T_peak–T_end interval and T_peak–T_end/QT ratio in patients with Brugada syndrome

Aims

Brugada syndrome (BrS) is characterized by a typical electrocardiogram (ECG) pattern in right precordial leads (V1–V3; so-called type 1 ECG) and an increased risk of sudden cardiac death due to ventricular fibrillation. Annual cardiac event rates vary from 0.5% in asymptomatic to 7.7% in high-risk patients. So far, spontaneous occurrence of the type 1 ECG, survived cardiac arrest, and/or documented ventricular arrhythmias are main risk predictors, whereas other factors (e.g. family history or genotype) are not applicable for risk stratification. In this study, we investigated the relationship between T_peak–T_end intervals (TpTe) as a novel ECG parameter for the occurrence of cardiac arrhythmias.

Methods and results

Clinical and genetic data of 78 unrelated BrS patients (male: n = 57, age: 45 ± 14 years) were retrospectively analysed for medical history, gene mutation, and ECG parameters (in particular heart rate, PQ, QRS, QT, and TpTe) as obtained after digital measurements. TpTe in ECG lead V1 (87 ± 30 vs. 71 ± 27 ms; P = 0.017) and the TpTe/QT ratio (0.24 vs. 0.19; P = 0.018) were significantly higher in high-risk BrS patients than in other BrS patients. In the other right precordial leads typically indicative for BrS, no significant difference was noted.

Conclusion

Assessment of the TpTe interval or the TpTe/QT ratio in lead V1 is potentially useful as a non-invasive risk marker for BrS patients with life-threatening arrhythmias.

Introduction

The Brugada syndrome (BrS) is a primary electrical heart disease that is characterized by a typical electrocardiographic pattern [type 1 electrocardiogram (ECG)] in the absence of structural cardiac or other mimicking abnormalities and predisposes to sudden cardiac death (SCD) secondary to polymorphic ventricular tachycardia (VT) or ventricular fibrillation (VF).^1,2 The estimated prevalence ranges from 1 to 5/10 000 in Europe up to 12/10 000 inhabitants in Southeast Asia and it is more common among young men (8:1) than women.² The mean age at diagnosis is 40–45 years and ECG changes may occur age-dependent and therefore making it rare during childhood. Patients with BrS are mainly symptomatic during a higher vagotonic state, i.e. at sleep, rest, after large meals, or during fever. The major genetic cause (∼20%) is due to loss-of-function mutations in the cardiac sodium channel gene SCN5A (subtype BRGDA1). Other related phenotypes due to loss-of-function mutations in SCN5A can be cardiac conduction disease (CCD), sinus node dysfunction or vice versa, long QT syndrome due to a gain-of-function mutation (LQTS; subform LQT3); in some patients with BrS, a phenotypic overlap can be seen. Recently, the diagnostic criteria for BrS have been updated in an expert consensus statement on the diagnosis and management of patients with inherited primary arrhythmia syndromes. In summary, the presence of a type 1 ECG in a single ECG lead (either fixed in normal or higher intercostal spaces) in the absence of phenocopies might be sufficient for diagnosis.

Risk stratification in patients with BrS is still challenging and controversial. It is well known that patients with severe cardiac events (i.e. documented VT, VF, or aborted SCD) are at a higher risk for recurrent lethal arrhythmic episodes and in nearly all cases, an implantable cardioverter defibrillator (ICD) implantation is recommended. In contrast, asymptomatic patients or patients with a history of syncope alone are believed to be on a low-to-moderate risk. So far, the family history (e.g. with SCD) or the presence of a mutation in the SCN5A gene has no significant influence on the risk for recurrent cardiac episodes.³ In consequence, asymptomatic patients typically do still not qualify for an ICD implantation as their life-threatening events are very low.

