Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Tholen, 2019

SR and meta-analysis of (SRs and RCTs

Literature search up to February 2018

Inclusion criteria:

- Adult patients with MS

- Movement therapy interventions

- Comparison with standard treatment, other interventions or placebo treatment

- Outcome improving mobility (walking, balance, falls)

- Study design systematic review, randomized controlled trial or meta-analyzes

- At least 5 out of 10 points on PEDRO scale

- sample size of at least 10 patients with MS

Exclusion criteria:

- Drug therapy

- Aid

- Functional Electrical Stimulation

- Interventions to change behaviour, training excluding the upper extremities, respiratory therapy

- Wheelchair mobility

See below

See below

See below

See below

- Voor ‘fall’ wordt verwezen naar Comber 2017, maar deze wordt niet uitgewerkt

- één exercised artikel wordt nergens aangehaald

- physiotherapy artikelen worden nergens aangehaald

- één virtual reality artikel wordt nergens aangehaald

Included SR

Author, year

Number of included studies

Intervention

Comparison/control

Outcome measure

Result

Taylor, 2017

8

Tai Chi

Physical function: balance; gait; flexibility;

and strength.

Psychosocial function: fatigue; quality

of life; disease symptoms; perception

of disease; and mood.

A total of 91 articles were retrieved, with 3 additional articles identified through

reviewing bibliographies of relevant articles. A total of 8 studies (randomized

controlled trials, n = 3; quasi-experimental, n = 5) enrolled 193 participants with

multiple sclerosis. Studies were conducted in the USA (n = 3), Europe (n = 3),

Iran, (n = 1), and India (n = 1). A total of 3 studies reported using the Yang style

of Tai Chi (not specified, n = 5 studies). The Tai Chi intervention averaged 27

sessions over 11 weeks. Study quality scores for the randomized controlled

trials had a mean score of 23 (range 19 to 26), while quality scores for quasiexperimental

studies had a mean score of 20 (range 13 to 26). Overall, participants

enrolled in Tai Chi had better balance, gait and flexibility, less fatigue

and depression, and better quality of life after the intervention; though mixed

results were reported.

Eastern philosphy (Tai Chi):
- Walking Speed

- Balance/Safety

Zou, 2017

10

Tai Chi

Health outcome measures (e.g., balance,

strength, mobility, flexibility,

pain, fatigue, depression, or quality

of life)

The existing evidence supports the effectiveness of Tai Chi on improving quality

of life (QOL) and functional balance in MS patients. A small number of these

studies also reported the positive effect of Tai Chi on flexibility, leg strength,

gait, and pain. The effect of Tai Chi on fatigue is inconsistent across studies.

Eastern philosphy (Tai Chi):
- Walking Speed

- Balance/Safety

Zhang

2017

10

whole-body

vibration

balance (e.g. Berg Balance Scale,

BBS), mobility (e.g. Timed Up and

Go, TUG), gait (e.g. gait analysis, GA),

spasticity (e.g. modified Ashworth

scale, MAS), muscle strength (e.g.

hand-held dynamometer) and other

related outcome measurements

Meta-analysis revealed no significant benefit of WBV in Berg balance scale

(standard mean difference (SMD) 0.06; 95% confidence interval (CI) -0.54 to

0.66; p = 0.85; I2= 69%) and Timed Up and Go test (SMD -0.15; 95% CI -0.41

to 0.10; p = 0.24; I2= 0%) when compared with outcomes in the control groups.

A significant difference in muscle strength was observed in knee extensor (SMD

0.43, 95% CI 0.05 to 0.81; p = 0.03; I2= 0%). There was no sufficient evidence

of benefit of WBV in reducing spasm, relieving fatigue, improving gait or for

enhancing well-being.

Vibration:

- Balance/Safety

Kang

2016

7

whole body

vibration

muscle performance and functional

mobility

Based on the inclusion and exclusion criteria, 7 trials including 201 subjects

were included in this meta-analysis. The results showed that the WBV significantly

improved the knee extensor strength (WMD=13.74, 95% CI 4.56 to

22.93, P=0.003). However, there were no significant differences between WBV

and control groups on knee flexor strength (95% CI -26.22 to 41.83, P=0.65),

Timed Up and Go test (TUG, WMD=0.31, 95% CI: -0.91 to 3.42, P=0.62) and

walking speed (SMD=0.42, 95% CI: -0.78 to 1.62, P=0.49).

Vibration:

- Walking speed

- Balance/Safety

Kantele

2015

7

whole body

vibration

timed up and go, walking speed, balance,

walking endurance

Relevant group-based data for analysis were available from 109 patients in WBV

groups and from 100 control patients; 41 patients withdrew from the studies.

Quality assessment revealed that the WBV training protocols were heterogeneous

and the methodological quality of the studies was generally poor. We

found borderline indication for improved 2–6 min walking endurance (ES = 0.25

(95% CI = -0.06–0.0.55)) favoring WBV training whereas no benefits were indicated

for short-distance (20m or less) walking speed or balance.

Vibration:

- Walking Speed

- Balance/Safety

Cramer

2014

7

Yoga

Primary outcomes: health-related

quality of life, fatigue, and/or mobility.

Secondary outcomes: mood, cognitive

function, and safety

Seven RCTs with a total of 670 patients were included. Evidence for short-term

effects of yoga compared to usual care were found for fatigue (standardized

mean difference (SMD) = -0.52; 95% confidence intervals (CI) = -1.02 to -0.02;

p = 0.04; heterogeneity: I2 = 60%; Chi2 = 7.43; p = 0.06) and mood (SMD = -0.55;

95%CI = -0.96 to -0.13; p = 0.01; heterogeneity: I2 = 0%; Chi2 = 1.25; p = 0.53),

but not for health-related quality of life, muscle function, or cognitive function.

Eastern philosophy (Yoga):

- Waling speed

- Endurance

Kjølhede

2012

12

progressive

resistance

training

strength, functional capacity, QoL,

fatigue, mood, balance et cetera.

Strong evidence regarding the beneficial effect of PRT on muscle strength was

observed. Regarding functional capacity, balance and self-reported measures

(fatigue, quality of life and mood) evidence is less strong, but the tendency is

overall positive. Indications of an effect on underlying mechanisms such as

muscle morphological changes, neural adaptations and cytokines also exist, but

the studies investigating these aspects are few and inconclusive.

Weight training:
- Walking speed

- Balance/Safety

Swinnen

2012

8

treadmill

training (TT),

body-weightsupported

TT

(BWSTT), and

robot-assisted

TT (RATT).

gait-related outcome measurements

In total 161 persons with MS were involved (TT, BWSTT, or RATT, 6 to 42 sessions;

2 to 5x/week; 3 to 21 weeks). Significant improvements in walking speed and

endurance were reported. Furthermore, improvements of step length, doublesupport

time, and Expanded Disability Status Scale were found.

RAGT. Conventional treadmill:

- Walking speed

- Endurance

Pau

2018

22

11/11

aerobic and strength training, 6 months

no structured

PA program

three-dimensional gait analysis

(Gait Profile Score, dynamic

range of motion (ROM) of hip,

knee, and ankle joints)

The training originated significant improvements in stride

length, gait speed and cadence in the APA group, while GPS

and GVS scores remained practically unchanged. A trend of

improvement was also observed as regard the dynamic ROM

of hip, knee, and ankle joints. No significant changes were

observed in the CG for any of the parameters considered.

Combined:

- Walking speed

Kalron

2017

50

25/25

Pilates, 12 weeks

standardized

physical

therapy

Time Up and Go Test, 2 and 6

minute walk test, Functional

Reach Test, Berg Balance Scale,

Four Square Step Test, Multiple

Sclerosis Walking Scale,

12-item patient-rated index,

questioning the perceived

impact of multiple sclerosis on

walking ability and the Modified

Fatigue Impact Scale

At the termination, both groups had significantly increased

their walking speed (p = 0.021) and mean step length

(p = 0.023). According to the 2-minute and 6-minute walking

tests, both groups at the end of the intervention program

had increased their walking speed. Mean (SD) increase in

the Pilates and physical therapy groups were 39.1 (78.3)

and 25.3 (67.2) meters, respectively. There was no effect of

group x time in all instrumented and clinical balance and gait

measures.

Pilates:

- Walking ability

- Walking speed

- Balance/Safety

Monjezi

2017

47

23/24

dual-task balance

training, 4 weeks

singletask

balance

training

10 Meter Walk test and Timed

Up and Go test under singletask

and dual-task conditions

Only 38 patients completed the treatment plan. There was no

difference in the amount of improvement seen between the two

study groups. In both groups there was a significant effect of

time for dual-10 Meter Walk Test (F(1,36) = 11.33, p = 0.002)

and dual-Timed Up-and-Go (F(1,36) = 14.27, p = 0.001) but not

for their single-tasks. Moreover, there was a significant effect

of time for Activities-specific Balance Confidence, Berg Balance

Scale, and Functional Gait Assessment (p < 0.01).

Balance:

- Walking Speed

- Balance/Safety

Seebacher

2017

112

38/36/38

1. motor imagery

with

music, 4 weeks

2. motor imagery

with

metronome

cues, 4 weeks

usual

treat-ment

and weekly

phone

calls

Timed 25-Foot Walk (T25FW);

6-Minute Walk Test (6MWT)

Of the 112 participants randomised, 101 completed the

study. Compared to controls, both interventions significantly

improved walking speed, distance and perception. Significant

improvements in cognitive but not psychosocial fatigue

were seen in the intervention groups, and physical fatigue

improved only in the music-based group. Both interventions

improved QoL; however, music-cued motor imagery was superior

at improving health-related QoL.

