Roesner et al.

Systematic Reviews (2024) 13:270 Systematic Reviews

https://doi.org/10.1186/s13643-024-02676-0

SYSTEMATIC REVIEW UPDATE Open Access

Effects of physical therapy modalities

for motor function, functional recovery,
and post‑stroke complications in patients
with severe stroke: a systematic review update
Katrin Roesner1,2* , Bettina Scheffler3, Martina Kaehler4, Bianca Schmidt‑Maciejewski5, Tabea Boettger6 and
Susanne Saal7,8

Abstract
Background Physical therapy interventions play a crucial role in the daily care of patients recovering from severe
stroke. However, the efficacy of these interventions and associated modalities, including duration, intensity, and fre‑
quency, have not been fully elucidated. In 2020, a systematic review reported the beneficial effects of physical therapy
for patients with severe stroke but did not assess therapeutic modalities. We aim to update the current evidence
on the effects of physical therapy interventions and their modalities in relation to the recovery phase in people
with severe stroke in a hospital or inpatient rehabilitation facility.
Methods We searched CENTRAL, MEDLINE, Web of Science, and three other relevant databases between Decem‑
ber 2018 and March 2021 and updated the search between April 2021 and March 2023. ClinicalTrials.gov and ICTRP
for searching trial registries helped to identify ongoing RCTs since 2023. We included individual and cluster rand‑
omized controlled trials in the English and German languages that compared physical therapy interventions to similar
or other interventions, usual care, or no intervention in a hospital or rehabilitation inpatient setting. We screened
the studies from this recent review for eligibility criteria, especially according to the setting. Critical appraisal was per‑
formed according to the Cochrane risk-of-bias tool 2.0. The data were synthesized narratively.
Results The update identified 15 new studies, cumulating in a total of 30 studies (n = 2545 participants) meet‑
ing the eligibility criteria. These studies reported 54 outcomes and 20 physical therapy interventions. Two studies
included participants during the hyperacute phase, 4 during the acute phase,18 during the early subacute phase,
and 3 in the late subacute phase. Three studies started in the chronic phase. Summarised evidence has revealed
an uncertain effect of physical therapy on patient outcomes (with moderate to low-quality evidence). Most stud‑
ies showed a high risk of bias and did not reach the optimal sample size. Little was stated about the standard care
and their therapy modalities.
Discussion There is conflicting evidence for the effectiveness of physical therapy interventions in patients
with severe stroke. There is a need for additional high-quality studies that also systematically report thera‑
peutic modalities from a multidimensional perspective in motor stroke recovery. Due to the high risk of bias

*Correspondence:
Katrin Roesner
katrin.roesner@uni-luebeck.de
Full list of author information is available at the end of the article

© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/.
Roesner et al. Systematic Reviews (2024) 13:270 Page 2 of 25

and the generally small sample size of the included studies, the generalizability of the findings to large and heteroge‑
neous volumes of outcome data is limited.
Systematic review registration PROSPERO CRD42021244285.
Keywords Physical therapy, Effects, Dose, Severe stroke, Systematic review update

Introduction an important determinant of the length of hospital stay

Stroke is the second leading cause of death and the third and is one of the most important indices for measuring
leading cause of death and disability worldwide accord- the use of hospital care [19]. Several predictors of stroke
ing to the Global Burden of Diseases Study Group [1]. recovery have been analyzed thus far, but therapy modal-
An increase in stroke deaths and disability-adjusted life ities, including duration, frequency, and intensity, have
years (DALYs) will accompany future population growth not been considered.
and increased life expectancy in many countries [2, 3]. An important factor in understanding study outcomes
Simultaneously, stroke care has been optimized through can be the therapeutic modality of an intervention [20].
specialized acute facilities (e.g., stroke units), and the In the literature, there is a lack of definitions and het-
advancement of recanalization therapies (i.e., thrombec- erogeneous use of the dose of intervention. In 2021, the
tomies and thrombolysis) has shown positive effects [4]. group around Hayward [21] proposed a framework for
In 2017, the Stroke Recovery and Rehabilitation Round- dose articulation in stroke recovery: duration can meas-
table Taskforce agreed on a standard definition of stroke ure the length of intervention in days, sessions per day,
phases [5]. This decision is based on current knowledge and session length with active or inactive episodes; an
and understanding of the biological repair processes of episode can be defined by the length of a task, its diffi-
the brain [6–8]. In animal models, a time dependency of culty, and its intensity. For this reason, among others,
recovery has already been observed [6]. It is hypothesized studies have been conducted to examine the influence of
that the biological repair process begins early after the duration and intensity of therapy and have shown posi-
stroke and then slowly subsides. A similar course has also tive treatment effects for a longer therapy duration. For
been observed in humans [6, 9]. example, depending on the author, the therapy time var-
However, this concept of stroke phases does not imply ied between upper extremity interventions post-stroke
that functional recovery occurs linearly; rather, it is of 10 h/week, as measured by the Motor Activity Log
meant to ensure better comparability between research (MAL) [22], and 30 h/week, as measured by the Fugl-
data sets. The proportion recovery rule (PRR) follows Meyer-Assessment Upper Limb (FMA-UL) [23]. Lit-
the hypothetical assumption that within the first 3–6 tle information is available about studies of people with
months, patients can improve on average by 70% (± 15%), severe stroke and associated therapy modalities, such as
and thus, the extent of recovery is highly predictable, duration and episodes.
regardless of the dose of therapy provided [10–12]. Some McGlinchey et al. (2020) [24] reported the beneficial
patients do not follow this rule, especially those who effects of rehabilitation interventions for severely affected
show severe deficits at baseline [13]. Recent approaches patients with stroke according to their stroke recovery
indicate that most people with stroke tend to experience phase and outcome measures according to the Interna-
a combination of constant recovery and proportional- tional Classification of Functioning, Disability and Health
to-spared function [14]. It was found that there are two (ICF). His review did not address therapy modalities
different patterns of recovery for patients with stroke and included all kinds of settings. This review builds on
and different severities of initial motor impairment [15]. and expands upon McGlinchey’s systematic review [24]
Severely and non-severely impaired patients exhibit indi- by additionally evaluating the duration and episodes of
vidual recovery trajectories [15]. interventions in people after a severe stroke in a hospital
Reports on patients with severe stroke indicate that or inpatient rehabilitative setting.
even after intensive and prolonged therapy, some patients
may show little to no improvement [16]. According to a Methods
systematic review, people with severe stroke often pro- This systematic review was conducted following the rec-
gress more slowly and with less functional improvement ommendations of the Cochrane Handbook for System-
during inpatient rehabilitation than people with mild atic Reviews of Interventions, version 6.3 [25] and was
impairment [17]. This could be due to the combina- reported according to the Preferred Reporting Items for
tion of significant sensory-motor and cognitive impair- Systematic Reviews and Meta-Analysis Statement 2020
ments induced by severe stroke [18]. Stroke severity is [26]. The protocol was registered prospectively at the
Roesner et al. Systematic Reviews (2024) 13:270 Page 3 of 25

Prospective Register of Systematic Reviews platform Studies were excluded if.

(CRD42021244285). There were no deviations from the
protocol. (1) patients suffered a mild to moderate stroke.
(2) Pharmacological, surgical, or complementary (e.g.,
Eligibility criteria acupuncture or non-invasive brain stimulation)
Studies were included when. interventions were used.
(3) Interventions were delivered in an outpatient set-
(1) they were a randomized or non-randomized ting (e.g., home or ambulatory).
controlled trials. A randomized controlled trial
can be an individually randomized trial (RCT), a Search strategy and selection criteria
cluster randomized controlled trial (cRCT) ran- Since the review builds up on the systematic review of
domized crossover trial, or a multicenter rand- McGlinchey et al. [24] a search was conducted for studies
omized study. A non-randomized controlled trial published between January 1987 and November 2018),
can be mono- or multicentric, with a prospec- and the search strategy was based on their report. For
tive study design. RCTs are considered the gold this reason, we have not consulted any other experts.
standard for demonstrating efficacy. Electronic searches were also conducted in the Medical
(2) Study participants were aged ≥ 18 years and Literature Analysis and Retrieval System Online (MED-
diagnosed with severe stroke (ischemic or hem- LINE), Excerpta Medica Database (EMBASE), Cumu-
orrhagic). For the purpose of this review, severe lative Index to Nursing and Allied Health Literature
stroke was defined as Functional Independ- (CINHAL), Physiotherapy Evidence Database (PEDro),
ence Measure (FIM) score < 54 or early FIM < 40 Web of Science, Database of Research in Stroke (DORIS)
or National Institute of Health Stroke Scale and the Cochrane Central Register of Controlled Trials
(NIHSS) ≤ 16 or modified Rankin Scale (mRS) ≤ 4 (CENTRAL) from December 2018 to March 2021 and
or Barthel Index (BI) ≤ 35 or Fugl-Meyer-Assess- an update from April 2021 to March 2023. The search
ment (FMA) ≤ 55 or Functional Ambulatory Cat- strategy included identifying the terms related to severe
egories (FAC) ≤ 2 [27–32]. stroke, physical therapy modalities, stroke rehabilitation,
(3) The placement of the intervention was within a motor recovery, and motor function. Information spe-
hospital or inpatient rehabilitation facility. cialists were consulted. An example of the search strategy
(4) Any kind of physical rehabilitation interven- can be found in Supplementary material S1. The search
tion was applied with the intention of managing was restricted to human studies. The International Trials
physical problems following a medical or surgical Registry Platform (ICTRP) and ClinicalTrials.gov helped
condition. to identify ongoing trials from January 2019 to March
(5) A comparison intervention was reported, includ- 2021 and were updated from April 2022 to March 2023.
ing any other type of physical rehabilitation or To identify further studies, reference lists were searched,
usual care. and a forward citation search was carried out on the Web
(6) At least one of the following outcomes was of Science, Google Scholar, and Scopus.
reported:
(7) Functional recovery was defined as a partial or Data collection and analysis
complete return to normal or proper physiologic Study selection
activity of an organ or body part following dis- All eligible studies were uploaded to a reference manage-
ease or trauma and was assessed using a recog- ment program. After duplicate removal, the remaining
nized outcome measure of functional ability or studies were uploaded to Rayyan [37]. Using the pre-
activities of daily living (ADLs) [33]. specified criteria for eligibility, two authors (KR, BSM)
(8) Motor function was defined as any activity of independently screened the studies for inclusion based
muscles due to stimulation by a motor neuron, on their titles and abstracts. Two independent reviewers
movement, or activation and was assessed using (KR, MK) screened the full texts of the studies identified
a recognized measure of motor function [34]. as eligible for inclusion during title and abstract screen-
(9) Post-stroke complications, adverse events, and ing. The reasons for the excluded studies are listed in a
medical or neurological problems necessitate table (Supplementary material S2). Reports from the
a physician’s order and require monitoring by same study population were linked to ensure that data
medical staff [35, 36]. from the same population were only included once in the
(10) The study was published in English or German review and analysis. Discrepancies were discussed and
to ensure feasibility. resolved by consensus with a third reviewer (SuS).
Roesner et al. Systematic Reviews (2024) 13:270 Page 4 of 25

