Systematic review of enriched enrolment, randomised withdrawal trial designs in chronic pain: a new framework for design and reporting.

Enriched enrolment, randomised withdrawal (EERW) pain trials select, before randomisation, patients who respond by demonstrating a predetermined degree of pain relief and acceptance of adverse events. There is uncertainty over the value of this design. We report a systematic review of EERW trials in chronic noncancer pain together with a critical appraisal of methods and potential biases in the methods used and recommendations for the design and reporting of future EERW trials. Electronic and other searches found 25 EERW trials published between 1995 and June 2014, involving 5669 patients in a randomised withdrawal phase comparing drug with placebo; 13 (median, 107 patients) had a randomised withdrawal phase of 6 weeks or less, and 12 (median, 334) lasted 12 to 26 weeks. Risks of bias included short duration, inadequate outcome definition, incomplete outcome data reporting, small size, and inadequate dose tapering on randomisation to placebo. Active treatment was usually better than placebo (22/25 trials). This review reduces the uncertainty around the value of EERW trials in pain. If properly designed, conducted, and reported, they are feasible and useful for making decisions about pain therapies. Shorter, small studies can be explanatory; longer, larger studies can inform practice. Current evidence is inadequate for valid comparisons in outcome between EERW and classical trials, although no gross differences were found. This systematic review provides a framework for assessing potential biases and the value of the EERW trials, and for the design of future studies by making recommendations for the conduct and reporting of EERW trials.

Evaluation of GW406381 for treatment of osteoarthritis of the knee: two randomized, controlled studies.

CONTEXT: GW406381 is an investigational, highly selective cyclooxygenase-2 (COX-2) inhibitor that is effective in animal models of central sensitization and of inflammatory pain. OBJECTIVE: To examine dose response for efficacy and safety of GW406381 in adults with osteoarthritis (OA) of the knee. DESIGN: Two randomized, double-blind, placebo- and positive-control studies: Study A, a 6-week nonflare design; Study B, a 12-week flare design. PATIENTS: 649 patients entered Study A; 1331 patients entered Study B. STUDY A: GW406381 10, 20, 35, or 50 mg, celecoxib 200 mg, or placebo. Study B: GW406381 1, 5, 10, 25, or 50 mg, celecoxib 200 mg, or placebo. MAIN OUTCOME MEASURES: Study A, co-primary endpoints were change from baseline in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscore and WOMAC question 1. Study B co-primary endpoints were change from baseline in WOMAC pain and function subscores and Patient Global Assessment of Arthritis Condition. A closed hierarchical test procedure was prespecified. RESULTS: Study A demonstrated that GW406381 50 mg was superior to placebo on WOMAC pain subscore (mean difference from placebo -6.9 mm; P= .012). No clear dose response was observed, and the results with celecoxib were no different from those of placebo. In Study B, no dose of GW406381 was superior to placebo on the co-primary endpoints. Celecoxib was superior to placebo on all co-primary endpoints. Dose-related blood pressure and renovascular effects were seen with GW406381. CONCLUSIONS: Overall, clinically meaningful efficacy in pain related to OA of the knee was not demonstrated for GW406381 despite its peripheral and central sites of action.