Motion visualisation:

- Walking speed

Pompa

2017

50

25/25

robot-assisted

gait training

(RAGT), 12 sessions

conventional

walking

training

(CWT)

walking capacity (2-minute

walking test) and the walking

ability (Functional Ambulatory

Category)

The number of subjects who achieved a clinical significant

improvement was significantly higher in RAGT than in CWT

(p < 0.05 for both primary outcome measures). RAGT also led

to an improvement in all the other clinical parameters (global

ability p < 0.001, global mobility p < 0.001, EDSS p = 0.014

and fatigue p = 0.001).

RAGT:

- Walking ability

- Endurance

Carling

2017

51

25/26

balance group

exercise programme, 7 weeks

waiting

list

Fatigue Scale for Motor and Cognitive

Functions (FSMC), Trunk

Impairment Scale (TIS), Timed

Sit-To-Stand test (TstS), Postural

Sway, Berg Balance Scale (BBS),

Falls Efficacy Scale-International

(FES-I), 12-item MS Walking

Scale (MSWS), Timed Up and Go

(TUG), 10-m walk test (10WT)

and 2-minute walk test (2MWT)

Post-intervention, the exercise group showed statistically

significant improvement (p = 0.015) in BBS and borderline

significant improvement in MS Walking Scale (p = 0.051),

both with large effect sizes (3.66; -2.89). No other significant

differences were found between groups. In the group starting

late, numbers of falls and near falls were statistically significantly

reduced after exercise compared to before (p < 0.001;

p < 0.004).

Balance:

Balance/Safety

Manca

2017

30

15/15

contralateral

strength training

(CST), 6 weeks

direct

strength

training

ankle dorsiflexors’ strength,

muscle endurance of the ankle

dorsiflexors to fatigue, and

mobility (2-minute walking

test, 6-minute walking test,

10-m timed walk, and Timed

Up and Go test)

In the more affected limb of both groups, pre- to postintervention

significant increases in maximal strength (P ≤ .006) and

fatigue endurance (P≤.04) were detected along with consistent

retention of these improvements at follow-up (P ≤ .04). At midintervention,

the direct strength training group showed significant

improvements (P ≤ .002), with no further increase at postintervention,

despite training continuation. Conversely, the CST group

showed nonsignificant strength gains, increasing to significance

at postintervention (P ≤ .003). In both groups, significant pre- to

postintervention improvements in mobility outcomes (P ≤ .03),

not retained at follow-up, were observed.

Weight training:
- Walking speed

- Endurance

- Balance/Safety

Afrasiabifar

2018

75

25/25/25

1. Cawthorne-

Cooksey

Exercise, 12 weeks

2. Frenkel

Exercise, 12 weeks

routine

care

Berg Balance Scale

Seventy-two patients completed the study. At the end of the

intervention, there was a statistically significant improvement

in Berg Balance Scale in the Cawthorne-Cooksey group

(n = 24) in comparison with the other two groups (p = 0.001).

In the Frenkel group (n = 23), the improvement was statistically

greater than the changes in the control group (n = 25), but

it did not appear to be clinically significant. The Berg Balance

Scale score increased to 8.9 in the Cawthorne-Cooksey group

and 2.3 in the Frenkel group, while it decreased to 1.2 in the

control group. When comparing inter-group changes, Berg

Balance Scale showed significant improvements in favor of

the Cawthorne-Cooksey group after the intervention (p < 0.05).

Balance:

Balance/safety

Peruzzi

2017

31

15/16

virtual reality

based treadmill

training

(VR-TT), 6 weeks

treadmill

training

(TT)

gait analysis under single and

dual task conditions, six-minute

walk test (6Mwt), 10mwt,

the timed up and go (TUG),

Berg balance scale (Berg); four

square step test – (FSST)

Subjects in both the groups significantly improved the walking

endurance and speed, cadence and stride length, lower

limb joint ranges of motion and powers, during single and

dual task gait. Moreover, subjects in the experimental group

also improved balance, as indicated by the results of the

clinical motor tests (p < 0.05). Between-group comparisons

revealed that the experimental group improved significantly

more than control group in hip range of motion and hip generated

power at terminal stance at post-training.

Conventional treadmill:

- Walking speed

- Endurance

- Balance/Safety

Sandroff

2017

83

43/40

Supervised, 6 months

multi-modal

exercise training

(progressive

aerobic,

resistance, and

balance)

stretching-

andtoning

activities

6MWT, timed 25-foot walk

(T25FW), Multiple Sclerosis

Walking Scale-12 (MSWS-12),

Functional Ambulation Profile

(FAP), velocity, cadence, step

time, VO2peak, peak power,

KE peak torque, Cognitive processing

speed, Symbol Digit

Modalities Test (SDMT), Paced

Auditory Serial Addition Test

(PASAT)

There were statistically significant improvements in six-minute

walk performance (F(2158) = 3.12, p = 0.05, ηp² = 0.04),

peak power output (F(2150) = 8.16, p < 0.01, ηp² = 0.10),

and Paced Auditory Serial Addition Test performance

(F(2162) = 4.67, p = 0.01, ηp² = 0.05), but not gait outcomes,

for those who underwent the intervention compared with

those who underwent the control condition.

Combined:

- Walking speed

- Endurance

Bulguroglu

2017

38

12/13/13

1. Mat Pilates, 8 weeks

2. Reformer

Pilates, 8 weeks

home

program

(relaxation

and

respiration

exercises)

Single Leg Stance, Timed Up

and Go Test (TUG), Activities

Specific Balance Confidence

Scale (ABC), Core stability, situps

test, modified push-ups

test, Fatigue Severity Scale,

Turkish version of the Multiple

Sclerosis Quality of Life Instrument

(MSQOL-54)

Balance, functional mobility, core stability, fatigue severity

and quality of life improved after Pilates in Mat and Reformer

Pilates groups (p < 0.05). On the other hand, we couldn’t

find any changing in the control group (p > 0.05). When the

gain obtained in the Pilates groups is compared, it has been

observed that progress has been more in trunk flexor muscle

strength in the Reformer Pilates group (p < 0.05) and that the

gain has been similar in the other parameters (p > 0.05).

Pilates:

- Balance/Safety

Vermöhlen

2018

70

32/38

Hippotherapy

+ standard care, 12 weeks

standard

care

Berg Balance Scale

Covariance analysis of the primary endpoint resulted in a

mean difference in BBS change of 2.33 (95% confidence

interval (CI): 0.03–4.63, p = 0.047) between intervention (

n = 32) and control ( n = 38) groups. Benefit on BBS was largest

for the subgroup with an Expanded Disability Status Scale

(EDSS) ⩾ 5 (5.1, p = 0.001). Fatigue (-6.8, p = 0.02) and spasticity

(-0.9, p = 0.03) improved in the intervention group. The

mean difference in change between groups was 12.0

(p < 0.001) in physical health score and 14.4 (p < 0.001) in

mental health score of Multiple Sclerosis Quality of Life-54

(MSQoL-54).

Hippotherapy:

Balance/Safety

Escudero-

Uribe

2017

55

19/18/18

1. Whole-body

vibration

(WBV), 12 weeks

2. Balance

Trainer (each

in addition

to a standard

exercise programme), 12 weeks

waiting

list

Fatigue Severity Scale (FSS),

Modified Fatigue Impact Scale

(MFIS), Beck Depression Inventory

(BDI-II), Multiple Sclerosis

International Quality of Life

(MusiQoL), gait parameters

(GAITRite electronic walkway)

Significant improvements in fatigue and mood were identified

for both intervention groups (p < 0.05). Gait parameters also

improved significantly in the WBV group: velocity and step

length increased (12.8% and 6.5%, respectively; p < 0.005),

and step time, stance time, double support time, and step

length asymmetry decreased (-5.3%, -1.4%, -5.9%, and

-43.7%, respectively; p < 0.005).

Vibration:

- Walking speed

Calabrò

2017

40

20/20

robot-assisted

gait training

(RAGT)

equipped with

virtual reality

(VR) system, 8 weeks

RAGT

without

VR

time up and go test (TUG), the

Berg Balance Scale (BBS), and

the Coping Orientation to Problem

Experienced (COPE)

Effect sizes were very small and non-significant between

the groups for Berg Balance Scale (-0.019, 95% CI -2.403

to 2.365) and TUG (-0.064, 95% CI -0.408 to 0.536) favoring

RAGT+VR. Effects were moderate-to-large and significant

for positive attitude (-0.505, 95% CI -3.615 to 2.604) and

problem-solving (-0.905, 95% CI -2.113 to 0.302) sub-items

of Coping Orientation to Problem Experienced, thus largely

favoring RAGT+VR.

RAGT:

-Balance/Safety

Braendvik

2016

28

13/15

1. treadmill, 8 weeks

2. strength

Training, 8 weeks

None

The Functional Ambulation

Profile

Two participants were lost to follow-up, and 11 (treadmill)

and 15 (strength training) were left for analysis. The treadmill

group increased their Functional Ambulation Profile score

significantly compared with the strength training group

(p = 0.037). A significant improvement in walking work economy

(p = 0.024) and a reduction of root mean square of vertical

acceleration (p = 0.047) also favoured the treadmill group.

