Journal of the American College of Cardiology Vol. 47, No. 1, by the American College of Cardiology Foundation ISSN /06/$32.

Transcription

1 Journal of the American College of Cardiology Vol. 47, No. 1, by the American College of Cardiology Foundation ISSN /06/$32.00 Published by Elsevier Inc. doi: /j.jacc Ratio of Late to Early T-Wave Peak Amplitude in 24-h Electrocardiographic Recordings as Indicator of Symptom History in Patients With Long-QT Syndrome Types 1 and 2 Matti Viitasalo, MD,* Lasse Oikarinen, MD,* Heikki Swan, MD,* Kathryn A. Glatter, MD, Heikki Väänänen, MSC, Heidi Fodstad, MSC, Nipavan Chiamvimonvat, MD, Kimmo Kontula, MD, Lauri Toivonen, MD, FACC,* Melvin M. Scheinman, MD, FACC Helsinki and Espoo, Finland; and Davis and San Francisco, California OBJECTIVES BACKGROUND METHODS RESULTS CONCLUSIONS We tested the hypothesis that in long-qt syndrome types 1 (LQT1) and 2 (LQT2), the diurnal maximal ratio between late and early T-wave peak amplitudes correlates with a history of symptoms better than QT interval durations. Genotype and phenotype studies have delineated clinical profiles of the most prevalent LQT1 and LQT2 subtypes of inherited LQT, but prediction of arrhythmia risk remains uncertain, the baseline QTc interval being the best predictor. In experimental long-qt syndrome models, the ratio between late and early T-wave peak amplitude predicts onset of torsade de pointes. We reviewed 24-h electrocardiographic recordings from 214 genotyped subjects 97 with LQT1, 62 with LQT2, and 55 unaffected to record maximal amplitude ratios between late and early T-wave peaks by use of a computer-assisted program. Maximal amplitude ratios between late and early T-wave peaks were higher in symptomatic than in asymptomatic patients both in LQT1 ( vs ; p 0.001) and LQT2 patients ( vs ; p 0.001). Although the QTc interval also was longer in symptomatic patients, only the maximal amplitude ratio between late and early T-wave peaks was independently associated with symptoms in both LQT1 and LQT2 patients. Maximal diurnal ratio between late and early T-wave peak amplitude improves noninvasive risk assessment both in LQT1 and LQT2 syndromes. We propose this new indicator in clinical evaluation of arrhythmia risk in LQT1 and LQT2. (J Am Coll Cardiol 2006;47: ) 2006 by the American College of Cardiology Foundation Torsades de pointes (TdP) is syncope-producing arrhythmia causing risk of sudden death in patients with inherited long-qt syndrome (LQT) (1). In experimental models of the LQT, prolonged repolarization, transmural dispersion of repolarization, and early afterdepolarizations are the three electrophysiologic components necessary to trigger and sustain TdP (2). Although studies of the genotype and phenotype have delineated clinical profiles of the most prevalent long-qt syndrome type 1 (LQT1) and type 2 (LQT2) subtypes of inherited LQT, estimation of arrhythmia risk within either genotype has been difficult. Of the three electrophysiologic elements of TdP, prolonged repolarization, measured as a heart rate-corrected QTc interval of 500 ms or more in 12-lead electrocardiogram (ECG), may assess arrhythmia risk (3). On the other hand, transmural dispersion of repolarization, as estimated clinically by the T-wave peak to T-wave end interval, does not help in From the Departments of *Cardiology and Medicine, Helsinki University Central Hospital, Helsinki, Finland; Department of Cardiology, University of California, Davis, California; Laboratory of Biomedical Engineering, Helsinki University of Technology, Espoo, Finland; and the Department of Medicine, Cardiac Electrophysiology, University of California, San Francisco, California. Supported by grants from the Finnish Foundation for Cardiovascular Research, the Finnish Academy, and the Sigrid Juselius Foundation, Helsinki, Finland. Manuscript received April 16, 2005; revised manuscript received July 14, 2005, accepted July 25, differentiating symptomatic and asymptomatic patients (4). Recently, Gbadebo et al. (5) suggested the ratio of U- to T-wave amplitude as the clinical counterpart of early afterdepolarizations, and a progressive increase in the ratio of U- to T-wave preceded the onset of TdP in an experimental model of LQT. In addition, Jackman et al. (6) suggested previously that the increment in U-wave amplitude after a premature ventricular beat is a marker for arrhythmia risk in pause-dependent LQT. In this report we prefer to use T2 (rather than U) waves, based on the prior observation of Yan and Antzelevitch (7) suggesting that in bifid T-wave the late component (T2) is due to same forces as those responsible for the T-wave whereas other forces produce the normal U-wave. Others also have emphasized the difference between the T2 and physiologic U waves (8). The objective of this study was to test the hypothesis that in LQT1 and LQT2 subtypes of LQT the diurnal maximal ratio between late and early T-wave peak amplitude correlates with a history of symptoms better than baseline QTc or diurnal maximal QT intervals. METHODS Study subjects. We reviewed 24-h EGC recordings in 214 individuals with either the LQT1 (97 patients) or the LQT2 genotype (62 patients) or unaffected (55 patients),

