Background: Sudden cardiac death is the single largest cause of mortality in dialysis patients. There are no published data on the use or survival impact of implantable cardioverter defibrillators (ICDs) in dialysis patients. The objective of this retrospective cohort study was to determine ICD use in dialysis patients and impact on survival.
Methods: Dialysis patients hospitalized from 1996 to 2001 for ventricular fibrillation/cardiac arrest, having ICD implantation within 30 days of admission, discharged alive, and surviving at least 30 days from admission were identified from the 100% end-stage renal disease (ESRD) sample of the Medicare database. Long-term survival was estimated by life-table method. Impact of independent predictors on survival was examined in a comorbidity-adjusted Cox model and a propensity model.
Results: There were 460 patients (7.6%) with ICD and 5582 patients (92.4%) without ICD. Estimated 1-, 2-, 3-, 4-, and 5-year survivals after day 30 of admission in the ICD group were 71%, 53%, 36%, 25%, and 22%, respectively; in the no-ICD group, 49%, 33%, 23%, 16%, and 12% (P < 0.0001). ICD implantation was independently associated with a 42% reduction in death risk [relative risk 0.58 (95% CI 0.50, 0.66)]. In the propensity model, the relative risks of death for the lower, middle, and upper third propensity groups were 0.45 (0.26, 0.81), 0.61 (0.45, 0.84), and 0.65 (0.55, 0.76), respectively. The C statistic for the propensity model equaled 0.81.
Conclusion: In dialysis patients, ICD therapy is apparently underused. ICD implantation in cardiac arrest survivors on dialysis is associated with greater survival.