QT variability paradox after premature ventricular contraction in patients with structural heart disease and ventricular arrhythmias

Available online at www.sciencedirect.com

Journal of Electrocardiology 45 (2012) 652–657www.jecgonline.com

QT variability paradox after premature ventricular contraction in patientswith structural heart disease and ventricular arrhythmias☆,☆☆

Durgesh Das, a Lichy Han, BS, a Ronald D. Berger, MD, PhD, b

Larisa G. Tereshchenko, MD, PhDb,⁎aWhiting School of Engineering, The Johns Hopkins University, Baltimore, MD, USA

bThe Division of Cardiology, Department of Medicine, Johns Hopkins Hospital, Baltimore, MD, USA

Received 11 April 2012

Abstract Background: Increased repolarization lability is known to be associated with the risk of ventricular

☆ Registration ide☆☆ Financial sup

as an investigator-initi⁎ Corresponding a

MD, 21287, USA.E-mail address: lt

0022-0736/$ – see frohttp://dx.doi.org/10.10

tachycardia (VT)/ventricular fibrillation (VF). Premature ventricular contractions (PVCs) are excludedfrom the analysis of QT variability. However, QT dynamics after PVCs is poorly understood.Methods: We analyzed data of 33 patients with structural heart disease (mean age 60.5±12.1; 24(73%) men; 26 (79%) whites; 22 (67%) ischemic cardiomyopathy) and single-chamber ICD implantedfor primary (28 patients, 85%) or secondary prevention of SCD. Arrhythmia group comprised 16patients with VT/VF/death outcomes. Alive patients (n=17) without VT/VF served as controls. Thebaseline far-field (FF) ICD electrogram (EGM) was recorded at rest. RR and QT intervals of 15 sinus

beats before and after PVC in 33 patients were analyzed. The prematurity index,Ci

MeanRR, where Ci is

coupling interval, was used to select the most premature PVC. QT variability index (QTVI) wascalculated. Difference in QTVI was calculated as QTVIdiff=QTVIafter−QTVIbefore.Results: In paired analysis QTVI significantly increased after PVC in controls (0.64±1.02 vs.0.26±1.15; P=0.046), but decreased in patients in the arrhythmia group (0.16±0.85 vs. 0.43±0.84;P=0.190). QTVIdiff was significantly lower in patients with VT/VF, as compared to controls(−0.197±0.650 vs. 0.207±0.723; P=0.030). In multivariate logistic regression after adjustment forthe type of cardiomyopathy and NYHA class the decrease in QTVI after PVC was associated withincreased risk of VT/VF (OR 9.24; 95% CI 1.11–76.82; P=0.040).Conclusion: Elevated at baseline QTVI is decreased during first 15 beats after PVC in patients atrisk for VT/VF.© 2012 Elsevier Inc. All rights reserved.

Keywords: Premature ventricular contraction; Repolarization lability; Ventricular arrhythmia; Implantable cardioverter defibrillator

Introduction

Sudden cardiac death (SCD) is responsible for approxi-mately 50% of deaths from cardiovascular disease in theUnited States and other developed countries. In the UnitedStates alone, it accounts for between 300,000 and 400,000deaths.1 Implantable cardioverter defibrillators (ICD) havebeen shown to be the most effective therapy for patients withventricular tachycardia (VT)/ventricular fibrillation (VF).Among patients in the Multicenter Automatic Defibrillator

ntification number: NCT00916435.port: The study was partially supported by Medtronicated research project (awarded to Dr. Tereshchenko).uthor. 600 N Wolfe Street, Carnegie 568, Baltimore,

[email protected]

nt matter © 2012 Elsevier Inc. All rights reserved.16/j.jelectrocard.2012.07.016

Implantation Trial (MADIT) who were given ICD devices orconventional medical therapy, after a 27-month follow-upimplanted patients had a 54% decreased rate of mortality.2 Itis known that the ICD is an effective therapy. However, amethod to identify patients at high risk of SCD has yet to befully developed. Guidelines of the American College ofCardiology and the American Heart Association3 recognizea low left ventricular ejection fraction (LVEF) as asignificant predictor of SCD. At the same time lowspecificity of LVEF has instigated a search for additionalpredictive markers to use for risk stratification of VT/VF.