Data extraction and management the intervention was delivered. In addition, further dose
Two review authors (KR, TB) independently extracted dimensions such as session density and episodes were
the data from all the included studies. A prespecified data extracted [42].
extraction form was developed based on the Cochrane
Handbook for Systematic Review of Interventions (ver-
Data analysis and synthesis
sion 6.3), the CONSORT statement for reporting ran-
Demographics and study results are reported as medi-
domized trials and extensions, and the Template for
ans (IQRs), minimum to maximum ranges, or num-
Intervention Description and Replication (TiDieR) [25,
bers of studies (percentages) as appropriate. Due to the
38, 39]. The data extraction form was pilot-tested dur-
heterogeneity of outcome measures, recovery phases,
ing two online face-to-face training sessions. The review
therapy modality outcomes, and the high proportion
by McGlinchey et al. [24] did not report the therapeutic
of studies where concerns regarding bias were present,
modalities; therefore, we screened all the studies accord-
pooled analyses were not performed. The extracted
ing to the setting, stroke severity cut-off points, and
data are summarised in tables as narrative descriptions
interventions. The following information was extracted:
of the intervention and therapy modalities by recovery
aim and focus of the studies, study design, details about
phase. If various outcomes were reported, the means
the intervention according to the TiDieR Checklist [39],
and standard deviations, including participant charac-
number and characteristics of participants, time post-
teristics and test results, were combined into one group
stroke (converted in days), outcomes, and outcome meas-
using the free accessible Statistics Toolkit (STATTOOLS,
ures. The individuals in each study were assigned to the
Palisade, Ithaca, NY) [43]. This procedure of combining
following post-stroke recovery phases: hyperacute (≤ 24
means (SDs) complies with the recommendations of the
h post-stroke), acute (> 24 h but ≤ 7 days post-stroke),
Cochrane Handbook for Systematic Reviews of Interven-
early subacute (> 7 days but ≤ 3 months (≤ 90 days) post-
tions (version 6.3) [25]. To judge the quality of evidence
stroke) and late subacute (> 3 months but ≤ 6 months
narratively, the Grading of Recommendations, Assess-
(≤ 180 days) post-stroke) [4]. Funding sources for the
ment, Development and Evaluation (GRADE) approach
studies were collected in tabular form (Supplementary
was used [44, 45].
material S3).

Results
Risk of bias Study selection
Two authors (BS, KR) independently rated the risk of Out of 3216 identified records, 564 full-title articles were
bias using the Cochrane Risk of Bias 2 (RoB2) tool for screened. The search was conducted between the 11th
individually randomized, parallel-group trials [40] and and 23rd of April 2021 and the 1st and 10th of April
recommended according to the “SHORT VERSION 2023. This review included 30 studies, 15 of which were
(CRIBSHEET)” [41]. An overall judgment of a high risk also included in McGlinchey’s analysis [46–69]. However,
of bias was given when the study was judged to be at high 13 studies included by McGlinchey were excluded in
risk of bias in one domain or if there was some concern this current review due to an incorrect setting, interven-
for multiple, in this review, two domains. Any differences tion, or different cut-offs for the severity of stroke. Two
in opinion were discussed between the two authors and reports [47, 48] were subanalyses from the AVERT study
were recorded in writing. There was no need to consult a [49], and two [50, 51] were follow-up studies to Kwakkel
third reviewer. McGlinchey [24] stated in his review that et al. (1999 [52]). No further results were found through
a high quality meant a low risk of bias, a moderate qual- searching reference lists or forward citations in the Web
ity a serious risk and a low to very low quality of evidence of Science, Google Scholar, or Scopus databases. Figure 1
was based on a high risk of bias. shows the results of the screening procedure.

Study characteristics
Reporting on intervention and dose Study characteristics, including the interventions used,
Extraction according to the TiDieR checklist [39] of the are provided in Table 1. All the studies were published
goal of the interventions included who and what was between 1999 [52, 53] and 2023 [54] and were mostly
provided, whether it was tailored or modified, how well conducted in Italy [55–58], Korea [46, 59–64], and China
it was planned, and whether there was economic infor- [54, 65–69]. A total of 2545 participants were included
mation available. Item eight of the TiDieR instrument across all the RCTs (range: n = 20 [70] to n = 294 [49]).
evaluates the number of times the intervention was deliv- The overall study duration ranged between one [71] and
ered and the duration, intensity, or dose during which 20 weeks [52].
Roesner et al. Systematic Reviews (2024) 13:270 Page 5 of 25

Fig. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analysis 2020 study selection flow diagram

Participant characteristics had a low risk of bias [49], four studies had a serious risk
The participants’ characteristics are summarised in [50–52, 72, 76, 79], and ten had a high risk of bias [53,
Table 2. Thirty studies reported on 2545 randomized 56–58, 61, 67, 68, 70, 73, 78].
patients with an average age of 67.25 years (SD ± 13.50).
Two studies [49, 57] included participants in the hyper- Types of interventions
acute phase, four studies [52, 67, 71, 79] in the acute The included studies used various types of interventions.
phase, 18 studies [53–56, 58, 60, 61, 64, 66, 68–70, 72– Studies have used active rehabilitation interventions
78] in the early subacute phase and three studies in the without technical support or devices, such as mirror
late subacute phase [59, 63, 65]. Three studies started in therapy [60, 62, 69], additional upper or lower limb ther-
the chronic phase [46, 62, 70]. apy [52], very early mobilization [49, 57], and interven-
tions with specialized therapists in the Bobath or Carr
Risk of bias and Shepard approaches [67, 68, 71]. Robotic-assisted
The results of the risk of bias assessment are shown in body weight supported (BWS) treadmill gait training
Figure 2. The individually randomised, parallel-group [61, 70, 76, 77], underwater gait training [46], BWS-sup-
studies[46, 55, 60, 62, 63, 65, 66, 69, 71, 80], showed an ported overground gait training [55, 75], BWS-supported
overall high risk of bias. All ten studies raised some con- treadmill [58], or gait training and leg cycle ergometry
cerns or were rated as having a high risk of bias in the [79] were used.
domains of deviation from intended interventions [46, Three studies used electrical stimulation, such as func-
62, 65, 69, 80], randomisation process [62, 71] or miss- tional electrical stimulation (FES) [66] and neuromus-
ing or retrospective study registration [55]. According to cular electrical stimulation (NMES) [65, 66]. Supportive
the quality assessments of McGlinchey’s 15 studies, one devices for verticalization, such as a standing frame [73,
Table 1 Result
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Global early mobilisation within 24 h

AVERT Trail Collabora‑ Parallel-group multi‑ n= 294 NIHSS Intervention PTs and nurses Favourable outcome Subgroup analysis
tion Group, 2015 [49]; centre RCT​ No information very early mobilisation (mRS 0-2) and mortality for severe stroke
Bernhardt et al., 2021 Australia, New Zealand, for severe stroke NIHSS >16 (n=147) at 3 months patients concern‑
Roesner et al. Systematic Reviews

[5], Cain et al., 2022 Malaysia, Singapore, patients Control ing mRS favours usual
[48, 47] and the UK usual care care over early mobi‑
NIHSS >16 (n=147) lisation with an Odds
Ratio of 0.35 (0.11–1.18).
88 (8%) patients died
in the very early mobilisa‑
(2024) 13:270

tion group and 72 (7%)

patients in the usual care
group.
Di Lauro et al., 2003 [57] Individual RCT​ n= 60 BI* Intervention Therapists and nursing BI, mNIHSS No differences
Italy Intervention Intensive rehab group staff between groups in BI
69,30 (±8,0) Mean BI 1.4 (±1.4) at 180 days (IG: 8.0±2.8;
♀ 62% (n=29) CG: 7.7±3.0; p>0.7)
Control Control or NIHSS at 14 days (IG:
67,6 (±9,3) Ordinary rehab group 8.1±3.0; CG: 8.4±2.6;
♀ 55% (n=31) p>0.6) and at 180 days
Mean BI 1.5 (±1.5) (IG: 6.2±2.8; CG: 6.5±2.7;
p>0.7).
Electrical Stimulation
Guo and Kang, 2018 Individual RCT​ n= 82 BI Intervention not described BI, ICIQ-SF, OBASS, Uro‑ The results of the inter‑
[65] China Intervention NMES (n=38) Mean BI dynamic outcome vention group are more
64.3 (±11.8) 10,5 (±2,2) promising than the con‑
♀ 46% Control trol group results in all
Control sham NMES (n=36) outcomes:
62.5 (±12.2) Mean BI 10,8 (±2,6) Urodynamic value
♀ 39% (p<0,01),
OABSS: IG: 8.1±3.4; CG:
12.3±3.0; p<0.01,
c: IG 7.8±3.3; CG: 10.5±3.1;
p<0.01,
BI: IG:15.7±3.1; CG:
11.1±3.4; p<0.01.
Page 6 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Rosewilliam et al., 2012 Individual RCT​ n=90 BI^ Intervention NMES- staff group ARAT, BI, wrist AROM, No differences in ARAT,
[72] USA entire sample NMES group (n=31) not reported, patients wrist strength, grip BI or wrist AROM
74,6 (±11,0) Mean BI 4.4 (±3.9) and carers, Usual care- strength between groups.
♀ 56% Control PTs Improvements in wrist
Roesner et al. Systematic Reviews

Usual care group (n=36) extensor and grip

Mean BI 2.5 (±2.9) strength in the NMES
group post-intervention
but not maintained
at follow-up.
Zheng et al., 2018 [66] Individual RCT​ n= 60 mBI Intervention PTs mBI, FMA, PASS, BBS, In favour of four-channel
(2024) 13:270

Intervention four-channel FES group BBA, mBI, fMRI FES group showed
59 (±11) (n=18) Mean mBI 22 fractional anisotropy
♀ 50% (±9) and increased fibre
Control 1 Control 1 bundles. No significance
60 (±9) dual-channel FES group between-group differ‑
♀ 40% (n=15) Mean mBI 23 ences.
Control 2 (±13)
59 (±9) Control 2
♀ 40% placebo group (n=15)
Mean mBI 24 (±13)
Mirror therapy
Cui et al., 2022 [69] Individual RCT​ n=32 mBI Intervention Therapists FMA-LE, BBS, mBI, A better effect
China Intervention mirror therapy on lower mRMI, rs-fMRI for the mirror therapy
61.5±9.93 leg (n=16), range mBI group concerning FMA-
♀ 44% 21.50 (20.00, 25.75) LE (Z= -4,526,p<0,01),
Control Control BBS (F = 36.985, p < 0.01),
58.5±11.15 routine rehabilitation mMRI (F = 27.171, p <
♀ 50% (n=16), range mBI 22.50 0.01), mBI (F = 9.830, p =
(10.00, 27.75) 0.004).
Lee et al., 2020 [60] Individual RCT​ n=21 K-mBI Intervention OTs K-mBI, FMA, MAL A favourable outcome
Korea Intervention multi-joint-based mirror for improving upper
50.91 (±8.73) therapy (n=11), Mean limb function (FMA-UE)
♀ 38% K-mBI 23,73(± 7,70) and ADL in IG compared
Control Control tcontrolrl.
61.5 (±9.93) single-joint based FMA-UE: IG:26.36±11.75;
♀ 28% mirror therapy (n=10), CG: 16.00 ±8.7; p= 0.034
Mean K-mBI 18,80 (± MAL-QoM: IG: 20.91±12.8;
7,22) CG: 12 ±6.2; p=0.034
MAL-AOU: IG: 17.64±72.8;
CG: 11.4±6; p=0.048
mBI: IG: 26.55±5.71; CG:
19,4±7.18; p= 0.031
Page 7 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Sim and Kwon, 2022 Individual RCT​ n=30 K-mBI Intervention OTs MVPT, K-MMSE, BIT, In favour of the interven‑
[62] Korea Intervention bimanual mirror K-CBS, K-mBI, SCT, LBT tion group was found
69.29 (±8.02) therapy (n=14), Mean for SCT (p<0,05), PST
♀ 33% K-mBI 35.64 (±16.08) (p<0,05), for LBT (p<0,05)
Roesner et al. Systematic Reviews

Control Control and for K-CBS (p<0,05).

