High prevalence of concealed Brugada syndrome in patients with atrioventricular nodal reentrant tachycardia Can Hasdemir, MD,* Serdar Payzin, MD,* Umut Kocabas, MD,* Hatice Sahin, RN,* Nihal Yildirim, RN,* Alpay Alp,† Mehmet Aydin, MD,‡ Ryan Pfeiffer, BSc,§ Elena Burashnikov, BSc,§ Yuesheng Wu, MS,§ Charles Antzelevitch, PhD, FHRS§ From the *Department of Cardiology, Ege University School of Medicine, Izmir, Turkey, †St. Jude Medical, Izmir, Turkey, ‡Tepecik Teaching and Research Hospital, Izmir, Turkey, and §Masonic Medical Research Laboratory, Utica, New York. BACKGROUND Atrioventricular nodal reentrant (AVNRT) may coexist with Brugada syndrome (BrS).

tachycardia

OBJECTIVES The present study was designed to determine the prevalence of drug-induced type 1 Brugada ECG pattern (concealed BrS) in patients presenting with clinical spontaneous AVNRT and to investigate their electrocardiographic, electrophysiological, and genetic characteristics. METHODS Ninety-six consecutive patients without any sign of BrS on baseline electrocardiogram undergoing electrophysiological study and ablation for symptomatic, drug-resistant AVNRT and 66 control subjects underwent an ajmaline challenge to unmask BrS. Genetic screening was performed in 17 patients displaying both AVNRT and BrS. RESULTS A concealed BrS electrocardiogram was uncovered in 26 of 96 patients with AVNRT (27.1%) and in 3 of 66 control subjects (4.5%) (P r .001). Patients with concealed BrS were predominantly female patients (n ¼ 23 [88.5%] vs n ¼ 44 [62.9%], P ¼ .015), had higher prevalence of chest pain (n ¼ 10 [38.5%] vs n ¼ 13 [18.6%], p ¼ 0.042), migraine headaches (n ¼ 10 [38.5%] vs n ¼ 10 [14.2%], p ¼ 0.008), and drug-induced initiation and/or worsening of duration and/or frequency of AVNRT (n ¼ 4 [15.4%] vs n ¼ 1 [1.4%], p ¼ 0.006) as compared to

Introduction Brugada syndrome (BrS) is a form of inherited arrhythmia syndrome characterized by a distinct ST-segment elevation in the right precordial leads in the absence of structural heart disease.1 BrS can be manifest, suspicious, or concealed on the basis of clinical and electrocardiographic (ECG) characteristics. BrS has been shown to be associated with atrial arrhythmias such as atrial fibrillation, atrial flutter, and paroxysmal supraventricular tachycardia (PSVT).2–5 In the presence of This study was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health (grant no. HL47678, to Dr Antzelevitch) and the Masons of New York, Florida, Massachusetts, Connecticut, Maryland, Delaware, Rhode Island, New Hampshire, and Wisconsin. Address reprint requests and correspondence: Dr Can Hasdemir, Department of Cardiology, Ege University School of Medicine, Bornova, Izmir 35100, Turkey. E-mail address: [email protected].

1547-5271/$-see front matter B 2015 Heart Rhythm Society. All rights reserved.

patients with AVNRT without BrS. Genetic screening identified 19 mutations or rare variants in 13 genes in 13 of 17 patients with both AVNRT and BrS (yield ¼ 76.5%). Ten of these 13 genotype-positive patients (76.9%) harbored genetic variants known or suspected to cause a loss of function of cardiac sodium channel current (SCN5A, SCN10A, SCN1B, GPD1L, PKP2, and HEY2). CONCLUSION Our results suggest that spontaneous AVNRT and concealed BrS co-occur, particularly in female patients, and that genetic variants that reduce sodium channel current may provide a mechanistic link between AVNRT and BrS and predispose to expression of both phenotypes. KEYWORDS Brugada syndrome; Atrioventricular nodal reentrant tachycardia; Supraventricular tachycardia; Genetics ABBREVIATIONS AV ¼ atrioventricular; AVNRT ¼ atrioventricular nodal reentrant tachycardia; BrS ¼ Brugada syndrome; ECG ¼ electrocardiogram/electrocardiographic; INa ¼ sodium channel current; PSVT ¼ paroxysmal supraventricular tachycardia (Heart Rhythm 2015;0:1–11) rights reserved.