It was previously shown that increased temporal variability ofQT interval duration (QT variability) is associated with life-threatening ventricular arrhythmias.4-6 It is known thatpremature ventricular contractions (PVC) can alter the QTinterval within several consecutive cardiac cycles.7 QT

653D. Das et al. / Journal of Electrocardiology 45 (2012) 652–657

variability analysis requires exclusion of one sinus beatimmediately after PVC. However, the effect of PVC on QTvariability has not been systematically studied. The goal of ourstudywas to compareQTvariability during 15 consecutive sinusbeats before and after single PVC in patients with structural heartdisease with and without VT/VF during follow-up.

Methods

Study patients population

We retrospectively analyzed the prospectively collected dataof 33 patients enrolled in the ICD-EGM study 4

(NCT00916435). The study enrolled patients with structuralheart disease who have been implanted with Medtronic(Minneapolis, MN, USA) or Boston Scientific Corporation(Natick, MA, USA) ICD device for primary or secondaryprevention of SCD. The design and selection criteria of theoriginal cohort have been previously described.4 Briefly, thestudy included patients with approved indications for ICDimplanted for primary or secondary prevention of SCD. Patientswith inherited channelopathies, children b18years, pregnantwomen and patients with concomitant non-cardiac diseaseswithhigh likelihood of death during the first year after enrollmentwere excluded from the study. Patients were usually enrolled1week after ICD implantation, during regular office visit. Afterenrollment study participants were followed at least once a yearin office and every 3months remotely via Carelink© orLatitude© remote monitoring. Programming of the ICD devicewas based on clinical evaluation of the attending electrophys-

Fig. 1. Example of QT variability measurement in 15

iologist. Appropriate ICD therapies (either shock or antitachy-cardia pacing [ATP]) served as primary end-point in ICD-EGMsstudy. All-cause mortality served as a secondary end-point.

We selected the sample of cases (n=16) and controls (n=17) out of the main study cohort. The sample of cases(arrhythmia group) either received appropriate ICD therapies(either shock or antitachycardia pacing [ATP]) for VT/VF, ordied from any cause, whichever came first. The sample ofcontrols comprised study subjects who did not experienceappropriate ICD therapies for VT/VF and was alive at the endof follow-up. Important condition for inclusion in this analysiswas recorded PVC on baseline EGM in sinus rhythm with atleast 15 consecutive sinus beats before and after PVC.No otherECG criteria of patient selection were applied.

Intracardiac electrogram recordings

Baseline digital far-field (FF) right ventricle (RV)intracardiac EGMs in sinus rhythm were recorded for 5 to10minutes at rest during regular office visits simultaneouslywith a single-lead (lead II) surface ECG via ICD programmeras previously described.4 FF RV EGM was recorded as avoltage difference between the ring of the RV lead and theICD Can, or as a voltage difference between the distal coiland the ICD Can. The EGM was sampled at 1000Hz.

QT variability analysis before and after prematureventricular contractions

Selections of 30 clean sinus beat intervals were madeusing custom software in MATLAB (MathWorks, Inc.

beats before PVC and in 15 beats after PVC.

Table 1Comparison of clinical characteristics of the arrhythmia group and controls.

Arrhythmiagroup (n=16)

Controls(n=17)

P

Mean age±SD, years 60.5±12.5 60.5±54.2 0.99Male, n (%) 14 (87.5) 10 (58.8) 0.065African Americans, n (%) 2 (12.5) 5 (29.4) 0.235Ischemic CM with MI history, n (%) 11 (68.8) 11 (64.7) 0.805

654 D. Das et al. / Journal of Electrocardiology 45 (2012) 652–657

Natick, MA, USA). Each interval used in the analysis wascentered upon a PVC with 15 consecutive sinus beats beforeand after PVC (Fig. 1). This selection criterion was defined tocapture the effect of the PVC without diluting the effect withregular sinus rhythm beats. Previous experience of measuringshort-term QT variability on 30 beats was reported.7

A dedicated computer algorithm was created inMATLAB (MathWorks, Inc, Natick, MA) in Tereshchen-ko's laboratory. QT and RR intervals were measuredautomatically and appropriateness of fiducial points' place-ment was verified by visual aid (D.D., L.G.T.). Then wecalculated mean QT interval (QTm), mean RR interval(RRm), QT variance (QTv), and RR variance (RRv) over 15beats before and 15 beats after PVC. We separatelyquantified normalized QT variability during 15 beats beforeand 15 beats after PVC. QT variability index (QTVI) wascalculated by the method of Berger et al.5 using thefollowing equation:

QTVI = log10QTvð Þ=ðQTmÞ2RRvð Þ=ðRRmÞ2

( )

In patients with multiple PVCs, the prematurity index wasused to select the most premature PVC as long as prematuritywas N20%. The prematurity index was calculated according

to equation:Ci

RRm, where Ci denotes coupling interval.

