932964

research-article2020
CRE0010.1177/0269215520932964Clinical RehabilitationVeldema and Jansen

CLINICAL
Original Article REHABILITATION

Clinical Rehabilitation

Resistance training in stroke 1­–25

© The Author(s) 2020
Article reuse guidelines:
rehabilitation: systematic review sagepub.com/journals-permissions
DOI: 10.1177/0269215520932964
https://doi.org/10.1177/0269215520932964

and meta-analysis journals.sagepub.com/home/cre

Jitka Veldema and Petra Jansen

Abstract
Objective: This systematic review and meta-analysis investigates the effects of resistance training in
supporting the recovery in stroke patients.
Data sources: PubMed, the Cochrane Central Register of Controlled Trials and the PEDro databases
were reviewed up to 30 April 2020.
Review methods: Randomized controlled trials were included, who compared: (i) resistance training
with no intervention, (ii) resistance training with other interventions and (iii) different resistance training
protocols in stroke rehabilitation.
Results: Overall 30 trials (n = 1051) were enrolled. The parameters evaluated were: (1) gait, (2) muscular
force and motor function, (3) mobility, balance and postural control, (4) health related quality of life,
independence and reintegration, (5) spasticity and hypertonia, (6) cardiorespiratory fitness, (7) cognitive
abilities and emotional state and (8) other health-relevant physiological indicators. The data indicates
that: (i) resistance training is beneficial for the majority of parameters observed, (ii) resistance training
is superior to other therapies on muscular force and motor function of lower and upper limbs, health
related quality of life, independence and reintegration and other health-relevant physiological indicators,
not significantly different from other therapies on walking ability, mobility balance and postural control
and spasticity and hypertonia, and inferior to ergometer training on cardiorespiratory fitness and (iii) the
type of resistance training protocol significantly impacts its effect; leg press is more efficient than knee
extension and high intensity training is superior than low intensity training.
Conclusion: Current data indicates that resistance training may be beneficial in supporting the recovery
of stroke patients. However, the current evidence is insufficient for evidence-based rehabilitation.

Keywords
Stroke, resistance training, rehabilitation

Received: 7 August 2019; accepted: 19 May 2020

Introduction Faculty of Psychology, Education and Sport Science, University

Stroke is one of the leading causes of long-term dis- of Regensburg, Regensburg, Germany
ability in adult persons worldwide,1 and the optimi-
Corresponding author:
zation of therapy management in these patients has
Jitka Veldema, Faculty of Psychology, Education and Sport
a high socio-economic importance. The evaluation Science, University of Regensburg, Universitätsstraße 31,
of fitness training and exercise on supporting func- Regensburg D-93053, Germany.
tion and quality of life and avoiding a subsequent Email: [email protected]
2 Clinical Rehabilitation 00(0)

stroke is one of the top research priorities related to review demonstrates that a regular application of
life after stroke.2 In this meta-analysis we will focus resistance exercises in elderly people significantly
on resistance training,3 which is a type of strength reduces the occurrence of a fall.9 Thus, it is possible
training that has gained popularity over the last dec- that this training may also be beneficial for persons
ades in rehabilitation after a stroke. This type of that suffered from stroke. The impact of resistance
training uses the resistance of free weights, machine training on cardiorespiratory fitness is also highly
weights, body weight or resistance band, which relevant. Because of motor impairment, energetic
have to be overcome by voluntary muscular effort.4 requirements in everyday activities are significantly
Depending on muscle contraction type, one can dif- enhanced after stroke.10 However, the cardiorespi-
ferentiate between (a) concentric, (b) eccentric and ratory fitness is extremely reduced in comparison to
(c) isometric resistance training.4 During concentric healthy persons.11 This decreases independence in
training, the muscle shortens in length while tension daily living activities and evocates secondary
increases to overcome or move some resistance. In stroke-related complications.12 Existing studies
an eccentric exercise, the resistance is greater than have already demonstrated that resistance training
the muscular force being produced, and the muscle induces cardiorespiratory benefits in neurological13
lengthens while producing tension. Concentric and as well as non-neurological14 cohorts. This type of
eccentric exercises are considered as isokinetic training also may impact cognition and emotional
exercises. During isometric exercises (also known wellbeing because physical activities support cog-
as static exercises) the trained muscle contracts to nitive recovery after stroke15 and relieve symptoms
produce tension, but there is no change in muscle of depression.16 Collectively, the present data indi-
length.4 The important variables of resistance train- cates that resistance training may help alleviate a
ing are (i) exercise intensity, (ii) the number of rep- broad spectrum of stroke-related disabilities.
etitions and the number of sets and (iii) the duration However, it is not clear if the present data is enough
of rest breaks between the sets and exercises. The for evidence-based rehabilitation.
exercise intensity (load of exercise) is usually Earlier reviews conducted until 2016 found
defined as the percentage of a one-repetition maxi- insufficient evidence for resistance training in
mum (the heaviest resistance that can be used to stroke rehabilitation.7,8,17 Since then, a lot of new
correctly complete only one repetition).3 Similarly, studies tested the effects of this therapy in a stroke
a 10-repetitions maximum, or a 25-repetition maxi- cohort, and a current overview of available litera-
mum may be used.3 It is also common to define a ture can bring new findings useful in stroke reha-
perceived exertion during exercise (e.g. ‘very bilitation. Furthermore, not only pure resistance
hard’).5 The number of repetitions, the number of training studies but also mixed interventions trials
sets, as well as the duration of rest breaks varies in combining resistance training with other interven-
dependence on trained population and the trainings tions were included in previous reviews. This
goals. As a rule, one to four sets of one to 30 repeti- approach makes it difficult to make a precise state-
tions are applied. A typical rest break is between ment about the effectiveness of resistance training.
tens of seconds up to some minutes.6 Here we perform a current overview and meta-
A frequently discussed topic is the effectiveness analysis of studies which evaluated pure resistance
of this type of training in supporting walking abil- training in supporting recovery after stroke (1) in
ity.2 The previous reviews and meta-analyses pro- comparison to no intervention, (2) in comparison
vide inconsistent results.7,8 One meta-analysis to other interventions and (3) by comparing differ-
shows that strength exercises support muscular ent resistance training protocols.
force and gait recovery after stroke.7 However,
another review found that strength exercises are
Methods
beneficial only for muscular force, but not for gait
ability in this cohort.8 A further important topic is The protocol for this systematic review and meta-
the recovery of balance and coordination.2 A recent analysis is not registered.
Veldema and Jansen 3

Data source PEDro scale. The overall score of this scale ranges
between 0 and 10. The higher the score, the better
The PubMed, the Cochrane Central Register of the methodological quality of the study (high qual-
Controlled Trials and the PEDro databases were ity = PEDro score 6–10, fair quality = PEDro score
searched from their inception through to 30 April 4–5, poor quality = PEDro score ⩽ 3).18
2020 for controlled trials evaluating the effects of
resistance training in supporting recovery after a
stroke. Search terms ‘stroke’, ‘resistance training’ Data synthesis and statistical analysis
and ‘stroke’, ‘strength exercises’ were used. The Effect size and the 95% confidence intervals were
screening was performed in accordance with the calculated for all parameters observed. On their
PRISMA guidelines by two independent reviewers. basis, overall outcomes (including whole outcomes)
First, titles and abstracts were screened for eligibility. and pooled outcomes (such as gait, cognitive abili-
After that, the full-text publications of the potentially ties etc. including corresponding outcomes) were
eligible trials were evaluated. In addition, reference calculated for each study and forest plots were con-
lists of previous relevant reviews and meta-analysis structed. For interpretation, the Cohen definition of
were searched for suitable trials. Disagreements were effect size was used (d = 0.2 ‘small’, d = 0.5
resolved by consensus. The actual search strategy is ‘medium’, d = 0.8 ‘large’).19 The homogeneity of
illustrated in the Online Appendix. effects across studies was evaluated using the
inconsistency test (I2), where values above 50%
Study selection were considered indicative of high heterogeneity.20
Trials matching the following criteria were included:
(1) human-studies, (2) prospective studies, (3) writ- Results
ten in English, (4) diagnosis of stroke, (5) resistance Overall, 30 trials corresponding to our inclusion cri-
training as intervention, (6) pre-and post-intervention teria were found. A total of 1051 patients with stroke
assessment, (7) two experimental groups at least and were enrolled. The trials show a large variety of par-
(8) five randomized patients at least. Only studies ticipants included, interventions applied and param-
applying pure resistance training were enrolled. eters evaluated. For the sake of simplicity, we
Mixed interventions (resistance training with an grouped the studies in three categories depending on
additional intervention) were excluded. study-protocol applied: (i) ‘resistance training’ com-
pared with ‘no intervention’, (ii) ‘resistance training’
Data extraction and risk of bias compared with ‘other intervention’ and (iii) ‘resist-
ance training’ compared with one another ‘resistance
One author extracted the following information training’. Serious adverse events were not reported.
from the selected publications: (1) subjects charac-
teristics (age, gender, time since stroke, stroke aeti-
ology and stroke location), (2) methodological Resistance training versus no
approach (number of participants, crossover/paral-
lel group design, number and scheduling of evalua-
intervention
tions), (3) intervention characteristics (type of Eleven studies investigated the effects of resistance
exercises, number of sets and repetitions, resistance training in comparison with no intervention (Table 1,
used, duration of rest period) and (4) outcomes Figure 1).21–31
(assessments applied, between group differences Participants and interventions: A total of 403
detected). If the information was missing in the persons were enrolled, between three months and
manuscripts, the study authors have been contacted 4.8 years after stroke. Between 12 and 40 interven-
for clarification. The methodological quality of tri- tion sessions of unilateral (affected side)22,24,26–30 or
als included (such as random allocation, baseline bilateral (both sides)21,23,25,26,31 resistance training
comparability, blinding etc.) was assessed using of lower limbs were applied. Two studies also
 4

Table 1.  Overview of studies investigating resistance training in comparison to no intervention in supporting stroke recovery.
Reference Subjects number/gender/ Time since Stroke etiology/lesion Study design/sessions Intervention Results/used assessments
age (years) stroke location/affected number and duration/
hemisphere evaluations/PEDro scale
(score)

Büyükvural et al.23 50/33 males, 17 3 (2–9) months 44 ischemic, 6 Parallel groups (25+25)/15 RT: bilateral concentric (1) knee flexion- RT significantly better: Ten Meter Walk Test, Six
females/53 ± 11 years haemorrhagic/lesion sessions, duration na/T1 extension and (2) ankle flexion-extension on Minute Walk Test, Timed Up and Go, Stair Climbing
location na/18 right, before treatment, T2 after training machine (one set of 15 or 20 repetitions Test, Berg Balance Scale, Rivermead Mobility Index,
22 left treatment/6 with increased angular velocity (90–150°/sec), Functional Independence Measure, Stroke Specific
followed by one set of 10 repetitions with Quality of Life Scale, muscular force (knee flexion, knee
constant angular velocity (180°/sec), submaximal extension, ankle flexion, ankle extension)
resistance, 10 sec rest)
  CG: no intervention  
Sims et al.31 45/27 males, 18 13 ± 5 months Etiology na/lesion Parallel groups (23+22)/20 RT: bilateral concentric (1) seated row, (2) lat No significant differences: Centre for Epidemiologic
females/67 ± 15 years location na/affected sessions, duration na/T1 pull-down, (3) chest press, (4) leg press, (5) Studies for Depression scale, Assessment of Quality of
hemisphere na before treatment, T2 after calf raise and (6) knee extension on training Life Instrument, Short Form (12) Health Survey, Stroke
treatment, T3 four months machines (three sets of 8–10 repetitions, Impact Scale, Satisfaction with Life Scale, Social Support
after treatment/6 resistance 80% of 1RM, rest na) Survey, Life Orientation Test-Revised, measure of
generalized dispositional optimism, Self-Esteem Scale,
Recovery Locus of Control Scale
  CG: no intervention  
Lee et al.29 39/25 males, 14 15 ± 9 months 23 ischemic, 16 Parallel groups RT1: unilateral concentric (1) leg press on RT1 significantly better than CG: plantar pressure
females/50 ± 8 years haemorrhagic/lesion (13+13+13)/30 sessions, training machine (three sets of 8–10 repetitions, distribution of the affected side (contact area –
location na/20 right, duration na/T1 before resistance 70% of 1RM, rest na); preceded by hindfoot; contact impulse – hindfoot)
18 left treatment, T2 after warm up (four repetitions with 25% of 1RM)
treatment/6
  RT2: unilateral concentric (1) knee extension on No significant differences between RT1 and CG:
training machine (three sets of 8–10 repetitions, plantar pressure distribution of the affected side
resistance 70% of 1RM, rest na); preceded by (contact area – forefoot, midfoot; peak contact
warm up (four repetitions with 25% of 1RM) force – forefoot, midfoot, hindfoot; contact impulse
– forefoot, midfoot)
  CG: no intervention No significant differences between RT2 and CG:
plantar pressure distribution of the affected side
(contact area – forefoot, midfoot, hindfoot; peak
contact force – forefoot, midfoot, hindfoot; contact
impulse – forefoot, midfoot, hindfoot)
Lee et al.30 28/17 males, 11 17 ± 10 16 ischemic, 12 Parallel groups (14+14)/30 RT: unilateral concentric (1) leg press on RT significantly better: gait parameters (gait velocity,
females/61 ± 8 years months haemorrhagic/lesion sessions à 30 min/T1 training machine (three sets of 8–10 repetitions, step length, stride length, double support, heel-to-heel
location na/13 right, before treatment, T2 after resistance 70% of 1RM, rest na); preceded by of support)
15 left treatment/6 warm up (four repetitions with 25% of 1RM)
  CG: no intervention  
Lee et al.28 33/20 males, 13 20 ± 8 months 20 ischemic, 13 Parallel groups RT1: unilateral concentric (1) leg press on RT1 significantly better than CG: EMG activity
females/59 ± 7 years haemorrhagic/lesion (11+11+11)/30 sessions, training machine (three sets of 8–10 repetitions, (rectus femoris, biceps femoris, gastrocnemius, tibialis
location na/17 right, duration na/T1 before resistance 70% of 1RM, 3 min rest); preceded by anterior), balance (anterio-posterior sway velocity with
16 left treatment, T2 after warm up (four repetitions with 25% of 1RM) open eyes and with closed eyes, medio-lateral sway
treatment/6 velocity with open eyes and with closed eyes)

