ARTICLE IN PRESS

Effects of Early Mobilization after Acute Stroke:

A Meta-Analysis of Randomized Control Trials

,
Zhuyue Li, MM,* † Xuemei Zhang, BSc,‡ Kang Wang, MM,§ and Jin Wen, PhD†

Background: Early mobilization is inconsistently associated with the recovery of

stroke. We aim to examine the effect of early mobilization on patients with acute
stroke. Methods: PubMed, EMBASE, and the Cochrane library were searched up
to April 2017. Randomized controlled trials that reported risk estimates or mean
with standard deviation were included. Primary outcomes were defined as modi-
fied Rankin scale score 0-2 and mortality, and secondary outcomes were Barthel
Index, length of stay, and incidence of complications. Summary relative risk, stan-
dardized mean difference (SMD), and weighted mean difference (WMD) were
calculated as needed. Sensitivity analyses were also conducted to test stability of
results. Results: Six studies (8 publications) were included to analyze the effects
of early mobilization after stroke. No differences between groups were observed
for modified Rankin scale 0-2 (relative risk [RR]: .80; 95% confidence interval [CI]:
.58-1.02; I2 = 45%) and the risk of death (RR: 1.21, 95% CI: .76-1.65; I2 = 0%). Com-
pared with conventional practice, early mobilization was superior in Barthel Index
(SMD: .66; 95% CI: .00-1.31; I2 = 85.9%), and shorter hospital stay for stroke pa-
tients (WMD: −1.97; 95% CI: −2.63 to −1.32; I2 = 15.3%). We found no significant
difference between groups on the incidence of complications. Conclusions: Current
evidence revealed that no statistical significant difference between early mobili-
zation and non-early mobilization was observed on modified Rankin scale score
0-2 and mortality. Interestingly, early mobilization is associated with an in-
creased Barthel Index and shorter hospital stay for patients. Further research is
necessary to verify the effect of early mobilization on patients with acute stroke.
Key Words: Early mobilization—stroke—modified Rankin scale—mortality—self-care
ability—rehabilitation—Barthel Index.
© 2017 National Stroke Association. Published by Elsevier Inc. All rights reserved.

From the *West China Hospital/West China School of Nursing, Sichuan University; †Institute of Hospital Management, West China Hos-
pital, Sichuan University; ‡Center of Gerontology and Geriatrics, West China Hospital, Sichuan University, Chengdu, China; and §Department
of the Endocrine and Breast Surgery, The First Affiliated hospital of Chongqing Medical university, Chongqing Medical University,
Chongqing, China.
Received October 24, 2017; revision received December 2, 2017; accepted December 17, 2017.
Zhuyue Li and Xuemei Zhang equally contributed to this article.
Conflict of interest: The authors declare that there is no conflict of interest.
Author contribution: L.Z.Y. wrote the main manuscript. W.J. provided the study idea. Z.X.M. and L.Z.Y. prepared tables and figures. W.K.
and L.Z.Y. completed article screening and abstracted information. Z.X.M. and L.Z.Y. performed literature search and quality assessment.
W.J. made critical comments and revision for the manuscript. All authors reviewed and approved the final manuscript.
Address correspondence to Jin Wen, PhD, Institute of Hospital Management, West China Hospital, Sichuan University. Guo Xue Xiang 37,
Chengdu, 610041, China. E-mail: huaxiwenjin@163.com.
1052-3057/$ - see front matter
© 2017 National Stroke Association. Published by Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.jstrokecerebrovasdis.2017.12.021

Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2017: pp ■■–■■ 1
 ARTICLE IN PRESS
2 L. ZHUYUE ET AL.

Introduction mobilization,” “early rehabilitation,” “early ambulation,”