2 JACC Vol. 47, No. 1, 2006 January 3, 2006: Viitasalo et al. T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk 113 Abbreviations and Acronyms ECG electrocardiogram/electrocardiographic LQT long-qt syndrome LQT1 long-qt syndrome type 1 LQT2 long-qt syndrome type 2 TdP torsades de pointes studied at the Helsinki University Central Hospital- Finland, the University of California-San Francisco, or the University of California-Davis. Holter recordings were done as primary investigations in symptomatic patients suspected to have an LQTS and in their family members and were included in the series consecutively after the genotypes became available. Baseline ECGs and 24-h recordings were obtained at the same visit. There were 13 different KCNQ1 mutations in the LQT1 group and 16 different HERG mutations in the LQT2 group. The patients were categorized as symptomatic if they had experienced cardiac arrest or unexplained syncope associated with specific triggering conditions shown in Table 1. Holter recordings were recorded before initiation of beta-blockers, and no subject took any other medication known to influence cardiac repolarization. All subjects were in sinus rhythm, did not show bundle-branch block, and did not have symptoms or signs of other cardiac disease on clinical examination. Normal daily activities were encouraged during the recordings. The ethical review committees of the institutes approved the study, and informed consent was obtained from all participants. Holter recordings and analyses. All study subjects underwent a 24-h ECG recording (model 8500, Marquette Electronics, Milwaukee, WI). The tapes were initially analyzed with a Marquette 8000 Holter Analysis system (version 5.8 software) to label the QRS complexes as normal, ventricular extrasystoles, or aberrant complexes. The ECG data were then transferred to a personal computer for further analysis of T waves and QT intervals. Measurement of the T1 and T2 peaks of the T-wave and the durations of QT intervals. We first reviewed the morphology of T waves in the 24-h ECG recordings using the superimposed scan. When two distinct measurable positive T-wave peaks according to grade III in the approach presented by Lehman et al. (9) were present, the first sample with a T2- to T1-wave amplitude ratio of more than 1 was printed on chart paper. Later in the analysis of each case, samples with the so-far-highest T2-wave amplitudes and with the so-far-highest T2- to T1-wave amplitude ratios were all printed on chart paper. Finally, the maximal values of these printed samples of the highest T2-wave amplitudes and the highest T2 to T1 amplitude ratios were included in the results. Both T1 and T2 waves were required to be 0.1 mv in amplitude and the deflection after the T1-wave to be down-sloping or horizontal before the takeoff point of the T2-wave for the strip to be included in the measurements. We analyzed the ratio of T2- to T1- Table 1. Clinical and Baseline ECG Characteristics of Study Subjects LQT1 (n 97) LQT2 (n 62) Symptomatic (n 29) p Value Asymptomatic (n 33) Symptomatic (n 33) p Value Asymptomatic (n 64) LQT1 LQT2 (n 159) p Value Unaffected (n 55) Age, yrs Male gender, n (%) 29 (53) 70 (44) 29 (45) 11 (33) 0.26* 22 (67) 8 (28) 0.002* Cardiac events, n (%) Cardiac arrest 3 (9) 4 (14) Syncope 30 (91) 25 (86) Triggers for cardiac events, n (%) Exercise 22 (67) 5 (17) Emotion/arousal 8 (24) 14 (48) Sleep or rest 1 (3) 6 (21) Other 2 (6) 4 (14) Baseline ECG Heart rate (beats/min) QTc interval (ms) Male: QTc interval (ms) Female: QTc interval (ms) Bifid T-wave, n (%) Obvious bifid T-wave 2 (4) 18 (11) 0.09* 4 (6) 1 (3) (12) 9 (31) 0.07* Subtle bifid T-wave 7 (13) 50 (31) 0.007* 12 (19) 5 (15) 0.66* 20 (61) 13 (45) 0.21* *Chi-square; Mann-Whitney U; Fisher exact test. ECG electrocardiogram; LQT1 long-qt syndrome type 1; LQT2 long-qt syndrome type 2.