(Continude)
Clinical Rehabilitation 00(0)
Table 1. (Continude)
Reference Subjects number/gender/ Time since Stroke etiology/lesion Study design/sessions Intervention Results/used assessments
age (years) stroke location/affected number and duration/
hemisphere evaluations/PEDro scale
(score)
Veldema and Jansen

  RT2: unilateral concentric (1) knee extension on RT2 significantly better than CG: EMG activity (rectus
training machine (three sets of 8–10 repetitions, femoris, biceps femoris)
resistance 70% of 1RM, 3 min rest); preceded by
warm up (four repetitions with 25% of 1RM)
  CG: no intervention No significant differences between RT2 and CG: EMG
activity (gastrocnemius, tibialis anterior), balance
(anterio-posterior sway velocity with open eyes and
with closed eyes, medio-lateral sway velocity with
open eyes and with closed eyes
Flansbjer et al.25 24/14 males, 10 20 ± 10 18 ischemic, 6 Parallel groups (15+9)/20 RT: bilateral, concentric (1) knee flexion and (2) RT significantly better: dynamic force (knee flexion and
females/61 ± 5 years months haemorrhagic/lesion sessions à 90 min/T1 knee extension on training machine (two sets of extension), Timed Up and Go, Stroke Impact Scale
location na/16 right, before treatment, T2 after 6–8 repetitions, resistance 80% of 1RM, 2 min on T3; dynamic force (knee flexion and extension),
8 left treatment, T3 five months rest); preceded by warm up (5 min cycling + 10 isokinetic force (knee extension) on T4
after treatment, T4 four repetitions with 0%–25% of 1RM); followed by
years after treatment/6 passive static stretching
  No significant differences: isokinetic force (knee flexion
and extension), fast gait speed, Six Minute Walk Test,
Modified Ashworth Scale on T3; isokinetic force (knee
flexion), Timed Up and Go, fast gait speed, Six Minute
Walk Test, Stroke Impact Scale on T4
  CG: no intervention  
Aidar et al.21 24/15 males, 9 >12 months 24 ischemic, 0 Parallel groups (13+11)/36 RT: bilateral concentric (1) squat, (2) bench RT significantly better: State-Trait Anxiety Inventory,
females/52 ± 8 years haemorrhagic/lesion sessions à 45–60 min/T1 press, (3) leg press, (4) shoulder press, (5) Beck Depression Inventory, Short Form (36) Health
location na/affected before treatment, T2 after crunch, (6) pull-downs and (7) lunges on training Survey, maximal force (squat, bench-press, leg-press,
hemisphere na treatment/8 machines (three sets of 8–10 repetitions, shoulder-press, pull-downs and lunges)
resistance ‘somewhat hard’, 2 min rest);
preceded by warm up (10–15 min walking)
  CG: no intervention  
Akbari and Karimi22 34/19 males, 15 35 ± 27 Etiology na/lesion Parallel groups (17+17)/12 RT: unilateral concentric (1) hip flexion, (2) hip RT significantly better: Modified Ashworth Scale
females/49 ± 5 years months location na/21 right, sessions, duration na/T1 extension, (3) hip abduction, (4) knee flexion, (quadriceps femoris, gastrocnemius), isometric force
13 left before treatment, T2 after (5) knee extension, (6) ankle dorsiflexion and of affected side (hip flexion, knee flexion, ankle
treatment/7 (7) ankle plantar flexion, equipment na (sets and dorsiflexion)
repetitions na, resistance 70% of 1RM, rest na)
  CG: no intervention No significant differences: isometric force of affected
side (knee extension)
Fernandez-Gonzalo 29/22 males, 7 3.9 ± 4.3 years 20 ischemic, 9 Parallel groups (15+14)/24 RT: unilateral eccentric (1) leg press on training RT significantly better: Timed Up and Go, Berg Balance
et al.24 females/64 years haemorrhagic/17 sessions, à 35 min/T1 machine (four sets of seven repetitions, Scale, maximal dynamic force and powerpeak (both
subcortical, 12 before treatment, T2 after resistance ‘maximal effort’, 3 min rest); preceded quadriceps femoris), maximal isometric force (affected
cortical/15 right, 14 left treatment/7 by standardized warm up quadriceps femoris), muscle volume and cross-sectional
area (affected quadriceps femoris, vastus lateralis,
intermedius and medialis), Wechsler Adult Intelligence
Scale (digits span), Verbal Fluency Test, Short Form
(36) Health Survey (mental health, pain)
5

(Continude)
 6

Table 1. (Continude)
Reference Subjects number/gender/ Time since Stroke etiology/lesion Study design/sessions Intervention Results/used assessments
age (years) stroke location/affected number and duration/
hemisphere evaluations/PEDro scale
(score)

  CG: no intervention No significant differences: Modified Ashworth Scale,

Talking-While-Walking test, maximal isometric force
(non-affected quadriceps femoris), muscle volume
and cross sectional area (non-affected quadriceps
femoris, vastus lateralis, intermedius and medialis,
both rectus femoris), Short Form (36) Medical Health
Survey (physical functioning, physical limitations,
social function, emotional limitations, vitality, general
perception), Wechsler Adult Intelligence Scale
(spatial span), Conners Continuous Performance Test
– reaction time and sustained attention index, Rey
Auditory Verbal Learning Test, Stroop Color and
Word Test, Trail Making Test
Inaba et al.26 77/37 males, 40 na 36 ischemic, 17 Parallel groups RT: unilateral concentric (1) leg press on training RT1 significantly better than CG: activities of daily
females/57 years haemorrhagic, 14 na/ (28+23+26)/20–40 machine (sets na, five repetitions at resistance living on T2, maximal isometric force (knee extension)
lesion location na/42 sessions à 15 min/T1 50% of 10RM, 10 repetitions at resistance on T2
right, 35 left before treatment, T2 10RM, rest na)
during treatment, T3 after
treatment/8
  RT2: bilateral concentric (1) knee flexion and No significant differences between RT1 and CG:
extension, (2) hip abduction and adduction, activities of daily living on T3, maximal isometric force
(3) lower limb coordination exercises and (knee extension) on T3
(4) trunk flexion, extension and rotation on
training machine (sets na, repetitions na, without
resistance, rest na)
  CG: no intervention No significant differences between RT2 and CG:
activities of daily living, maximal isometric force (knee
extension)
Lee et al.27 20/13 males, 7 na 13 ischemic, 7 Parallel groups (10+10)/18 RT: laterality na, eccentric (1) hip flexion and RT significantly better: Timed Up and Go, Ten Meter
females/54 ± 10 years haemorrhagic/lesion sessions à 60 min/T1 (2) hip extension on training machine (four sets Walk Test, Stair Up and Down, muscular force (hip
location na/11 right, before treatment, T2 of eight repetitions, resistance na, rest 30 sec); flexion, hip extension)
9 left during treatment, T3 after preceded by warm up (5 min ergometer)
treatment/7
  CG: no intervention  

CG: control group; na: not available, not applicable; min: minute; RT: resistance training; T: test; RM: repetition maximum.
Clinical Rehabilitation 00(0)
Veldema and Jansen 7

Assessments Effect Lower Upper Rela

ve
size limit limit weight
Overall (n=276)
Sims et al., 2009 CES-D -0.28 -0.87 0.31 15.00
Lee et al., 2013c (KE) gait analysis 0.04 -0.73 0.81 8.67
Lee et al., 2013a TUG, 10MWT, SUD, MF 0.15 -0.72 1.03 6.67
Lee et al., 2013c (LP) gait analysis 0.23 -0.59 1.06 8.67
Fernandez-Gonzalo et al., 2016 TUG, BBS, MAS, SF-36, MF, MS, 0.24 -0.49 0.97 9.67
WAIS, FAS, RWW, CPT, RAVLT,
SCWT, TMT
Lee et al., 2013b (KE) standing analysis, EMG 0.53 -0.33 1.38 7.33
Flansbjer et al., 2012 TUG, 10MWT, gait speed, SIS, MF 0.62 -0.24 1.47 8.00
Akbari et al., 2006 MAS 0.88 0.18 1.58 11.33
Aidar et al., 2012 BDI, STAI, SF-6, MF 1.42 0.49 2.34 8.00
Lee et al., 2013b (LP) standing analysis, EMG 1.88 0.87 2.89 7.33
Lee et al., 2013d gait analysis 3.14 1.98 4.30 9.33
Total 0.75 -0.08 1.57 100.00
Subgroup heterogenity: I2=89.4%
-2 0 2 4 6
Gait (n=127)
Lee et al., 2013c (KE) gait analysis 0.04 -0.73 0.81 16.35
Flansbjer et al., 2012 TUG, 10MWT, gait speed 0.13 -0.70 0.96 18.87
Fernandez-Gonzalo et al., 2016 TUG 0.13 -0.60 0.86 18.24
Lee et al., 2013a TUG, 10MWT, SUD 0.19 -0.69 1.07 12.58
Lee et al., 2013c (LP) gait analysis 0.23 -0.59 1.06 16.35
Lee et al., 2013d gait analysis 3.14 1.98 4.30 17.61
Total 0.67 -0.19 1.54 100.00
Subgroup heterogenity: I2= 96.2%
-2 0 2 4 6
Muscular force and motor func…on of lower limbs (n=97)
Lee et al., 2013a MF (HF, HE) 0.10 -0.78 0.98 19.42
Fernandez-Gonzalo et al., 2016 MF (QF) 0.57 -0.17 1.31 28.16
Flansbjer et al., 2012 MF (KF, KE) 0.83 -0.04 1.69 29.13
Aidar et al., 2012 MF (S, BP, LP, SP, PD, L) 2.17 1.13 3.21 23.30
Total 0.93 0.05 1.80 100.00
Subgroup heterogenity: I2= 94.1%
-1 0 1 2 3 4
Other health-relevant physiological indicators (n=62)
Fernandez-Gonzalo et al., 2016 MS (QF, RF, VI, VL, VM) 0.14 -0.59 0.87 39.73
Lee et al., 2013b (KE) EMG (RF, BF, G, TA) 0.60 -0.26 1.47 30.14
Lee et al., 2013b (LP) EMG (RF, BF, G, TA) 1.43 0.49 2.36 30.14
Total 0.67 -0.16 1.50 100.00
Subgroup heterogenity: I2= 93.0%
-1 0 1 2 3
Quality of life, independence and reintegra…on (n=77)
Fernandez-Gonzalo et al., 2016 SF-36 0.32 -0.42 1.05 34.94
Flansbjer et al., 2012 SIS 0.39 -0.45 1.22 36.14
Aidar et al., 2012 SF-36 1.26 0.38 2.14 28.92
Total 0.62 -0.20 1.43 100.00
Subgroup heterogenity: I2= 88.2%
-1 0 1 2 3
Mobility, balance and postural control (n=73)
Lee 2013b (LP) standing analysis 0.45 -0.39 1.30 30.14
Fernarnderz-Gonzalo 2016 BBS 0.56 -0.18 1.30 39.73
Lee 2013b (KE) standing analysis 2.33 1.25 3.42 30.14
Total 1.06 0.19 1.94 100.00
Subgroup heterogenity: I2= 97.1%
-1 0 1 2 3 4
Spas…city and hypertonia (n=63)
Fernarnderz-Gonzalo 2016 MAS 0.11 -0.62 0.84 46.03
Akbari et al., 2006 MAS 0.88 0.18 1.58 53.97
Total 0.53 -0.19 1.24 100.00
Subgroup heterogenity: I2= 93.2%
-1 0 1 2
Emo…onal status (n=69)
Sims et al., 2009 CES-D -0.28 -0.87 0.31 65.22
Aidar et al., 2012 BDI, STAI 0.33 -0.48 1.14 34.78
Total -0.07 -0.73 0.60 100.00
Subgroup heterogenity: I2= 88.3%
-1 0 1 2
Cogni…ve abili…es (n=29)
Fernarnderz-Gonzalo 2016 WAIS, FAS, RWW, CPT, RAVLT, 0.23 -0.50 0.96 100.00
SCWT, TMT
-1 0 1 2
favours no interven…on favours resistance training

