This study tested whether galcanezumab, a humanized monoclonal antibody with efficacy against migraine, was superior to placebo for the treatment of mild or moderate osteoarthritis (OA) knee pain. In a multicenter, double-blind, placebo- and celecoxib-controlled trial, patients with moderate to severe OA pain were randomized to placebo; celecoxib 200mg daily for 16 weeks; or galcanezumab 5, 50, 120, and 300mg subcutaneously every 4 weeks, twice. The primary outcome was change from baseline at Week 8 in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscore measured by 100mm visual analog scale (VAS). The trial was considered positive if ≥1 dose of galcanezumab demonstrated ≥95% Bayesian posterior probability of superiority to placebo and ≥50% posterior probability of superiority to placebo by ≥9mm. A planned interim analysis allowed termination of the study if posterior probability of superiority to placebo by ≥9mm was ≤5%. Secondary endpoints included WOMAC function subscore and Patient Global Assessment (PGA) of OA. Safety and tolerability were also assessed. The study was terminated after interim analysis suggested inadequate efficacy. Celecoxib significantly reduced WOMAC pain subscore compared with placebo [-12.0mm; 95% confidence interval (CI)-23 to-2mm]. None of the galcanezumab arms demonstrated clinically meaningful improvement (range: 1.5 to-5.0mm) or met the prespecified success criteria. No improvement in any secondary objective was observed. Galcanezumab was well tolerated by OA patients. This study failed to demonstrate sufficient statistical evidence that galcanezumab was efficacious for treating OA knee pain. NCT02192190.