Jcp325927.qxd

Baseline Values and Sotalol-Induced Changes of Ventricular Repolarization Duration,
Heterogeneity, and Instability in Patients With a History of Drug-Induced Torsades de Pointes
Jean-Philippe Couderc, Stefan Kaab, Martin Hinterseer, Scott McNitt, Xiaojuan Xia, Anthony Fossa, Britt M.
Beckmann, Slava Polonsky and Wojciech Zareba 2009; 49; 6 originally published online Oct 28, 2008; The online version of this article can be found at: http://www.jclinpharm.org/cgi/content/abstract/49/1/6 Additional services and information for
can be found at:
The Journal of Clinical Pharmacology
CLINICAL STUDIES
Baseline Values and Sotalol-Induced Changes of
Ventricular Repolarization Duration,
Heterogeneity, and Instability in Patients With a
History of Drug-Induced Torsades de Pointes
Jean-Philippe Couderc, PhD, MBA, Stefan Kaab, MD, PhD, Martin Hinterseer, MD, Scott McNitt, MS, Xiaojuan Xia, MS, Anthony Fossa, PhD, Britt M. Beckmann, MD, Slava Polonsky, MS, and Wojciech Zareba, MD, PhD, FACC The authors investigated whether computerized parame- patients with a history of Torsades de Pointes had a longer ters quantifying ventricular repolarization delay, hetero- QTc and an increased repolarization heterogeneity of geneity, and instability characterize individuals who : 44 ± 13 vs 35 ± 8 ms, developed drug-induced Torsades de Pointes. Assessing P = .02). On sotalol, the electrocardiograms from individ- an individual’s propensity to Torsades de Pointes when uals with Torsades de Pointes revealed a delay of the ter- exposed to a QT-prolonging drug is challenging because minal part of the T wave that was not present in patients baseline QT prolongation has limited predictive value. without Torsades de Pointes (TpTe: 27 ± 40 vs –2 ± 21 ms, Five-minute digital 12-lead electrocardiograms were : 20 ± 29 vs 2 ± 4 ms, P = .04). Results sug- acquired at baseline and after a sotalol challenge in 16 gest that the electrocardiogram abnormalities characteriz- patients who had a history of Torsades de Pointes in the ing patients with a history of Torsades de Pointes are (1) context of a QT-prolonging drug and 17 patients who did an increased repolarization heterogeneity at baseline and not have such history. Computerized measurements of QTc, (2) a sotalol-induced prolongation of the terminal part of T peak to T end intervals (TpTe), TpTe/QTc, and QT vari- ability were implemented, and novel quantifiers of ventric-ular repolarization heterogeneity from the early (ERD) and Keywords: QT interval; Torsades de Pointes; electrocar-
late (LRD) part of the T wave were investigated. Compared with electrocardiograms of patients without a history of Journal of Clinical Pharmacology, 2009;49:6-16
Torsades de Pointes, the baseline electrocardiograms of 2009 the American College of Clinical Pharmacology
From the Heart Research Follow-Up Program, Cardiology Department,University of Rochester Medical Center, Rochester, New York (Dr Couderc, Mr Drug-induced Torsades de Pointes (TdP) have been
associated with an increasing number of cardiac McNitt, Mr Xia, Mr Polonsky, Dr Zareba); Ludwig-Maximilians-University, and noncardiac marketed drugs commonly affecting Munich, Klinikum Grosshadern, Department of Medicine 1, Munich, the rapid components of the delayed rectifier potas- Germany (Dr Kaab, Dr Hinterseer, Dr Beckmann); and iCardiac sium current (I ) of the myocardial cells. However, if Technologies Inc, Rochester, New York (Dr Fossa). Submitted for publication I inhibition and QT interval prolongation are asso- April 2, 2008; revised version accepted September 1, 2008. Address for cor- ciated with the occurrence of drug-induced TdP, respondence: Jean-Philippe Couderc, PhD, MBA, Box 653, Heart Research they should not constitute a reason for considering a Follow-Up Program, Cardiology Department, University of Rochester drug to be proarrhythmic.1 The dissociation between Medical Center, Rochester, NY 14642; e-mail: jean-philippe.couderc@heart.rochester.edu.