Figure 1.  (Continude)

8 Clinical Rehabilitation 00(0)

Figure 1.  Forest plot of studies comparing resistance training with no intervention in supporting stroke
recovery.
BBC: Berg Balance Scale; BDI: Beck Depression Inventory; BF: biceps femoris; BP: bench-press; G: gastrocnemius; CES-D:
Centre for Epidemiologic Studies for Depression scale; CPT: Conners Continuous Performance Test; FAS: Verbal Fluency
Test; HE: hip extension; HF: hip flexion; I2: inconsistency test; KE: knee extension; KF: knee flexion; L: lunges; LP: leg-press;
MAS: Modified Ashworth Scale; MF: muscular force; MS: muscle size; PD: pull-down; QF: quadriceps femoris; RAVLT: Rey
Auditory Verbal Learning Test; RF: rectus femoris; S: squat; SCWT: Stroop Color and Word Test; SF-36: Short Form (36)
Medical Health Survey; SIS: Stroke Impact Scale; SP: shoulder-press; STAI: State-Trait Anxiety Inventory; SUD: Stair Up and
Down; TA: tibialis anterior; TMT: Trail Making Test; TUG: Timed Up and Go; VI: vastus intermedius; VL: vastus lateralis;
VM: vastus medialis; TWW: Talking-While-Walking test; WAIS: Wechsler Adult Intelligence Scale; 10MWT: Ten Meter
Walk Test.

performed resistance exercises of upper part of the Depression scale, Satisfaction with Life
body.21,31 Most of the studies performed concentric Scale, Social Support Survey, Life
exercises. Eccentric training was applied in two tri- Orientation Test-Revised, measure of gener-
als only.24,27 alized dispositional optimism, Self-Esteem
Parameters assessed: The effectiveness of resist- Scale, Recovery Locus of Control Scale).
ance training has been evaluated on: - Cognitive abilities24 (Wechsler Adult
Intelligence Scale, Verbal Fluency Test,
- Gait:23,24,27,29,30 (gait velocity, step length, Conners Continuous Performance Test, Rey
stride length, double support, heel-to-heel of Auditory Verbal Learning Test, Stroop Color
support, plantar pressure distribution, Timed and Word Test, Trail Making Test).
Up and Go, Ten Meter Walk Test, Stair Up - Other health-relevant physiological indica-
and Down, Six Minute Walk Test, Stair tors24,28 (EMG activity, volume and cross-
Climbing Test). sectional area of lower limbs muscles).
- Muscular force and motor function of lower
limbs21–27 (muscular force, dynamic force, Effectiveness: Collectively, the data indicates
isokinetic force, maximal force, isometric that resistance training is superior to no interven-
force, power peak of lover extremity muscles tion in supporting recovery after stroke. Most of
(quadriceps femoris, knee flexion and exten- the studies investigating muscular force of lower
sion, ankle flexion and extension, hip flexion limbs, quality of life, independence and reintegra-
and extension, squat, bench-press, leg-press, tion, mobility, balance and postural control and
shoulder-press, pull-downs and lunges)). other health-relevant physiological indicators dem-
- Quality of life, independence and reintegra- onstrate significant effects of resistance training on
tion21,23–26,31 (Short Form (36) Health Survey, parameters assessed. For gait, spasticity and hyper-
Short Form (12) Health Survey, Stroke tonia and emotional state, only some data shows a
Impact Scale, Assessment of Quality of Life better improvement with resistance training in
Instrument, Stroke Specific Quality of Life comparison to no intervention. Eccentric training
Scale, Functional Independence Measure, leads to a smaller improvement of parameters
activities of daily living). assessed than concentric exercises. No evident dif-
- Mobility, balance and postural control23,24,28 ferences were detected between unilateral und
(Berg Balance Scale, sway velocity with bilateral training.
open eyes and with closed eyes, Rivermead
Mobility Index). Resistance training versus other
- Spasticity and hypertonia22,24 (Modified
Ashworth Scale).
intervention
- Emotional state21,31 (State-Trait Anxiety Fifteen trials compared the effectiveness of resist-
Inventory, Beck Depression Inventory, ance training with other interventions in stroke
Centre for Epidemiologic Studies for rehabilitation (Table 2, Figure 2).32–46
Table 2.  Overview of studies investigating resistance training in comparison to other interventions in supporting stroke recovery.
Reference Subjects number/ Time since Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) stroke location/affected and duration/evaluations/PEDro
hemisphere scale (score)

Knox et al.40 144/72 males, 72 9 ± 7 weeks Etiology na, lesion Parallel groups (45+51+48)/6 RT: 10 laterality na, concentric (1) lower limb CT1 significantly better than RT and CT2: Six Minute
females/50 ± 14 years location na, affected sessions à 60 min (RT, CT1) or 1 exercises using gravity, free weights, elastic bands Walk Test, Timed Up and Go, Berg Balance Scale,
hemisphere na session à 90 min (CT2)/T1 before and balls (three sets of 10 repetitions, resistance habitual gait speed, fast gait speed
treatment, T2 after treatment, T3 na, rest na)
Veldema and Jansen

12 weeks after treatment/8

  CT1: Six task exercises focusing strength, balance  
and task performance while standing and walking
  CT2: education of stroke management + exercises RT significantly better than CT2: Six Minute Walk
Test (T2)
36
Glasser 20/10 males, 10 3–6 months Etiology na/lesion Parallel groups (10+10)/25 RT: bilateral concentric (1) lower extremity No significant differences: Functional Ambulation
females/40–75 years location na/10 right, sessions à 10–30 min/T1 before exercise on training machine (6–10 sets of 25 Profile Score, gait speed
10 left treatment, T2 after treatment/5 repetitions, resistance na, rest 1 min)
  CT: therapeutic exercise program and gait training  
Kim and Lee39 24/12 males, 12 10 months Etiology na/lesion Parallel groups (12+12)/20 RT: laterality na, concentric (1) lunges using CT significantly better: gait parameters (gait speed,
females/47 years location na/11 right, sessions, duration na/T1 before bodyweight (five sets of 10 repetitions, resistance cadence, step length – affected side, stride length –
13 left treatment, T2 after treatment/7 na, rest na) affected side, single-support time – affected side)
  CT: Mulligas’s mobilization (1) ankle dorsoflexion No significant differences: gait parameters (step
(five sets of 10 repetitions) length – non-affected side, stride length – non-
affected side)
Marzolini et al.42 68/44 males, 24 12 ± 11 53 ischemic, 12 Parallel groups (33+35)/42 RT: bilateral, concentric (1) lunges, (2) squats, RT significantly better: Sit-to-Stand Test, total body
females/64 ± 12 months haemorrhagic, 3 na/ sessions à 20–60 min/T1 before (3) abdominal curl-up, (4) heel raise, (5) biceps lean mass, ergometer/treadmill test (oxygen uptake
(34–92) years lesion location na/36 treatment, T2 after treatment/7 curl, (6) supine triceps extension and unilateral at the ventilatory threshold), total body percentage
right, 30 left, 2 bilateral concentric, (7) hip flexion, (8) hip extension, (9) fat, maximal isometric force (elbow flexion and
ankle dorsiflexion, (10) knee extension and (11) extension – affected and non-affected side, knee
knee flexion using dumbbell, elastic band, own extension – non-affected side)
body weight and training machines (1–2 sets of
10–15 repetitions, resistance 50%–70% of 1RM/
perceived exertion ‘hard’, rest na)
  CT: aerobic training (walking or ergometer), (1) No significant differences: Six Minute Walk Test,
moderate intensity continuous training (intensity stair climbing, ergometer/treadmill test (HRrest,
60%–80% of HR reserve/perceived exertion systolic BPpeak, diastolic BPpeak, oxygen uptakepeak,
‘somewhat hard’ to ‘hard’) alternating with (2) respiratory exchange ratiopeak), Body Mass Index,
high intensity interval training (15–30 sec of high maximal isometric force (knee extension – affected
intensity/perceived exertion ‘very hard’ alternating side)
with 1–2 min of low intensity)
Zou et al.46 56/22 males, 34 12 ± 6 months 37 ischemic, 19 Parallel groups (28+28)/24 RT: bilateral concentric (1) leg press (2) knee RT significantly better: maximal concentric force
females/52 ± 7 years haemorrhagic/31 sessions à 40 min/T1 before extension and (3) knee flexion on training (leg press – affected and non-affected side, knee
subcortical, 25 treatment, T2 after treatment/8 machines (three sets of 15 repetitions, resistance extension – affected and non-affected side), fasting
cortical/32 right, 24 left causing muscle failure between repetitions 10 and insulin, homeostasis model assessment of insulin
12, rest na) resistance, two-hour blood glucose levels, total
cholesterol, high-density lipoprotein cholesterol,
low-density lipoprotein cholesterol
  CT: active and passive stretching exercises of the No significant differences: Fugl-Meyer Assessment
upper and lower body – upper extremities, lower extremities, Body Mass
Index, fasting glucose, glycosylated haemoglobin,
total triglycerides

(Continude)
9
 10

Table 2. (Continude)
Reference Subjects number/ Time since Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) stroke location/affected and duration/evaluations/PEDro
hemisphere scale (score)

Patten et al.44 19/15 males, 4 13 ± 4 months 14 ischemic/4 Crossover (19−19)/24 (12+12) RT: unilateral eccentric and concentric (1) RT significantly better: Wolf Motor Function Test,
females/69 ± 10 years haemorrhagic, 1 sessions à 35 min/T1 before shoulder abduction-adduction, (2) shoulder Functional Independence Measure, isometric joint
others/5 subcortical, 9 first treatment, T2 after first flexion-extension, (3) shoulder external-internal torque, EMG activity, maximal voluntary contraction,
cortical, 5 na/10 right, treatment, T3 before second rotation and (4) transverse plane elbow flexion- joint power and stretch reflex of affected upper
9 left treatment, T4 after second extension on training machine (three sets (1. extremity (elbow flexion and extension, shoulder
treatment/8 set concentric, 2. and 3. set eccentric) of 10 flexion, abduction and external rotation)
repetitions, resistance na, rest na)
  CT: functional task training of upper limbs No significant differences: Fugl-Meyer Assessment –
upper extremities, Modified Ashworth Scale
Corti et al.33 14/12 males, 2 15 ± 7 months 12 ischemic, 2 Crossover (14−14)/60 (30+30) RT: laterality na, concentric and eccentric (1) No significant differences: Fugl-Meyer Assessment –
females/60 ± 15 years haemorrhagic/5 sessions à 90 min/T1 before shoulder abduction-adduction, (2) shoulder upper extremities, European Stroke Scale, Chedoke
subcortical, 9 cortical/6 treatment, T2 after first flexion-extension, (3) shoulder external-internal Mc Master Hand and Arm Inventory, Reintegration
right, 8 left treatment, T3 after second rotation, (4) transverse plane elbow flexion- to Normal Living index, kinematics of functional
treatment/8 extension and (5) wrist flexion-extension on reach to grasp
training machine (three sets of 10 repetitions,
resistance na, rest na)
  CT: functional task training of upper limbs  
Severinsen et al.45 43/31 males, 12 16 (8–38) 43 ischemic, 0 Parallel groups (14+16+13)/36 RT1: bilateral concentric (1) leg press, (2) hip RT1 significantly better than CT: isometric force of
females/68 (50–80) months haemorrhagic/lesion sessions à 60 min/T1 before extension, (3) hip flexion, (4) knee extension, (5) the affected and of the non-affected knee extensors,
years location na/19 right, treatment, T2 after treatment, T3 ankle dorsal flexion and (6) ankle plantar flexion on Ten Meter Walk Test (at T3)
24 left one year after treatment/8 training machines (three sets of eight repetitions,
resistance 80% of 1RM, rest na); preceded by
warm up (5 min ergometer)
  CT significantly better than RT1 and RT2: graded
test on ergometer (oxygen uptakepeak; at T2)
  RT2: bilateral concentric (1) elbow flexion, (2) No significant differences between RT and CT: Six
elbow extension, (3) shoulder abduction and Minute Walk Test, Ten Meter Walk Test (after
(4) combined shoulder movement using a pulley treatment), graded test on ergometer (oxygen
(three sets of 15 repetitions, resistance 60% uptakepeak; at T3)
of 1RM, rest na); preceded by warm up (5 min
ergometer)
  CT: ergometer training at 75% of heart rate  
reserve (3 × 15 min); preceded by warm up (5 min
ergometer)
Folkerts et al.34 11/9 males, 2 23 ± 9 months 10 ischemic, 1na/lesion Crossover (11−11)/24 (12+12) RT: unilateral eccentric (1) shoulder flexion, (2) CT significantly better: maximal isometric force of
females/56 ± 11 location na/5 right, 6 left sessions, à 30–60 min/T1 shoulder extension, (3) shoulder adduction, (4) the affected side (elbow flexion and extension)
(35–73) years before treatment, T2 after first shoulder abduction, (5) shoulder external rotation,
treatment, T3 after second (6) shoulder internal rotation, (7) elbow flexion,
treatment/6 (8) elbow extension, (9) wrist palmar flexion
and (10) wrist dorsal flexion using dumbbells
and rubber bands (two to three sets of 5–15
repetitions, resistance 0–2 kg dumbbell, extra-
extra light–light rubber bands, rest na)

