Response criteria for clinical trials on osteoarthritis of the knee and hip: a report of the Osteoarthritis Research Society International Standing Committee for Clinical Trials response criteria initiative

The domains of pain, function and patient's global assessment are identified as core variables and frequently measured in clinical trials of patients with osteoarthritis (OA) of the hip and knee. To develop response criteria for OA of hip and knee based on the domains of pain, function and patient's global assessment. A methodology was developed by an interaction of the Osteoarthritis Research Society International Standing Committee on Clinical Trials, biostatisticians, pharmaceutical company representatives and health agency representatives. Data from previously conducted placebo-controlled clinical trials were normalized and collated. Data were subset by location of OA (knee, hip), active agent used in the clinical trial (non-steroidal anti-inflammatory drug, other agent) and route of administration (oral, intra-articular). Statistical analysis identified response criteria which best discriminate active agent from placebo. Based on the analysis of data from 14 studies (totaling 1886 patients) and consensus opinion, the optimal responder criteria set differed for location of OA, active agent to be used, and route of administration. Because of nearly identical statistical results, two sets of responder criteria are proposed: (1) 'high' pain response or, alternatively, a 'moderate' response for at least two of three domains: pain, function and patient's global assessment; (2) 'high' response for either pain or function or, alternatively, a 'moderate' response for at least two of three domains: pain, function and patient's global assessment. The sensitivity (i.e., the percentage of responders in the active group) ranged from 52 to 96% and the specificity (i.e., the percentage of nonresponders in the control group) from 47 to 73%. Based on data from clinical trials, two sets of responder criteria have been developed that can categorize an individual's responses to treatment in a clinical trial. These responder criteria require validation in additional datasets.