Upper gastrointestinal adverse drug reactions and cyclo-oxygenase-2 inhibitors (celecoxib and rofecoxib): a case/non-case study from the French Pharmacovigilance Database.

To evaluate the gastrointestinal safety of cyclo-oxygenase-2 inhibitors under their real conditions of use. DESIGN. Case/non-case study. Adverse drug reactions (ADRs) in adults recorded in the French Pharmacovigilance Database between 25 May 2000 and 31 December 2002. Cases were all reports of "serious" oeso-gastro-duodenal ADRs (oeso-gastro-duodenal ulcers, oesophagitis, gastritis, duodenitis). Five non-cases were randomly selected for one case from all other non oeso-gastro-duodenal reports in the database after matching them for age, gender and period of occurrence. Coxib exposure was compared among cases and non-cases, with adjustment for matching factors: French Regional Pharmacovigilance Centres that collected ADRs, reporter health professional's characteristics and exposures to non-selective non-steroidal anti-inflammatory, aspirin, anticoagulant, antiplatelet and gastroprotective drugs. Included in the study were 505 cases and 2,525 non-cases. A positive association was found between occurrence of oeso-gastro-duodenal ADRs and coxib (adjusted odds ratio 14.9 [95% CI 9.3-23.7]), diclofenac (9.2 [3.8-22.2]), ibuprofen (7.3 [3.2-16.6]) or oxicam (25.3 [11.9-53.6]) use. Despite the compulsory limits of the case/non-case methodology, the present study shows that coxibs did induce "serious" gastrointestinal ADRs in real clinical practice. These results underline the need for pharmacoepidemiological studies under real conditions of use in order to verify (or not) the conclusions of clinical trials.