I

2015 Heart Rhythm Society. All

increased prevalence of spontaneous and/or inducible atrioventricular nodal reentrant tachycardia (AVNRT) in patients with manifest BrS compared with the general population (7% vs 0.135%), we hypothesized that spontaneous AVNRT may co-exist with a concealed or manifest BrS phenotype.4–6 The aim of this study was first to determine the prevalence of concealed BrS in patients presenting with clinical spontaneous AVNRT and second to investigate the clinical, ECG, electrophysiological, and genetic characteristics of these patients.

Methods Study populations One hundred three consecutive patients without any signs of the Brugada pattern on baseline standard 12-lead ECGs who underwent electrophysiological study and catheter ablation http://dx.doi.org/10.1016/j.hrthm.2015.03.015

2 for symptomatic, drug-resistant AVNRT between July 2011 and November 2013 were retrospectively included. Seven patients were excluded owing to ischemic cardiomyopathy (n = 3), coexisting concealed accessory pathway (n = 2), secundum-type atrial septal defect (n = 1), and rheumatic heart disease (n = 1). The remaining 96 patients (67 women and 29 men; mean age 46 ⫾ 15 years; range 18–72 years) formed the patient population, and their characteristics were compared with those of the control group (n = 66 [42 women and 24 men]; mean age 42.7 ⫾ 9.5 years; range 25–63 years). All patients and control subjects underwent the ajmaline challenge test. Genetic screening and analysis was performed in a total of 17 patients: 10 with concealed BrS and clinical AVNRT from our present study and 7 with previous diagnosis of BrS and clinical AVNRT from our observational BrS cohort consisting of 71 patients referred to our tertiary referral hospital between January 2004 and December 2013. The study protocol was approved by the ethics committee of the Ege University School of Medicine. Written informed consent was obtained from all patients for ajmaline challenge test and genetic testing.

Data acquisition A detailed medical history including age of onset of AVNRT, associated symptoms (palpitation, chest pain, syncope, and cardiac arrest), presence of systemic diseases (systemic hypertension, diabetes mellitus, and migraine headaches), and initiation of AVNRT (in the presence of high sympathetic or parasympathetic tone or both) was obtained in all patients. Syncope was classified as reflex (vasovagal or situational), orthostatic hypotension, arrhythmia-related cardiac syncope, and unexplained syncope. The diagnosis of migraine headache was based on the criteria of the International Classification of Headache Disorders, 2nd edition.7 Presence of additional atrial (focal atrial tachycardia and atrial fibrillation) and ventricular (frequent [410 per hour], monomorphic premature ventricular contractions, and/or ventricular tachycardia) arrhythmias, family history of AVNRT (documented AVNRT in Z2 family members), and sudden unexpected death (r50 years of age) in the firstdegree relatives were noted in all patients with AVNRT. Brugada pattern–inducing drug exposure and its effect on AVNRT frequency and duration were noted in all patients. All study subjects underwent transthoracic echocardiography for the evaluation of valves and right and left ventricular size and function.

Definition of ECG parameters All subjects had a baseline 12-lead ECG with leads in the standard lead position and a high precordial lead ECG, with leads V1 and V2 moved up to the third and the second intercostal space.3 Type 1 Brugada pattern, QRS fragmentation, and early repolarization patterns were defined according to previously described criteria.8–10

Heart Rhythm, Vol 0, No 0, Month 2015 Every patient had a 12-lead ECG during their index clinical arrhythmia. Twelve-lead ECGs were analyzed for the rate of clinical AVNRT, presence of pseudo-r0 deflection in lead V1, pseudo-S wave in inferior leads, P-in-QRS pattern (absence of pseudo-r0 deflection in lead V1 and/or pseudo-S wave in inferior leads), and QRS alternans.11 Five patients were excluded for the analysis: 3 with slow/slow AVNRT, 1 with fast/slow AVNRT, and 1 with left variant slow/fast AVNRT.

