There is insufficient knowledge concerning long-QT (LQT) 3 in the pediatric population to determine whether recommendations for more aggressive therapy in these patients are appropriate. An international multicenter review of 43 children with cardiac sodium channel (SCN5A) mutations and clinical manifestations of LQT syndrome without overlap of other SCN5A syndromes was undertaken to describe the clinical characteristics of LQT3 in children. Patients were aged 7.6 ± 5.9 years at presentation and were followed for 4.7 ± 3.9 years. There was significant intrasubject corrected QT interval (QTc) variability on serial electrocardiography. Forty-two percent presented with severe symptoms or arrhythmia and exhibited longer QTc intervals compared to asymptomatic patients. None of the 14 patients who underwent primary prevention implantable cardioverter-defibrillator (ICD) implantation received appropriate shocks in 41 patient-years of follow-up, while 2 of 6 patients who underwent secondary prevention ICD implantation received appropriate shocks in 30 patient-years of follow-up. Half of patients who underwent ICD implantation experienced inappropriate shocks or ICD-related complications. Mexiletine significantly shortened the QTc interval, and QTc shortening was greater in patients with longer pretreated QTc intervals. Two ICD patients with frequent appropriate ICD shocks showed immediate clinical improvement, with elimination of appropriate ICD shocks after mexiletine loading. In conclusion, severe symptoms are common in children with LQT3 and are associated with longer QTc intervals. ICD implantation is associated with significant morbidity. Mexiletine shortens the QTc interval, and it may be beneficial.
Although β blockers are highly effective therapy for long-QT (LQT) 1, their efficacy is limited in patients with LQT3. The higher likelihood that cardiac events experienced by patients with LQT3 are lethal and the reduced efficacy of medical therapy led to the class IIb recommendation regarding the use of implantable cardioverter-defibrillators (ICDs) as prophylaxis for sudden cardiac death in patients with LQT3. These recommendations have not gone unnoticed by the international medical community, resulting in ICD implantation in a larger proportion of patients with LQT3 compared to those with LQT1 and LQT2. Although this approach has come under intense scrutiny, the underrepresentation of patients with LQT3 in the published research has complicated the interpretation of clinical data. In some reports, only a limited number of unrelated LQT3 families have been studied. Hence, there are limited data available regarding the clinical outcomes of patients with pediatric LQT3. Within this context, we embarked on a multicenter study of the clinical characteristics of LQT3 in the pediatric population.
Methods
A multicenter retrospective chart review of patients with pediatric LQT3 was conducted according to the guidelines of the institutional review boards of 14 participating centers. Inclusion criteria included age <18 years at presentation, a clinical course consistent with LQT syndrome including a corrected QT (QTc) interval >450 ms, and an identified cardiac sodium channel (SCN5A) mutation. Genetic analysis was performed by established research and commercial laboratories using standard techniques. We excluded patients with family histories of Brugada syndrome, electrocardiographic evidence of Brugada pattern at any point during their courses, or family histories of sodium channel overlap syndrome. Significant symptoms were defined as syncope, seizures, and cardiac arrest, and significant rhythm abnormalities included Torsades de Pointes (TdP) and sustained ventricular tachycardia. The presence of sinus bradycardia and functional 2:1 second-degree atrioventricular (AV) block were noted but for our purposes were not considered severe arrhythmias. A single standardized data collection form was used by all centers to provided uniform data for comparison. The QTc interval was manually calculated using established methods by participating investigators at each institution. Therapy was prescribed at the discretion of the treating physician in a nonstandardized manner.
For analyses involving the entire cohort, quantitative variables are presented as mean ± SD, while categorical values are presented as percentages. For certain subpopulation analyses, quantitative variables are presented as medians and ranges. Continuous data were compared using Student’s t tests, whereas categorical comparisons were made using Fischer’s exact tests or chi-square tests as appropriate. A p value <0.05 was considered significant.
