identify the effect of corticosteroids on ICUAW. In this standardized data collection form. The following data
review, we provide a meta-analysis of randomized con- were collected from each study: author information,
trolled trials (RCTs) and prospective cohort studies to publication year, study design, study location, inclusion
assess the association between corticosteroid use and and exclusion criteria, tools of neuromuscular evalu-
ICUAW development. ation, number of participants, ICUAW incidence, and
No universal recommendation or consensus on the def- number of ICUAW patients who were given and not
inition or classification of the disease exists; after consult- given corticosteroids. Disagreements in study selection
ing the literature [14], the relatively broad term “intensive or data extraction were resolved by either consensus or
care unit-acquired weakness (ICUAW)” was selected for a third-party decision. Authors of the included studies
use in this review. Although there was no diagnostic gold were contacted when data required clarification.
standard for ICUAW, three diagnostic methods were rec-
ommended to identify ICUAW [14, 15]: manual muscle Study quality assessment
testing (Medical Research Council (MRC) weakness scale), Two reviewers (TY and ZqL) independently assessed the
electrophysiological studies, and the histopathology of methodological quality of each study using the Newcas-
muscle or nerve tissue. However, muscle or nerve tissue tle–Ottawa scale [17] for prospective studies and the
biopsy was rarely used in the studies. This review explores Cochrane Collaboration tool [17] for RCTs.
the adverse effect of corticosteroids on ICUAW develop-
ment, from patients with clinical weakness to patients Data analysis
with clinically undetectable neuromuscular dysfunction. Meta-analysis was performed using Stata version 12.0
(StataCorp, College Station, TX, USA), and the results
Methods were analyzed using odds ratios (ORs) and 95% confi-
This study was conducted according to the Preferred dence intervals (CIs). Data were pooled using the DerSi-
Reporting Items for Systematic Reviews and Meta-Ana- monian and Laird random effects model. Heterogeneity
lyses: the PRISMA statement [16]. was assessed using the χ2 statistic with P ≤ 0.1 consid-
ered statistically significant. The impact of statistical het-
Search strategy erogeneity on the study results was estimated by
A systematic literature review of all of the pertinent Eng- calculating the I2 statistic. Values of the I2 statistic above
lish language studies was undertaken in the following 50% were regarded as a cutoff point for considerable
Yang et al. Critical Care (2018) 22:187 Page 3 of 10

heterogeneity. Subgroup analyses examined: RCT and Results

prospective cohort studies; studies using clinical muscle Study search and selection
testing and electrophysiology as a diagnostic method; The electronic search yielded a total of 10,789 citations
studies using mechanical and nonmechanical ventilation (Fig. 1). Twenty-one additional articles were identified
as inclusion criteria; studies using sepsis and nonsepsis through other sources. After screening the titles and ab-
as inclusion criteria; and studies with relatively large stracts, 48 articles were selected for full-text review.
(n ≥ 100) and small (n < 100) sample sizes. Publication Thirty articles did not meet the inclusion criteria and
bias was examined using funnel plots for qualitative as- were excluded, and therefore 18 studies were included in
sessment, using Begg’s rank correlation test and Egger’s this review.
linear regression test for quantitative assessment.
Study characteristics and quality
Summary of findings The characteristics of the included studies in this sys-
The Grading of Recommendations, Assessment, Devel- tematic review are presented in Table 1. They included
opment and Evaluation (GRADE) assessment method one RCT [18] and 17 prospective cohort studies [1, 19–
was employed to determine the quality of evidence in 34]. The number of participants in each study ranged
our review associated with the main outcome (incidence from 20 to 412. The studies were carried out in the
of ICUAW). Two reviewers (TY and ZqL) independently United States [21, 27], India [19], Vietnam [1], Belgium
graded the evidence prior to agreement and created the [28], the Netherlands [20, 33], Germany [18, 25],
‘Summary of findings’ table using GRADE software [17]. Switzerland [24], France [23, 29, 31], Spain [32], Greece
We considered risk of bias, directness of evidence, het- [22, 26], Canada [30], and England [34]. Diagnosis of
erogeneity of the data, precision of effect, and risk of ICUAW was accomplished in eight studies [18, 20, 21,
publication bias as the factors influencing assessment of 23, 24, 26, 27, 31] using the MRC scale and in 10 studies
the review. [1, 19, 22, 25, 28–30, 32–34] using electrophysiological

Fig. 1 Flow diagram of the study selection process. CIHAL Cumulative Index of Nursing and Allied Health Literature, ICUAW intensive care unit-acquired
weakness, RCT randomized controlled trial
Yang et al. Critical Care (2018) 22:187 Page 4 of 10