Electrophysiological study All patients with AVNRT underwent electrophysiological study and catheter ablation. The diagnosis of AVNRT was made on the basis of previously described criteria.12 Baseline AH and HV intervals, type and rate of AVNRT, time intervals (His-V, V-AHis, V-Aright atrial appendage, and VAdistal coronary sinus) during AVNRT were determined in each patient.

Definition of BrS Type 1 Brugada pattern in at least 1 right precordial lead was considered to be diagnostic of BrS.8 Manifest BrS was defined as the presence of diagnostic type 1 Brugada pattern on baseline 12-lead ECGs before the drug challenge test in the presence of a presenting symptom. Suspicious BrS was defined as the presence of type 2 or 3 Brugada pattern on baseline 12lead ECGs before the drug challenge test in the presence of a presenting symptom. Concealed BrS was defined as the absence of any signs of Brugada pattern on baseline 12-lead ECGs before the drug challenge test and the development of type 1 Brugada pattern after the drug challenge test regardless of symptomatic status, consistent with the current guidelines.8

Definition of the control group The control group consisted of unrelated subjects with structurally normal hearts and no history of any type of atrial arrhythmia including AVNRT, premature ventricular contractions, or ventricular tachycardia.

Ajmaline challenge test The ajmaline challenge test was performed according to the second BrS consensus conference report.3 Ajmaline was administered as continuous intravenous infusion at a rate of 1 mg/kg bodyweight over 5 minutes. Criteria for discontinuation of ajmaline infusion were the development of diagnostic type 1 Brugada pattern, ventricular arrhythmias, and QRS widening to Z130% of baseline. Patients with AVNRT underwent an ajmaline challenge after electrophysiological study and catheter ablation. All tests were performed by the same investigator in the same center using the same equipment with the same standard settings. Twelve-lead ECGs were recorded by using a standard electrocardiograph (ECG9132K, Nihon Kohden Corporation, Nakano-Ku, TKY, Japan) with standard settings (paper speed 25 mm/s and gain setting 10 mm/mV) in all subjects. PR and corrected QT intervals, QRS duration, and QRS axis were automatically

Hasdemir et al

Spontaneous AVNRT and Brugada Syndrome

analyzed, and P-wave duration and amplitude were measured manually in lead II at baseline and after ajmaline challenge. Presence and type of intraventricular conduction disturbances after ajmaline challenge were defined according to slightly modified criteria drawn from the current guidelines for ECG interpretation.13

3 2.98; 95% confidence interval 1.029–8.664; P ¼ .044) in patients with BrS as compared with patients without BrS. This relationship was stronger in female patients (odds ratio 3.67; 95% confidence interval 0.974–13.858; P ¼ .055). None of the patients had ST-T wave changes during their chest pain episodes. None of the patients had a history of unexplained syncope or cardiac arrest.

Genetic screening and analysis Genetic screening and analysis was performed in 17 patients (13 women and 4 men; mean age 46 ⫾ 14 years): 10 with concealed BrS and 7 with previous diagnosis of manifest or suspicious BrS and clinical AVNRT from our observational BrS cohort. Genomic DNA was extracted from peripheral blood leukocytes and amplified. Libraries were constructed using Ion AmpliSeq Custom Panel to amplify all exons and intron borders of a targeted next-generation 30-gene panel on a Veriti Thermal Cycler. Details are presented in the Online Supplement.

Statistical analysis

Baseline 12-lead ECG and electrophysiological characteristics All subjects were in normal sinus rhythm. None of the patients with AVNRT had any signs of Brugada pattern on baseline standard 12-lead ECGs. Four patients with AVNRT (2 with concealed BrS and 2 without BrS) had type 3 Brugada pattern on baseline 12-lead ECGs with leads V1 and V2 recorded at the third intercostal space. A comparison of baseline 12-lead ECG and electrophysiological characteristics is presented in Table 1.

Twelve-lead ECG characteristics during AVNRT

Values are expressed as range and mean ⫾ SD. Differences between means were calculated using the t test or the MannWhitney test, where appropriate. Categorical variables were compared using the χ2 test or the Fisher exact test. The Pearson correlation coefficient was calculated for the relationship of 2 numeric data. The association between clinical and 12-lead ECG predictors of concealed BrS was examined by multivariate stepwise logistic regression analysis. P o .05 (2-sided) was considered statistically significant.