Coupling interval was measured as a time interval betweenthe PVC and the sinus beat preceding PVC. To comparedifference in QTVI before and after PVC, we calculateddifference in QTVI as follows:

QTVIdiff¼ QTVIafter � QTVIbefore

Statistical analysis

Results are presented as mean±standard deviation (SD)for normally distributed variables, and as median andinterquartile range for skewed variables. Continuous vari-ables that were normally distributed were compared usingindependent sample t test. For skewed continuous variables,a Wilcoxon rank-sum test was used. Mean heart rate, QTcand QTVI before and after PVC were compared using pairedt test in the arrhythmia group and controls. Univariate andmultivariate logistic regression analysis was used todetermine if changes in QTVI after PVC are associatedwith the risk of VT/VF or death. We considered a probabilityvalue b0.05 to be significant. STATA 12 (StataCorp LP,College Station, TX) was used for calculations.

Primary prevention of SCD, n (%) 13 (81.3) 15 (88.2) 0.576LVEF at ICD implantation±SD, % 33.3±10.1 34.5±10.2 0.746NYHA class I, % 18.2 36.4 0.506NYHA class II, % 63.6 45.5 0.506Diabetes mellitus, % 36.4 27.3 0.647Hypertension, % 81.8 90.9 0.534History of CABG, % 9.1 27.3 0.269History of PTCA, % 36.4 45.5 0.665Class III antiarrhythmics 0 18.2 0.138

CM: cardiomyopathy; MI: myocardial Infarction; NYHA: New York HeartAssociation; CABG: coronary artery bypass grafting; PTCA: percutaneoustransluminal coronary angioplasty.

Results

Study population comprised patients with ICD implantedfor primary prevention of SCD (85%) and secondaryprevention (15%). Majority of patients had ischemiccardiomyopathy (67%) with mild heart failure (meanejection fraction 33.9±10.0), predominantly NYHA classII. Mean age was 60.5±12.2years (range 33–81years),73% were men. Mean heart rate was 83.6±13.7bpm. On

average patients had 1.3±0.9PVCs per minute. Therewere no statistically significant differences in baselineclinical characteristics of the arrhythmia group and controls(Table 1). Patients in the arrhythmia group were borderline-likely to be male. Over a period of 24.1±16.1months, 13 of16 patients in the arrhythmia group sustained VT/VF withappropriate ICD therapy and 3 patients died.

QT variability before and after prematureventricular contractions

At the first step of analysis on average 2 PVCs and asso-ciate 30 beats were analyzed in each study participant (67PVCs in 33 patients). At the second step PVCs were sortedbased on prematurity index, and the most premature PVCand associate 30 beats were selected and further analyzed.

Fig. 2 shows the representative examples of the heart rateand QT interval dynamics during 15 beats before and 15beats after PVC in a case and control patient. There was nostatistically significant difference in baseline heart rate, heartrate variability, mean QT interval and QT variability in thearrhythmia group and controls (Table 2). At the same time,dynamics of QT variability after PVC in patients with andwithout sustained VT/VF with appropriate ICD therapieswas dramatically different. In paired analysis QTVIsignificantly increased after PVC in controls (0.64±1.02vs. 0.26±1.15; P=0.046), but decreased in patients in thearrhythmia group (0.16±0.85 vs. 0.43±0.84; P=0.190).Fig. 3 shows a complex relationship between QTVI beforeand after PVC and two major trends in QTVI changes afterPCV: in controls QTVI before PVC positively correlatedwith QTVI after PVC, while in the arrhythmia groupcorrelation of QTVI before and after PVC was predomi-nantly negative. QTVIdiff was significantly lower in patientswith VT/VF and appropriate ICD therapies, as compared topatients without arrhythmia (−0.197±0.650 vs. 0.207±0.723; P=0.030). The mean difference in QTVI wasnegative in patients with VT/VF, i.e., QTVI decreasedafter PVC. In contrast, a positive difference signifiedincreased QTVI after PVC in patients without arrhythmia.