Conventional treadmill:

- Walking speed

Cattaneo

2018

119

78/41

treatment

aimed at

improving

balance and

mobility, 6-10 weeks

treatments

to reduce

limitations

at

activity

and body

function

level

frequency of fallers (> 1 fall in

two months) and responders

(> 3 points improvement) at the

Berg Balance Scale (BBS))

One hundred and nineteen participants were randomised.

Following treatment frequency of fallers was 22% in the intervention

group and 23% in the control group, odds ratio (OR)

and (confidence limits) 1.05 (0.41 to 2.77). Responders on

the BBS were 28% in the intervention group and 33% in the

control group, OR 0.75 (0.30 to 1.91). At follow up ORs for

fallers and responders at BBS were 0.98 (0.48 to 2.01) and

0.79 (0.26 to 2.42), respectively.

Balance:

- Balance/Safety

- Falls

Fox

2016

100

33/35/32

1. Pilates, 12 weeks

2. standard

Exercises, 12 weeks

3. relaxation, 12 weeks

none

10-Meter Timed Walk Test

(10MTW)

100 participants (mean age 54 years, 74% female) were randomised.

Six relapsed (withdrawn), leaving 94 for intention

to treat analysis. There was no significant difference between

Pilates and relaxation (p > 0.05) in mean 10MTW. At 12 weeks

there was a mean reduction of 4.2 seconds for SE compared

with relaxation (95% confidence interval (relax minus SE)

0.0 to 8.4) and 3.7 seconds compared with Pilates (95% CI

(Pilates minus SE) -0.4 to 7.8). At 16 weeks, mean 10MTW for

SE remained quicker than Pilates and relaxation, although

non-significant (both p > 0.05). There were no significant differences

between Pilates and relaxation for any secondary

outcome (p > 0.05).

Pilates:
Walking speed

Kalron

2016

32

16/16

virtual reality

(VR) balance

training program, 6 weeks

conventional

balance

exercise

functional reach test (FRT);

Berg balance scale (BBS);

Four Square Step Test (FSST);

Falls Efficacy Scale International

(FES-I);

instrumented posturography

outcome measures

Final analysis included 30 patients (19 females, 11 males;

mean age (SD) 45.2 (11.6) years; mean EDSS (SD) 4.1 (1.3),

mean disease duration (SD) 11.0 (8.9) years). Both groups

showed a main effect of time on the center of pressure (CoP)

path length with eyes open (F = 5.278, p = 0.024), sway rate

with eyes open (F = 5.852, p = 0.035), Functional Reach Test

(F = 20.841, p = 0.001), Four Square Step Test (F = 9.011,

p = 0.031) and the Fear of Falls self-reported questionnaire

(F = 17.815, p = 0.023). In addition, significant differences

in favor of the VR program were observed for the group x

time interactions of the Functional Reach Test (F = 10.173,

p = 0.009) and fear of falling (F = 6.710, p = 0.021).

Balance:

Balance/Safety

Samaei

2016

34

17/17

1. downhill

treadmill

walking

(10%), 4 weeks

2. uphill treadmill

walking

(10%), 4 weeks

None

2-Minute Walk Test, Timed

25-Foot Walk test, and Timed

Up and Go test

After the intervention, significant improvement was found

in the downhill group versus the uphill group in terms of

fatigue, mobility, and disability indices; functional activities;

balance indices; and quadriceps isometric torque (P < .05).

The results were stable at 4-week follow-up.

Conventional treadmill:

- Walking speed

- Endurance

- Balance/Safety

Sangelaji

2016

40

10/10/10/

10

1. 1 aerobic

exercise

training and

3 resistance

exercise

training sessions

per

week., 8 weeks

2. 2 aerobic

exercise

trainings

and 2 resistance

exercise

training

sessions per

week., 8 weeks

3. 3 aerobic

exercise

trainings

and 1 resistance

exercise

training

session per

week., 8 weeks

no intervention

one repetition maximum (1RM)

test; Berg balance scale (BBS);

Timed up and go (TUG) test;

10-minute walk test (10MWT)

and 20MWT; 6MWT; Fatigue

severity scale (FSS); BMI;

For most tests, post-intervention values of the group 1, with

3-aerobic and 1-resistance exercises, were significantly

higher compared to control group (p < 0.050). However, no

significant progression was observed in the other two intervention

groups.

Combined:
- Walking speed

- Endurance

- Balance/Safety

Straudi

2016

58

30/28

robot-assisted

gait training

(RAGT), 6 weeks

conventional

walking

therapy

(CWT)

Ten-Meter Walk Test (10MWT),

6MWT

Walking endurance (p < 0.01) and balance (p < 0.01) were

improved among those in the RAGT group. Positive effects

on depression in both treatment groups were highlighted.

However, only among those in the RAGT group was perceived

physical functioning QoL increased. No significant effects on

fatigue were found.

Combined:

-Walking ability

- Walking speed

- Endurance

- Balance/Safety

Brichetto

2015

32

16/16

tailored

rehabilitation

treatments for

balance disorders

based

on visual,

somatosensory

and vestibular

deficits (PRG), 4 weeks

traditional

rehabilitation

exercises

(TRG)

Berg Balance Scale (BBS), the

composite score (CS) obtained

by computerized dynamic posturography

(CDP) test, Modified

Fatigue Impact Scale (MFIS)

BBS score showed a significant difference between pre- and

post-treatment scores of 6.3 and 2.0 points respectively for

PRG and TRG. CS showed a significant difference between

pre- and post-treatment scores of 16.6 and 7.6 points respectively

for PRG and TRG. No interaction effect was found for

MFIS score.

Balance:

Balance/Safety

Gandolfi

2015

80

39/41

specific training

to improve

central integration

of afferent

sensory inputs, 5 weeks

conventional

rehabilitation

Berg Balance Scale (BBS)

The experimental training program produced greater improvements

than the control group training on the BBS (p < 0.001),

the FSS (p < 0.002), number of falls (p = 0.002) and SOT

(p < 0.05).

RAGT:

- Walking speed

- Balance/Safety

Karpatkin

2015

29

intermittent

walking, 1 session

continuous

walking

6MWT; 100-mm-long line

visual analog scale of fatigue

(VAS-F)

Participants walked greater distances in the INT condition

compared to the CONT condition (P = 0.005). There was a significant

interaction of walking condition and time (P < 0.001),

indicating that the distances walked in the INT condition

changed across time. ΔVAS-F was significantly lower in the

INT condition than in the CONT condition (P = 0.036).

Conventional treadmill:

- Endurance

Kjølhede

2015

35

18/17

supervised PRT, 48 weeks

waiting

list

Timed 25 ft Walk Test (T25FWT),

Two-Minute Walk Test (2MWT),

5-Time Sit-To-Stand Test (5STS),

ascending stair climb test,

12-item MS Walking Scale

(MSWS-12), Isokinetic dynamometry,

surface electromyography

(EMG)

The training group significantly improved neuromuscular

function of the knee extensors and flexors, which translated

to improvements in functional capacity. Furthermore, the

improved functional capacity was maintained after 24 weeks

of self-guided physical activity. The waitlist group produced

similar patterns of changes after PRT.

Weight training

- Walking speed

- Balance/Safety

Moradi

2015

20

10/10

progressive

resistancetraining

program, 8 weeks

no intervention

EDSS; 10-meter timed walk test

(10TW); three minute step test;

Timed Up and Go test; muscle

strength testing; Flamingo

Stand test

Two patients of group E left the program. The other eight

subjects completed the program with no MS-related exacerbations/

complications. There was a significant change in

2 of 3 aspects of ambulatory function (three minutes step

test (p = 0.001), Timed Up and Go test (p = 0.009)), muscle

strength (p = 0.000), and EDSS (p = 0.014). Comparing the two

groups, we did not observe any significant change in »balance

« (p = 0.407).

Weight training

- Walking speed

- Balance/Safety

Robinson

2015

56

20/18/18

1. Wii Fit™

balance

exercises, 4 weeks

2. traditional

balance

training, 4 weeks

no intervention

postural sway (measured using

a force plate), gait (measured

using the GAITRite™ walkway),

technology acceptance

(measured using the UTAUT

questionnaire) and flow experience

(measured using the FSS

questionnaire)

There were significant improvements in bipedal postural sway

in both intervention groups when compared to the control

group; and no effects of either intervention on gait. There

were no significant differences between the interventions

in technology acceptance but on several dimensions of flow

experience the Wii Fit was superior to traditional balance

training. Both interventions showed improvements in disability

compared to control.

Balance:

Balance/Safety

Aydin

2014

40

20/20

Calisthenic

exercises hospital-

based, 12 weeks

Calisthenic

exercises

homebased

EDSS, MS International Quality

of Life Scale, 10-meter walking

test, Berg Balance Scale,

Fatigue Severity Scale and the

Hospital Anxiety Depression

Inventory

36 participants completed the exercise programme (hospital

based = 16, home based = 20). The mean age was

32.83 ± 3.64 years. The mean duration of disease was

6.97 ± 3.15 years. Hospital-based and home-based exercise

groups had significant improvements in the balance,

10-meter walking test, anxiety, and the quality of life after

the 12-week exercise programme. There was a significant

improvement in the hospital-based patients in terms of the

depression scores. No significant improvement was observed

in terms of fatigue in any of the groups. When both groups

were compared, the improvement in the balance and depression

scores of the hospital-based patients was significantly

higher than the home-based patients.

