Repetitive gait training early after stroke
83
1. Categorization 2. Results
Activity level: The measurements assessing the abi-
lity to walk are classified under the activity domain
“walking” (ICF d450).
• Walking independence: Independence, the primary
outcome, is either measured with the FAC, a 5-item
scale measuring the degree of assistance required to
walk, or by dichotomized outcome where the number
of patients achieving independence (e.g. FAC ≥ 4)
is scored.
• Walking speed: The time is measured while partici-
pants walk over a 5- or 10-m distance at a comfor-
table pace to calculate walking speed.
• Walking endurance: Endurance is assessed by asking
the participant to walk the greatest possible distance
during a period of 6 min. Activity level:
• Walking independence: Post-intervention FAC scores
were reported in 10 RCTs. Pooling yielded a nonsig-
nificant heterogeneous SES (10 RCTs; n = 671 [exp
338; ctr 333]; MD = 0.38 [random]; 95% CI –0.03
to 0.78; p = 0.07; I 2 = 78%). There is a significant
subgroup difference (p = 0.002) and the sub-analysis
revealed end-effector training to be effective only (5
RCTs, n = 385 [exp 187; ctr 195]; MD = 0.73 [ran-
dom]; 95% CI 0.17–1.30; p = 0.01; I 2 = 75%).
• At follow-up (≥ 3 months), pooling of 7 RCTs re-
sulted in a significant heterogeneous SES (7 RCTs;
n = 538 [exp 266; ctr 272]; MD = 0.57 [random]; 95%
CI 0.14–1.01; p = 0.01; I 2 = 66%). If pooling outcome
of end-effector studies only, a heterogeneous SES
is identified (5 RCTs; n = 381 [exp 186; ctr 195];
MD = 0.72 [random]; 95% CI 0.16–1.28; p = 0.007;
I 2 = 69%).
• In addition, dichotomous outcome was pooled to
calculate an odds ratio. Follow-up data was entered
if provided and otherwise post-intervention data
was used. This yielded significant heterogeneous
results (8 RCTs; n = 627 [exp 312; ctr 315]; OR = 1.99
[random]; 95% CI 1.13–3.53; p = 0.02; I 2 = 60%).
If pooling end-effector studies only, a greater, but
non-significant, SES is identified (5 RCTs; n = 382
[exp 187; ctr 195]; OR = 2.15 [random]; 95% CI
0.88–5.28; p = 0.1; I 2 = 75%). Taking the inconsis-
Body function level:
• Motor control: The motor subscale for the lower
extremity of the Fugl Meyer Assessment (FM-L) is
classified under the domain “control of voluntary
movements functions” (ICF b760).
• Muscle strength: The (Medical Research Council)
Motricity Index subscale for the lower limb (MI-L)
measures the strength of major leg muscle groups and
is classified as “muscles power functions” (ICF b730).
The results of the meta-analysis for each outcome, as
defined above (see Table IV), are described below. Fo-
rest plots are derived from RevMan (see Figs S1–S7 1 ).
Table IV. Results of the quantitative analysis on the comparative effectiveness of repetitive gait training. This includes gait-specific
outcome on both body function and activity level. Sub-analyses based on the intervention type are performed for each comparison and
results are analysed if at least 3 RCTs could be included
Activity
Body function
Walking independence (FAC)
Post-intervention
(FAC)
Follow-up (FAC)
MD = 0.57
[0.14, 1.01];
p = 0.01;
n = 266/272
Repetitive gait training MD = 0.38
[–0.03, 0.78];
p = 0.07;
n = 338/333 a
RAGT Exo MD = –0.27
?
[–0.57, 0.03];
p = 0.08; n = 63/56
RAGT EE MD = 0.73
[0.17, 1.30];
p = 0.01;
n = 187/195
?
BWSTT
MD = 0.72
[0.16, 1.28];
p = 0.01;
n = 186/195
?
Walking speed
Walking endurance
(5/10 mWT, m/s) (6 minWT, m)
Motor control Muscle strength
(FM-L)
(MI-L)
OR = 1.99
[1.13, 3.53];
p = 0.02; n = 312
(65%)/315
(50%)
? MD = 0.05
[–0.00, 0.11];
p = 0.06;
n = 342/330 MD = 24.36
[3.58, 45.14];
p = 0.02;
n = 206/200 MD = 0.52
[–1.54, 2.59];
p = 0.62;
n = 91/88
? ? OR = 2.15
[0.88, 3.53];
p = 0.1; n = 187
(61%)/195 (43%)
? MD = 0.08
[0.01, 0.15];
p = 0.03;
n = 171/176
MD = 0.00
[–0.10, 0.10];
p = 0.99;
n = 122/111 MD = 32.08
[8.30, 55.86];
p = 0.008;
n = 154/155
? MD = 0.76
?
[–1.83, 3.36];
p = 0.56;
n = 63/56
?
MD = 8.00
[2.08, 13.93];
p = 0.008;
n = 113/117
?
?
MD = 3.64
[–2.88, 10.57];
p = 0.27;
n = 185/179
Effect sizes in bold present statistically significant results.
a
a significant (p < 0.05) subgroup differences was identified.
RCT: randomized controlled trial; RAGT: robot-assisted gait training; Exo: exoskeleton; EE: end-effector; BWSTT: body-weight supported treadmill training; FAC:
Functional Ambulatory Categories; 5/10mWT: 5 or 10 m Walk Test, 6minWT: 6-min walk test; FM-L: Fugl-Meyer Assessment motor subscale for the lower limb;
MI-L: Motricity Index subscale for the lower limb; MD: mean difference; OR: odds ratio; ?: unknown effect due to lack of data (< 3 RCTs).
J Rehabil Med 51, 2019