drug-induced QT interval prolongation2 and an increased risk of arrhythmias is supported by the 6 • J Clin Pharmacol 2009;49:6-16
VENTRICULAR REPOLARIZATION AND HISTORY OF TORSADES DE POINTES existence of drugs associated with a QT interval pro- documented TdP in the context of a drug with QT- longation but a limited history of TdP such as tamox- prolonging potential: sotalol, sumatriptan, amiodarone, ifen,3 carvedilol,4 and, more recently, ranolazine, bisacodyl, cipramil, furosemide, clarithromycin, which seems to prolong the QT interval duration while erythromycin, or roxythromicin. The patients were reducing ventricular heterogeneity.5 Consequently, the enrolled for an evaluation of the individual level of rejection of a novel drug because of its QT-prolong- repolarization reserve, and all patients signed informed ing effect is a rather dubious strategy because it consent to receive doses of sotalol as described pre- might deprive patients of a valuable medication.
viously.12 They were all genetically tested for the pres- Improving the assessment of drug cardiotoxicity ence of a mutation of the major LQTS genes using linked to the ventricular repolarization process standard genotyping techniques (genomic DNA was could depend on the development of better electro- prepared from lymphocytes, and amplification of cardiographic markers than QT prolongation.
KCNQ1, KCNH2, KCNE1, KCNE2, and SCN5A using The triggering mechanism(s) of drug-induced TdP polymerase chain reactions was performed, followed remains to be elucidated, but there are several inter- by direct sequencing of these major LQT-disease esting alternatives currently proposed: Hondeghem genes). The control group consisted of patients who et al suggested the TRiaD concept, emphasizing the were started on sotalol for the prevention of parox- role of action potential triangulation, reverse use ysmal atrial fibrillation and had given informed dependence of the drug, and repolarization instabil- ity.6,7 The triangulation of the action potential andthe heterogeneity of electrical properties of the cells Study Protocol and Electrocardiogram Recordings
across the myocardium are consistent with theproarrhythmic factors described by Belardinelli The study protocol was described by Kaab et al.12 et al8: the transmural dispersion of repolarization and Briefly, dl-sotalol was given intravenously at a constant the promoting role of early after-depolarization.9,10 rate over a 20-minute interval at a dose of 2 mg/kg body Finally, the concept of repolarization reserve described weight in 50 mL of a 0.9% saline solution in a group of by Roden11 emphasizes the role of the interplay individuals with (+TdP) and without (-TdP) a history of ion currents involved in cardiac repolarization.
of drug-induced TdP. Tests were performed in the These currents provide functional redundancy, or morning. Sotalol was injected to unmask latent repo- “reserve,” and can protect an individual against larization abnormalities while patients were closely excessive QT prolongation by drugs. Also, gender, and continuously monitored in the intensive care unit.
hypokalemia, predisposing DNA polymorphism, Continuous 5-minute surface 12-lead ECG recordings and environmental factors are recognized to be (Mortara Instrument, Milwaukee, Wisconsin) were potential modulators of the ventricular repolariza- acquired at rest in the supine position at baseline and tion process. They can lead to a reduced repolariza- at 20-minute steady-state phase after injection. We tion reserve and an increased propensity to obtained access to 2 ECG tracings per individual at baseline and on peak concentration of the drug.
In this study, we hypothesize that patients with a The measurements of the PR and QRS durations history of drug-induced TdP have a certain level of from the 5-minute ECGs were provided by the repolarization impairment (heterogeneity, reduced Mortara SuperECG software (SuperECG, Mortara repolarization reserve, and instability) that can be Instrument). The RR intervals and repolarization measured from their digital surface electrocardio- intervals were based on technology developed at the gram (ECG). Increased QT duration, repolarization University of Rochester Medical Center (Rochester, heterogeneity, and QT variability are investigated at New York). The COMPAS software provided the loca- baseline and when the patients are exposed to a tor- tion of the end of the T wave based on a technique identifying the crossing point between the baselineand the descending slope of the T wave (least squares technique).13 The apex of the T wave relied on amethod using a parabola fit of the T wave where the Study Population
maximum of the parabola identified the location ofthe apex. Baseline wandering was adjusted using The patients were enrolled after being admitted to Spline interpolation.13 The amplitude of the T wave the University Hospital of Munich, Germany, for was measured at the apex of the T wave.14 Description of the vectocardiographic measurements of the early (ERD) and late (LRD) repolarization measurement. The , with 30% representing the threshold used for identifying the length of these intervals from the T loop (upper panel) and its corresponding intervals on the first eigenvector (lower panel). ERD encompasses an interval spreading towardthe QRS complex, whereas LRD encompasses an interval toward the end of the T-wave. QT Interval Measurements From the
repolarization signal, and (2) it is independent from Scalar Electrocardiograms
the determination of the end of the T wave. Suchapproach requires that the patients remain at rest The computer-based end of the T wave was visually during the ECG recording to avoid high heart rates checked by trained technicians and manually (ie, short RR intervals in which the T wave would be adjusted using the on-screen caliper available in the encompassed 220 ms prior to the next R peak).
