Hospitalization for gastrointestinal adverse events attributable to the use of low-dose aspirin among patients 50 years or older also using non-steroidal anti-inflammatory drugs: a retrospective cohort study


Dr E. Rahme, Division of Clinical Epidemiology, McGill University Health Centre, Division of Clinical Epidemiology, 687 Pine Avenue West, V Building, Montreal, Quebec H3A 1A1, Canada.


Background  Use of aspirin with non-steroidal anti-inflammatory drugs increases the risk of gastrointestinal ulcers; however, it is not clear if this risk varies with the non-steroidal anti-inflammatory drug used.

Aim  To assess the risk of gastrointestinal hospitalizations attributable to aspirin in patients 50 years or older also using non-steroidal anti-inflammatory drugs.

Methods  Administrative data of patients 50 years or older who received a non-steroidal anti-inflammatory drug or acetaminophen prescription between 1998 and 2004 were used.

Results  Study patients received 7 412 992 non-steroidal anti-inflammatory drug prescriptions and 5 614 044 acetaminophen prescriptions among which 23% and 32%, respectively, were dispensed to aspirin users. Time-dependent Cox regression models revealed that, compared to patients using acetaminophen (without aspirin), the adjusted hazard ratio (95% CI) among non-users of aspirin were: rofecoxib 1.3 (1.2, 1.5), celecoxib 0.7 (0.6, 0.8), diclofenac 1.5 (1.2, 1.7), ibuprofen 0.9 (0.6, 1.4), naproxen 2.5 (2.1, 3.0) and piroxicam 1.5 (0.8, 2.8); among users of aspirin: rofecoxib 3.2 (2.8, 3.7), celecoxib 1.8 (1.5, 2.1), diclofenac 2.8 (2.2, 3.5), ibuprofen 1.4 (0.8, 2.7), naproxen 2.2 (1.6, 3.0) and piroxicam 2.0 (0.8, 5.4). The risk attributable to aspirin varied from none with naproxen to 61% (53%, 68%) with celecoxib.

Conclusion  The increase in gastrointestinal hospitalization attributable to aspirin differed with the non-steroidal anti-inflammatory drug used, and seemed higher with cyclo-oxygenase-2 inhibitors than with non-selective non-steroidal anti-inflammatory drugs.