The goal of this meta-analysis was to assess the frequency and magnitude of placebo and nocebo responses in placebo-controlled randomized controlled trials (RCTs) of Janus kinase (JAK) inhibitor monotherapy for rheumatoid arthritis (RA) METHODS: We performed a meta-analysis on the rates of placebo response, adverse effects (AEs), severe AEs (SAEs) and withdrawal due to AEs in placebo-controlled randomized clinical trials (RCTs) of JAK inhibitor therapy for RA.
Five RCTs contained a total of 1422 patients (746 trial participants and 676 controls). The pooled incidence of an American College of Rheumatology 20% (ACR20) response rate was 33.0% (95% CI 19.6-44.9%) in placebo-treated patients and 68.3% (95% CI 61.4-74.1%) in active drug-treated patients. A strong negative correlation was observed between drug efficacies (ACR20 response) and AE rates in the placebo arm, indicating that the stronger the placebo response, the weaker the nocebo response (r = -0.906, P = 0.034). The pooled estimate of at least one AE was 54.1% (95% CI 44.6-63.4%) in placebo-treated patients and 54.5% (95% CI 46.2-62.6%) in active drug-treated patients. The pooled SAE rate was 3.9% (95% CI 2.7-5.7%) in placebo-treated patients and 3.8% (95% CI 2.5-5.7%) in active comparator-treated patients. The pooled estimate of withdrawal owing to an AE was 4.1% (95% CI 1.4-11.3%) in placebo-treated patients and 2.1% (95% CI 0.8-5.4%) in active drug-treated patients. However, there were no differences in the pooled risk of AE, SAEs, or withdrawal owing to AEs between the active comparator and placebo groups. A strong positive correlation was observed in AE rates between the placebo and active arms, indicating that the stronger the nocebo response, the higher the AE rate in the active arm (r = 0.957, P = 0.012).
The frequency of placebo and nocebo responses was 33.0 and 54.1%, respectively, in JAK monotherapy trials for RA. The findings indicated that the strengths of placebo and nocebo responses are inversely proportional and that clinically significant differences were absent between AE, SAE, and dropout owing to AEs.