OUP user menu

Heart rate-dependent variability of cardiac events in type 2 congenital long-QT syndrome

Iori Nagaoka, Wataru Shimizu, Yuka Mizusawa, Tomoko Sakaguchi, Hideki Itoh, Seiko Ohno, Takeru Makiyama, Ken-ichiro Yamagata, Hisaki Makimoto, Yoshihiro Miyamoto, Shiro Kamakura, Minoru Horie
DOI: http://dx.doi.org/10.1093/europace/euq342 1623-1629 First published online: 29 September 2010

Abstract

Aims We aimed to examine the validity of heart rate (HR) at rest before β-blocker therapy as a risk factor influencing cardiac events (ventricular fibrillation, torsades de pointes, or syncope) in long QT type 2 (LQT2) patients.

Methods and results In 110 genetically confirmed LQT2 patients (45 probands), we examined the significance of variables [HR at rest, corrected QT (QTc), female gender, age of the first cardiac event, mutation site] as a risk factor for cardiac events. We also evaluated frequency of cardiac events in four groups classified by the combination of basal HR and QTc with cutoff values of 60 b.p.m. and 500 ms to estimate if these two electrocardiographic parameters in combination could be a good predictor of outcome (mean follow-up period: 50 ± 39 months). Logistic regression analysis revealed three predictors: HR <60 b.p.m., QTc ≥500 ms, and female gender. When the study population was divided into four groups using the cutoff values of 60 b.p.m. for HR and 500 ms for QTc, the cumulative event-free survival by the Kaplan–Meier method was significantly higher in the group with HR ≥60 b.p.m. and QTc <500 ms than in the group with HR <60 b.p.m. and QTc <500 ms or that with HR <60 b.p.m. and QTc ≥500 m (P< 0.05). Irrespective of QTc interval, LQT2 patients with basal HR <60 b.p.m. were at significantly higher risk.

Conclusion The basal HR of <60 b.p.m. is a notable risk factor for the prediction of life-threatening arrhythmias in LQT2 patients.

  • Long QT syndrome
  • Arrhythmia
  • Genetics
  • Heart rate
  • Torsades de pointes

Introduction

Long QT syndrome (LQTS) is a primary electrical disease characterized by an abnormality in myocardial repolarization that leads to the prolongation of QT interval, morphological changes in T waves, and torsades de pointes (TdP) type of ventricular tachyarrhythmias on surface electrocardiogram (ECG).1 Studies on genotype–phenotype correlation identified the clinical characteristics in each genetic subgroup, which made it possible to diagnose and introduce β-blocker therapy (BBT) appropriately in LQTS patients.24 In patients with LQTS type 1 (LQT1), β-blockers are quite effective, whereas they are less effective in suppressing arrhythmic events in LQT2 and 3.2

Previous studies have demonstrated the importance of evaluating patients by clinical symptoms, gender, causative mutations, the type or biophysical function of mutations, and corrected QT (QTc) interval to stratify the arrhythmic risk in LQTS.3,514 Heart rate (HR) has been recognized since the establishment of LQTS as a clinical entity, and a low HR for age was included in the diagnostic criteria.15 A recent study by Schwartz et al.16 demonstrated that a lower resting HR and a relatively low baroreflex sensitivity in KCNQ1 A341V carriers are protective factors, whereas HR at rest in other subtypes of LQTS has not been fully investigated. In clinical practice, we have noted that in some cases of LQT2 that TdP was triggered by HR of <60 b.p.m. and suppressed by pacing at 80 b.p.m., which made us evaluate the importance of HR in arrhythmic events of LQT2 patients. For these reasons, we aimed to analyse whether HR at rest before BBT could be a novel risk factor for cardiac events besides gender, genetic locus, and prolonged QT interval in LQT2. We also evaluated the relationship between HR at rest and arrhythmic events before and after BBT through the analysis of clinical data on patients with LQT2.