COMPAS software, if the automatic algorithm failed The method is based on the singular value to correctly identify the end of the T wave (semi- decomposition (SVD) of the RIs from the 12-lead computerized method). The QT interval measure- signals. SVD is used to reduce the dimension of ments were done in 3 cardiac beats in sinus rhythm the ECG lead systems from 12 leads to 2 leads.16 from lead V5 (or II), and the median value from these We refer to the resulting 2 leads as the eigenvectors 3 measures was computed. We report both the scalar 1 (ev ) and 2 (ev ). We measured the QT, QTapex, computerized and semi-computerized QT interval and the TpTe intervals (TpTe = QT – QTapex) from measurements expressed in milliseconds. Recently, ev . The apex and the end of the T wave were iden- Liu et al9 reported the presence of an increased T tified in a fully computerized manner using the peak to T end interval (TpTe)/QT ratio prior to the method described above for scalar measurements.
development of TdP in a rabbit wedge preparation.
We called the T loop the representation of the RI in This novel parameter was included in our analysis the 2-dimensional space, defined by ev and ev to investigate its interest when measured from (upper panel of Figure 1). The vector describing the T loop path across time is the repolarizationvector.
T Loop Measurements
The early repolarization duration (ERD) and the late repolarization duration (LRD) are measurements In our analysis, the repolarization interval (RI) is of interval duration based on the T loop. The start- defined between the J point and the point located ing point of these intervals is the time at which the 220 ms before the next R peak. The determination of length of the repolarization vector is maximized the J point was based on an algorithm developed in the upper panel of Figure 1). The ending by Zong at al.15 This ensures that (1) the analysis point of these intervals is delimited by a circle of encompasses all components of the ventricular 8 • J Clin Pharmacol 2009;49:6-16
VENTRICULAR REPOLARIZATION AND HISTORY OF TORSADES DE POINTES (Figure 1). Consequently, these parameters mea- Statistical Analysis
sure the time needed for the heart vector to varyfrom its maximum length to a time point corre- Differences between groups were expressed as mean sponding to a 30% reduction of its maximum length ± standard deviation. The analysis of correlation between values of various parameters was based on sure toward the end of the RI, and ERD is directed Spearman rank correlation, and we report its associ- toward the J point (see Figure 1, lower panel). The ated coefficients (ρ). P values less than or equal to duration of these time intervals increases when the .05 were considered statistically significant. We heart vector slows down or/and the roundness of the used logistic regression models to describe the asso- T loop increases. Consequently, these parameters ciation between baseline ECG measures and the measure repolarization duration (reflected in the level of drug-induced ECG changes. When we inves- velocity of the heart vector) and the repolarization tigated the presence of a history of TdP as the pri- heterogeneity (reflected in the path of the heart vec- mary endpoint, binary logistic regression models were used, and both the best subsets regression pro-cedure and the stepwise procedure were used to Heart Rate Correction
select the optimal models. The statistical analyseswere done using SAS (SAS Institute, Cary, North All repolarization measurements were heart rate cor- rected using the pooled technique. A linear regres-sion analysis was used to model the relationship between repolarization measurements and RR inter-vals during baseline periods. The slope (β) charac- Study Population
terizing this relationship was used to correct therepolarization measurements such as QTc = QT + The clinical characteristics of the study population β(1 – RR) for the QT interval. The same heart rate are provided in Table I. The average ages of the pop- correction technique was applied to all other ulations were not significantly different between the 2 groups: 59 ± 13 years versus 61 ± 12 years. Thenumber of women was slightly higher (n = 12) in QT Variability
the group of individuals without TdP than in thegroup with TdP (n = 9). Presence of a history of The instability of the repolarization was estimated myocardial infarction, coronary artery disease, and using the median absolute deviation (MAD) of the hypertension was similar between the study groups.
beat-to-beat measurements of the semi-computerized There were several patients with a history of atrial QT and QTapex parameters after heart rate correc- fibrillation in both groups (+TdP: n = 11 and –TdP: tion based on the pooled formula. To further control n = 17, P < .05). One of the patients had atrial fibril- for the effect of RR variation, we divided these MAD lation during the ECG recording. This ECG was values by the MAD of the RR intervals (MAD removed from the analysis, resulting in a group of 16 patients with a history of TdP and 17 individualsfree of such history.
Heart Rate Variability
The group of individuals with a history of drug- induced TdP has been reported using a heteroge- The heart rate variability (HRV) was estimated neous list of medications. Seven of the patients had from the 5-minute recordings using an autoregres- sotalol-induced TdP. None of the patients in the sive method. The normalized high- (HFnorm) study experienced episodes of TdP during the sotalol and low-frequency (LFnorm) components, expre- challenge, and none of them carried a mutation ssed in percentages, were computed using the linked to the major congenital forms of the LQTS.
