Physiotherapy 113 (2021) 12–28

Systematic review

Aerobic, resistance and combined exercise training for patients

with amyotrophic lateral sclerosis: a systematic review and
meta-analysis
Masoud Rahmati ∗ , Fatemeh Malakoutinia
Department of Physical Education and Sport Sciences, Faculty of Literature and Human Sciences, Lorestan University, Khoramabad,
Iran

Abstract
Objective The aim of this systematic review and meta-analysis was to assess the effect of aerobic, resistance and combined exercise training
in patients with ALS.
Data source A comprehensive systematic search of CENTRAL, CINAHL, SPORTDiscuss, Embase, Scopus, ProQuest was performed
from inception to February 2021.
Eligibility criteria The systematic review included all studies that examined the effect of exercise training in ALS patients. Meta-analysis
was also carried out on randomized controlled trials (RCTs).
Data extraction and data synthesis Data related to primary outcomes (functional ability, respiratory function, fatigue, pain, quality
of life, upper-body strength, lower-body strength and Vo2 peak) and secondary outcomes (adverse events and feasibility of exercises) was
extracted from all studies and systematically reviewed.
Results 16 trials including 532 patients met the inclusion criteria; of these, eight studies were included in this meta-analysis. The meta-
analysis found a statistically significant difference in favor of exercise in functional ability (P = 0.001), overall quality of life (P = 0.03) and
Vo2 peak (P = 0.01). The included trials were generally of poor quality and had a risk of bias. However, the results of sensitivity analysis,
after omitting studies with high risk of bias, showed no statistically significant difference in functional ability (P = 0.05), overall quality of
life (P = 0.12) and Vo2peak (P = 0.13). Finally, no significant difference was found in respiratory function, fatigue, pain, and upper-body and
lower-body strength.
Conclusions The safety and effectiveness of exercise therapy in ALS patients remains unclear and further high quality RCTs with larger
sample size are needed.
Systematic Review Registration Number PROSPERO CRD42019140011
© 2021 Chartered Society of Physiotherapy. Published by Elsevier Ltd. All rights reserved.

Contribution of the Paper

• Exercise causes no serious adverse events in ALS patients and has positive effects on functional ability, quality of life and Vo2peak.
• Comparing the different modes of exercises, combined exercise training provides a significant difference in functional ability, upper-body
strength, lower-body strength and Vo2 peak.

Keywords: Exercise; Amyotrophic lateral sclerosis; Systematic review; Meta-analysis

∗ Correspondence: Department of Physical Education and Sport Sciences, Faculty of Humanities, Lorestan University, Khorramābād, Iran. Tel.: +98

6633223253; fax: +98 6633223253; mobile: +98 9124525538.

E-mail address: rahmati.mas@lu.ac.ir (M. Rahmati).

https://doi.org/10.1016/j.physio.2021.04.005
0031-9406/© 2021 Chartered Society of Physiotherapy. Published by Elsevier Ltd. All rights reserved.
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 13

Introduction Methods

Amyotrophic lateral sclerosis is a progressive neuro- The present systematic review and meta-analysis has
muscular disease described first in 1869 [1]. The primary been conducted according to the Preferred Reporting Items
hallmark of ALS is progressive degeneration of motor neu- for Systematic Reviews and Meta-Analyses (PRISMA)
rons in the brainstem, spinal cord, and cerebral cortex that guidelines [14] and registered with the international prospec-
leads to weakness, spasticity, fatigue, skeletal muscle atrophy, tive register of systematic reviews (PROSPERO; Reg. n:
paralysis and even death [2]. In this regard, functional ability CRD42019140011).
has been altered in people with ALS, due to muscle weakness
and spasticity, which are the primary symptoms reported in
patients with ALS [3]. In addition, muscle weakness leads to Search strategy
respiratory dysfunction, which is the most common cause of
death in ALS patients [4]. The literature was searched using the following electronic
Exercise training as a non-pharmacological tool may databases: CENTRAL, Embase, SPORTDiscuss, CINAHL,
generally improve muscle strength and cardiovascular func- Scopus, as well as ProQuest, from the start of each database
tion; it also has a positive effect on the quality of life [5]. until February 2021. The search strategy was performed using
However, there are conflicting reports about the effect of a combination of the following keywords
exercise training in ALS. The possible deleterious effects (e.g. “Amyotrophic lateral sclerosis,” or “Lou-Gehrig dis-
of exercise described in the literature may be related to ease”, or “ALS,” or “Motor neuron disease”), AND (e.g.
high-intensity training [6–8], even though this is not con- “physical activity,” or “exercise training,” or “physical train-
sensual and additional high quality studies are needed to ing,” or “exercise activity”) AND (e.g. “Random*,” or
advance our understanding of the effects of exercise inten- “clinical trial,” or “quasi* random*”). Forward (citation
sity in ALS. Mild-to-moderate intensity exercise appears the searching) and backward tracking (reference lists search-
most advantageous mode of exercise in ALS, improving body ing) were also conducted to identify relevant studies and
function and slowing disease progression [8]. In addition, an no restriction was imposed in terms of language or date
improved Amyotrophic Lateral Sclerosis Functional Rating of publication on the retrieved articles. Moreover, our sec-
Scale score (ALSFRS), muscle strength, and quality of life ondary search strategy included scanning bibliographies of
in ALS patients is reported following a period of exercise the retrieved articles and searching for unpublished studies
intervention [9–11]. It is certain that any form of exercise using key trial registries (clinicaltrials.gov, Current Con-
has an advantageous impact on the quality of life of ALS trolled Trials Registry, WHO hosted International Clinical
patients [8], but the best way to exercise for patients with Trials Registry Platform -ICTRP). The primary search strat-
ALS has not been determined and the effectiveness of exer- egy is outlined in the Supplementary Material S1.
cise training as a treatment for other symptoms of ALS in
human studies is not yet clear. Moreover, the effects of respi-
ratory training programs in ALS patients have been addressed
Selection criteria
in a previous meta-analysis [12]. Therefore, this systematic
review aims to analyze the effects of aerobic, resistance and
First, two review authors (MR, FM) independently read
combined training exercise modalities in patients with ALS.
the titles and abstracts generated by the electronic searches.
In 2013, a Cochrane review was conducted to investigate
Second, they studied the full texts of the remaining publi-
the effectiveness of exercise in patients with ALS on the
cations for further assessment and inclusion and exclusion
ALSFRS, fatigue muscle strength, and quality of life with-
criteria. Disagreement was resolved through reaching con-
out regard to the type of exercise intervention[13]. Results
sensus among the authors. The authors contacted the authors
from this systematic review and meta-analysis were incon-
of the selected articles when necessary to get more informa-
clusive due to the number of studies included being too small
tion about the methodology of trials and unreported results.
(n = 2) and lack of understanding about which types of exer-
The eligibility of the studies was selected according to the
cise training might be beneficial. Since then, many more
following criteria.
exercise trials in ALS have been conducted. Therefore, the
authors carried out an updated systematic review and meta-
analysis to clarify the effects of exercise interventions on
ALS patients in terms of functional score, strength, fatigue, Study design
quality of life as well as respiratory function. In addition, First, the systematic review included randomized con-
the authors aimed to investigate adverse events and the fea- trolled trials, as well as uncontrolled clinical trials, written
sibility of exercise as secondary outcomes. To do that, the in English. This systematic review excluded observational
authors focused on three frequently used exercise interven- studies, including prospective cohort, retrospective cohort,
tions, including aerobic, resistance and combined training cross-sectional, and case-control studies. Second, the authors
exercises. only included RCTs in our meta-analysis.
14 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