3 114 Viitasalo et al. JACC Vol. 47, No. 1, 2006 T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk January 3, 2006: Figure 1. Electrocardiographic signals before, at times of maximal T2- to T1-wave amplitude ratios, and after the events of maximal T2- to T1-wave amplitude ratio in an unaffected subject (first panel), a long-qt syndrome type 1 (LQT1) patient (second panel), and an LQT2 patient (third panel). The two lower panels with a continuous electrocardiographic signal during a 20-s period show an episode of a prominent T2-wave in an LQT1 patient (modified lead V 5 ). Note that the T2-wave is present in 24 beats only, the arrows showing the first and the last beat with the T2-wave. wave irrespective of the time interval between T1 and T2 waves (Fig. 1). Finally, we measured the maximal T2 to T1 amplitude ratios manually from the ECG strips printed on chart paper at the speed of 50 mm per second and with the amplitude calibration of 0.1 mv/mm using modified lead V 5. All the highest T2 waves and T2 to T1 amplitude ratios were determined and checked visually by use of the unaveraged ECG signal. We recorded the maximal T2 to T1 amplitude value as the mean of five consecutive beats for each individual if 1.1; otherwise we used the value of 1 as the individual maximum. The coefficient of variation for determination of the maximal T2 to T1 amplitude ratio was 5.7%. Pause-induced T2 to T1 amplitude ratios 1 were measured from one beat after any pause and recorded separately if 1.1. To measure the QT interval durations from the 24-h ECG recordings, we used a previously described method for determination of the QT intervals (10). In this study, we determined the highest amplitude peak of the deflections during repolarization as the peak of the T-wave. In QT interval measurements, bifid T waves exhibiting a time interval of 0.15 s between the highest peak and the later lower peak were calculated into the QT duration; otherwise, the later lower peak was not included into the QT duration (9). Thus, QT interval measurements always include T2 waves that were of higher amplitude than T1 waves. We recorded diurnal maximal QT peak, QT end, and T-wave peak to T-wave end intervals at specified RR intervals, determined with 3 consecutive acceptable measurements. QT end intervals were recorded also at stable heart rates (11). All analyses were done without the investigator knowing the genotypes or presence of symptoms. The baseline QT interval adjusted for heart rate (QTc) was measured in lead II from 12-lead ECGs using Bazett s formula. Bifid T waves in baseline ECGs (recorded at a paper speed of 50 mm/s) were classified as obvious or subtle according to Zhang et al. (12). Statistical analyses. All continuous data are presented as mean SD. Mean values were compared between groups by use of Mann-Whitney U tests. Proportions were compared by chi-square or Fisher exact tests as appropriate. The strength of correlation between continuous parameters was determined by Spearman correlation coefficients. To analyze the association of maximal T2- to T1-wave amplitude ratio and QTc interval with the LQT-related symptoms, we created genotype-specific tertiles of both parameters and related symptom history within the genotype to these ordered tertiles using chi-square test for trend. Logistic regression analysis was used to calculate odds ratios and 95% confidence intervals to estimate the relative increase in the risk of symptoms in the higher tertiles as compared with the lowest tertile (referent). To identify independent predictors of symptom history within LQT1 and LQT2 genotypes, age, gender, QTc interval, and maximal T2 to T1 amplitude ratio were included in a multivariate stepwise logistic regression analyses. Two-tailed p 0.05 was considered statistically significant. RESULTS Duration of QT intervals. In both LQT1 and LQT2 patients, baseline QTc intervals were significantly longer in symptomatic than in asymptomatic patients (Table 1). Diurnal maximal QT intervals distinguished between symp-

4 Table 2. Heart Rates, QT Intervals, T-Wave Amplitudes, and T2- to T1-Wave Amplitude Ratios (Mean SD) of the 214 Study Subjects During 24-h ECG Recordings Unaffected (n 55; Male 29) LQT1 LQT2 (n 159; Male 70) p Value Asymptomatic (n 64; Male 29) LQT1 (n 97) LQT2 (n 62) Symptomatic (n 33; Male 11) p Value Asymptomatic (n 33; Male 22) Symptomatic (n 29; Male 8) p Value Measure Minimal heart rate (beats/min) Maximal heart rate (beats/min) Maximal QT peak interval (ms) Maximal QT end interval (ms) Maximal T-wave peak to T-wave end interval (ms) Maximal T2-wave amplitude (mv) T1-wave amplitude (mv), at the time of maximal T2- to T1-wave amplitude ratio Maximal T2- to T1-wave amplitude ratio Mean Range Male: mean Female: mean Pause induced maximal T2 to T1 amplitude ratio Mean Range Abbreviations as in Table 1. JACC Vol. 47, No. 1, 2006 Viitasalo et al. January 3, 2006: T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk 115