Methods

Study population

From September 1996 to July 2009, 587 probands with QT prolongation underwent genetic testing in three institutes in Japan, Shiga University of Medical Science, Kyoto University Graduate School of Medicine, and the National Cardiovascular Center. One hundred and fifty-two probands (26%) were genotyped as LQT2. We also screened mutations in KCNQ1, SCN5A, KCNE1–3, and KCNJ2 using the standard genetic tests1720 and excluded 20 probands with compound mutations and/or modifier single-nucleotide polymorphisms known to affect the QT interval (KCNH2 K897T and KCNE1 D85N).21,22 The remaining 132 probands were found to have a single KCNH2 mutation, and among them, we excluded from analyses patients under 15 years and those without detailed clinical information or with medication (except for β-blocker) which could influence baseline ECG measurements at the first medical contact and thereafter. Children <15 years old were not studied because they had relatively high basal HR. Family members of the 152 probands were recruited for the analysis if we could obtain necessary clinical information and if they were over 15 years old. As a result, the study population became 110 patients (45 probands and 65 family members) from 74 unrelated Japanese LQT2 families.

Both symptomatic and asymptomatic patients were included in the groups of probands and family members. Regardless of being probands or family members, patients were defined as symptomatic when they had a history of cardiac events (defined as ventricular fibrillation, TdP, or syncope due to ventricular arrhythmia) at the first medical contact or at the time of yearly follow-up. Patients with an apparent history of vasovagal syncope were not included in the study. The protocol for genetic analysis complied with the Declaration of Helsinki and was approved by the institutional ethics committees and performed under their guidelines. All individuals or their guardians gave written informed consent to genetic and clinical data analyses. Follow-up data were obtained from patients’ regular hospitals working with the authors in case patients lived far from our institutions or hospitals and were not able to visit us.

Genetic analysis and characterization

Genomic DNA was isolated from venous blood lymphocytes using the QIAamp DNA blood midikit (Qiagen, Hilden, Germany). Established primer settings were used to amplify the entire coding regions of known LQTS genes from genomic DNA.1720 Denaturing high performance liquid chromatography (DHPLC) or direct sequencing techniques were employed as described elsewhere.11 Polymerase chain reaction fragments presenting abnormal signals in DHPLC analysis were subsequently sequenced by the dideoxynucleotide chain termination method with fluorescent dideoxynucleotides on an ABI 3113xl genetic analyzer (PE Applied Biosystems).

The pore region of the KCNH2 channel was defined as the area extending from S5 to the mid-portion of S6 involving amino acid residues from 550 through 650 according to the previous report.10 The non-pore region included the N-terminus region, transmembrane domains apart from the pore region and the C-terminus region.

Clinical characterization

Routine demographic data and basal 12-lead ECGs were obtained from all subjects at the first medical contact as well as at yearly follow-up. In 104 patients, ECG parameters were measured before BBT was introduced. The remaining six patients, in whom BBT was started after the first cardiac event by an attending physician in other hospitals, visited a university hospital for further diagnostic confirmation of the symptoms. One of the six patients experienced aborted sudden cardiac death, four had documented TdP, and one had a syncopal attack. After obtaining informed consent, BBT was discontinued for more than five times the half life and examinations were performed, including a blood test, basal ECG, chest X ray, echocardiogram, and treadmill test for the diagnosis of congenital long QT syndrome.

Electrocardiograph parameters measured in the study were HR and QT interval. Rate-dependent QT intervals were corrected for HR using Bazett's method. QT interval was manually measured in lead V5 using the tangent method4 with an average of 2 or 3 consecutive beats by three investigators who were completely unaware of the patients’ clinical and genetic state. There were no significant differences in the measured data among three investigators. Bifid T waves, but not U waves, were included in the QT measurements. In the presence of bifid T waves, the end of the second T wave was defined as the end of the QT interval. If ECG recordings were obtained during a cardiac event, such as the appearance of frequent ventricular tachycardia, TdP, or cardiac arrest, they were requested to perform another examination after patient's general status had improved.