“modified Rankin scale,” “mortality,” “Barthel Index,”
Stroke is one of the most important causes of death
“functional independence measure,” “length of stay,” and
and long-term disability around the world, and devel-
“outcome assessment.” We do not attempt to get unpub-
oping countries show heavier burden of stroke than
lished papers, and we manually checked the reference
developed countries.1,2 Given the burden of stroke,
lists from included studies and relevant reviews to iden-
nearly 800,800 individuals were affected by stroke an-
tify additional citations.
nually and many survivors suffer constant difficulty
with daily tasks.3 Immediate admission, tissue plasmino-
gen activator. and early mobilization (EM) have been Study Selection
proposed, which were beneficial for prognosis of pa-
Trails were included if they met the following char-
tients with stroke.4 EM has attracted much attention
acteristics: (1) population—patients with acute ischemia
recently in the clinical research field,5,6 which is a pro-
or hemorrhagic stroke; (2) intervention—EM within 24
cedure to accelerate the ability of patients to walk or
hours post stroke (consensus definition of start time for
move, which is characterized by a shorter period of
EM was inadequate, therefore we definite it as 24 hours.);
hospitalization or recumbency than normally practiced,
(3) control—non-EM protocols (e.g., delayed mobiliza-
such as early sitting out of bed, transfer, standing, and
tion, usual care); (4) outcomes—mRS score 0-2 and
walking.7-16
mortality were considered as primary outcomes for this
Some studies showed a positive effect of EM,14,15,17
meta-analysis, secondary outcomes included Barthel Index
whereas others revealed inconsistent and controversial
(BI), length of stay (LOS), and incidence of complica-
findings.7,8 A 2008 phase II clinical trial conducted by
tions; and (5) study design—RCT. Study screening was
Bernhardt et al17 revealed that EM within 24 hours ap-
conducted by a 2-stage method. First, titles and ab-
peared to be safe and feasible for stroke patients and
stracts were scrutinized to exclude ineligible studies. Second,
improved their independence in activities of daily living.
investigators read the full texts for including eligible studies.
Meanwhile, an individual patient data meta-analysis made
Any disagreements on study screening were resolved by
a comparable conclusion by precise data source.18 Un-
discussions.
expectedly, an international randomized controlled trial
(RCT) involving more than 2000 stroke patients7 found
that EM within 24 hours reduced the odds of a favor- Data Extraction and Quality Assessment
able outcome defined as modified Rankin scale (mRS) score
In the present study, data extraction was performed
0-2, which might change clinical decision. EM has both
by investigators (L.Z.Y. and W.K.) as follows: first author,
potential harm and potential benefits on recovery process
publication year, study location, participant’s mean age,
post stroke, such as potential aggravation of neurologic
sample size, stroke type, intervention protocol, control
deficits 19,20 and preventing immobility-related
protocol, the first out-of-bed activities time, outcomes, mean
complications,21 arising confusion on the topic for pa-
and standard deviation for continuous variable, and risk
tients with acute stroke. Some researchers found that it
estimate with corresponding 95% confidence interval (CI)
is not adequate to confirm that EM is beneficial to the
for binary variable. Median with range was extracted
recovery of patients with stroke,22,23 and the evidences
when initial study did not report mean and standard
need to be further explored.
deviation.
For these reasons we conducted a systematic review
We appraised quality of each research by the “risk of
to study this controversial topic and explore the effect
bias” tool, which was recommended by the Cochrane
of EM post stroke and its impact on prognosis of pa-
Handbook,25 and quality scores were assigned for (1)
tients with stroke.
random sequence, (2) allocation concealment, (3) blind-
ing of participants and personnel, (4) blinding of outcome
Methods assessment, (5) incomplete outcome data, (6) selective re-
porting, and (7) other bias.
Search Strategy
The quality of evidence for each outcome was graded
The present study was conducted according to the Pre- according to GRADE guideline.26 We assessed down-
ferred Reporting Items for Systemic Reviews and Meta- grading factors including risk of bias, inconsistency of
Analyses statement.24 We conducted a comprehensive results, indirectness of evidence, imprecision, as well as
electronic search to obtain relevant RCTs published from publication bias, and upgrading factors comprising
1970 to April 2017 on PubMed, EMBASE, and the Co- large magnitude of effect, plausible confounding factors
chrane Library. Supplement Table S1 in the online-only would change the effect, and dose-response gradient,
data showed the search strategy in detail. Both free text respectively. According to evaluation, quality of evi-
and explored Medical Subject Headings terms were used dence was judged as “very low,” “low,” “moderate,” or
as follows: “stroke,” “cerebrovascular accident,” “early “high.”
 ARTICLE IN PRESS
EARLY MOBILIZATION AND ACUTE STROKE