Combined:

- Walking speed

- Balance/Safety

Briken

2014

47

12/12/12/

11

1. arm ergometry, 12 sessions

2. rowing 3.

bicycle

ergometry, 12 sessions

no intervention

VO2, 6MWT, Symbol Digit Modalities

Test (SDMT), Verbal Learning

and Memory Test (VLMT),

Test Battery of Attention (TAP),

Achievement Testing System

(Leistungsprüfsystem (LPS)),

Regensburg Verbal Fluency Test

(RWT), Inventory of Depressive

Symptoms (IDS-SR), Modified

Fatigue Impact Scale (MFIS).

A total of 42 patients completed the trial (10.6% drop-out

rate). Significant improvements were seen in aerobic fitness.

In addition, exercise improved walking ability, depressive

symptoms, fatigue and several domains of cognitive function.

Ergometer:

- Endurance

Gandolfi

2014

26

14/12

end-effector

robot-assisted

gait training

(RAGT), 6 weeks

sensory

integration

balance

training

(SIBT)

gait speed (cm/s);

Berg balance scale (BBS)

Between groups comparisons showed no significant differences

on primary and secondary outcome measures over

time. Within group comparisons showed significant improvements

in both groups on the Berg Balance Scale (p = 0.001).

Changes approaching significance were found on gait speed

(p = 0.07) only in the RAGT group. Significant changes in balance

task-related domains during standing and walking conditions

were found in the SIBT group.

RAGT:

- Walking speed

- Balance/Safety

Gervasoni

2014

30

15/15

conventional

therapy including

15 min

of treadmill

training, 12 treatment sessions over a period of 2 weeks

conventional

therapy

Borg Scale = rate of perceived

exertion (RPE), heart rate (HR),

the Fatigue Severity Scale, Berg

Balance Scale, Dynamic gait

index (DGI)

Rate of perceived exertion (RPE), heart rate (HR), the Fatigue

Severity Scale, and the Berg Balance Scale were assessed

before and after rehabilitation. Despite a low HR (107 beats/

min), the RPE score was high (15.8) at baseline assessment.

In the experimental group, RPE decreased significantly to

12.8 (P = 0.04) after training. Treadmill training also had a

positive effect on HR, but no changes in the Fatigue Severity

Scale and balance were observed.

Conventional treadmill:

- Balance/Safety

Schatz

2014

29

15/14

Hippotherapie, 3 weeks

traditional

physiotherapy

6-minute walking test (6min-

WT), 10-meter walking test

(10mWT) and timed-get-upand-

go test (TGUG)

The 6minWT was significantly improved after three weeks in

both groups, whereas gait analysis was significantly different

only in HG. In addition, only HG but not CG resulted in

positive effects directly after the intervention. Accordingly,

pooled analysis showed significant improvements in 6minWT,

10mWT and TGUG from Tuesday to Wednesday only in HG,

with no differences in CG.

Hippotherapy:

- Walking speed

- Endurance

- Balance/Safety

Straudi

2014

24

12/12

High-intensity

rehabilitative

task-oriented

circuit training

(TOCT), 14 weeks

usual care

10 Meter Walk Test (10MWT),

6 Minute Walk Test (6MWT),

Timed Up and Go test (TUG),

Dynamic Gait Index (DGI),

Fatigue Severity Scale (FSS),

Multiple Sclerosis Walking

Scale-12 (MSWS-12), Multiple

Sclerosis Impact Scale-29

(MSIS-29), Feasibility patientreported

questionnaire

After TOCT subjects reported a positive global rating on the

received treatment. At 3 months, we found a 58.33% of

adherence to the home-exercise program. After TOCT, walking

ability and health-related quality of life were improved

(p < 0.05) with minor retention after 3 months. The control

group showed no significant changes in any variables.

RAGT:

- Walking speed

- Endurance

Brichetto

2013

36

18/18

Wii® Balance

Board®, 4 weeks

traditional

rehabilitation

strategies

in

improving

balance

Berg Balance Scale (BBS)

BBS showed a statistically significant improvement (from

49.6 to 54.6 points, p < 0.05) in the Wii group.

Balance:

Balance/Safety

Nilsagård

2013

84

42/42

Nintendo Wii

Fit® balance

exercise programme, 6-7 weeks

no intervention

Timed Up and Go test

After the intervention, there were no statistically significant

differences between groups but effect sizes for the TUG,

TUGcognitive and, the Dynamic Gait Index (DGI) were moderate

and small for all other measures. Statistically significant

improvements within the exercise group were present for

all measures (large to moderate effect sizes) except in walking

speed and balance confidence. The non-exercise group

showed statistically significant improvements for the Four

Square Step Test and the DGI.

Balance:

Balance/Safety

Prosperini

2013

36

18/18

Nintendo

Wii Balance

Board System

(WBBS), 12 weeks

no intervention

force platform-based measures

of static standing balance,

4-step square test (FSST),

25-Foot Walking Test (25-FWT),

29-item MS Impact Scale

(MSIS-29)

The 2 groups did not differ in baseline characteristics.

Repeated-measures analyses of variance showed significant

time × treatment effects, indicating that WBBS was effective

in ameliorating force platform measures (F = 4.608, P = .016),

FSST (F = 3.745, P = .034), 25-FWT (F = 3.339, P = .048), and

MSIS-29 (F = 4.282, P = .023). Five adverse events attributable

to the WBSS training (knee or low back pain) were

recorded, but only 1 patient had to retire from the study.

Balance:

- Walking speed

- Balance/Safety

Tarakci

2013

110

55/55

group exercise

programme

under the

physical therapists‘

supervision, 12 weeks

waiting

list

Berg Balance Scale; 10-metre

walk test; 10-steps climbing

test

99 patients completed the study. There were statistically

significant improvements for all outcome measures in the

group exercise group (n = 51) (p < 0.01). In the control group

(n = 48), there were statistically significant negative change

in the Berg Balance Scale and 10-metre walk test measures

(p = 0.002, p = 0.001) and statistically significant increment

only in the Fatigue Severity Scale score (p = 0.002). The Berg

Balance Scale score was increased 4.33 in the exercise group,

while a decreased of 2.33 in control group. The 10-metre

walk test duration (second) was decreased 2.72 in exercise

group, while increased 1.44 in control group. In comparing

inter-groups changes, both primary and secondary outcome

mesures showed significant improvements in favour of the

exercise group after the training (p < 0.05).

Combined:

- Walking speed

- Balance/Safety

Learmonth

2012

32

20/12

communitybased

group

exercise intervention, 12 weeks

usual care

Timed 25-Foot Walk Test

The intervention made no statistically significant difference

to the results of participants‘ 25-foot walk time. However

the intervention led to many improvements. In the intervention

group levels of physical activity improved statistically

between baseline and week 8 (P < 0.001) and baseline and

week 12 (P = 0.005). Balance confidence results showed a significant

difference between baseline and week 12 (P = 0.013).

Good effect sizes were found for dynamic balance (d = 0.80),

leg strength (d = 1.33), activity levels (d = 1.05) and perceived

balance (d = 0.94).

Combined:

- Walking speed

Schwartz

2012

32

15/17

Robotassisted

gait training

(RAGT) with a

LokomatTM, 4weeks

conventional

walking

treatment

(CWT)

10MWT, 6MWT, TUG

Fifteen and 17 patients were randomly allocated to RAGT and

CWT, respectively. Both groups were comparable at baseline

in all parameters. As compared with baseline, although

some gait parameters improved significantly following

the treatment at each time point there was no difference

between the groups. Both FIM and EDSS scores improved

significantly post-treatment with no difference between the

groups. At 6 months, most gait and functional parameters

had returned to baseline.

RAGT:

- Walking speed

- Endurance

- Balance/Safety

Vaney

2012

67

34/33

Robot-assisted

gait training

(RAGT) with a

LokomatTM, until end of rehabilitation

walking

over

-round

end of

rehabilitation

Well-Being Visual Analogue

Scale (VAS; question 17 of the

General Well-Being Schedule),

EQ-5D European VAS value

set, Activity level with a single

axis accelerometer (ActiGraph,

Pensacola, FL), 10-m walking

speed, 3-minute walking speed

In all, 49 patients finished the interventions. Mean age was

56 years (range 3–74 years), mean EDSS was 5.8 (3.0–6.5),

and the preferred walking speed at baseline was 0.56 m/s

(0.06–1.43 m/s). Before rehabilitation, participants spent on

average 68 min/d at an MET ≥ 3. The walking group improved

gait speed nonsignificantly more than the RAGT; the upper

bound of the confidence interval (CI) did not exclude a clinically

relevant benefit (defined as a difference of 0.05 m/s)

in favor of the walking group; the lower bound of the CI did

exclude a clinically important benefit in favor of the Lokomat.

Quality of life improved in both groups, with a nonsignificant

between-group difference in favor of the walking group. Both

groups had reduced their activity by 8 weeks after the rehabilitation.

