Development of responder definitions for fibromyalgia clinical trials

Abstract

Objective: To develop responder definitions for fibromyalgia (FM) clinical trials using key symptom and function domains.

Methods: Twenty-four candidate responder definitions were developed by expert consensus and were evaluated in 12 randomized, placebo-controlled trials of 4 medications for the treatment of FM. For each definition, the treatment effects of the medication compared with placebo were analyzed using Cochran-Mantel-Haenszel tests or chi-square tests. A meta-analysis of the pooled results for the 4 medications established risk ratios to determine the definitions that best favored medication over placebo.

Results: Two definitions performed best in the analyses. Both definitions included ≥30% reduction in pain and ≥10% improvement in physical function. The definitions differed in that one (≥30% improvement in FM [FM30] short version) included ≥30% improvement in sleep or fatigue, and the other (FM30 long version) required ≥30% improvement in 2 of the following symptoms: sleep, fatigue, depression, anxiety, or cognition. In the analysis of both versions, the response rate was ≥15% for each medication and was significantly greater compared with placebo. The risk ratio favoring drug over placebo in the pooled analysis for FM30 version 3 (short version) was 1.50 (95% confidence interval [95% CI] 1.24-1.82; P ≤ 0.0001); the risk ratio for FM30 version 6 (long version) was 1.60 (95% CI 1.31-1.96; P ≤ 0.00001).

Conclusion: Among the 24 responder definitions tested, 2 were identified as most sensitive in identifying response to treatment. The identification of responder definitions for FM clinical trials that include assessments of key symptom and function domains may improve the sensitivity of clinical trials to identify meaningful improvements, leading to improved management of FM.

Figure 1

Forest plots of FM30 version 3 (A) and FM30 version 6 (B). Risk ratios and 95% confidence intervals (CI) are presented, along with heterogeneity results. Overall effect was obtained using a random effects model with Mantel-Haenszel (M–H) test. All drugs show statistically significant superiority against placebo using both versions.