(164) Early identification of the association between rofecoxib and cardiovascular adverse events using real-world data from the UK: A case-control study

Background: Spontaneous reporting systems have been the mainstay for detecting adverse drug reactions (ADRs) of new drugs. However, they have important limitations, such as the inability to detect common ADRs, as was the case with higher risk of cardiovascular adverse events with rofecoxib use in the 2000s. Real-world data (RWD) and pharmacoepidemiologic methodologies might address this gap.

Objective: To study whether cardiovascular adverse effects of rofecoxib could have been detected earlier using RWD, comparing case-control and case-crossover designs.

Methods: We included all adult patients who ever used rofecoxib from the UK Clinical Practice Research Datalink (CPRD) between 20.05.1999 (rofecoxib’s market uptake in the UK) and 30.09.2004 (market withdrawal). The outcome of interest was the first occurrence of a major adverse cardiovascular event (MACE) including myocardial infarction, stroke and heart failure. Cases who experienced the first-ever MACE (index-date) were matched with up to four controls based on age, sex, general practitioner region, and calendar time. Rofecoxib exposure was stratified based on the most recent prescription before index-date into current (3 months), recent (3-6), and past use (>6). In the case-crossover design, odds of exposure were compared between a 3-month risk window before the index-date with up to four prior 3-month reference windows among cases only. Conditional logistic regression models were fitted to estimate the associations. A cumulative uptake time analysis was used to identify the shortest time required to detect MACE among rofecoxib users.

Results: A total of 3,218 cases (mean age of 73.8 years, 66% female) were matched to 10,745 controls. Current use of rofecoxib was associated with an increased risk of MACE (adjusted odds ratio [aOR] 1.19, 95% CI 1.08-1.30), compared with past use, while the case-crossover design yielded a higher association (aOR 1.83, 95%CI 1.53-2.18). Cumulative uptake time analysis revealed that the association between current rofecoxib use and MACE was detectable within three years after introduction with a case-control design (aOR 1.40, 95% CI 1.14-1.72) and already after two years with a case-crossover design (aOR 3.94, 95% CI 1.88-8.25).


Conclusion: Cardiovascular adverse effects of rofecoxib could have been detected already after two years with a case-crossover and after three years with a case-control design using RWD from the UK.