Don Williamson

Key Points: Sudden cardiac death is a major concern for hemodialysis patients. Mortality is higher on dialysis days and is associated with higher dialysate bicarbonate (D_BIC). Contrary to our hypothesis, there was no consistent association of higher D_BIC with a higher risk of arrhythmia. Further research is needed to assess the optimal D_BIC and mechanisms by which it may improve outcomes for maintenance hemodialysis patients. Background: Sudden death accounts for approximately 25% of deaths among maintenance hemodialysis patients, occurring more frequently on hemodialysis days. Higher dialysate bicarbonate (D_BIC) may predispose to alkalemia and arrhythmogenesis. Methods: We conducted a 12-month analysis of session-level data from 66 patients with implantable loop recorders. We fit logistic regression and negative binomial mixed-effects regression models to assess the association of D_BIC with clinically significant arrhythmia (ventricular tachycardia ≥115 beats per minute [BPM] for at least 30 seconds, bradycardia ≤40 BPM for at least 6 seconds, or asystole for at least 3 seconds) and reviewer confirmed arrhythmia (RCA—implantable loop recorder-identified or patient-marked event for which a manual review of the stored electrocardiogram tracing confirmed the presence of atrial fibrillation, supraventricular tachycardia, sinus tachycardia with rate >130 BPM, ventricular tachycardia, asystole, or bradycardia). Models adjusted for age, sex, race, hemodialysis vintage, vascular access, and prehemodialysis serum bicarbonate and additionally for serum and dialysate potassium levels. Results: The mean age was 56±12 years, 70% were male, 53% were Black, and 35% were Asian. Fewer RCA episodes were associated with D_BIC >35 than 35 mEq/L (incidence rate ratio 0.45 [0.27 to 0.75] and adjusted incident rate ratio 0.54 [0.30 to 0.97]), but the association was not significant when adjusting for serum and dialysate potassium levels (adjusted incident rate ratio, 0.60 [0.32 to 1.11]). Otherwise, no associations between D_BIC and arrhythmia were identified. Conclusions: We observed a lower frequency of RCA with higher D_BIC, compared with D_BIC of 35 mEql/L, contrary to our original hypothesis, but this association was attenuated in fully adjusted models. Validation of these findings in larger studies is required, with a further need for interventional studies to explore the optimal D_BIC concentration.