812 Synergistic Effects of Concomitant Use of Non-Selective NSAIDs, Coxibs and Low-Dose Aspirin on Risk of Upper Gastrointestinal Bleeding

Abstract

Aim: Concomitant use of non-steroidal anti-inflammatory drugs (NSAIDs) and low-dose aspirin (LDA) increases the risk of upper gastrointestinal bleeding (UGIB). Although clinical guidelines suggest to avoid certain drug combinations, little is known about the magnitude of interaction between different NSAIDs and specific drug groups. This is partly due to the need of a large sample for studying drug interactions. Understanding drug synergism is important to develop strategies to minimize the risk of UGIB. Design: A self-controlled case series analysis was performed using data from seven population-based healthcare databases located in three countries (Denmark, Italy, and the Netherlands). UGIB cases were identified based on disease classification codes. Drug exposure was classified into mutually exclusive groups. Interaction of the three main groups (non-selective (ns)NSAIDs, coxibs and LDA) with other drugs was examined. Incidence rate ratios (IRRs) of UGIB during exposure (and corresponding 95% CI) were obtained by Poisson regression modeling. Interaction was assessed on additive (with relative excess risk due to interaction, RERI) and multiplicative scale. Results: In total 114,835 UGIB cases, with corresponding follow-up of 930,888 personyears, were analyzed. Monotherapy with nsNSAIDs was associated with a higher IRR than monotherapy with coxibs or LDA (IRR 4.3; 2.9 and 3.1 respectively). The IRR was highest for concomitant use of nsNSAIDs and steroids (IRR 12.8; 95% CI: 11.2-14.7), which also demonstrated the highest additive interaction (RERI 5.5)(Figure 1). The IRR for nsNSAIDs and aldosterone antagonists was 11.0 (95%CI 8.6 -14.0) and RERI 4.5. Selective serotonin re-uptake inhibitors (SSRIs) and anticoagulants combined with either nsNSAIDs, coxibs or LDA increased the risk of UGIB significantly and also to a greater extent than expected based on the sum of the individual drugs (RERI 1.6;1.9 and 0.5 for SSRIs and 2.4;0.1 and 1.9 for anticoagulants, respectively). Only for the combination of LDA and antiplatelets was apart from additive, also multiplicative interaction observed (IRR 5.4; 95%CI: 4.7-6.4; RERI 1.7). Increasing age was associated with a higher IRR. Stratifying by age showed that for all drugs older persons (.60 years) had a higher IRR compared to younger persons ( ,60 years), except for combinations of nsNSAIDs+anticoagulants and coxibs+steroids. Conclusion: Concomitant use of SSRIs with nsNSAIDs, coxibs or LDA significantly increases the risk of UGIB up to seven-fold. Concomitant use of steroids, anticoagulants or antiplatelets with nsNSAIDs or LDA, but not with coxibs, showed an increased risk of UGIB, up to thirteen-fold. These increased risks were greater than the sum of the risks of individual drugs. This knowledge is of clinical relevance and can help clinicians in tailoring therapy to minimize UGI adverse events.