69.14 (±6.92) unimanual mirror No differences were
♀ 36% therapy (n=14), Mean found for K-mBI.
K-mBI 30.92 (±11.57)
Neurodevelopmental techniques
Bai et al., 2014 [67] Individual RCT​ n=165 mBI Intervention PTs and OTs mBI, mAS IG demonstrated higher
(2024) 13:270

China Intervention 1.-3. month of staged mBI scores than the rou‑
67,63 (±9,52) rehabilitation group tine care group at 1,
♀ 38% (n=83) Mean BI 28 3- and 6 months post-
Control (range 24-31) stroke. IG: (M1 vs. M0,
66.04 (±10,13) Control M3 vs. M1, M6 vs. M3,
♀ 38% Routine care group p< 0.01); CG: (M1 vs. M0,
(n=82) Mean BI 23 p <0.01; M3 vs. M1, p =
(range 19-27) 0.026)
42.9% of patients
in the CG demonstrated
spasticity in at least one
body part compared
to 36.4% of patients
in the staged rehab
group.
Rahayu et al., 2020 [71] Individual RCT​ n=64 BI Intervention Research Assistant BI, BNDF-Biomarker, BBS In favour of intervention
Indonesia Intervention Neurorestoration inter‑ group for functional
58.84 (±8.68) vention (n=32) Mean BI performance (BI: IG 67.
♀ 38% 25.81 (±15.77) 47 (58.99-75.94; CG:
Control Control 46.41 (37.77-55.04);
59.93 (±10.65) standard procedure p=0.008) and bal‑
♀ 50% (n=32) Mean BI 19,00 ance (BBS: IG: 28.38
(± 10.29) (21.74-35.01; CG:17.16
(12.62-21.69; p=0.016)
in between group dif‑
ference. No differences
in neuroplasticity regen‑
eration (p=0.07).
Page 8 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Tang et al., 2014 [68] Individual RCT​ n=48 STREAM, BBS Intervention PTs STREAM, BBS Improvements in STREAM
China Intervention Early contemporary (F (1, 46) = 11.7, η2 =
68,2 (±4,1) Bobath group (n=24) 0.203, p < .01) and BBS (F
♀ 29% Mean STREAM 1.4 (± (1, 46) = 35.4, ŋ2 = 0.435,
Roesner et al. Systematic Reviews

Control 1.0), Mean BBS 0 (± 0) p < .001) in the contem‑

66,9 (±4,1) Control porary Bobath approach
♀ 33% Contemporary group with early mobilisation
(n=24) Mean STREAM group.
1.3 (± 0.9), Mean BBS
0 (± 0)
(2024) 13:270

Interventions for verticalization

Bagley et al., 2005 [73] Individual RCT​ n= 140 BI Intervention PTs and nurses RMI, BI, HADS NEADL, No differences
UK Intervention Oswestry group (n=71) RMA, MAS (balance, sit- between groups for all
75,8 (±11,5) Median BI 1 (IQR 0-3) to-stand sections), TCT, outcome measures. No
♀ 29% Control CSI, GHQ-28 difference in the number
Control Control group (n=69) of treatment sessions
75,1 (±9,4) Median BI 2 (IQR 1-3) or stuff required for treat‑
♀ 31% ment.
Calabró et al., 2015 [56] Individual RCT​ n=32 PASS, FMA-LL Intervention PTs PASS, FMA-LL, RCPM, Both interventions
Italy Intervention Robotic verticalization MRC, vertical posture were well, tolerated.
71 (±3) group (n=10) Mean tolerance The robotic group
♀ 60% PASS 3 (±1), Mean LL demonstrated greater
Control FMA 13 (±3) improvements compared
70 (±5) Control to the physiotherapy
♀ 50% Physiotherapy group group in:
(n=10) Mean PASS 3 MRC (IG: 2±1; CG: 1±1;
(±3), Mean LL FMA 12 p=0,03); FMA (IG: 92±10;
(±6) CG: 58±7; p=0.008)
and PASS (IG: 166±30;
CG: 66±2; p=0.008).
Between-group differ‑
ences for cognition were
measured with RCPM
of p=0.03.
Page 9 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Logan et al., 2022 [74] Individual RCT​ n= 45 mRS Intervention PTs Edmans ADL, BI, It was a feasibility trial.
UK Intervention Functional standing Goniometer, muscle It is not feasible in its
81.7 (±11.7) frame programme strength, MAS, TCT, current design. The
Roesner et al. Systematic Reviews

♀ 27% (n=22) VAS for fatigue, PHQ-9, intervention group

Control mRS 4= 17; mRS 5= 5 SADQ-10, EQ-5D 5L, showed some promis‑
78.9 (±10.5) Control Stroke and Aphasia QoL ing results on the BI
♀ 31% usual physiotherapy Scale-39 for example at the ­55th
(n=23) week with a Mean dif‑
mRS 4= 19; mRS 5= 4 ference (95% CI) of 0,86
(2024) 13:270

[-4.76, 6.49], and showed

over time a ≥ 1.85 point
minimal clinically impor‑
tant difference.
Gait training interventions without electrical support
Brunelli et al., 2019 [55] Individual RCT​ n= 37 FAC Intervention PTs FAC, RMI, BI 6MWT Patients in both groups
Italy Intervention BWS overground gait improved continu‑
69.64 (±10.88) training (n=16), Mean BI ously. No difference
Control 14.35 (±14.62) between groups (p>0.05)
72.05 (±10.08) Control in independence in walk‑
Overall ♀ 52% Gait training with‑ ing (FAC), or any second‑
out BWS (n=21), Mean ary outcome (p>0.05).
BI 14.42 (±15.72)
Kim et al., 2020 [46] Individual RCT​ n= 22 FAC Intervention PTs FAC, PASS, Balancia No favourable outcome
Korea Intervention underwater gait train‑ 2.0 program, GAITRite for PASS, Postural control
65.2 (11.9) ing (n=10) FAC<3 system between the groups
♀ 45% Control (p>0.05). The step
Control overground gait train‑ length difference
61.4 (10.9) ing (n=11) FAC <3 varied between groups,
♀ 18% increased in IG
and decreased in CG
(IG: 4.55±6.68; CG: -1.25
±3.56; p<0.05).
Page 10 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics
Robotic-assisted gait training
Chang et al., 2012 [61] Individual RCT​ n= 48 FAC, FMA-LL Intervention PTs FAC, LL MI, FMA-LL, Peak Between-group differ‑
Korea Intervention Robot-assisted group chang ences for intervention
55,5 (±12) (n=20) Mean FAC 0.5 group in FMA-LL (IG:
Roesner et al. Systematic Reviews

♀ 29% (±0.5) 22.7 ±5.7; CG 19.6±5.6;

Control Control p=0.037) and peak VO2
59,7 (±12.1) Conventional group (l/min, IG:1.23±0.44; CG:
♀ 29% (n=17) Mean FAC 0.4 1.11±0.46; p=0.025).
(±0.5) No improvements
in LL MI (IG: 56.2±11.0;
(2024) 13:270

CG: 53.5±12; p= 0.200)

and FAC (IG 0.5±0.5; CG:
1.4±0.8; p=0.232).
Francesschini et al., Individual multicentre n= 97 BI* Intervention PTs MI, TCT, mRS, BI, FAC, No differences
2009 [58] RCT​ Intervention Treadmill training group AS, LL proprioception, between groups.
Italy 65,5 (±12,2) (n=52) Median BI 6 (IQR 6MWT, 10MWT, BS, WHS
♀ 46% 3-9), Median FAC 0 (IQR
Control 0-0)
70,9 (±11,8) Control
♀ 51% Conventional group
(n=45) Median BI 5 (IQR
3-7), Median FAC 0 (IQR
0-0)
Page 11 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Louie et al., 2021 [75] Individual multicentre n= 36 FAC Intervention PTs FAC, 5MWT, 6MWT, No significant between-
RCT​ Intervention Exoskelet group (n=19), FMA-LL, BBS, MoCA, group differences for FAC.
Canada 59.6 (15.8) Median FAC 0 (0-1) SF-36 But significant effects
♀ 16% Control for the intervention
Roesner et al. Systematic Reviews

Control usual physiotherapy group on FMA-LL(as-

55.3 (10.6) group (n=17), Median treated adjusted group
♀ 41% FAC 0 (0-1) difference: 3.9, 95% CI
1.3–6.6, F(1,33) = 9.33,
p = 0.004; per-protocol
adjusted group differenci:
(2024) 13:270

3.7, 95% CI 0.9–6.5,

F(1,28) = 7.29, p = 0.01)
and MoCA (as-treated
adjusted group differ‑
ence: 2.1, 95% CI 0.6–3.7,
F(1,29) = 7.96, p = 0.009;
per-protocol adjusted
group difference: 2.0, 95%
CI 0.4–3.6, F(1,25) = 6.62,
p = 0.02). No further sig‑
nificant between-group
differences in secondary
outcomes.
Ochi et al., 2015 [76] Individual RCT​ n= 26 FIM mobility, FAC Intervention Robot-assisted gait FAC, FMA, LL mus‑ Robot-assisted gait train‑
Japan Intervention Robot-assisted treadmill training not reported, cle torque, 10MWT, ing group demonstrated
61,8 (±7,5) gait training group conventional gait train‑ FIM(mobility scores) greater improvements
♀ 15% (n=13) Median FAC 0 ing PTS in FAC (IG: 3 (3-4); CG: 3
Control (IQR 0-1), Median FIM (3-3); p=0.02) and peak
65,5 (±12,1) mobility 7 (IQR 6-10) LL muscle torque (IG: 0.37
♀ 31% Control (0.2-0.52); CG: 0.18 (0.09-
Conventional group 0.23; p=0.05compared
(n=13) Median FAC 1 to the conventional
(IQR 0-1), Median FIM group.
mobility 7 (IQR 7-9)
Page 12 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Rodrigues et al., 2017 Individual RCT​ n=20 FMA-LL, FAC Intervention Not reported FAC, TUG, 6MWT, Improvements in FAC,
[70] USA Intervention Robot-assisted BWS 10MWT, BBS, FMA-LL FMA-LL, TUG and 6MWT
59,3 (±13,8) treadmill gait training in the slow group com‑
♀ 50% with increasing speed pared to the fast group.
Roesner et al. Systematic Reviews

Control fast group (n=10) The fast group led to bet‑

50,6 (±14,4) Median FAC 1.5 (1–2), ter outcomes on BBS. No
♀ 40% Mean FMA-LL 19.5 between group differ‑
(±4.6) ence is recorded.
Control
Slower speed group
(2024) 13:270