(Continude)
Clinical Rehabilitation 00(0)
Table 2. (Continude)
Reference Subjects number/ Time since Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) stroke location/affected and duration/evaluations/PEDro
hemisphere scale (score)

  CT: bilateral computer-based task training of No significant differences: Action Research Arm
Veldema and Jansen

upper limbs Test, Intrinsic Motivation Inventory, maximal

isometric force of the affected side (shoulder flexion,
extension, adduction, abduction, external rotation
and internal rotation and wrist palmar flexion and
dorsal flexion)
Ouellette et al.43 42/gender na/45 ± 13 29 ± 4 months Etiology na/lesion Parallel groups (21+21)/36 RT: bilateral concentric (1) leg press, (2) knee RT significantly better: maximal concentric force (leg
years location na/affected sessions, duration na/T1 before extension, (3) ankle dorsiflexion and (4) ankle press, knee extension – affected and non-affected
hemisphere na treatment, T2 after treatment/6 plantar flexion using training machine and weight side, ankle dorsiflexion – affected side, ankle plantar
stack-pulley system (three sets of 8–10 repetitions, flexion – affected side)
resistance 70% of 1RM, rest na); preceded by
warm up (four repetitions with 25% of 1RM)
  CT: upper body stretching exercises No significant differences: maximal concentric force
(ankle dorsiflexion – non-affected side, ankle plantar
flexion – non-affected side), Six Minute Walk Test,
stair climbing test, Sit-to-Stand Test, maximal gait
velocity, habitual gait velocity, Late Life Function and
Disability Instrument
Coroian et al.32 20/16 males, 4 16 ischemic, 4 Parallel groups (10+10)/18 RT: unilateral concentric (1) elbow flexion- No significant differences: Fugl-Meyer Assessment
females/64 ± 12 years haemorrhagic/20 sessions à 45 min/T1 before extension (2) wrist flexion-extension using – upper extremities, Box and Block Test, Modified
subcortical/4 right, treatment, T2 after treatment, T3 dynamometer (six sets of eight repetitions, Ashworth Scale, concentric and eccentric strength
16 left three months after treatment, T4 resistance increased progressively from 40% to (elbow flexion and extension, wrist palmar and
six months after treatment/8 70% of 1RM, 30 sec rest); preceded by warm up dorsal flexion – affected side)
(three sets of six repetitions with 20% of 1RM)
  CT: passive mobilization unilateral (1) elbow  
flexion-extension (2) wrist flexion-extension
Kim et al.38 20/14 males, 6 4.1 ± 2.3 years 11 ischemic, 7 Parallel groups (10+10)/18 R: unilateral concentric (1) hip flexion-extension, No significant differences: maximal concentric
females/61 ± 9 years haemorrhagic, 2 na/ sessions à 30 min/T1 2–4 days (2) knee flexion-extension and (3) ankle force of lower extremities, gait speed (self-selected
lesion location na/9 before treatment, T2 2–4 days dorsiflexion-plantar flexion on training machine and maximal), stair climbing speed (self-selected,
right, 11 left after treatment/9 (three sets of 10 repetitions, resistance ‘maximal maximal), Short Form (36) Health Survey
effort’, rest na)
  CT: passive range of motion exercises on training  
machine
Lee et al.41 48/28 males, 20 4.8 ± 4.5 years 33 ischemic, 6 Parallel groups RT active + CT active: bilateral concentric RT active significantly better than CT active: stair
females/63 ± 9 years haemorrhagic, 6 other/ (12+12+12+12)/30 sessions à (1) leg press, (2) knee extension, (3) knee climbing power, concentric force of lower limb –
lesion location na/27 60 min/T1 before treatment, T2 flexion and (4) ankle plantar flexion on training affected and non-affected side, concentric endurance
right, 21 left after treatment/8 machines, and bilateral isotonic (5) abduction of lower limb – affected side
and (6) dorsiflexion using free weights (two sets
of eight repetitions, resistance 50% of 1RM/
perceived exertion ‘very hard’, rest na; 30 min),
+ wheelchair ergometer training (50%–70% of
oxygen uptakepeak (30 min)

(Continude)
11
 12

Table 2. (Continude)
Reference Subjects number/ Time since Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) stroke location/affected and duration/evaluations/PEDro
hemisphere scale (score)

  RT active + CT passive: bilateral concentric (1) CT active significantly better than RT active: graded
leg press, (2) knee extension, (3) knee flexion and test on ergometer (oxygen uptakepeak), treadmill
(4) ankle plantar flexion on training machines, and walking (physical cost index), Ewart’s physical self-
bilateral isotonic (5) abduction and (6) dorsiflexion efficacy scales
using free weights (two sets of eight repetitions,
resistance 50% of 1RM/perceived exertion ‘very
hard’, rest na; 30 min), + wheelchair ergometer
training passive (30 min)
  RT passive + CT active: bilateral concentric (1) No significant differences between RT active and
leg press, (2) knee extension, (3) knee flexion and CT active: Six Minute Walk Test, fast and habitual
(4) ankle plantar flexion on training machines, and gait velocity, graded test on ergometer (power
bilateral isotonic (5) abduction and (6) dorsiflexion outputpeak, heart ratepeak), treadmill walking (oxygen
using free weights (two sets of eight repetitions, cost), concentric endurance of lower limb – non-
without resistance, rest na; 30 min), + wheelchair affected side, Short Form (36) Health Survey
ergometer training at 50%–70% of peak oxygen
uptake (30 min)
  RT passive + CT passive: bilateral concentric (1)  
leg press, (2) knee extension, (3) knee flexion and
(4) ankle plantar flexion on training machines, and
bilateral isotonic (5) abduction and (6) dorsiflexion
using free weights (two sets of eight repetitions,
without resistance, rest na; 30 min), + wheelchair
ergometer training passive (30 min)
Ivey et al.37 30/21 males, 9 5.5 ± 4.5 years Etiology na/lesion Parallel groups (14+16)/30 RT: bilateral concentric (1) leg press, (2) knee RT significantly better: Six Minute Walk Test, Ten
females/56 ± 12 years location na/affected sessions à 45 min/T1 before extension and (3) knee flexion on training Meter Walk Test (fastest comfortable speed),
hemisphere na treatment, T2 after treatment/7 machines (two sets of 20 repetitions, resistance VO2max, maximal concentric force (leg press –
causing muscle failure between repetitions 10 and affected and non-affected side)
15, rest na)
  CT lower body stretching exercises No significant differences: Ten Meter Walk Test
(self-selected speed)
35
Gambassi et al. 22/9 males, 13 5.8 ± 4.6 years Etiology na/lesion Parallel groups (11+11)/16 RT: bilateral concentric (1) seated row, (2) squat RT significantly better: Ten Meter Walk Test, Timed
females/62 ± 11 years location na/16 right, sessions, duration na/T1 before on the chair, (3) vertical chest press and (4) knee Up and Go, Sit-to-Stand Test, isometric force (hand
6 left treatment, T2 after treatment/6 extension using elastic bands and ankle wrist grip – affected and non-affected), heart rate, double
weights (three sets of 6–12 repetitions, perceived product, heart rate variability, oxidative stress
exertion ‘moderate to hard’, rest na) markers
  CT: neurological physical therapy (ADL, balance, No significant differences: systolic blood pressure,
gait) diastolic blood pressure

BP: blood pressure; CT: control therapy; HR: heart rate; na: not available, not applicable; min: minute; RT: resistance training; sec: second; T: test; RM: repetition maximum; ADL: activities of daily
living; VO2: oxygen consumption.
Clinical Rehabilitation 00(0)
Veldema and Jansen 13

Assessments Effect Lower Upper Relave

size limit limit weight
Overall (n=567)
Kim et al., 2016 gait analysis -0.75 -1.59 0.09 3.68
Knox et al., 2018 (TE) 6MWT, TUG, gait speed, BBS -0.25 -0.65 0.15 14.70
Glasser et al., 1986 FAPS, gait speed -0.22 -1.10 0.66 3.06
Folkerts et al., 2017 ARAT, MF -0.14 -1.02 0.74 3.06
Kim et al., 2001 gait speed, stair climbing speed, SF- -0.03 -0.92 0.86 3.06
36, MF
Severinsen et al., 2014 (UL) 6MWT, 10MWT, ergometer test, MF 0.04 -0.70 0.78 4.44
Coroian et al., 2017 FM-UL, B&B, MF, MAS 0.19 -0.69 1.07 3.06
Knox et al., 2018 (E) 6MWT, TUG, gait speed, BBS 0.21 -0.19 0.62 14.24
Marzolini et al., 2018 6MWT, stair climbing speed, 0.35 -0.13 0.84 10.41
ergometer/treadmill test, MF, BMI,
body lean mass, body fe
Lee et al., 2008 6MWT, gait speed, stair climbing 0.38 -0.47 1.23 7.35
power, ergometer test, treadmill
test, MF
Ouellee et al., 2004 6MWT, gait speed, stair climbing 0.41 -0.21 1.02 6.43
speed, LLFDI, Sit-to-Stand-Test, MF
Severinsen et al., 2014 (LL) 6MWT, 10MWT, ergometer test, MF 0.43 -0.36 1.22 4.13
Paen et al., 2013 WMFT,FM-UL, joint torque, FIM, 0.52 -0.13 1.17 5.82
MAS, EMG
Gambassi et al., 2019 TUG, Sit-to-Stand-Test, MF, HR, BP, 0.87 -0.02 1.76 3.37
HRV, OSM
Zhou et al., 2015 FM-UL, FM-LL, MF, BMI, blood 1.08 0.51 1.66 8.58
glucose level, serum lipids profiles
Ivey et al., 2017 10MWT, 6MWT, MF, treadmill test 1.30 0.49 2.11 4.59
Total 0.28 -0.35 0.94 100.00
Subgroup heterogenity: I2=32.2%
-2 -1 0 1 2 3
Gait (n=461)
Kim et al., 2016 gait analysis -0.75 -1.59 0.09 4.62
Kim et al., 2001 gait speed, stair climbing speed -0.51 -1.40 0.39 3.85
Knox et al., 2018 (TE) 6MWT, TUG, gait speed -0.25 -0.65 0.16 18.50
Glasser et al., 1986 FAPS, gait speed -0.22 -1.10 0.66 3.85
Lee et al., 2008 6MWT, gait speed, stair climbing -0.19 -0.99 0.62 9.25
power
Marzolini et al., 2018 6MWT, stair climbing speed -0.08 -0.56 0.39 13.10
Ouellee et al., 2004 6MWT, gait speed, stair climbing 0.08 -0.52 0.69 8.09
speed
Severinsen et al., 2014 (LL) 6MWT, 10MWT 0.10 -0.66 0.85 5.20
Severinsen et al., 2014 (UL) 6MWT, 10MWT 0.18 -0.56 0.91 5.59
Knox et al., 2018 (educaon) 6MWT, TUG, gait speed 0.22 -0.19 0.63 17.92
Gambassi et al., 2019 TUG 0.34 -0.50 1.18 4.24
Ivey et al., 2017 10MWT, 6MWT 0.76 0.01 1.52 5.78
Total -0.02 -0.62 0.58 100.00
Subgroup heterogenity: I2= 0.0%
-2 -1 0 1 2
Muscular force and motor funcon of lower limbs (n=329)
Severinsen et al., 2014 (UL) MF (KE) 0.27 -0.47 1.00 8.48
Marzolini et al., 2018 MF (KE) 0.42 -0.06 0.90 19.88
Kim et al., 2001 MF (HF, HE, KF, KE, ADF, APF) 0.63 -0.27 1.53 5.85
Ouellee et al., 2004 MF (KF, KE, ADF, APF) Sit-to-Stand- 0.74 0.10 1.37 12.28
Test
Gambassi et al., 2019 Sit-to-Stand-Test 1.16 0.26 2.06 6.43
Zhou et al., 2015 FM-LL, MF (LP, KE) 1.25 0.68 1.83 16.37
Severinsen et al., 2014 (LL) MF (KE) 1.27 0.44 2.11 7.89