CaseControlA

B

C

Fig. 2. Representative example of QT variability in a case and control patient. (A) Sequence of analyzed NF RV EGM heart beats. (B) Graph of 15 RR intervalsbefore and 15 RR intervals after PVC. (C) Graph of 15 QT intervals before PVC and 15 QT intervals after PVC.

655D. Das et al. / Journal of Electrocardiology 45 (2012) 652–657

Logistic regression analysis for prediction of VT/VF

Increased QTVI, as measured before PVC, was border-line-predictive for sustained VT/VF with appropriate ICDtherapies. Surprisingly, decreased QTVI after PVC was a

significant predictor of VT/VF in univariate and multivariatelogistic regression, after adjustment for age, sex, type ofcardiomyopathy and NYHA heart failure class (Table 3).Decrease in QTVI after PVC by 0.1 was associated withabout 70% increase of the VT/VF risk. Any decrease in QTVI

Table 2Comparison of QT variability before and after PVC in the arrhythmia groupand Controls.

Arrhythmiagroup (n=16)

Controls(n=17)

P

Mean heart rate before PVC±SD,bpm

78.3±22.7 88.6±24.3 0.215

Mean heart rate after PVC±SD, bpm 77.4±21.8 90.2±17.3 0.146Mean QTc before PVC±SD, ms 468±86 433±77 0.228Mean QTc after PVC±SD, ms 460±85 428±60 0.231QTVI before PVC±SD 0.241±1.04 0.411±1.05 0.643QTVI after PVC±SD 0.141±1.00 0.740±0.867 0.078QTVI difference±SD −0.100±0.762 0.330±0.813 0.128Coupling interval±SD, ms 429±158 422±132 0.886Prematurity index±SD 0.53±0.19 0.59±0.12 0.340

PVC: premature ventricular contraction; QTc: QT corrected; QTVI: QTvariability index.

656 D. Das et al. / Journal of Electrocardiology 45 (2012) 652–657

after PVC as compared to QTVI before PVC was associatedwith 9-fold higher odds of having VT/VF in multivariateanalysis after adjustment for the type of cardiomyopathy andNYHA class. Interestingly, QTVIdiff was less predictivetoward combined end point (VT/VF or all-cause death): afteradjustment for the type of cardiomyopathy and NYHA classOR 2.32 (95% CI 0.39–13.72); P=0.352.

Discussion

In the present study we described QT variability paradoxafter PVC. We showed that QTVI elevated at baseline isdecreased during the first 15 beats after PVC in patients atrisk for VT/VF. The steeper decrease of QTVI after PCV thehigher the odds of having VT/VF. At the same time, patientswith ICD who did not have VT/VF during follow-up werecharacterized by increased QTVI immediately after PCV.Difference in QTVI before and after PVC was stronglyassociated with sustained VT/VF with appropriate ICDtherapies after adjustment for demographics (age, sex) andimportant clinical variables (NYHA class, ischemic or non-ischemic cardiomyopathy).

Fig. 3. Correlation between QT variability index before and after PVC.Median spline-fitted lines are shown.

QT variability Index

QTVI was proposed by Berger et al5 in 1997 to quantifythe magnitude of QT interval variability, using normalizedmean QT duration and the magnitude of heart ratevariability. Increased QT variability observed in heart failurepatients has been considered a sign of an increasedsympathetic tone in the ventricles of the heart. Direct andindirect evidence of correlation between QTVI and integrat-ed vagus nerve activity, as well as integrated left stellate-ganglion nervous activity was shown in experiment8 and inclinical studies.9

Several prospective observational studies showed apredictive value of QTVI for risk stratification of suddencardiac death (SCD) and ventricular arrhythmias. Our recentfinding of increased intracardiac QT variability in this studypopulation confirmed that repolarization lability may bepresent throughout the ventricles.4 Elevated QTVI predictedSCD and VT/VF in patients with ischemic and non-ischemiccardiomyopathy,6,10,11 hypertrophic cardiomyopathy,12

myocardial ischemia,13 and long-QT syndrome.14 However,in some studies QTVI did not predict VT/VF. In particular,QTVI fails to predict VT/VF in acute intensive cardiac careunit (CCU) settings,15 in patients with frequent PVCs. Webelieve that our research partially explains previous work ofSachdev et al.15 They observed frequent PVCs (6–7PVCs/min) and slightly elevated QTVI in acute CCU patients withand without VT/VF. Moreover, Sachdev et al. demonstratedabrupt decrease in QTVI, which mirrored an increasednumber of PVCs before VT/VF event (Fig. 3 of the citedpaper). Thus, in patients with frequent PVCs QTVI isconfounded by the effect of PVCs on QTVI.