5 116 Viitasalo et al. JACC Vol. 47, No. 1, 2006 T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk January 3, 2006: Figure 2. Data from 24-h electrocardiographic recordings during normal daily activities in the long-qt syndrome type 1 (LQT1) and 2 (LQT2) groups showing maximal QT intervals (two higher lines) and QT intervals at stable heart rates (two lower lines) in symptomatic (solid lines) and asymptomatic (broken lines) patients. Separate data points show the QT versus RR intervals at the moment of maximal T2- to T1-wave amplitude ratios in symptomatic (closed symbols) and asymptomatic (open symbols) patients. tomatic and asymptomatic patients in the LQT1 patient group but not in the LQT2 patient group (Table 2). Figure 2 shows that QT intervals determined at stable heart rates were similarly shortened in symptomatic and asymptomatic patients in both patient groups when heart rate increased. In further analyses, from QT parameters the baseline QTc interval was used because it differentiated best between symptomatic and asymptomatic patients. Bifid T waves in baseline ECGs. In baseline ECGs, bifid T waves were observed in 9 out of 55 (16%) unaffected subjects, in 22 out of 97 (23%) LQT1 patients, and in 46 out of 62 (74%) LQT2 patients (p 0.001). An obvious bifid T-wave occurred in 2 (4%) unaffected subjects and in 5 (5%) LQT1 and 13 (21%) LQT2 patients (p 0.001) (Table 1). Maximal T2- to T1-wave amplitude ratio during 24-h ECG recordings. The T2- to T1-wave amplitude ratio of more than 1 was observed in 39 out of 55 (71%) unaffected subjects, in 91 out of 97 (94%) LQT1 patients, and in 58 out of 62 (94%) LQT2 patients (p 0.001). Amplitude ratios of more than 2 were observed in 3 (6%) unaffected subjects, in 65 (67%) LQT1 patients, and in 28 (45%) LQT2 patients (p 0.001) (Fig. 3). The episodes with the amplitude ratio of T2- to T1-wave of more than 2 lasted 5 2 s in unaffected subjects, s in LQT1 patients, and s in LQT2 patients (Fig. 1). Pause-induced T2- to T1-wave amplitude ratio of more than 1 was observed in 5 unaffected subjects (9%), in 17 (18%) LQT1 patients, and in 34 (55%) LQT2 patients (p 0.001). At the time of the maximal T2- to T1-wave amplitude ratios, the average heart rates were beats/min in unaffected subjects, beats/min in LQT1 patients, and beats/min in LQT2 patients (Fig. 2). The maximal T2- to T1-wave amplitude ratio of more than 1 occurred during daytime in 34 out of 39 (87%) unaffected subjects, in 84 out of 91 (92%) LQT1 patients, and in 31 out of 58 (53%) LQT2 patients (p 0.001). At the time of the maximal T2 to T1 ratio, LQT2 patients exhibited a heart rate acceleration in 76% of cases, whereas both unaffected and LQT1 patients showed an acceleration in 51% of cases (p 0.005). The maximal T2 to T1 amplitude ratios correlated with baseline QTc intervals in both LQT1 patients (r 0.39; p 0.001) and LQT2 patients (r 0.35; p 0.005) but not in unaffected subjects (r 0.11; p 0.46). In cases with no bifid T-wave and with an obvious bifid T-wave in the baseline ECG, the maximal T2 to T1 amplitude ratios were and (p 0.61) in LQT1 patients, and and (p 0.001) in LQT2 patients. Symptomatic versus asymptomatic patients. Among LQT1 patients the maximal T2- to T1-wave amplitude ratio was in symptomatic and in asymptomatic patients (p 0.001), symptomatic female patients showing the highest values (Table 2). Among LQT2 patients the corresponding values were in symptomatic and in asymptomatic subjects (p 0.001). In further analyses of the association of the maximal T2 to T1 amplitude ratio to the history of symptoms, we found that the proportion of symptomatic patients increased from 13% in the lowest T2 to T1 amplitude ratio tertile (1.0 to 2.0) to 58% in the highest tertile (3.0 to 5.2) in LQT1 patients (p for trend). Among LQT1 patients those with a T2 to T1 amplitude ratio in the highest tertile had a risk of symptoms that was increased by a factor of 9.7 (95% confidence interval 2.7 to 34) as compared with those with a T2 to T1 amplitude ratio in the lowest tertile. In comparison, those LQT1 patients with a QTc in the highest tertile (480 to 615 ms) had a 2.9-fold risk of symptoms (95% confidence interval 1.03 to 8.4) as compared with those with a QTc in the lowest tertile (380 to 450 ms) (Fig. 4). In LQT2 patients the proportion of symptomatic patients increased from 20% in the lowest T2 to T1 amplitude ratio tertile (1.0 to 1.5) to 68% in the highest tertile (2.4 to 4.7) (p for trend). Among