Pseudo-r0 deflection in lead V1 was more common in patients with concealed BrS than in patients without BrS (92.3% vs 72.3%; P ¼ .047). The presence of pseudo-S wave in inferior leads (53.8% vs 58.4%; P ¼ .835), P-in-QRS pattern (7.7% vs 16.9%; P ¼ .33), and QRS alternans (11.5% vs 5.7%; P ¼ .387) were not statistically significant in patients with concealed BrS compared with patients without BrS, respectively.

Results

Brugada pattern–inducing drug exposure and family history

Prevalence of concealed BrS and patient characteristics Concealed BrS was detected in 26 of 96 patients with clinical spontaneous AVNRT (27.1%) and in 3 of 66 control subjects (4.5%) (odds ratio 7.80; 95% confidence interval 2.251– 27.024; P r .001). The patient population was matched to the control group in terms of age (46 ⫾ 15 years vs 42.7 ⫾ 9.5 years; P ¼ .089), female sex (69.8% vs 63.6%; P ¼ .412), and left ventricular ejection fraction (64.1% ⫾ 5.1% vs 64.7% ⫾ 3.8%; P ¼ .594). The patient population presented with a higher prevalence of palpitations (100% vs 6.1%; P r .001), chest pain (24.5% vs 7.6%; P ¼ .006), diabetes mellitus (20% vs 3%; P ¼ .002), and systemic hypertension (30.5% vs 9.1%; P ¼ .001) than did the control group. The prevalence of syncope (18.8% vs 10.6%; P ¼ .159) and migraine headaches (21.3% vs 16.7%; P ¼ .468) were similar between the patient population and the control group. Patients with AVNRT and concealed BrS were predominantly female patients (88.5% vs 62.9%; P ¼ .015) and had higher prevalence of chest pain (38.5% vs 18.6%; P ¼ .042) and migraine headaches (38.4% vs 14.2%; P ¼ .008) compared with patients without concealed BrS (Table 1). Multivariate logistic regression analysis showed a significant association between BrS and migraine headaches (odds ratio

Drug-induced initiation (paroxetine [n ¼ 1]) and/or worsening (fluoxetine [n ¼ 1], flecainide [n ¼ 1], and carbamazepine [n ¼ 1]) of duration and/or frequency of palpitations (along with documented AVNRT) were observed in 4 patients with concealed BrS and in 1 patient (worsening with propafenone) without concealed BrS (15.4% vs 1.4%; P ¼ .006). Family history of AVNRT was documented in 2 patients with concealed BrS (7.7%) and in 2 patients without BrS (2.9%). None of the patients in the study population had family history of sudden unexpected death in the first-degree relatives.

Ajmaline challenge test characteristics A comparison of 12-lead ECG characteristics before and after the ajmaline challenge test in patients with AVNRT with and without concealed BrS and control subjects is presented in Table 2. In the patient population, type 1 Brugada pattern after ajmaline challenge was uncovered with leads V1 and V2 in the fourth intercostal space in 10 patients (38.5%), in the third intercostal space in 14 patients (53.8%), and in the second intercostal space in 2 patients (7.7%). Representative examples of 12-lead ECG characteristics at baseline and after ajmaline challenge in 2 genotypepositive patients are illustrated in Figures 1 and 2.

4

Heart Rhythm, Vol 0, No 0, Month 2015

Table 1 Comparison of clinical, electrocardiographic, and electrophysiological characteristics of patients with spontaneous AVNRT with and without concealed BrS Variable Age (y) Age of onset of AVNRT (y) Sex: female Presenting symptoms Palpitations Chest pain Syncope Vasovagal AVNRT-related Presence of systemic hypertension Presence of diabetes mellitus Migraine headaches Left ventricular ejection fraction (%) Initiation of AVNRT During high sympathetic tone (daily activity) During high parasympathetic tone (sleep/postprandial) Baseline 12-lead ECG characteristics QRS axis (deg) QRS fragmentation (inferior/lateral/V1–V3) (%) Early repolarization pattern (inferior/lateral) (%) Baseline electrophysiology and AVNRT characteristics Baseline AH interval (ms) Baseline HV interval (ms) Spontaneous AVNRT rate (cycle length in ms) Inducible AVNRT rate (cycle length in ms) Type of AVNRT Slow/fast Slow/slow Fast/slow Slow/fast (left atrial variant) AVNRT His-V interval (ms) AVNRT V-AHis ınterval (ms) AVNRT V-Aright atrial appendage interval (ms) AVNRT V-Adistal coronary sinus interval (ms) Coexisting atrial and ventricular arrhythmias Atrial arrhythmias (AT/AF) Ventricular arrhythmias (PVC and/or VT)