Statistical Analysis
For this meta-analysis, relative risk (RR) and 95% CI
were employed as the effect size of binary variables such
as mRS score 0-2, mortality, and incidence of complica-
tions. Standardized mean differences (SMD) with 95% CI
were adopted for BI, and weighted mean differences with
95% CI for LOS. Estimating method proposed and de-
veloped by Hozo and colleagues27 was applied to estimate
continuous outcomes, which reported median and range
rather than mean value and standard deviation. Q-statistic
(significance level at P ≤ .10) and the I2-statistic (high het-
erogeneity: I2 ≥ 75%; moderate heterogeneity: 50% ≤ I2 <
75%; low heterogeneity: I2 < 50%)28 were respectively used
to qualitatively and quantitatively evaluate the between-
study heterogeneity.
Omitting a single study in turn and repeating meta-
analysis through a fixed-effects model were employed to
perform sensitivity analyses. To test publication bias for
the present study, we simultaneously used Begg’s rank
correlation test29 and Egger’s linear regression test30 with
funnel plot (P ≤ .10 was found to be statistically signif-
icant). Sensitivity analyses and publication bias were
conducted only on BI, mRS score 0-2, mortality, and in-
cidence of complications because of quantity limitations
of included studies. Our study used subgroup analyses
to clarify the differences in results regarding varied RCT
types, and a Pinteraction between subgroups was calculated
by meta-regression.31
Data synthesis was performed by STATA 12.0 (StataCorp,
College Station, TX) software. We used random effect
analysis32 to obtain pooled values because of inevitable
heterogeneity; P value less than .05 in 2-sided test was
considered as statistically significant.
Results
Literature Search
Figure 1 showed the flowchart of literature identifica-
tion. The initial literature retrieval respectively identified
975, 814, and 58 citations from PubMed, EMBASE, and
the Cochrane Library, respectively. A total of 1706 rele-
vant citations remained after removing duplications. In
the process of screening, 1670 citations were excluded after
scrutinizing their titles and abstracts. Among 36 eligible
studies, 5 literatures were included through reading full
texts, and 3 studies were added by reviewing relevant
reference lists.13-15 The details of excluded studies and their
reasons are given in the online-only data Supplement
Table S2. As our primary outcome, 5 of 8 included studies
supplied analysis for mRS score 0-2 7,8,11,15,17
and 4 for mortality.7,8,15,16 In addition, for secondary
outcomes, 4 studies were included for BI,8,9,14,15 4 for in-
cidence of complications (e.g., immobility-related
complications and secondary changes),7,8,15,16 and 3 for
LOS,8,9,15 respectively.
3
Study Characteristics
Table 1 showed the results of study characteristics.
Based on the same study subjects, 3 studies14,16,17 re-
ported different outcomes, which were available for
synthesis. Our study included 8 publications from 6
unique studies, which were published between 200817
and 2016,8 2 studies were conducted in European
countries,8,11 1 in Asia,9 2 in Australia,14,15 and the re-
maining one is an international RCT.7 The sample
size ranged from 3215 to 2104,7 and mean age of
participants changed from 59.32/60.57 years (interven-
tion group/control group) 15 to 76.4 years. 11 Five
studies with 1646 participants were employed to
analyze the outcome of mRS score 0-2.7,8,11,15,17 Four
studies reported 181 death cases among 2440 partici-
pants, which were eligible for analysis of mortality
post stroke. 7,8,15,16 The analysis of BI included 4
trials with 285 participants. 8,9,14,15 The analyses
included 47,8,15,16 and 38,9,15 for LOS and complications,
respectively.
Analyses of Outcomes
Primary Outcomes
Figure 2 to Figure 6 exhibited all forest plots. Five
trials provided information on mRS score 0-2 in a total
of 1646 participants. Meta-analysis showed no signifi-
cant effect between groups (RR: .80; 95% CI: .58-1.02;
Fig 2), with low heterogeneity (I2 = 45%, Pheterogeneity = .12).
Summary effect size did not change in sensitivity anal-
ysis, and effect sizes ranged from RR: .83; 95% CI (.62,
1.26) to RR:1.11; 95% CI (.51, 2.40) for mRS score 0-2.
Clear evidence of publication bias was not present by
Begg’s test and Egger’s test (all P > .1).
A total of 4 trials reported death cases, and the risk
of death was similar between groups (RR: 1.21, 95% CI:
.76-1.65; Fig 3), with no between-study heterogeneity
(I2 = 0%, Pheterogeneity = .54). The finding was robust in sen-
sitivity analysis, and the effect size ranged from RR: 1.07;
95%CI (.31, 3.72) to RR: 1.39; 95%CI (.94, 2.08). In addi-
tion, clear evidence of publication bias was not detected
(all P > .1).
Secondary Outcomes
The relationship between EM and self-care ability
was evaluated in 4 studies, comprising 285 participants.
The pooled SMD of BI was .66 (95% CI .00-1.31; Fig 4)
with high heterogeneity (Pheterogeneity < .001, I2 = 85.9%),
indicating that the intervention of EM increased score
of BI and reduced dependence of patients on daily
tasks. We performed sensitivity analysis when pooling
data from these 4 trials, which failed to change the
statistical significance on SMD of BI with EM compared
 ARTICLE IN PRESS
4 L. ZHUYUE ET AL.