(n=10) Median FAC 1

(1–2), Mean FMA-LL
17.5 (±2.8)
Thimabut et al., 2022 Individual RCT​ n=26 BI Intervention PTs FIM walking, 6MWT, In favour of the interven‑
[77] Japan Intervention Robotic-assisted gait mBI, gait parameters tion group for FIM walk
52.8 (±12.6) device (n=13), Mean BI (Xsens) score in between-group
♀ 23% 10 (±2.61) comparison at the end
Control Control of the ­15th session
62.8 (±8.5) Control group (n=13), (5.00±1.29 vs 3.46±1.76,
♀ 40% Mean BI 11.23 (±2.31) P=.012), but no differ‑
ences were detected
at the second half. No
differences for the 6MWT.
But between-group
differences for ADLs (BI:
(7.31±1.89 vs 4.62±0.96,
P<.001).
Diverse interventions
An et al., 2021 [64] Individual RCT​ n=30 k-mBI Intervention PT BLS, PASS,K-mBI, BBS, All outcomes showed
Korea Intervention whole-body tilting FMA-LL a significant between-
60,5 (±6,0) postural training (n=15), group difference
♀ 27% Mean K-mBI 23,4 (±8,0) in favour for the interven‑
Control Control tion group.
64.7 (±6,9) General postural train‑
♀ 30% ing group (n=15), Mean
K-mBI 17,8 (±11,1)
Page 13 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Chen et al., 2011 [78] Individual RCT​ n=35 FAC, FMA-LL Intervention Thermal-stimulation- FMA-LL, MRC-LL, mMAS, Thermal stimulation
Taiwan Intervention Thermal stimulation PTs PASS (trunk control group demonstrated
58,0 (±11,5) group (n=17) Median items), BBS, FAC greater recovery gains
♀ 23% FAC 0 (IQR 0-1), Median compared to standard
Roesner et al. Systematic Reviews

Control LL FMA 7 (4-11.5) care in all outcomes

62,3 (±11,35) Control except PASS.
♀ 43% Standard rehab group FMA-LL: IG: 14(10.5-15.5);
(n=16) Median FAC 0 CG:6.0(3-9.8); p <0.001
(IQR 0-1), Median LL MRC-LL: IG: 6(4-7);CG:
FMA 6 (4.3-12.0) 3(1.3-4); p<0.001
(2024) 13:270

mMAS: IG 16(12.5-18.5);
CG: 10.5(5.3-14); p=0.01
BBS: IG: 28(20.5-33.5); CG:
15.5(9.3-23.5); p=0.007
FAC: IG: 2(2—2); CG: 1(1-
1); p<0.001
Choi et al., 2021 [59] Individual RCT​ n=24 mBI Intervention Therapists LBT, CBS, MVPT-V, Digital practice with VR
Korea Intervention Digital Practice group head-tracking sensor rehabilitation system
63.00 (±10.02) (n=12) Mean mBI 37.42 data, mBI led to greater recov‑
♀ 58% (±8.73) ery of self-awareness
Control Control of behavioural neglect,
61.58 (±9.99) Control group (n=12) cognitive and visual
♀ 50% Mean mBI 38.08 (±9.80) perception. Between-
group differences in LBT-
score (IG: 11.75±5.83;
CG: 9.67±6.61; p=0.02).
No differences in mBI
(p=0.52) and CBS
(p=0.143).
Katz-Leurer et al., 20003 Individual RCT​ n=92 SSS Leg cycle ergometer Leg cycle ergometer- FAI No differences in FAI
[79] Israel Intervention and regular rehabilita‑ PTs between groups.
65,5 (±12,2) tion groups- actual
♀ 48% number of patients
Control with severe stroke (SSS
70,9 (±11,8) <30) not reported
♀ 46%
Page 14 of 25
Table 1 (continued)
Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics

Kim et al., 2022 [63] Individual RCT​ n=41 K-mBI Intervention Not further defined X-Ray, FMA, K-mBI, VAS There was a signifi‑
Korea Intervention Elastic dynamic shoul‑ pain, MAD, MMT cant between-group
64.76 (±12.80) der sling group (n=21) difference (-0,80±3,11
♀ 38% Mean K-mBI 35.00 vs. 2,28±3,66, p=0,006)
Roesner et al. Systematic Reviews

Control (±17.85) in favour of the interven‑

63.60 (±14.46) Control tion group. No further
♀ 50% Bobath sling group differences had been
(n=20) Mean K-mBI detected.
30.90 (±20.50)
Kwakkel et al., 1999 [52], Individual RCT​ n=101 BI Intervention 1 (IG-A) PTs and OTs BI, FAC, ARAT, 10MWT, UL training group had
(2024) 13:270

2002a [51], 2002b [50] Netherlands Intervention 2 UL training group SIP, NHP, FAI significantly higher ARAT
64,1 (±15,0) (n=33) Median BI 5 (IQR than the splint control
♀ 62% 3-7) group post-intervention.
Intervention 2 Intervention 2 (IG-L) LL training group had sig‑
69 (±9,8) LL training (n=31), nificantly higher BI, FAC,
♀ 52% Median BI 6 (IQR 3-8) walking speed and ARAT
Control Control than the splint control
64,5 (±9,7) Splint control group group post-intervention.
♀ 58% (n=37), Median BI 5.5 No significant differences
(IQR 3-7) in all outcomes were
seen between groups
from 6 months onwards
up until the 12-month
follow-up.
ADL 6 weeks: IG-L:
13(8.75-19); IG-A: 10(5-
13); p< 0.05)
FAC: IG-L: 3(2-4); IG-A:
2(1-3); p<0.05
ARAT: IG-A: 3 (0-34), CG: 0
(0-1); p<0.05
Lincoln et al., 1999 [53] Individual RCT​ n=282 BI^ Intervention PTs/ PTAs RMA- arm scale, ARAT, No differences
UK Intervention Qualified PT group THPT, grip strength, between the groups
73(64-80) (n=94) Median BI 6 (IQR mAS, BI, MCA across all outcomes.
♀ 53% 3-9)
Control 1 Control 1
73 (65-91) PTA group (n=93)
♀ 46% Median BI 6 (IQR 4-8)
Control 2 Control 2
73 (66-80) Standard PT group
♀ 48% (n=95) Median BI 7 (IQR
3-9)
Page 15 of 25
Table 1 (continued)
Roesner et al. Systematic Reviews

Reference Study design, country sample, age in Stroke severity Intervention Delivered by Outcome Results
years, percent measure and severity
female (♀%) characteristics
Shao et al., 2023 [54] Individual RCT​ n=139 NIHSS Intervention Not further defined BBS, 6MWT, mBI, max. There are significant
China Intervention Strength training group muscle strength between-group differ‑
(2024) 13:270

64.56 (±7.08) (n=69) Median NIHSS ences for BBS (adjusted:

♀ 23% 16.25 (±3.69) 40.30±0.75 vs. 33.47±0.74;
Control Control mean difference (95%
65.72 (±5.95) Usual physiotherapy CI) 6.83 (4.71 8.94); ɳ2=
♀ 40% group (n=70) Median 0.24; p <0.001) and 6MWT
NIHSS 15.97 (±3.30) (adjusted: 196.82±3.48
vs. 146.45±3.45, mean
difference (95% CI) 50.32
(40.58 60.05); ɳ2= 0.45; p
<0.001). The comparison
of the gain of muscle
strength of hemiplegic
limbs was in favour
of the intervention group
(p=0,01).
Abbreviations: BBA Brunel Balance Assessment, BLS Burke Lateropulsion Scale, ARAT​Action research Arm Test, aROM Active Range of motion BBA, Brunel Balance Assessment, BBS Berg Balance Scale, BDNF Biomarker
Brain-derived neurotrophic factor – Biomarker, BI Barthel Index, BWS Body weight supported, CBS Catherine Borgego Scale, CG Control group, CI Confidence Interval, CSI Caregiver Strain Index, EQ-5D 5L EuroQol – 5
Dimension – 5 Level, FAC Functional Ambulation Classification, FAI Frenchay activities Index, FES Functional electrical stimulation, FIM Functional independence measure, FMA Fugl-Meyer Assessment, FMA LL Fugl-Meyer
Assessment lower extremity, FMA UE Fugl-Meyer Assessment upper extremity, fMRI Functional magnetic resonance images, GHQ-28 General Health Questionnaire 28, HADS Hospital anxiety and depression scale, HTSD
head-tracking sensor data, ICIQ-SF International Consultation on Incontinence Questionnaire-Short Form, IG Intervention group, K-mBi Korean Modified Barthel Index, K-MMSE Korean Minimental State Exam, LBT Line
bisection Test, mBI Modified Barthel Index, MAL Motor Activity Log, MAS Modified Ashworth scale, mBI Modified Barthel Index, MI Motricity Index, MAS Motor-Assessment-Scale, MMT Manual muscle testing, MoCA
Montreal Cognitive Asessment, MRCS Medical Research Council Scale, mRS Modified Rankin Scale, MVPT-V Motor-Free Visual Perception Test Vertical, NHP Nine-hole peg test, NEADL Nottingham extended Activities of
daily living scale, NIHSS National Institute of Health Stroke Scale, NMES neuromuscular electrostimulation, PASS Postural Assessment Scale for Stroke, OBASS Overactive Bladder Symptom Score, OT Occupational Therapist,
PHQ-9 Patient Health Questionnaire, PT Physiotherapist, PTA Physiotherapy Assistend, RCPM Raven´s Coloured Progressive Matrices, RMI Rivermead Mobility Index, rs fMRI Resting state functional magnetic resonance
imaging, SADQ-10 Stroke Aphasia Depression Questionnaire-10, SF-36 Medical Outcomes Short-Form 36, SCT Star cancelation Test, SSS Scandinavian Stroke Scale, TCT​Trunk Control Test, VAS Visual analogue scale, 5MWT 5
meter walking test, 6MWT 6-minute Walk Test
Page 16 of 25
Roesner et al. Systematic Reviews (2024) 13:270 Page 17 of 25

Table 2 Study demographics alone [46, 55, 58, 68, 70, 72, 73, 76, 78]. The intensity of
n = 30
the interventions and what the control group performed
in terms of content were inconsistently reported. An
Total number randomized 2545 overview can be found in Table 2.
Age, years ° 67.25 (± 13.5)
Sex, n (%) Reporting on quality of evidence and dose
Female 928 (36%) The 30 studies used 54 outcome measures. Those out-
Male 1617 (64%) come measures were categorized by the ICF into body
Stroke type, n (%) (n = 2500; 29 studies) function (n = 26), activity (n = 23), and participation
Ischemic 974 (39%) (n = 5). Supplementary material S4 provides an overview.
Hemorrhagic 1526 (61%) Details on the GRADE criteria are reported in Supple-
Side of stroke, n (%) (n = 2459; 28 studies) mentary material S5.
Left 740 (30%)
Right 1719 (70%) Functional impairment in global early mobilization
Time since stroke, * without electrical supportive devices within 24 h
Hyperacute (≤ 24 h post-stroke) 2 Two studies [49, 57] compared early mobilization within
Acute (> 24 h but ≤ 7days post-stroke) 4 24 h to usual care referred to the outcome of independ-
Early subacute (> 7 days but ≤ 3 months post-stroke) 18 ence. The quality of evidence was judged to be moderate
Late subacute (> 3 months but ≤ 6 months post-stroke) 3 due to concerns about inconsistency and imprecision.
Chronical (> 6 months) 3 Inconsistency and imprecision in the results and their
⁰Source of calculated mean age [43] direction had been found and the threshold of 400 par-
*1month = 30 days ticipants was not reached.
Abbreviations: h-hour
Basic ADLs for neurodevelopmental interventions
without electrical supportive devices
74] or robotic verticalization with the help of an Erigo Neurodevelopmental interventions without electri-
[56], were used in 2 studies. One study involved digital cal supportive devices compared to usual care for basic
practice with virtual reality [59], an elastic sling for the ADLs in severe stroke patients. The quality was judged
upper extremity [63], strength training [54], a whole- with low quality of evidence due to concerns of incon-
body tilt apparatus for postural training [64], differently sistency and imprecision [53, 71]. Wide confidence inter-
qualified therapists [53], and thermal stimulation [78] as vals (CG CI 95% 46.41 (37.77–55.04)); IG CI 95% (67
interventions. An overview of the interventions can be (58.99–75.94)) and appreciable benefits, and no differ-
found in Table 1. ence between groups were found. Further on there were
different directions of effects.
Types of comparators
Interventions were compared to standard care [49, 57, Basic ADL in interventions with NMES
61, 66], delivered by physiotherapists (PTs) [53, 55, 56, Two studies [65, 72] compared NMES to standardized
66, 70, 71, 74, 76–79], occupational therapists (OTs) [52, upper limb therapy and sham NMES in basic ADL. The
60, 62, 78] or speech-language therapists (SLTs) [76, 79], quality of evidence was judged to be low due to inconsist-
as well as by nursing staff [57]. Only one study [67] used ency and imprecision. Different control groups and vary-
routine medication as a control intervention. ing ages in the population may affect the consistency. The
Standard care was often not described in detail. Infor- required threshold of participants was not reached with
mation on the control intervention was incomplete in 172 participants.
five studies [59, 60, 66, 71, 73] and was missing in two
studies [24, 55]. Extended ADLs in interventions with verticalization
support
Therapy modalities Interventions with verticalization support compared to
The study duration ranged from  7 days [71] to 20 weeks usual care in extended ADLs were found in two studies
[50–52]. Therapy sessions were offered twice a day [55, [73, 74]. The quality of evidence was judged to be low due
67, 72], daily [71], and most often five times per week to imprecisions and limitations in the design and imple-
[50–53, 58, 60, 61, 65, 66, 68, 70, 73, 76, 78, 81]. Session mentation of the available studies. The threshold of 400
length differs per day from 10 min [79] to 120 min [57] participants is not reached and the results are impre-
up to statements that refer to the time spent in therapy cise. Both studies showed no significant between-group
Roesner et al. Systematic Reviews (2024) 13:270 Page 18 of 25

