The primary endpoint was American College of Rheumatology 20% improvement (ACR20) at week 24 in all patients per assigned treatment group using non-responder imputation.