Table 1 Characteristics of the selected studies

Study Study Country Setting Population n Examination ICUAW Use of ICU mortality
design CS a (%)a
Keh et al., 2016 [18] RCT Germany MSICU Severe sepsis 375 Clinical 82 46 vs 140 NR
Gupta and Mishra, Prospective India MICU Sepsis 100 EMG 37 26 vs 12 NR
2016 [19] cohort
Nguyen The and Prospective Vietnam MSICU ICU LOS ≥ 10 days 133 EMG 73 29 vs 18 49% vs 30%
Nguyen Huu, 2015 [1] cohort
Patel et al., 2014 [21] Prospective America MICU MV ≥ 24 h 104 Clinical 41 35 vs 44 NR
cohort
Wieske et al., Prospective Netherlands MSICU MV ≥ 2 days 212 Clinical 103 81 vs 63 34% vs 9%
2014 [20] cohort
Anastasopoulos et al., Prospective Greece MSICU ICU LOS ≥ 7 days 190 EMG 40 28 vs 102 32.5% vs NR
2011 [22] cohort
Sharshar et al., Prospective France MICU, MV > 7 days 86 Clinical 39 29 vs 22 NR
2010 [23] cohort SICU
Brunello et al., Prospective Switzerland MSICU MV > 48 h and SIRS 39 Clinical 13 4 vs 0 62% vs 23%
2010 [24] cohort
Weber-Carstens Prospective Germany SICU MV and SAPS II ≥ 20 56 EMG 34 21 vs 5 NR
et al., 2009 [25] cohort
Nanas et al., 2008 [26] Prospective Greece MSICU LOS > 10 days 185 Clinical 44 7 vs 31 36% vs 20%
cohort
Ali et al., 2008 [27] Prospective America MICU, MV > 5days 136 Clinical 35 16 vs 42 31.4% vs 6%
cohort NICU
Hermans et al., 2007 Prospective Belgium MICU MV > 7 days 412 EMG 188 123 vs NR
[28] cohort 156
Khan et al., 2006 [30] Prospective Canada MSICU Sepsis 20 EMG 10 4 vs 4 55% vs 0%
cohort
Lefaucheur et al., Prospective France MICU MV > 7 days, diffuse weakness 30 EMG 26 15 vs 1 NR
2006 [29] cohort
De Jonghe et al., Prospective France MICU, MV > 7 days and awake 95 Clinical 24 13 vs 13 17% vs 6%
2002 [31] cohort SICU
de Letter et al., Prospective Netherlands MSICU MV ≥ 4 days 97 EMG 34 9 vs 18 NR
2001 [33] cohort
Garnacho-Montero Prospective Spain MSICU MV > 10 days and sepsis with 73 EMG 50 7 vs 4 66% vs 52%
et al., 2001 [32] cohort MOF
Coakley et al., Prospective England MSICU MV and ICU LOS > 7 days 44 EMG 37 11 vs 2 NR
1998 [34] cohort
ICUAW intensive care unit-acquired weakness, CS corticosteroids, ICU intensive care unit, RCT randomized controlled trial, MICU medical ICU, MSICU medical–
surgical ICU, SICU surgical ICU, NR not reported, EMG electromyography, LOS length of stay, MV mechanical ventilation, SIRS systemic inflammatory response
syndrome, SAPS Simplified Acute Physiology Score, MOF multiple organ failure
a
Comparison between ICUAW and no ICUAW

evaluation. Participant inclusion criteria included mech- comparability. Four studies did not report whether the as-
anical ventilation in 12 studies [20, 21, 23–25, 27–29, sessments were independently blinded for clinicians or
31–34], systemic inflammatory response syndrome physical therapists.
(SIRS) or sepsis in five studies [18, 19, 24, 30, 32], and
length of ICU stay in four studies [1, 22, 26, 34]. ICU Corticosteroids and ICUAW
mortality differed across the studies. When the 18 studies were pooled together (Fig. 2), the effect
The methodological quality assessment of the included size analysis (OR 1.84; 95% CI 1.26–2.67; P = 0.002) indi-
reports is presented in Table 2. The risk of bias of the ran- cated that the use of corticosteroids was significantly associ-
domized trial was low, and the overall risk of bias of the ated with increased odds of developing ICUAW. Data were
prospective studies was acceptable in general. Three of the pooled using a random effects model considering the ob-
17 observational studies made statistical comparisons with served heterogeneity (τ2 = 0.38; χ2 = 51.87, df = 17 (P < 0.001);
multivariable regression analysis for corticosteroids, and I2 = 67.2%). The overall incidence of ICUAW was 43% in the
therefore the other 14 studies received no scores for corticosteroid group versus 34% in the control group.
Yang et al. Critical Care (2018) 22:187 Page 5 of 10