6 JACC Vol. 47, No. 1, 2006 January 3, 2006: Viitasalo et al. T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk 117 Figure 3. (Top panels) High T2- to T1-wave amplitude ratios during 24-h electrocardiographic (ECG) recordings in an unaffected subject (a), in three long-qt syndrome type 1 (LQT1) patients (b, c, and d), and in an LQT2 patient (e) (paper speed 25 mm/s). Baseline ECG recordings (leads V 1 to V 6, paper speed 50 mm/s) of the same patients are shown in the lower panels. Note that only the LQT2 patient exhibits bifid T waves in the baseline ECG recording. LQT2 patients those with a T2 to T1 amplitude ratio in the highest tertile had an 8.6-fold greater risk of symptoms (95% confidence interval 2.1 to 35), whereas those with a QTc in the highest tertile (480 to 618 ms) had a 3.7-fold greater risk of symptoms (95% confidence interval 1.02 to 14), as compared with those with a T2 to T1 amplitude ratio and QTc in the lowest tertile (402 to 441 ms), respectively (Fig. 4). Selected cutoff values for maximal T2 to T1 amplitude ratios of 3.0 in LQT1 and 2.4 in LQT2 patients identified 18 symptomatic patients from the LQT1 group (sensitivity 0.55) with 0.80 specificity and 15 symptomatic patients from the LQT2 group (sensitivity 0.52) with 0.79 specificity, respectively. Pause-induced T2- to T1-wave amplitude ratio of more than 1 was present in 8 out of 33 (24%) symptomatic and in 9 out of 64 (14%) asymptomatic LQT1 patients (p 0.21) and in 16 out of 29 (55%) symptomatic and in 18 out of 33 (55%) asymptomatic LQT2 patients (p 0.96) (Table 2). The presence of bifid T waves in baseline ECGs did not differentiate between symptomatic and asymptomatic patients among either LQT1 or LQT2 patients (Table 1). To assess the significance and independence of the predictors of symptoms within LQT1 and LQT2 genotypes, we included age, gender, baseline QTc, and maximal T2 to T1 amplitude ratio in multivariate logistic regression analyses. The analyses showed that within the LQT1 genotype, only the maximal T2 to T1 ratio was independently associated with symptoms (p 0.001). Within the LQT2 genotype both maximal T2 to T1 ratio (p 0.001) and female sex (p 0.003) were independently associated with symptom history. DISCUSSION Main findings. Our results show that the diurnal maximal ratio between late and early T-wave peak amplitude correlates with a history of LQT-related symptoms better than the baseline QTc interval in both LQT1 and LQT2

7 118 Viitasalo et al. JACC Vol. 47, No. 1, 2006 T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk January 3, 2006: Figure 4. Odds ratios and 95% confidence intervals for the risk of long-qt syndrome (LQT)-related symptoms, according to the tertile of maximal T2- to T1-wave amplitude ratio and the QTc interval. The y axis is on a log scale. The reference group is tertile 1. (A) In LQT type 1 (LQT1) patients, the maximum T2 to T1 amplitude ratio was 3 or more in tertile 3 and 2 or less in tertile 1. (B) In LQT type 2 (LQT2) patients, the corresponding ratio was 2.4 or more in tertile 3 and 1.5 or less in tertile 1. patients. In both genotypes, multivariate analyses showed that only the diurnal maximal ratio between late and early T-wave peak amplitude was independently associated with the symptom history. Therefore, in these patient groups, diurnal maximal T2- to T1-wave amplitude ratio is the best available noninvasive test to assess the risk of symptoms. The frequency of abrupt diurnal episodes with the ratio of T2- to T1-wave amplitude of more than 1 is higher than previous reports using averaged T-wave templates have suggested (13) and do occur in the majority of subjects. In LQT1 and LQT2 patients, diurnal distributions of maximal T2- to T1-wave amplitude ratios are similar to reported corresponding distributions of cardiac events (14). The T2-wave in experimental models of TdP. In an experimental LQT2 model, Gbadebo et al. (5) demonstrated that the ratio of U-wave (our T2) to T-wave (our T1) amplitude could be used as an ECG parameter to predict the acute onset of TdP. Earlier, Mazur et al. (15) presented examples of a large U-wave (our T2) preceding the onset of TdP, suggesting at least a temporal