- Email: [email protected]
Repolarization analysis in children with the long QT syndrome
Journal of Electrocardiology Vol. 36 Supplement 2003
Repolarization Analysis in Children With the Long QT Syndrome
Konrad Brockmeier, MD,* Markus Khalil, MD,† Narayanswami Sreeram, MD,* and Herbert E. Ulmer, MD†
Abstract: QT duration changes were analysed in 26 children with long QT syndrome (median age 9 years). With 12-lead Holter recordings, we found that macroscopically QT adaptation to heart rate changes was slow and only 3 of 26 patients showed abrupt changes in QT interval duration. These changes in the time domain were always accompanied by changes of STT morphology in the 3 of 26 patients. Application of Bazett⬘s heart rate correction algorithm was inappropriate to show true QT changes in this young patient group due to temporary typical large cycle length variations, where moreover an overcorrection of the heart rate impact resulted in reflection of RR interval changes rather than true QT duration changes. Key words: LQTS, Bazett, QT/RR relations.
tion may result at low and high heart rates, respectively (2). With all its documented limitations, Bazett’s formula may be helpful as a first approach in detection of a prolonged QT interval. In clinical practice, however, the use of percentiles of QT values have shown advantages (3). In the patient with suspected long QT syndrome (LQTS), clinical parameters and noninvasive testing remain the basic workup for diagnostic and therapeutic decisions. However, repolarization duration may vary from extremely prolonged to normal in LQTS patients. In a recent work of Perkiomaki et al, it was shown that in LQTS patients the variability of QT duration seems to play an important role to identify symptomatic patients (4). When we routinely reviewed Holter data of LQTS patients to screen for QT interval and T-wave morphology changes, displaying them digitally by superimposing adjacent beats, we moreover found a rather stable relation of the beat-to-beat QRS to T wave relation in most of
The complexity of heart rate and repolarization duration interaction is a well described phenomenon. Adaptation of ventricular repolarization to heart rate changes is slow in the normal heart. In the atrially paced normal subject adaptation of repolarization to abrupt changes in heart rate takes more than 120 seconds until an adequate steady state relation of beat rate and repolarization duration is reached (1). To provide comparable values for time domain calculations of the ventricular repolarization, the Bazett formula to correct heart rate influence is still very frequently used. Typically, by applying the Bazett formula overestimation and underestima-
From the Pediatric Cardiology, University of Cologne, Ko¨ln, Germany; and Pediatric Cardiology, University of Heidelberg, Heidelberg, Germany. Reprint requests: Konrad Brockmeier, MD, Head of Department of Pediatric Cardiology, Universitaetsklinikum Koeln, 50924 Koeln, Germany; e-mail: [email protected]. © 2003 Elsevier Inc. All rights reserved. 0022-0736/03/360S-0058$30.00/0 doi:10.1016/j.jelectrocard.2003.09.061
209
210 Journal of Electrocardiology Vol. 36 Supplement 2003
Fig. 1. An example of the inappropriateness of Bazett⬘s correction formula during periods of substantial RR interval changes. In the displayed episode (lead II) all QT intervals were similar in duration, however, pathologically prolonged. Only the amount of RR changes in the displayed time series is reflected in the QTc values as QT changes do not occurr. Numbers are given in milliseconds, upper row represents cycle length, then lower QT duration, and lowest row QTc values calculated by use of the Bazett formula.
the patients without macroscopic changes of the STT morphology (6). In the present study, we tried to evaluate the occurrence of abrupt changes of QT interval duration in Holter ECG recordings of young LQTS patients.
Materials and Methods Digital 12-lead Holter recordings from 26 LQTS patients (age median 9 years, range 1 -20 years), 15 women and 11 men were analyzed (H12 cont. 2.00, Mortara Instruments, Milwaukee, WI). Seven of 26 patients were under -blocker medication. All patients were clinically identifyed as LQTS patients with score values of ⱖ 3, using the criteria of Schwartz et al (5). RR and QT values as well as beat rate corrected QTc values, using the Bazett correction formula, were automatically processed with the software of the Holter system. QTc values were averaged and displayed with median and quartiles thus eliminating data outliers. Additionally, data analysis was performed using software to investigate dynamics of cycle length mediated repolarization changes (6). Data of an appropriate lead were interactively processed by displaying adjacent beats superimposed on the workstation screen to qualify changes of the QRS– STT pattern, thus visualizing the recordings hour by hour. In all patients, genotyping for analysis of the underlying ion channel mutation was performed.