Figure 1. Flow chart of study selection in meta-analysis of early mobilization after stroke.

with delayed mobilization or standard care. When
omitting a study in turn, pooled values ranged
from SMD: .46; 95%CI (−.25, 1.16) to SMD: .93; 95%CI
(.25, 1.62) for BI. Publication bias was not revealed
(all P > .1).
In total of 236 participants, 3 studies were eligible for
analysis of the association between hospital stay and EM.
The result derived from these studies indicated that stroke
patients with EM were discharged nearly 2 days earlier
than the control group (weighted mean difference: −1.97,
95% CI: −2.63 to −1.32; Fig 5). Heterogeneity between
studies was low (I2 = 15.3%, Pheterogeneity = .31).
Four trials that randomized 2428 stroke patients
compared incidence of complications between EM
patients and non-EM patients. No difference was evident
for the incidence of complications (RR: .99; 95% CI:
.80-1.17; Fig 6) and no heterogeneity was present
(I2 = 0%,Pheterogeneity = .45). The finding withstood sensitiv-
ity analysis. Clear evidence of publication bias was not
present (all P > .1).
 Table 1. Characteristics of included studies regarding early mobilization and prognosis of patients with stroke

EARLY MOBILIZATION AND ACUTE STROKE

Author;
publication year; Mean Sample Stroke FOAT(h) Intervention
study location age (y) size (I/C) type (I/C) protocol Control protocol Outcomes

Herisson et al; 68.1/71.2 82/85 Ischemic 26.4 ± 4.8/ Started time: within 24 h. day 0: positioned in bed at 30°; Primary outcome: the proportion of
20168; France 72 ± 4.8* Duration: 15-60 min. day 1: positioned in bed at 45°; mRS score 0-2.
day 2:positioned in bed at 60°; Secondary outcomes: NIHSS scores,
day 3: sitting out of bed. BI, complications, LOS.
Duration:15-60 min.
Chippala and 59.32/60.57 43/43 NR 18(16.62-19.75)/ Started time: within 24 h. Standard care. Duration: BI.
Sharma, 30.5(29.0-35)† Duration: 5-30 min. approximately 45 min.
20169; India Frequency: minimum

ARTICLE IN PRESS
2 times per day.
Bernhardt et al; 72.3/72.7 1054/1050 Ischemic or 18.5(12.8-22.3)/ Started time: within 24 h. Started time: 24-48 h. Primary outcome: the proportion of
20157; hemorrhagic 2. 4(16.5-29.3) Duration: approximately Duration: at least 10 min. mRS (0-2). Secondary outcome:
Australia, 31 minutes. Frequency: 3 times per week. walking 50 m unassisted; mRS
New Zealand, Frequency: at least twice category.
Malaysia, per day, 6 days per
Singapore, week.
and the UK
Sundseth et al; 76.4 123/123 Ischemic or 10.5(8.5-22.3)/ Started time: within 24 h. Started time: 24-48 h. the proportion of mRS (0-2).
201411; hemorrhagic 35.8(28.0-41.0)
Norway
‡Cumming 74.6/74.9 38/33 NR 18/31 Started time: within 24 h. Standard care. Primary outcome: walking 50 m
et al; 201114; Frequency: at least twice unassisted. Secondary outcome:
Switzerland a day. BI, Rivermead Motor Assessment.
Langhorne et al; 64/71 16/16 Ischemic or 27.3(26.0-29.0)/ Started time: within 24 h. Standard care. Duration: Primary outcome: dead or disabled
201015; hemorrhagic 32.0(22.5-47.3) Frequency: at least 4 30-60 min. (modified Rankin Score 3-5).
Australia times per day. Secondary outcome: complications,
BI, LOS.
‡Sorbello et al; 74.6/74.9 38/33 NR 18/31 Started time: within 24 h. Standard care. Complications, death.
200916; Frequency: at least twice
Australia a day.
‡Bernhardt 74.6/74.9 38/33 NR 18/31 Started time: within 24 h. Standard care. Primary outcome: death. Secondary
et al; 200922; Frequency: at least twice outcome: serious adverse events,
Australia a day. deterioration, Borg Perceived
Exertion scale.

Abbreviations: BI, Barthel Index; C, control; day 0, the day of stroke onset; day 1, the day after day 0; day 2, the day after day1; day 3, the day after day 2; FOAT, The First Out-of-bed Ac-
tivities Time; h, hours; NIHSS, National Institutes of Health Stroke Scale; I, intervention; LOS, length of stay.; mRS, modified Rankin scale; RCT, randomized control trial.
*Exhibited as mean ± standard deviation.
†Exhibited as median with range.

