Responder analysis and correlation of outcome measures: pooled results from two identical studies comparing etoricoxib, celecoxib, and placebo in osteoarthritis

Abstract

To determine the proportion of responders in two identical osteoarthritis (OA) trials using Outcome Measures in Arthritis Clinical Trials-Osteoarthritis Research Society International (OMERACT-OARSI) criteria and to assess the comparability and correlation of individual component measurements. Data were pooled from two identical 26-week, double-blind, randomized, parallel, multicenter trials comparing once daily etoricoxib 30 mg (N=475), celecoxib 200 mg (N=488), and placebo (N=244) in patients with OA of the knee or hip. OMERACT-OARSI criteria were (1) improvement in pain or physical function > or =50% and an absolute change > or =20 mm on a 100-mm visual analog scale (VAS); or (2) improvement of > or =20% and with an absolute change > or =10 mm in at least two of the following three categories: pain, physical function, and patient's global assessment. Correlations were assessed between endpoints measured as time-weighted average change from baseline over 12 weeks using Pearson's correlation coefficient (r). There were significantly greater proportions of responders in the etoricoxib (66.2%) and celecoxib (63.5%) groups compared with the placebo group (43.0%; P<0.001). There was no difference between the two active treatment groups. There was high correlation between pain and physical function (r=0.903), pain and global assessment (r=0.778), and physical function and global assessment (r=0.820). There was high sensitivity (75-87%) and specificity (80-96%) for changes in individual component measurements to predict OMERACT-OARSI responders. Significantly more patients receiving etoricoxib or celecoxib than placebo were OMERACT-OARSI responders. The high correlation between individual scales composing this composite response measurement suggests some redundancies between individual components, particularly between pain and physical function.