SuperECG software (Mortara Instrument). The def- Tables II and III provide the ECG-based parame- inition of the frequency bands for the HF and LF ters across populations for the baseline recordings components was recommended by the European and for the sotalol-induced changes, respectively. PR Task Force.17 The standard deviation from normal- and QRS durations were not significantly different to-normal intervals (SDNN) was also computed in between groups at baseline and after drug. As shown in Table III, the RR intervals were significantly results using the semi- or fully computerized method in baseline conditions (Table II). The groupof +TdP patients had a longer QT interval duration Drug Inducing TdP
(~25 ms) than –TdP patients. We identified 6 patientswith a QTc duration above the gender-specific Group With History of Torsades de Pointes
QTc >470 ms in men). One of them was a woman from the control group (-TdP); the remaining ones were from the +TdP group and included 3 men and Vectorial measurements. At baseline, the vectorial QT measurements (from ev ) were slightly longer than the scalar QT intervals, but the difference between the study groups remained consistent (26 ms). According to the vectorial parameters, this pro- longation was localized within the early part of the (P = .02), and this prolongation reached 14 ms with (P = .03). Interestingly, this delay in the early phase of the repolarization segment was not cap- Group Without History of Torsades de Pointes
tured by the QTapex interval (from ev ), suggesting that the morphology of the T loop (ventricular het- erogeneity) primarily drives this delay.
Our investigation of QT variability reveals a trend toward larger variability in baseline ECGs of +TdP patients, but this difference did not reach statistical Sotalol-Induced Prolongation, Heterogeneity,
and Instability of Repolarization
Scalar measurements. Sotalol is associated with strong prolongation of the QTc interval duration, and this was true for the 2 groups (+TdP: 85 ± 42 and –TdP: 65 ± 47 ms). These changes were statistically different from zero (P < .0001) but not statisticallydifferent between groups (P = .22) when considering Y/N, yes/no; EF, left ventricular ejection fraction; CAD, coronary artery single lead-based measurements. Similar results disease; MI, myocardial infarction; HT, history of hypertension; AF,history of atrial fibrillation; TdP, Torsades de Pointes.
were found using the scalar computerized technique(+TdP: 63 ± 57 and –TdP: 56 ± 41 ms, P = .70).
The ratio of the terminal part of the T wave to the QTc interval was not significantly different between longer after sotalol (+TdP: 201 ± 101 ms and -TdP: 175 ± 98 ms, P < .05), but the bradycardic effect ofthe drug was not different between groups (P = .45).
Vectorial measurements. There was a significantsotalol-induced QTc and QTc apex prolongation (P < Baseline Duration, Heterogeneity,
.01) within the 2 study groups. It is noteworthy that and Instability of Repolarization
QTc measured from ev did reveal statistically sig- nificant prolongation in the +TdP group (75 ± 44 vs Scalar measurements. In comparison to the scalar 37 ± 26 ms, P = .008). This observation is consistent QTc interval measurements, we found similar with Kaab and coworkers’ results12 evidencing 10 • J Clin Pharmacol 2009;49:6-16
VENTRICULAR REPOLARIZATION AND HISTORY OF TORSADES DE POINTES Description of Baseline Values of Electrocardiographic Parameters Baseline (Absolute Values)
−TdP (n = 17)
+TdP (n = 16)
P Values
-TdP, patients without a history of TdP; +TdP, patients with a history of TdP; TpTe, T peak to T end interval in lead II; ERD, early repolarization dura-tion; LRD, late repolarization duration (for definition of the ERD and LRD parameters, see text); HF, high frequency; LF, low frequency; MAD, medianabsolute deviation; HRV, heart rate variability; SDNN, standard deviation from normal-to-normal intervals; TdP, Torsades de Pointes. Values associatedwith P < .05 are in bold. NU, no unit.
a. These measurements are corrected using the pooled formula and expressed in milliseconds.
significantly larger sotalol-induced QT prolongation levels of QT variability was found in ECGs recorded between patients with and without a history of TdP using the maximum QT interval from all availableleads. More interestingly, sotalol significantly pro- Characterizing Patients With a
longed the late part of the repolarization in the History of Torsades de Pointes
group of patients with a history of TdP: their TpTeinterval prolongation was longer (23 ± 27 vs 4 ± 12 Binary logistic regressions were implemented to find which baseline information could help predict III) values were more prolonged (20 ± 29 vs 2 ± 14 ms, the presence of a history of TdP in a multivariate fashion. The QTc, QTc apex, TpTe, TpTe/QTc, No significant changes in SDNN values and LF norm values were found after sotalol in any of the the design. Based on both stepwise and best subsets, study groups. But the high-frequency norm revealed a trend toward increased parasympathetic innervations predictor of a history of TdP. For each incremental 1- in the group of patients with a history of TdP (–TdP: 6.7% ± 11.7% vs +TdP: 14.5% ± 8.1%, P = .05).