Results
The patient population included 43 patients with LQT3 (22 female) aged 7.6 ± 5.9 years at presentation from 31 families (81% Caucasian, 11% African American, and 8% Hispanic) followed for 4.7 ± 3.9 years. The initial 12-lead electrocardiograms at presentation demonstrated sinus rhythm in 29 (67%), functional 2:1 AV block in 6 (14%), sinus bradycardia in 5 (12%), nonsustained TdP in 2 (4%), and functional 2:1 AV block with nonsustained TdP in 1 (2%). The 7 patients with 2:1 AV block were significantly younger (2.6 ± 5.7 vs 8.6 ± 5.5 years, p = 0.01) than those without 2:1 AV block, with 4 presenting at birth or in utero. The mean QTc interval on the initial presenting electrocardiogram for the entire cohort was 498 ± 63 ms.
Most patients presented with symptoms or rhythm abnormalities (24 of 43 [56%]; Figure 1 ) , and severe symptoms or arrhythmias were noted at presentation in 18 of 43 (42%), including cardiac arrest, TdP, sustained ventricular tachycardia, syncope, and seizures. Six subjects presented with cardiac arrest ( Table 1 ) and 5 with TdP. The presenting symptom was syncope in 6 patients and seizures in 1. Three patients who presented with TdP in fetal life subsequently died or went on to heart transplantation ( Table 1 ). Of the 12 patients identified through family screening, 11 were asymptomatic at presentation. One of these patients had seizures believed to be arrhythmic in origin (described previously). The ages and gender of patients with severe symptoms or arrhythmias are compared to those of patients without in Table 2 . Family histories of LQT syndrome were present in 16 subjects (38%), and family histories of sudden, unexpected death in 10 (24%). On the basis of the exclusion criteria, no patient had a family history of Brugada syndrome or a known sodium channel overlap syndrome.
| Patient | Gender | Age at Presentation (Years) | Symptoms at Presentation | Maximum QTc interval (ms) | Age at CA | Age at Death/OHT | Age at ICD Implantation (Years) | Age at PM Implantation | BB | NaCB |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | M | 2.5 | CA | 470 | 2.5 yrs | Alive | 2.8 | NA | Post CA | Ø |
| 2 | F | 2.1 | CA | 630 | 2.1 yrs | Alive | 2.1 | NA | Post | Post CA |
| 3 | F | 5.8 | CA | 536 | 5.8 yrs | Alive | 6.1 | NA | Post | Post ICD |
| 4 | M | 15.6 | CA | 508 | 15.6 yrs | Alive | 15.6 | NA | Post | No |
| 5 | F | 0 (fetus) | TdP, 2:1 AVB | 690 | NA | 5 mo | NA | NA | + | In utero |
| 6 | F | 9.1 | CA | 546 | 9.1 yrs | Alive | 9.2 | NA | Post | + |
| 7 | F | 2.7 | CA | 675 | 2.7 yrs | Alive | NA | NA | Post | + |
| 8 | F | 0 (fetus) | TdP | 530 | 5 d | 19 yrs | NA | 2 weeks | At birth | Ø |
| 9 | M | 0 (fetus) | TdP | 630 | NA | 7 mos | NA | 1 mo | At birth | + |
| Presenting Symptoms/Arrhythmias | Cardiac Arrest or Death | |||||
|---|---|---|---|---|---|---|
| Severe (n = 18) | Mild or None (n = 25) | p Value | Yes (n = 8) | No (n = 35) | p Value | |
| Age at presentation (years) | 6.4 ± 6.2 | 8.4 ± 5.7 | NS | 4.7 ± 5.3 | 8.2 ± 5.9 | 0.13 |
| QTc interval (ms) | ||||||
| Presenting | 523 ± 82 | 479 ± 38 | 0.02 | 525 ± 104 | 491 ± 50 | NS |
| Minimum | 484 ± 72 | 457 ± 29 | 0.11 | 487 ± 95 | 463 ± 36 | NS |
| Maximum | 563 ± 72 | 508 ± 40 | 0.003 | 573 ± 81 | 520 ± 52 | 0.03 |
In our cohort, QTc intervals varied significantly throughout the patients’ clinical courses. In comparing serial electrocardiograms before the initiation of pacing or sodium channel blocker therapy, the maximum QTc interval measured on an individual electrocardiogram was significantly longer than the shortest documented QTc interval measured on other electrocardiograms for the same patient ( Figure 2 ) . The maximum QTc interval in patients presenting with severe symptoms or arrhythmias was significantly longer than in the group without severe symptoms or arrhythmias ( Table 2 ). In addition, the maximum QTc interval was longer in patients who experienced cardiac arrest or sudden death ( Table 2 ).