Fig. 2 Risk of bias of individual domains for the updated studies. Randomisation process Deviations from the intended interventions Missing
outcome data Measurement of the outcome Selection of the reported result
Roesner et al. Systematic Reviews (2024) 13:270 Page 19 of 25

differences when using a standing frame as verticaliza- function in thermal stimulation [69], walking capacity
tion support. in BWS overground gait training [55], walking capacity
in BWS overground gait training [70], spatiotemporal
Balance skills in neurodevelopmental interventions gait parameters in underwater gait training [46], walk-
The Berg Balance Scale (BBS) was used to assess balance ing speed in RAGT [77], balance in four-channel FES [66]
skills in two RCTs [68, 71] using neurodevelopmental and in strength training off the non-hemiplegic lower
interventions compared to usual care and a conventional extremity [54], upper extremity function in multijoint
Bobath approach. The quality of evidence was judged mirror therapy [60], independence in daily and social
to be very low due to imprecisions, inconsistency of the activities in leg cycle ergometer [79], functional mobility
results, and a high risk of bias in both studies due to seri- in the use of an Oswestry standing [73], muscle strength
ous methodological limitations. The optimal information while using an elastic dynamic sling [63] and pusher syn-
size (OIS) was not reached and differences in population drome in whole-body tilt apparatus [82].
had been detected.
Discussion
Walking capacity in robotic‑assisted gait training The goal of this systematic review was to summarise the
Three studies [61, 75, 76] assessed walking capacity most recent research on physical therapy interventions
using the FAC comparing robotic-assisted gait training and their dosage requirements for patients with severe
(RAGT) to conventional therapy and overground gait stroke who are treated in hospitals or inpatient rehabilita-
training. The quality of evidence was judged to be low. tion facilities. Thirty studies [46–79] were included, with
We had serious concerns about the inconsistency of the 54 outcomes and various types of interventions. Despite
results based on the different reporting and directions of being an update of a recent systematic review, the over-
the effects. Regarding the OIS and the given effect sizes, all evidence remains insufficient. However, the evidence
we had serious concerns regarding imprecision. is not robust enough to determine the effect of physical
therapy interventions for patients with severe stroke. Due
Motor function in robotic‑assisted gait training to limitations in the design, inconsistencies in results,
Three studies [61, 75, 76] compared RAGT to conven- and subjective interpretations, much of the evidence has
tional therapy and overground gait training for assessing been rated as low quality. There were not enough indi-
motor function in the lower limb with the Fugl-Mayer- vidual studies to obtain trustworthy evidence for the out-
Assessment of the lower extremity (FMA-LE). The qual- comes considered.
ity of evidence was judged to be low due to concerns
about inconsistency and imprecision. Serious concerns Robotic gait interventions
were raised regarding the OIS and the missing effect sizes The review included six studies using robotic, BWS,
as well as different directions of effect and slight differ- treadmill, or overground training to undertake a form of
ences in the intervention and control groups. gait training, such as robotic-assisted body weight sup-
ported (BWS) treadmill gait training and underwater
Dexterity in highly intensive active interventions (without gait training [46, 47, 57, 60, 66, 67]. Unexpectedly, there
electric support) was poor-quality evidence for all gait therapy outcomes.
Two studies investigating highly intensive active inter- A systematic review of the current guidelines showed
ventions compared to restriction and usual care for the that robotic gait interventions are now recommended
dexterity of participants with severe stroke, measured [83]. Eight out of 11 guidelines included supporting
using the Action Research Arm Test (ARAT), were RAGT, which was shown to improve walking speed, step
judged low-quality evidence. This is due to some con- length, and balance. Previously, in a meta-regression
cerns about the risk of bias and the number of partici- study, Moucheboeuf et al. (2020, [84]) showed no correla-
pants and the authors’ overall assessment, inconsistency tion between stroke severity and age, time since stroke,
in the results and their direction had been detected as rehabilitation intensity, or treatment success. Stroke
well as not researching the threshold of 400 participants severity was measured using the FAC. In contrast, other
and imprecise results [52, 53]. studies have shown that the severity of paresis influ-
The following outcomes were each examined in one ences the ability to predict the recovery of walking abil-
study, and therefore no reliable conclusions can be drawn ity [85]. However, severity was measured by the presence
regarding their effects: spasticity in a staged rehabilitation of hemiparesis or hemiplegia. The Cochrane Review
intervention [67], neglect in digital training with virtual on treadmill training and BWS for walking after stroke
reality [59], and bimanual mirror therapy [62], cognitive [86] rated walking speed with moderate evidence qual-
function in robotic verticalization [81], sensorimotor ity, while our rating was low. They included 26 studies,
Roesner et al. Systematic Reviews (2024) 13:270 Page 20 of 25

compared to 2 in our review, and found that patients able from ineffective [92] to inconclusive [94] to effective [95].
to walk at the start improved more than those who could Clinicians seek evidence-based, practical research that
not. These findings align with our results, despite differ- reflects real-world rehabilitation [96], underscoring the
ing stroke severity levels in the studies. The differences in need for more transparent reporting in future studies,
outcomes may caused by the different search approaches particularly regarding comparison interventions.
used. The current Cochrane Review on Electromechan-
ical-assisted training for walking after stroke [87] still Therapy dimensions
highlights recent findings. In contrast to our findings, The frequency and dosage of therapeutic interventions
they showed high-quality evidence that the use of elec- like Bobath or RAGT are often unclear. For example,
tromechanical devices increased the likelihood of walk- there is no consensus on the number, duration, timing,
ing independently at the end of the intervention phase or appropriate patient profile for RAGT [83]. A Cochrane
for survivors of stroke. Studies with a focus on patients review indicates that treadmill gait training may be more
with severe stroke were included instead of being inter- beneficial in the first three months post-stroke than in
vention-specific. Nevertheless, the authors of the guide- the chronic phase, but the precise dosage remains uncer-
line review recommend the use of the RAGT for people tain [86]. The review also found no significant increase
who would not otherwise perform gait training [83]. in walking speed for dependent stroke survivors at treat-
However, RAGT should not be used in place of conven- ment onset (95% CI [− 0.06 to 0.03]; P = 0.52) [86]. A
tional therapy [83]. Cochrane review indicated that increased therapy dura-
tion generally improves outcomes, particularly for lower
Usual care extremity functional impairments and ADLs [97].
It is possible to hypothesize that the intervention or One study approached this issue by varying the inter-
control group can be as effective only as the underlying vention start time [49]. However, the functional mobility
standard therapies are if standard therapy (usual care) is outcome of early mobilization after stroke showed mod-
used. In most of the studies included the interventions erate evidence that early activation had no advantage
were carried out in addition to standard care. Accord- over the control group. These interventions both took
ing to the TiDieR checklist, many of the studies lack place in the acute phase after stroke. In a large multi-
a description of their interventions. Regardless of the centre study, 75% of all patients were mobilized within
term “standard therapy”, “standard care”, etc., refer to the 18 h [49], but earlier mobilization was linked to higher
standard of care at the local institution. Usual care and mortality when the mRS score was evaluated (adjusted
control groups are still insufficiently reported in inter- odds ratio (OR) 0.73, 95% CI [0.59 to 0.90]; p = 0.004). A
vention studies. For example, an intervention study may recent systematic review supports starting mobilization
be based on very good usual care, and intervention in 24 h post-stroke [98]. Although there was some evidence
the study, however poor, may achieve good results. The that patients with severe stroke and intracerebral hemor-
reverse may also be true. A recent systematic review [88] rhage would have worse outcomes with very early mobi-
reignited the discussion on control intervention groups lization, these differences were not statistically significant
and highlighted the importance of the therapies on which (p > 0.05). Other RCTs [99, 100] with patients with mild
study interventions are based. Although usual care may to moderate strokes also indicated that more interven-
be referred to by various terms, e.g., rehabilitative ther- tion did not necessarily lead to better outcomes (daily
apy [57], standard intervention [71], or conventional amount per person IG: 31 min (16.5–50.5 min); CG: 10
therapy [61], precise details regarding the frequency, min (0–18 min)). These findings suggest that early mobi-
intensity, and methods of patient treatment are rarely lization within 24 h may be disadvantageous, and factors
recorded. Using tools like the TidieR checklist [39], facili- like intensity and duration, which depend on recovery
tate accurate therapy descriptions [89, 90], but this issue phases, also affect outcomes [97]. Dromerick et al. [101]
exists beyond neurorehabilitation [91]. It is important demonstrated positive effects in the subacute phase post-
to remember that each nation’s standard of treatment stroke, with a treatment window of two and three months
is shaped by its healthcare system. In Germany, outpa- and a daily intensity of 2 h. Of the included studies, only
tient stroke therapy can involve specialized methods three [50–52, 61, 76] provided more therapy, and one
like Bobath or proprioceptive neuromuscular facilitation [66] met Dromerick et al. [101] 600 min/week threshold.
(PNF), which allow for higher payments [92]. Although A systematic review reported a 240% increase in usual
Bobath is not guideline-recommended, it is widely used rehabilitation aimed at reducing ADL limitations [102],
in the UK, with 67% of clinicians employing it for peo- though it did not account for stroke severity. The inter-
ple with stroke [93]. Its use is evident in our findings [53, ventions in weekly duration and total therapy amount are
61, 68]. Studies on the effectiveness of the Bobath vary shown in Table 2.
Roesner et al. Systematic Reviews (2024) 13:270 Page 21 of 25

