Clinical outcomes of rehabilitation with a robotic anthropomorphic exoskeleton in patients with motor-incomplete spinal cord injury: a multicenter randomized controlled trial.

Small trial: 36 people enrolled instead of the 40 planned. The authors state the effect size they assumed for the sample size calculation (90% responders with the exoskeleton versus 50% without) was optimistic, so the study may have been underpowered to detect smaller but real differences. Everyone in the study could already walk, and most used aids or an orthosis. The authors suggest a ceiling effect may have limited any extra benefit that could be shown. The exoskeleton block was short and low dose (four weeks, 30 to 40 minutes, three times a week); the authors say this may have limited carry-over to unsupported walking. All participants were 1 to 5 years after injury, so the results say nothing about complete injuries or the acute and subacute phases, and they may not transfer to other exoskeleton devices. Participants and treating therapists could not be blinded, although the people measuring walking outcomes were.

Funded by the Italian Ministry of Health through a peer-reviewed grant (project code GR-2018-12367485). The authors declare no conflict of interest with any financial organisation regarding the material in the manuscript. The exoskeleton's manufacturer and its Italian distributor are named in the methods section, but no role for them in the study is stated.

Do not conclude from this that exoskeleton gait training does not work. The trial enrolled 36 of the 40 patients it needed, added only a short four-week block on top of an already intensive programme, and every participant could already walk. It cannot say anything about people with complete injuries, people early after injury, or longer or more intensive exoskeleton programmes. The safe reading is that this dose added nothing beyond conventional rehabilitation in this particular group.