Background: Falls are a frequent and serious complication of Parkinson's disease and are related partly to an underlying cholinergic deficit that contributes to gait and cognitive dysfunction in these patients. Gait dysfunction can lead to an increased variability of gait from one step to another, raising the likelihood of falls. In the ReSPonD trial we aimed to assess whether ameliorating this cholinergic deficit with the acetylcholinesterase inhibitor rivastigmine would reduce gait variability.

Methods: We did this randomised, double-blind, placebo-controlled, phase 2 trial at the North Bristol NHS Trust Hospital, Bristol, UK, in patients with Parkinson's disease recruited from community and hospital settings in the UK. We included patients who had fallen at least once in the year before enrolment, were able to walk 18 m without an aid, had no previous exposure to an acetylcholinesterase inhibitor, and did not have dementia. Our clinical trials unit randomly assigned (1:1) patients to oral rivastigmine or placebo capsules (both taken twice a day) using a computer-generated randomisation sequence and web-based allocation. Rivastigmine was uptitrated from 3 mg per day to the target dose of 12 mg per day over 12 weeks. Both the trial team and patients were masked to treatment allocation. Masking was achieved with matched placebo capsules and a dummy uptitration schedule. The primary endpoint was difference in step time variability between the two groups at 32 weeks, adjusted for baseline age, cognition, step time variability, and number of falls in the previous year. We measured step time variability with a triaxial accelerometer during an 18 m walking task in three conditions: normal walking, simple dual task with phonemic verbal fluency (walking while naming words beginning with a single letter), and complex dual task switching with phonemic verbal fluency (walking while naming words, alternating between two letters of the alphabet). Analysis was by modified intention to treat; we excluded from the primary analysis patients who withdrew, died, or did not attend the 32 week assessment. This trial is registered with ISRCTN, number 19880883.

Findings: Between Oct 4, 2012 and March 28, 2013, we enrolled 130 patients and randomly assigned 65 to the rivastigmine group and 65 to the placebo group. At week 32, compared with patients assigned to placebo (59 assessed), those assigned to rivastigmine (55 assessed) had improved step time variability for normal walking (ratio of geometric means 0.72, 95% CI 0.58-0.88; p=0.002) and the simple dual task (0.79; 0.62-0.99; p=0.045). Improvements in step time variability for the complex dual task did not differ between groups (0.81, 0.60-1.09; p=0.17). Gastrointestinal side-effects were more common in the rivastigmine group than in the placebo group (p<0.0001); 20 (31%) patients in the rivastigmine group versus three (5%) in the placebo group had nausea and 15 (17%) versus three (5%) had vomiting.

Interpretation: Rivastigmine can improve gait stability and might reduce the frequency of falls. A phase 3 study is needed to confirm these findings and show cost-effectiveness of rivastigmine treatment.

Funding: Parkinson's UK.