Several studies highlight that the physical therapy (PT) timing is important in rehabilitation process from physical disability after stroke. The purpose of this meta-analysis was to investigate the effect of all PT on balance after stroke and to evaluate whether changes are related to stroke stage of recovering [acute and subacute (ASS) or chronic stage (CS)].

Randomised controlled trials (RCT), without language restriction, until October 2015, assessing the Berg Balance Scale (BBS), the Postural Assessment Scale for Stroke, the posturographic parameters in static condition and measures of independence in activities of daily living were included. Two independent authors (AH and JDM) led this selection following the Cochrane recommendations (Hugues et al., 2017).

One hundred and ninety-three parallel and crossover RCT from 9337 records were included, involving 8018 subjects [study sample: mean 41.5 (SD 49.7) subjects, age: pooled mean 61.7 (pooled SD 12.78) years]. A significant difference in favor of PT compared to no treatment was immediately found after the intervention for the BBS at ASS and CS (ASS: 11 studies, mean difference (MD) 3.14; 95% confidence interval (CI) [2.21; 4.06]. CS: 12 studies, MD 1.31; 95% CI [0.79; 1.82]. Mixed: 2 studies, MD 0.89; 95% CI [−0.63; 2.40]). Only at CS, PT is more effective than sham treatment or usual care immediately after the intervention for the BBS (ASS: 12 studies, MD 2.21; 95% CI [−0.37; 4.78]. CS: 13 studies, MD 2.50; 95% CI [1.00; 3.99]. Mixed: 2 studies, MD −1.17; 95% CI [−2.94; 0.60]).

These results show immediately after the intervention:

–that balance can be improved by PT compared to no treatment whatever the stage of stroke;

–an improvement of balance by PT compared to sham treatment or usual care at chronic stage.