RAGT:

- Walking speed

Collett

2011

61

20/20/21

1. intermittent

Cycling, 12 weeks

2. continuous

Cycling, 12 weeks

3. 30 s cycling

followed by

30 s rest

(intermittent), 12 weeks

None

2MWT

55 participants were included in the analysis (n=continuous

20, intermittent 18, combined 17). No differences were found

between groups. After 6 weeks, considering all participants,

2 min walk distance increased by 6.96 ± 2.56 m (95% CI: 1.81

to 12.10, effect size (es): 0.25, p < 0.01). The continuous

group increased by 4.71 ± 4.24 m (95% CI: -3.80 to 13.22,

es: 0.06), intermittent by 12.94 ± 4.71 m (95% CI: 3.97 to

21.92, es: 0.28) and combined by 3.22 ± 4.60 m (95% CI:

-6.01 to 12.46, es: 0.04). Two minute walk did not significantly

change between further assessments. Between 6 and

12 weeks there was a drop in attendance that seemed to be

associated with the intermittent and combined groups; these

groups also had a greater number of adverse events (leg pain

during cycling most common) and dropouts (n=continuous 1,

intermittent 5, combined 10). Considering all participants,

6 weeks of cycling exercise produced benefits in mobility that

were maintained with further sessions.

Ergometer:

- Endurance

Hayes

2011

22

11/11

high-intensity

resistance

training

(RENEW), 12 weeks

standard

exercise

(STAND)

Lower extremity strength (hip/

knee flexion and extension,

ankle plantar and dorsiflexion,

and the sum of these individual

values (sum strength)); Timed

Up and Go (TUG), 10-m walk,

self-selected pace (TMWSS) and

maximal-pace (TMWMP), stair

ascent (S-A) and descent (S-D)

and 6-Minute Walk Test (6MWT),

Berg Balance Scale (BBS),

Fatigue Severity Scale (FSS)

No significant time effects or interactions were observed for

strength, TUG, TMWSS, TMWMP, or 6MWT. However, the mean

difference in sum strength in the RENEW group was 38.60

(representing a 15% increase) compared to the sum strength

observed in the STAND group with a mean difference of 5.58

(a 2% increase). A significant interaction was observed for

S-A, S-D, and BBS as the STAND group improved whereas the

RENEW group did not improve in these measures.

Weight training:

- Walking speed

- Endurance

- Balance/Safety

Hebert

2011

38

12/13/13

1. vestibular

Rehabilitation, 6 weeks

2. bicycle

endurance

and stretching

exercises, 6 weeks

usual

medical

care

Modified Fatigue Impact Scale

(MFIS), The Smart Balance

Master System, Six-Minute

Walk Test (6MWT)

Following intervention, the experimental group had greater

improvements in fatigue, balance, and disability due to

dizziness or disequilibrium compared with the exercise control

group and the wait-listed control group. These results

changed minimally at the 4-week follow-up. Limitations The

study was limited by the small sample size. Further investigations

are needed to determine the underlying mechanisms

associated with the changes in the outcome measures due to

the vestibular rehabilitation program.

Balance:

- Endurance
- Balance/Safety

Çakit

2010

45

15/15/15

1. cycling

progressive

resistance

training +

balance

exercises, 8 weeks

2. balance

Exercises alone, 8 weeks

no exercise

Timed Up and Go test (TUG),

Dynamic Gait Index (DGI), functional

reach (FR), Falls Efficacy

scale (FES), 10-m walking test,

Fatigue Severity scale (FSS),

Beck Depression Inventory

(BDI), Short Form 36 (SF-36)

After dropouts, the whole study group consisted of 20 women

and 13 men (mean age 37.9 ± 10.43 yrs). In training group 1,

duration of exercise, tolerated maximum workload, timed Up

and Go test, Dynamic Gait Index, functional reach, falls efficacy

scale, 10-m walk test, Fatigue Severity scale, and Beck Depression

Inventory scores, and in group 2, the mean duration of

exercise, tolerated maximum workload, and Falls Efficacy scale

scores were significantly improved after the training program

(p < 0.05). There were no significant improvements in any of

the outcome measurements in the control group (p > 0.05). In

between-group comparisons, improvements in outcome measures

of group 1 patients were significantly higher than those

in other groups, except for 10-m walking test. Group 1 patients

showed statistically significant improvement in physical functioning

and role-physical functioning scales of the Short Form

36 (p < 0.01 and p < 0.05, respectively), and group 2 patients

showed statistically significant improvement in only physical

functioning scale of Short Form 36 (p < 0.05) after 8 wks.

Weight training

- Walking speed

- Balance/Safety

Dettmers

2009

30

15/15

endurance

exercise, 3 weeks

stretching,

balance

training

and coordination

walking distance

Maximal walking distance before intervention averaged

1043 ± 568 and 1163 ± 750 m in the two groups. The intervention

group increased its maximal walking distance by

650 ± 474 m. The control group extended its walking distance

by 96 ± 70 m.

Conventional treadmill:

- Endurance

Beer

2008

35

19/16

robot-assisted

gait training

(RAGT) with

Lokomat, 3 weeks

conventional

walking

trai-ning

(CWT)

20-m timed walking

Nineteen patients were randomly allocated to RAGT and 16

patients to CWT. Groups were comparable at baseline. There

were 5 drop-outs (2 related directly to treatment) in the RAGT

group and 1 in the CWT group, leaving 14 RAGT patients and 15

CWT patients for final analysis. Acceptance and convenience of

RAGT as rated by patients were high. Effect sizes were moderate

to large, although not significant, for walking velocity (0.700,

95% CI -0.089 to 1.489), walking distance (0.401, 95% CI

-0.370 to 1.172) and knee-extensor strength (right: 1.105,

95% CI 0.278 to 1.932, left 0.650, 95% CI -0.135 to 1.436)

favouring RAGT. Prepost within-group analysis revealed an

increase of walking velocity, walking distance and knee-extensor

strength in the RAGT group, whereas in CWT group only

walking velocity was improved. In both groups outcome values

returned to baseline at follow-up after six months (n = 23).

RGAT:

- Walking speed

Cattaneo

2007

50

23/12/15

1. balance

rehabilitation

to improve

motor and

sensory

strategies, 3 weeks

2. balance

rehabilitation

to improve

motor strategy, 3 weeks

treatments

not specifically

aimed at

improving

balance

Berg Balance Scale, Dynamic

Gait Index, fall frequency, Dizziness

Handicap Inventory,

Activities-specific Balance

Confidence

Frequency of falls post treatment was statistically different

among groups (P=0.0001); The Berg Balance Scale showed an

overall statistically significant difference (P=0.0008) among

groups. Change pre-post scores were 6.7, 4.6 and 0.8 points for

groups 1, 2 and 3. Dynamic Gait Index showed an overall near

statistically significant difference among groups (P=0.14), with

change pre-post scores of 3.85, 1.6 and 1.75 points for groups

1, 2 and 3; after the exclusion of drop-outs a statistically significant

difference was observed (P=0.04). The self-administered

tests (Activities-specific Balance Confidence and Dizziness Handicap

Inventory) did not show clinically relevant improvements.

Balance:

- Balance/Safety

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Abasiyanik, 2019

Type of study: RCT

Setting and country: Turkey

Funding and conflicts of interest: This study was financed by the Northern Norway

Regional Health Authority (Project Grant 1240). The authors declare no conflict of interest.

Inclusion criteria:

definite MS diagnosis according to the revised McDonald criteria, age over 18 years, walking 100 m

independently, and willingness to participate.

Exclusion criteria:

any neurological disease other than MS, relapse within 3

months, orthopedic disorders that could negatively affect gait and balance, cardiopulmonary problems that could affect performing

exercises, diagnosed psychiatric problems or cognitive decline that render the patient incapable of performing tests and exercise training,

and current or recent (within the prior 6 months) participation

in a core stability-based exercise program.

N total at baseline:

Intervention: 21

Control: 21

Important prognostic factors:

age ± SD:

I: 42.5 (6.76)

C: 48.24 (11.79)

Sex:

I: 25% M

C: 35.5% M

Groups comparable at baseline? There were no group differences in the demographic data including age, EDSS, disease duration, and body mass index.

Describe intervention (treatment/procedure/test):

The exercise program was given one session a week for 8 weeks plus a two-day home exercise program. The duration of each session was 55-60 min.

Describe control (treatment/procedure/test):

The control group subjects were given written standardized home

exercises for 3 times per week for 8 weeks. These exercises reflect routine clinical practice and were aimed at improving spinal column, upper, and lower extremity flexibility, strength, trunk and pelvic stability,

and balance.

Length of follow-up:

Until the end of treatment.

Loss-to-follow-up:

Intervention:

5 (23.8%) due to difficulties travelling (2), going on vacation, weather conditions, and relapse.

Control:

4 (19%) due to health problems of a child, difficulties in travelling for the final assessment (2) and out of province.

Incomplete outcome data:

NA

Outcome measures and effect size (include 95%CI and p-value if available):

Difference between groups (mean with SD)

T25FW

2.95 (1.31), p=0.096

TUG

1.04 (1.31), p=0.316

6MWT

13.57 (1.31), p=0.001

MSWS-12

2.02 (1.31), p=0.165

FES-I

0.78 (1.31), p=0.385

ABC

2.64 (1.31), p=0.114

Clinical Pilates training may be preferred

as an alternative method to improve balance, walking, respiratory, and cognitive functions in pwMS.

Combined intervention.

