Purpose: Various population-based cohort studies have shown that antimicrobial agents increase the risk of overanticoagulation in patients using coumarins. In this study, we assessed this association in hospitalized patients.
Methods: We included all patients hospitalized in the Spaarne Gasthuis (Haarlem/Hoofddorp, the Netherlands), who started using an antimicrobial agent during acenocoumarol treatment or vice versa between 1 January 2015 and 1 July 2019. Patients were followed from start of concomitant therapy until 48 h after termination of the concomitant therapy or discharge, whichever came first. We analyzed the association between the antimicrobial agents and the risk of overanticoagulation, defined as an interpolated INR above 6, using Cox regression analysis. We corrected for multiple testing with the Bonferroni correction. Patients who started using acenocoumarol and amoxicillin/clavulanic acid were used as reference group.
Results: In the study population, sixteen antimicrobial agents were started frequently concomitantly with acenocoumarol treatment. We included 2157 interaction episodes in 1172 patients. Patients who started using the combination of co-trimoxazole (HR 3.76; 95% CI 1.47–9.62; p = 0.006), metronidazole (HR 2.55; 95% CI 1.37–4.76; p = 0.003), or itraconazole (HR 4.11; 95% CI 1.79–9.45; p = 0.001) concomitantly with acenocoumarol treatment had an increased risk of overanticoagulation compared with patients using acenocoumarol and amoxicillin/clavulanic acid concomitantly. The associations for metronidazole (p = 0.045) and itraconazole (p = 0.015) remained statistically significant after correction for multiple testing.
Conclusion: Co-trimoxazole, metronidazole, and itraconazole increase the risk of overanticoagulation in patients using acenocoumarol. These combinations should be avoided if possible or otherwise acenocoumarol doses should be reduced and INR measured more frequently.