Implantable cardioverter defibrillator lead performance: A systematic review and individual patient data Meta-analysis.

Background: Reliable post-approval surveillance of implantable cardioverter-defibrillator (ICD) lead performance remains a challenge. In the past, two ICD leads were recalled due to a high frequency of failures. In this meta-analysis, we sought to provide a combined estimate of failure-free rate for ICD leads by reconstructing individual patient data from published Kaplan-Meier (KM) curves and to investigate whether estimates could be influenced by the characteristics of the study.
Methods: Observational studies assessing failure-free estimates of transvenous ICD leads with KM method, were identified through a systematic search up to November 2021.
Results: Forty-four studies were eligible that included 41,870 (63.1%) non-recalled leads and 24,493 (36.9%) recalled leads. The 8-year cumulative failure-free rate was 94.1% (CI, 93.6% - 94.6%) for contemporary non-recalled leads and 81.2% (80.3% - 82.0%) for recalled leads (hazard ratio [HR], 3.15 [2.85-3.47], p < 0.001). Failure-free rate was lower in single-center studies in both the non-recalled (HR, 0.28 [0.15-0.51], p < 0.001) and recalled (HR, 0.54 [0.33-0.88], p = 0.014) group compared with multicenter studies. Similarly, estimates were significantly lower in small (i.e. extracted KM curve with <312 leads) versus large studies (HR non-recalled group, 0.54 [CI, 0.33-0.89], p = 0.015; HR recalled group, 0.62 [CI, 0.43-0.89], p = 0.009).
Conclusions: In this meta-analysis including >66,000 leads, we provide pooled survival curves that may play a role in generating evidence-based standards for assessing clinically acceptable failure rates for ICD leads. Lead performance was underestimated with single-center and small-sized studies; multicenter studies remain the main tool to reliably conduct post-market surveillance of ICD leads.
Competing Interests: Declaration of Competing Interest D.G. and A.G. are employees of BIOTRONIK Italia. All the remaining authors have no conflicts of interest to disclose.
(Copyright © 2022 Elsevier B.V. All rights reserved.)