Iatrogenic effects of COX-2 inhibitors in the US population: findings from the Medical Expenditure Panel Survey

Abstract

Background: Selective cyclo-oxygenase 2 inhibitors ('coxibs') have been demonstrated to increase cardiovascular risk, but the cumulative burden of adverse effects in the US population is uncertain.

Objective: To quantify cardiovascular and gastrointestinal (GI) haemorrhage disease burden from coxibs and traditional 'non-selective' non-steroidal anti-inflammatory drugs (t-NSAIDs) in the US population.

Design, setting and participants: Adult respondents from the 1999-2003 Medical Expenditure Panel Survey, a representative sample of the US population which first became available in December 2006, were included. Respondents were followed for 2 years. Exposure was defined by two or more prescriptions of rofecoxib, celecoxib or a t-NSAID in the first year.

Main outcome measures: Acute myocardial infarction (AMI), stroke and/or GI haemorrhage in the year following exposure.

Results: Exposure to rofecoxib was associated with an adjusted odds ratio (OR) of 3.30 for AMI (95% CI 1.41, 7.68; p=0.01) and 4.28 for GI haemorrhage (95% CI 1.33, 13.71; p=0.02). Celecoxib was not associated with a statistically significant effect on AMI (OR 1.44; 95% CI 0.57, 3.69; p=0.44), but there was an OR of 2.43 for stroke (95% CI 1.05, 5.58; p=0.04) and 4.98 for GI haemorrhage (95% CI 2.22, 11.17; p<0.001). The group of t-NSAIDs was not associated with a significant adverse effect on AMI (OR 1.47; 95% CI 0.76, 2.84; p=0.25) or stroke (OR 1.26; 95% CI 0.42, 3.81; p=0.68), and was associated with an OR of 2.38 for GI haemorrhage (CI 1.04, 5.46; p=0.04). In the 1999-2004 period rofecoxib was associated with 46 783 AMIs and 31 188 GI haemorrhages; celecoxib with 21 832 strokes and 69 654 GI haemorrhages; resulting in an estimated 26 603 deaths from both coxibs. The t-NSAID group was associated with an excess of 87 327 GI haemorrhages and 9606 deaths in the same period.

Conclusions: Iatrogenic effects of coxibs in the US population were substantial, posing an important public health risk. Drugs that were rapidly accepted for assumed safety advantages proved instead to have caused substantial injury and death.