Combining nonselective NSAIDs, COX-2 inhibitors, or low-dose aspirin with SSRIs led to excess risks of upper gastrointestinal bleeding of 1.6 (95% CI, 0.5-2.6), 1.9 (95% CI, 0.2-3.4), and 0.49 (–0.05-1.03), respectively, the researchers reported. "From a biological point of view, this interaction seems plausible because SSRIs decrease the serotonin level, resulting in impaired thrombocyte aggregation and an increased risk of bleeding in general," they said.
Corticosteroids combined with nonselective NSAIDs led to the greatest increases in bleeding risk, with an incidence ratio of 12.8 (95% CI, 11.1-14.7), compared with nonuse of any drug studied, and an excess risk of 5.5 (3.7-7.3), compared with NSAID use alone, said the researchers. Adding aldosterone antagonists to nonselective NSAIDs led to an excess risk of 4.46, compared with using nonselective NSAIDs alone, they reported (95% CI, 1.79-7.13).
Because the study did not capture over-the-counter NSAID prescriptions, it could have underestimated use of these drugs, the investigators said. Also, changes in health or NSAID use during the study could have created residual confounding, although sensitivity analyses did not reveal problems, they reported. They added that misclassification of some data could have led them to underestimate risks. "Finally, we did not take any carryover effect or dose of drug exposure into account, which potentially limits the generalizability concerning causality of the associations," they concluded.
Five authors reported employment or other financial support from Erasmus University Medical Center, AstraZeneca, Janssen, PHARMO Institute, and the European Medicines Agency. The other authors reported no relevant conflicts of interests.