Nonsteroidal Anti-inflammatory Drugs vs Cognitive Behavioral Therapy for Arthritis Pain: A Randomized Withdrawal Trial.

Importance: Nonsteroidal anti-inflammatory drugs (NSAIDs) are commonly prescribed for knee osteoarthritis. However, they are associated with uncertain long-term clinical benefit and significant toxic effects. Objective: To evaluate whether discontinuing NSAIDs and engaging in a telephone-based cognitive behavioral therapy (CBT) program is noninferior to continuing NSAIDs for patients with knee osteoarthritis. Design, Setting, and Participants: The Stopping NSAIDs for Arthritis Pain multicenter randomized withdrawal trial was conducted for 364 patients taking NSAIDs for knee osteoarthritis pain on most days of the week for at least 3 months between September 1, 2013, and September 30, 2018. Analysis was performed on an intent-to-treat basis. Interventions: Participants discontinued their current NSAID and took 15 mg per day of meloxicam daily during a 2-week run-in period. Those who remained eligible were randomized in a 1:1 ratio to receive meloxicam or placebo for 4 weeks (blinded phase 1). Participants receiving meloxicam then continued this medication for 10 weeks, while those receiving placebo participated in a 10-week CBT program (unblinded phase 2). Main Outcomes and Measures: The primary outcome measure was the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain score at 4 weeks with the noninferiority margin set at 1. Secondary outcomes included the area under the curve of the pain score after 4 weeks as well as the WOMAC pain score, area under the curve of the pain score, WOMAC disability score, and global impression of change after treatment at 14 weeks. Results: A total of 180 participants (161 men; mean [SD] age, 58. 2 [11.8] years) were randomized to receive placebo followed by CBT, and a total of 184 participants (154 men; mean [SD] age, 58.5 [10.0] years) were randomized to receive meloxicam. After adjustment for baseline pain and study site, the estimated mean difference in WOMAC pain score between the placebo and meloxicam groups after 4 weeks was 1.4 (95% CI, 0.8-2.0; noninferiority test P = .92). At week 14, the adjusted mean difference in WOMAC pain score between the placebo (followed by CBT) and meloxicam groups was 0.8 (95% CI, 0.2-1.4; noninferiority P = .28). There was no statistically significant difference in the global impression of change (mean difference in scores, -0.2; 95% CI, -0.4 to 0.1; P = .15) or lower extremity disability (mean difference in scores, 0.9; 95% CI, -1.4 to 3.2; P = .45) between the 2 groups after 14 weeks. Conclusions and Relevance: Among patients with knee osteoarthritis, placebo and CBT (after placebo) are inferior to meloxicam. However, the WOMAC pain score differences between the 2 groups were small, and there were no statistically significant differences in participants' global impression of change or function after 14 weeks. Trial Registration: ClinicalTrials.gov Identifier: NCT01799213.