odds of having a history of TdP (P = .016). The sec- The variability of the QTc and QTc apex interval durations, adjusted for heart rate, was measured ated with a 41% increase for each 0.1 increase in value (P = .066). Baseline QTc or TpTe intervals did Table III
Description of Sotalol-Induced Changes in Values of Electrocardiographic Parameters Sotalol Challenge (Sotalol-Induced Changes)
−TdP (n = 17)
+TdP (n = 16)
P Values
6.7 ± 11.7
14.5 ± 8.1*
-TdP, patients without a history of TdP; +TdP, patients with a history of TdP; TpTe, T peak to T end interval in lead II; ERD, early repolarization dura-tion; LRD, late repolarization duration (for definition of the ERD and LRD parameters, see text); HF, high frequency; LF, low frequency; MAD, medianabsolute deviation; HRV, heart rate variability; SDNN, standard deviation from normal-to-normal intervals; TdP, Torsades de Pointes. Values associatedwith P < .05 are in bold. Testing if the average is different from 0: *P < .01. NU, no unit.
a. These measurements are corrected using the pooled formula and expressed in milliseconds.
not contribute to the model despite the presence of None of the other ECG measurements entered the 5 patients with a prolonged QTc interval at baseline model—that is, neither baseline QT interval dura- tion nor baseline TpTe interval contributed to the A second logistic model was implemented consid- prediction of sotalol-induced TpTe prolongation.
ering the sotalol-induced TpTe interval prolongation On the basis of these results, we report in the as a primary continuous endpoint and baseline ECG upper panel of Figure 2 the scatterplots for ERD30% measurements as covariates. Again, both ERD and QT variability values in the 2 study groups at history of TdP, one can separate the 2 groups with was a 1.8-ms increment in TpTe interval value with a 100% specificity and 69% sensitivity. The 6 indi- strong statistical significance (P = .0002). A univariate viduals presented prolonged QT intervals on their baseline ECGs (LQTS) based on the following clini- induced prolongation of TpTe was significant (r2 = 31%, cal criterion: QTc >480 ms in women and QTc >470 ms in men (see Figure 2). One may note that 7 patients without clinically identifiable LQTS were detected by our novel parameters. Among them, 3 in the TpTe interval on sotalol (P = .01).
patients have borderline QTc (450 ms > QTC >470 ms), 12 • J Clin Pharmacol 2009;49:6-16
VENTRICULAR REPOLARIZATION AND HISTORY OF TORSADES DE POINTES Distribution of heart rate–corrected values for ERD and the values of QT variability (upper panel) in baseline electrocar- diograms (ECGs) for our study populations marking the patients with long QT interval duration (based on gender-specific criteria of thelong QT syndrome). The sotalol-induced changes of TpTe and QTc interval (computerized) are reported in the lower panel. The grayareas in both panels represent the range of values in which most patients without history of Torsades de Pointes (TdP) are located. Thevertical and horizontal lines define the values of the parameter thresholds. and 4 have normal QTc interval duration.
TdP had a specific repolarization profile similar to Consequently, using a criterion based on QTc >450 the one we observed in ECGs of healthy participants ms to identify patients with increased risk would on moxifloxacin—namely, changes in morphology provide a sensitivity of 50% and a specificity of of the T wave prior to the T wave apex.16 It is note- 76%. This univariate analysis confirms our observa- worthy that the T wave from LQT2 patients with tions from the multivariate analysis that our novel borderline QTc interval duration (390-440 ms) also ECG parameters are bringing complementary infor- shows an abnormal early portion of the T wave (quantified using the left slope of the T wave). This Figure 2 also provides the scatterplot of values of information helps to better identify patients carrying sotalol-induced changes for TpTe and QTc intervals.
the KCNH2 mutation from noncarrier family We defined thresholds for maximizing the separa- members.14 In this study group, patients did not tion between the 2 groups as the sotalol-induced carry any of the major LQTS mutations, but our changes in TpTe >50 ms and in QTc >83 ms, and the observations might also reveal the presence of a groups can be separated with 94% specificity and reduced repolarization reserve. Indeed, unrecog- 75% sensitivity based on these thresholds.
nized repolarization modulators could be presentsuch as nondocumented drugs, underlying cardiac DISCUSSION
disease, and predisposing genetic factors (nonidenti-fied congenital long QT syndrome).