Effect of PVC on repolarization

Very few studies explored the effect of PVCs onrepolarization dynamics. Savelieva et al.16 measured QTturbulence in patients with and without ventricular dysfunc-tion. In contrast to the noticeable heart rate turbulence afterPVC (early acceleration and then late deceleration in heartrate), QT turbulence after PVC was negligible. Savelievaet al.16 noticed particularly blunted response to PVC inpatients with LV dysfunction, as compared to subjects withnormal LVEF, which is consistent with our finding ofdecreased QT variability in cardiomyopathy patients withVT/VF. However, previously no comparison of QTVI beforeand after PVC has been made, as turbulence assessmentsuggests measurement of two cardiac cycles before PVC only.

The mechanism of blunted QT variability after PVC inpatients at risk of VT/VF is not clearly defined. Severalmechanisms potentially could govern QT variability afterPVC. Electrical restitution reflects adaptation of the actionpotential duration to changes in cycle length. 17 QThysteresis explains delayed (50–60s) adaptation of QTinterval to heart rate changes.18 Decreased QT variabilityafter PVC might reflect impaired post-PVC adaptation toabrupt changes in cardiac output. Further studies are neededto explain our findings.

Our findings potentially carry several important practicalimplications. First, decreased QTVI after PVC should be

Table 3Univariate and multivariate odds ratio of having sustained VT/VF with appropriate ICD therapies.

Predictor Unadjusted odds ratio (95% CI) P Adjusted odds ratio (95% CI) P

QTVI before PVC 3.1 (0.93–10.64) 0.067 175.9 (0.42–73634.9)a 0.093QTVI after PVC 0.24 (0.07–0.85) 0.027 0.008 (0.00008–0.93)a 0.046Difference in QTVI 0.28 (0.09–0.91) 0.035 0.011 (0.0002–0.59)a 0.026Decreased QTVI after PVC 3.73 (0.86–16.25) 0.079 9.24 (1.11–76.82)b 0.040

Variables “QTVI before PVC” and “QTVI after PVC” were entered in one model.a Adjusted for age, sex, New York Heart Association heart failure class, and type of cardiomyopathy (ischemic vs. non-ischemic).b Adjusted for New York Heart Association heart failure class, and type of cardiomyopathy.

657D. Das et al. / Journal of Electrocardiology 45 (2012) 652–657

taken into account and ECG recordings with and withoutfrequent PVCs might require different threshold definition.Presence of PVCs should be taken into account whendetermining threshold of abnormal QTVI. Alternatively,requirements of eligibility for QT variability analysis mightbe revised. Second, observed phenomenon could be furtherdeveloped as a separate prognostic test, which could beadministered in patients during invasive EP procedure (todeliver single PVC), or Holter ECG recording could beanalyzed for that purpose.

LimitationsThe study population was very small and this study was

retrospective. Statistical power was limited. The resultsshould be validated in a larger prospective cohort study.Analyzed 15-s time segments are nonstationary, andtherefore observed results could be random, by chance.Reproducibility studies are needed to confirm consistenteffect of PVC on QTVI in patients with and withoutarrhythmia. QT duration was measured on a single lead andtherefore, might carry a measurement error, as projection ofthe heart vector on a single ECG lead is always imperfect.19

Conclusion

In patients with structural heart disease, those that showsustained VT/VF are characterized by decreased QTVI overa 15-beat interval following a PVC compared to an increasedQTVI in similar patients with structural heart disease who donot experience VT/VF.

References

1. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and strokestatistics–2012 update: a report from the American Heart Association.Circulation 2012;125:e2.

2. Moss AJ, Zareba W, Hall WJ, et al. Prophylactic implantation of adefibrillator in patients with myocardial infarction and reduced ejectionfraction. N Engl J Med 2002;346:877.

