Ventricular stimulus site influences dynamic dispersion of repolarization in the intact human heart

Abstract

The spatial variation in restitution properties in relation to varying stimulus site is poorly defined. This study aimed to investigate the effect of varying stimulus site on apicobasal and transmural activation time (AT), action potential duration (APD) and repolarization time (RT) during restitution studies in the intact human heart. Ten patients with structurally normal hearts, undergoing clinical electrophysiology studies, were enrolled. Decapolar catheters were placed apex to base in the endocardial right ventricle (RVendo) and left ventricle (LVendo), and an LV branch of the coronary sinus (LVepi) for transmural recording. S1-S2 restitution protocols were performed pacing RVendo apex, LVendo base, and LVepi base. Overall, 725 restitution curves were analyzed, 74% of slopes had a maximum slope of activation recovery interval (ARI) restitution (Smax) > 1 (P < 0.001); mean Smax = 1.76. APD was shorter in the LVepi compared with LVendo, regardless of pacing site (30-ms difference during RVendo pacing, 25-ms during LVendo, and 48-ms during LVepi; 50th quantile, P < 0.01). Basal LVepi pacing resulted in a significant transmural gradient of RT (77 ms, 50th quantile: P < 0.01), due to loss of negative transmural AT-APD coupling (mean slope 0.63 ± 0.3). No significant transmural gradient in RT was demonstrated during endocardial RV or LV pacing, with preserved negative transmural AT-APD coupling (mean slope -1.36 ± 1.9 and -0.71 ± 0.4, respectively). Steep ARI restitution slopes predominate in the normal ventricle and dynamic ARI; RT gradients exist that are modulated by the site of activation. Epicardial stimulation to initiate ventricular activation promotes significant transmural gradients of repolarization that could be proarrhythmic.

Keywords: APD heterogeneity; dispersion of repolarization; human; restitution; whole heart.

Fig. 1.

Orientation of catheters in the heart for recording. A: schematic showing positioning of catheters in an apicobasal orientation in the left ventricle (LV) and right ventricle (RV) endocardium (endo), and transmurally across the lateral base of the LV epicardium (EPI) via the CS, with corresponding unipolar electrograms recorded. B: catheter positions were checked via fluoroscopy to ensure adequate apicobasal and transmural apposition. RA, right atrium; His, bundle of His.

Fig. 2.

Analysis of cardiac restitution studies. A: example of simultaneous transmural unipolar electrogram recording at the LV_endo base and LV_epi base during LV endocardial restitution. Activation moments (circles) and repolarization moments (inverted triangles) are shown. B: the Wyatt method was used to determine activation time (AT) and repolarization time (RT), with activation recovery interval (ARI) taken as a surrogate marker of APD. UEG, unipolar electrogram; au, arbitrary units. C: example of an ARI restitution curve of one patient, from a series of catheter poles in the RV. Each pole is represented by a different color.

Fig. 3.

Boxplots of S_max for all patients, with S_max on y-axis and recording region on the x-axis. A: all recorded S_max in the endocardium (Endo) of the right ventricle (RV), left ventricle (LV), and epicardium of the left ventricle (Epi), regardless of pacing site. B, C, and D: S_max for RV_endo, LV_endo, and LV_endo pacing, respectively. The median S_max was >1 in all instances. Bars above the graphs represent ANOVA comparisons of statistical difference between regional S_max, based on pacing site. *P < 0.05 for regional S_max >1, based on one-sampled T-test.

Fig. 4.

Lowes regression of activation time (AT) restitution in all patients, with 95% confidence interval. Pacing site for each row is shown: endocardium of the right ventricle (RV_endo; A–C), endocardium of the left ventricle (LV_endo; D–F), and LV_epi (G–I). Comparisons are made between apex and base in the RV_endo (A, D, and G), LV_endo (B, E, and H), and transmurally across the LV base (C, F, and I), represented by each column heading (top). Significant differences were assumed if quantile regression was <0.05 across the whole of the restitution curve, and statistical significance is shown inset within the graphs. Arrows with time in miliseconds (ms) represent difference at the 50th quantile, where curves were significantly different. ns, No significance; CL, cycle length.

Fig. 5.

Lowes regression of activation recovery interval (ARI) restitution in all patients, with 95% confidence interval. Pacing site for each row is shown: endocardium of the right ventricle (RV_endo; A–C), endocardium of the left ventricle (LV_endo; D–F), and LV_epi (G–I). Comparisons are made between apex and base in the RV_endo (A, D, and G), LV_endo (B, E, and H), and transmurally across the left ventricular (LV) base (C, F, and I), represented by each column heading (top). Significant differences were assumed if quantile regression was <0.05 across the whole of the restitution curve, and statistical significance shown inset within the graphs. Arrows with time in miliseconds (ms) represent difference at the 50th quantile, where curves were significantly different. There is a consistent transmural difference in ARI, regardless of pacing site (C, F, and I), with a significant LV apicobasal difference in ARI on LV_endo pacing only (E). DI, diastolic interval; ns, no significance.

Fig. 6.

Lowes regression of repolarization time (RT) restitution in all patients, with 95% confidence interval. Pacing site for each row is shown: endocardium of the right ventricle (RV_endo; A–C), endocardium of the left ventricle (LV_endo; D–F), and LV_epi (G–I). Comparisons are made between apex and base in the RV_endo (A, D, and G), LV_endo (B, E, and H), and transmurally across the left ventricular (LV) base (C, F, and I), represented by each column heading (top). Significant differences were assumed if quantile regression was <0.05 across the whole of the restitution curve, and statistical significance is shown inset within the graphs. Arrows with time in milliseconds (ms) represent difference at the 50th quantile, where curves were significantly different. There is a significant apicobasal gradient in RT on RV and LV pacing (A and E), while CS pacing creates a significant transmural gradient in RT (I), as well as an LV apicobasal gradient in RT (H). CL, cycle length; ns, no significance.