Lee et al., 2008 MF (HE, KF, KE, ADF, APF) 1.34 0.43 2.24 14.04
Ivey et al., 2017 MF (LP) 2.17 1.26 3.07 8.77
Total 0.99 0.28 1.70 100.00
Subgroup heterogenity: I2= 71.6%
-1 0 1 2 3 4
favours other intervenon favours resistance training

Figure 2.  (Continude)

14 Clinical Rehabilitation 00(0)

Effect Lower Upper Relave

size limit limit weight
Muscular force and motor funcon of upper limbs (n=385)
Folkerts et al., 2017 ARAT, MF (SF, SE, SAB, SAD, SER, -0.14 -1.02 0.74 8.93
SIR, EF, EE, WPF, WDF)
Zhou et al., 2015 FM-UL 0.14 -0.39 0.66 25.00
Coroian et al., 2017 FM-UL, B&B, MF (EF, EE, WPF, WDF) 0.22 -0.66 1.11 8.93
Gambassi et al., 2019 MF (handgrip) 0.31 -0.54 1.15 9.82
Pa–en et al., 2013 WMFT,FM-UL, joint torque (EF, EE, 0.64 -0.02 1.30 16.96
SF, SAB, SER)
Marzolini et al., 2018 MF (EF) 1.15 0.63 1.66 30.36
Total 0.53 -0.11 1.17 100.00
Subgroup heterogenity: I2= 76.1%
-2 -1 0 1 2
Cardiorespiratory fitness (n=91)
Severinsen et al., 2014 (UL) ergometer test (VO2max) -0.67 -1.43 0.08 14.36
Severinsen et al., 2014 (LL) ergometer test (VO2max) -0.61 -1.38 0.16 13.37
Lee et al., 2008 ergometer test (powermax, HRmax, -0.31 -1.12 0.50 23.76
VO2max), treadmill test (physical cost
index, oxygen cost)
Marzolini et al., 2018 ergometer/treadmill test (HRrest, 0.24 -0.24 0.71 33.66
BPrest, VO2max, VO2VT, respiratory
exchange rao)
Ivey et al., 2017 treadmill test (VO2max) 1.18 0.41 1.96 14.85
Total -0.25 -0.65 0.15 100.00
Subgroup heterogenity: I2= 90.7%
-2 -1 0 1 2 3
Other health-relevant physiological indicators (n=195)
Marzolini et al., 2018 BMI, body lean mass, body fat 0.29 -0.19 0.77 36.96
Pa–en et al., 2013 EMG (BB, TB, DA, DM, DP, IS, BR, PM) 0.57 -0.08 1.22 20.65
Gambassi et al., 2019 HR, BP, HRV, OSM 0.94 0.04 1.84 11.96
Zhou et al., 2015 BMI, blood glucose level, serum 1.07 0.50 1.64 30.43
lipids profiles
Total 0.66 0.07 1.25 100.00
Subgroup heterogenity: I2= 73.2%
-1 0 1 2
Quality of life, independence and reintegraon (n=81)
Kim et al., 2001 SF-36 0.11 -0.62 0.84 20.00
Pa–en et al., 2013 FIM 0.56 -0.09 1.21 38.00
Ouelle–e et al., 2003 LLFDI 0.88 0.18 1.58 42.00
Total 0.60 -0.08 1.29 100.00
Subgroup heterogenity: I2= 75.6%
-1 0 1 2
Spascity and hypertonia (n=39)
Pa–en et al., 2013 MAS -0.05 -0.68 0.59 48.72
Coroian et al., 2017 MAS 0.00 -0.88 0.88 51.28
Total -0.02 -0.78 0.74 100.00
Subgroup heterogenity: I2= 0.0%
-1 0 1
Mobility, balance and postural control (n=144)
Knox et al., 2018 (TE) BBS -0.22 -0.62 0.19 50.79
Knox et al., 2018 (E) BBS 0.07 -0.34 0.48 49.21
Total -0.07 -0.48 0.33 100.00
Subgroup heterogenity: I2= 51.6%
-1 0 1
favours other intervenon favours resistance training

Figure 2.  (Continude)

Veldema and Jansen 15

Figure 2.  Forest plot of studies comparing resistance training with other interventions in supporting stroke
recovery.
ADF: ankle dorsi flexion; APF: ankle plantar flexion; ARAT: Action Research Arm Test; BB: biceps brachii; BBC: Berg Balance
Scale; BMI: body mass index; BP: blood pressure; BR: brachioradialis; B&B: Box and Block Test; DA: deltoid anterior; DM: deltoid
middle; DP: deltoid posterior; E: education; EE: elbow extension; EF: elbow flexion; FAPS: Functional Ambulation Profile Score;
FIM: Functional Independence Measure; FM-UL: Fugl-Meyer Assessment – upper extremities; HE: hip extension; HF: hip flexion;
HR: heart rate; HRV: heart rate variability; IS: infraspinatus; I2: inconsistency test; KE: knee extension; KF: knee flexion; LL: lower
limbs; LLFDI: Late Life Function and Disability Instrument; LP: leg-press; MAS: Modified Ashworth Scale; MF: muscular force;
OSM: oxidative stress markers; PM: pectoralis major; SAB: shoulder abduction; SAD: shoulder adduction; SE: shoulder extension;
SER: shoulder external rotation; SF: shoulder flexion; SIR: shoulder internal rotation: TB: triceps brachii; TE: task exercises; TUG:
Timed Up and Go; UL: upper limbs; VO2max: maximal oxygen uptake; VO2VT: oxygen uptake at the ventilatory threshold; WDF:
wrist dorsal flexion; WPF: wrist palmar flexion; 6MWT: Six Minute Walk Test; 10MWT: Ten Meter Walk Test.

Participants and interventions: A total of 581 Action Research Arm Test, Wolf Motor
patients were randomized between nine weeks and Function Test, Box and Block Test, con-
5.8 years after incident. The probands received centric, eccentric and isometric force of
between six and 66 sessions of unilateral (affected diverse upper limb muscles (elbow flexion
side)32,34,38,42,44 or bilateral (both sides)35–37,41–43,45,46 and extension, wrist flexion and extension,
resistance training. Three studies did not specify hand grip, shoulder flexion, extension,
the trained side.31,39,40 Most of the studies exclu- adduction, abduction, external and internal
sively applied lower body resistance training. Only rotation, wrist palmar and dorsal flexion),
five studies evaluated the effectiveness of upper isometric joint torque, maximal voluntary
body exercises.31,32,34,40,45 Two studies used com- contraction, joint power of several upper
bined training of the upper and lower body.35,42 limb muscles (elbow flexion and exten-
Concentric exercises were applied in the major part sion, shoulder flexion, abduction and
of the trials. Only one trial investigated the effects external rotation), kinematics of functional
of eccentric exercises.34 Two trials combined con- reach to grasp).
centric and eccentric training.31,44 One study per- - Cardiorespiratory fitness41,42,45 (ergometer
formed concentric and isotonic training.41 and treadmill test (physical cost index, oxy-
Parameters assessed: The studies tested: gen cost, power output peak, heart rate peak,
oxygen uptake peak, oxygen uptake at the
- Gait35–43,45 (Six Minute Walk Test, Ten Meter ventilatory threshold, heart rate rest, systolic
Walk Test, Functional Ambulation Profile blood pressure peak, diastolic blood pressure
Score, Timed Up and Go, fast and habitual peak, respiratory exchange ratio peak)).
gait speed, gait cadence, step length, stride - Quality of life, independence and reintegra-
length, single-support time, stair climbing tion31,41, 43,44 (Short Form (36) Health Survey,
test). Late Life Function and Disability Instrument,
- Muscular force and motor function of lower Ewart’s physical self-efficacy scales,
limbs35,37,38,41–43,45,46 (Fugl-Meyer Assessment Functional Independence Measure,
lower extremities, Sit-to-Stand Test, isomet- Reintegration to Normal Living index).
ric force, concentric force, concentric endur- -  Mobility, balance and postural control40
ance of diverse lower limb muscles (knee (Berg Balance Scale).
flexion and extension, ankle dorsiflexion and -  Spasticity and hypertonia32,44 (Modified
plantar flexion, leg-press). Ashworth Scale, stretch reflex).
- Muscular force and motor function of - Neurological deficits31 (European Stroke
upper limbs31,32,34,35,42,44,46 (Chedoke Mc Scale).
Master Hand and Arm Inventory, Fugl- - Emotional status34 (Intrinsic Motivation
Meyer Assessment upper extremities, Inventory).
16 Clinical Rehabilitation 00(0)

- Other health-relevant physiological indica- the effects of different resistance exercises (leg
tors35,37,42,44,46 (glycosylated haemoglobin, press vs knee extension,28,29 leg press vs diverse
total triglycerides, fasting insulin, fasting upper limbs exercises,26 lower body vs upper body
glucose, homeostasis model assessment of exercises45). Three studies evaluated the effects of
insulin resistance, two-hour blood glucose different muscle contraction types (concentric vs
levels, total cholesterol, high-density lipo- eccentric48,49 and concentric + eccentric vs iso-
protein cholesterol, low-density lipopro- metric47) on the effectiveness of identical exer-
tein cholesterol, total body lean mass, total cises. Three trials compared the influence of
body percentage fat, Body Mass Index, training intensity (high vs low) during identical41,50
EMG activity, heart rate, double product, or different26 exercises, on supporting stroke
heart rate variability, oxidative stress mark- recovery. Both, unilateral (affected side)26,28,29,48–50
ers, systolic blood pressure, diastolic blood as well as bilateral (both sides)26,41,45,47 training
pressure). were applied.
Parameters assessed: The studies evaluated:
Effectiveness: Collectively, the data indicates
that resistance training is more effective in sup-
porting the recovery after stroke than most other - Gait26,29,41,45,47–50 (Six Minute Walk Test, Ten
therapies. Especially muscular force and motor Meter Walk Test, Timed Up and Go, Stair
function of lower and upper limbs, quality of life, Climbing Power, fast, self-selected and
independence and reintegration and other health habitual walking speed, plantar pressure dis-
relevant physiological indicators improve with tribution, swing phase duration).
resistance training significantly better than with - Muscular force and motor function of lower
other interventions. Only for cardiorespiratory fit- limbs26,41,45,47–49 (isometric force, dynamic
ness, resistance training shows inferior effects force, eccentric and concentric force of
than ergometer training. Regarding the improve- lower limb muscles (knee flexion and
ment of the gait ability, resistance training is extension)).
superior to stretching training and stroke manage- - Mobility, balance and postural control28,49,50
ment education, but less efficient than passive (Berg Balance Scale, anterio-posterior and
mobilization and range of motion exercises. The medio-lateral postural balance with open
direct comparison with therapeutic gait and bal- eyes and with closed eyes, body weight dis-
ance training shows equivocal results. The current tribution in rising and in sitting down).
data indicates no differential effects of resistance - Cardiorespiratory fitness41,45 (ergometer and
training in dependence on trained side (unilateral, treadmill test (power output peak, heart rate
bilateral) or muscle contraction types (eccentric, peak, oxygen uptake peak, physical cost
concentric). index, oxygen cost)).
- Quality of life, independence and reintegra-
tion26,41,47 (Short Form (36) Health Survey,
Resistance training versus other Ewart’s physical self-efficacy scales).
resistance training - Other health-relevant physiological indica-
tors28,47–49 (EMG activity, tumour necrosis
Nine studies compared the effectiveness of differ- factor-α, high sensitivity C-reactive protein,
ent resistance training protocols in rehabilitation interleukin-6).
after stroke (Table 3, Figure S1).26,28,29,41,45,47–50
Participants and interventions: A total of 286
patients between three months and five years after Effectiveness: The data indicates that the type of
acute incident were enrolled. About 12–40 train- resistance training protocol may significantly
ing sessions were applied. Four studies compared impact its effect on recovery after stroke. Generally,
Table 3.  Overview of studies comparing different resistance training protocols in supporting stroke recovery.
Reference Subjects number/ Time since stroke Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) location/affected hemisphere and duration/evaluations/PEDro
scale (score)