Population Statistical analysis

Studies conducted on participants with ALS were included
with no restriction placed on the age, gender, severity or Statistical analyses were performed with Review Manager
duration of disease, and regardless of pharmacotherapy. Software (RevMan 5.3, Cochrane Collaboration, Copen-
hagen, Denmark) and P < 0.05 was considered to represent a
statistically significant difference. For dichotomous data, the
Intervention type
data was summarized using relative risk with 95% CI and
Exercise intervention is defined as any purposive, repeti-
for pooled continuous data the authors expressed results as
tive and structured physical activity that has an objective of
mean difference (MD) or standardized MD (SMD) with cor-
health and/or improving or maintaining physical fitness [15].
responding 95% confidence intervals (CI). The pooled effect
The interventions were included that matched this definition.
was calculated using a Random-effect model if significant
heterogeneity was detected. Otherwise, a Fixed-effect model
Outcome was applied. Data was pooled if outcomes were reported by
The outcome measures were functional ability, respiratory at least two studies. When data were not presented in table
function, fatigue severity, pain, Vo2 peak and overall quality or text, data was extracted using Web Plot Digitizer from the
of life as primary outcomes, and adverse events and feasibility figures. The heterogeneity across trial results was tested with
of exercises as secondary outcomes. The studies that assessed Cochran’s Q test and I2 statistic, considering I2 values of less
at least one of the outcome measures were included. Articles than 25%, 25–75%, and more than 75% as small, moder-
were excluded if they were animal studies, reviews, case, ate, and high heterogeneity, respectively. Subgroup analysis
or case series studies. A detailed description of the exercise on exercise types was conducted in outcomes that had a
programs of selected studies is reported in Supplementary sufficient number of studies. Also, a sensitivity analysis of
Material S2. the primary outcomes excluding lower quality studies (score
≤4) [10,17–20] was performed to ensure the reliability and
Data collection process and data items robustness of results.

Two independent reviewers (MR and FM) extracted the

following data: first author names, publication year, study Results
design features, sample size, type and duration of exercise
intervention, the number and duration of training sessions Study identification
per week and primary and secondary outcomes.
The primary outcomes were standardized mean differ- 4175 records were identified from electronic databases
ence, or mean difference in measures of functional ability in accordance with the predesigned search strategy and
(measured using ALSFRS or ALSFRS-Revised or functional 687 additional research reports were identified by a man-
independent measure (FIM)), respiratory function (measured ual search. The two review authors (MA and FM) screened
by FVC), fatigue severity (assessed by a validated outcome all the records based on the abstract or the title, and 65 full
tool; e.g. fatigue severity scale), pain (measured using Visual text articles were retrieved for further investigation of cri-
analog scale pain), overall quality of life (SF-36 total score) teria. Sixteen studies fulfilling the eligibility criteria were
and upper and lower limb strength (MVIC). The secondary considered for this systematic review (Fig. 1).
outcomes were data about adverse events and the feasibility
of exercises for a descriptive synthesis. Study characteristics

Across the 16 studies, 532 patients with ALS were

Risk of bias in individual studies included in the systematic review. Of this, 271 and 261 were
allocated to exercise and control groups, respectively. The
The authors evaluated the quality of the selected studies mean age reported in studies varied from 45 to 70 years.
using Cochrane risk of bias tool according to the “Cochrane Studies were conducted in Portugal [17,21,22], US [9,23–25],
Handbook for Systematic Reviews of Interventions” [16]. The Japan [26], Italy [10,19,20,27], Israel [18,28], Denmark [11]
tool comprises of the following domains: random sequence and in Netherlands [29]. All studies reported that their experi-
generation, allocation concealment, blinding of participants mental procedures and protocols were approved by their local
and personnel, blinding of outcome assessment, incomplete committee and all participants provided written informed
outcome data, selective outcome reporting, and other bias. consent. The characteristic of the included studies in the sys-
Sensitivity analysis was performed by repeating the analysis tematic review is described in detail in Table 1. Of 16 studies,
excluding trails judged to be at a high risk of bias and studies the authors included all RCTs in the meta-analysis. However,
with score 4–7 were classified as higher quality. The details two RCTs did not provide the appropriate data [24,29], and
of the criteria used to assess each domain are summarized in the authors tried to contact the study authors by email to
Fig. 3. obtain further information, but the authors received only one
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 15

Fig. 1. PRISMA flow diagram.

response [29]. Therefore, the authors had to exclude one RCT Effects of interventions on primary outcome
from the meta-analysis [24].
Functional ability
16 studies reported the effects of exercise training on
functional ability in participants with ALS measured by ALS-
Risk of bias of included studies FRS, ALSFRS-R, and FIM [9–11,17,18,20–27,29]. Due to
incomplete data, one RCT study could not be entered in the
The randomization process was described and appropriate meta-analysis [24] and then eight RCTs [9,10,17–20,28,29]
for four studies and in no study was it stated that allocations involving 233 participants (127 of whom performed exercise
were concealed. The risk for performance bias was high in all training, while the 106 remaining served as control) were
included studies because it was not feasible to blind partic- included. The exercise training period ranged from 5 [20] to
ipants to the exercise intervention. All the included studies, 24 [9,10,17,18] weeks with an average of 3 sessions per week.
except one [18], reported blinding of the outcome assessors; The results showed that participants in the exercise group had
so, their risk of bias was judged as low. In addition, the rate a statistically significant difference in terms of SMD scores
of drop out in six studies was reported to be high, hence of the functional ability compared with the control group in
their attrition bias was considered high [9,10,17,18,24,29]. favor of the intervention group (SMD = 0.84, 95% CI 0.33 to
By checking the available protocol of included studies, the 1.36, P = 0.001, I2 = 67%, the random-effect model; Fig. 2A).
selective reporting bias in the majority of studies was judged Subgroup analyses of exercise types were pooled into aerobic
as low risk. No risk of other biases was identified in all exercise [10,17,29], resistance exercise [9,18], and com-
included studies. To prevent the risk of collider-stratification bined exercise (aerobic–resistance) [19,20,28] groups. The
bias [30], all RCTs were included in the meta-analysis, irre- results demonstrated that aerobic exercise had a positive
spective of their scores in the Cochrane risk of bias tool. effect, but it did not reach a statistically significant difference
However, sensitivity analysis based on high-quality studies (SMD = 0.32 (−0.04, 0.68), P = 0.08, I2 = 0%, Fig. 2A). In
was conducted to test the stability of our results. addition, resistance exercise did not have a significant effect
 16
Table 1
Characteristics of the selected studies.
Study Population Duration Study design Intervention Outcome Main finding Compliance with
measure training
Karlon et al., EX: N = 14 3 month Randomized EX: Aerobic (cycling, ALSFRSR Significant difference NR
2021 [28] Age: 58.5 (±13.2) controlled trial 20–30 minutes) + range of motion and FSS between groups in
Drop-out: 2 (13%) stretching exercises FVC ALSFRS
CON: N = 14 (10 minutes) + strength program (by Pain
Age: 60.4 (±14.7) individual’s body weight, 20 minutes), 2
Drop-out: 2 (13%) session/week
CON: Stretching exercise of the upper
and lower limb (20 minutes), 5

