Implantable Cardioverter Defibrillators in Primary and Secondary Prevention

BACKGROUND: Sudden cardiac death is common in persons with cardiovascular disease. PURPOSE: To assess the efficacy of implantable cardioverter defibrillators (ICDs) in persons at increased risk for sudden cardiac death. DATA SOURCES: MEDLINE (1980-2002), EMBASE (1980-2002), Cochrane Controlled Clinical Trial Registry (2002, Volume 3), other databases, and conference proceedings. Primary study authors and device manufacturers were contacted, and bibliographies of relevant papers were hand searched. STUDY SELECTION: Randomized, controlled clinical trials evaluating ICDs versus usual care were selected. DATA EXTRACTION: Two reviewers extracted data independently. DATA SYNTHESIS: Eight trials were included in the final analysis (4909 patients, 1154 deaths). Compared with usual care (most commonly amiodarone therapy), ICDs significantly reduced sudden cardiac death (relative risk RR, 0.43 95% CI, 0.35 to 0.53) and all-cause mortality (RR, 0.74 CI, 0.67 to 0.82). The included trials were divided a priori into two categories: secondary prevention (involving patients resuscitated after cardiac arrest or unstable ventricular tachycardia or ventricular fibrillation n = 1963) and primary prevention (involving patients at increased risk for sudden cardiac death but without documented cardiac arrest, ventricular fibrillation, or ventricular tachycardia n = 2946). Regardless of baseline risk, ICDs were equally efficacious in preventing sudden cardiac death in both types of trials (RR, 0.50 CI, 0.38 to 0.66 for secondary prevention vs. 0.37 CI, 0.27 to 0.50 for primary prevention). However, the magnitude of benefit in total mortality varied within the primary prevention trials depending on baseline risk for sudden cardiac death. CONCLUSIONS: Implantable cardioverter defibrillators prevent sudden cardiac death regardless of baseline risk. However, their impact on total mortality is sensitive to baseline risk for arrhythmic death. Decisions about resource allocation for ICDs depend on accurate stratification of patients according to risk.