Limitations Compared to those of patients who had mild or moder-

A limitation of this review was restricting the language to ate stroke, the results of the interventions in these studies
German and English. To address this, we searched mul- had varying quality of evidence.
tiple databases using a broad strategy to identify RCTs The included interventions reflected daily clinical prac-
for severe stroke treatment. Our strict assessment of the tice. Until now, this certainty of evidence has been hin-
high risk of bias using the RoB 2 tool ensured consistency dered by heterogeneous study populations and control
but may have contributed to heterogeneous evaluations groups that are difficult to compare. This has prevented
among different authors. us from drawing any practical conclusions. More research
Our given cut-off score for stroke severity may have is needed. Certainty can be gained when there are more
excluded studies using different measures to define comparable interventions through a better description of
severe stroke [18, 103]. However, broad inclusion crite- the interventions, through a comparable control group,
ria resulted in a diverse range of studies with populations and through a clear description of the severity of stroke
that were difficult to compare. The BI and mBI were most that has been studied. Better transparency will allow for
commonly used (twelve times in all studies) to define better comparability between studies and their respective
stroke severity. The heterogeneity in defining severe outcomes. The analysis of the therapy dimension of an
stroke may change with the adoption of Stroke Recovery intervention can be a key component in explaining study
and Rehabilitation Roundtable Taskforce recommenda- outcomes for patients. Here, in particular, it is interesting
tions [5], which suggest using the NIHSS score. Yet, no to take a closer look at which intervention is effective for
studies before 2018 reported this assessment. This could which degree of severity of stroke and time post-stroke
lead to the large heterogeneity of the included studies. so that it does not turn into a watering can principle.
Other reasons for this heterogeneity were also a high There is a need for additional high-quality studies in the
rated risk of bias, generally small sample size of included early to late subacute phase that systematically articulate
studies, and large and heterogeneous volume of outcome intervention doses from a multidimensional perspective
data. We did not include outpatient or home-based set- in motor stroke recovery. This requires the implementa-
tings because we wanted to focus on the early health care tion of the recommendation of stroke recovery and reha-
of patients with severe stroke. bilitation roundtables for the use of the NIHSS score as
an assessment of stroke severity [5, 106].
Abbreviations
Future research ADL Activities of daily living
Future research should: (a) address, how to assess stroke ARAT​ Action Research Arm Test
severity [5] and especially during baseline assessments BBS Berg Balance Scale
BI Barthel Index
[5]; (b) consider detailed reporting of interventions and BWS Body weight support
control intervention as well as usual care [104, 105] CENTRAL Cochrane Central Register of Controlled Trials
including the amount of dosage according to the report- CI Confidence interval
CINHAL Cumulative Index to Nursing and Allied Health Literature
ing guidelines and (c) studies should include follow-ups cRCT​ Cluster randomized controlled trial
to assess long-term outcomes. DALYs Disability-adjusted life years
The interventions identified in this review are present DORIS Web of Science, Database of Research in Stroke
EMBASE Excerpta Medica Database
in clinical practice, but certainty of their effectiveness is FAC Functional Ambulatory Categories
lacking for daily application. (d) A structured recording FES Functional electrical stimulation
[42], would help accurately describe both the interven- FIM Functional Independence Measure
FMA Fugl-Meyer-Assessment
tion and standard care, improving study comparability, FMA-UL Fugl-Meyer-Assessment Upper Limb
certainty, and theory–practice transfer. FMA-LE Fugl-Meyer-Assessment Lower Extremity
Investigating whether optimal, timely, and targeted GRADE Grading of Recommendations, Assessment, Development and
Evaluation
therapy can reduce long-term costs is essential. ICF International Classification of Functioning, Disability and Health
ICTRP International Trials Registry Platform
IQR Interquartile range
MAL Motor Activity Log
Conclusion MAS Modified Ashworth Scale
This systematic review revealed mostly low- and mod- MEDLINE Medical Literature Analysis and Retrieval System Online
erate-quality evidence related to physical therapy inter- mRS Modified Rankin Scale
NIHSS National Institute of Health Stroke Scale
ventions for patients with severe stroke. Although this NMES Neuromuscular electrical stimulation therapy
is an update of an existing systematic review, there is OIS Optimal information size
still insufficient and little evidence to support the effec- OR Odds ratio
OT Occupational therapist
tiveness of physical therapies in an inpatient setting. PEDro Physiotherapy Evidence Database
Roesner et al. Systematic Reviews (2024) 13:270 Page 22 of 25

PNF Proprioceptive neuromuscular facilitation subarachnoid haemorrhage: a systematic analysis of the global burden
PRR Proportion recovery rule of disease study 2017. Neuroepidemiology. 2020;54(Suppl. 2):171–9.
PT Physiotherapist 2. Feigin VL, Nichols E, Alam T, Bannick MS, Beghi E, Blake N, et al. Global,
RAGT​ Robotic-assisted gait therapy regional, and national burden of neurological disorders, 1990–2016: a
RCT​ Randomised controlled trial systematic analysis for the Global Burden of Disease Study 2016. The
RMA Rivermead Motor Assessment Lancet Neurology. 2019May;18(5):459–80.
RoB2 Risk of Bias Tool 2 3. Feigin VL, Stark BA, Johnson CO, Roth GA, Bisignano C, Abady GG, et al.
SLT Speech-language therapist Global, regional, and national burden of stroke and its risk factors,
TiDieR Template for Intervention Description and Replication 1990–2019: a systematic analysis for the Global Burden of Disease Study
UK United Kingdom 2019. The Lancet Neurology. 2021Oct;20(10):795–820.
4. Grefkes C, Fink GR. Recovery from stroke: current concepts and future
perspectives. Neurol Res Pract. 2020Dec;2(1):17.
Supplementary Information 5. Bernhardt J, Hayward KS, Kwakkel G, Ward NS, Wolf SL, Borschmann K,
The online version contains supplementary material available at https://​doi.​ et al. Agreed definitions and a shared vision for new standards in stroke
org/​10.​1186/​s13643-​024-​02676-0. recovery research: The Stroke Recovery and Rehabilitation Roundtable
taskforce. Int J Stroke. 2017Jul;12(5):444–50.
Supplementary material 1: S1. Search strategy example. S2. Reasons for 6. Buma F, Kwakkel G, Ramsey N. Understanding upper limb recovery after
excluded studies. S3. Funding sources of the included studies. S4. Out‑ stroke. Restor Neurol Neurosci. 2013;31(6):707–22.
comes according to ICF domains, S5. Grade Judgement 7. Grefkes C, Ward NS. Cortical reorganization after stroke: how much and
how functional? Neuroscientist. 2014Feb;20(1):56–70.
8. Krakauer JW, Carmichael ST, Corbett D, Wittenberg GF. Getting
Acknowledgements neurorehabilitation right: what can be learned from animal models?
The authors have nothing to acknowledge. Neurorehabil Neural Repair. 2012Oct;26(8):923–31.
We would like to extend our sincere thanks to Kerstin Luedtke for the critical 9. Carmichael ST. Cellular and molecular mechanisms of neural repair after
revision and supporting thoughts. stroke: making waves. Ann Neurol. 2006May;59(5):735–42.
10. Krakauer J, Marshall R. The proportional recovery rule for stroke revis‑
Authors’ contributions ited: the proportional recovery rule for stroke revisited. Ann Neurol.
KR and SuS conceived the idea for the systematic review. KR conceptualized 2015Dec;78(6):845–7.
the research strategy for this systematic review and wrote the draft of the 11. Prabhakaran S, Zarahn E, Riley C, Speizer A, Chong JY, Lazar RM, et al.
manuscript. BS, TB, MK, BSM, and SuS contributed to the critical revision of the Inter-individual variability in the capacity for motor recovery after
manuscript. All the authors read and approved the final manuscript. ischemic stroke. Neurorehabil Neural Repair. 2008Jan;22(1):64–71.
12. Stinear CM. Prediction of motor recovery after stroke: advances in
Funding biomarkers. The Lancet Neurology. 2017Oct;16(10):826–36.
Open Access funding enabled and organized by Projekt DEAL. We acknowl‑ 13. van der Vliet R, Ribbers GM, Vandermeeren Y, Frens MA, Selles RW. BDNF
edge financial support from Land Schleswig–Holstein within the funding Val66Met but not transcranial direct current stimulation affects motor
program Open Access Publikationsfonds. This research received no further learning after stroke. Brain Stimul. 2017Oct;10(5):882–92.
funding. 14. Bonkhoff AK, Hope T, Bzdok D, Guggisberg AG, Hawe RL, Dukelow SP,
et al. Bringing proportional recovery into proportion: Bayesian model‑
ling of post-stroke motor impairment. Brain. 2020Jul 1;143(7):2189–206.
Declarations 15. Bonkhoff AK, Hope T, Bzdok D, Guggisberg AG, Hawe RL, Dukelow SP,
et al. Recovery after stroke: the severely impaired are a distinct group. J
Competing interests Neurol Neurosurg Psychiatry. 2021 Dec 22;jnnp-2021–327211.
The authors declare that they have no competing interests. 16. Nakayma H, Jørgensen HS, Raaschou HO, Olsen TS. Compensation in
recovery of upper extremity function after stroke: The Copenhagen
Author details Stroke Study. Arch Phys Med Rehabil. 1994Aug;75(8):852–7.
1
Department of Physiotherapy, Pain and Exercise Research Luebeck (P.E.R.L.), 17. Pereira S, Ross Graham J, Shahabaz A, Salter K, Foley N, Meyer M, et al.
Institute of Health Sciences, Universität zu Lübeck, Ratzeburger Allee 160, Rehabilitation of individuals with severe stroke: synthesis of best
23562 Lübeck, Germany. 2 International Graduate Academy (InGrA), Institute evidence and challenges in implementation. Top Stroke Rehabil.
of Health and Nursing Sciences, Medical Faculty of Martin Luther University 2012Mar;19(2):122–31.
Halle‑Wittenberg, University Medicine Halle, Magdeburger Straße 8, 06112 18. Teasell R, Pereira S, Cotoi A. Evidence-based review ofstroke rehabilita‑
Halle (Saale), Germany. 3 Department of Therapy Science I, Brandenburg tion [Internet]. The Rehabilitation of Severe Stroke. 2021. Available from:
University of Technology Cottbus–Senftenberg, Universitätsplatz 1, 01968 http://​www.​ebrsr.​com/​evide​nce-​review/​22-​rehab​ilita​tion-​severe-​stroke
Senftenberg, Germany. 4 Schön Klinik Hamburg Eilbek, Dehnhaide 120, 22081 19. García-Rudolph A, Cegarra B, Opisso E, Tormos JM, Bernabeu M, Saurí
Hamburg, Germany. 5 Executive Department for Nursing Competencies, Wil‑ J. Predicting length of stay in patients admitted to stroke rehabilitation
helmsburger Krankenhaus Hamburg Großsand, Groß‑Sand 3, 21107 Hamburg, with severe and moderate levels of functional impairments. Medicine.
Germany. 6 Department of Occupational Therapy, Institute of Health Sciences, 2020Oct 23;99(43): e22423.
Universität zu Lübeck, Ratzeburger Allee 160, 23562 Lübeck, Germany. 20. Lang CE, Lohse KR, Birkenmeier RL. Dose and timing in neuroreha‑
7
Ernst-Abbe-Hochschule Jena-University of Applied Science, Carl‑Zeiß‑Prome‑ bilitation: prescribing motor therapy after stroke. Curr Opin Neurol.
nade 2, 07745 Jena, Germany. 8 Institute of Health and Nursing Science, Medi‑ 2015Dec;28(6):549–55.
cal Faculty of Martin Luther University Halle‑Wittenberg, University Medicine 21. Hayward KS, Churilov L, Dalton EJ, Brodtmann A, Campbell BCV, Cop‑
Halle, Magdeburger Straße 8, 06112 Halle (Saale), Germany. land D, et al. Advancing stroke recovery through improved articulation
of nonpharmacological intervention dose. Stroke. 2021Feb;52(2):761–9.
Received: 19 February 2024 Accepted: 4 October 2024 22. Winstein C, Kim B, Kim S, Martinez C, Schweighofer N. Dosage matters:
a phase iib randomized controlled trial of motor therapy in the chronic
phase after stroke. Stroke. 2019Jul;50(7):1831–7.
23. Ward NS, Brander F, Kelly K. Intensive upper limb neurorehabilitation in
chronic stroke: outcomes from the Queen Square programme. J Neurol
References Neurosurg Psychiatry. 2019May;90(5):498–506.
1. Krishnamurthi RV, Ikeda T, Feigin VL. Global, regional and country- 24. McGlinchey MP, James J, McKevitt C, Douiri A, Sackley C. The effect of
specific burden of ischaemic stroke, intracerebral haemorrhage and rehabilitation interventions on physical function and immobility-related
Roesner et al. Systematic Reviews (2024) 13:270 Page 23 of 25