M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

session/week
Van EX: N = 27 Age: 69 (±10) 4 month Randomized EX: Aerobic cycling exercise program ALSFRS- Significant difference In all, 41% of the
Groenestijn Drop-out: 17 (63%) controlled trial (on a cycle ergometer and a step R between groups in patients were able to
et al., 2019 CON: N = 30 board) + Usual care, 3session/week FVC FVC attend 75% or more
[29] Age: 59.9 (±10.7) CON: Usual care (consist of a CIS- treatment sessions of
Drop-out: 8 (27%) rehabilitation medicine consultant, an fatigue AET program
occupational therapist, physical therapist, VAS pain
speech therapist, dietician, social worker, SF-36
psychologist, and consultant physicians) Upper and
Lower
body
strength
Sivaramakrishnan EX: N = 9 1 month Single group EX: 65–70 steps per minute for ALSFRS- No significant 8 participants
et al., 2019 Age: 59.2 (±12.3) repeated- 40 minutes (5 minutes warm up, R differences were completed all session
[23] Drop-out: 1 (11%). measures 30 minutes of exercise) with 50–70% FSS found (total 96 session),
CON: No control design maximal heart rate (moderate intensity), 100% compliance
3 session/week
Pegoraro EX: N = 18 Age: 61.1 6 week Single group EX: Depending on the level of ability of ALSFRS- Significant difference NR
et al., 2019 (±12.8) repeated- ALS patients, they received an R between groups in
[27] Drop-out: Not any CON: No measures individualized daily progressive training FSS FIM ALSFRS-R, FSS,
control design of muscular strengthening and aerobic FIM
endurance exercises (cycle ergometer or
arm-leg ergometer or treadmill, or
standard rehabilitation)
Ferri et al., EX: N = 7 3 month Randomized EX: 15 minutes of cycling + 25 minutes ALSFRS Significant difference NR
2019 [19] Age: 50.7 (±3.3) controlled trial of strength exercises with 60% Vo2-peak between groups in
Drop-out: 1 (13%) RM + 10 minutes of proprioceptive ALSFRS
CON: N = 4 exercises + 10 minutes of upper and
Age: 55.5 (±5.9) lower body strength exercises, 3
Drop-out: 3 (38%) session/week
CON: Usual care (continued their usual
standard of care), 3 session/week
Table 1 (Continued)
Study Population Duration Study design Intervention Outcome Main finding Compliance with
measure training
Braga et al., EX: N = 24 6 month Randomized EX: Aerobic exercise ALSFRS- Significant difference NR
2018 [17] Age: 63.2 (±13.0) Drop-out: controlled trial (treadmill) + standard care, R between groups in
5 (21%) 2session/week ALSFRS-R
CON: N = 24 CON: Daily Standard care (range of
Age: 62 (±12.1) motion exercise, limbs relaxation, trunk
Drop-out: 19 (79%) balance, gait training)
Merico et al., EX: N = 23 5 week Randomized EX: Individualized strengthening (using FIM Significant difference NR
2018 [20] Age: 61.6 (±10.6) controlled trial rubber bands, 3 reps for bilateral muscle, FSS between groups in
Drop-out: 8 (26%) and 80% max contraction) and aerobic FIM
CON: N = 15 endurance (using cycle ergometer or No significant
Age: 59.8 (±14.7) ergometry arm-leg or treadmill, 65% differences in FSS

M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

Drop-out: 8 (35%) MHR, 15–20 minutes) 7 session/week,
60 minutes
CON: Standard care (stretching
exercise + active mobilization + general
muscle reinforcement), 7 session/week,
total duration 60 minutes
Kitano et al., EX: N = 21 6 month Non- EX: Strength training, stretching ALSFRS- Significant difference NR
2018 [26] Age: 62.8 (±10.2) Drop-out: randomized exercise, functional training + Usual care R between groups in
6 (23%) controlled trial CON: Usual care ALSFRS-R
CON: N = 84
Age: 62.7 (±12.1)
Drop-out: Not any
Braga et al., EX: N = 10 6 month Single group EX: Aerobic training (symptom-limited FVC% Significant difference Average sessions
2018 [35] Age: 57 (±9.1) repeated- treadmill-ramp protocol) ALSFRS- within baseline and 6 number was 29
Drop-out: Not any measures R month intervention in
CON: No control design ALSFRS-R
Clawson et al., EX: Endurance: 6 month Randomized EX: 2 groups: ALSFRS There were no After 6 month,
2018 [24] N = 20 controlled trial Resistance exercise: 3session/week, FSS significant different compliance was 73%,
Age: 57.82 (±11.8) 40–70% 1 RM FVC between estimated 60% and 46% in
Drop-out: 9 (31%) Endurance exercise (upper and lower variables control, resistance and
EX: Resistance: limb exercise), 3session/week, 40–70% endurance groups;
N = 18 target HR respectively
Age: 63.65 (±10.5) CON: Stretching/range of motion
Drop-out: 4 (19%)
CON: N = 21
Age: 57.68 (±9.72)
Drop-out: 2 (9%)
Jensen et al., EX: N = 6 3 month Open-label EX: Resistance exercise (leg press, knee ALSFRS- No significant NR
2017 [11] Age: 62.5 (±8.8) trial, repeated extension, leg curl, calf raises, lateral arm R difference between
Drop-out: 1 (17%) measures pull, seated rows, chest press, shoulder groups in ALSFRS
CON: No control design press, abdominal crunches, and back
extension), familiarization period consist
of 3 sets of 12 repetitions at 15-repetition
maximum Load, which followed 2 sets of
5 repetitions at 6 RM, 2–3 session/week

17
 18
Table 1 (Continued)
Study Population Duration Study design Intervention Outcome Main finding Compliance with
measure training
Lunetta et al., EX: N = 30 6 month Randomized EX: Three strictly monitored exercise ALSFRS- Significant difference NR
2016 [10] Age: 61.1 (±10.1) controlled trial programs (SMEP) R between groups in
Drop-out: 8 (27%) SMEP1: Active exercise in 6 muscle FVC ALSFRS-R
CON: N = 30 group against gravity combined with No significant
Age: 60.3 (±9.9) cycloergometer activity difference in FVC
Drop-out: 5 (17%) SMEP2: Active exercise in 6 muscle
group against gravity
SMEP3: Passive exercise
(flexion–extension movements in upper
and lower muscle groups)