3. Epstein AE, DiMarco JP, Ellenbogen KA, et al. ACC/AHA/HRS 2008guidelines for device-based therapy of cardiac rhythm abnormalities: areport of the American College of Cardiology/American HeartAssociation Task Force on Practice Guidelines (Writing Committee toRevise the ACC/AHA/NASPE 2002 Guideline Update for Implantationof Cardiac Pacemakers and Antiarrhythmia Devices): developed incollaboration with the American Association for Thoracic Surgery andSociety of Thoracic Surgeons. Circulation 2008;117:e350.

4. Tereshchenko LG, Fetics BJ, Domitrovich PP, Lindsay BD, Berger RD.Prediction of ventricular tachyarrhythmias by intracardiac repolariza-tion variability analysis. Circ Arrhythm Electrophysiol 2009;2:276.

5. Berger RD, Kasper EK, Baughman KL, Marban E, Calkins H,Tomaselli GF. Beat-to-beat QT interval variability: novel evidence forrepolarization lability in ischemic and nonischemic dilated cardiomy-opathy. Circulation 1997;96:1557.

6. Haigney MC, Zareba W, Gentlesk PJ, et al. QT interval variability andspontaneous ventricular tachycardia or fibrillation in the MulticenterAutomatic Defibrillator Implantation Trial (MADIT) II patients. J AmColl Cardiol 2004;44:1481.

7. Oosterhoff P, Tereshchenko LG, van der Heyden MA, et al. Short-termvariability of repolarization predicts ventricular tachycardia and suddencardiac death in patients with structural heart disease: a comparison withQT variability index. Heart Rhythm 2011;8:1584.

8. Piccirillo G,Magri D,OgawaM, et al. Autonomic nervous system activitymeasured directly and QT interval variability in normal and pacing-induced tachycardia heart failure dogs. J Am Coll Cardiol 2009;54:840.

9. Baumert M, Schlaich MP, Nalivaiko E, et al. Relation between QTinterval variability and cardiac sympathetic activity in hypertension. AmJ Physiol Heart Circ Physiol 2011;300:H1412.

10. Atiga WL, Calkins H, Lawrence JH, Tomaselli GF, Smith JM, BergerRD. Beat-to-beat repolarization lability identifies patients at risk forsudden cardiac death. J Cardiovasc Electrophysiol 1998;9:899.

11. Piccirillo G, Magri D, Matera S, et al. QT variability strongly predictssudden cardiac death in asymptomatic subjects with mild or moderateleft ventricular systolic dysfunction: a prospective study. Eur Heart J2007;28:1344.

12. Atiga WL, Fananapazir L, McAreavey D, Calkins H, Berger RD.Temporal repolarization lability in hypertrophic cardiomyopathy causedby beta-myosin heavy-chain gene mutations. Circulation 2000;101:1237.

13. Murabayashi T, Fetics B, Kass D, Nevo E, Gramatikov B, Berger RD.Beat-to-beat QT interval variability associated with acute myocardialischemia. J Electrocardiol 2002;35:19.

14. Bilchick K, Viitasalo M, Oikarinen L, et al. Temporal repolarizationlability differences among genotyped patients with the long QTsyndrome. Am J Cardiol 2004;94:1312.

15. Sachdev M, Fetics BJ, Lai S, Dalal D, Insel J, Berger RD. Failure inshort-term prediction of ventricular tachycardia and ventricularfibrillation from continuous electrocardiogram in intensive care unitpatients. J Electrocardiol 2010;43:400.

16. Savelieva I, Wichterle D, Camm JA. QT-interval turbulence induced byatrial and ventricular extrastimuli in patients with ventricular tachycar-dia. Pacing Clin Electrophysiol 2005;28(Suppl 1):S187.

17. Franz MR, Swerdlow CD, Liem LB, Schaefer J. Cycle lengthdependence of human action potential duration in vivo. Effects ofsingle extrastimuli, sudden sustained rate acceleration and deceleration,and different steady-state frequencies. J Clin Invest 1988;82:972.

18. Lau CP, Freedman AR, Fleming S, Malik M, Camm AJ, Ward DE.Hysteresis of the ventricular paced QT interval in response to abruptchanges in pacing rate. Cardiovasc Res 1988;22:67.

19. Kors JA, van HG, van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation 1999;99:1458.