Sensory retraining of the leg after stroke: systematic review and meta-analysis.

The review included 16 studies with 430 participants, and none of the studies received a strong quality rating. Ten studies were weak and six were moderate, with incomplete reporting of recruitment, randomisation, blinding, and confounders in several trials. The positive pooled results for sensation and balance were heterogeneous. The studies did not agree well with one another. Subgroup analyses for joint position sense and light touch were not significant, so the positive primary result is not stable across sensation types. Motor impairment and leg function could not be pooled because of clinical heterogeneity and insufficient data. The interventions varied widely, including sensory retraining, proprioceptive training, vibration, TENS, and gait training, so the review cannot show which component worked. Some studies combined sensory retraining with other rehabilitation, so the effect of sensory retraining alone cannot be separated. The participants were from seven days to nearly 16 years after stroke, so findings may not apply equally to all stages. Gait symmetry and other gait variables were not analysed, so a null result for gait speed does not prove no effect on walking. Many somatosensory measures had unestablished or unreported psychometric properties, and some measures may have been insensitive to change.

Do not conclude that any single leg sensory retraining method improves walking speed or independence. The review combined different interventions, pooled gait speed was not significant, motor impairment and leg function could not be pooled, and the positive sensation and balance results came from small, heterogeneous, low-quality studies.