Arntzen, 2019

Type of study: prospective RCT

Setting and country: assessor blinded RCT, Norway

Funding and conflicts of interest:

Inclusion criteria:

a diagnosis of MS in accordance with the McDonald criteria, registered at the MS outpatient clinic, living in one of the six selected municipalities, aged 18 years or older, capable of providing signed written informed consent; and, an EDSS value between 1 and 6.5 (1 = minor

disability and 6.5 = able to walk 20 m with or without a walking aid).

Exclusion criteria:

pregnancy at the time of examination, exacerbation within 2 weeks prior to enrolment,

other acute conditions resulting in compromised balance (such as

acute neurological conditions, including stroke).

N total at baseline:

Intervention: 40

Control: 40

Important prognostic factors:

age ± SD:

I: 48 (8.75)

C: 52.2 (12.9)

Sex:

I: 27.5 % M

C: 30.8 % M

Groups comparable at baseline? yes

Describe intervention (treatment/procedure/test):

Group-based, individualized, comprehensive core stability and balance intervention (GroupCoreDIST).

Describe control (treatment/procedure/test):

Standard care on walking

Length of follow-up:

30 weeks after the intervention.

Loss-to-follow-up:

Intervention: 1 (2.5%) unable to contact.

Control: 1 (2.5%) ill

Incomplete outcome data:

18 weeks:

Intervention: no missings.

Control: 3 missings; afraid to fly, family circumstances, and holiday.

30 weeks:
intervention: 2 missings; broken hip and died.

Control: 2 missings; low back pain and unable to contact.

Outcome measures and effect size (include 95%CI and p-value if available):

These results demonstrate statistically significant between-group differences in favour of GroupCoreDIST at

all follow-up time points for the 2MWT; between-group differences at 7 and 18 weeks for the 10MWT-fast and the MSWS-

12.

Mean differences (95%CI)

2MWT:

16.7 (8.15 – 25.25) SE 4.3

10MWT – fast speed

0.48 (0.11 – 0.85) SE 0.19

10MWT – preferred speed

0.20 (-0.29 – 0.69) SE 0.25

10MWT – slow speed

0.14 (-0.97 – 1.26) SE 0.56

MSWS-12

9.77 (3.19 – 16.35) SE 3.3

Compared with standard care, GroupCoreDIST significantly improved walking immediately after the intervention for up to 24 weeks of follow-up.

Balance intervention.

Callesen, 2019

Type of study: cluster RCT

Setting and country: multicenter, single-blind, Denmark

Funding and conflicts of interest: The work was supported by the

Danish foundation TrygFonden.

J.C., D.C., J.B., and M.-L.K.J. declare that they have no conflict

of interests associated with this paper. U.D. has received research support, travel grants, and/or teaching

honorary from Biogen Idec, Merck Serono,

Novartis, Bayer Schering, and Sanofi Aventis as well as honoraria from serving on scientific advisory

boards of Biogen Idec and Genzyme.

Inclusion criteria:

>age 18, confirmed diagnosis of MS, Expanded

Disability Status Scale (EDSS): 2.0–6.5, Six Spot > Step Test (SSST) score

> 8 seconds or Timed 25-Foot Walk (T25FW) > 5 seconds, relapse-free within the past 8 weeks, and no adjustment of disease-modifying

medication or medication that affects gait performance

and spasticity within the past 8 weeks.

Exclusion criteria:

Exclusion criteria were as follows: co-morbidity in

terms of cognitive disorders or alcohol abuse (based on clinical judgment), pathologies that did not allow systematic PRT

> 1 session/week within the last 3

months. Eligibility according to criteria that concerned

co-morbidity, disease activity, medication, and

EDSS score was provided by neurologists based on journal records.

N total at baseline:

Intervention 1: 23

Intervention 2: 28

Control: 20

Important prognostic factors:

age ± SD:

I1: 52 (38 – 64)

I2: 51 (31-75)

C: 56 (30 – 73)

Sex:

I1: 70% F

I2: 82% F

C: 80% F

Groups comparable at baseline? At baseline, no difference between the groups was found.

Describe intervention (treatment/procedure/test):

20 1-hour sessions over 10 weeks (2 sessions/week). All sessions started with a 10-minute warm-up.

Interventie 1: progressive resistance training (PRT)

The program predominantly targeted knee and hip flexion and extension where the exercises progressed from three sets of 10 repetitions at 15 RM

toward four sets of 8 repetitions at 8 RM. The exercises were conducted in machines that targeted

the specified muscle groups, but type of machines could vary between centers. All training sessions were supervised by physiotherapists who

were trained to deliver the intervention.

Interventie 2: balance and motor control training (BMCT): addressing salient tasks including sitting (5 minutes),

standing (5 minutes), stepping (10 minutes), walking

×

(2

10 minutes), and eye-movement training (10

minutes; framework illustrated in Supplemental

Appendix). To ensure that exercises were sufficiently

challenging, the relative complexity level of an exercise was maintained by variation and by progression

obtained by alteration of geometry of the base of sup

port, by changing movement speed, by adding sensory conditions to promote better use of proprioceptive and visuo-vestibular information, and by addition of segmental movement. Furthermore, as a means for progression, and to promote cognitive load related to divided attention, cognitive multitask challenges were

added to some of the exercises.

Describe control (treatment/procedure/test):

Waitlist controls.

Participants in the control waited for 10 weeks,

where they were encouraged to maintain usual care and level of physical activity. Thereafter, they received an intervention with one weekly session of BMCT

and one weekly session of PRT.

Length of follow-up:

The follow-up test was performed within 7 days after the last intervention or after the last day of being on the waitlist.

Loss-to-follow-up:

In total, 12 participants (17%) were lost to follow-up,

with the highest dropout rate found in the PRT group.

I1: 6 (1 unknown reason, 1 illness, 1 intense fatigue, 1 fall unrelated to training).

I2: 4 (3 unknown reason, 1 logistic issues).

C: 2 (unknown reason + started training).

Incomplete outcome data:

Intervention:

N (%)

Reasons (describe)

Control:

N (%)

Reasons (describe)

Outcome measures and effect size (include 95%CI and p-value if available):

Adverse events: Two adverse events occurred in the PRT group. One participant experienced exacerbation of fatigue following PRT sessions and dropped out; one experienced intermittent low back pain during PRT, but continued in the study.

T25FW: BMCT significantly improved the T25FW (0.10 m/s, p = 0.04) and SSST (–2.2 seconds, p< 0.01), while no improvement was found for

PRT for neither T25FW (0.2 m/s, p =0.66) nor SSST (–0.5 seconds, p=0.52).

Straudi, 2019

Type of study:

RCT

Setting and country: Italy

Funding and conflicts of interest: this trial was supported by a grant from the 2010–2012 Research Programme of

Emilia Romagna Region (grant number 1786/2012).

The funding source had no role in the design of the study and did not have any role in its execution, analyses

and interpretation of the data. The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or

publication of this article.

Inclusion criteria:

men and women aged between 18 and 65 years; primary or secondary progressive MS and severe gait

impairments, defined by an expanded disability status scale (EDSS) score ranging from 6 to 7.

Exclusion criteria:

inability to perform the T25FW; MS

worsening in the 3 months before intervention;

impaired

cognitive function

(mini-mental state

< examination score 24); modified Ashworth scale

> score 3; clinical conditions in addition to MS that

may interfere with the ability to complete the study

protocol safely; changes in drug therapy during the

study; rehabilitation treatment or botulinum toxin

injections in the 3 months preceding the enrolment.

N total at baseline:

Intervention: 36

Control: 36

Important prognostic factors:

age ± SD:

I: 56 (11)

C: 55 (11)

Sex:

I: 33% M

C: 31% M

Groups comparable at baseline? The two study groups at baseline did not differ in terms

of anthropometrics, disease duration, severity, drug

therapy or any of the outcome measures (Table 1).

Describe intervention (treatment/procedure/test):

Robot-assisted gait training

Patients randomized to this group per formed RAGT on a Lokomat treadmill lasting around 40 minutes (Hocoma, Volketswil, Switzerland). Considering

the time to set up the device and a brief familiarization period, the real walking time was about

30 minutes. Training parameters, including gait speed, body weight support, guidance force and torque of the knee and hip drives, were individually set for each patient during the first training session by the same

expert physiotherapist (starting with 100% guidance and 50% body weight support). As the training progressed,

adjustments (10% each) in these parameters were done according to the patient's performance.

Describe control (treatment/procedure/test):

Conventional training

Patients included in the CT group underwent

individual physiotherapy sessions, composed of

assisted overground walking for a total of approximately

40 minutes, inserted between 10 minutes warm

Up and cool-down periods. The patients continuously

walked on an 80-m indoor flat corridor with their habitual walking device. Resting pauses were allowed as necessary. Gait speed was set on patients’ tolerance

Length of follow-up:

3 months after completion of the sessions.

Loss-to-follow-up:

Six patients (RAGT n = 2; CT n = 4) did not start the allocated treatment because of worsening MS (n = 1),

acute intercurrent disease (n = 2) or withdrawal of

consent to participate (n = 3).

CT: Two patients discontinued the intervention and were lost to follow-up because of intercurrent disease requiring hospitalization.

Incomplete outcome data:

Intervention: 5 patients missed 1 out of the 12 RAGT sessions because

of work or transportation issues.