5
‡The 3 studies from the same dataset with different outcomes reported.
 ARTICLE IN PRESS
6 L. ZHUYUE ET AL.

Figure 2. Forest plot of outcome of modified Rankin scale after 3 months of stroke.

Figure 3. Forest plot of mortality after 3 months of stroke.

Subgroup Analyses difference, effect of EM on mortality and incidence of

complications were similar between subgroup. The
The results of traditional RCTs were comparable pragmatic RCT and traditional RCTs showed different
with pragmatic RCT, which were exhibited in Table 2. result on mRS score 0-2. All Pinteraction were greater than
Although the sample sizes showed significant .05.
 ARTICLE IN PRESS
EARLY MOBILIZATION AND ACUTE STROKE 7

Figure 4. Forest plot of outcome of Barthel Index after 3 months of stroke.

Figure 5. Forest plot of outcome of length of stay.

Quality Assessment and Recommendations
Determination
Figure 7 displayed the results of the quality assess-
ment of included studies. In the domain of selection bias,
random sequence generation and allocation concealment
were adequately generated in all trials, and the risk of
bias was assessed to be low. Some studies blinded par-
ticipants, whereas evidence was inadequate to judge other
studies and the performance bias was judged to be low.
Outcome assessments were successfully blinded in most
studies and the detection bias was considered to be low.
 ARTICLE IN PRESS
8 L. ZHUYUE ET AL.

Figure 6. Forest plot of incidence of complications after 3 months for stroke patients.

Table 2. Subgroup analysis of pragmatic RCT versus traditional RCT

Pragmatic trial1 Traditional RCTs*

Sample size (I/C) Effect size Sample size (I/C) Effect size Pinteraction

mRS score 0-2 1038/1045 RR0.73(.59, .9) 142/148 RR0.81(.38, 1.23) .57
Mortality 1048/1050 RR1.34(.93, 1.93) 166/182 RR0.68(−.32, 1.68) .96
Complication 1048/1050 RR0.99(.77, 1.21) 165/179 RR1.00(.62, 1.38) .80

Abbreviations: C, control group; I, intervention group; mRS, modified Rankin Scale.

*Studies for subgroup analysis of mRS score 0-22-5; for subgroup analysis of mortality2,4,6; for subgroup analysis of the incidence of
complications.2,4,6

