Michael R Gold1; Dominic A Theuns2; Bradley P Knight3; Lacy J Sturdivant4; Kenneth A Ellenbogen5; Mark A Wood5; Martin C Burke6

    1 Med Univ of South Carolina， Charleston， SC
    2 Erasmus Med Cntr， Rotterdam， Netherlands
    3 Univ of Chicago， Chicago， IL
    4 Med Univ of South Carolina， Charleston， SC
    5 Virginia Commonwealth Univ， Richmond， VA
    6 Univ of Chicago， Chicago， IL

    Introduction: The development of a totally subcutaneous implantable defibrillator (SQ) system requires a new approach for arrhythmia detection. The Subcutaneous vs. Transvenous Arrhythmia Recognition Testing (START) study was a prospective， multi-center trial designed to compare the performance of the SQ system with dual-chamber transvenous ICD (DC) systems for arrhythmia discrimination.

    Methods: At ICD implantation， induced ventricular and atrial arrhythmias were recorded simultaneously in transvenous and cutaneous configurations. Recordings of induced ventricular (n=46) and atrial arrhythmias (n=50) with ventricular rates >= 170 bpm from 64 patients were used to analyze detection offline by Medtronic (MDT)， Boston Scientific (BSC)， St. Jude Medical (SJM) and Cameron Health SQ systems. Sensitivity analysis of VT/VF was performed in single-zone (>= 170 bpm) and dual-zone configurations (VF >240 bpm; VT >= 170 bpm). Specificity analysis (atrial fibrillation (AF)， flutter or supraventricular tachycardia) was performed in a dual-zone configuration (VT zone: 170–240 bpm， VF zone: >240 bpm) in order to enable discrimination algorithms. Morphology-based algorithms in conjunction with interval-based discriminators (BSC and SJM) or a timing-based algorithm (MDT) were enabled at nominal settings with the exception of sustained rate timers (programmed OFF).

    Results: Detection accuracy of ventricular arrhythmias by the SQ system was comparable to DC systems. Sensitivity in single-zone configuration was 100% for all devices; for dual-zone configuration the sensitivity was 100% for S-ICD vs. 98.6% for DC (range: 95.7%–100%). Specificity was better for the SQ system compared to DC systems (98.0% SQ vs. 68.0% DC， range: 32.7%–89.8%; p<0.001). The majority of misclassifications were observed during AF. Specificity for AF was 97.5% (SQ) vs. 69.2% (DC); p<0.01.

    Conclusion: Ventricular arrhythmia detection sensitivity of the SQ system using subcutaneous signals is comparable to transvenous ICD systems; however， specificity of arrhythmia discrimination by the SQ system is better than morphology discrimination by DC systems. Moreover， atrial information in DC systems does not significantly improve specificity compared to the SQ system.