Study location, patient population, and background treatment may affect American College of Rheumatology (ACR) placebo response rates in clinical trials of rheumatoid arthritis (RA) medications, according to research results published in Clinical Drug Investigation.
Researchers sought to estimate ACR placebo response rates after 3 months in clinical trials for RA treatments (20% improvement in response rate according to ACR criteria: ACR20) as well as to investigate what factors may affect these rates. A systematic review and meta-analysis of randomized, double-blind, placebo-controlled trials of both unlicensed and licensed pharmacologic agents was conducted.
The investigators searched literature databases (PubMed, Embase, and the Cochrane Central Register of Controlled Trails) up to April 2019 to identify relevant studies. Then, 2 authors independently selected studies for inclusion and extracted data; discrepancies were resolved through discussion. The primary outcome was ACR20 at 3 months and the secondary outcome was ACR50 at 3 months.
The investigators identified 6590 articles in an initial literature search and 88 were ultimately included in the meta-analysis. All included studies were multicenter, randomized, double-blind, and placebo-controlled and included patients with RA and data on ACR20 at 3 months. A total of 8406 patients who received a placebo were included in the meta-analysis.
Across all 88 trials, the pooled estimate of ACR20 at weeks 12 through 14 was 29.0% (95% CI, 27.2-30.9), with a range of 10.0% to 46.2% among the individual studies. The investigators observed substantial heterogeneity among the studies (I² =69.3%).
Univariable meta-regression analysis results demonstrated that statistically significant factors affecting ACR20 were publication year, treatment administration route, background disease-modifying antirheumatic drug (DMARD) treatment, patient population, study location, and baseline C-reactive protein level. These factors were statistically significantly associated with placebo responses and were used in the multivariable meta-regression analysis. In the multivariable model, statistically significant factors for ACR20 were publication year (odds ratio [OR], 1.027), study location in a non-Western country (OR, 0.589), and inadequate response to biologic DMARDs in the patient population (OR, 0.705). Year of publication was moderately correlated with baseline C-reactive protein value, and baseline C-reactive protein levels in monotherapy studies were higher than those in add-on studies.
ACR50, determined from pooled data from 80 studies, was 9.8% (95% CI, 8.8-10.8). Substantial heterogeneity was noted among studies (I² =52.0%). Statistically significant factors affecting ACR50 were publication year, treatment administration route, study location, and baseline C-reactive protein level.
Subgroup analyses, grouped by study location, were performed for 13 studies conducted in non-Western countries and 75 studies conducted in Western countries only or multiregional locations. Non-Western location was a statistically significant factor that contributed to placebo responses, according to the primary multivariable meta-regression analysis. A univariable meta-regression analysis of the non-Western studies found that conventional DMARD-inadequate response population and percentage of women were both statistically significant factors that affected ACR20.
Study limitations included the large number of excluded studies, the potential for publication bias among the included studies, and the use of aggregated data, which may not reflect the true relationship between outcomes and patient characteristics.
“Our meta-analysis indicated that researchers should consider study location, patient population, and presence of background DMARD treatment as potential factors influencing a placebo ACR response when designing a clinical trial for RA patients,” the researchers concluded. “Further studies with individual data are needed to expand our knowledge of the placebo effect and enable more efficient placebo-controlled studies for RA patients.”
Disclosure: Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.
Nagai N, Matsubayashi K, Ide K, Seto K, Kawasaki Y, Kawakami K. Factors influencing placebo responses in rheumatoid arthritis clinical trials: a meta-analysis of randomized, double-blind, placebo-controlled studies. Clin Drug Investig. 2020;40(3):197-209.