Comprehensive T wave morphology assessment in a randomized clinical study of dofetilide, quinidine, ranolazine, and verapamil

Abstract

Background: Congenital long QT syndrome type 2 (abnormal hERG potassium channel) patients can develop flat, asymmetric, and notched T waves. Similar observations have been made with a limited number of hERG-blocking drugs. However, it is not known how additional calcium or late sodium block, that can decrease torsade risk, affects T wave morphology.

Methods and results: Twenty-two healthy subjects received a single dose of a pure hERG blocker (dofetilide) and 3 drugs that also block calcium or sodium (quinidine, ranolazine, and verapamil) as part of a 5-period, placebo-controlled cross-over trial. At pre-dose and 15 time-points post-dose, ECGs and plasma drug concentration were assessed. Patch clamp experiments were performed to assess block of hERG, calcium (L-type) and late sodium currents for each drug. Pure hERG block (dofetilide) and strong hERG block with lesser calcium and late sodium block (quinidine) caused substantial T wave morphology changes (P<0.001). Strong late sodium current and hERG block (ranolazine) still caused T wave morphology changes (P<0.01). Strong calcium and hERG block (verapamil) did not cause T wave morphology changes. At equivalent QTc prolongation, multichannel blockers (quinidine and ranolazine) caused equal or greater T wave morphology changes compared with pure hERG block (dofetilide).

Conclusions: T wave morphology changes are directly related to amount of hERG block; however, with quinidine and ranolazine, multichannel block did not prevent T wave morphology changes. A combined approach of assessing multiple ion channels, along with ECG intervals and T wave morphology may provide the greatest insight into drug-ion channel interactions and torsade de pointes risk.

Clinical trial registration: URL: https://ichgcp.net/clinical-trials-registry/NCT01873950" title="See in ClinicalTrials.gov">NCT01873950.

Keywords: electrocardiography; ion channels; long‐QT syndrome; pharmacology; torsade de pointes.

© 2015 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley Blackwell.

Figures

Figure 1.

Illustration showing ECG biomarkers assessed in this study: (A) QT and subintervals (QRS, J‐Tpeak and Tpeak‐Tend); (B) flat, asymmetric and notched T waves (solid lines) vs. normal T waves (dotted lines); (C) 30% of early (ERD30%) and late (LRD30%) repolarization duration in the preferential plane formed by the two‐first Eigen leads from principal component analysis (PCA1 and PCA2); (D) other vectorcardiographic biomarkers (QRS‐T angle, ventricular gradient and maximum magnitude of the T vector). See methods for further details.

Figure 2.

Patch clamp experiment results for (A) dofetilide, (B) quinidine, (C) ranolazine, and (D) verapamil for hERG (red dots), calcium (green triangles), and late sodium (blue squares). The lines in each plot correspond to a fit between the observed means per concentration and data (see text). The error bars denote±SEM. The dashed lines in each panel correspond to the range of observed clinical concentrations and the solid line is the population average maximum concentration (Cmax). hERG indicates hERG potassium channel; Calcium, L‐type calcium current; Late sodium, late sodium current.

Figure 3.

Mean baseline‐ and placebo‐corrected plasma drug concentration‐dependent changes (lines) and 95% confidence intervals (shaded areas) for T wave (A) flatness, (B) asymmetry, (C) probability of presence of notch, (D) maximum magnitude of T vector, (E) 30% of early repolarization duration, (F) 30% of late repolarization duration, (G) angle between the mean QRS and T vectors (QRS‐T angle), (H) total cosine of R to T, (I) spatial ventricular gradient, (J) Fridericia's heart rate corrected QT (QTc) interval, (K) early repolarization interval (J‐Tpeakc), and (L) late repolarization interval (Tpeak‐Tend). x axis shows % of population Cmax for each drug (mean±SD): dofetilide (2.7±0.3 ng/mL, solid gray), quinidine (1.8±0.4 μg/mL, dotted orange line), ranolazine (2.3±1.4 μg/mL, blue dashed line) and verapamil (130.4±75.8 ng/mL, dashed green line). See Table 3 for detailed drug concentration ranges. Mean slopes together with 95% confidence interval values are reported in Table S1. Figures S3 through S5 show similar information in a drug‐by‐drug basis together with relative block of hERG, calcium, and late sodium currents from the patch clamp experiments. Corresponding time profile plots of drug plasma concentration and time‐matched drug‐induced changes are shown in Figures S6 through S8.

Figure 4.

Lead V3 of median beats of time‐matched placebo (gray), dofetilide (black), quinidine (orange), ranolazine (blue), and verapamil (green). Top panel (A): ECGs at 0.5, 2.5, 4, and 12 hours post‐dose from a woman with notches present in dofetilide and quinidine but not in ranolazine. Bottom panel (B): ECGs at 1, 3, 6, and 12 hours post‐dose from a woman with notches induced by ranolazine. See Table S2 for heart rate and QT/QTc measurements. ECG indicates electrocardiogram.

Figure 5.

Relationship between predicted drug‐induced baseline‐ and placebo‐corrected changes in QTc (x axis) and T wave (A) flatness, (B) asymmetry, (C) probability of presence of notch, (D) maximum magnitude of T vector, (E) 30% of early repolarization duration, (F) 30% of late repolarization duration, (G) angle between the mean QRS and T vectors (QRS‐T angle), (H) total cosine of R to T, (I) spatial ventricular gradient, (J) early repolarization interval (J‐Tpeakc), and (K) late repolarization interval (Tpeak‐Tend) (y axis). Average predictions (dots) and 95% confidence intervals (horizontal and vertical lines) from concentration dependent models for QTc (x axis) and the different T wave morphology biomarkers (y axis) at baseline and 25% increments of population's Cmax for dofetilide (dashed gray line) quinidine (dotted orange line) and ranolazine (solid blue line). QTc, Fridericia's heart rate corrected QT.