Control: Two patients discontinued the intervention and were lost to follow-up because of intercurrent disease requiring hospitalization. Four patients missed one training session, and three patients missed two sessions

because of work or transportation issues.

Outcome measures and effect size (include 95%CI and p-value if available):

At the end of treatment with respect to baseline, both

groups significantly improved gait speed (p<0.001) and most secondary outcomes without between group

 differences. Outcome values returned to baseline at follow-up.

Young, 2019

Type of study: RCT

Setting and country: community-based fitness facility, USA

Funding and conflicts of interest: Supported by the National Institute on Disability, Independent Living, and Rehabilitation

Research (NIDILRR grant no. 90DP0059-01-00). NIDILRR is a center within the Administration

for Community Living (ACL), Department of Health and Human Services (HHS). The contents of

this article do not necessarily represent the policy of NIDILRR, ACL, or HHS, and you should not

assume endorsement by the Federal Government.

No conflicts of interests are disclosed.

Inclusion criteria:

Interested participants were enrolled if they self

Reported a diagnosis of MS, with a Patient Determined Disease Steps (PDDS) score between 0 and 6. Additional inclusion criteria included ages of 18-65 years, ability to exercise with arms and/or legs, and physician

clearance.

Exclusion criteria:

Exclusion criteria included participation in a similar

intervention in the last 6 months, use of tobacco products in the last 6 months, unstable weight, cognitive impairment (Mini-

Mental State Exam score <24), active pressure ulcer, and any contra-indications to exercise based on the American College of

Sports Medicine guidelines.

N total at baseline:

Intervention: 26

Control: 28

Important prognostic factors:

age ± SD:

I: 48 (10)

C: 47 (10)

Sex:

I: 23 % M

C: 14 % M

Groups comparable at baseline? There were no statistical differences between groups with respect

to all baseline characteristics.

Describe intervention (treatment/procedure/test):

Yoga

The AY intervention was taught by yoga instructors who were YogaFit level 1 certified and had experience adapting yoga for people with disabilities.

Participants received a yoga session of 60 minutes for 3 times a week.

Describe control (treatment/procedure/test):

Waitlist (no intervention)

WC participants received biweekly newsletters that provided educational information on living with MS.

Length of follow-up:

12-weeks (during the intervention program)

Loss-to-follow-up:

Intervention:

5 (19%)

Reasons (describe)

Control: 5

5 (18%)

Reasons (describe)

Incomplete outcome data:

Intervention:

1 (4%)

Reasons (describe)

Control:

4 (14%)

Reasons (describe)

Outcome measures and effect size (include 95%CI and p-value if available):

TUG: No significant TUG

improvement was found in AY compared to WC.

6MWT: No significant TUG

improvement was found in AY compared to WC.

FTSST (five times sit-to-stand tes): No significant TUG

improvement was found in AY compared to WC.

AE related to study: none

AE not related to the study: 1 stroke in the yoga group.

Self-reported MS diagnosis.

Cattaneo, 2018

Type of study: RCT

Setting and country: bi-center randomised rater-blinded controlled trial, Italy and Czech Republic.

Funding and conflicts of interest:

Inclusion criteria:

ability to walk (also with aid) for 6 m and ability to maintain standing position with open eyes for at least 30 s.

Exclusion criteria:

Ability to maintain monopodalic-standing position for 10 s, ability to maintain tandem position for 30 s, cognitive disorders hampering the execution of the exercises/assessment

N total at baseline:

Intervention: 78

Control: 41

Important prognostic factors:

age ± SD:

I: 48.9 (11.1)

C: 46.7 (11.4)

Sex:

I: 31% M

C: 29% M

Groups comparable at baseline?

The baseline

characteristics of the participants were broadly similar between the two groups

Describe intervention (treatment/procedure/test):

Participants in both groups received 20 treatments sessions lasting 45 min 2/3 times per week by experienced physical therapists trained for the study

Participants in the intervention group

received at least 25–45 min of balance treatment.

Describe control (treatment/procedure/test):

Participants in the control group were treated to reduce limitations at body function and activity levels,

while treatment for balance disorders was restricted to a maximum

of 10 min per session.

Length of follow-up:

Two months after the treatment

Loss-to-follow-up:

1 – post treatment

Intervention: 9 (11.5%)

Control: 5 (12.2%)

2 – follow-up

Intervention: 11 (15.9%)

Control: 10 (27.7%)

Among these 35 dropouts 10

withdrew from the study because of new co-morbidities. Non-medical

reasons for withdrawal were lack of motivation (3), long

holiday (2), other/unknown reasons (20).

Incomplete outcome data:

NA

Outcome measures and effect size (include 95%CI and p-value if available):

Falls

1- post intervention

Intervention: 16 (23%)

Control: 8 (22%)

OR: 1.05 (0.41-2.77)

2 – follow-up

Intervention: 18 (31%)

Control: 1 (4%)

OR: 11.3 (1.41 – 89.58)

BBS (mean, SE)

1- post intervention

Intervention: 49.2 (0.6)

Control: 48.9 (0.8)

Difference (95%CI): -0.4 (-2.4 – 1.6)

2 – follow-up

Intervention: 47.8 (0.6)

Control: 47.8 (0.9)

Difference (95%CI): 0.0 (-2.2 – 2.2)

DGI (mean, SE)

1- post intervention

Intervention: 17.6 (0.4)

Control: 18.1 (0.6)

Difference (95%CI): 0.6 (-0.9 – 2.0)

2 – follow-up

Intervention: 16.9 (0.5)

Control: 17.0 (0.8)

Difference (95%CI): 0.1 (-1.8 – 2.1)

TUG (mean, SE)

1- post intervention

Intervention: 12.4 (0.65)

Control: 12.5 (0.65)

Difference (95%CI): 0.1 (-1.5 – 1.7)

2 – follow-up

Intervention: 16.9 (0.5)

Control: 17.0 (0.8)

Difference (95%CI): 0.1 (-1.8 – 2.1)

ABC (mean, SE)

1- post intervention

Intervention: 59.1 (1.6)

Control: 62.9 (2.1)

Difference (95%CI): 3.8 (-1.6 – 9.2)

2 – follow-up

Intervention: 55.2 (3.4)

Control: 66.2 (5.3)

Difference (95%CI): 10.8 (-2.5 – 19.1)

Twenty sessions 2–3 times/week of balance specific rehabilitation did not reduce fall frequency

nor improve balance suggesting the need for more frequent and challenging interventions.

Duff, 2018

Type of study: RCT

Setting and country: single blinded, Canada

Funding and conflicts of interest: his study was funded by a Hermes Canada

MS Society of Canada Wellness Research Innovation grant. The other authors have no conflicts of

interest to disclose.

Inclusion criteria:

The inclusion criteria were purposely kept broad and

consisted of a definite diagnosis of MS, not restricted to

a wheelchair or scooter, and the ability to travel to the assessment (University of Saskatchewan) and intervention

(Lead Pilates and Integrative Therapies) sites, both located in Saskatoon, Saskatchewan, Canada.

Exclusion criteria:

Not reported.

N total at baseline:

Intervention: 15

Control: 15

Important prognostic factors:

age ± SD:

I: 45.7 (9.4)

C: 45.1 (7.4)

Sex:

I: 20% M

C: 27% M

Groups comparable at baseline? Yes

Describe intervention (treatment/procedure/test):

The Pilates intervention consisted of two 50-minute

Pilates sessions per week. The Pilates sessions were conducted in a group setting (n = 5-10) but were individualized to each participant’s ability level and focused on proper

breathing, correct body alignment, and core stabilization.

Participants in the Pilates group also received a

weekly 1-hour massage therapy session with a registered massage therapist specially trained in massage for MS.

Describe control (treatment/procedure/test):

Participants in the control group received only the

once-weekly 1-hour massage therapy session.

Length of follow-up:

Until 1 week after the intervention.

Loss-to-follow-up:

Intervention:

1 (6,66%) discontinued the intervention due to medical reasons.

Control:

2 (13,3%) discontinued the intervention due to medical reasons.

Incomplete outcome data:

NA

Outcome measures and effect size (include 95%CI and p-value if available):

Capacity

6MWT:

Intervention: 52.4m (32.7-72.1)

Control: 15.0m (-4.7-34.7)

p-value: 0.01.

TUG left turn:

Intervention: -1.5 (-2.7— -0.4)

Control: 0.3 (-0.7-1.4)

P=0.03

TUG right turn:

Intervention: -1.1 (-2.1 - -0.1)

Control: 0.3 (-0.7 – 1.4

P=0.06

Adverse events

No adverse events were reported in either group during

the intervention.

Pilates improved walking performance and functional ability in persons with MS and is a viable exercise option to help manage the disease.

Jonsdottir, 2018

Type of study:

RCT

Setting and country:

Pilot, Italy

Funding and conflicts of interest: Funding fromthe Ministero della Salute, Italia was given to JJ and

funding from Fondazone Italiana Sclerosi Multipla to ET (Grant # 2013/R/26).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Inclusion criteria:

diagnosis of

multiple sclerosis according to McDonald’s criteria (16), EDSS score

7 (free from relapses and steroid treatment for at

least 3 months), able to stand 30 s, able to walk at least 10

meters independently or with a walking aid, aged between 18

and 80 years, able to understand and follow instructions, stable

neurological conditions, and willingness to participate in the

study.