Results The number of successfully genotyped individuals was 9 of 26 (35%): LQT1 n ⫽ 6, LQT2 n ⫽ 2, and LQT3 n ⫽ 1. QT duration of the patients was prolonged in all patients under resting conditions. Circadian changes of the QT interval differed, varying from extremely prolonged to normal in the investigated LQTS patients. The averaged and heart rate corrected QT intervals showed a remarkably broad range with a median of QTc (Bazett) 498 (range, 400 – 630). In the 7 patients with -blocker medication, a similar range of QT intervals was typical, however, presenting lower mean heart rates: median QTc 504 (range, 420 –580); median of circadian heart rate 65 bpm in the -blocker treated group versus 74 bpm in the untreated patients. The pattern of QRS–STT behavior macroscopically showed little changes in all of the patients when leads were displayed superimposed and tracked hour by hour. In 3 of 26 patients, we found episodes with abrupt QT duration changes. These changes in the time domain were invariably accompanied by changes of T-wave morphology. The episodes were found in 2 patients with LQT2 and in one patient with LQT3, respectively. QTc values obtained, by using Bazett⬘s heart rate correction formula, provided an inappropriate picture of the QT/RR relation in all patients, as during periods of sinus node modulated heart rate changes–a typical behavior in young individuals– the variation of the RR intervals were overrepresented in the results of QTc calculations. Figure 1 is
Repolarization Analysis in LQTS •
Brockmeier et al. 211
Fig. 2. An episode with a singular premature ventricular beat in another LQTS patient. All QT intervals are pathologically prolonged (530 ms). The intermittend extrasystole, however, does not change the QT duration in the displayed cycles. A possible marginal additional prolongation can be seen in the adjacent beat after the extrasystole. The following cycle then shows again a QT interval of 530 ms. Lead II and V5, numbers are given in milliseconds.
showing an example of the inappropriateness of Bazett⬘s correction formula during periods of substantial RR interval changes. In Figure 1, the displayed QT intervals were all unchanged, however, pathologically prolonged. Only the amount of RR changes in the displayed time series is reflected in the QTc values as QT changes do not occur. Figure 2 shows an episode with a singular premature ventricular beat in another LQTS patient. All QT intervals are pathologically prolonged (530 ms). The intermittent extrasystole, however, does not change the QT duration in the displayed cycles. A possible marginal additional prolongation can be seen in the adjacent beat after the extrasystole. The following cycle then shows again a QT interval of 530 ms.
Discussion Our results show that a rather slow variability of the QT duration in consequence to substantial heart rate changes in LQTS patients is typical. This is a similar behavior as seen in healthy patients. All investigated patients had score values greater or equal to 3 according to the Schwartz criteria and had prolonged QT intervals. The use of beta-blocker did not modify repolarization duration. Beta-
blocker medication did slow the heart rate but not QT variability. A beat-to-beat variability of the QT to RR ratio is a rare finding, as most of the repolarization changes are very slow-also in a majority of patients with ion channel mutations. In the investigated LQTS patients abrupt macroscopic changes of RR/QT relation were detected in 3 patients only. This behavior was seen in 2 individuals with mutations of the HERG gene (LQT2) and in 1 patient with a mutation of the SCN5A sodium channel (LQT3). The fast changes of repolarization duration seen in these patients represented forms of T-wave alternans episodes. The slow reaction of the ventricular repolarization to rapid changes of heart rate is a typical finding also during exercise testing of LQTS patients. Exercise testing is used in the setting of potential candidates of LQTS to uncover QT prolongation after stress testing in patients with borderline QT duration under baseline conditions. The highest proportion of ion channel mutations among LQTS patients is found in the KVLQ1 gene, which is sensitive to exercise related torsades. In a publication by Dillenburg et al., it was shown that in comparison to healthy controls, patients with LQTS showed significant changes of QT prolongation not earlier than 3 minutes after cessation of exercise (7). These findings underline the typical slow reac-
212 Journal of Electrocardiology Vol. 36 Supplement 2003 tion of repolarizaton duration adaptation to heart rate changes in LQTS patients. Heart rate is the strongest variable influencing the repolarization duration and is typically coupled. Larger ranges of high QTc values are typical in symptomatic LQTS, but do not nessessarily represent real QT changes or substantial QT/RR short time changes.Ranges of heart rate corrected QT duration, as provided by Bazett⬘s QTc correction (8), may therefore not neccessarily reflect a higher instability/vulnerability of the ventricular repolarization– but may be heavily influenced by short time RR interval changes for instance in children. Time domain analysis is very probably inferior to algorithms, which quantify T-wave morphology changes or combine both approaches.
