Withholding ICDs after EF improves to > 35% can not be justified in all patients with CAD and left ventricular systolic dysfunction (LVSD)

Aurora Affiliations

Aurora Cardiovascular Services

Aurora Sinai/Aurora St. Luke’s Medical Centers

Aurora Research Institute

Center for Integrative Research on CV Aging - Aurora Health Care

Presentation Notes

Poster presented at: Heart Rhythm Scientific Sessions; May 11, 2017; Chicago, IL.


BACKGROUND: When EF≤35% is identified, ICDs are considered for SCD protection, but are withheld if EF improves to >35% within 90 days. This practice is not supported by direct evidence and may leave a significant segment of the population exposed to SCD risk.

OBJECTIVE: Evaluate relationship of EF to timing of appropriate ICD therapy, as well as impact of withholding ICD after EF improves to >35%, in CAD patients with at least moderate LVSD.

METHODS: Aurora Healthcare patients with EF≤35% and subsequent improvement at any time to EF>35% were included; excluding patients with: EF recovery in ≤7 days, CRT, LVAD, transplant, and inherited sudden death syndromes. After propensity matching, the study cohort (n=798) was segregated by presence/absence of ICD. SCD events and appropriate ICD therapies (ApprRx) were tabulated in ICD patients, as well as EF at the time of the event.

RESULTS: Of 133 ICD recipients (48.1%>65 years, 31.6% female) 6% suffered ApprRx over 23±13 months. In 62.5% of patients with ApprRx the EF was confirmed >35%, including in 75% of primary prevention patients with AppRx. In secondary prevention (2°) patients, a greater percentage (25%) of patients with EF>35% received ApprRx, compared to those with EF≤35% (7.1%). Alarmingly, 22.2% of 2° patients did not have an ICD at the time of their initial SCD event because they were disqualified due to EF recovery.

CONCLUSION: In patients with CAD, the ICD provides sudden death protection even after EF improves to >35%. The practice of withholding ICD therapy for EF recovery to >35% was associated with sudden death events in this analysis, and requires reexamination.

Document Type




Unaffiliated Users