Beta blockers were used in 34 patients (79%) ( Table 3 ). Side effects attributed to β blockers were noted in 8 patients (24%), including fatigue (n = 3), sinus bradycardia (n = 3), 2:1 AV block (n = 1), and behavioral problems (n = 1). In addition to β-blocker therapy, devices were implanted in 16 patients (47%) treated with β blockers. Of those treated with β blockers, 5 patients (15%) underwent pacemaker placement, and in 11 (32%), ICDs were implanted. Despite β blockers, 3 patients (9%) had significant events, including appropriate ICD shock (n = 1), death (n = 1), and persistent seizures (n = 1). Beta blockers were ultimately discontinued in 12 patients (35%): 9 after sodium channel blockers were started, 2 after the diagnosis of LQT3 was genetically confirmed, and 1 because of the development of 2:1 AV block.
| β Blocker | Initial Drug (n) | Final Drug (n) |
|---|---|---|
| Propranolol | 13 | 8 |
| Nadolol | 10 | 13 |
| Atenolol | 9 | 6 |
| Metoprolol | 1 | 2 |
| Unspecified | 1 | 1 |
Pacemaker implantation was undertaken in 5 patients (12%) at a median age of 3 months (range 0 to 13). The indication for pacemaker implantation was 2:1 AV block in 3 patients (1 with associated TdP) and TdP and sinus bradycardia in 1 patient each. The initial programming was ventricular-inhibited pacing in 3, dual-chamber pacing in 1, and unknown in 1 patient. The pacemaker was upgraded to a dual-chamber from a single-chamber ventricular device in 1 patient. After a median follow-up period of 4.2 years (range 2.5 months to 19 years), 2 patients with pacemakers had fewer episodes of TdP, and 3 were unchanged. One patient with an initial clinical improvement after pacemaker implantation died suddenly after 19 years from a presumed ventricular arrhythmia.
ICDs were implanted in 20 patients (47%) in the LQT3 cohort, with a median age of 11.5 years (range 0 to 18.2). The indication for ICD implantation was primary prevention in 14 patients and secondary prevention in 6 (5 with cardiac arrest and 1 with TdP). Patients were followed for a median of 3 years (range 0 to 11.9). During 41 patient-years of follow-up, no patients treated for primary prevention have yet received an appropriate ICD discharge. This is in contrast to 33% of the ICD patients treated for secondary prevention who experienced appropriate shocks after 30 patient-years of follow-up. One patient received 2 shocks for TdP, the first episode occurring 2.9 years after implantation. No information is available regarding the duration after implantation of the appropriate shock for the second patient.
Inappropriate ICD shocks occurred in 2 patients (14.2%) treated for primary prevention and in 1 (16.7%) treated for secondary prevention ( Figure 3 ) . One subject had multiple inappropriate shocks due to T-wave oversensing and was effectively treated by placement of a new pace-sense lead. The second patient had an inappropriate shock related to a problematic Sprint Fidelis lead (Medtronic, Inc., Minneapolis, Minnesota) that was extracted and replaced. The third patient underwent ICD implantation after cardiac arrest and received >20 inappropriate shocks beginning within 1 month of implantation because of T-wave oversensing from lead fracture. Despite lead replacement, system dysfunction continued, and device therapies were eventually disabled. Other complications occurred in 4 primary prevention patients (28.6%) and in 3 secondary prevention patients (50%). These complications included lead dislocation, lead dislocation and fracture requiring revision, pericardial effusion requiring pericardiocentesis, vascular tear during lead extraction requiring operation, wound infection, wound dehiscence requiring reoperation, and venous thrombosis. Overall, 10 of 20 patients (50%) who underwent ICD implantation had adverse events.