We report the analysis of ventricular repolarization When patients with a history of TdP are exposed duration, heterogeneity, and instability in a group of to the torsadogenic compound sotalol, the repolar- individuals with and without a history of drug- ization abnormalities are not limited to the early part induced TdP. We investigated ECG abnormalities of the T wave but extend to the late portion of the T that could be linked to arrhythmogenic factors con- wave. Our results show that the late part of the T tributing to trigger and maintain drug-induced TdP.
wave, measured either by the TpTe interval or the Repolarization heterogeneity and instability were parameters, is more significantly prolonged assessed from surface ECGs based on the T loop during an infusion of sotalol in patients with a morphology (ERD and LRD parameters) and QT vari- history of TdP than in patients without such history.
ability. The assessment of reverse use dependency of In human studies, Smetana et al22 investigated the sotalol using QT/RR modeling was not included in TpTe interval duration in the European Myocardial our analysis because the ECG recordings were too Infarction Amiodarone Trial (EMIAT) population, short to reliably assess the QT/RR relationship.18 comparing the length of this interval between The baseline ECGs of patients with a history of patients who died or did not die of cardiac arrhyth- TdP revealed more pronounced repolarization mic events. The results suggested a significant pro- abnormalities in comparison to the ECGs of patients longation of the TpTe interval in patients who died without such history. This increased repolarization in the placebo group (71 ± 3 vs 66 ± 1 ms, P = .04).
variability measured from baseline ECGs was not Interestingly, this difference was not found in the significantly different between the 2 study groups, group of patients on amiodarone (both groups had but the multivariate analysis suggested that this long TpTe intervals of 79 ± 6 vs 73.2 ± 2 ms, P = .17).
variability contributed to better classify these The ratio between TpTe and QT interval was not groups. Such observation is consistent with the beat- longer in patients with a history of TdP. This obser- to-beat variability of QT described in the study vation is not consistent with Liu et al’s work9 report- reported by Hinterseer et al.19 This instability of ing increased values of this ratio prior to the repolarization is 1 of 3 components of the TriAD occurrence of TdP in a rabbit model.
concept, and its proarrhythmic role has been docu- Our study suggests that the prolongation of the mented in several clinical studies that have reported QT intervals at baseline and on sotalol in patients their independent predicting value for appropriate with a history of TdP is associated with an unevenly implantable cardio-defibrillator therapy in postin- distributed delay across the repolarization interval.
These observations fit the arrhythmogenic concept, At baseline, our investigation revealed that the enhancing the role of the TpTe interval prolongation repolarization delay was prominently located in the as an important proarrhythmic factor.9,10 Animal and early part of the T wave prior to its apex. Thus, a clinical investigations have emphasized that a pro- large set of patients with a history of drug-induced longation of the QT interval might be more or less 14 • J Clin Pharmacol 2009;49:6-16
VENTRICULAR REPOLARIZATION AND HISTORY OF TORSADES DE POINTES malignant according to the location of the abnormal- In this study, we used individuals with prior ity inside the T wave in erythromycin-induced documented TdP induced by various types of QT- LQTS, in arterially perfused wedges from the canine prolonging drugs. We do not have information about left ventricle,23 and in cases of Brugada syndrome.24 drug level and triggering events in these patients.
Finally, a univariate analysis of the correlation Also, none of the patients in the group with a history between repolarization parameters at baseline and of TdP had an episode of TdP while on sotalol. Even if one does not fully understand the mechanisms significantly correlated with sotalol-induced TpTe involved in the triggering of drug-induced TdP, one prolongation (r2 = 31%, P < .0001). Our logistic could speculate that if sotalol strongly impaired the models confirm this strong relationship and suggest repolarization process, it might not set up all compo- that a prolongation of the late and the early part of nents needed for triggering TdP in our patients with the repolarization signal are not independent. The a torsadogenic predisposition. It has been shown in mechanism underlying this dependency remains to clinical studies of patients with the congenital long QT syndrome that there are crucial environmental The values of the time domain HRV parameter factors known to trigger the occurrence of cardiac and of the high-frequency components were very arrhythmias. Schwartz et al28 observed that in LQT2 similar to the HRV indices reported in normal partic- patients (patients with reduced I kinetics), most car- ipants for short-term recordings using an autoregressive diac events occur following an emotional stress event method.25 But our study groups were characterized (abrupt neurally mediated release of norepinehrine), by a vagally driven regulation of the heart rate at whereas LQT1 patients have events during exercise.