complications in severe stroke: a systematic review. BMJ Open. recommendations. Updated October 2013. The GRADE Working Group,
2020Feb;10(2): e033642. 2013. Available from: guidelinedevelopment.org/handbook. [cited 2022
25. Higgins J, Thomas J, Chandler J, Cumpston M, Li T, Page M, et al. Apr 30].
Cochrane Handbook for Systematic Reviews of Interventions version 46. Kim NH, Park HY, Son JK, Moon Y, Lee JH, Cha YJ. Comparison of under‑
6.3 (updated February 2022) [Internet]. Cochrane; 2022. Available from: water gait training and overground gait training for improving the
www.​train​ing.​cochr​ane.​org/​handb​ook walking and balancing ability of patients with severe hemiplegic stroke:
26. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, a randomized controlled pilot trial. Gait Posture. 2020Jul;80:124–9.
et al. The PRISMA 2020 statement: an updated guideline for reporting 47. Bernhardt J, Borschmann K, Collier JM, Thrift AG, Langhorne P, Middle‑
systematic reviews. BMJ. 2021Mar;29: n71. ton S, et al. Fatal and nonfatal events within 14 days after early, inten‑
27. Quinn T, Harrison, McArthur. Assessment scales in stroke: clinimetric sive mobilization poststroke. Neurology. 2021Feb 23;96(8):e1156–66.
and clinical considerations. CIA. 2013 Feb;201. 48. Cain S, Churilov L, Collier J, Carvalho L, Borschmann K, Moodie M, et al.
28. Ancheta J, Husband M, Law D, Reding M. Initial functional independ‑ Factors associated with paid employment 12 months after stroke
ence measure score and interval post stroke help assess outcome, in A Very Early Rehabilitation Trial (AVERT). Ann Phys Rehabil Med.
length of hospitalization, and quality of care. Neurorehabil Neural 2022;65(3): 101565.
Repair. 2000Jun;14(2):127–34. 49. AVERT Trial Collaboration group. Efficacy and safety of very early mobi‑
29. Holloway G, Prognosis R. and decision making in severe stroke. JAMA. lisation within 24 h of stroke onset (AVERT): a randomised controlled
2005Aug 10;294(6):725. trial. The Lancet. 2015Jul;386(9988):46–55.
30. Rollnik JD, Janosch U. Current trends in the length of stay in neurologi‑ 50. Kwakkel G, Wagenaar RC. Effect of duration of upper- and
cal early rehabilitation. Deutsches Ärzteblatt international. 2010 Apr 23; lower-extremity rehabilitation sessions and walking speed on
Available from: https://www.aerzteblatt.de/https://​doi.​org/​10.​3238/​ recovery of interlimb coordination in hemiplegic gait. Phys Ther.
arzte​bl.​2010.​0286[cited 2022 Mar 13] 2002May;82(5):432–48.
31. Duncan PW, Goldstein LB, Horner RD, Landsman PB, Samsa GP, Matchar 51. Kwakkel G, Kollen BJ, Wagenaar RC. Long term effects of intensity of
DB. Similar motor recovery of upper and lower extremities after stroke. upper and lower limb training after stroke: a randomised trial. J Neurol
Stroke. 1994Jun;25(6):1181–8. Neurosurg Psychiatry. 2002Apr;72(4):473–9.
32. Mehrholz J, Wagner K, Rutte K, Meiβner D, Pohl M. Predictive 52. Kwakkel G, Wagenaar RC, Twisk JW, Lankhorst GJ, Koetsier JC. Intensity
validity and responsiveness of the functional ambulation cat‑ of leg and arm training after primary middle-cerebral-artery stroke: a
egory in hemiparetic patients after stroke. Arch Phys Med Rehabil. randomised trial. The Lancet. 1999Jul;354(9174):191–6.
2007Oct;88(10):1314–9. 53. Lincoln NB, Parry RH, Vass CD. Randomized, controlled trial to evaluate
33. Ko SB, Yoon BW. Mechanisms of functional recovery after stroke. In: increased intensity of physiotherapy treatment of arm function after
Naritomi H, Krieger DW, editors. Frontiers of Neurology and Neurosci‑ stroke. Stroke. 1999Mar;30(3):573–9.
ence. Basel: S. KARGER AG; 2013. p. 1–8. Available from: https://​www.​ 54. Shao C, Wang Y, Gou H, Xiao H, Chen T. Strength training of the
karger.​com/​Artic​le/​FullT​ext/​346405. [cited 2022 Mar 13]. nonhemiplegic side promotes motor function recovery in patients
34. Hatem SM, Saussez G, della Faille M, Prist V, Zhang X, Dispa D, et al. with stroke: a randomized controlled trial. Arch Phys Med Rehabil.
Rehabilitation of motor function after stroke: a multiple systematic 2023;104(2):188–94.
review focused on techniques to stimulate upper extremity recovery. 55. Brunelli S, Iosa M, Fusco F, Pirri C, Di Giunta C, Foti C, et al. Early body
Front Hum Neurosci. 2016 Sep 13;10. Available from: http://journal. weight-supported overground walking training in patients with stroke
frontiersin.org/Article/https://​doi.​org/​10.​3389/​fnhum.​2016.​00442/​abstr​ in subacute phase compared to conventional physiotherapy: a rand‑
act. [cited 2022 Mar 13]. omized controlled pilot study. International journal of rehabilitation
35. Chen CM, Hsu HC, Chang CH, Lin CH, Chen KH, Hsieh WC, et al. Age- research Internationale Zeitschrift fur Rehabilitationsforschung Revue
based prediction of incidence of complications during inpatient stroke internationale de recherches de readaptation. 2019;42(4):309–15.
rehabilitation: a retrospective longitudinal cohort study. BMC Geriatr. 56. Calabrò RS, Naro A, Russo M, Leo A, Balletta T, Saccá I, et al. Do post-
2014Dec;14(1):41. stroke patients benefit from robotic verticalization? A pilot-study
36. Dromerick A, Reding M. Medical and neurological complications during focusing on a novel neurophysiological approach. RNN. 2015Oct
inpatient stroke rehabilitation. Stroke. 1994Feb;25(2):358–61. 5;33(5):671–81.
37. Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan—a web 57. Di Lauro A, Pellegrino L, Savastano G, Ferraro C, Fusco M, Balzarano F,
and mobile app for systematic reviews. Syst Rev. 2016Dec;5(1):210. et al. A randomized trial on the efficacy of intensive rehabilitation in the
38. Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 acute phase of ischemic stroke. J Neurol. 2003Oct 1;250(10):1206–8.
Statement: updated guidelines for reporting parallel group randomised 58. Franceschini M, Carda S, Agosti M, Antenucci R, Malgrati D, Cisari
trials. BMC Med. 2010 Mar 24;8:18. C. Walking After Stroke: What Does Treadmill Training With Body
39. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Weight Support Add to Overground Gait Training in Patients Early
Better reporting of interventions: template for intervention descrip‑ After Stroke?: A Single-Blind, Randomized. Controlled Trial Stroke.
tion and replication (TIDieR) checklist and guide. BMJ. 2014 Mar 2009Sep;40(9):3079–85.
7;348(mar07 3):g1687–g1687. 59. Choi HS, Shin WS, Bang DH. Application of digital practice to improve
40. Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. head movement, visual perception and activities of daily living for
RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. subacute stroke patients with unilateral spatial neglect: Preliminary
2019Aug;28: l4898. results of a single-blinded, randomized controlled trial. Medicine.
41. Revised Cochrane risk-of-bias-tool for randomized controlled trails (RoB 2021Feb 12;100(6): e24637.
2) – SHORT VERSION (CRIBSHEET). Available from: https://​drive.​google.​ 60. Lee SH, Prak JS, Choi JB, Yoo WG. Improving upper extremity motor
com/​file/d/​1Q4Fk​3HCuB​RwIDW​TGZa5​oH11O​dR4Gb​hdo/​view function in stroke patients using a complex task with multi-joint-
42. Dalton EJ, Churilov L, Lannin NA, Corbett D, Campbell BCV, Hayward based mirror therapy: a randomized controlled trial. Neurology Asia.
KS. Multidimensional phase I dose ranging trials for stroke recovery 2020;25(3):245–51.
interventions: key challenges and how to address them. Neurorehabil 61. Chang WH, Kim MS, Huh JP, Lee PKW, Kim YH. Effects of robot-
Neural Repair. 2021Aug;35(8):663–79. assisted gait training on cardiopulmonary fitness in subacute stroke
43. Chang A, Sahota D. Statistics Toolkit (StatsToDo). Available from: https://​ patients: a randomized controlled study. Neurorehabil Neural Repair.
www.​stats​todo.​com/​Combi​neMea​nsSDs.​php. [cited 2021 May 22]. 2012May;26(4):318–24.
44. Murad MH, Mustafa RA, Schünemann HJ, Sultan S, Santesso N. Rating 62. Sim T, Kwon J. Comparing the effectiveness of bimanual and unimanual
the certainty in evidence in the absence of a single estimate of effect. mirror therapy in unilateral neglect after stroke: a pilot study. NeuroRe‑
Evid Based Med. 2017Jun;22(3):85–7. habilitation. 2022;50(1):133–41.
45. Schünemann H, Brożek J, Guyatt G, Oxman A, editors. GRADE handbook 63. Kim M, Lee S, Park E, Choi M, Kim J, Sohn M, et al. Elastic dynamic sling
for grading quality of evidence and strength of recommendations. on subluxation of hemiplegic shoulder in patients with subacute
GRADE handbook for grading quality of evidence and strength of stroke: a multicenter randomized controlled trial. International journal
Roesner et al. Systematic Reviews (2024) 13:270 Page 24 of 25