Data on patients who received BBT after the initial check-up were evaluated, including the dose of each drug, HR under medication, and recurrent arrhythmic episodes. Other treatments, such as implantable cardioverter-defibrillator (ICD) implantation and surgical left cardiac sympathetic denervation, were also evaluated. Follow-up data, including the occurrence of cardiac events and therapeutic changes, were collected retrospectively.

Statistical analysis

Student's t-test was employed to compare continuous data. Differences in frequencies were analysed by the χ2 test or Fisher's exact test. Analysis of variance was used to test differences of variables among more than three groups. Stepwise regression analysis was performed to determine predictors of cardiac events. Variables with P< 0.05 on univariate analysis were included in a logistic regression model with cardiac events as dependent variables. To determine the connection of the selected clinical variables with the occurrence of cardiac events, odds ratios for unadjusted data and their 95% confidence intervals were calculated. The cumulative probability of the first cardiac event between 15 and 50 years old was estimated using the Kaplan–Meier method. The Cox proportional-hazards survivorship model was used to investigate whether there were any prognostic factors that could influence the occurrence of cardiac events. Data are reported as the mean ± SD. Two-sided probability values <0.05 were considered significant. Statistical calculations were performed using SPSS software (version 18.0J).

Results

Clinical and genetic characteristics

The study population consisted of 110 consecutive patients from 74 unrelated Japanese LQT2 families (Table 1). The baseline ECG showed that the mean HR of probands tended to be lower than that of family members (P=0.06).

View this table:
Table 1

Basal characteristics of the study population

All (n=110)Proband (n=45)Family member (n=65)
Clinical characteristics
 Age (years)40.8 ± 17.5 (15–87)31.2 ± 15.6 (15–77)47.4 ± 15.6 (16–87)**
 Sex (male/female)40/7010/3530/35*
Symptomatic patients [n (%)]48 (44)38 (84)10 (15)**
 Cardiac arrest (n)743
 Syncope (n)46388
 Both (n)541
ECG
 HR (b.p.m.)62 ± 1060 ± 963 ± 11
 QTc (ms)483 ± 58508 ± 60467 ± 50**
  • *P < 0.05 vs. proband.

  • **P< 0.001 vs. proband.

All patients were genotyped to be a heterozygous carrier of 70 different KCNH2 mutations (18 in the N-terminus, 15 in non-pore regions, 13 in pore regions, and 24 in the C-terminus). Forty-three mutations were missense mutations, 15 were deletion/insertions, 9 were frameshifts, and 3 were nonsense mutations.

Factors determining cardiac events in LQT2 patients

We first evaluated whether HR and other variables (age at onset of cardiac events, female gender, site of mutation, missense mutation, and QTc) served as risk factors for cardiac events in LQT2 patients. Univariate analysis (Table 2) showed that HR of <60 b.p.m. per se was a significant risk for cardiac events (P< 0.01). In addition, female gender, HR as a continuous variable, a QTc interval of ≥500 ms, and pore site mutation were associated with an increased risk for cardiac events (P< 0.05). Other variables such as age at onset of cardiac events, sites of mutation (non-pore transmembrane, N-terminal, and C-terminal), and missense mutation were not statistically significant.