baseline. A parasympathetically driven regulation is In LQT1 patients, the risk for arrhythmic events is known to increase QT interval duration: Viitasalo increased when the protective effect of the I current and Karjalainen26 have shown an 18-ms QT prolon- does not “kick in” at a high heart rate. Such cardiac gation during night compared with day recordings stress events were not included in our experiment for the same level of heart rate (60 bpm). The pro- but could have been crucial triggering events of TdP longation of the QT interval under vagal influence has been confirmed by Bexton et al,27 who investi- Finally, Hong et al29 suggested that patients with gated the influence of the autonomic nervous system atrial fibrillation may have a shortening of the QT (ANS) on the QT interval. For these reasons, we interval. The KCNH3-K897T polymorphism associ- believe it is important to combine information about ated with atrial fibrillation30 is suggested to be also the repolarization changes and the presence of an associated with QT shortening based on large “atypical” regulation of the heart by the ANS. In our cohorts of patients from the MONICA, KORA, and study, the group of patients with a history of TdP Framingham Heart Studies.31 One must acknowl- was associated with a statistically significant edge that our study groups are both primarily con- sotalol-induced increased parasympathetic regula- stituted by patients with a history of atrial tion of the heart rate that could mean that these fibrillation: all -TdP patients and 11 of 16 patients in patients might have an increased sensibility to a beta-adrenergic blocking property of dl-sotalol,enhancing their increased propensity to repolariza- Conclusion
tion delay and to ventricular heterogeneity.
It is important for clinicians and for pharmaceutical Limitations of the Study
companies to be able to assess the level of predispo-sition to TdP of an individual. When comparing the The size of the study population was rather small ECGs from patients with and without a history of but contained a large set of ECG recordings from TdP, our results suggest that patients with a history patients with history of TdP. As far as we know, it is of drug-induced TdP have specific T wave mor- the largest set of digital ECGs in a group of patients phologies on their baseline ECGs. When challenged with a history of such rare arrhythmias. The logistic by sotalol, the patients with a history of TdP have model developed in our study has a limited value as a significantly longer late portion of the T wave a predictive tool until it is validated on an indepen- than the patients without such history. We believe this information could help optimize therapeutic strategies for cardiologists and improve the design of 15. Zong W, Moddy G, Jiang D. A robust open-source algorithm to
detect onset and duration of QRS complexes. Comput Cardiol.
2003;30:737-740.
We thank Meijian Zhou from iCardiac Technologies, Inc for 16. Couderc JP, Vaglio M, Xia X, McNitt S, Hyrien O.
valuable support during the analysis of these data. Electrocardiographic method for identifying drug-induced repo-larization abnormalities associated with a reduction of the rapidly Financial disclosure: This work has been partially funded activating delayed rectifier potassium current. Conf Proc IEEE from unrestricted grants from iCardiac Technologies, Inc and Eng Med Biol Soc. 2006;1:4010-4015.
17. Heart rate variability: standards of measurement, physiologi-
cal interpretation and clinical use. Task Force of the European
Society of Cardiology and the North American Society of Pacing
REFERENCES
and Electrophysiology. Circulation. 1996;93:1043-1065.
18. Couderc JP, Xiaojuan X, Zareba W, Moss AJ. Assessment of the
1. Shah RR, Hondeghem LM. Refining detection of drug-induced
stability of the individual-based correction of QT interval for proarrhythmia: QT interval and TRIaD. Heart Rhythm. 2005;2: heart rate. Ann Noninvasive Electrocardiol. 2005;10:25-34.
19. Hinterseer M, Thomsen MB, Beckmann BM, et al. Beat-to-beat
2. Roden DM. Drug-induced prolongation of the QT interval. N
variability of QT intervals is increased in patients with drug- Engl J Med. 2004;350:1013-1022.
induced long-QT syndrome: a case control pilot study. Eur Heart 3. Liu XK, Katchman A, Ebert SN, Woosley RL. The antiestrogen
tamoxifen blocks the delayed rectifier potassium current, IKr, in 20. Couderc JP, Zareba W, Maison-Blanche P, Moss AJ.
rabbit ventricular myocytes. J Pharmacol Exp Ther. 1998;287: Repolarization variability in the risk stratification of MADIT II patients. Europace. 2007;9:717-723.