There was a low risk of attrition bias, for most data came
from studies. As for selective outcome reporting, the bias
was assessed to be low. Other risk from included studies
remained unclear.
The quality of evidence was judged by the GRADE
profiler and exhibited in Table 3. Mortality was consid-
ered as high-quality outcome, self-care ability measured
with BI and mRS score 0-2 was assessed to be of mod-
erate quality.
Discussion
On the controversial topic, this study failed to reveal
the associations between EM and the risk of death, better
neurologic prognosis (namely mRS score 0-2), and the in-
cidence of complications as well. However, findings
supported that EM after stroke increased BI and improved
self-care ability, which is comparable with the result of
the trial conducted by Herisson et al.8
Before this review, we understood that EM could help
improve stroke recovery, and EM has been practiced widely
or recommended generally with insufficient evidence.33,34
However, the multicenter international AVERT trial7 sug-
gested that an earlier, more frequent and higher dose
mobilization reduced the odds of mRS score 0-2. Nev-
ertheless, this pragmatic RCT 7 has some important
limitations, which might reduce the efficacy of its find-
ings. First of all, only 4 hours was the difference on first
mobilization between the intervention group and the control
group; besides, continuously shortened interval between
stroke onset and mobilization in the control group de-
creased between-group difference (median 18.5 hours versus
22.4 hours post stroke onset). Second, the intervention
was more frequent (6.5 per day) and more intensive (31
 ARTICLE IN PRESS
EARLY MOBILIZATION AND ACUTE STROKE
Figure 7. Risk of bias graph of included studies.
minutes per day) than usual care, which might be a
manmade risk for the stroke patients. According to the
latest AVERT Phase III report,35 researchers concluded that
adopting shorter, more frequent mobilization early post
stroke may be associated with a more favorable outcome
for patients, which revealed that dose and frequency may
be the most important factor on effect of EM post stroke.
Our subgroup analysis of traditional RCTs revealed that
there are no significant differences regarding mortality
and the incidence of complications between the inter-
vention and the control groups, which is generally
consistent with the pragmatic RCT.7 In our sensitivity anal-
ysis, heterogeneity decreased from 45% to 31.4% when
the large sample size RCT was removed on the outcome
of mRS score 0-2.7 Although the study by Bernhardt et al7
occupied a large weight of pooled value (45.71% for mRS
score 0-2, 79.93% for mortality, and 68.21% for inci-
dence of complications) because of the large sample size,
our results were still stable according to sensitivity anal-
yses. Namely, the present study is robust.
The following arguments contribute to potential mech-
anism of EM on BI. On the one hand, EM prevents some
bed rest-related effects on patient’s musculoskeletal, car-
diovascular, respiratory, and immune systems, which hinder
the recovery of self-care ability21; on the other hand, an
enhanced self-efficacy gave patients increased confidence36
and accelerated discharge. Paradoxically, previous meta-
analysis conducted by Lynch and colleagues37 suggested
that there was no significant difference between groups
on BI. Because Lynch and colleagues obtained unpub-
lished information from requesting, there is lack of
feasibility to assess reliability of data and repeat their meta-
analysis. The former also reported that there was no
significant difference between groups regarding mRS score
0-2 and the incidence of complications, which are com-
parable with this meta-analysis. When contrasting the
sample sizes of 2 meta-analyses (Lynch et al 159 versus
this study 2706), the CI of effect value in the present study
was less narrow and accuracy was increased. Even without
statistical significance, both the 2014 meta-analysis and
9
this study observed a larger proportion of death among
EM testers, and it is probable because of upright posi-
tion and head position changing, which affect blood
pressure and cerebral blood flow velocity a lot. Accord-
ing to a high-quality RCT,38 blood pressure reduction after
stroke could prevent further vascular events. Mean-
while, increased blood pressure and cerebral blood flow
might lead to further bleeding, especially among pa-
tients treated with recombinant tissue-type plasminogen
activator and patients with intracranial hemorrhage.39
Quality of included studies was good, and the analyses
failed to observe substantial heterogeneities except on BI.
Although there is a significant heterogeneity for BI, similar
findings through sensitivity analyses indicate that our result
is robust.
To the best of our knowledge, this is the currently best
evidence on this topic. Although preventing complica-
tions (e.g., immobility-related complications and further
changes from stroke-related inactivity, such as falls, deep
vein thromboembolism, loss of cardiovascular fitness,
muscle atrophy21,40) and improving prognosis were the
main rationale for EM to be recommended by many
guidelines,34,41 we failed to find favorable evidence sup-
porting the guidelines. Unexpectedly, significant effects
of EM on self-care ability and hospital stay were firstly
confirmed. Effects of EM were roundly assessed in this
meta-analysis with comprehensive measurements. More-
over, GRADE was performed to judge the quality of
evidence and its results appeared to be pretty good. This
meta-analysis also has several limitations. First, we did
not discuss dose-response effect for optimal timing, fre-
quency, and duration of EM because of insufficiency of
initial data. Second, differences on intervention proto-
cols were inevitable. Third, there is a small amount of
included studies on each outcome.
Conclusions
In conclusion, current evidence showed no differ-
ences between EM and control groups on mortality,
 10
Table 3. Grade evidence profile

Quality assessment Summary of Findings

Study event Anticipated

rates (%) absolute effects

Risk difference

ARTICLE IN PRESS
Participants Overall Relative with early
(studies) Risk of Publication quality of With With early effect Risk with mobilization
follow-up bias Inconsistency Indirectness Imprecision bias evidence control mobilization (95% CI) control (95% CI)

A favorable outcome (CRITICAL OUTCOME; assessed with: modified Rankin Scale (0-2))
2373 (5 studies) No serious Serious No serious No serious Undetected ⊕⊕⊕⊝ 621/1193 563/1180 RR .80 521 per 1000 104 fewer per 1000
3 mo risk of indirectness imprecision MODERATE (52.1%) (47.7%) (.58-1.02) (from 219 fewer to
bias because of 10 more)
inconsistency
Mortality (CRITICAL OUTCOME; assessed with: death)
2446 (4 studies) No serious Serious No serious No serious Undetected ⊕⊕⊕⊕ HIGH 82/1232 99/1214 RR 1.21 67 per 1000 14 more per 1000
3 mo risk of indirectness imprecision because of (6.7%) (8.2%) (.76-1.65) (from 16 fewer to
bias inconsistency, 43 more)
plausible
confounding would
change the effect
Self-care ability (IMPORTANT OUTCOME; measured with: Barthel Index; range of scores: 0-100; Better indicated by higher values)
303 (4 studies) No serious Serious No serious Serious Undetected ⊕⊕⊕⊝ 154 149 — The mean self- The mean self-care
3 mo risk of indirectness MODERATE care ability in ability in the
bias because of the control intervention groups
inconsistency, groups was was .66 standard
imprecision, large .66 SMD deviations higher
effect (.00-1.31 higher)