Table 2 Methodology and reporting assessment

Cochrane Collaboration tool for assessing risk of bias
Study Sequence Allocation Blinding of participants, Incomplete Selective Other potential Risk of
generation concealment personnel, and outcome outcome outcome threats to validity bias
assessors data reporting
Keh et al., 2016 Y Y Y Y Y Y Low
[18]
Newcastle–Ottawa quality assessment scale for cohort studies
Study Selection Comparability Outcome Score
Exposed Nonexposed Ascertainment Outcome of Assessment Adequate Completeness
representative? representative? of exposure interest not of outcome duration of of follow-up
present at start follow-up
Gupta and Y Y Y Y Y, Y Y Y Y 9
Mishra, 2016 [19]
Nguyen The and Y Y Y Y N, N Y Y Y 7
Nguyen Huu,
2015 [1]
Wieskeet al., Y Y Y Y N, N Y Y Y 7
2014 [20]
Patel et al., 2014 Y Y Y Y N, N Y Y Y 7
[21]
Anastasopoulos Y Y Y Y N, N Y Y Y 7
et al., 2011 [22]
Brunello et al., Y Y Y Y Y, Y N Y N 7
2010 [24]
Sharshar et al., Y Y Y Y N, N N Y Y 6
2010 [23]
Weber-Carstens Y Y Y Y N, N Y Y Y 7
et al., 2009 [25]
Nanas et al., Y Y Y Y N, N N Y Y 6
2008 [26]
Ali et al., 2008 Y Y Y Y N, N Y Y Y 7
[27]
Hermans et al., Y Y Y Y N, N Y Y Y 7
2007 [28]
Khan et al., 2006 Y Y Y Y N, N Y Y N 6
[30]
Lefaucheur et al., Y Y Y Y N, N Y Y Y 7
2006 [29]
De Jonghe et al., Y Y Y Y Y, Y Y Y Y 9
2002 [31]
de Letter et al., Y Y Y Y N, N N Y Y 6
2001 [33]
Garnacho- Y Y Y Y N, N Y Y Y 7
Montero et al.,
2001 [32]
Coakley et al., Y Y Y Y N, N Y Y Y 7
1998 [34]
Y criteria satisfied, N criteria not satisfied

Subgroup analyses (Additional file 2). The meta-analysis of 17 prospective co-

RCTs versus prospective cohort studies hort studies (OR 1.90; 95% CI 1.25–2.89; P = 0.003) showed
The subgroup analyses are presented in Table 3. One RCT a significant association with a random effects model con-
revealed no significant association between corticosteroids sidering the observed heterogeneity (τ2 = 0.46; χ2 = 51.66,
and ICUAW (OR 1.40; 95% CI 0.85–2.29; P = 0.184), and df = 16 (P < 0.001); I2 = 69.0%). The incidence of ICUAW
the GRADE quality of evidence was high for this trial was 46% in the corticosteroid group versus 36% in the
Yang et al. Critical Care (2018) 22:187 Page 6 of 10

Fig. 2 Forest plot of associations between corticosteroid use and ICUAW. CI confidence interval, OR odds ratio

control group; however, the GRADE quality of evidence patients with clinical weakness and demonstrated an in-
was very low (Additional file 2). There was no significantly cidence of 39% in the corticosteroid group and 23% in
statistical heterogeneity found between the subgroups the control group. The overall effect size (OR 2.06; 95%
based on a test of interaction (P = 0.35). CI 1.27–3.33; P = 0.003) demonstrated a significant asso-
ciation with a random effects model considering the
Clinical assessment versus electrophysiology observed heterogeneity (τ2 = 0.27; χ2 = 17.78, df = 7 (P =
Eight studies [18, 20, 21, 23, 24, 26, 27, 31] examined 0.013); I2 = 60.6%). Ten observational studies [1, 19, 22,
the association between the use of corticosteroids and 25, 28–30, 32–34] reported an association between the

Table 3 Subgroup analyses

Analysis Study n I2 (%) Ph OR 95% CI Pe Pi Incidence Incidence
corticosteroid (%) control (%)
Study type
RCT [18] 375 1.40 0.85–2.29 0.184 25 19
Prospective cohort studies [1, 19–34] 2012 69.0 < 0.001 1.90 1.25–2.89 0.003 0.35 46 36
Diagnostic method
Clinical assessment [18, 20, 21, 23, 24, 26, 27, 31] 1232 60.6 0.013 2.06 1.27–3.33 0.003 39 23
Electrophysiology [1, 19, 22, 25, 28–30, 32–34] 1155 70.6 < 0.001 1.65 0.92–2.95 0.093 0.56 46 46
Inclusion criterion
Sepsis [18, 19, 30, 32] 568 80.8 0.001 1.96 0.61–6.30 0.260 34 30
Nonsepsis [1, 20–29, 31, 33, 34] 1819 63.0 0.001 1.77 1.18–2.64 0.006 0.87 45 35
MV [20, 21, 23–25, 27–29, 31–34] 1384 66.0 0.001 2.00 1.23–3.27 0.006 50 40
Non-MV [1, 18, 19, 22, 26, 30] 1003 74.4 0.002 1.61 0.83–3.13 0.161 0.61 31 26
Sample size
n ≥ 100 [1, 18–22, 26–28] 1847 74.9 < 0.001 1.62 1.02–2.53 0.042 39 30
n < 100 [23–25, 29–34] 540 49.3 0.046 2.32 1.21–4.42 0.011 0.36 62 43
I2 I-squared statistic test for heterogeneity, Ph P value for test of heterogeneity, OR odds ratio, CI confidence interval, Pe P value for the effect estimate for each
subgroup, Pi P value for interaction tests of heterogeneity between subgroups, RCT randomized controlled trial, MV mechanical ventilation
Yang et al. Critical Care (2018) 22:187 Page 7 of 10

