Background: The efficacy of dual-chamber ICD arrhythmia classification algorithms is crucial to prevent inappropriate shocks. We report our experience from a meta-analysis of five prospective clinical studies with inclusion phases ranging between 1997 and 2003.
Methods: Dual-chamber ICD using standard dual-chamber arrhythmia classification algorithms were implanted in 802 patients (mean age = 64 +/- 11 years, 88% men) in 74 medical centers. The ICD indication was secondary prevention in 95% of patients. Supraventricular tachyarrhythmias (SVT) were previously documented in 26% of patients. All spontaneous tachyarrhythmic events documented by the device memories were analyzed by a adjudicating committee. The episodes lasting > 12 seconds and/or treated by the ICD were analyzed.
Results: Over a mean follow-up of 302 +/- 113 days, 9,690 events were reported. Mean heart rate at the time of events was 131 +/- 45 bpm (100-430). Events were classified as oversensing in 1.4%, sinus tachycardia (ST) in 66%, SVT in 13%, slow (< 150 bpm) ventricular tachycardia (VT) in 8.7%, and VT or ventricular fibrillation (VF) in 10.3%. The sensitivity of slow VT detection was 94%, and of VT/VF detection 99.3%. The specificity of sinus rhythm/ST/SVT recognition was 94%, positive predictive value 79.3%, and negative predictive value 99.2%. A total of 1,918 episodes were treated in 330 patients: 1,472 appropriately in 213, and 446 inappropriately in 117 (15% of the overall population) patients. Only 62 episodes were inappropriately treated by shocks in 40 patients, representing 5% of the overall population.
Conclusions: In this conventional ICD population, the overall specificity of standard dual-chamber arrhythmia detection settings reached 94%. This feature allows efficient detection of fast as well as slow VT events with a very low rate of inappropriate shocks.