L. ZHUYUE ET AL.
 ARTICLE IN PRESS
EARLY MOBILIZATION AND ACUTE STROKE
complications, and mRS score 0-2. In addition, we found
an interesting finding that EM could improve self-care
ability of patients post stroke and promote earlier dis-
charge. In this highly focused and controversial field,
further research is essential to explore the effect of EM
post stroke. Future research could focus on dose-
response study and confirm the optimal timing, frequency,
and duration to perform EM.
Appendix: Supplementary Material
Supplementary data to this article can be found online
at doi:10.1016/j.jstrokecerebrovasdis.2017.12.021.
References
1. Feigin VL, Lawes CM, Bennett DA, et al. Stroke
epidemiology: a review of population-based studies of
incidence, prevalence, and case-fatality in the late 20th
century. Lancet Neurol 2003;2:43-53.
2. Krishnamurthi RV, Feigin VL, Forouzanfar MH, et al.
Global and regional burden of first-ever ischaemic and
haemorrhagic stroke during 1990-2010: findings from the
Global Burden of Disease Study 2010. Lancet Glob Health
2013;1:e259-e281.
3. Buntin MB, Colla CH, Deb P, et al. Medicare spending
and outcomes after postacute care for stroke and hip
fracture. Med Care 2010;48:776-784.
4. Winstein CJ, Stein J, Arena R, et al. Guidelines for adult
stroke rehabilitation and recovery: a guideline for
healthcare professionals from the American Heart
Association/American Stroke Association. Stroke
2017;48:e78.
5. Hodgson C, Bellomo R, Berney S, et al. Early mobilization
and recovery in mechanically ventilated patients in the
ICU: a bi-national, multi-centre, prospective cohort study.
Crit Care 2015;19:81.
6. Guerra ML, Singh PJ, Taylor NF. Early mobilization of
patients who have had a hip or knee joint replacement
reduces length of stay in hospital: a systematic review.
Clin Rehabil 2015;29:844-854.
7. Bernhardt J, Langhorne P, Lindley RI, et al. Efficacy and
safety of very early mobilisation within 24 h of stroke
onset (AVERT): a randomised controlled trial. Lancet
2015;46-55.
8. Herisson F, Godard S, Volteau C, et al. Early sitting in
ischemic stroke patients (SEVEL): a randomized controlled
trial. PLoS ONE 2016;11:e0149466.
9. Chippala P, Sharma R. Effect of very early mobilisation
on functional status in patients with acute stroke:
a single-blind, randomized controlled trail. Clin Rehabil
2016;30:669-675.
10. Poletto SR, Rebello LC, Valens MJM, et al. Early
mobilization in ischemic stroke: a pilot randomized trial
of safety and feasibility in a public hospital in Brazil.
Cerebrovasc Dis Extra 2015;31-40.
11. Sundseth A, Thommessen B, Rønning OM. Early
mobilization after acute stroke. J Stroke Cerebrovasc Dis
2014;496-499.
12. Liu N, Cadilhac DA, Andrew NE, et al. Randomized
controlled trial of early rehabilitation after intracerebral
hemorrhage stroke: difference in outcomes within 6
months of stroke. Stroke 2014;3502-3507.
11
13. Diserens K, Moreira T, Hirt L, et al. Early mobilization
out of bed after ischaemic stroke reduces severe
complications but not cerebral blood flow: a randomized
controlled pilot trial. Clin Rehabil 2012;26:451-459.
14. Cumming TB, Thrift AG, Collier JM, et al. Very early
mobilization after stroke fast-tracks return to walking:
further results from the phase II AVERT randomized
controlled trial. Stroke 2011;153-158.
15. Langhorne P, Stott D, Knight A, et al. Very early
rehabilitation or intensive telemetry after stroke: a pilot
randomised trial. Cerebrovasc Dis 2010;29:352-360.
16. Sorbello D, Dewey HM, Churilov L, et al. Very early
mobilisation and complications in the first 3 months after
stroke: further results from phase II of A Very Early
Rehabilitation Trial (AVERT). Cerebrovasc Dis 2009;28:378-
383.
17. Bernhardt J, Dewey H, Thrift A, et al. A very early
rehabilitation trial for stroke (AVERT): phase II safety and
feasibility. Stroke 2008;39:390-396.
18. Craig LE, Bernhardt J, Langhorne P, et al. Early
mobilization after stroke: an example of an individual
