Cycle length-associated modulation of the regional dispersion of ventricular repolarization in a canine model of long QT syndrome.

Previous tridimensional activation mapping showed that the development of functional conduction block at the onset of torsades de pointes was regionally heterogeneous; conduction block was frequently observed in the LV and the interventricular septum (IVS) but not in the RV, in the canine anthopleurin-A (AP-A) model of long QT syndrome (LQTS). This may be related to the distribution of myocytes with M celllike electrophysiological characteristics. To better understand the regional difference of arrhythmogenicity in LQTS, the authors investigated cycle length related modulation of ventricular repolarization among three different layers: the endocardium (End), mid-myocardium (Mid), and epicardium (Epi) of the LV and RV and at two different areas: the Epi and septum (Sep) in the IVS. The LQT3 model was produced by AP-A in dogs. Using constant pacing and single premature stimulation (S1S2), the ventricular repolarization pattern was analyzed from 256 unipolar electrograms. Activation-recovery intervals (ARIs) were used to estimate local repolarization. In seven experiments, AP-A increased regional ARI dispersion to 88.1 +/- 36.0 ms in the LV, to 72.9 +/- 35.7 ms in the IVS, and to 23.0 +/- 8.7 ms in the RV at the pacing cycle length (CL) of 1,000 ms. Development of the large ARI dispersion was due to greater ARI prolongation at the Mid site in the LV and at Sep site in the IVS. As the S1S2 interval was shortened, regional ARI dispersion decreased gradually, and finally, ARI dispersion showed a reversal gradient of repolarization between the Mid and Epi sites in the LV and between the Sep and Epi sites in the IVS. Two factors contributed to create the reversal gradient of repolarization: (1) a difference in restitution kinetics at the Mid site in the LV and at the Sep site in the IVS, characterized by a larger delta ARI and slower time constant (tau), and (2) a difference in diastolic intervals at each site resulting in different input to restitution at the same CL. However, the RV showed small alteration in the transmural dispersion of repolarization in the S1S2 protocol. S2 created heterogeneous functional conduction block in the LV and IVS but not in the RV. In the LQT3 model, the arrhythmogenicity of torsades de pointes is primarily due to dispersion of repolarization in the LV and IVS because of prominent distribution of M cells. The RV seems to participate passively in reentrant excitation during torsades de pointes.