use of corticosteroids and patients with abnormal elec- the pooled effect size (OR 1.62; 95% CI 1.02–2.53; P =
trophysiology and showed an event rate of 46% in the 0.042) still demonstrated a significant association
corticosteroid group and 46% in the control group. The between corticosteroid use and ICUAW with a random
pooled effect size (OR 1.65; 95% CI 0.92–2.95; P = 0.093) effects model considering the observed heterogeneity
revealed no significant association. Data were pooled (τ2 = 0.35; χ2 = 31.92, df = 8 (P < 0.001); I2 = 74.9%),
using a random effects model considering the observed with an event rate of 39% in the corticosteroid group and
heterogeneity (τ2 = 0.53; χ2 = 30.63, df = 9 (P < 0.001); 30% in the control group. The remaining nine studies
I2 = 70.6%). No statistically significant heterogeneity [23–25, 29–34] with relatively small sample sizes (n < 100)
between the subgroups was found based on a test of showed an unadjusted event rate in the corticosteroid
the interaction (P = 0.56). group of 62% versus 43% in the control group. The pooled
effect size (OR 2.32; 95% CI 1.21–4.42; P = 0.011) demon-
Sepsis versus nonsepsis strated a significant association with a random effects
Four trials [18, 19, 30, 32] with sepsis as the inclusion cri- model considering the observed heterogeneity (τ2 = 0.44;
terion reported an association between the use of cortico- χ2 = 15.77, df = 8 (P = 0.046); I2 = 49.3%). No statistically
steroids and ICUAW, and demonstrated an incidence of significant heterogeneity between the subgroups was
34% in the corticosteroid group and 30% in the control found based on a test of the interaction (P = 0.36).
group. The pooled effect size (OR 1.96; 95% CI 0.61–6.30;
P = 0.260) revealed no significant association. Data were Heterogeneity
pooled using a random effects model considering the ob- Methodological heterogeneity
served heterogeneity (τ2 = 1.08; χ2 = 15.65, df = 3 (P = Methodological heterogeneity was found among the in-
0.001); I2 = 80.8%). The remaining 14 studies [1, 20–29, 31, cluded studies. Two study design types were utilized
33, 34] without sepsis as an inclusion criterion showed an among the included studies, and two diagnostic methods
unadjusted event rate in the corticosteroid group of 45% were used in the included studies. Sample sizes differed
versus 35% in the control group. The pooled effect size across the included studies; small and large studies were
(OR 1.77; 95% CI 1.18–2.64; P = 0.006) demonstrated a sig- delineated by a cutoff value of 100 subjects. This meth-
nificant association with a random effects model consider- odological heterogeneity led to three comparisons in the
ing the observed heterogeneity (τ2 = 0.32; χ2 = 35.18, df = 13 review: RCTs versus prospective cohort studies, clinical
(P = 0.001); I2 = 63.0%). No statistically significant hetero- assessment versus electrophysiology, and sample size
geneity between the subgroups was found based on a test analysis (n ≥ 100 versus n < 100).
of the interaction (P = 0.87).
Clinical heterogeneity
Mechanical ventilation versus nonmechanical ventilation Clinical heterogeneity was also observed in the included
Twelve observational studies [20, 21, 23–25, 27–29, 31–34] studies. The study cohorts were differed due to different in-
using mechanical ventilation as an inclusion criterion ex- clusion criteria among the included studies, which led to
amined the association between the use of corticosteroids two comparisons in the review: sepsis versus nonsepsis, and
and ICUAW, and showed an event rate of 50% in the cor- mechanical ventilation versus nonmechanical ventilation.
ticosteroid group and 40% in the control group. The overall
effect size (OR 2.00; 95% CI 1.23–3.27; P = 0.006) demon- Statistical heterogeneity
strated a significant association with a random effects There were high levels of statistical heterogeneity in the
model considering the observed heterogeneity (τ2 = 0.42; χ2 review, and statistical heterogeneity remained substantial
= 32.32, df = 11 (P = 0.001); I2 = 66.0%). The remaining six within each of the five comparisons described (presented
studies [1, 18, 19, 22, 26, 30] without mechanical ventilation in Table 3).
as an inclusion criterion showed an unadjusted event rate
in the corticosteroid group of 31% versus 26% in the con- Assessment of publication biases
trol group. The pooled effect size (OR 1.61; 95% CI 0.83– Funnel plots were used to estimate the publication bias.
3.13; P = 0.161) revealed no significant association with a As depicted in Fig. 3a, b, there was no significant asym-
random effects model considering the observed heterogen- metry found in the funnel plots. Begg’s test (z = 1.06, P =
eity (τ2 = 0.48; χ2 = 19.54, df = 5 (P = 0.002); I2 = 74.4%). No 0.289) and Egger’s test (t = 1.77, P = 0.095) were adopted
statistically significant heterogeneity between the subgroups to detect publication bias in the meta-analysis, and no
was found based on a test of the interaction (P = 0.61). significant biases were found.