Chen et al.47 24/13 males, 11 3.2 ± 1.5 months 20 ischemic, 4 Parallel groups (12+12)/20 RT1: bilateral concentric and eccentric (1) knee RT1 significantly better: dynamic force
Veldema and Jansen

females/66 ± 13 haemorrhagic/19 subcortical, sessions, duration na/T1 before flexion-extension on training machine (three sets (knee flexion and extension – affected side),
years 5 cortical/12 right, 12 left treatment, T2 after treatment/6 of five repetitions concentric, followed from five isometric force (knee flexion – non-affected
repetitions eccentric, resistance na, rest na) side), tumour necrosis factor-α, high sensitivity
C-reactive protein
  RT2: bilateral isometric (1) knee flexion and (2) No significant differences: isometric force
knee extension on training machine (three sets of (knee extension – non-affected side), Short
10 repetitions, resistance 60% of 1RM, rest na) Form (36) Health Survey, Timed Up and Go,
interleukin-6
Clark and Patten48 34/26 males, 9 12 ± 5 (6–18) Etiology na/lesion location Parallel groups (16+18)/15 RT1: unilateral concentric (1) ankle dorsiflexion, R1 significant better: dynamic force (knee
females/62 ± 11 months na/21 right, 13 left sessions à 90 min/T1 before (2) ankle plantar flexion, (3) knee flexion, (4) knee extension – affected side), EMG activity (vastus
years treatment, T2 after treatment, extension, (5) hip abduction and (6) multi-segmental medialis – affected side)
T3 three weeks after task involving hip flexion/extension, knee extension/
treatment/7 flexion and ankle plantar flexion/dorsiflexion on
training machine (three sets of 10 repetitions,
resistance ‘maximal effort’, rest na)
  RT2: unilateral eccentric (1) ankle dorsiflexion, No significant differences: walking speed (fast,
(2) ankle plantar flexion, (3) knee flexion, (4) self-selected), dynamic force (knee extension
knee extension, (5) hip abduction and (6) multi- – affected side), EMG activity (rectus femoris –
segmental task involving hip flexion/extension, affected and non-affected side, semitendinosus
knee extension/flexion and ankle plantar flexion/ – affected and non-affected side, biceps
dorsiflexion on training machine (three sets of 10 femoris – affected and non-affected side, vastus
repetitions, resistance ‘maximal effort’, rest na) medialis – non-affected side)
Lee et al.29 26/16 males, 10 15 ± 9 months 15 ischemic, 11 Parallel groups (13+13)/30 RT1: unilateral concentric (1) leg press on training RT1 significantly better: plantar pressure
females/50 ± 8 haemorrhagic/lesion location sessions, duration na/T1 before machine (three sets of 8–10 repetitions, resistance distribution of the affected side (contact area
years na/13 right, 13 left treatment, T2 during treatment, 70% of 1RM, rest na); preceded by warm up (four – hindfoot; contact impulse – hindfoot)
T3 after treatment/6 repetitions with 25% of 1RM)
  RT2: unilateral concentric (1) knee extension on No significant differences: plantar pressure
training machine (three sets of 8–10 repetitions, distribution of the affected side (contact
resistance 70% of 1RM, rest na); preceded by warm area – forefoot, midfoot; peak contact force
up (four repetitions with 25% of 1RM) – forefoot, midfoot, hindfoot; contact impulse
– forefoot, midfoot)
Severinsen et al.45 30/22 males, 8 18 (8–38) months 30 ischemic, 0 haemorrhagic/ Parallel groups (14+16)/36 RT1: bilateral concentric (1) leg press, (2) hip RT1 significantly better: isometric force
females/67 (52–80) lesion location na/13 right, sessions à 60 min/T1 before extension, (3) hip flexion, (4) knee extension, (5) (knee extension – affected and non-affected
years 17 left treatment, T2 after treatment, ankle dorsal flexion and (6) ankle plantar flexion on side), Ten Meter Walk Test (one year after
T3 one year after treatment/8 training machines (three sets of eight repetitions, treatment)
resistance 80% of 1RM, rest na); preceded by warm
up (5 min ergometer)
  RT2: bilateral concentric (1) elbow flexion, (2) No significant differences: Six Minute Walk
elbow extension, (3) shoulder abduction and (4) Test, Ten Meter Walk Test (after treatment),
combined shoulder movement using a pulley (three graded test on ergometer (oxygen uptakepeak)
sets of 15 repetitions, resistance 60% of 1RM, rest
na); preceded by warm up (5 min ergometer)

(Continude)
17
 18

Table 3. (Continude)
Reference Subjects number/ Time since stroke Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
gender/age (years) location/affected hemisphere and duration/evaluations/PEDro
scale (score)

Son et al.50 28/15 males, 13 19 months 15 ischemic, 13 Parallel groups (14+14)/30 RT1: unilateral concentric (1) leg press on training No significant differences: Timed Up and Go,
females/57 years haemorrhagic/lesion location sessions à 30 min/T1 before machine (three sets of 8–10 repetitions, resistance Berg Balance Scale, postural balance (anterio-
na/15 right, 13 left treatment, T2 after treatment/7 70% of 1RM, rest na); preceded by warm up (four posterior, medio-lateral)
repetitions with 25% of 1RM)
  RT2: unilateral concentric (1) leg press on training  
machine (three sets of 8–10 repetitions, without
resistance, rest na); preceded by warm up (four
repetitions with 25% of 1RM)
Lee et al.28 22/14 males, 8 20 ± 8 months 13 ischemic, 7 haemorrhagic/ Parallel groups (11+11)/30 RT1: unilateral concentric (1) leg press on training RT1 significantly better: EMG activity
females/59 ± 7 lesion location na/11 right, sessions, duration na/T1 before machine (three sets of 8–10 repetitions, resistance (gastrocnemius, tibialis anterior), balance
years 11 left treatment, T2 after treatment/6 70% of 1RM, 3 min rest); preceded by warm up (anterio-posterior sway velocity with open
(four repetitions with 25% of 1RM) eyes and with closed eyes, medio-lateral sway
velocity with open eyes and with closed eyes
  RT2: unilateral concentric (1) knee extension on No significant differences: EMG activity (rectus
training machine (three sets of 8–10 repetitions, femoris, biceps femoris)
resistance 70% of 1RM, 3 min rest); preceded by
warm up (four repetitions with 25% of 1RM)
Engardt et al.49 20/15 males, 5 27 ± 11 months Etiology na/lesion location Parallel groups (10+10)/12 RT1: unilateral concentric (1) knee extension RT1 significantly better: walking speed (fast,
females/63 ± 7 na/12 right, 8 left sessions, duration na/T1 before on training machine (maximal 15 sets at angular self-selected), swing phase duration (self-
years treatment, T2 after treatment/5 velocities of 60, 120, 180, 120, 60, 120, 180°/sec or selected walking speed)
more, 10 repetitions, resistance na, 60 sec rest)
  RT2 significantly better: antagonist EMG
activity during concentric knee extension
(angular velocity of 120 and 180°/sec; affected
side), body weight distribution in rising
  RT2: unilateral eccentric (1) knee extension on No significant differences: eccentric and
training machine (maximal 15 sets at angular concentric force (knee extension-affected
velocities of 60, 120, 180, 120, 60, 120, 180°/sec or side), agonist EMG activity during concentric
more, 10 repetitions, resistance na, 60 sec rest) and eccentric knee extension (affected side),
antagonist EMG activity during eccentric
knee extension and during concentric knee
extension (angular velocity of 60°/sec; affected
side), body weight distribution in sitting down,
swing phase duration (fast walking speed)
Lee et al.41 48/28 males, 20 4.8 ± 4.5 years 33 ischemic, 6 haemorrhagic, Parallel groups RT active + CT active: bilateral concentric (1) RT active significantly better than RT
females/63 ± 9 6 other/lesion location na/27 (12+12+12+12)/30 sessions à leg press, (2) knee extension, (3) knee flexion and passive: Stair Climbing Power, graded test
years right, 21 left 60 min/T1 before treatment, T2 (4) ankle plantar flexion on training machines, and on ergometer (power outputpeak), maximal
after treatment/8 bilateral isotonic (5) abduction and (6) dorsiflexion concentric force of lower limbs – affected and
using free weights (two sets of eight repetitions, non-affected side, concentric endurance of
resistance 50% of 1RM/‘very hard’, rest na; 30 min), lower limbs – affected and non-affected side,
+ wheelchair ergometer training at 50%–70% of Ewart’s physical self-efficacy scales
oxygen uptakepeak (30 min)

(Continude)
Clinical Rehabilitation 00(0)
Table 3. (Continude)
Reference Subjects number/ Time since stroke Stroke etiology/lesion Study design/sessions number Intervention Results/used assessments
Veldema and Jansen

gender/age (years) location/affected hemisphere and duration/evaluations/PEDro

scale (score)

  RT active + CT passive: bilateral concentric (1) No significant differences between RT active

leg press, (2) knee extension, (3) knee flexion and and RT passive: Six Minute Walk Test, fast and
(4) ankle plantar flexion on training machines, and habitual gait velocity, graded test on ergometer
bilateral isotonic (5) abduction and (6) dorsiflexion (heart ratepeak, oxygen uptakepeak), treadmill
using free weights (two sets of eight repetitions, walking (physical cost index, oxygen cost),
resistance 50% of 1RM/‘very hard’, rest na; 30 min), Short Form (36) Health Survey
+ wheelchair ergometer training passive (30 min)
  RT passive + CT active: bilateral concentric (1)  
leg press, (2) knee extension, (3) knee flexion and
(4) ankle plantar flexion on training machines, and
bilateral isotonic (5) abduction and (6) dorsiflexion
using free weights (two sets of eight repetitions,
without resistance, rest na; 30 min), + wheelchair
ergometer training at 50%–70% of peak oxygen
uptake (30 min)
  RT passive + CT passive: bilateral concentric (1)  
leg press, (2) knee extension, (3) knee flexion and
(4) ankle plantar flexion on training machines, and
bilateral isotonic (5) abduction and (6) dorsiflexion
using free weights (two sets of eight repetitions,
without resistance, rest na; 30 min), + wheelchair
ergometer training passive (30 min)
Inaba et al.26 54/26 males, 28 na 33 ischemic, 13 parallel groups (28+26)/20–40 RT1: unilateral concentric (1) leg press on training RT significantly better: activities of daily
fames/57 years haemorrhagic, 8 na/lesion sessions à 15 min/T1 before machine (sets na, five repetitions at resistance 50% living at T2, maximal isometric force (knee
location na/27 right, 27 left treatment, T2 during treatment, of 10RM, 10 repetitions at resistance 10RM, rest na) extension) at T2
T3 after treatment/8
  RT2: bilateral concentric (1) knee flexion and No significant differences: activities of daily
extension, (2) hip abduction and adduction, (3) living at T3, maximal isometric force (knee
lower limb coordination exercises and (4) trunk extension) at T3
flexion, extension and rotation on training machine
(sets na, repetitions na, without resistance, rest na)