M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

Daily exercise for 2week each month
CON: Passive exercise
(flexion–extension movements in upper
and lower muscle group), 20 minutes, 2
session/week.
Sanjak et al., EX: N = 9 2 month Non-random, EX: Repetitive rhythmic exercise-STAT ALSFRS- Significant difference NR
2010 [25] Age: 62 (±14.06) Interventional (30 minutes total; 6 × 5 minutes interval R between groups in
Drop-out: 3 (33%) with repeated- of exercise intercalated with 5 minutes of FSS ALSFRS-R
CON: No control measures rest), mild to moderate intensity but not MMT No significant
design to exceed 12–13/20 Borg scale, 3 MVIC difference in MVIC
session/week. (L/E and score
U/E)
Dal -Haas EX: N = 13 6 month Randomized EX: Resistance exercise + standard care, ALSFRS Significant difference High and moderate
et al., 2007 Age: 56.0 (±7.3) controlled trial 3session/week FSS between groups in
[9] Drop-out: 5 (38.4) CON: Standard care (stretching FVC ALSFRS and L/E
CON: N = 14 exercise), daily MVIC MVIC score
Age: 51.8 (±12.6) Drop-out: (L/E and No significant in U/E
4 (28.5) U/E) MVIC score
Drory et al., EX: N = 14 6 month Randomized EX: Muscle endurance exercise (four ALSFRS At 3 month, NR
2001 [18] Age: 58.0 (±13.2) Drop-out: controlled trial limbs and trunk), 15 minutes, 2 FSS Significant between
6 (42.8) session/week at home VAS pain groups in ALSFRS
CON: N = 11 CON: No exercise (daily routine Quality of At 6 month, no
Age: 60.7 (±16.4) Drop-out: activity) life significant difference
5 (54.5)
Pinto at al., EX: N = 8 12 month NR EX: Endurance exercise was applied FVC Significant difference NR
1999 [21] Age: 62 (±14) with the goal of attaining anaerobic FIM between groups in
Drop-out: Not any threshold in 5–15 minutes in Bruce or FIM
CON: N = 12 Naughton protocol
Age: 64 (±16) CON: No exercise (daily routine
Drop-out: Not any activity)
NR = not reported; ALSFRS-R = ALS Functional Rating Scale Revised; FVC = forced vital capacity; VAS = Visual Analog Scale; SF-36 = 36-Item Short Form Survey; CIF fatigue = the Checklist Individual
Strength subscale fatigue.
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 19

in favor of the intervention group (SMD = 0.61 (−0.11, 1.34), significant level and decreased heterogeneity from 86% to
P = 0.10, I2 = 0%, Fig. 2A). The effect size for combined pro- 62% (n = 103, SMD = 0.71, 95% CI 0.01 to 1.42), P = 0.05,
grams had a statistically significant effect in favor of the I2 = 62%, the random-effect model, Supplementary Material
intervention group (SMD = 1.70 (1.09, 2.32), P = 0.00001, S3A).
I2 = 18%, Fig. 2A). Also, subgroup analysis of exercise types
showed a statistically significant difference among the three Respiratory function
subgroups which indicated that a combined exercise pro- Respiratory function was examined in 7 studies (8 tri-
gram is the most appropriate form of exercise for improving als) [9,10,21,22,24,28,29], 4 of which were RCTs. Of these
functional ability in ALS patients (P = 0.0007; I2 = 86%, studies, four of the trials used aerobic exercise and in only
Fig. 2A). The sensitivity analysis results of higher quality one of them was observed statistically significant improve-
studies [9,28,29] for functional ability changed to a non- ment from exercise in FVC [29]. Three further trials using

Fig. 2. Meta-analysis of pre-post change means for (A) Functional ability, (B) Respiratory function, (C) Fatigue severity, (D) Pain, (E) SF-36 total score, (F)
Upper body strength, (G) Lower body strength, (H) Vo2 peak and (I) 6-month follow-up functional ability after the end of intervention. CI = confidence interval;
IV = inverse-variance method; SD = standard deviation; SMD = standardized mean difference; MD = mean difference.
20 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

Fig. 2. (Continued)
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 21

Fig. 2. (Continued)
22 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

Fig. 2. (Continued)

Fig. 3. (A) Risk of bias summary and (B) Risk of bias graph: review authors’ judgements about each risk of bias items for each included study.
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 23