Therefore, it might be very useful to identify additional markers for risk stratification. In this line, signal-averaged ECG (late potentials), fragmented QRS complexes, QRS duration, and ventricular effective refractory period (ERP) <200 ms recorded during electrophysiological studies as well as an HV interval >60 ms are currently being considered for risk stratification.^3–5 Among other risk indicators, QRS duration in BrS patients and the occurrence of cardiac events are still controversial.⁴ Besides this, there is also no definite consensus on the value of invasive electrophysiological studies (EPS) concerning their positive predictive value. To date, EPS showed controversial results concerning inducibility of VT/VF in BrS patients, and apart from differences in the study protocol applied, some studies propose non-inducibility associated with a negative predictive value for cardiac events.^3,6

The present study investigated the relation of ECG markers for right ventricular repolarization including TpTe and the TpTe/QT ratio and the occurrence of symptoms in patients with BrS.

Methods

Patient population

In this single-centre, retrospective study, clinical and genetic data of 78 index patients with BrS being part of the database of the Institute for Genetics of Heart Diseases (>7000 entries for inherited arrhythmogenic disorders) were analysed. All patients met the diagnostic criteria for BrS (presence of a type 1 ECG) according to the second consensus conference and to a recent report.^13,14 Overall, patients were included if (i) a type 1 ECG was documented (either spontaneously or pharmacologically induced) and (ii) medical documentation was completed. Besides this, clinical data were evaluated including initial symptoms of the patient (asymptomatic, syncopal episodes, VT, VF/sudden cardiac arrest), results of the genetic testing, family history, type of Brugada ECG, and results of drug challenge test, if performed. However, none of the patients were treated with drugs that influence re- and depolarization, e.g. ‘Brugada drugs’.

Electrocardiogram measurements

In all patients, rest ECGs (in the absence of a drug challenge test and faraway of a cardiac event) were digitalized by scanning in a high-resolution format and were imported into a graphic programme (DatInf^® Measure, Germany) for accurate measurements. Duration of the PQ, QRS, QT, and TpTe as well as the heart rate (RR) intervals were measured in two consecutive beats in selected ECG leads (I, II, III, V1–V3, and V5). The endpoint of the QT segment was identified by Lepeschkin's method (‘teach the tangent, avoid the tip’); a tangential line is drawn from the peak of the T-wave so that the endpoint of the QT segment is obtained from the intersection with the isoelectric line. An example of the method of measurements in leads V1 (coved type), V3 (saddleback type), and V5 is shown in Figure 1. Further parameters were calculated including the heart rate, the TpTe/QT ratio (TpTe/QT), and the predicted QT value (QTp) as well as its 88% value (QTp88%). The TpTe interval was defined by the distance between the end of the T-wave identified (see above) and the intersection between the perpendicular of the highest point of the T-wave and the isoelectric line (see also Figure 1). In this study, a short QT interval was defined in relation to the study of Gussak et al. as QTp88%—two standard deviations of the mean QT interval calculated via the formula of the Framingham study.¹⁵ An ECG was determined as a type 1 Brugada ECG pattern if, according to the updated expert consensus statement, the typical ECG pattern was identified in one lead in standard position in V1–V3 or in the higher positioned leads V1 or V2.

Genetic studies

Genetic analysis was performed as previously described.¹⁶ In brief, each patient gave written informed consent before genetic analysis. The study was in accordance with the latest revised version (1998) of the Declaration of Helsinki¹⁷ and with recommendations of the local ethics committee. After isolation of genomic DNA, all 28 exons and adjacent intronic sequences of the cardiac sodium channel gene (SCN5A; BRGDA1) were analysed by direct sequencing. Obtained sequences were compared with the human wild-type sequence (nucleotide sequence: GenBank, NM_000335; protein sequence: Swiss-Prot, Q14524; Locus Genome Reference ID: LRG_289). If the QTc interval was in the lower range, sequence analysis was expanded to the calcium channel genes CACNA1C (NM_001129840.1; LRG_341) and CACNB2 (NM_201590.2; no LRG ID so far) (both genes were reported to overlapping ECG phenotypes) and to other candidate genes predisposing for BrS, respectively.

