Abstract
The objective of the present analysis was to determine whether changes in Brief Pain Inventory (BPI) average pain scores by patient global impression of improvement (PGI-I) category and the cut-off for clinically important difference (CID) were different between Asian and Caucasian patients with chronic pain due to osteoarthritis. This analysis used data from 3 (Caucasian) and 2 (Asian) randomized, placebo-controlled, 10- to 14-week duloxetine studies for the treatment of patients ≥40 years of age with osteoarthritis pain. The receiver operating characteristic (ROC) analysis was used to characterize the association between changes in BPI average pain scores and PGI-I levels at study endpoint. The CID was characterized by PGI-I, and the cut-off point for CID in BPI average pain scores was determined by the intersection of a 45-degree tangent line with each ROC curve. Data from 668 Asian and 868 Caucasian patients were available for analysis. Baseline BPI average pain ratings including worst and least pain were comparable between Asians and Caucasians. Ratings for percentage change from baseline to endpoint for BPI average pain scores in Asian patients and Caucasian patients were similar across the 7 PGI-I categories, regardless of age, gender, study, and treatment. The ROC analysis results of cut-off points in BPI average pain scores demonstrated the raw change cut-off was -3.0, and percentage change cut-off was -40% for both Asian and Caucasian patients. Overall, the present analysis concludes changes in BPI average pain scores by PGI-I category and the cut-off for CID were similar for Asian and Caucasian patients with chronic pain due to osteoarthritis.