View this table:
Table 2

Predictors of cardiac events (syncope, aborted cardiac arrest, or sudden cardiac death) in univariate and multivariate analyses

Univariate analysisMultivariate analysis
Odds ratio (95% CI)P-valueOdds ratio (95% CI)P-value
Age at onset1.08 (0.78–1.49)0.639
Female gender3.56 (1.51–8.38)0.0044.54 (1.72–12.00)0.002
HR <60 b.p.m.2.83 (1.30–6.16)0.0094.46 (1.77–11.24)0.001
HR (continuous variable)0.95 (0.91–0.99)0.022
QTc ≥500 ms2.65 (1.18–6.00)0.0192.93 (1.13–7.59)0.026
Mutation location
Pore2.45 (1.07–5.60)0.0341.77 (0.70–4.48)0.230
Transmembrane, non-pore0.91 (0.27–3.08)0.914
N-terminal0.83 (0.33–2.04)0.677
C-terminal0.57 (0.26–1.27)0.169
Missense mutation2.10 (0.91–4.85)0.081

Multivariate analysis (Table 2) was subsequently performed using female gender, HR of <60 b.p.m., QTc of ≥500 ms, and pore site mutation. As for HR, we chose HR of <60 b.p.m. for multivariate analysis because we aimed to clarify if low HR of <60 b.p.m. was a significant risk factor for cardiac events. As shown in Table 2, female gender, HR <60 b.p.m., and QTc ≥500 ms were revealed to be significant risk factors for cardiac events (P < 0.05).

Bradycardia as an arrhythmic risk factor in LQT2 patients

We employed two ECG parameters, HR and QTc, to scrutinize who were more prone to have cardiac events in our LQT2 cohort. Using cutoff values of 60 b.p.m. for HR without β-blockers and 500 ms for QTc, we classified 110 LQT2 patients into four groups (Figure 1). Closed and open circles in the figure indicate symptomatic and asymptomatic patients, respectively (including both probands and family members). There were only eight symptomatic patients (23%) in the quadrant area of HR ≥60 b.p.m. and QTc <500 ms. In contrast, in the quadrant area defined as HR <60 b.p.m. and QTc ≥500 ms, 12 subjects (86%) experienced cardiac events (P < 0.05, vs. HR ≥60 b.p.m. and QTc <500 ms).

Figure 1

Distribution of KCNH2 mutation carriers according to the resting HR and QTc duration. Closed and open circles indicate symptomatic and asymptomatic patients, respectively. Two solid lines in the graph are drawn using the cutoff values of 60 b.p.m. and 500 ms. QTc was measured in lead V5. Groups A-D in the graph correspond to those in the text, Table 3 and Figure 2.

Table 3 summarizes the baseline characteristics of four groups divided by HR and QTc. The group of HR ≥60 b.p.m. and QTc <500 ms was defined as Group A, the group of HR <60 b.p.m. and QTc <500 ms as Group B, HR ≥60 b.p.m. and QTc ≥500 ms as Group C, and HR <60 b.p.m. and QTc ≥500 ms as Group D. There were no significant differences among four groups regarding age at baseline ECG recording, age at the first event, percentages of female gender, and BBT. In Group A, the number of proband was significantly lower than that in Groups B and D. The incidence of syncope or aborted cardiac arrest in Group A was significantly lower than in the Groups B and C. In groups of HR <60 b.p.m. (B and D), patients with QTc ≥500 ms (Group D) had more arrhythmic events than those with QTc <500 ms (Group B).

View this table:
Table 3

Baseline clinical characteristics of four subgroups defined by QTc and basal HR

QTc <500 msQTc ≥500 ms
Group A: HR ≥60 b.p.m. (n = 35)Group B: HR <60 b.p.m. (n = 39)Group C: HR ≥60 b.p.m. (n = 22)Group D: HR <60 b.p.m. (n = 14)
Age (years) at ECG (range)43 ± 18 (16–87)39 ± 17 (15–71)42 ± 18 (16–77)39 ± 17 (15–64)
Age (years) at first event (range, number of patients)25 ± 10 (13–42, n = 8)27 ± 15 (15–71, n = 19)26 ± 19 (15–77, n = 10)26 ± 15 (13–54, n = 10)
Female gender [n (%)]23 (66)22 (55)16 (73)9 (64)
Proband [n (%)]8 (23)*18 (46)12 (55)7 (50)
Pore site mutation [n (%)]6 (17)**11 (28)10 (46)7 (50)
Schwarz score3.1 ± 2.0§3.6 ± 1.7§5.5 ± 1.76.2 ± 1.2
Syncope or aborted cardiac arrest [n (%)]8 (23)19 (49)10 (46)11 (79)
β-Blockers [n (%)]7 (20)13 (33)9(41)6 (43)
  • Values are given as the mean ± SD where indicated. HR = heart rate.