Statistical analysis

Statistical analyses were performed with IBM SPSS Statistics Version 20. Results are reported as mean ± standard deviation. The Mann–Whitney U test was used to compare interval scale characteristics, i.e. ECG parameters in symptomatic and asymptomatic BrS patients. The Kruskal–Wallis test was used to compare ECG parameters in patient groups of different symptom severity (syncope, VT, VF/sudden cardiac arrest). χ² test was used to compare nominal characteristics, i.e. gender (m/f) and genetic testing (SCN5A mutation +/−). P-values of <0.05 are considered significant. To examine prognostic value from TpTe and to determine cut-off values, analysis of receiver operating characteristic (ROC) curves were made according to standard procedures. The Youden index was used to derive a reasonable cut-off value.

Results

Clinical characteristics

Clinical and genetic data are summarized in Table 1 and Supplementary material online, Table S1. In this study, clinical and genetic data of 78 index patients with BrS (mean age was 45 ± 14 years; 57 males, 73.1%; 21 females, 26.9%) were analysed. Overall, most of the patients had no or mild symptoms; in detail, 27 patients did not show any symptoms (34.6%), whereas 27 suffered from syncopal episodes. Six patients had documented episodes of VT (7.7%) and 16 patients survived SCD due to VF (20.5%). In two patients, the presence of symptoms could not be determined. A family history of sudden death was found in 16 patients (20.5%) and a spontaneous type 1 ECG was documented in 23 patients or during ajmaline drug challenge in 55 patients.

Genetic studies

Sequence analysis of the SCN5A coding region was performed in all 78 patients, in 17 (21.8%) a heterozygous mutation has been identified, mostly non-synonymous mutations (52.9%), whereas nonsense and other mutations (deletions, insertions and splice site) accounted for 47.1%. However, patients with a SCN5A mutation did not differ in their clinical course (symptomatic vs. asymptomatic; P = 0.576).

Electrocardiogram parameters

Summarized data are presented in Table 2 and in Supplementary material online, Tables S1 and S2. Overall, in the rest ECG, a Brugada type 1 ECG was present in 23 patients (29.5%), in 8 patients (10.3%) a type 2 ECG, and in further 12 patients (15.4%) a type 3 ECG. In the remaining 35 patients (44.9%), no significant changes could be identified. However, mainly men showed pathological ECG changes in the resting ECG (n = 35 or 44.9% in men vs. n = 8 or 10.3% in female).

Therefore, TpTe ≥77 ms was associated with an odds ratio (OR) for life-threatening events (VT/VF) of 5.0 (95% CI 1.7–14.4; P = 0.003) and a TpTe/QT ratio of ≥0.205 with an OR of 5.8 (95% CI 1.9–17.4; P = 0.002), respectively. No significant changes in TpTe and TpTe/QT ratio could be seen in the other right precordial lead V2 or V3. Besides this, all observations were independent of the mutation state (SCN5A+ vs. SCN5A−) or the Brugada ECG pattern.

Of note, patients with a SCN5A mutation had a longer PQ interval than patients without mutations in SCN5A in all leads except V5 (lead I, P = 0.026; lead II, P = 0.026; lead III, P = 0.019; lead V1, P = 0.001; lead V2, P = 0.008; lead V3, P = 0.005; lead V5, P = 0.108), see also Supplementary material online, Table S2.

The mean QRS duration in lead V1 was 99 ± 17 ms; 53.8% of the subjects had a normal QRS duration (<100 ms), 32.1% showed a slightly extended QRS complex, and in 14.1% QRS duration was longer than 120 ms. There were no significant differences between male and female subjects or the ones with or without symptoms (P = 0.457), independently of the ECG lead, respectively, with or without a mutation in SCN5A (see also Supplementary material online, Table S3).

QT intervals did now show significant differences between male and female patients or patients with and without symptoms or regarding mutation status in SCN5A. The same holds for heart rate corrected QTc intervals. A prolongation of QT intervals in the right precordial leads (but not any other) has been identified in some of the patients. Besides this, 5 out of 78 patients had a shortened QTc interval. These 5 patients may have a phenotypical overlap of short QT syndrome and BrS, but no mutations were found in the genes that predispose for short QT syndrome.

