Early Improvement in Pain Predicts Pain Response at Endpoint in Patients With Fibromyalgia

An unanswered, but clinically important question is whether there are early indicators that a patient might respond to duloxetine treatment for fibromyalgia pain. To address this question, pooled data from 4 double-blind, placebo-controlled trials in duloxetine-treated patients (N = 797) with primary fibromyalgia as defined by the American College for Rheumatology were analyzed. Classification and Regression Tree (CART) analysis was used to determine what level of early pain improvement as measured by the 24-hour average pain severity question on the Brief Pain Inventory (BPI) best predicted later response. The predictor variables tested were 10, 15, 20, 25, and 30% decrease in BPI 24-hour average pain from baseline to Week 1 and Week 2. The results of the CART analysis showed that for patients with >= 15% improvement in pain at Week 1 and >= 30% improvement at Week 2, the probability of response at 3 months was 75%. For patients with <15% improvement at both Week 1 and Week 2, the probability of not responding at 3 months was 86%. Quantifiable early improvement in pain during the first 2 weeks of treatment with duloxetine was highly predictive of response or nonresponse after 3 months of treatment. Perspective: This article presents early indicators that can highly predict later pain response or nonresponse in fibromyalgia patients treated with duloxetine. The results may aid clinicians to predict the likelihood of response at 3 months within the first 2 weeks of treatment. (C) 2011 by the American Pain Society