2.
3.
4.
5.
6.
7.
References 1. Vainer J, van der Steld B, Smeets JL, et al: Beat-to-beat behavior of QT interval during conducted supraven-
8.
tricular rhythm in the normal heart. PACE 17:1469, 1994 Rautaharju PM, Warren JW, Calhoun HP. Estimation of QT prolongation: A persistent, avoidable error in computer electrocardiography. J Electrocardiol 23: 111, 1991 (suppl) Eberle T, Brockmeier K, Heling G, et al. Prediction of normal QT intervals in children. J Electrocardiol 31: 121, 1999 (Suppl) Perkiomaki JH, Zareba W, Nomura A, et al: Repolarization dynamics in patients with long QT syndrome. J Cardiovasc Electrophysiol 13:651, 2002 Schwartz PJ, Moss AJ, Vincent GM, et al: Diagnostic criteria for the long QT syndrome: An update. Circulation 88:782, 1993 Brockmeier K, Aslan I, Hilbel T, et al: T-wave alternans in LQTS: Repolarization-rate dynamics from 12 lead Holter data. J Electrocardiol 34:93, 2001 Dillenburg RF, Hamilton RM: Is exercise testing useful in identifying congenital long QT syndrome? Am J Cardiol 89:233, 1994 Kligfield P, Lax KG, Okin PM. QTc behavior during treadmill exercise as a function of the underlying QT-heart rate relationship. J Electrocardiol 28:206, 1995 (suppl)
Repolarization Analysis in Children With the Long QT Syndrome
Konrad Brockmeier, MD,* Markus Khalil, MD,† Narayanswami Sreeram, MD,* and Herbert E. Ulmer, MD†
Abstract: QT duration changes were analysed in 26 children with long QT syndrome (median age 9 years). With 12-lead Holter recordings, we found that macroscopically QT adaptation to heart rate changes was slow and only 3 of 26 patients showed abrupt changes in QT interval duration. These changes in the time domain were always accompanied by changes of STT morphology in the 3 of 26 patients. Application of Bazett⬘s heart rate correction algorithm was inappropriate to show true QT changes in this young patient group due to temporary typical large cycle length variations, where moreover an overcorrection of the heart rate impact resulted in reflection of RR interval changes rather than true QT duration changes. Key words: LQTS, Bazett, QT/RR relations.