Exclusion criteria:

cardiac pacemaker,

any heart condition that their medical doctor considered risky for intense aerobic activity, any pre-existing conditions that

affected walking function, diagnosis of depression or psychotic disorder.

N total at baseline:

Intervention: 28

Control: 14

Important prognostic factors:

age ± SD:

I: 51.4 (10.7)

C: 56.7 (5.7)

Sex:

I: 9/26 % M

C:1/12 % M

Groups comparable at baseline? Yes.

Describe intervention (treatment/procedure/test):

Participants in both groups received 15–20 treatments sessions lasting 30min 5 times per week by experienced physical

therapists trained for the study.

Participants in the experimental group received supervised

treadmill training, 4–5 sessions per week. The treatment protocol was aimed at improving participants’ resistance, walking velocity, balance and cognitive functions during motor and cognitive

(dual) tasks. The treadmill training was carried out without body weight support but if needed the participants were

attached to a safety harness.

Describe control (treatment/procedure/test):

Participants in the Strength training group received supervised muscle resistance training, 4–5 sessions per week. The exercises were specifically aimed at improving strength in muscles involved

in walking (hip abductors, quadriceps, plantar flexors, dorsal flexors) and were progressed according to current guidelines from the American College of Sports Medicine (25).

Length of follow-up:

End of treatment.

Loss-to-follow-up:

None reported.

Incomplete outcome data:

Intervention: 2 (7%)

Control: 2 (14%)

Excluded from analysis due to early dismissal from hospital.

Outcome measures and effect size (include 95%CI and p-value if available):

Adverse events

There were no adverse events reported during or

following either treatment except muscle and general fatigue that was resolved in a couple of hours after the session.

2MWT (>15% clinically relevant)

Intervention: 19 (73%)

Control: 2 (17%)

P = 0,001

Difference in proportions 56% (22.4% - 74.2%).

Between group differences (mean + 95%CI)

2MWT: 28.3 (13.04 – 43.60)

Gait speed: 0.2 (0.04 – 0.30)

TUG: -2.83 (-0.9 - -4.7)

BBS: 1.1 (-1.4 – 3.7)

DGI: 0.2 (-1.95 – 2.27)

Tramontano, 2018

Type of study: preliminary RCT

Setting and country:

Hospitalized setting, single blind, Italy

Funding and conflicts of interest: The authors report no financial support. The authors declare no competing financial interests.

Inclusion criteria:

clinically definite MS, age > 20 and <65 years, disability between 5 and 7 on the Expanded Disability Status Scale (EDSS), walking

ability according to the EDSS score and a minimal score

leg spasticity

=1 on the Modified Ashworth spasticity scale (score range: 0–4; 0 indicates no spasticity and 4 indicates leg rigidity in all directions; 1

indicates a slight increase in muscle tone).

Exclusion criteria:

the presence of neurological,

orthopaedic and severe cardiac comorbidities

and peripheral vestibular disorders (i.e., positional

paroxysmal vertigo, Meniere’s disease and acoustic

neuroma), legal blindness in one or both eyes documented

MS-related exacerbation in the past three

months and being involved in other research studies

N total at baseline:

Intervention: 15

Control: 15

Important prognostic factors:

age ± SD:

I: 50.64 (11.73)

C: 45.77 (10.91)

Sex:

I: 6/15 M

C: 7/15 M

Groups comparable at baseline? Yes

Describe intervention (treatment/procedure/test):

Both groups performed two daily 40-minute sessions

5x/wk for 4 weeks of conventional neurorehabilitation

therapy for MS.

VR consisted of two types of exercises, i.e., those for gaze stability and those for postural stability in a

standing position on a foam cushion.

Describe control (treatment/procedure/test):

Standard neurorehabilitation consisted of muscle

stretching, postural alignment especially at the height of the axial segments, active-assisted mobilizations and neuromuscular facilitations to improve motor

recruitment.

Length of follow-up:

Until the end of treatment.

Loss-to-follow-up:

Intervention:

2 (13,3%); released before the end of the intervention, referred a side effect.

Control:

5 (33,3%): 4 released before the end of the intervention, 1 refused to continue for reasons not related to the trial.

Incomplete outcome data:

Not reported.

Outcome measures and effect size (include 95%CI and p-value if available):

Between group:

BBS: p=0.0001

% Increase:

16.8 (4.4) intervention

2.5 (1.5) control

2MWT: p=0.201

T25FW: p=0.037

% Increase:

15.1 (4.9) intervention

4.7 (1.3) control

Study

First author, year

Appropriate and clearly focused question?

Yes/no/unclear

Comprehensive and systematic literature search?

Yes/no/unclear

Description of included and excluded studies?

Yes/no/unclear

Description of relevant characteristics of included studies?

Yes/no/unclear

Appropriate adjustment for potential confounders in observational studies?

Yes/no/unclear/notapplicable

Assessment of scientific quality of included studies?

Yes/no/unclear

Enough similarities between studies to make combining them reasonable?

Yes/no/unclear

Potential risk of publication bias taken into account?

Yes/no/unclear

Potential conflicts of interest reported?

Yes/no/unclear

Tholen, 2019

Yes

Yes

Unclear.

The reasons for exclusion are mentioned, however references of the excluded studies were not included.

Yes

Not applicable

Yes, PEDro-scale

Not applicable, data was not pooled.

No

No

Study reference

(first author, publication year)

Describe method of randomisation

Bias due to inadequate concealment of allocation?

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?

(unlikely/likely/unclear)

Bias due to loss to follow-up?

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?

(unlikely/likely/unclear)

Abasiyanik, 2019

Those who were eligible and agreed to participate in this study

were allocated into two groups with stratified randomization according to gender and EDSS. The first woman participating in the study was enrolled in the Clinical Pilates group and the second was included in the home exercise program group. In the same way, the first male was included in the Clinical Pilates group and the second was enrolled in the home exercise program group. Each group was then randomized again according to the subjects’ EDSS scores.

Unlikely

Likely, participants were not blinded.

Likely, care provides were not blinded.

Likely, outcome assessors were not blinded.

Unlikely

Unlikely, the loss to follow-up was similar.

Unlikely

Arntzen, 2019

Random allocation by electronic concealed randomization using a web-based system.

Unlikely

Likely, participants were not blinded.

Likely, care provides were not blinded.

Unlikely, outcome assessors were blinded.

Unlikely

Unlikely, the loss to follow-up was similar.

Unlikely

Callesen, 2019

Allocation of

interventions was done by a cluster randomization

conducted whenever three groups could be established

at three different participating clinics.

Unlikely.

Likely, participants were not blinded.

Likely, care providers were not blinded.

Unlikely, A research assistant (L.A.) assigned patients with an ID number and pseudonyms, which were used during baseline and follow-up testing to ensure blinding of the two experienced assessors (primary investigator (J.C.) and assistant

(M.-L.K.J.)).

Unlikely

Unlikely, the loss to follow-up was similar.

Unlikely

Strada, 2019

Participants were allocated 1:1 to RAGT or CT using computer-generated allocation sequence managed by

an external administrator. The randomization was per

formed in blocks of four patients, stratified by EDSS.

Unlikely.

Likely, participants were not blinded.

Likely, care provides were not blinded.

Unlikely.

Researchers blinded to the patients’ allocation groups performed outcome measures and statistical analyses.

Unlikely

Likely, three patients withdrew their consent to participate. Unclear which group.

Unlikely

Young, 2019

Eligible participants were randomly assigned using computer-generated randomization by a statistician who was not involved with the interventions.

Unlikely.

Likely, participants were not blinded.

Likely, care providers were not blinded.

Unlikely, the outcome measures were administered by blinded assessors.

Unlikely

Unlikely, the loss to follow-up was similar.

Unlikely

Cattaneo, 2018

In each clinical centre, an independent clinician allocated participants to either the control or intervention group according to a randomisation

list made before the beginning of the study.

Unlikely

Likely, participants were not blinded.

Likely, care providers were not blinded.

Unlikely, evaluation was done by a blinded assessor in each center and group allocation was kept concealed from them throughout the study.

Unlikely

Unlikely, loss to follow up was higher in the control group.

The participants who dropped out before follow-up did not show

statistically significant differences from the rest of the sample

except for number of frequent fallers.

Unlikely

Duff, 2018

Randomization

was performed using a computer-generated allocation schedule, with a block size of two, by a researcher

(C.D.E.) who was not involved in baseline or outcome

assessment or in the statistical analysis.

Unlikely

Likely, participants were not blinded.

Likely, care providers were not blinded.

Unlikely, researchers involved in outcome assessment and analysis were blinded to group assignment. With statistical analysis blinded through coding of the study groups.

Unlikely

Unlikely, the loss to follow-up was similar.

Unlikely

Jonsdottir, 2019

Participants were consecutively randomized to the experimental

(DT-group) or control group (S-Group) in a 2:1 ratio, using a computerized automated algorithm. To ensure a concealed allocation randomization was done after determining

eligibility.

Unlikely.

Likely, participants were not blinded.

Unlikely, care providers were blinded.

Unlikely, outcome assessors were blinded to group assignment.

Unlikely

Unlikely, rates loss to follow-up were similar.

Unlikely

Tramontano, 2018

Not described.

Unclear.

Likely, participants were not blinded.

Likely, care providers were not blinded.

Unlikely, outcome assessors were blinded.

Unlikely

Likely, high rates of loss to follow-up.

Likely