Safety profile of celecoxib as used in general practice in England: results of a prescription-event monitoring study.

AIMS: A post-marketing surveillance study using the technique of Prescription Event Monitoring was undertaken to monitor the safety of celecoxib, a cyclo-oxygenase (COX)-2 inhibitor, as prescribed in primary care in England.
METHODS: Patients were identified from dispensed British National Health Service prescription data supplied in confidence by the Prescription Pricing Authority for celecoxib between May and December 2000. Simple questionnaires were sent to the prescribing general practitioner at least 6 months after the date of the first dispensed prescription for each individual patient. Event incidence densities (IDs) [the number of 1st reports per 1000 patient-months of exposure (pme)] were calculated. ID differences for events reported in month 1 (ID1) and months 2-6 (ID2) were examined for temporal changes in event rate. Information on suspected adverse drug reactions (ADRs), reasons for stopping treatment, outcome of pregnancies and cause of death were also requested. Data were gathered on potential gastrointestinal (GI) risk factors [recent use of other non-steroidal anti-inflammatory drugs (NSAIDs), past history of upper GI disorders and concomitant gastro-irritant agents or anti-ulcer drugs]. Crude IDs per 1000 pme and ID ratios were calculated according to potential risk factors, and age (> or = 65 years, < or = 64 years).
RESULTS: The cohort comprised of 17,458 patients [median age 62 years (IQR 51,73); 68.3% female]. The most common specified indication was osteoarthritis (28.1%, n = 4905). Not effective was the event with the highest ID1 (139.9 per 1000 pme). The clinical events with the highest ID1 were dyspepsia (25.4 per 1000 pme) followed by abdominal pain (10.6). These were also given frequently as reasons for stopping (551 and 174 of 9126 reports). Of 436 events in 325 patients (1.9% of total cohort) that were reported as ADRs, the most frequent were events within the alimentary system (186 reports). Uncommon events reported during treatment (not necessarily as ADRs) included allergy (0.10%, n = 17), anaphylaxis (0.01%, n = 2), angioneurotic oedema (0.02%, n = 3) and bronchospasm (0.05%, n = 9). There were 103 reports of events associated with thromboembolism and 111 reports of serious GI events [90 GI bleeds (upper and lower); 21 peptic ulcers] received during treatment or within 1 month of stopping. A past history of dyspeptic/other upper GI conditions and use of concomitant gastro-protective drugs were each associated with a significantly increased risk of dyspepsia and abdominal pain.
CONCLUSIONS: Frequently reported adverse events were those GI events commonly associated with treatment with other NSAIDS. Stratification by identified risk factors suggested that channelling of high-risk patients is likely. Serious upper and lower GI events, and thromboembolic events did occur during this study, although the incidence was low (< 1%). Doctors should continue to prescribe NSAIDs, including COX-2-specific inhibitors, with caution.