Sodium channel blockers were initiated in 19 patients (44%) at a median age of 8.8 years (range 3 days to 18 years) for symptoms (n = 7) or primary prevention related to confirmed diagnosis of LQT3 (n = 12). After follow-up of 1.2 years (range 0 to 7.4), 16 patients were on either mexiletine (n = 15) or flecainide (n = 1). The median dose of mexiletine was 7 mg/kg/day (range 3.5 to 12.5). Sodium channel blockers were discontinued in 3 subjects, after cardiac sympathectomy (n = 1) and secondary to gastrointestinal side effects (n = 1); in 1 subject, the rationale for discontinuation was not clear. Eleven patients (58%) taking sodium channel blockers received other therapies, including β blockers (n = 8), pacemakers (n = 2), and/or ICDs (n = 7). No patient developed AV conduction disturbance or electrocardiographic changes consistent with Brugada syndrome while receiving sodium channel blockade. One patient died suddenly from presumed arrhythmia while taking mexiletine. This patient harbored a R1623Q mutation that has been reported to be effectively treated with sodium channel blockade.
We analyzed the effect of mexiletine on QTc interval duration in the 13 patients without ventricular pacing. Mexiletine administration significantly shortened the QTc interval for the drug-treated cohort (570 ± 75 vs 476 ± 43 ms before vs after mexiletine, p = 0.001). The degree of QTc interval shortening was dependent on the duration of the pretreated QTc interval (see Figure 4 ) . In addition to QTc interval shortening, 2 ICD patients with frequent appropriate ICD shocks showed immediate clinical improvement, with elimination of appropriate ICD shocks after loading with mexiletine. The mexiletine dose was higher in these 2 patients compared to the others receiving mexiletine (11.8 ± 1.1 vs 6.6 ± 1.7 mg/kg/day, p = 0.002).
One patient underwent left cardiac sympathetic denervation, which was unsuccessful in controlling ventricular arrhythmias. The patient subsequently underwent heart transplantation.
Two of the 43 patients (4.7%) died (mutations R1623Q and N406K), and 1 (2.3%) underwent heart transplantation (unspecified SCN5A mutation). The 2 deaths occurred in patients who presented in fetal life with TdP. The first subject was treated with propranolol and mexiletine but died at 5 months of age from presumed arrhythmia. The second patient survived a cardiac arrest 5 days after birth and underwent pacemaker implantation as an infant. He was started on propranolol, which was discontinued when LQT3 was confirmed. The patient was not receiving a sodium channel blocker. This patient died at 19 years of age. A third patient underwent heart transplantation because of refractory ventricular arrhythmias despite a β blocker, a sodium channel blocker, pacemaker placement, and left cardiac sympathetic denervation.
When death was added to the composite of severe symptoms and arrhythmias, those who died or had severe symptoms at some point in their courses had significantly longer maximum QTc intervals than those who did not ( Table 2 ). Patient characteristics and treatment modalities for the 8 patients (18.6%) who died or had cardiac arrest are detailed in Table 1 . Although more female patients died or had cardiac arrest compared to male patients, this did not reach statistical significance.
Twenty missense SCN5A mutations were identified in 27 families ( Table 4 , Figure 5 ) . Eight missense mutations are novel. Five of the 8 novel mutations localize to the linker between domains I and II. Three of the subjects with I-II linker mutations experienced syncope or cardiac arrest. One patient carried 2 mutations that affected the same amino acid position, A572V and A572S. Another subject was a compound heterozygote carrying A572F and a second mutation in the potassium channel gene KCNQ1 G269D, considered a pathogenic mutation. The remaining novel missense mutations localize to the S1 segment of domain III (A1221V) and the turret preceding the S6 segment of domain IV (R1739W). Four LQT3 subjects harbored unspecified SCN5A mutations. These 4 patients underwent genetic testing as part of a research effort, and the treating physicians were not provided with the specific mutations. All 4 subjects exhibited a phenotype consistent with LQT syndrome: 2 with prolonged QTc intervals and cardiac arrest at presentation and 2 with 2:1 AV block and QTc intervals >500 ms. This includes a patient with refractory ventricular arrhythmias who underwent heart transplantation.
