Methods: We constructed a cohort of 185,010 Medicare beneficiaries diagnosed with an AMI. We followed the patients for one year after their index AMI. For each week a patient was followed, we determined whether a patient suffered another AMI, an arrhythmia or any important cardiac event. Also, in each week, we determined if a prescription was filled for any of the following antimicrobials: azithromycin, levofloxacin, clarithromycin, amoxicillin, amoxicillin-clavulanate, doxycycline or moxifloxacin. For each of the 3 cardiac outcomes and 7 antimicrobials, we created a logistic regression model with robust standard errors. The main independent variable was antimicrobial prescription in a week, and the dependent variable was a cardiac event in the following week. Each of the 21 models included patients’ comorbidities, medications, procedures, demographics, and insurance status.
Results: For 20 models, an antimicrobial prescribed in the previous week was positively associated with the cardiac events studied. Only doxycycline was not associated with AMI . Estimated odds ratios, from smallest to largest, were doxycycline (1.37 to 1.68), amoxicillin (1.72 to 2.34), azithromycin (1.72 to 2.34), amoxicillin-clavulanate (2.23 to 2.34), moxifloxacin (2.50 to 2.78), clarithromycin (3.16 to 3.70) and levofloxacin (2.99 to 2.36).
Conclusion: We found that all 7 antimicrobials were positively associated with adverse cardiac events. However, the elevated risk for non-macrolide and non-fluoroquinolone agents suggests that some of this cardiac risk may be due to the condition for which the antimicrobial agent is being prescribed. Thus, further work is needed to determine if antimicrobials themselves or the conditions they were meant to treat are the cause of these adverse events.
J. Cavanaugh, None
M. Schroeder, None
P. M. Polgreen, None