Discussion

Transmural dispersion of repolarization is represented by the TpTe duration¹⁸ and has been proposed to be responsible for arrhythmogenic events in BrS, J-wave syndrome, and long and short QT syndrome. Although the detailed underling mechanism of the TpTe interval is still unclear, there are at least two models that try to explain the underlying mechanism:

1. Electrophysiological and pharmacological studies identified three predominant ventricular myocardium cell types (epicardial, endocardial, and mid-layer/M cells) with typical action potential configurations.¹² In relation to the ECG T-wave morphology, the end of the action potential of the M cells coincide with the end of the T-wave, whereas the action potential of the epicardial cells ends at the peak of the T-wave¹²; action potential of the endocardial cells are between those of the epicardial and M cells. This altogether may represent the TDR and the TpTe interval in the surface ECG.

2. The above-mentioned explanation, however, does not reflect regional, apex-base, and interventricular differences, which are at least as important for the formation of the T-wave in the beating heart.¹⁹ In an experimental and model study, it could be shown that activation-recovery intervals increased from apex to base, from the left ventricle to the right ventricle, and in the apical portion of the left ventricle from epicardium to endocardium and from the right side of the septum to the left side.¹⁹ The simulated T_peak corresponded to the earliest end of repolarization, whereas the T_end corresponded to the latest end of repolarization.¹⁹ This study concluded that the TpTe interval corresponds to the global dispersion of repolarization with distinct contributions of the apicobasal and transmural action potential duration gradients and apicobasal difference in activation times.¹⁹ Nevertheless, in several studies, TpTe has been proven as a non-invasive risk marker of arrhythmogenesis and life-threatening arrhythmias.¹¹

In a pilot study of 29 BrS patients by Castro Hevia et al. it has been first demonstrated that TpTe and TpTe dispersion significantly correlated with the occurrence of life-threatening arrhythmic events.⁷ Brugada syndrome patients with recurrent episodes had a significantly longer TpTe interval (104.4 ms) than less severe patients (87.4 ms) and a prolonged TpTe dispersion (35.6 vs. 23.2 ms).⁷ Interestingly, the TpTe interval was the maximum measured value from all precordial leads. In contrast, we could not replicate these findings in a larger available patient sample [TpTe values for severely affected patients: 87 vs. 71 ms (lead V1) or 97 vs. 86 ms (lead V2), 95 vs. 90 ms (lead V3), 86 vs. 80 ms (lead V5), see Supplementary material online, Table S3]. In addition, we could show similar effects in the relevant right precordial lead V1 (TpTe difference of 16.3 ms). Recently, Letsas et al. investigated 23 BrS patients who underwent programmed ventricular stimulation⁸; those patients with inducible VT/VF displayed a significantly increased TpTe interval in lead V2 (88.8 vs. 78.3 ms) and V6 (76.3 vs. 66.7 ms) as well as a significantly greater TpTe/QT ratio in lead V6 (0.214 vs. 0.180).⁸ Again, in the present study encompassing the largest patient set, we found a greater TpTe/QT ratio in lead V1 rather than in other (left precordial) leads (0.24 or 24%).

In order to find normal values for TpTe/QT ratio and to characterize TpTe variability with heart rate, Gupta et al. evaluated ECGs of 60 normal healthy individuals.²⁰ The TpTe interval and the TpTe/QT ratio, however, were measured in lead V6. This study demonstrated that TpTe interval decreased linearly with increase in heart rate, but TpTe/QT ratio remained relatively constant with a median and mean of 0.21. This finding is also supported by Bieganowska et al., who analysed 131 healthy children (mean age 9.1 years).²¹ They also showed that TpTe intervals may vary between ECG leads and upon gender (boys > girls), but should be measured in precordial leads.²¹ Furthermore, Gupta et al. studied the relationship of TpTe/QT ratio in 7 patients with BrS and also demonstrated a significantly higher TpTe/QT ratio of 0.32 ± 0.04 (P < 0.01) in the right precordial leads than in the control group.²⁰ The findings were explained by the significantly prolonged TpTe duration in the right precordial leads due to the formation of the typical coved type ST-segment elevation. Hence, the TpTe/QT ratio enlarges especially in the right precordial leads.²⁰