relationship between the two. The hypothesis that TdP and enhanced U waves (our T2) are initiated by early afterdepolarizations links abnormal U waves as a trigger for TdP. Our finding that high T2 to T1 amplitude ratio was independently associated with symptoms when measured in nonacute situations shows that this parameter carries a sign of the susceptibility for TdP in LQT1 and LQT2 patients. In our material, other causes of syncope than TdP are unlikely in view of the absence of cardiac disease or associated recurrent vasodepressor symptoms. The T2-wave in LQT1 and LQT2. We found that the highest T2 to T1 amplitude ratios occurred near the heart rate level of 100 beats/min. Although many of the LQT2 patients exhibited their highest T2 to T1 amplitude ratios during nighttime, most of these patients showed the highest T2 to T1 amplitude ratios associated with a heart rate acceleration. Thus, at the time of the maximal T2 to T1 amplitude ratios, a moderate level of preceding sympathetic activity may be involved in both LQT1 and LQT2 patients. Accordingly, Shimizu et al. (16) reported previously that isoproterenol induces early afterdepolarizations and electrocardiographic TU complex abnormalities in clinically diagnosed LQT patients. Noda et al. (17), by using epinephrine infusion, described a similar T2-wave as we observed in an LQT1 patient, whereas Takenaka et al. (18) reported a T2-wave appearance, very similar to those in our patients, in an LQT2 patient during the exercise test. Of note is also that acceleration-induced early afterdepolarizations presented by Burashnikov and Antzelevitch (19) in their experimental LQT2 model, and the typical clinical association of LQT2-related symptoms with arousals (20), predicts the appearance of high T2 to T1 ratios in LQT2 patients during heart rate accelerations. Recently, Nakagava et al. (21) showed that isoproterenol-induced transient TU abnormalities in normal subjects quite similar to those seen in patients with LQT. According to the sympathetic stimulation tests in both LQT1 and LQT2 patients (17,18) and in experimental LQT2 models (19), the present findings show the capacity of both LQT1 and LQT2 patients to transiently and dramatically produce high T2 waves. Previously Lupoglazoff et al. (13) observed T2- to T1- wave amplitude ratios of more than 1 in 63% of their LQT2 patients but in no LQT1 patient or control subject in 24-h ECG recordings. They classified QRS-T complexes according to the value of the preceding RR interval every 25 ms and then averaged the thousands of QRS-T complexes, whereas we used the unaveraged ECG signal. Our study s strikingly higher prevalence of T2- to T1-wave amplitude ratios of more than 1, especially in LQT1 patients, highlights the abrupt nature of this phenomenon. On the other hand, our observation that the majority of LQT2 patients but only few LQT1 patients show bifid T waves in the

8 JACC Vol. 47, No. 1, 2006 January 3, 2006: Viitasalo et al. T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk 119 baseline ECG is in accordance with previous studies (12,13). Our results in LQT2 patients show further that the presence of an obvious bifid T-wave in the baseline ECG suggests a tendency to show a high T2- to T1-wave amplitude ratio in 24-h recordings. Under fluctuation in autonomic nervous activity the presence of I Ks (LQT1 patients) and I Kr (LQT2 patients) defects may modulate the effects of repolarization transients, leading to abrupt episodes of T2- to T1-wave amplitude ratio of even more than 2 in both LQT1 and LQT2 patients, as observed in this study. Patients with longest baseline QTc intervals showed highest T2 to T1 amplitude ratios. We observed, however, that maximal T2 to T1 amplitude ratios associated poorly with the diurnal maximal QT durations. Thus, long baseline QT duration shows the susceptibility to experience high T2 to T1 amplitude ratios, but QT duration does not determine the appearance of high T2 waves. These findings are in accordance with previous experimental studies by Burashnikov and Antzelevitch (19), who observed that the appearance of heart rate acceleration-induced early afterdepolarizations was accompanied by either an abbreviation or a prolongation of action potentials. The T2-wave and other symptom indicators in LQT1 and LQT2. Our finding that in LQT1 and LQT2 the high T2 to T1 amplitude ratio is an even more important indicator of symptoms than the baseline QTc interval is consistent with the experimental idea that prolonged repolarization per se is not the