4. Akdeniz B, Guneri S, Savas IZ, et al. Effects of carvedilol ther-
21. Haigney MC, Zareba W, Gentlesk PJ, et al. QT interval vari-
apy on arrhythmia markers in patients with congestive heart fail- ability and spontaneous ventricular tachycardia or fibrillation in ure. Int Heart J. 2006;47:565-573.
the Multicenter Automatic Defibrillator Implantation Trial 5. Antzelevitch C, Belardinelli L, Zygmunt AC, et al. Electro-
(MADIT) II patients. J Am Coll Cardiol. 2004;44:1481-1487.
physiological effects of ranolazine, a novel antianginal agent with 22. Smetana E, Pueyo K, Hnatkova V, et al. Effect of amiodarone on
antiarrhythmic properties. Circulation. 2004;110:904-910.
the descending limb of the T wave. Am J Cardiol. 2003;92:742-746.
6. Hondeghem LM, Carlsson L, Duker G. Instability and triangu-
23. Antzelevitch C, Sun ZQ, Zhang ZQ, Yan GX. Cellular and ionic
lation of the action potential predict serious proarrhythmia, but mechanisms underlying erythromycin-induced long QT intervals action potential duration prolongation is antiarrhythmic.
and Torsade de Pointes. J Am Coll Cardiol. 1996;28:1836-1848.
Circulation. 2001;103:2004-2013.
24. Castro HJ, Antzelevitch C, Tornes BF, et al. Tpeak-Tend and
7. Hondeghem LM. TRIad: foundation for proarrhythmia (triangu-
Tpeak-Tend dispersion as risk factors for ventricular tachycar- lation, reverse use dependence and instability). Novartis Found dia/ventricular fibrillation in patients with the Brugada syn- drome. J Am Coll Cardiol. 2006;47:1828-1834.
8. Belardinelli L, Antzelevitch C, Vos MA. Assessing predictors of
25. Pitzalis MV, Mastropasqua F, Massari F, et al. Short- and long-
drug-induced Torsade de Pointes. Trends Pharmacol Sci. 2003;24: term reproducibility of time and frequency domain heart rate variability measurements in normal subjects. Cardiovasc Res.
9. Liu T, Brown BS, Wu Y, Antzelevitch C, Kowey PR, Yan GX.
Blinded validation of the isolated arterially perfused rabbit ven- 26. Viitasalo M, Karjalainen J. QT intervals at heart rates from 50
tricular wedge in preclinical assessment of drug-induced proar- to 120 beats per minute during 24-hour electrocardiographic rhythmias. Heart Rhythm. 2006;3:948-956.
recordings in 100 healthy men: effects of atenolol. Circulation.
10. Yan GX, Wu Y, Liu T, Wang J, Marinchak RA, Kowey PR.
Phase 2 early afterdepolarization as a trigger of polymorphic ven- 27. Bexton RS, Vallin HO, Camm AJ. Diurnal variation of the QT
tricular tachycardia in acquired long-QT syndrome: direct evi- interval: influence of the autonomic nervous system. Br Heart J.
dence from intracellular recordings in the intact left ventricular wall. Circulation. 2001;103:2851-2856.
28. Schwartz PJ, Priori SG, Spazzolini C, et al. Genotype-phenotype
11. Roden DM. Long QT syndrome: reduced repolarization
correlation in the long-QT syndrome: gene-specific triggers for reserve and the genetic link. J Intern Med. 2006;259:59-69.
life-threatening arrhythmias. Circulation. 2001;103:89-95.
12. Kaab S, Hinterseer M, Nabauer M, Steinbeck G. Sotalol testing
29. Hong K, Bjerregaard P, Gussak I, Brugada R. Short QT syn-
unmasks altered repolarization in patients with suspected drome and atrial fibrillation caused by mutation in KCNH2. J acquired long-QT-syndrome: a case-control pilot study using i.v.
Cardiovasc Electrophysiol. 2005;16:394-396.
sotalol. Eur Heart J. 2003;24:649-657.
30. Sinner MF, Pfeufer A, Akyol M, et al. The non-synonymous
13. Lepeschkin E, Surawicz B. The measurement of the Q-T inter-
coding IKr-channel variant KCNH2-K897T is associated with val of the electrocardiogram. Circulation. 1952;6:378-388.
atrial fibrillation: results from a systematic candidate gene-based 14. Couderc JP, McNitt S, Xia J, Zareba W, Moss AJ. Repolarization
analysis of KCNH2 (HERG). Eur Heart J. 2008;29:907-914.
morphology in adult LQT2 carriers with borderline prolonged QTc 31. Ellinor PT, Milan DJ. Polymorphisms and atrial fibrillation:
interval. Heart Rhythm. 2006;3:1460-1466.
sorting the wheat from the chaff. Eur Heart J. 2008;29:843-845.
16 • J Clin Pharmacol 2009;49:6-16

Source: http://www.icardiac.com/Portals/0/PDFs/jclinicalpharmacol%20tdps.pdf