DIFFERENTIATION OF SINUS TACHYCARDIA FROM VENTRICULAR-TACHYCARDIA WITH 1 1-VENTRICULOATRIAL CONDUCTION IN DUAL-CHAMBER IMPLANTABLE CARDIOVERTER-DEFIBRILLATORS - FEASIBILITY OF A CRITERION BASED ON THE ATRIOVENTRICULAR INTERVAL/

Tachycardia discrimination in future implantable cardioverter defibril lators (ICDs) is likely to be enhanced by the addition of an atrial se nsing/pacing lead. However, differentiation of sinus tachycardia (ST) from ventricular tachycardia (VT) with 1:1 VA conduction will remain p roblematic. we assessed the use of the AV interval as a potential crit erion for correctly differentiating ST from VT. Incremental V pacing a t the right ventricular (RV) apex served as a ''VT'' model in each of 41 patients with 1:1 VA conduction to pacing cycle lengths less than o r equal to 450 msec. High right atrial and RV apical electrograms duri ng normal sinus rhythm (NSR) and during incremental V pacing were digi tized (simulating ICD sensing). From these signals, AV interval versus pacing cycle length plots were computer generated to identify crossov er cycle lengths, each defined as the cycle length at which the AV int erval during V pacing equals the AV interval during NSR. At cycle leng ths longer than the crossover value, the AV interval during ''VT'' exc eeds the AV interval during NSR. In contrast, the AV interval during S T is physiologically shorter than the AV interval during NSR. Thus, ST can be readily differentiated from ''VT'' over a range of cycle lengt hs greater than the crossover value. The overall mean calculated cross over cycle length was 371 +/- 52 msec. In 11 patients paced multiple t imes, each crossover cycle length was reproducible (mean coefficient o f variation was 1.2% +/- 0.9% per patient). AV intervals measured at t he RV apex were also analyzed with incremental V pacing during catecho lamine stimulation (isoproterenol, n = 5) and during alternate site '' VT'' (RV outflow tract [n = 8] and left ventricle [n = 2]). In all the se cases, the new ''VT'' plots of AV interval versus pacing cycle leng th coincided with or fell to the left of those obtained during control RV apical pacing and recording (i.e., these AV interval values crosse d the NSR baseline at cycle lengths less than or equal to the crossove r cycle length). Thus, the cycle length range for recognizable differe ntiation of ST from ''VT'' remained valid. The data suggest that the d escribed AV interval criterion relying on the crossover cycle length: (1) is a promising approach to improve differentiation of ST from rela tively slow VTs with 1:1 VA conduction, and (2) can readily be automat ed in future dual chamber ICDs, given ifs computational simplicity.