patient data meta-analysis of a complex intervention.
Stroke 2010;41:2632-2636.
19. Jauch EC, Saver JL, Adams HP Jr, et al. Guidelines for
the early management of patients with acute ischemic
stroke: a guideline for healthcare professionals from the
American Heart Association/American Stroke Association.
Stroke 2013;44:870-947.
20. Quinn TJ, Paolucci S, Sunnerhagen KS, et al. Evidence-
based stroke rehabilitation: an expanded guidance
document from the European Stroke Organisation (ESO)
guidelines for management of ischaemic stroke and
transient ischaemic attack 2008. J Rehabil Med 2009;41:99-
111.
21. Allen C, Glasziou P, Del Mar C. Bed rest: a potentially
harmful treatment needing more careful evaluation. Lancet
1999;354:1229-1233.
22. Bernhardt J, Thuy MNT, Collier JM, et al. Very early
versus delayed mobilisation after stroke. United Kingdom:
John Wiley and Sons Ltd (Southern Gate, Chichester, West
Sussex PO19 8SQ, United Kingdom), 2009 p. no
pagination.
23. van Wijk R, Cumming T, Churilov L, et al. An early
mobilization protocol successfully delivers more and
earlier therapy to acute stroke patients: further results
from phase II of AVERT. Neurorehabil Neural Repair
2012;26:20-26.
24. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting
items for systematic reviews and meta-analyses: the
PRISMA statement. Ann Intern Med 2009;151:264-269,
w64.
25. Collaboration TC. Cochrane handbook for systematic
reviews of interventions version 5.1.0 [updated March
2011]. Cochrane Collab, 2011.
26. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines:
1. Introduction-GRADE evidence profiles and summary
of findings tables. J Clin Epidemiol 2011;64:383-394.
27. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean
and variance from the median, range, and the size of a
sample. BMC Med Res Methodol 2005;5:13.
28. Higgins J, Thompson S, Deeks J, et al. Measuring
inconsistency in meta-analysis. BMJ 2003;327:4.
29. Begg C, Mazumdar M. Operating characteristics of a rank
correlation test for publication bias. Biometrics 1994;50:14.
30. Egger M, Davey Smith G, Schneider M, et al. Bias in
meta-analysis detected by a simple, graphical test. BMJ
1997;315:6.
 ARTICLE IN PRESS
12
31. Higgins JP, Thompson SG. Controlling the risk of spurious
findings from meta-regression. Stat Med 2004;23:1663-1682.
32. DerSimonian R, Laird N. Meta-analysis in clinical trials.
Control Clin Trials 1986;7:12.
33. Glober NK, Sporer KA, Guluma KZ, et al. Acute stroke:
current evidence-based recommendations for prehospital
care. West J Emerg Med 2016;17:104-128.
34. Winstein CJ, Stein J, Arena R, et al. Guidelines for adult
stroke rehabilitation and recovery: a guideline for
healthcare professionals from the American Heart
Association/American Stroke Association. Stroke
2016;47:e98-e169.
35. Langhorne P, Wu O, Rodgers H, et al. Very early
rehabilitation trial after stroke (AVERT): a Phase III,
multicentre, randomised controlled trial. Health Technol
Assess 2017;21:1-120.
36. Hellstrom K, Lindmark B, Wahlberg B, et al. Self-efficacy
in relation to impairments and activities of daily
L. ZHUYUE ET AL.
living disability in elderly patients with stroke:
a prospective investigation. 5th ed. Norway: Foundation
for Rehabilitation Information (Tradardsgatan 14, Uppsala
SE-753 09, Sweden), 2003:202-207.
37. Lynch E, Hillier S, Cadilhac D. When should physical
rehabilitation commence after stroke: a systematic
review. 4th ed. United Kingdom: Blackwell Publishing
Ltd, 2014:468-478.
38. Group PC. Randomised trial of a perindopril-based
blood-pressure-lowering regimen among 6,105 individuals
with previous stroke or transient ischaemic attack. Lancet
2001;358:1033-1041.
39. Saver JL. Blood pressure management in early ischemic
stroke. JAMA 2014;311:469-470.
40. Ivey FM, Hafer-Macko CE, Macko RF. Exercise
rehabilitation after stroke. NeuroRx 2006;3:439-450.
41. Party tISW. National clinical guideline for stroke. 5th ed.
Royal College of Physicians, 2016.