EMG: electromyography; na: not available, not applicable; min: minute; RT: resistance training; sec: second; T: test; 1RM: 1-repetition maximum.
19
20 Clinical Rehabilitation 00(0)

leg press shows to be more efficient than knee muscular force and the walking ability.7 Further
extension exercise. Lower body exercises lead to studies on large patient cohorts are needed for more
greater amelioration of parameters assessed than clarity on this important topic. The ability to walk
upper body exercises. High intensity training sup- independently is the most common rehabilitation
ports the recovery more effective than low inten- goal after stroke.51 Despite this, about 50% of stroke
sity training. Eccentric and concentric exercising survivors suffer from impaired walking ability six
are more efficient than isometric training. months after the cerebrovascular incident.52 Thus,
the development of innovative therapy strategies
for improving walking ability is one of the top
Discussion research priorities in stroke rehabilitation.2 Future
This systematic review and meta-analysis examine studies could evaluate resistance training coupled
the effects of resistance training on supporting the with non-invasive brain stimulation. Current data
recovery after a stroke. The data indicates that: (1) indicates that non-invasive brain stimulation may
repetitive application of resistance training may support the effectiveness of other therapies in gait
significantly support the recovery after stroke; (2) rehabilitation after stroke.53 Furthermore, more data
the effects of resistance training on muscular force is necessary about the neural mechanism of gait
and motor function of upper and lower limbs, qual- recovery in patients with stroke. Gait is a complex
ity of life, independence and reintegration, and sensorimotor function controlled by integrated cor-
other health relevant physiological indicators is tical, subcortical and spinal networks,54,55 and its
superior to others therapies and (3) different resist- neural background have not been sufficiently inves-
ance training protocols may lead to significantly tigated in this cohort up to now.55
different effects. Thus, resistance training can be a Muscular force and motor function: Our data
safe and effective tool for supporting the recovery shows that resistance training is highly efficient in
after a stroke. However, the available data is insuf- supporting the muscular force and motor function
ficient for evidence-based rehabilitation. Important in both the affected and the non-affected hemi-
aspects are high inconsistency of the effects body, superior to all other therapies. Furthermore,
detected, possibly caused by the large variability of the selection of resistance training protocol plays
interventions and populations, as well as inhomo- an important role. High intensity training, eccen-
geneity of parameters assessed. tric training and eccentric + concentric training
Gait: Only a part of our data demonstrates the show a better efficiency than low intensity train-
beneficial effects of resistance training in stroke ing, concentric training and isometric training.
gait rehabilitation. A direct comparison with other Current reviews confirm clinically relevant posi-
interventions shows that resistance training is supe- tive effects of strength and resistance training on
rior to stretching-oriented training, but inferior muscular force in persons after stroke,7,8 in
compared to passive mobilization and passive range accordance with our results. The potential of
of motion exercises. No unambiguous results were resistance training in improving the muscular
found for comparison with therapeutic gait and bal- function and muscular force in both the affected
ance training. Furthermore, the combination of and the non-affected limbs is highly relevant in
eccentric and concentric exercises is more efficient neurorehabilitation. Available data shows that
than isometric training and in addition concentric stroke patients suffer not only from reduced mus-
exercises are more efficient than eccentric training. cular force and muscle mass in the affected but
The currently opinions are divided when it comes to also in the non-affected hemi body.56 Insufficient
resistance exercises in gait rehabilitation.7,8 A recent muscular strength, as well as muscular imbal-
review detects positive effects for resistance train- ances within the affected leg are considered to be
ing only on muscular force, but not on activities the most important reasons for hemiparetic gait
such as walking or motor function of lower and abnormalities.57,58 Furthermore, paretic muscle
upper limbs.8 In contrast, a meta-analysis shows atrophy strongly correlates with reduced fitness
beneficial effects of strength training on both the levels.59 Resistance training has the potential to
Veldema and Jansen 21

support normal muscle functioning within the interventions.65,66 For example, a meta-analysis
affected leg and may counteract the stroke-related indicates that resistance training positively
decrease of physical fitness59 as well as the stroke- impacts health related quality of life in patients
related sarcopenia.56 Future studies should with chronic heart failure.64 An interventional
increasingly examine the relationship between study shows that resistance training significantly
resistance training-induced increase of muscular improved health related quality of life in elderly
force and progress in activities of daily life. people.65 However, the effects of team sports are
Mobility, balance and postural control: Only a comparable.65 Similarly, an interventional study
few studies tested the potential of resistance train- found no significant differences between resist-
ing on mobility, balance and postural control in ance training and balance training on independ-
patients after stroke. However, resistance exer- ence in Parkinson’s disease.66
cises may significantly reduce these disabilities. Spasticity and hypertonia: There are only lim-
Its effectiveness did not differ relevantly from ited data available about the influence of resistance
other therapies. The direct comparison of diverse exercises on spasticity and muscle tone disturbance
resistance training protocols demonstrates that leg after stroke. However, the available results indicate
press exercise may support the recovery of bal- that resistance training may reduce these disabili-
ance ability significantly better than knee exten- ties. Future studies should devote more attention to
sion exercise. In the future, comprehensive this relevant topic. About 25% of stroke victims
investigations should be carried out to examine developed an increase of muscle tone within two
the potential of resistance training in this area. weeks after stroke.67 These patients had signifi-
Balance and coordination deficits are common cantly higher incidences of pain and nursing home
post-stroke complications.60 They are associated placement as well as lower quality-of-life and daily
with impeded participation in activities of daily functioning.67,68 Spastic symptoms can induce
living, with difficult reintegration back into the pain, ankylosis, tendon retraction or muscle weak-
community,61 as well as with an increased fall ness in patients which may limit the potential suc-
risk.62 A current review demonstrates that regu- cess of rehabilitation.69
larly performing of resistance exercises is for Cardiorespiratory fitness: The evidence of
elderly people living in the community associated ergometer training on cardiorespiratory fitness in
with decreased fall risk.9 This supports the opin- stroke is insufficient until today. Available data
ion that resistance training may be useful for the shows that resistance training is less useful than
balance ability. On the other hand, a meta-analy- ergometer cycling, however, no study evaluated the
sis investigating the potential of diverse interven- effects of resistance training in comparison to no
tions on postural control in older adults found no intervention. The lack of data is striking, consider-
beneficial effects for resistance training.63 ing the fact that the cardiorespiratory fitness of
Quality of life, independence and reintegra- stroke persons is extremely reduced (by 30%–70%)
tion: Our data indicates that regular implementa- in comparison to healthy peers.11 This is coupled
tion of resistance exercises after stroke with enhanced energetic requirements during activ-
significantly increases health-related quality of ities of daily living, as consequence of motor defi-
life, independence in activities of daily living and cit.10 These limitations leads to an inability to carry
reintegration, superior to other therapies. out the activities of daily living in a large part of
Unfortunately, the data is too limited to make a patients.10,12 Reduced cardiorespiratory fitness also
definite statement about the effects of resistance limits the progress of rehabilitation processes and is
training in these fundamental areas. Future stud- associated with initial occurrence of stroke as well
ies should devote more attention to these relevant as with secondary stroke risks.12 Current data dem-
topics. Recent data demonstrates in several onstrates the beneficial effects of resistance training
cohorts that resistance training positively impacts on cardiorespiratory capacity in several cohorts
independence and quality of life.64,65 However, it such as critical illness polyneuropathy patients13 or
seems to be not more beneficial than other in elderly people.14
22 Clinical Rehabilitation 00(0)

Cognitive abilities and emotional state: There Strengths and limitations

exist little data about the impact of resistance train-
ing on cognition and emotional state in stroke There are obvious strengths and limitations of this
patients up to now. Most previous trials did not article. An important aspect is that this is the first
regard the potential of ‘physical exercises’ on ‘non- meta-analysis since 2006, which evaluates the
physical outcomes’. However, our data gives a hint benefits of resistance training in stroke patients,
that resistance training may impact these areas. and includes several papers published over the
Future studies should focus on this relevant topic. last three years. In addition, this is so far the first
Post-stroke cognitive impairment is a common con- meta-analysis comparing the effectiveness of dif-
sequence of stroke, leading to a reduced quality of ferent resistance training protocols in this cohort.
life.70 Post-stroke depression occurs in a significant Overall, the included trials show fair to high
number of patients and constitutes an important methodological quality. The most frequent meth-
complication of stroke, leading to a greater disability odological deficiencies are absence of investiga-
as well as increased mortality.71 Current reviews tors’ and subjects’ blinding, as well as intention to
indicate that physical activity and exercising may treat analysis. Because of limited data on this
support the cognitive recovery after a stroke15 and topic, no methodological quality depending selec-
reduce the occurrence of depression symptoms.16 tion of the manuscripts was performed for our
Other health-relevant physiological indicators: meta-analysis. This may hamper the interpreta-
The meta-analysis presented here shows that resist- tion of the data. Another weakness of our meta-
ance training is more efficient than other interven- analysis is the inconsistency of studies regarding
tions on improving health-relevant metabolic their methodological approach (longitudinal,
processes (e.g. glucose metabolism), haemody- crossover), interventions (different resistance
namic parameters (e.g. blood pressure, heart rate training types, different control interventions, dif-
variability, oxidative stress), body composition (e.g. ferent intervention duration) and outcomes (more
BMI, body fat) and muscles activity (EMG). than 100 outcomes were pooled in 11 areas).
Furthermore, significant protocol-relevant differ- These may be the reasons for the high inconsist-
ences were detected. Leg press, eccentric + concen- ency of effect sizes detected. Interpretation of the
tric training and concentric training are more results is limited by the small sample sizes of the
effective than knee extension, isometric exercises included studies with the highest being 77 partici-
and eccentric exercises in improving cardiac risk, pants, as well as by the fact that only a few trials
immunity-related factors and EMG activity. All examined long-term preservation of the effects.
these areas are relevant for neurorehabilitation. For
example, the stroke related sarcopenia56 impairs the Clinical messages
restoration of a normal gait pattern,57,58 the fitness
level59 and is associated with an increased fat mass.72 •• Resistance training after stroke
The so-called sarcopenic obesity is associated with improves recovery.
major health risks, functional limitations and meta- •• The effect is superior to other thera-
bolic dysregulation.72 Hyperglycaemia and dyslipi- pies on muscular force and motor
daemia are closely associated with stroke morbidity function, health related quality of
and mortality.73 Future studies also should also life, independence and several
investigate the influence of resistance training on health-relevant physiological
neurophysiological processes. A better understand- indicators.
ing of the neurophysiological backgrounds of •• Type of resistance training signifi-
diverse therapy strategies may contribute to optimi- cantly impacts its effects.
zation of neurorehabilitation processes. Animal •• The evidence is insufficient to
studies prove that moderate forced exercises effects change practice.
brain repair processes early after stroke.74
Veldema and Jansen 23

Declaration of conflicting interests a mixed-methods systematic review and meta-analysis.

Brain Behav 2018; 8(7): e01000.
The author(s) declared no potential conflicts of interest 11. Smith AC, Saunders DH and Mead G. Cardiorespiratory
with respect to the research, authorship and/or publica- fitness after stroke: a systematic review. Int J Stroke 2012;
tion of this article. 7(6): 499–510.
12. Ploughman M and Kelly LP. Four birds with one stone?
Reparative, neuroplastic, cardiorespiratory, and metabolic
Funding
benefits of aerobic exercise poststroke. Curr Opin Neurol
The author(s) received no financial support for the 2016; 29(6): 684–692.
research, authorship and/or publication of this article. 13. Veldema J, Bösl K, Kugler P, et al. Cycle ergometer train-
ing vs resistance training in ICU-acquired weakness. Acta
Neurol Scand 2019; 140(1): 62–71.
ORCID iD 14. de Resende-Neto AG, do Nascimento MA, De Sá CA,
Jitka Veldema https://orcid.org/0000-0002-9184 et al. Comparison between functional and traditional
-9776 resistance training on joint mobility, determinants of
walking and muscle strength in older women. J Sports
Med Phys Fitness 2019; 59(10). DOI: 10.23736/S0022-
Supplemental material 4707.19.09751-2.
Supplemental material for this article is available online. 15. Constans A, Pin-Barre C, Temprado JJ, et al. Influence
of aerobic training and combinations of interventions on
cognition and neuroplasticity after stroke. Front Aging
References Neurosci 2016; 8:164.
1. Mozaffarian D, Benjamin EJ, Go AS, et al. Heart dis- 16. Stanton R and Reaburn P. Exercise and the treatment of
ease and stroke statistics – 2015 update: a report from the depression: a review of the exercise program variables. J
American Heart Association. Circulation 2015; 131(4): Sci Med Sport 2014; 17(2): 177–182.
e29–322. 17. Saunders DH, Sanderson M, Hayes S, et al. Physical fit-
2. Pollock A, St George B, Fenton M, et al. Top ten research ness training for stroke patients. Cochrane Database Syst
priorities relating to life after stroke. Lancet Neurol 2012; Rev 2016; 3(3): CD003316.
11(3): 209. 18. Macedo LG, Elkins MR, Maher CG, et  al. There
3. Steele J, Fisher J, Giessing JG, et al. Clarity in reporting was evidence of convergent and construct validity of
terminology and definitions of set endpoints in resistance Physiotherapy Evidence Database quality scale for physi-
training. Muscle Nerve 2017; 56(3): 368–374. otherapy trials. J Clin Epidemiol 2010; 63(8): 920–925.
4. Prentice WE. Regaining muscular strength, endur- 19. Campbell MJ, Machin D and Walters SJ. Medical statis-
ance, and power. In: Prentice WE (ed.) Rehabilitation tics: a textbook for the health sciences. 4th ed. Chichester:
techniques for sports medicine and athletic training. John Wiley and Sons, 2007.
Thorofare, NJ: SLACK Incorporated, pp. 243–266. 20. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring
5. Grant S, Aitchison T, Henderson E, et al. A comparison of inconsistency in meta-analyses. BMJ 2003; 327(7414):
the reproducibility and the sensitivity to change of visual 557–560.
analogue scales, Borg scales, and Likert scales in normal 21. Aidar FJ, de Oliveira RJ, Silva AJ, et al. The influence
subjects during submaximal exercise. Chest 1999; 116(5): of resistance exercise training on the levels of anxiety in
1208–1217. ischemic stroke. Stroke Res Treat 2012; 2012: 298375.
6. Campos GE, Luecke TJ, Wendeln HK, et al. Muscular 22. Akbari A and Karimi H. The effect of strengthening exer-
adaptations in response to three different resistance-train- cises on exaggerated muscle tonicity in chronic hemipare-
ing regimens: specificity of repetition maximum training sis following stroke. J Med Sci 2006; 6(3): 382–388.
zones. Eur J Appl Physiol 2002; 88(1–2): 50–60. 23. Büyükvural Şen S, Özbudak Demir S, Ekiz T, et al.
7. Wist S, Clivaz J and Sattelmayer M. Muscle strengthening Effects of the bilateral isokinetic strengthening training
for hemiparesis after stroke: a meta-analysis. Ann Phys on functional parameters, gait, and the quality of life
Rehabil Med 2016; 59(2): 114–124. in patients with stroke. Int J Clin Exp Med 2015; 8(9):
8. Dorsch S, Ada L and Alloggia D. Progressive resistance 16871–16879.
training increases strength after stroke but this may not 24. Fernandez-Gonzalo R, Fernandez-Gonzalo S, Turon M,
carry over to activity: a systematic review. J Physiother et al. Muscle, functional and cognitive adaptations after
2018; 64(2): 84–90. flywheel resistance training in stroke patients: a pilot ran-
9. Sherrington C, Fairhall NJ, Wallbank GK, et al. Exercise domized controlled trial. J Neuroeng Rehabil 2016; 13
for preventing falls in older people living in the commu- (1): 37.
nity. Cochrane Database Syst Rev 2019; 1(1):CD012424. 25. Flansbjer UB, Lexell J and Brogårdh C. Long-term ben-
10. Lloyd M, Skelton DA, Mead GE, et al. Physical fit- efits of progressive resistance training in chronic stroke: a
ness interventions for nonambulatory stroke survivors: 4-year follow-up. J Rehabil Med 2012; 44(3): 218–221.
24 Clinical Rehabilitation 00(0)