proximate cause of the initiation of TdP arrhythmia (5). The facts that diurnal maximal QT intervals did not clearly differ between symptomatic and asymptomatic patients in either patient group and that QT intervals shortened similarly in symptomatic and asymptomatic patients with increasing heart rates obscure further the importance of QT duration as a sign of the risk of symptoms. Pause-induced T2 to T1 amplitude ratio of more than 1, on the other hand, seems to be typical of LQT2 patients, but the association of this measure with the history of symptoms is not clear. In agreement with the study by Lupoglazoff et al. (13) we did not find a correlation between the presence of obvious bifid T waves in the baseline ECG and the history of symptoms in either LQT1 or LQT2 patients. Also in the present study, female sex showed an independent correlation with symptoms among LQT2 patients, emphasizing the importance of gender difference as an arrhythmia risk factor in LQT2, as recently pointed out by Priori et al. (3). Study limitations. Our study was a retrospective analysis of 24-h ECG recordings collected in three different centers in Finland and in California. It should be noted, however, that the recordings were done before the beta-blocking medication was started and that a prospective study without a beta-blocking medication in symptomatic patients might be unethical. The observations derived from this study must be validated in a second cohort of similar patients. We compared symptomatic patients with similar patients who had not yet developed symptoms. However, the ages of symptomatic and asymptomatic patients were similar. Moreover, the mean age of the asymptomatic patients was over 30 years, and the majority of patients who are destined to develop symptoms will become symptomatic before the age of 30 years (22). The T-wave morphology is likely to vary with body position (23), and we do not know what exactly the patients were doing at the time of the maximal T2 to T1 amplitude ratios. However, during an ECG recording of 24 h, each ambulatory subject may use the most usual body positions, and body position-related ECG changes might be more sudden than the gradual appearance and disappearance of prominent T2-wave shown in Figure 1. Thus, body positions may not explain the observed high T2 waves. The current practical limitation of the proposed approach is that it is time consuming, and therefore computerized software specifically to detect maximal T2 amplitude and maximal T2 to T1 amplitude ratio from 24-h ECG recordings would be helpful. Clinical implications. Our retrospective study encourages clinicians to include careful analyses of the diurnal maximal T2 to T1 amplitude ratio along with QT durations in consideration of the arrhythmia risk for asymptomatic patients with known type 1 or 2 LQT genotype. In LQT1 patients, a maximal diurnal T2 to T1 amplitude ratio of 3 or higher suggests a high probability of being symptomatic whereas a ratio of 2 or lower suggests a low probability. Among LQT2 patients, ratios suggestive of being symptomatic or asymptomatic are, respectively, 2.4 or higher and 1.5 or lower. Further studies are needed to evaluate the prospective value of the maximal T2 to T1 amplitude ratio in asymptomatic patients with LQT1 and LQT2. Further studies are also needed to evaluate the benefit of this measure in predicting the patient response of the betablocker therapy in these patient groups. Reprint requests and correspondence: Dr. Matti Viitasalo, Department of Cardiology, University Central Hospital, Haartmaninkatu 4, Helsinki, Finland. matti.viitasalo@ hus.fi. REFERENCES 1. Schwartz PJ, Priori SG, Napolitano C. The long QT syndrome. In: Zipes DP, Jalife J, editors. Cardiac Electrophysiology: From Cell to Bedside. 3rd edition. Philadelphia, PA: Saunders, 2000: Ebert SN, Liu X-K, Woosley RL. Gender as a risk factor for acquired torsades de pointes. In: Franz MR, editor. Monophasic Action Potentials. Bridging Cell and Bedside. New York, NY: Futura, 2000: Priori SG, Schwartz PJ, Napolitano C, et al. Risk stratification in the long-qt syndrome. N Engl J Med 2003;348: Viitasalo M, Oikarinen L, Swan H, et al. Ambulatory electrocardiographic evidence of transmural dispersion of repolarization in patients with long-qt syndrome type 1 and 2. Circulation 2002;106: Gbadebo TD, Trimble RW, Khoo MSC, Temple J, Roden DM, Anderson ME. Calmodulin inhibitor W-7 unmasks a novel electrocardiographic parameter that predicts initiation of torsade de pointes. Circulation 2002;105:770 4.

