An 8-week aerobic cycling intervention elicits improved gait velocity and biomechanics in persons with Parkinson's disease.

Small numbers: 28 people in total, 14 per group, and they were a subset of a larger trial, so this is a secondary analysis rather than the trial's main question. Everyone had mild to moderate Parkinson's disease. The authors say they do not know whether people with advanced disease (Hoehn and Yahr stage IV) could complete the programme, or whether they would gain walking speed if they did. Walking speed fell by 0.09 m/s in the control group over the 8 weeks, and the authors say they cannot tell whether that was disease progression or something else they did not measure, so part of the gap between the groups may not come from the exercise. The improvements in joint angles and in the forces at the ground were described in the cycling group rather than confirmed by significant statistical tests. Gait was recorded with participants off their Parkinson's medication, in a motion capture laboratory, over a 60-second walk.

The trial was registered (NCT01636297) and reports N.I.H. extramural research support under grant R01NS673717. One author, Jay Alberts, holds intellectual property on the algorithm used to control the forced-exercise bicycle; the other authors declared no conflicts of interest. The paper does not say whether the funder had any role in the design, analysis or writing.

Do not read the 0.14 m/s gain as the size of the training effect on its own. The control group lost 0.09 m/s over the same eight weeks, and the authors say they cannot tell whether that decline was disease progression or something else they did not measure.