Abstract
OBJECTIVES: Numerous indicators have been proposed to evaluate the efficacy for randomized clinical trials (RCTs) of psoriasis (Pso) and psoriatic arthritis (PsA), but their comparability and correlation remain unknown. We aim to evaluate the preference and relative sensitivity of the most widely used indicators that report response rate, and to offer guidance for the primary endpoint selection for Pso and PsA trials. METHODS: We conducted a systematic search, including five databases and four registries, to identify all pharmacological intervention-controlled RCTs. A Bayesian hierarchical linear mixed model was employed to assess relative discriminations and provide a ranking of these indicators. This model, considered the gold standard for sparse and heterogeneous data, was applied to estimate differences between control and intervention groups and assess the preference and relative sensitivity of outcome indicators in Pso and PsA. RESULTS: Altogether, 386 RCTs met our inclusion criteria. We included 9 and 8 commonly used response rate indicators for Pso and PsA trials, respectively, all of which were treated as primary endpoints. We found evidence of significant differences among indicators. PASI 50, PASI 75 and IGA 0,1 proved to be robust indicators for assessing pharmacological efficacy in the majority of RCTs of Pso. Conversely, PASI 125, DIQI 0,1 and NRS-4 were not preferred under different circumstances. Furthermore, PASI 50, PASI 75 and PASI 90 appeared to be highly effective in almost all categories of pharmacological RCTs of PsA. However, due to their extreme sensitivity, it was advisable to use ACR 20 to prevent an overestimation of the therapeutic benefits of interventions. ACR 50, ACR 70 and MDA were less sensitive, but they were supposed to be more cautious in evaluating disease changing. The choice of indicators was slightly influenced by disease severity, intervention type and administration method. CONCLUSIONS: The notable efficacy discrimination ability of indicators underscores the importance of flexibility and comprehensiveness in selecting primary outcome(s). Our findings provide practical implications for optimizing indicator selection in future trial design, ensuring better alignment with trial objectives and disease characteristics. PROSPERO number: CRD42022337725.