Longitudinal Outcomes Associated With Non-Evidence-Based Implantable Cardioverter-Defibrillators Among Medicare Beneficiaries (From the National Cardiovascular Data Registry)

Abstract

Primary prevention implantable cardioverter-defibrillators (ICDs) in patients with recent myocardial infarction or coronary revascularization and those with newly diagnosed or severe heart failure (HF) are considered non-evidence-based, as defined by pivotal randomized clinical trials. Although non-evidence-based ICDs have been associated previously with greater risk of in-hospital adverse events, longitudinal outcomes are not known. We used Medicare-linked data from the National Cardiovascular Data Registry's ICD Registry to identify patients discharged alive following first-time primary prevention ICD implantations performed between 2010 and 2013. We compared longitudinal outcomes, including all-cause mortality and all-cause hospital readmission among patients receiving non-evidence-based versus evidence-based ICDs, up to 4.75 years after implantation, using multivariable time-to-event analyses. Of 71,666 ICD implantations, 9,609 (13.4%) were classified as non-evidence-based. Compared to patients receiving evidence-based ICDs, non-evidence-based ICD recipients had greater mortality risk at 90 days (HR=1.44, CI: 1.37 - 1.52, p <0.0001) and at 1 year (HR=1.19, CI: 1.15 - 1.24, p <0.0001), but similar mortality risk at 3 years (HR=1.03, CI: 0.98 - 1.08, p=0.2630). Risk of all-cause hospitalization was higher in patients with non-evidence-based ICDs at 90 days (HR=1.17, CI: 1.14 - 1.20, p <0.0001), but the difference diminished at 1 year (HR=1.04, CI 1.00 - 1.07, p=0.0272) and at 3 years (HR=0.94, CI: 0.90 - 0.99, p=0.0105). In conclusion, among patients undergoing primary prevention ICD implantations between 2010 and 2013, those with non-evidence-based ICDs were at increased risk of mortality and readmission during longitudinal follow-up. Differences in the risk of mortality and hospitalization were highest in the first year following device implantation.