9 120 Viitasalo et al. JACC Vol. 47, No. 1, 2006 T2- to T1-Wave Amplitude Ratio and Arrhythmia Risk January 3, 2006: Jackman WM, Friday KJ, Anderson JL, Aliot EM, Clark M, Lazzara R. The long QT syndromes: a critical review, new clinical observations and a unifying hypothesis. Prog Cardiovasc Dis 1988;31: Yan G-X, Antzelevitch C. Cellular basis for the normal T-wave and the electrocardiographic manifestations of the long-qt syndrome. Circulation 1998;98: Viskin S, Zeltser D, Antzelevitch C. When U say U waves, what do U mean? Pacing Clin Electrophysiol 2004;27: Lehman MH, Suzuki F, Fromm BS, et al. T-wave humps as a potential electrocardiographic marker of the long-qt syndrome. J Am Coll Cardiol 1994;24: Oikarinen L, Paavola M, Montonen J, et al. Magnetocardiographic QT interval dispersion in postmyocardial infarction patients with sustained ventricular tachycardia: validation of automated QT measurements. Pacing Clin Electrophysiol 1998;21: Viitasalo M, Oikarinen L, Väänänen H, et al. Differentiation between LQT1 and LQT2 patients and unaffected subjects using 24-hour electrocardiographic recordings. Am J Cardiol 2002;89: Zhang L, Timothy KW, Vincent GM, et al. Spectrum of ST-T-wave patterns and repolarization parameters in congenital long-qt syndrome. ECG findings identify genotypes. Circulation 2000;102: Lupoglazoff JM, Denjoy I, Berthet M, et al. Notched T waves on Holter recordings enhance detection of patients with LQT2 (HERG) mutations. Circulation 2001;103: Schwartz PJ, Priori SG, Spazzoli C, et al. Genotype-phenotype correlation in the long QT syndrome. Gene-specific triggers for life-threatening arrhythmias. Circulation 2001;103: Mazur A, Roden DM, Anderson ME. Systemic administration of calmodulin antagonist W-7 or protein kinase A inhibitor H-8 prevents torsade de pointes in rabbits. Circulation 1999;100: Shimizu W, Ohe T, Kurita T, et al. Early afterdepolarizations induced by isoproterenol in patients with congenital long QT syndrome. Circulation 1991;84: Noda T, Takaki H, Kurita T, et al. Gene-specific response of dynamic ventricular repolarization to sympathetic stimulation in LQT1, LQT2 and LQT3 forms of congenital long QT syndrome. Eur Heart J 2002;23: Takenaka K, Ai T, Shimizu W, et al. Exercise stress test amplifies genotype-phenotype correlation in the LQT1 and LQT2 forms of the long-qt syndrome. Circulation 2003;107: Burashnikov A, Antzelevitch C. Acceleration-induced action potential prolongation and early afterdepolarizations. J Cardiovasc Electrophysiol 1998;9: Wilde AAM, Jongbloed RJE, Doevendans PA, et al. Auditory stimuli as a trigger for arrhythmic events differentiate HERG-related (LQTS 2 ) patients from KVLQT1-related patients (LQTS 1 ).JAm Coll Cardiol 1999;33: Nakagava M, Ooie T, Ou B, et al. Gender differences in autonomic modulation of ventricular repolarization in humans. J Cardiovasc Electrophysiol 2005;16: Zareba W, Moss AJ, Locati EH, et al. Modulating effects of age and gender on the clinical course of long QT syndrome by genotype. J Am Coll Cardiol 2003;42: Krasnow AZ, Bloomfield DK. Artifacts in portable electrocardiographic monitoring. Am Heart J 1976;91: