Nocebo response in Parkinson's disease: A systematic review and meta-analysis

To estimate the magnitude of the nocebo response in Parkinson's disease and explore possible associations with study characteristics.

Databases were searched up to February 2017. Placebo-controlled, parallel-group randomized controlled trials investigating pharmacological interventions in people with Parkinson's disease were included. Data were derived from the last measured within-group response in the placebo and intervention arms of randomized controlled trials, after independent extraction. A random-effects model was used to pool study data. The main outcome was the nocebo response, measured as the proportion of placebo-treated participants experiencing any adverse events (AEs). We also measured the proportion of patients with serious AEs (SAEs), and the rates of study dropouts (including due to AEs) and death. PROSPERO registration number is CRD42017070471.

We included 236 randomized controlled trials, with a combined population of 17,381 participants allocated to placebo. The nocebo response was 56.0% (95% CI, 51.7%-60.4%; 148 trials; I² = 98%). SAEs were reported in 4.0% (95% CI, 3.4%-4.6%, 157 trials; I² = 73%) of placebo-treated patients, dropouts in 14.0% (95% CI, 12.5%-15.5%, 225 trials; I² = 91%), dropouts due to AEs in 5.7% (95% CI, 5.1%-6.4%, 219 trials; I² = 73%). Deaths occurred in 0.6% (95% CI, 0.5%-0.7%, 227 trials; I² = 0%). Similar proportions were identified in patients in intervention arms.

The magnitude of the nocebo response in parallel-designed randomized controlled trials in Parkinson's disease is substantial and should be considered in the interpretation of safety results and in the design and interpretation of future clinical trials.