  • *P<0.05 vs. Groups B and C.

  • **P<0.05 vs. QTc ≥500 ms (Groups C and D).

  • §P<0.001 vs. QTc ≥500 ms (Groups C and D).

  • P<0.05 vs. Group D.

  • P<0.05 vs. HR <60 b.p.m. (Groups B and D).

We then estimated the cumulative probability of the first cardiac event between the age of 15 and 50 in four groups (Groups A–D, Figure 2). The Kaplan–Meier analysis of all subjects (Figure 2A) showed that cumulative event-free survival was significantly different (P=0.007 by the log-rank test) and when adjusted for multiple comparisons, cumulative event-free survival was higher in Group A than in groups of HR <60 b.p.m. (P=0.014 vs. Group B, P=0.001 vs. Group D). In contrast, the survival rate was not statistically different among Groups B–D.

Figure 2

The Kaplan–Meier cumulative cardiac event-free survival curves from the age of 15–50 among each of four subgroups defined by cutoff values for HR of 60 b.p.m. and QTc of 500 ms. Panels AC show the Kaplan–Meier curves of 110 patients, 45 probands and 65 family members, respectively. The group of HR ≥60 b.p.m. and QTc <500 ms was defined as Group A, the group of HR <60 b.p.m. and QTc <500 ms as Group B, HR ≥60 b.p.m. and QTc ≥500 ms as Group C, and HR <60 b.p.m. and QTc ≥500 ms as Group D.

In Figure 2B and C, we examined the clinical course of 45 probands and 65 family members separately. The Kaplan–Meier analysis revealed no statistical difference in probands (Figure 2B, P=0.206 by the log-rank test), whereas in family members, cumulative event-free survival was significantly different among the subgroups (Figure 2C, P=0.017 by the log-rank test, P=0.058 for Group A vs. Group B, P=0.002 for Group A vs. Group D in multiple comparisons). Thus, the statistical difference in overall subjects may result from the prognosis of family members in our study population.

Finally, in order to assess the significance and independence of HR and QTc for cardiac events, we evaluated the parameters with the Cox proportional-hazards survival model (Table 4). The values of HR and QTc were centred at 60 b.p.m. and 500 ms for ease of interpretation. Compared with patients in Group A, patients in groups of HR <60 b.p.m. (Groups B and D) showed a higher risk for cardiac events by 2.6–4.4-fold. Although the hazard ratio in Group C was 2.16, there was no statistical difference between Groups A and C.

View this table:
Table 4

Contribution of QTc duration and HR to COX survival model

Number of patientsHazard ratio95% CIP-value
QTc <500 ms
 HR ≥60 b.p.m. (Group A)351
 HR <60 b.p.m. (Group B)392.601.14–5.970.023
QTc ≥500 ms
 HR ≥60 b.p.m. (Group C)222.160.85–5.470.105
 HR <60 b.p.m. (Group D)144.391.76–10.920.001

Treatment

β-Blocker therapy was introduced in 35 patients (29 probands) after diagnosis of LQT2 was made. Mean HR on medication was 56 ± 8 b.p.m. Metoprolol was used in 3 patients (90 ± 52 mg, 30–120), carvedilol in 3 (15 ± 9 mg, 5–20), atenolol in 4 (50 ± 0 mg, 50), propranolol in 21 (42 ± 16 mg, 30–80), and bisoprolol in 4 (4 ± 1 mg, 2.5–5).

