Gastrointestinal safety of coxibs: systematic review and meta-analysis of observational studies on selective inhibitors of cyclo-oxygenase 2.

Prior meta-analyses have shown a higher gastrointestinal risk of nonselective NSAIDs versus placebo and a lower gastrointestinal risk of coxibs versus nonselective NSAIDs. However, the available data about gastrointestinal risk for coxibs versus placebo are scarce. The aim of this study was to review the current evidence on the use of coxibs and to evaluate the risk of gastrointestinal adverse outcomes (GAO) associated with coxibs versus nonexposed. Search was conducted on PubMed and Embase databases. We selected cohort observational, case-control, nested case-control and case-crossover studies that reported the risk of GAO associated with coxibs versus nonexposed as relative risk (RR), odds ratio (OR), hazard ratio (HR) or incidence rate ratio (IRR). It was estimated the pooled RR and the 95% confidence interval (CI) for coxibs both individually and as a whole by the DerSimonian and Laird method. Twenty-eight studies met inclusion criteria. Overall, coxibs were associated with a significant increment in the risk of GAO [RR 1.64 (95% CI 1.44-1.86)]. The analysis by individual drugs showed that etoricoxib [RR 4.85 (95% CI 2.64-8.93)] presented the highest gastrointestinal risk, followed by rofecoxib [RR 2.02 (95% CI 1.56-2.61)] and celecoxib [RR 1.53 (95% CI 1.19-1.97)]. Gastrointestinal risk was also high for the subgroups aged <65 years and low-dose coxibs. The use of coxibs is associated with a statistically significant increased risk of GAO, which would be high even for low-dose coxibs and <65-year-old subgroups. The risk would be higher for etoricoxib than for celecoxib and rofecoxib.