QT-Prolonging Antibiotics, Serum-to-Dialysate Potassium Gradient, and Risk of Sudden Cardiac Death among Patients Receiving Maintenance Hemodialysis

Rationale & Objective: Treatment with certain QT interval-prolonging antibiotics is associated with a higher risk of sudden cardiac death among individuals with hemodialysis-dependent kidney failure. Concurrent exposure to large serum-todialys ate potassium gradients, which promote large potassium shifts, may augment the proarrhythmic effects of these medications. The primary objective of this study was to examine whether the serum-to-dialys ate gradient modifies the cardiac safety of azithromycin, and separately, levofloxacin/moxifloxacin. Study Design: Retrospective observational cohort study using a new-user study design. Setting & Population: Adult in-center hemodialysis patients with Medicare coverage in the US Renal Data System (2007-2017). Exposure: Initiation of azithromycin (or levofloxacin/moxifloxacin) as compared to amoxicillinbased antibiotics (exposure). Serum-to-dialys ate potassium gradient (effect modifier). Individual patients could contribute multiple study antibiotic treatment episodes to the analyses. Outcomes: Sudden cardiac death (14 days). Analytical Approach: Inverse probability of treatment-weighted survival models to estimate HRs and robust 95% CIs. Results: The azithromycin versus amoxicillinbased antibiotic cohort included 89,379 unique patients with 113,516 azithromycin and 103,493 amoxicillin-based treatment episodes. Azithromycin versus amoxicillinbased antibiotic treatment was associated with a higher risk of sudden cardiac death overall, HR, 1.68; 95% CI, 1.31-2.16. The risk was numerically higher when the baseline serum-todialys ate potassium gradient was & GE;3 mEq/L compared with <3 mEq/L (HR, 2.22; 95% CI, 1.46-3.40 vs HR, 1.43; 95% CI. 1.04-1.96, Pinteraction = 0.07). Analogous analyses in a respiratory fluoroquinolone (levofloxacin/ moxifloxacin) versus amoxicillin-based antibiotic cohort with 79,449 unique patients and 65,959 respiratory fluoroquinolone and 103,776 amoxicillin-based treatment episodes yielded similar results. Limitations: Residual confounding. Conclusions: Although treatment with azithromycin and, separately, respiratory fluo-roquinolones were each associated with a heightened risk of sudden cardiac death, this risk was augmented in the setting of larger serum-todialys ate potassium gradients. Minimizing the potassium gradient may be an approach to reduce the cardiac risk of these antibiotics.