tion may result at low and high heart rates, respectively (2). With all its documented limitations, Bazett’s formula may be helpful as a first approach in detection of a prolonged QT interval. In clinical practice, however, the use of percentiles of QT values have shown advantages (3). In the patient with suspected long QT syndrome (LQTS), clinical parameters and noninvasive testing remain the basic workup for diagnostic and therapeutic decisions. However, repolarization duration may vary from extremely prolonged to normal in LQTS patients. In a recent work of Perkiomaki et al, it was shown that in LQTS patients the variability of QT duration seems to play an important role to identify symptomatic patients (4). When we routinely reviewed Holter data of LQTS patients to screen for QT interval and T-wave morphology changes, displaying them digitally by superimposing adjacent beats, we moreover found a rather stable relation of the beat-to-beat QRS to T wave relation in most of
The complexity of heart rate and repolarization duration interaction is a well described phenomenon. Adaptation of ventricular repolarization to heart rate changes is slow in the normal heart. In the atrially paced normal subject adaptation of repolarization to abrupt changes in heart rate takes more than 120 seconds until an adequate steady state relation of beat rate and repolarization duration is reached (1). To provide comparable values for time domain calculations of the ventricular repolarization, the Bazett formula to correct heart rate influence is still very frequently used. Typically, by applying the Bazett formula overestimation and underestima-
From the Pediatric Cardiology, University of Cologne, Ko¨ln, Germany; and Pediatric Cardiology, University of Heidelberg, Heidelberg, Germany. Reprint requests: Konrad Brockmeier, MD, Head of Department of Pediatric Cardiology, Universitaetsklinikum Koeln, 50924 Koeln, Germany; e-mail: [email protected]. © 2003 Elsevier Inc. All rights reserved. 0022-0736/03/360S-0058$30.00/0 doi:10.1016/j.jelectrocard.2003.09.061
209
210 Journal of Electrocardiology Vol. 36 Supplement 2003
Fig. 1. An example of the inappropriateness of Bazett⬘s correction formula during periods of substantial RR interval changes. In the displayed episode (lead II) all QT intervals were similar in duration, however, pathologically prolonged. Only the amount of RR changes in the displayed time series is reflected in the QTc values as QT changes do not occurr. Numbers are given in milliseconds, upper row represents cycle length, then lower QT duration, and lowest row QTc values calculated by use of the Bazett formula.
the patients without macroscopic changes of the STT morphology (6). In the present study, we tried to evaluate the occurrence of abrupt changes of QT interval duration in Holter ECG recordings of young LQTS patients.
Materials and Methods Digital 12-lead Holter recordings from 26 LQTS patients (age median 9 years, range 1 -20 years), 15 women and 11 men were analyzed (H12 cont. 2.00, Mortara Instruments, Milwaukee, WI). Seven of 26 patients were under -blocker medication. All patients were clinically identifyed as LQTS patients with score values of ⱖ 3, using the criteria of Schwartz et al (5). RR and QT values as well as beat rate corrected QTc values, using the Bazett correction formula, were automatically processed with the software of the Holter system. QTc values were averaged and displayed with median and quartiles thus eliminating data outliers. Additionally, data analysis was performed using software to investigate dynamics of cycle length mediated repolarization changes (6). Data of an appropriate lead were interactively processed by displaying adjacent beats superimposed on the workstation screen to qualify changes of the QRS– STT pattern, thus visualizing the recordings hour by hour. In all patients, genotyping for analysis of the underlying ion channel mutation was performed.
Results The number of successfully genotyped individuals was 9 of 26 (35%): LQT1 n ⫽ 6, LQT2 n ⫽ 2, and LQT3 n ⫽ 1. QT duration of the patients was prolonged in all patients under resting conditions. Circadian changes of the QT interval differed, varying from extremely prolonged to normal in the investigated LQTS patients. The averaged and heart rate corrected QT intervals showed a remarkably broad range with a median of QTc (Bazett) 498 (range, 400 – 630). In the 7 patients with -blocker medication, a similar range of QT intervals was typical, however, presenting lower mean heart rates: median QTc 504 (range, 420 –580); median of circadian heart rate 65 bpm in the -blocker treated group versus 74 bpm in the untreated patients. The pattern of QRS–STT behavior macroscopically showed little changes in all of the patients when leads were displayed superimposed and tracked hour by hour. In 3 of 26 patients, we found episodes with abrupt QT duration changes. These changes in the time domain were invariably accompanied by changes of T-wave morphology. The episodes were found in 2 patients with LQT2 and in one patient with LQT3, respectively. QTc values obtained, by using Bazett⬘s heart rate correction formula, provided an inappropriate picture of the QT/RR relation in all patients, as during periods of sinus node modulated heart rate changes–a typical behavior in young individuals– the variation of the RR intervals were overrepresented in the results of QTc calculations. Figure 1 is
Repolarization Analysis in LQTS •
Brockmeier et al. 211
Fig. 2. An episode with a singular premature ventricular beat in another LQTS patient. All QT intervals are pathologically prolonged (530 ms). The intermittend extrasystole, however, does not change the QT duration in the displayed cycles. A possible marginal additional prolongation can be seen in the adjacent beat after the extrasystole. The following cycle then shows again a QT interval of 530 ms. Lead II and V5, numbers are given in milliseconds.
showing an example of the inappropriateness of Bazett⬘s correction formula during periods of substantial RR interval changes. In Figure 1, the displayed QT intervals were all unchanged, however, pathologically prolonged. Only the amount of RR changes in the displayed time series is reflected in the QTc values as QT changes do not occur. Figure 2 shows an episode with a singular premature ventricular beat in another LQTS patient. All QT intervals are pathologically prolonged (530 ms). The intermittent extrasystole, however, does not change the QT duration in the displayed cycles. A possible marginal additional prolongation can be seen in the adjacent beat after the extrasystole. The following cycle then shows again a QT interval of 530 ms.
