The APC and PreSAP trials: A post hoc noninferiority analysis using a comprehensive new measure for gastrointestinal tract injury in 2 randomized, double-blind studies comparing celecoxib and placebo

Background: Previous gastrointestinal (GI) outcomes of nonsteroidal anti-inflammatory drug (NSAID) trials have focused on upper GI events, although recent evidence suggests NSAID-related lower GI effects are important and clinically relevant. Objective: We assessed the long-term GI adverse event (AE) profile of celecoxib in a nonarthritis population. The aim of this post hoc analysis was to determine the incidence of serious GI AEs, using a new Clinically Significant Upper and/or Lower GI Events end point. Methods: Patients from 2 colorectal adenoma recurrence studies were included. Patients received celecoxib 200 mg/400 mg BID, 400 mg once daily, or placebo over 3 years. The analysis measured noninferiority, using a prespecified definition of noninferiority. Celecoxib was predefined to be noninferior to placebo if the upper limit of the 95% CI for the hazard ratio (HR) with celecoxib was <1.25, at any dose, compared with the placebo (calculated using the Cox proportional hazards model). Results: A total of 3588 patients were included; in the primary analysis, the HR for celecoxib (any dose) compared with placebo was 1.22 (95% CI: 0.69-2.18; P = 0.4948). In the secondary dose analyses, the HR associated with a 400-mg daily dose, compared with placebo, was 1.04 (95% CI: 0.55-1.96; P = 0.9149); for 800 mg/d, the HR was 1.79 (95% CI: 0.82-3.89; P = 0.1427). In a third covariate analysis, low-dose aspirin use (HR = 2.33; 95% CI: 1.33-4.08) and age ≥65 years (HR = 1.82; 95% CI, 1.05-3.15) was suggested to have a statistically significant association with increased risk of GI AEs. Study limitations include retrospective evaluation and small sample size of patients with GI AEs. Conclusions: The noninferiority of celecoxib to placebo was not established because the HR for the time to the first Clinically Significant Upper and/or Lower GI Event was greater than the prespecified upper limit of 95% CI for noninferiority. In addition, HRs associated with daily doses of 400 or 800 mg celecoxib compared with placebo were not significant. However, a significantly increased risk of clinically significant upper and/or lower GI events was observed in low-dose aspirin users (≤162.5 mg average daily use) and in patients ≥65 years of age. ClinicalTrials.gov identifiers: NCT00005094 and NCT00141193.