of environmental research and public health. 2022;19(16). Available 81. Calabro RS, Accorinti M, Porcari B, Carioti L, Ciatto L, Billeri L, et al.
from: https://www.cochranelibrary.com/central/doi/https://​doi.​org/​10.​ Does hand robotic rehabilitation improve motor function by
1002/​centr​al/​CN-​02456​101/​full rebalancing interhemispheric connectivity after chronic stroke?
64. An CM, Ko MH, Kim D hyun, Kim GW. Effect of postural training using a Encouraging data from a randomised-clinical-trial. Clin Neurophysiol.
whole-body tilt apparatus in subacute stroke patients with lateropul‑ 2019May;130(5):767–80.
sion: a single-blinded randomized controlled trial. Annals of Physical 82. An CM, Ko MH, Kim DH, Kim GW. Effect of postural training using a
and Rehabilitation Medicine. 2021 Mar;64(2):101393. whole-body tilt apparatus in subacute stroke patients with lateropul‑
65. Guo G yan, Kang Y gang. Effectiveness of neuromuscular electrical sion: a single-blinded randomized controlled trial. Ann Phys Rehabil
stimulation therapy in patients with urinary incontinence after stroke: a Med. 2021;64(2): 101393.
randomized sham controlled trial. Medicine. 2018 Dec;97(52):e13702. 83. on behalf of Italian Consensus Conference on Robotics in Neuroreha‑
66. Zheng X, Chen D, Yan T, Jin D, Zhuang Z, Tan Z, et al. A randomized clini‑ bilitation (CICERONE), Calabrò RS, Sorrentino G, Cassio A, Mazzoli D,
cal trial of a functional electrical stimulation mimic to gait promotes Andrenelli E, et al. Robotic-assisted gait rehabilitation following stroke:
motor recovery and brain remodeling in acute stroke. Behav Neurol. a systematic review of current guidelines and practical clinical recom‑
2018Dec;18(2018):1–10. mendations. Eur J Phys Rehabil Med [Internet]. 2021 Jul;57(3). Available
67. Bai Y long, Hu Y shan, Wu Y, Zhu Y lian, Zhang B, Jiang C yu, et al. from: https://​www.​miner​vamed​ica.​it/​index2.​php?​show=​R33Y2​021N0​
Long-term three-stage rehabilitation intervention alleviates spasticity 3A0460. [cited 2022 Aug 24].
of the elbows, fingers, and plantar flexors and improves activities of 84. Moucheboeuf G, Griffier R, Gasq D, Glize B, Bouyer L, Dehail P, et al.
daily living in ischemic stroke patients: a randomized, controlled trial. Effects of robotic gait training after stroke: a meta-analysis. Ann Phys
NeuroReport. 2014 Sep 10;25(13):998–1005. Rehabil Med. 2020Nov;63(6):518–34.
68. Tang Q, Tan L, Li B, Huang X, Ouyang C, Zhan H, et al. Early sitting, stand‑ 85. Matsunaga A, Kurokawa Y, Kanda T, Sakai F. Prediction of locomotion
ing, and walking in conjunction with contemporary bobath approach function at one month after stroke. J Phys Ther Sci. 1997;9(2):93–7.
for stroke patients with severe motor deficit. Top Stroke Rehabil. 86. Mehrholz J, Thomas S, Elsner B. Treadmill training and body weight sup‑
2014Mar;21(2):120–7. port for walking after stroke. Cochrane Stroke Group, editor. Cochrane
69. Cui W, Huang L, Tian Y, Luo H, Chen S, Yang Y, et al. Effect and mecha‑ Database of Systematic Reviews. 2017 Aug 17.;2017(8). Available from:
nism of mirror therapy on lower limb rehabilitation after ischemic http://doi.wiley.com/https://​doi.​org/​10.​1002/​14651​858.​CD002​840.​
stroke: a fMRI study. NeuroRehabilitation. 2022;51(1):65–77. pub4. [cited 2022 Aug 24].
70. Rodrigues TA, Goroso DG, Westgate PM, Carrico C, Batistella LR, 87. Mehrholz J, Thomas S, Kugler J, Pohl M, Elsner B. Electromechani‑
Sawaki L. Slow versus fast robot-assisted locomotor training after cal‐assisted training for walking after stroke. Cochrane Database of
severe stroke: a randomized controlled trial. Am J Phys Med Rehabil. Systematic Reviews [Internet]. 2020;(10). Available from: https://​doi.​org/​
2017Oct;96(10):S165–70. 10.​1002/​14651​858.​CD006​185.​pub5
71. Rahayu UB, Wibowo S, Setyopranoto I, Hibatullah RM. Effectiveness of 88. Arienti C, Buraschi R, Pollet J, Lazzarini SG, Cordani C, Negrini S, et al. A
physiotherapy interventions in brain plasticity, balance and functional systematic review opens the black box of ‘usual care’ in stroke rehabili‑
ability in stroke survivors: a randomized controlled trial. NRE. 2020Dec tation control groups and finds a black hole. Eur J Phys Rehabil Med
22;47(4):463–70. [Internet]. 2022 Jul;58(4). Available from: https://​www.​miner​vamed​ica.​
72. Rosewilliam S, Malhotra S, Roffe C, Jones P, Pandyan AD. Can surface it/​index2.​php?​show=​R33Y2​022N0​4A0520. [cited 2022 Aug 28].
neuromuscular electrical stimulation of the wrist and hand combined 89. Nicholls SG, Zwarenstein M, Taljaard M. The importance of describing as
with routine therapy facilitate recovery of arm function in patients with well as defining usual care. Am J Bioeth. 2020Jan 2;20(1):56–8.
stroke? Arch Phys Med Rehabil. 2012Oct;93(10):1715-1721.e1. 90. Macklin R, Natanson C. Misrepresenting, “usual care” in research: an ethi‑
73. Bagley P, Hudson M, Forster A, Smith J, Young J. A randomized trial cal and scientific error. Am J Bioeth. 2020Jan 2;20(1):31–9.
evaluation of the Oswestry Standing Frame for patients after stroke. 91. Young HM, Miyamoto S, Tang-Feldman Y, Dharmar M, Balsbaugh T,
Clin Rehabil. 2005Jun;19(4):354–64. Greenwood D. Defining usual care in clinical trials. Res Gerontol Nurs.
74. Logan A; Freeman J; Kent B; Pooler J; Creanor S; Enki D; Vickery J; Barton 2020May;13(3):125–9.
A; Marsden J. Functional standing frame programme early after severe 92. Pt SH, Pt PGP, Dutta A, Pt AK. Comparison of effectiveness of trunk PNF
sub-acute stroke (SPIRES): a randomised controlled feasibility trial. Pilot versus NDT on trunk stability in stroke patients: life sciences-physiother‑
and Feasibility Studies 2022 Mar 3;8(50):Epub. 2022; apy. Int J Life Sci Pharm Res. 2022Jul;1:L78-92.
75. Louie DR, Mortenson WB, Durocher M, Schneeberg A, Teasell R, Yao 93. Lennon S. Physiotherapy practice in stroke rehabilitation: a survey.
J, et al. Efficacy of an exoskeleton-based physical therapy program Disabil Rehabil. 2003Jan;25(9):455–61.
for non-ambulatory patients during subacute stroke rehabilitation: a 94. Pathak A, Gyanpuri V, Dev P, Dhiman N. The Bobath Concept (NDT) as
randomized controlled trial. J Neuroeng Rehabil. 2021;18(1):149. rehabilitation in stroke patients: a systematic review. J Family Med Prim
76. Ochi M, Wada F, Saeki S, Hachisuka K. Gait training in subacute non- Care. 2021;10(11):3983.
ambulatory stroke patients using a full weight-bearing gait-assistance 95. Yazıcı G, Güçlü Gündüz A, Batur Çağlayan HZ, Özkul Ç, Volkan Yazıcı M,
robot: a prospective, randomized, open, blinded-endpoint trial. J Neurol Nazlıel B. Investigation of early term neurodevelopmental treatment-
Sci. 2015Jun;353(1–2):130–6. bobath approach results in patients with stroke. Türk Beyin Damar
77. Thimabut N, Yotnuengnit P, Charoenlimprasert J, Sillapachai T, Hirano Hast Derg. 2021;27(1):27–33.
S, Saitoh E, et al. Effects of the robot-assisted gait training device plus 96. Cott CA, Graham JV, Brunton K. When will the evidence catch up with
physiotherapy in improving ambulatory functions in patients with clinical practice? Physiother Can. 2011Jul;63(3):387–90.
subacute stroke with hemiplegia: an assessor-blinded, randomized 97. Clark B, Whitall J, Kwakkel G, Mehrholz J, Ewings S, Burridge J. The effect
controlled trial. Arch Phys Med Rehabil. 2022;103(5):843–50. of time spent in rehabilitation on activity limitation and impairment
78. Chen JC, Lin CH, Wei YC, Hsiao J, Liang CC. Facilitation of motor and after stroke. Cochrane Stroke Group, editor. Cochrane Database of
balance recovery by thermal intervention for the paretic lower limb Systematic Reviews. 2021 Oct 25;2021(10). Available from: http://doi.
of acute stroke: a single-blind randomized clinical trial. Clin Rehabil. wiley.com/https://​doi.​org/​10.​1002/​14651​858.​CD012​612.​pub2. [cited
2011Sep;25(9):823–32. 2022 Aug 31].
79. Katz-Leurer M, Carmeli E, Shochina M. The effect of early aerobic 98. Rethnam V, Langhorne P, Churilov L, Hayward KS, Herisson F, Poletto
training on independence six months post stroke. Clin Rehabil. SR, et al. Early mobilisation post-stroke: a systematic review and
2003Nov;17(7):735–41. meta-analysis of individual participant data. Disabil Rehabil. 2022Apr
80. Choi HS, Shin WS, Bang DH. Application of digital practice to improve 10;44(8):1156–63.
head movement, visual perception and activities of daily living for 99. Sundseth A, Thommessen B, Rønning OM. Outcome after mobilization
subacute stroke patients with unilateral spatial neglect: Preliminary within 24 hours of acute stroke: a randomized controlled trial. Stroke.
results of a single-blinded, randomized controlled trial. Medicine (Balti‑ 2012Sep;43(9):2389–94.
more). 2021Feb 12;100(6): e24637.
Roesner et al. Systematic Reviews (2024) 13:270 Page 25 of 25

100. Yelnik AP, Quintaine V, Andriantsifanetra C, Wannepain M, Reiner P,

Marnef H, et al. AMOBES (Active Mobility Very Early After Stroke): a
randomized controlled trial. Stroke. 2017Feb;48(2):400–5.
101. Dromerick AW, Geed S, Barth J, Brady K, Giannetti ML, Mitchell A, et al.
Critical Period After Stroke Study (CPASS): a phase II clinical trial testing
an optimal time for motor recovery after stroke in humans. Proc Natl
Acad Sci USA. 2021Sep 28;118(39): e2026676118.
102. Schneider EJ, Lannin NA, Ada L, Schmidt J. Increasing the amount of
usual rehabilitation improves activity after stroke: a systematic review. J
Physiother. 2016Oct;62(4):182–7.
103. Sulter G, Steen C, Jacques De Keyser. Use of the Barthel Index and Mod‑
ified Rankin Scale in Acute Stroke Trials. Stroke. 1999 Aug;30(8):1538–41.
104. Negrini S, Armijo-Olivo S, Patrini M, Frontera WR, Heinemann AW,
Machalicek W, et al. The randomized controlled trials rehabilitation
checklist: methodology of development of a reporting guideline
specific to rehabilitation. Am J Phys Med Rehabil. 2020Mar;99(3):210–5.
105. Negrini S, Arienti C, Kiekens C. Usual care: the big but unmanaged prob‑
lem of rehabilitation evidence. The Lancet. 2020Feb;395(10221):337.
106. Kwakkel G, Meskers CG, Ward NS. Time for the next stage of stroke
recovery trials. Lancet Neurol. 2020;19(8):636–7.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub‑
lished maps and institutional affiliations.