Discussion Our results show that a rather slow variability of the QT duration in consequence to substantial heart rate changes in LQTS patients is typical. This is a similar behavior as seen in healthy patients. All investigated patients had score values greater or equal to 3 according to the Schwartz criteria and had prolonged QT intervals. The use of beta-blocker did not modify repolarization duration. Beta-
blocker medication did slow the heart rate but not QT variability. A beat-to-beat variability of the QT to RR ratio is a rare finding, as most of the repolarization changes are very slow-also in a majority of patients with ion channel mutations. In the investigated LQTS patients abrupt macroscopic changes of RR/QT relation were detected in 3 patients only. This behavior was seen in 2 individuals with mutations of the HERG gene (LQT2) and in 1 patient with a mutation of the SCN5A sodium channel (LQT3). The fast changes of repolarization duration seen in these patients represented forms of T-wave alternans episodes. The slow reaction of the ventricular repolarization to rapid changes of heart rate is a typical finding also during exercise testing of LQTS patients. Exercise testing is used in the setting of potential candidates of LQTS to uncover QT prolongation after stress testing in patients with borderline QT duration under baseline conditions. The highest proportion of ion channel mutations among LQTS patients is found in the KVLQ1 gene, which is sensitive to exercise related torsades. In a publication by Dillenburg et al., it was shown that in comparison to healthy controls, patients with LQTS showed significant changes of QT prolongation not earlier than 3 minutes after cessation of exercise (7). These findings underline the typical slow reac-
212 Journal of Electrocardiology Vol. 36 Supplement 2003 tion of repolarizaton duration adaptation to heart rate changes in LQTS patients. Heart rate is the strongest variable influencing the repolarization duration and is typically coupled. Larger ranges of high QTc values are typical in symptomatic LQTS, but do not nessessarily represent real QT changes or substantial QT/RR short time changes.Ranges of heart rate corrected QT duration, as provided by Bazett⬘s QTc correction (8), may therefore not neccessarily reflect a higher instability/vulnerability of the ventricular repolarization– but may be heavily influenced by short time RR interval changes for instance in children. Time domain analysis is very probably inferior to algorithms, which quantify T-wave morphology changes or combine both approaches.
2.
3.
4.
5.
6.
7.
References 1. Vainer J, van der Steld B, Smeets JL, et al: Beat-to-beat behavior of QT interval during conducted supraven-
8.
tricular rhythm in the normal heart. PACE 17:1469, 1994 Rautaharju PM, Warren JW, Calhoun HP. Estimation of QT prolongation: A persistent, avoidable error in computer electrocardiography. J Electrocardiol 23: 111, 1991 (suppl) Eberle T, Brockmeier K, Heling G, et al. Prediction of normal QT intervals in children. J Electrocardiol 31: 121, 1999 (Suppl) Perkiomaki JH, Zareba W, Nomura A, et al: Repolarization dynamics in patients with long QT syndrome. J Cardiovasc Electrophysiol 13:651, 2002 Schwartz PJ, Moss AJ, Vincent GM, et al: Diagnostic criteria for the long QT syndrome: An update. Circulation 88:782, 1993 Brockmeier K, Aslan I, Hilbel T, et al: T-wave alternans in LQTS: Repolarization-rate dynamics from 12 lead Holter data. J Electrocardiol 34:93, 2001 Dillenburg RF, Hamilton RM: Is exercise testing useful in identifying congenital long QT syndrome? Am J Cardiol 89:233, 1994 Kligfield P, Lax KG, Okin PM. QTc behavior during treadmill exercise as a function of the underlying QT-heart rate relationship. J Electrocardiol 28:206, 1995 (suppl)