Implantable cardioverter-defibrillator was implanted in 12 patients (VF: five patients, syncope: seven patients) during the first hospitalization or follow-up. In seven patients with a history of cardiac arrest due to VF (Table 1), two patients were treated with an ICD, three with both ICD and β-blocker, one with a pacemaker, and one with β-blocker alone (because the patient rejected ICD implantation). In a patient with a pacemaker, TdP was observed repeatedly whenever she fell asleep and sinus rhythm became <60 b.p.m. during her first admission to the hospital. After pacemaker implantation, atrial pacing at 80 b.p.m. completely suppressed TdP. None of the patients received surgical left cardiac sympathetic denervation in our study population.

Recurrence of arrhythmic events during follow-up

For the follow-up data, 36 patients (22 patients on BBT) followed more than 3 months were recruited and 86% of patients (31 patients: 18 patients on BBT, 8 patients with an ICD, 10 patients without treatment) completed the mean follow-up period of 50 ± 39 months (40 ± 35 months for 18 patients on BBT and 63 ± 42 months for 13 patients without BBT).

Eighteen subjects on BBT consisted of 14 symptomatic (due to syncope) and 4 asymptomatic patients. Arrhythmic events during follow-up were observed only in symptomatic patients (seven patients: VF was observed in one patient, syncope in six patients). Analysis of the relationship between HR of <60 b.p.m. and recurrent events was also performed. Cardiac events during follow-up were observed in three of nine patients who showed HR <60 b.p.m. before BBT and four of eight patients with HR <60 b.p.m. after BBT (P = 0.60 and 0.06, respectively). Therefore, low HR of <60 b.p.m. at rest before or after β-blockers did not predispose ventricular arrhythmia, although the statistical insignificance could be due to a small number of patients for analysis. Details of treatment after recurrence in each individual were described below.

A 16-year-old male patient with a history of syncope experienced VF and was resuscitated. He underwent ICD implantation and dosage of bisoprolol was increased from 2.5 to 5 mg/day, which prevented any cardiac events for a follow-up period of 34.5 months. Recurrent syncope or documented TdP on BBT were observed in six patients: two patients who took metoprolol (one did not comply with the drug regimen and one with a syncopal episode on medication), one patient with atenolol (syncope twice and electrical storm due to TdP twice on medication), and three patients with propranolol (one did not comply with the drug regimen, two experienced a recurrent syncopal episode on medication). In those who did not comply with medication, syncope or TdP was suppressed by resuming BBT. Recurrent episodes of syncope in one patient on metoprolol (120 mg/day) have been suppressed by changing BBT to bisoprolol (2.5 mg/day) for 20 months. Implantable cardioverter-defibrillator implantation was also performed in this patient. Episodes of one patient on atenolol (50 mg/day) were not suppressed with additional prescription of mexiletine (400 mg/day), and ICD was implanted. He experienced an electrical storm after ICD implantation. While adjusting BBT, he was diagnosed with oesophageal cancer and died after 19.8 months follow-up. Syncope in one patient on propranolol (60 mg/day) was suppressed with combined medication of propranolol and diazepam. The other patient on propranolol (30 mg/day) was implanted with an ICD after recurrent episodes of syncope. Atrial pacing of 84 b.p.m. prevented arrhythmic events.

In 13 patients without BBT, 5 were symptomatic (1 VF and 4 syncope) at the first medical contact. In these patients, only one patient with a history of VF experienced an appropriate ICD shock following recurrent VF. To note, pacing using ICD leads was introduced during the first hospitalization in three of five symptomatic patients in whom TdP was repeatedly observed under HR of 60 b.p.m. In these patients, pacing prevented recurrent cardiac events during follow-up.

Discussion

The present study demonstrated that basal HR of <60 b.p.m. was an apparent risk factor for cardiac events in LQT2 patients. Corrected QT ≥500 ms and female gender were also useful for risk stratification in LQT2. The Kaplan–Meier analysis in total study population revealed that cumulative event-free survival was significantly higher in the subgroup with HR ≥60 b.p.m. and QTc <500 ms than in the two groups with HR <60 b.p.m. (P< 0.05). The same trend was observed in the analysis of family members. On the other hand, there was no significant difference in basal HR irrespective of cardiac events in probands. Because, first, the number of probands (n = 45) was relatively smaller than that of family members (n = 65), and second, there was an entry bias: 84% of probands were referred for genetic testing as they were symptomatic, which influenced the evaluation of basal HR and cardiac events. Our examination of family members therefore suggested that KCNH2 mutation carriers associated with more severe bradycardia may show a stronger penetrance.

Mutations in KCNH2 are causative of LQT2, and KCNH2 encodes for the rapid component of the delayed rectifier K-current (IKr). In electrophysiological studies, IKr was shown to be present in rabbit23 and mouse24 sinoatrial node cells. Pharmacological inhibition of IKr by E-4031 markedly suppressed the spontaneous activity of sinoatrial node cells, suggesting that IKr activation plays a key role in maintaining an adequate HR. In other experimental models,25 IKr blockade has also been shown to cause bradycardia. In clinical studies, bradycardia is more frequently observed in LQT2.3,26 However, no previous studies have demonstrated the validity of bradycardia as a predictor of prognosis.

As for pore site mutations of KCNH2, known as a risk factor for cardiac events in LQT2, they were correlated with cardiac events in univariate but not multivariate analysis in our study cohort (Table 2). This contrasts with the previous report of Moss et al.10 and is probably due to the difference in the number of studied mutations as well as the exclusion of patients who had their first cardiac events before 15 years old.

β-Blockers are first line therapy for prevention of TdP in LQT2 because it suppresses early afterdepolarizations carried by L-type Ca channels or Ca2+ channels.2729 The result of our study, however, may cause concerns that BBT-induced HR-reduction could lead to recurrence of ventricular arrhythmias. To answer the question, we analysed the patient group on BBT during follow-up, but low HR of <60 b.p.m. at rest before or after β-blockers did not predict recurrence of cardiac events (P= 0.60 and 0.06, respectively). Our study cohort may be too small to clarify this issue and therefore, further clinical evaluation with a large number of patients will be required to conclude the significance of low HR on/off β-blockers in LQT2. On the basis of our findings, however, it is reasonable to hypothesize that pacing could be used as an adjunctive therapy in LQT2 patients showing HR <60 b.p.m. irrespective of QTc values. Our combined risk-evaluating scales (Figure 1) would help physicians estimate long-term therapy in asymptomatic KCNH2 mutation carriers, both probands and family members.

Limitations

In some symptomatic patients, there was a long period between the average age at onset of symptoms and the average age at ECG recording. Regarding this issue, the risk evaluation should be carefully considered. In addition, it was difficult to gather ECG recordings of the first event, because many patients suffered syncope without a doctor witnessing the first event. However, among the four subgroups, there was no significant difference in age at ECG recording and age at the first event (Table 3). Therefore, we evaluated cardiac risk using the HR recorded by ECG at the first medical contact. As for the effect of BBT on HR as a risk factor for cardiac events, our cohort was too small to lead a relevant conclusion because follow-up of patients was insufficient. Hence, it awaits a further study with a larger number of genotyped LQT2 patients.

Funding

This work was supported by the Uehara Memorial Foundation, the Ministry of Education, Culture, Sports, Science (Technology Leading Project for Biosimulation) and the Ministry of Health, Labour and Welfare, Japan (Research Grant for the Cardiovascular Diseases, H18-Research on Human Genome, 21C-8, 22-4-7).

Acknowledgements

The authors would like to thank the Japanese LQT2 families for their willingness to participate in this study and Ms Arisa Ikeda for her excellent technical support.

Conflict of interest: none declared.

References

View Abstract