26. Inaba M, Edberg E, Montgomery J, et al. Effectiveness health-care system of South Africa. Clin Rehabil 2018;
of functional training, active exercise, and resistive exer- 32(8): 1057–1068.
cise for patients with hemiplegia. Phys Ther 1973; 53(1): 41. Lee MJ, Kilbreath SL, Singh MF, et al. Comparison of
28–35. effect of aerobic cycle training and progressive resistance
27. Lee SB and Kang KY. The effects of isokinetic eccentric training on walking ability after stroke: a randomized
resistance exercise for the hip joint on functional gait of sham exercise-controlled study. J Am Geriatr Soc 2008;
stroke patients. J Phys Ther Sci 2013; 25(9): 1177–1179. 56(6): 976–985.
28. Lee NK, Kwon JW, Son SM, et  al. The effects of closed and 42. Marzolini S, Brooks D, Oh P, et al. Aerobic with resistance
open kinetic chain exercises on lower limb muscle activ- training or aerobic training alone poststroke: a secondary
ity and balance in stroke survivors. NeuroRehabilitation analysis from a randomized clinical trial. Neurorehabil
2013; 33(1): 177–183. Neural Repair 2018; 32(3): 209–222.
29. Lee NK, Kwon JW, Son SM, et al. Changes of plantar 43. Ouellette MM, LeBrasseur NK, Bean JF, et al. High-
pressure distributions following open and closed kinetic intensity resistance training improves muscle strength,
chain exercise in patients with stroke. NeuroRehabilitation self-reported function, and disability in long-term stroke
2013; 32(2): 385–390. survivors. Stroke 2004; 35(6): 1404–1409.
30. Lee NK, Son SM, Nam SH, et al. Effects of progressive 44. Patten C, Condliffe EG, Dairaghi CA, et al. Concurrent
resistance training integrated with foot and ankle com- neuromechanical and functional gains following upper-
pression on spatiotemporal gait parameters of individuals extremity power training post-stroke. J Neuroeng Rehabil
with stroke. J Phys Ther Sci 2013; 25(10): 1235–1237. 2013; 10(1): 1.
31. Sims J, Galea M, Taylor N, et al. Regenerate: assessing 45. Severinsen K, Jakobsen JK, Pedersen AR, et al. Effects
the feasibility of a strength-training program to enhance of resistance training and aerobic training on ambulation
the physical and mental health of chronic post stroke in chronic stroke. Am J Phys Med Rehabil 2014; 93(1):
patients with depression. Int J Geriatr Psychiatry 2009; 29–42.
24(1): 76–83. 46. Zou J, Wang Z, Qu Q, et al. Resistance training improves
32. Coroian F, Jourdan C, Bakhti K, et al. Upper limb isoki- hyperglycemia and dyslipidemia, highly prevalent among
netic strengthening versus passive mobilization in patients nonelderly, nondiabetic, chronically disabled stroke
with chronic stroke: a randomized controlled trial. Arch patients. Arch Phys Med Rehabil 2015; 96(7): 1291–1296.
Phys Med Rehabil 2018; 99(2): 321–328. 47. Chen CL, Chang KJ, Wu PY, et al. Comparison of the
33. Corti M, McGuirk TE, Wu SS, et al. Differential effects effects between isokinetic and isotonic strength training in
of power training versus functional task practice on com- subacute stroke patients. J Stroke Cerebrovasc Dis 2015;
pensation and restoration of arm function after stroke. 24(6): 1317–1323.
Neurorehabil Neural Repair 2012; 26(7): 842–854. 48. Clark DJ and Patten C. Eccentric versus concentric resist-
34. Folkerts MA, Hijmans JM, Elsinghorst AL, et  al. ance training to enhance neuromuscular activation and
Effectiveness and feasibility of eccentric and task- walking speed following stroke. Neurorehabil Neural
oriented strength training in individuals with stroke. Repair 2013; 27(4): 335–344.
NeuroRehabilitation 2017; 40(4): 459–471. 49. Engardt M, Knutsson E, Jonsson M, et al. Dynamic mus-
35. Gambassi BB, Coelho-Junior HJ, Dos Santos CP, et al. cle strength training in stroke patients: effects on knee
Dynamic resistance training improves cardiac autonomic extension torque, electromyographic activity, and motor
modulation and oxidative stress parameters in chronic function. Arch Phys Med Rehabil 1995; 76(5): 419–425.
stroke survivors: a randomized controlled trial. Oxid Med 50. Son SM, Park MK and Lee NK. Influence of resistance
Cell Longev 2019; 2019: 5382843. exercise training to strengthen muscles across multiple
36. Glasser L. Effects of isokinetic training on the rate of joints of the lower limbs on dynamic balance functions of
movement during ambulation in hemiparetic patients. stroke patients. J Phys Ther Sci 2014; 26(8): 1267–1269.
Phys Ther 1986; 66(5): 673–676. 51. Bohannon RW, Andrews AW and Smith MB.
37. Ivey FM, Prior SJ, Hafer-Macko CE, et al. Strength train- Rehabilitation goals of patients with hemiplegia. Int J
ing for skeletal muscle endurance after stroke. J Stroke Rehabil Res 1988; 11(2): 181–184.
Cerebrovasc Dis 2017; 26(4): 787–794. 52. Jørgensen HS, Nakayama H, Raaschou HO, et  al.
38. Kim CM, Eng JJ, MacIntyre DL, et al. Effects of isoki- Recovery of walking function in stroke patients: the
netic strength training on walking in persons with stroke: a Copenhagen Stroke Study. Arch Phys Med Rehabil 1995;
double-blind controlled pilot study. J Stroke Cerebrovasc 76(1): 27–32.
Dis 2001; 10(6): 265–273. 53. Vaz PG, Salazar APDS, Stein C, et al. Noninvasive brain
39. Kim SL and Lee BH. Effect of Mulligan's mobiliza- stimulation combined with other therapies improves gait
tion with movement technique on gait function in stroke speed after stroke: a systematic review and meta-analysis.
patients. J Phys Ther Sci 2016; 28(8): 2326–2329. Top Stroke Rehabil 2019; 26(3): 201–213.
40. Knox M, Stewart A and Richards CL. Six hours of 54. Fukuyama H, Ouchi Y, Matsuzaki S, et al. Brain func-
task-oriented training optimizes walking competency tional activity during gait in normal subjects: a SPECT
post stroke: a randomized controlled trial in the public study. Neurosci Lett 1997; 228(3): 183–186.
Veldema and Jansen 25

55. Swinnen SP. Intermanual coordination: from behav- 65. Pedersen MT, Vorup J, Nistrup A, et al. Effect of team
ioural principles to neural-network interactions. Nat Rev sports and resistance training on physical function, quality
Neurosci 2002; 3(5): 348–359. of life, and motivation in older adults. Scand J Med Sci
56. Hunnicutt JL and Gregory CM. Skeletal muscle changes Sports 2017; 27(8): 852–864.
following stroke: a systematic review and comparison 66. Santos SM, da Silva RA, Terra MB, et al. Balance ver-
to healthy individuals. Top Stroke Rehabil 2017; 24(6): sus resistance training on postural control in patients with
463–471. Parkinson's disease: a randomized controlled trial. Eur J
57. Reynard F, Dériaz O and Bergeau J. Foot varus in stroke Phys Rehabil Med 2017; 53(2): 173–183.
patients: muscular activity of extensor digitorum longus dur- 67. Wissel J, Schelosky LD, Scott J, et al. Early development
ing the swing phase of gait. Foot (Edinb) 2009; 19(2): 69–74. of spasticity following stroke: a prospective, observa-
58. Souissi H, Zory R, Boudarham J, et al. Muscle force strat- tional trial. J Neurol 2010; 257(7): 1067–1072.
egies for poststroke hemiparetic patients during gait. Top 68. Doan QV, Brashear A, Gillard PJ, et al. Relationship
Stroke Rehabil 2019; 26(1): 58–65. between disability and health-related quality of life and
59. Ryan AS, Dobrovolny CL, Smith GV, et al. Hemiparetic caregiver burden in patients with upper limb poststroke
muscle atrophy and increased intramuscular fat in stroke spasticity. PM R 2012; 4(1): 4–10.
patients. Arch Phys Med Rehabil 2002; 83(12): 1703–1707. 69. Thibaut A, Chatelle C, Ziegler E, et al. Spasticity after
60. Minet LR, Peterson E, von Koch L, et al. Occurrence and stroke: physiology, assessment and treatment. Brain Inj
predictors of falls in people with stroke. Stroke 2015; 2013; 27(10): 1093–1105.
46(9): 2688–2690. 70. Sexton E, McLoughlin A, Williams DJ, et al. Systematic
61. Kong KH and Lee J. Temporal recovery of activities of review and meta-analysis of the prevalence of cognitive
daily living in the first year after ischemic stroke: a pro- impairment no dementia in the first year post-stroke. Eur
spective study of patients admitted to a rehabilitation unit. Stroke J 2019; 4(2): 160–171.
NeuroRehabilitation 2014; 35(2): 221–226. 71. Robinson RG and Jorge RE. Post-stroke depression: a
62. Hamzat TK, Olaleye OA and Akinwumi OB. Functional review. Am J Psychiatry 2016; 173(3): 221–231.
ability, community reintegration and participation restric- 72. English C, Thoirs K, Coates A, et al. Changes in fat mass
tion among community-dwelling female stroke survivors in stroke survivors: a systematic review. Int J Stroke 2012;
in Ibadan. Ethiop J Health Sci 2014; 24(1): 43–48. 7(6): 491–498.
63. Low DC, Walsh GS and Arkesteijn M. Effectiveness of 73. Gast KB, Tjeerdema N, Stijnen T, et al. Insulin resist-
exercise interventions to improve postural control in older ance and risk of incident cardiovascular events in adults
adults: a systematic review and meta-analyses of centre of without diabetes: meta-analysis. PLoS ONE 2012; 7(12):
pressure measurements. Sports Med 2017; 47(1): 101–112. e52036.
64. Giuliano C, Karahalios A, Neil C, et al. The effects of 74. Austin MW, Ploughman M, Glynn L, et al. Aerobic exer-
resistance training on muscle strength, quality of life and cise effects on neuroprotection and brain repair following
aerobic capacity in patients with chronic heart failure – a stroke: a systematic review and perspective. Neurosci Res
meta-analysis. Int J Cardiol 2017; 227: 413–423. 2014; 87: 8–15.