Apart from these studies, there are further reports that investigated TpTe and TpTe/QT ratio in ion channel and non-ion channel heart diseases like in long and short QT syndrome patients, patients with ST-segment elevation myocardial infarction, in cardiac resynchronization therapy patients, or patients with SCD.^{9–11,20,22,23} All studies have in common that TpTe interval and/or the TpTe/QT ratio are significantly prolonged in patients with adverse cardiac events (e.g. ventricular arrhythmias, in-hospital death, cardiac death). TpTe/QT ratio ranged from 0.24 in cardiac resynchronization therapy patients⁹ up to 0.28–0.38 in short QT syndrome patients^20,22 as well as 0.28 in acquired LQTS²³ and 0.29 in patients with ST-segment elevation myocardial infarction.¹⁰

For a better risk stratification in patients with BrS, we performed ROC curve analysis (Figure 3). As a key finding, BrS patients with a TpTe duration ≥77 ms (sensitivity 63.6%, specificity 74.1%; OR 5.0, 95% CI 1.7–14.4; P = 0.003) and a TpTe/QT ratio of ≥0.205 (sensitivity 72.7%, specificity 65.5%; OR 5.8, 95% CI 1.9–17.4; P = 0.002) have a higher risk for cardiac events (VT/VF).

In comparison to the study of Castro Hevia et al., the TpTe durations were far shorter in our population (104 vs. 87 ms).⁷ This might be due to the fact that Castro Hevia et al. used the maximum TpTe interval measured in all precordial leads, but especially in the BrS the most informative leads are the right precordial ones (lead V1–V3). This approach has been applied in another, very recent study by Mauri et al. (2015) that analysed only type 1 ECGs of 325 BrS patients; here, TpTe intervals and TpTe/QT ratio were significantly higher in the right precordial leads (mainly V1 and V2).²² However, the type 1 ECG is present only in a minority of patients at baseline and our population included all patients with a spontaneous and with a pharmacologically induced Brugada type 1 ECG. Besides this, the TpTe values of the study of Letsas et al. are comparable with ours (89 vs. 87 ms).⁸ Interestingly, Letsas et al. performed measurements in the precordial leads only in leads V2 and V6. Although the TpTe/QT ratio had similar values (0.21 vs. 0.24 in our study), they are not compatible due to the fact that Letsas et al. found significant changes in lead V6 and not in the right precordial leads as we did in lead V2. Finally, the TpTe/QT ratio of Gupta et al. is much higher than ours (0.32 vs. 0.24).²⁰ Both were measured in the right precordial leads. The different values might be explained by the different sizes of the populations (7 vs. 78 patients in our study). Furthermore, it is not clear whether the measurements in the study of Gupta et al. were performed only in patients with a type 1 BrS ECG and what definition of the T-wave peak was used. In the Figure 5 of Gupta et al.,²⁰ it seems like the elevated J-point was defined as the peak of the T-wave and this would be discordant to your understanding (Figure 1).

Study limitation

The study was performed in a retrospective manner and therefore does not contain follow-up information.

Conclusion

In baseline ECG, the TpTe interval as well as TpTe/QT ratio in lead V1 are potentially non-invasive risk markers for life-threatening arrhythmias in patients with BrS. Therefore, BrS patients with a TpTe duration of ≥77 ms or a TpTe/QT ratio of ≥0.205 might be on a higher risk for life-threatening arrhythmias. Further prospectively oriented studies are needed to confirm these findings.

Supplementary material

Supplementary material is available at Europace online.

Funding

This work was supported by grants of the Foundation Leducq, Paris, France, the German Research Foundation, Bonn, Germany (DFG; SFB 656-C1), and the Interdisciplinary Center for Clinical Research, Münster, Germany (IZKF; Schu01-012-11) to E.S.-B. as well as by grants of the Innovative Medicine Research, Münster, Germany (IMF; ST121119) to B.S. L.E. is supported by the Peter Osypka foundation, Rheinfelden-Herten, Germany, and holds the Peter Osypka professorship for experimental and clinical electrophysiology.

Conflict of interest: none declared.

Acknowledgements

We kindly thank the patients for their participation.

References

Author notes