resistance and anaerobic exercise observed no statistically improvement from exercise in fatigue [27]. When lower qual-
significant improvement in FVC [9,21,24]. However, data ity studies [18,20] were removed from sensitivity analysis for
for one study was not available [24]; therefore, 4 RCTs fatigue severity assessment, the pooled estimate was changed
[9,10,28,29] involving 137 participants (62 of whom per- and heterogeneity decreased (n = 103, SMD = −0.24, 95%
formed exercise training, while the remaining 75 served as CI −0.63 to 0.15, P = 0.23, I2 = 27%, the fixed-effect model,
a control) were included for respiratory function in meta- Supplementary Material S3C).
analysis. The exercise training period ranged from 12 [28]
to 24 [9,10] weeks with an average of 2 sessions per Pain
week. The pooled SMD was calculated and did not show Three RCTs [18,28,29] with 91 participants (44 of whom
positive effect size (SMD = 0.22, 95% CI −0.10 to 0.55, performed exercise training, while the 47 remaining served
P = 0.18, I2 = 37%, the fixed-effect model; Fig. 2B). Sub- as a control) assessed pain using the visual analog. The
group analyses of exercise types were pooled into aerobic exercise training period for these RCTs ranged from 12
exercise [10,29], resistance exercise [9], and combined exer- [28] to 24 [18] weeks with an average of 3 sessions per
cise (aerobic–resistance) [28] groups. Results demonstrated week. Meta-analysis showed no statistically significant dif-
that aerobic exercise did not have a positive effect size ferences between the exercise training and control groups
(SMD = 0.11 (−0.28, 0.50), P = 0.58, I2 = 58%, Fig. 2B). (SMD = 0.15, 95% CI −0.25 to 0.55, P = 0.46, I2 = 0%, the
Also, resistance exercise did not have a significant effect fixed-effect model; Fig. 2D). Subgroup analyses of exercise
(SMD = 0.96 (−0.04, 1.95), P = 0.06, Fig. 2B). Finally, the types were pooled into aerobic exercise [29], resistance exer-
effect size for combined programs did not show statistical cise [18], and combined exercise (aerobic–resistance) [28]
significance (SMD = 0.23 (−0.51, 0.97), P = 0.54, Fig. 2B). groups. The results demonstrated that aerobic exercise had
Also, subgroup analysis of exercise types showed no sta- no positive effect size (SMD = 0.02 (−0.50, 0.54), P = 0.95,
tistically significant differences among the three subgroups Fig. 2D). In addition, resistance exercise had no significant
(P = 0.30; I2 = 18%, Fig. 2B). Sensitivity analysis of higher effect (SMD = 0.09 (−0.97, 1.15), P = 0.87, Fig. 2D). More-
quality studies for respiratory function [9,28,29] changed the over, the effect size for combined programs did not show
pooled estimate results (n = 103, SMD = 0.12, 95% CI −0.27 statistical significance (SMD = 0.46 (−0.30, 1.21), P = 0.23,
to 0.51, P = 0.54, I2 = 49%, the fixed-effect model, Supple- Fig. 2D). Also, subgroup analysis of exercise types showed no
mentary Material S3B). statistical significant differences among the three subgroups
(P = 0.64; I2 = 0%, Fig. 2A). When a lower quality study [18]
Fatigue severity was removed from the sensitivity analysis for pain assess-
The effects of exercise training on fatigue severity were ment, the pooled estimate was changed (n = 85, SMD = 0.16,
evaluated in eight studies (9 trials) using the fatigue severity 95% CI −0.27 to 0.59, P = 0.47, I2 = 0%, the fixed-effect
scale (FSS) [9,18,20,23–25,27,28] and the Checklist Individ- model, Supplementary Material S3D).
ual Strength subscale fatigue (CIF-F) [29]; although, data for
one study was not available [24]. 5 RCTs [9,18,20,28,29] Overall quality of life
involving 147 participants (76 of whom performed exer- Two RCTs [18,29] involving 71 participants (35 of whom
cise training, while the 71 remaining served as control) performed exercise training, while the 36 remaining served
were included in the meta-analysis. The exercise training as control) examined the effects of exercise interventions on
period for these RCTs ranged from 5 [20] to 24 [9,18] the overall quality of life using a SF-36 total score. The exer-
weeks with an average of 2 sessions per week. The results cise training period for these RCTs ranged from 16 [29] to
showed no significant difference between the exercise train- 24 [18] weeks with an average of 3 sessions per week. The
ing and control groups (SMD = 0.94, 95% CI −0.75 to results showed that SMD was statistically significant in terms
2.63, P = 0.27, I2 = 94%, the random-effect model; Fig. 2C). of overall quality of life in patients with ALS in favor of
Subgroup analyses of exercise types were pooled into aero- the intervention group (SMD = 0.52, 95% CI 0.04 to 1.00,
bic exercise [29], resistance exercise [9,18], and combined P = 0.03, I2 = 0%, the fixed-effect model; Fig. 2E). The num-
exercise (aerobic–resistance) [20,28] groups. The results ber of studies was too small to permit subgroup analyses of
demonstrated that aerobic exercise did not have a positive exercise types in the quality of life outcome. Sensitivity anal-
effect size (SMD = −0.29 (−0.82, 0.23), P = 0.27, Fig. 2C). ysis results of a higher quality study [29] for overall quality of
In addition, resistance exercise had no significant effect life, changed to a non-significant level (n = 57, SMD = 0.42,
(SMD = −0.42 (−2.19, 1.35), P = 0.64, I2 = 81%, Fig. 2C). 95% CI −0.11 to 0.95), P = 0.12, Supplementary Material
Also, the effect size for combined programs did not show S3E).
statistical significance (SMD = 3.49 (−4.55, 11.53), P = 0.4,
I2 = 98%, Fig. 2C). Also, subgroup analysis of exercise types Upper body strength
showed no statistical significant differences among the three Of the four studies [9,20,25,29], three RCTs [9,20,29]
subgroups (P = 0.65; I2 = 0%, Fig. 2A). It should be noted with 113 participants (58 of whom performed exercise train-
that only one single group with a repeated-measures study ing, while the 55 remaining served as a control) assessed
design using aerobic exercise showed statistical significant the effects of exercise interventions on muscle strength using
24 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

upper body maximal voluntary isometric contraction. The SMD = −0.04, 95% CI −0.43 to 0.34), P = 0.82, I2 = 0%, the
exercise training period in these RCTs ranged from 5 [20] fixed-effect model, Supplementary Material S3G).
to 24 [9] weeks with an average of 3 sessions per week.
Meta-analysis showed no significant differences between the
Vo2 peak
exercise groups rather than control groups (SMD = 0.26, 95%
Four RCTs [17,19,20,28] involving 107 participants (68
CI to −0.49 to 1.00, P = 0.50, I2 = 71%, the random-effect
of whom performed exercise training, while the 39 remaining
model; Fig. 2F). Subgroup analyses of exercise types were
served as control) were included for Vo2 peak in meta-
pooled into aerobic exercise [29], resistance exercise [9],
analysis. The exercise training period ranged from 5 [20]
and combined exercise (aerobic–resistance) [20] groups. The
to 24 [17] weeks with an average of 3 sessions per week.
results demonstrated that aerobic exercise had no positive
The pooled SMD was calculated and showed a positive
effect size (SMD = 0.06 (−0.46, 0.58), P = 0.83, Fig. 2F).
effect size in favor of the intervention group (SMD = 1.26,
In addition, resistance exercise had no significant effect
95% CI 0.27 to 2.25, P = 0.01, I2 = 77%, the random-effect
(SMD = −0.38 (−1.32, 0.56), P = 0.43, Fig. 2F). Moreover,
model; Fig. 2H). Subgroup analyses of exercise types were
the effect size for combined programs showed statistical
pooled into aerobic exercise [17] and combined exercise
significance (SMD = 1.00 (0.31, 1.70), P = 0.005, Fig. 2F).
(aerobic–resistance) [19,20,28] groups. The results demon-
Also, subgroup analysis of exercise types showed statis-
strated that aerobic exercise did not have a positive effect size
tically significant differences among the three subgroups
(SMD = 0.61 (−0.30, 1.52), P = 0.19, Fig. 2H). The effect
which indicates that a combined exercise program is the
size for combined programs showed a statistically signifi-
most appropriate form of exercise for improving upper limb
cant positive effect size in favor of the intervention group
strength in ALS patients (P = 0.03; I2 = 71%, Fig. 2F). Sen-
(SMD = 1.49 (0.17, 2.81), P = 0.03, I2 = 81%, Fig. 2H). Also,
sitivity analysis results of higher quality studies [9,29] for
subgroup analysis of exercise types showed no statistical
upper body strength changed and heterogeneity decreased
significant differences among the two subgroups (P = 0.28;
from 71% to 0% (n = 75, SMD = −0.05, 95% CI −0.50 to
I2 = 14%, Fig. 2H). Sensitivity analysis results of a higher
0.41), P = 0.85, I2 = 0%, the fixed-effect model, Supplemen-
quality study [28] for Vo2 peak, changed to a non-significant
tary Material S3F).
level (n = 28, SMD = 0.59, 95% CI −0.17 to 1.35), P = 0.13,
Supplementary Material S3H).

Lower body strength Six-month follow-up functional ability scale after the end
Of the six studies [9,19,20,25,28,29], five RCTs of trials phase
[9,19,20,28,29] with 152 participants (79 of whom per- Two RCTs [10,29] involving 104 participants (49 of whom
formed exercise training, while the 73 remaining served as performed exercise training, while the 55 remaining served
a control) explored the effects of exercise interventions on as a control) reported 6-month follow-up functional ability
muscle strength using lower body maximal voluntary iso- after the end of trial phase measures by ALSFRS-r. A sig-
metric contraction. The exercise training period in these nificant difference was observed for this outcome in favor
RCTs ranged from 5[20] to 24 [9] weeks with an average of the intervention group (MD = 3.40, 95% CI 0.81, 5.99,
of 3 sessions per week. Meta-analysis showed no signif- P = 0.01, I2 = 0%, the fixed-effect model; Fig. 2I). The num-
icant differences between the exercise groups rather than ber of studies was too small to permit subgroup analyses of
control groups (SMD = 0.19, 95% CI to −0.23 to 0.61, exercise types in the upper limb strength outcome. Sensitiv-
P = 0.37, I2 = 34%, the fixed-effect model; Fig. 2G). Sub- ity analysis results of a higher quality study [29] for 6-month
group analyses of exercise types were pooled into aerobic follow-up functional ability after the end of the trial phase,
exercise [29], resistance exercise [9], and combined exercise changed to a non-significant level (n = 57, SMD = 3.40, 95%
(aerobic–resistance) [19,20,28] groups. The results demon- CI −0.13 to 6.93), P = 0.06, Supplementary Material S3I).
strated that aerobic exercise didn’t have a positive effect
size (SMD = −0.05 (−0.57, 0.47), P = 0.85, Fig. 2G). In Effects of interventions on secondary outcome
addition, resistance exercise didn’t have a positive effect
size (SMD = −0.05 (−1.45, 0.45), P = 0.30, Fig. 2G). The Adverse effects
effect size for combined programs showed statistical signifi- Seven studies reported information on adverse events
cance in favor of the intervention group (SMD = 0.54 (0.07, involving 276 patients. Five studies reported no adverse
1.00), P = 0.02, I2 = 0%, Fig. 2G). Also, subgroup analysis events [9,23,25,26]. However, there were adverse events in
of exercise types showed that compared with other types of another two studies. In one study, 1 case of myalgia, 2 cases
exercises, combined exercise had a positive effect size but it of fasciculation as well as 2 cases of cramps were observed
did not reach a statistically significant difference (P = 0.08; in the aerobic exercise intervention [22,29]. In another study
I2 = 60%, Fig. 2G). Sensitivity analysis results of higher (three-arm trial), the number of common adverse events in
quality studies for lower body strength [9,28,29] changed the usual care group (stretching), resistance training and
and decreased heterogeneity from 34% to 0% (n = 103, endurance training were 27, 17, and 16, respectively. In
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 25

addition, this study reported four serious adverse events in should be interpreted with caution, as it has been suggested
exercise groups; in the resistance group, one incidence of that estimating the effect size in fewer than five studies may
pulmonary embolism was reported related to the consump- be imprecise [31]. There is, however, some exciting RCT data
tion of oral contraceptive medicine a few days before the that, in combination with future research, has the potential to
event; another event was the discovery of a long neoplasm. allow for more definitive conclusions in this field of study.
In addition, pneumonia and respiratory distress occurred in The results of secondary outcomes of this systematic
the endurance group [24]. review indicated that different included RCTs had a high
dropout in the intervention group, causing an attrition bias.
Feasibility of exercise Patients remaining in the intervention group were possibly
Of the RCTs, 8 studies reported drop-out rates, separately, patients with a better prognosis than patients who dropped
in control and exercise groups [9,10,18–20,22,24,28,29]. out. Therefore, although the most frequent or serious adverse
Total dropout was 34% (103/300 participants). In addi- events have mostly been reported as being equivalent in the
tion, dropout from exercise group was 34% (57/170), while intervention group compared to the control group, due to
dropout from the control group was almost the same, at the high rate of drop out in the intervention group, it is not
32% (49/152). Across eight RCTs, the compliance percent possible to determine exactly whether exercise training is
was reported in three studies [9,24,29]. In one RCT, after 6 safe. Thus, further research is necessary to clearly warrant
months, compliance was reported at 73.22, 60.22 and 46.19 the safety of exercise training for ALS patients.
percent in a standard care group (stretch and range of motion This systematic review and meta-analyses is the first to
exercise), resistance and aerobic exercise, respectively [24]. focus on respiratory function, ALSFRS score after six-month
In another study, 41% of the patients were able to attend 75% follow-up, Vo2 peak and feasibility of exercise training in
or more treatment sessions of an aerobic exercise therapy ALS patients. The benefits of exercise on functional ability
(AET) program [29]. at the end of interventions is in line with a previous review
[13] that there was a significant difference in functional abil-
ity. The present review is inconsistent with a previous review,
Discussion which found no statistically significant difference in the qual-
ity of life. It is noteworthy that the previous systematic review
The current systematic review aimed to obtain all the and meta-analysis examined the effect of exercise training
pertinent studies of exercise training in amyotrophic lateral after three months; however, due to the progressive nature
sclerosis to provide a complete description of the research. of the disease, investigation of the effect of exercise training
Sixteen studies were incorporated in the systematic review; at the end of the trials was required. In addition, the current
eight of them met the inclusion criteria for the meta-analysis. meta-analysis includes further studies and therefore extends
This enabled us to conduct meta-analysis of the effect of and validates the previous results.
exercise training on primary outcomes (functional ability, The SMD in functional ability which was statistically sig-
respiratory function, fatigue, pain, overall quality of life, nificant corresponds to an SMD of 0.84 (95% CI 0.33 to
upper and lower body strength and Vo2 peak). In addition, 1.36). A common interpretation of the SMD value is based
the authors intended to investigate the effectiveness of dif- on Cohen’s comments [32], that SMDs of around 0.2, 0.5 and
ferent exercise types on these factors. The results from the 0.8 are small, medium, and large effect sizes, respectively.
present meta-analysis demonstrate statistically significant Accordingly, our results indicated that exercise training exerts
differences in functional ability after the end of intervention a large effect on functional ability in patients with ALS. This
and after 6-month follow-up, Vo2 peak and overall quality of finding is in line with the previous studies that investigated the
life in patients with ALS. Results from our subgroup analy- effects of aerobic, endurance and resistance training on the
sis on different exercise training types suggest that combined functional ability. Of these, seven studies observed a signif-
training had a statistically significant effect on functional abil- icant difference in functional ability from exercise training
ity, whereas aerobic training, although somewhat effective, groups at the end of interventions [19,21,22,25–28], How-
did not reach a statistically significant difference. In addi- ever, three other studies reported no statistically significant
tion, combined training had a statistically significant effect differences among the patients who participated in aerobic
on upper and lower body strength and Vo2 peak. and resistance training programs [11,23,24]. Since exercise
However, the sensitivity analysis for functional ability prescriptions differed among trials, subgroup analysis was
after the end of intervention and after 6-month follow-up, conducted to better understand the effects of the different
Vo2 peak and overall quality of life did not demonstrate a sig- types of exercise (aerobic, resistance, combined exercise pro-
nificant difference with the removal of the studies judged to grams) on functional ability. Results of the subgroup analyses
be at a high risk of bias [10,17–20]. Therefore, the differ- indicated a positive trend in favor of combined training and
ences between the meta-analysis and the sensitivity analysis a significant large effect size was found (SMD = 1.70).
in these outcomes suggest that those results were not quite In addition, the mean difference of six months’ follow-up
consistent, which may be due to the high level of heterogene- functional ability after the end of the exercise interventions
ity of the studies. It should be noted that sensitivity analysis phase was statistically significant. These results indicated that
26 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

exercise training decreased the rate of decline of functional cise training is beneficial for Vo2 peak improvement in ALS
ability, even 6 months after the end of intervention (while patient
functional ability reduction in the control group was faster). The overall methodological quality of the studies that were
The current systematic review is the first to include included in the meta-analysis was considered low to medium
meta-analysis for respiratory function, which demonstrated with an approximate mean score of 3; the methodology score
a positive effect size in respiratory function, but did not show ranged from 2 to 4 on the Cochrane Collaboration Risk of
a statistically significant difference. In addition, our meta- Bias Tool. Less than half the studies (3 studies) did not report
analysis results did not show a significant difference in pain using an adequate sequence generation and none of the stud-
and in fatigue. The cause of fatigue in patients with ALS ies concealed allocation which can influence selection bias
is not precisely defined and may have several causes with of patients with a slow or fast progression in the selected
peripheral or central origin [33]. For this reason, it has not studies and lead to overestimation of effect sizes. In addition,
been conclusively stated whether exercise training affects the although it is desirable to blind participants and personnel,
patient’s fatigue. However, some studies reported that fatigue due to the nature of interventions of the included studies this
is an independent symptom and it can be treated after treat- was impossible. In addition, there was high attrition bias in
ment of other symptoms, such as pain etc. [34]. Thus, it can all of the included studies that leads to an instability in the
be interpreted that longer periods of exercise training may be effect size. The authors conducted a sensitivity analysis and
needed to reduce pain and fatigue in ALS patients. excluded low quality studies, but this analysis did not con-
The findings on the quality of life are supported by a firm robustness or the reliability of the results. Although, it is
recent study which showed that exercise interventions have a necessary to mention that sensitivity analysis was restricted
positive effect on quality of life of ALS patients [18]. How- to only two trials, the small number of studies and patients
ever, Dal Bello-hass and Florence (2013) investigated SF-36 may have influenced the overall effect. Moreover, the hetero-
subscale scores, and found no significant difference in any geneity levels of the included studies were considered high
subscales (except for Physical Function subscale). It should for functional ability, fatigue severity, upper body strength
be noted that since just two studies were included, this finding and Vo2peak. The divergent findings regarding the effective-
should be viewed with more caution. ness of exercise training in ALS patients may be due to the
In the present meta-analysis there was no significant dif- difference in physiotherapy schedule that was applied to the
ference in the upper and lower body strength. In addition, control groups of the included studies including usual care,
no significant differences were found in all of the included stretching and range of motion exercises. However, due to the
studies of systematic review (except for one open trial study) small number of included studies the authors could not per-
for muscle strength scales, suggesting that exercise training form meta-regression to find out the source of heterogeneity,
has a subtle effect on ALS patients. The open trial study this finding should thus be treated with considerable caution.
demonstrated that hand grip and maximal knee extensor was The other challenge of developing exercise training pro-
significantly improved [11]; however, in this study, data was tocols for ALS patients based on a review study is that in the
obtained from too small sample size. Moreover, the results published literature there was variability in the type (aerobic
of the subgroup analyses for upper body strength indicated a training, resistance training, muscle endurance training) and
statistically significant effect in favor of combined training; dose (frequency, intensity, length of intervention) of exercise
however, there was only one RCT for combined training in programs prescribed (frequency of intervention ranging from
this analysis and future studies should be more focused on 2 to 7 session per week, length of intervention ranging from 5
this concern to provide more solid results. In addition, the weeks to 6 month), and a lack of sufficient detail reported to
authors found a positive trend in favor of combined training conduct comparative subgroup meta-analyses. Thus, further
on lower body strength, although this was not statistically research is necessary to clearly establish the optimal type and
significant. It seems that exercise training cannot completely dose of exercise training for use in this population.
reduce the level of muscle weakness in patients with ALS
and more studies are required with different doses and types
of exercise training programs. Conclusion
The current systematic review is the first to include meta-
analysis for Vo2 peak, that demonstrated a positive effect size Based on current evidence, it remains difficult to deter-
in Vo2 peak in favor of the intervention group. The calculated mine whether exercise training in ALS patients is safe and
pooled SMD was 1.26 which showed a very large posi- feasible. Moreover, the results of the present meta-analysis
tive effect size. Interestingly, aerobic exercise has not been should be interpreted with caution as included trials were
able to increase the Vo2 peak in these patients, while com- generally of poor quality and risk of bias may be present
bined exercise has successfully increased the Vo2 peak, with in this review. Also, considering the diversities in the types,
a medium positive effect size (SMD = 0.54). Taken together, frequencies and durations of exercise interventions, and the
this finding indicates that combined exercise is a good strat- limitations in the methodological quality of included studies,
egy to improve reduced Vo2 peak in ALS patients. However, the number of included studies and their small sample size,
more research is needed to elucidate which kind of exer- the findings of this review must be interpreted cautiously.
 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28 27

The high level of heterogeneity in the studies included in [10] Lunetta C, Lizio A, Sansone VA, Cellotto NM, Maestri E, Bettinelli M,
the meta-analysis may explain why the results of the meta- et al. Strictly monitored exercise programs reduce motor deterioration
in ALS: preliminary results of a randomized controlled trial. J Neurol
analysis and the sensitivity analysis were not consistent with
2016;263(1):52–60.
respect to these outcomes. The authors of this meta-analysis [11] Jensen L, Djurtoft JB, Bech RD, Nielsen JL, Jørgensen LH, Schrøder
recommend further high quality RCTs to provide robust data HD, et al. Influence of resistance training on neuromuscular func-
on feasibility, safety and the effectiveness of various types of tion and physical capacity in ALS patients. J Neurodegener Dis
exercise interventions in patients with ALS. 2017;2017:1436519.
[12] Ferreira GD, Costa ACC, Plentz RD, Coronel CC, Sbruzzi G.
Conflict of interest: The authors declare that there is no
Respiratory training improved ventilatory function and respiratory
conflict of interest.Ethical approval: Not applicable. muscle strength in patients with multiple sclerosis and lateral amy-
otrophic sclerosis: systematic review and meta-analysis. Physiotherapy
2016;102(3):221–8.
Acknowledgments [13] Dal Bello-Haas V, Florence JM. Therapeutic exercise for people
with amyotrophic lateral sclerosis or motor neuron disease. Cochrane
Database Syst Rev 2013;(5).
All authors have declared: no support from any organiza- [14] Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M,
tion for the submitted work; no financial relationships with et al. Preferred reporting items for systematic review and meta-analysis
any organizations that might have an interest in the submit- protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4(1):1.
ted work in the previous three years; no other relationships or [15] Caspersen CJ, Powell KE, Christenson GM. Physical activity, exer-
cise, and physical fitness: definitions and distinctions for health-related
activities that could appear to have influenced the submitted
research. Public Health Rep 1985;100(2):126–31.
work. [16] Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD,
et al. The Cochrane Collaboration’s tool for assessing risk of bias in
randomised trials. BMJ 2011;343:d5928.
Appendix A. Supplementary data [17] Braga ACM, Pinto A, Pinto S, de Carvalho M. The role of moderate
aerobic exercise as determined by cardiopulmonary exercise testing in
ALS. Neurol Res Int 2018;2018:8218697.
Supplementary material related to this article can be [18] Drory VE, Goltsman E, Reznik JG, Mosek A, Korczyn AD. The value of
found, in the online version, at https://doi.org/10.1016/ muscle exercise in patients with amyotrophic lateral sclerosis. J Neurol
j.physio.2021.04.005. Sci 2001;191(1–2):133–7.
[19] Ferri A, Lanfranconi F, Corna G, Bonazzi R, Marchese S, Magnoni A,
et al. Tailored exercise training counteracts muscle disuse and atten-
uates reductions in physical function in individuals with amyotrophic
References lateral sclerosis. Front Physiol 2019;10:1537.
[20] Merico A, Cavinato M, Gregorio C, Lacatena A, Gioia E, Piccione F,
[1] Cleveland DW, Rothstein JD. From Charcot to Lou Gehrig: deci- et al. Effects of combined endurance and resistance training in Amy-
phering selective motor neuron death in ALS. Nat Rev Neurosci otrophic Lateral Sclerosis: a pilot, randomized, controlled study. Eur J
2001;2(11):806. Transl Myol 2018;28(1).
[2] Byrne S, Walsh C, Lynch C, Bede P, Elamin M, Kenna K, et al. [21] Pinto AC, Alves M, Nogueira A, Evangelista T, Carvalho J, Coelho
Rate of familial amyotrophic lateral sclerosis: a systematic review A, et al. Can amyotrophic lateral sclerosis patients with respiratory
and meta-analysis, Journal of Neurology. Neurosurg Psychiatry insufficiency exercise? J Neurol Sci 1999;169(1–2):69–75.
2011;82(6):623–7. [22] Braga A, Pinto A, Pinto S. Tele-monitoring of a home-based exercise
[3] Jensen L, Jørgensen L, Bech R, Frandsen U, Schrøder H. Skeletal mus- program in amyotrophic lateral sclerosis: a feasibility study. Eur J Phys
cle remodelling as a function of disease progression in amyotrophic Rehabil Med 2018;54(3):501–3.
lateral sclerosis. BioMed Res Int 2016;2016:5930621. [23] Sivaramakrishnan A, Madhavan S. Recumbent stepping aerobic exer-
[4] Corcia P, Pradat PF, Salachas F, Bruneteau G, Le Forestier N, Seilhean cise in amyotrophic lateral sclerosis: a pilot study. Neurol Sci 2019:1–8.
D, et al. Causes of death in a post-mortem series of ALS patients. [24] Clawson LL, Cudkowicz M, Krivickas L, Brooks BR, Sanjak M, Allred
Amyotroph Lateral Scler 2008;9(1):59–62. P, et al. A randomized controlled trial of resistance and endurance
[5] Garber CE, Blissmer B, Deschenes MR, Franklin BA, Lamonte MJ, exercise in amyotrophic lateral sclerosis. Amyotroph Lateral Scler
Lee I-M, et al. Quantity and quality of exercise for developing and Frontotemporal Degener 2018;19(3–4):250–8.
maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness [25] Sanjak M, Bravver E, Bockenek WL, Norton HJ, Brooks BR. Supported
in apparently healthy adults: guidance for prescribing exercise. Med Sci treadmill ambulation for amyotrophic lateral sclerosis: a pilot study.
Sports Exerc 2011;43(7):1334–59. Arch Phys Med Rehabil 2010;91(12):1920–9.
[6] Harwood CA, McDermott CJ, Shaw PJ. Physical activity as an exoge- [26] Kitano K, Asakawa T, Kamide N, Yorimoto K, Yoneda M, Kikuchi
nous risk factor in motor neuron disease (MND): a review of the Y, et al. Effectiveness of home-based exercises without supervision
evidence. Amyotroph Lateral Scler 2009;10(4):191–204. by physical therapists for patients with early-stage amyotrophic lateral
[7] Mahoney DJ, Rodriguez C, Devries M, Yasuda N, Tarnopolsky sclerosis: a pilot study. Arch Phys Med Rehabil 2018;99(10):2114–7.
MA. Effects of high-intensity endurance exercise training in the [27] Pegoraro V, Merico A, Angelini C. MyomiRNAs dysregulation in ALS
G93A mouse model of amyotrophic lateral sclerosis. Muscle Nerve rehabilitation. Brain Sci 2019;9(1):8.
2004;29(5):656–62. [28] Kalron A, Mahameed I, Weiss I, Rosengarten D, Balmor GR, Heching
[8] Tsitkanou S, Della Gatta P, Foletta V, Russell A. The role of exercise M, et al. Effects of a 12-week combined aerobic and strength training
as a non-pharmacological therapeutic approach for amyotrophic lateral program in ambulatory patients with amyotrophic lateral sclerosis: a
sclerosis: beneficial or detrimental? Front Neurol 2019;10:783. randomized controlled trial. J Neurol 2021:1–10.
[9] Dal Bello-Haas V, Florence J, Kloos A, Scheirbecker J, Lopate G, [29] van Groenestijn AC, Schröder CD, van Eijk RP, Veldink JH,
Hayes S, et al. A randomized controlled trial of resistance exercise in Kruitwagen-van Reenen ET, Groothuis JT, et al. Aerobic exercise ther-
individuals with ALS. Neurology 2007;68(23):2003–7.
28 M. Rahmati, F. Malakoutinia / Physiotherapy 113 (2021) 12–28

apy in ambulatory patients with ALS: a randomized controlled trial. [33] Rabkin JG, Gordon PH, McElhiney M, Rabkin R, Chew S, Mitsumoto
Neurorehabil Neural Repair 2019;33(2):153–64. H. Modafinil treatment of fatigue in patients with ALS: a placebo-
[30] Stone J, Gurunathan U, Glass K, Munn Z, Tugwell P, Doi SA. Strati- controlled study. Muscle Nerve 2009;39(3):297–303.
fication by quality induced selection bias in a meta-analysis of clinical [34] Lou J-S. Fatigue in amyotrophic lateral sclerosis. Phys Med Rehabil
trials. J Clin Epidemiol 2019;107:51–9. Clin N Am 2008;19(3):533–43.
[31] Ahn E, Kang H. Introduction to systematic review and meta-analysis. [35] Braga AC, Pinto A, Pinto S, de Carvalho M. Tele-monitoring of a home-
Korean J Anesthesiol 2018;71(2):103–12. based exercise program in amyotrophic lateral sclerosis: a feasibility
[32] Cohen J. Statistical power analysis for the behavioral sciences. Aca- study. Eur J Phys Rehabil Med 2018;54(3):501–3.
demic press; 2013.

Available online at www.sciencedirect.com

ScienceDirect