Sample sizes (n ≥ 100 versus n < 100) Discussion

After the results of the nine studies [1, 18–22, 26–28] This review synthesized data on the relationship between
with sample sizes greater than 100 were incorporated, corticosteroids and ICUAW. We identified 18 studies with
Yang et al. Critical Care (2018) 22:187 Page 8 of 10

Fig. 3 Funnel plots. a Begg’s funnel plots of included studies. b Egger’s funnel plots of included studies. s.e. standard error

a total of 2387 enrolled patients. When the studies were significantly higher in patients with ICUAW than in those
pooled together, the effect size analysis showed that cor- without ICUAW in two studies [23, 25] based on univariate
ticosteroid use was a significant risk for developing analysis. Thus, exposure to corticosteroids should be lim-
ICUAW. ited or the dose lowered in clinical practice to reduce the
Corticosteroid therapy was still an essential treatment op- risk of ICUAW.
tion in selected critically ill patients, such as those with re- Our subgroup analyses revealed a stronger association in
fractory septic shock and ARDS. Similar muscle changes to patients with clinical weakness but not in patients with ab-
those of animals as a result of corticosteroid therapy had normal electrophysiology. The use of corticosteroids was
been found in ICU patients [35]. Corticosteroid therapy found to be significantly associated with muscle weakness
was found to cause changes in specific gene expression to in the review. However, within the electrophysiology sub-
indicate the inhibition of protein synthesis resulting in pro- group, the incidences of ICUAW in the corticosteroid and
moting muscle wasting [36, 37]. Evaluating the effect of control groups were higher than those found in the clinical
corticosteroid therapy on ICUAW development is critical. assessment subgroup. ICUAW is essentially a clinically de-
Thus, this systematic review synthesized data on the rela- tectable weakness, and clinical examinations are easier,
tionship between the use of corticosteroids and ICUAW in timelier, and more convenient to perform than electro-
ICU patients. In addition, the effect of corticosteroid ther- physiology examinations. However, clinical examinations
apy on ICUAW is complex and may also depend on the usually cannot be conducted in the early disease course due
duration and cumulative dosage of the corticosteroids. Of to suboptimal levels of consciousness or attentiveness. Elec-
the included studies, duration of the corticosteroids was trophysiologic studies may have been more sensitive for de-
not found to be an independent risk factor for ICUAW [28, tecting subclinical ICUAW in both the corticosteroid and
31], but the cumulative doses of corticosteroids were control groups, thus resulting in a nonsignificant effect of
Yang et al. Critical Care (2018) 22:187 Page 9 of 10

corticosteroid use on ICUAW in this subgroup. These con- reported the degree of missing data and how missing data
siderations may represent an alternative explanation for the were processed, and thus only a form of per-protocol ana-
different outcome. lysis was performed.
Our subgroup analyses showed that there was no signifi-
cant association between the use of corticosteroids and Conclusion
ICUAW in patients with sepsis. Corticosteroids are a crit- First, our review demonstrates a statistically significant
ical treatment for patients with sepsis, and the incidence association between corticosteroid use and ICUAW. Cli-
of this condition’s adverse event, ICUAW, was not signifi- nicians should limit exposure to corticosteroids or
cantly different in this review. A therapeutic benefit of shorten the administration time to decrease the inci-
early low-dose corticosteroid therapy for decreasing mor- dence of ICUAW. Second, we did not find a significant
tality was found in septic shock patients with the highest association between the use of corticosteroids and
severity of illness [9]. Low-dose and short-term cortico- ICUAW in patients with sepsis. Third, our review sug-
steroid therapy could improve the prognosis of specific gests a significant association between corticosteroid use
critically ill populations without increasing the risk of and ICUAW in patients with mechanical ventilation. For
ICUAW. Our subgroup analyses demonstrated that stud- specific critically ill patients, clinicians should target
ies limited to patients with mechanical ventilation still re- low-dose and short-term corticosteroid therapy in clin-
vealed the significant association between corticosteroids ical practice to limit the adverse effects of the drugs. Fu-
and ICUAW. ICUAW significantly increases the duration ture research should focus on RCTs or prospective
of mechanical ventilation [2, 38, 39], and thus the benefits cohort studies by performing multivariable adjustment
of corticosteroids should be weighed against the adverse for confounders to identify the associations between the
effect in ICUAW. Our subgroup analyses revealed that use, duration, and total doses of corticosteroids and
studies limited to relatively large sample sizes still demon- ICUAW.
strated a significant association between corticosteroid use
and ICUAW, and this result partly demonstrates the sta- Additional files
bility of the overall effect size.
Studies were excluded for the following common rea- Additional file 1: PubMed search strategy. (DOCX 15 kb)
sons: the study design was not a RCT or prospective co- Additional file 2: Summary of findings for the main comparison.
hort, insufficient data were reported, and clear (DOCX 14 kb)
diagnostic criteria were lacking. Only RCTs and pro-
spective cohort studies were included in the review. Abbreviations
ARDS: Acute respiratory distress syndrome; CI: Confidence interval;
However, only three studies controlled for other add- GRADE: Grading of Recommendations, Assessment, Development and
itional factors based on multivariate analysis. We dem- Evaluation; ICU: Intensive care unit; ICUAW: Intensive care unit-acquired
onstrated a modest association between the use of weakness; MRC: Medical Research Council; OR: Odds ratio; RCT: Randomized
controlled trial; SIRS: Systemic inflammatory response syndrome
corticosteroids and ICUAW, without adjustment for po-
tential confounders. Availability of data and materials
There are limitations to our review. The included stud- All data generated or analyzed during this study are included in this
published article.
ies were not population-based cohort studies. Temporal
trends were not examined in the included studies. Baseline Authors’ contributions
exposure to corticosteroids was not reported in the in- TY and ZqL contributed equally to the study design, study selection, data
cluded studies and thus could not be examined via extraction, quality assessment, data analysis, and writing or revising the
manuscript for important intellectual content, gave approval of the final
meta-regression. Because different risk factors existed manuscript, and served as principal authors. XmX and LJ contributed to the
across the included studies and because few studies were study conception, design, and data interpretation, revised the manuscript for
designed to adjust for other independent risk factors, pri- critical intellectual content, and supervised the study. All authors read and
approved the final manuscript.
mary analysis was performed using a univariate approach
without adjustment for potential confounders. High levels Ethics approval and consent to participate
of heterogeneity were identified for all of the outcomes. Not applicable.
We analyzed the outcomes in subgroups classified by Consent for publication
study design, diagnostic methods, sample sizes, and study Not applicable.
participants in an effort to reduce methodological and
Competing interests
clinical heterogeneity; however, substantial statistical het-
The authors declare that they have no competing interests.
erogeneity remained despite these attempts. Therefore, a
random effects model rather than a fixed effects model
Publisher’s Note
was selected to address the observed heterogeneity. Add- Springer Nature remains neutral with regard to jurisdictional claims in
itionally, none of the included prospective cohort studies published maps and institutional affiliations.
Yang et al. Critical Care (2018) 22:187 Page 10 of 10

Author details 19. Gupta S, Mishra M. Acute Physiology and Chronic Health Evaluation II score
1
Department of Critical Care Medicine, Fu Xing Hospital, Capital Medical of ≥15: a risk factor for sepsis-induced critical illness polyneuropathy. Neurol
University, 20A Fu Xing Men Wai Da Jie, Xicheng District, Beijing 100038, India. 2016;64(4):640–5.
China. 2Department of Critical Care Medicine, North China University of 20. Wieske L, Witteveen E, Verhamme C, Dettling-Ihnenfeldt DS, van der Schaaf
Science and Technology Affiliated Hospital, Tangshan, China. M, Schultz MJ, van Schaik IN, Horn J. Early prediction of intensive care unit-
acquired weakness using easily available parameters: a prospective
Received: 30 January 2018 Accepted: 29 June 2018 observational study. PLoS One. 2014;9(10):e111259.
21. Patel BK, Pohlman AS, Hall JB, Kress JP. Impact of early mobilization on
glycemic control and ICU-acquired weakness in critically ill patients who are
mechanically ventilated. Chest. 2014;146(3):583–9.
References 22. Anastasopoulos D, Kefaliakos A, Michalopoulos A. Is plasma calcium
1. Nguyen The L, Nguyen Huu C. Critical illness polyneuropathy and myopathy concentration implicated in the development of critical illness
in a rural area in Vietnam. J Neurol Sci. 2015;357(1–2):276–81. polyneuropathy and myopathy? Crit Care. 2011;15(5):R247.
2. Hermans G, Van Mechelen H, Clerckx B, Vanhullebusch T, Mesotten D, 23. Sharshar T, Bastuji-Garin S, De Jonghe B, Stevens RD, Polito A, Maxime V,
Wilmer A, Casaer MP, Meersseman P, Debaveye Y, Van Cromphaut S, et al. Rodriguez P, Cerf C, Outin H, Touraine P, et al. Hormonal status and ICU-
Acute outcomes and 1-year mortality of intensive care unit-acquired acquired paresis in critically ill patients. Intensive Care Med. 2010;36(8):1318–26.
weakness. A cohort study and propensity-matched analysis. Am J Respir Crit 24. Brunello AG, Haenggi M, Wigger O, Porta F, Takala J, Jakob SM. Usefulness of a
Care Med. 2014;190(4):410–20. clinical diagnosis of ICU-acquired paresis to predict outcome in patients with
3. Penuelas O, Muriel A, Frutos-Vivar F, Fan E, Raymondos K, Rios F, Nin N, SIRS and acute respiratory failure. Intensive Care Med. 2010;36(1):66–74.
Thille AW, Gonzalez M, Villagomez AJ, et al. Prediction and outcome of 25. Weber-Carstens S, Koch S, Spuler S, Spies CD, Bubser F, Wernecke KD, Deja M.
intensive care unit-acquired paresis. J Intensive Care Med. 2018;33(1):16–28. Nonexcitable muscle membrane predicts intensive care unit-acquired paresis
4. Cho YJ, Moon JY, Shin ES, Kim JH, Jung H, Park SY, Kim HC, Sim YS, Rhee in mechanically ventilated, sedated patients. Crit Care Med. 2009;37(9):2632–7.
CK, Lim J, et al. Clinical practice guideline of acute respiratory distress 26. Nanas S, Kritikos K, Angelopoulos E, Siafaka A, Tsikriki S, Poriazi M,
syndrome. Tuberc Res Dis. 2016;79(4):214–33. Kanaloupiti D, Kontogeorgi M, Pratikaki M, Zervakis D, et al. Predisposing
5. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, Kumar A, factors for critical illness polyneuromyopathy in a multidisciplinary intensive
Sevransky JE, Sprung CL, Nunnally ME, et al. Surviving Sepsis Campaign: care unit. Acta Neurol Scand. 2008;118(3):175–81.
International Guidelines for Management of Sepsis and Septic Shock: 2016. 27. Ali NA, O'Brien JM Jr, Hoffmann SP, Phillips G, Garland A, Finley JC, Almoosa
Intensive Care Med. 2017;43(3):304–77. K, Hejal R, Wolf KM, Lemeshow S, et al. Acquired weakness, handgrip
6. Venkatesh B, Finfer S, Cohen J, Rajbhandari D, Arabi Y, Bellomo R, Billot L. strength, and mortality in critically ill patients. Am J Respir Crit Care Med.
Adjunctive glucocorticoid therapy in patients with septic shock. N Engl J 2008;178(3):261–8.
Med. 2018;378(9):797–808. 28. Hermans G, Wilmer A, Meersseman W, Milants I, Wouters PJ, Bobbaers H,
7. Annane D, Renault A, Brun-Buisson C, Megarbane B, Quenot JP, Siami Bruyninckx F, Van den Berghe G. Impact of intensive insulin therapy on
S, Cariou A, Forceville X, Schwebel C, Martin C, et al. Hydrocortisone neuromuscular complications and ventilator dependency in the medical
plus fludrocortisone for adults with septic shock. N Engl J Med. 2018; intensive care unit. Am J Respir Crit Care Med. 2007;175(5):480–9.
378(9):809–18. 29. Lefaucheur JP, Nordine T, Rodriguez P, Brochard L. Origin of ICU acquired
8. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y. Corticosteroids paresis determined by direct muscle stimulation. J Neurol Neurosurg
for treating sepsis. Cochrane Database Syst Rev. 2015;12:CD002243. Psychiatry. 2006;77(4):500–6.
9. Funk D, Doucette S, Pisipati A, Dodek P, Marshall JC, Kumar A. Low-dose 30. Khan J, Harrison TB, Rich MM, Moss M. Early development of critical illness
corticosteroid treatment in septic shock: a propensity-matching study. Crit myopathy and neuropathy in patients with severe sepsis. Neurology. 2006;
Care Med. 2014;42(11):2333–41. 67(8):1421–5.
10. Confalonieri M, Urbino R, Potena A, Piattella M, Parigi P, Puccio G, Della 31. De Jonghe B, Sharshar T, Lefaucheur JP, Authier FJ, Durand-Zaleski I, Boussarsar
Porta R, Giorgio C, Blasi F, Umberger R, et al. Hydrocortisone infusion for M, Cerf C, Renaud E, Mesrati F, Carlet J, et al. Paresis acquired in the intensive
severe community-acquired pneumonia: a preliminary randomized study. care unit: a prospective multicenter study. JAMA. 2002;288(22):2859–67.
Am J Respir Crit Care Med. 2005;171(3):242–8. 32. Garnacho-Montero J, Madrazo-Osuna J, Garcia-Garmendia JL, Ortiz-Leyba C,
11. Steinberg KP, Hudson LD, Goodman RB, Hough CL, Lanken PN, Hyzy R, Jimenez-Jimenez FJ, Barrero-Almodovar A, Garnacho-Montero MC, Moyano-
Thompson BT, Ancukiewicz M. Efficacy and safety of corticosteroids for Del-Estad MR. Critical illness polyneuropathy: risk factors and clinical
persistent acute respiratory distress syndrome. N Engl J Med. 2006;354(16): consequences. A cohort study in septic patients. Intensive Care Med. 2001;
1671–84. 27(8):1288–96.
12. Annane D, Sebille V, Bellissant E. Effect of low doses of corticosteroids in 33. de Letter MA, Schmitz PI, Visser LH, Verheul FA, Schellens RL, Op de Coul
septic shock patients with or without early acute respiratory distress DA, van der Meche FG. Risk factors for the development of polyneuropathy
syndrome. Crit Care Med. 2006;34(1):22–30. and myopathy in critically ill patients. Crit Care Med. 2001;29(12):2281–6.
13. Meduri GU, Golden E, Freire AX, Taylor E, Zaman M, Carson SJ, Gibson M, 34. Coakley JH, Nagendran K, Yarwood GD, Honavar M, Hinds CJ. Patterns of
Umberger R. Methylprednisolone infusion in early severe ARDS: results of a neurophysiological abnormality in prolonged critical illness. Intensive Care
randomized controlled trial. Chest. 2007;131(4):954–63. Med. 1998;24(8):801–7.
14. Stevens RD, Marshall SA, Cornblath DR, Hoke A, Needham DM, de Jonghe B, 35. Massa R, Carpenter S, Holland P, Karpati G. Loss and renewal of thick
Ali NA, Sharshar T. A framework for diagnosing and classifying intensive myofilaments in glucocorticoid-treated rat soleus after denervation and
care unit-acquired weakness. Crit Care Med. 2009;37(10 Suppl):S299–308. reinnervation. Muscle Nerve. 1992;15(11):1290–8.
15. Fan E, Cheek F, Chlan L, Gosselink R, Hart N, Herridge MS, Hopkins RO, 36. Di Giovanni S, Molon A, Broccolini A, Melcon G, Mirabella M, Hoffman EP,
Hough CL, Kress JP, Latronico N, et al. An official American Thoracic Society Servidei S. Constitutive activation of MAPK cascade in acute quadriplegic
clinical practice guideline: the diagnosis of intensive care unit-acquired myopathy. Ann Neurol. 2004;55(2):195–206.
weakness in adults. Am J Respir Crit Care Med. 2014;190(12):1437–46. 37. Aare S, Radell P, Eriksson LI, Akkad H, Chen YW, Hoffman EP, Larsson L.
16. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred Reporting Items for Effects of corticosteroids in the development of limb muscle weakness in a
Systematic Reviews and Meta-Analyses: the PRISMA statement. PLoS Med. porcine intensive care unit model. Physiol Genomics. 2013;45(8):312–20.
2009;6(7):e1000097. 38. De Jonghe B, Bastuji-Garin S, Sharshar T, Outin H, Brochard L. Does ICU-
17. Higgins JPT, Green S: Cochrane handbook for systematic reviews of acquired paresis lengthen weaning from mechanical ventilation? Intensive
interventions Version 5.1.0 [updated March 2011]. The Cochrane Care Med. 2004;30(6):1117–21.
Collaboration, 2011 London, UK. Available at: http://handbook.cochrane.org. 39. Garnacho-Montero J, Amaya-Villar R, Garcia-Garmendia JL, Madrazo-Osuna J,
Accessed 15 July 2015. Ortiz-Leyba C. Effect of critical illness polyneuropathy on the withdrawal
18. Keh D, Trips E, Marx G, Wirtz SP, Abduljawwad E, Bercker S, Bogatsch H, from mechanical ventilation and the length of stay in septic patients. Crit
Briegel J, Engel C, Gerlach H, et al. Effect of hydrocortisone on development Care Med. 2005;33(2):349–54.
of shock among patients with severe sepsis: the HYPRESS randomized
clinical trial. JAMA. 2016;316(17):1775–85.