Patients with BrS (n = 26)

Patients without BrS (n = 70)

P

43 ⫾ 13 32 ⫾ 13 23 (88.5)

47 ⫾ 16 34 ⫾ 17 44 (62.9)

.272 .611 .015

26 (100) 10 (38.5) 7 (26.9) 5 (19.2) 2 (7.7) 6 (23.1) 3 (11.5) 10 (38.4) 65.3 ⫾ 4.7

70 (100) 13 (18.6) 11 (15.7) 6 (8.6) 5 (7.1) 23 (32.9) 16 (22.9) 10 (14.2) 63.4⫾5.2

1 .042 .211 .162 .927 .354 .216 .008 .216

24 (92.3) 10 (38.5)

68 (97.1) 21 (30)

.292 .431

20 (35 to 80) 84 12

26 (34 to 88) 69.6 14.5

.422 .162 .757

67 ⫾ 48 ⫾ 339 ⫾ 344 ⫾

20 7 33 38

26 (100) 0 0 0 45 ⫾ 6 5.5 (15 to 22) 113.1 ⫾ 19.2 94.7 ⫾ 17.1 1 (3.8) 1 (3.8)

69 ⫾ 47 ⫾ 345 ⫾ 361 ⫾

15 8 50 44

65 (92.9) 3 (4.3) 1 (1.4) 1 (1.4) 45 ⫾ 9 7 (18 to 44) 98.1 ⫾ 19.3 83.7 ⫾ 16.4 5 (7.1) 5 (7.1)

.634 .463 .520 .196 .213 1.0 1.0 1.0 .946 . .009 .026 1.0 1.0

Data are given as mean ⫾ SD, n (%), and median (range). P o .05 was considered statistically significant. AF = atrial fibrillation; AT = atrial tachycardia; AVNRT = atrioventricular nodal reentrant tachycardia; BrS = Brugada syndrome; PVC = premature ventricular contraction; VT = ventricular tachycardia.

Genetic characteristics Genetic screening and analysis identified 19 mutations or rare variants in 13 of 17 patients with both AVNRT and BrS for a yield of 76.5%. We identified 19 mutations or rare variants in 13 genes (SCN10A:5, SCN1B:3, ANK2:1, HEY2:1, SCN5A:1, CACNA1C:1, GPD1L:1 KCNE1:1, KCNE2:1, KCNJ11:1, KCNJ8:1, RyR2:1, and PKP2:1). Clinical and genetic characteristics of genetically screened patients are summarized in Tables 3 and 4, respectively. Six of 17 patients (35.3%) were also found to carry a second potentially pathogenic BrS mutation (Table 4). Owing to the lack of availability of family members, it is uncertain which genetic variant (or both) is causative in those cases where 2 variants were uncovered. Interestingly, 10 of these 13 genotype-positive patients (76.9%) were found to harbor genetic variants that are either known or suspected to cause a loss of function of cardiac sodium channel current (INa) (SCN5A, SCN10A, SCN1B, GPD1L, PKP2, and HEY2).14–17

Patient 14 carried mutations in both GPD1L, which is suspected to cause a loss of function in INa, and KCNJ8, known to cause a gain of function in ATP-sensitive potassium channel current (IK-ATP). While both mechanisms have been shown to contribute to the BrS phenotype, only the GPD1L mutation is likely to contribute to the AVNRT phenotype. Patient 15 was found to carry mutations in both SCN1B, which have been implicated in loss of function in INa, and RyR2, which is known to modulate release of calcium from the sarcoplasmic reticulum. We hypothesize that the SCN1B variant is the one that contributes to both AVNRT and BrS.

Discussion AVNRT is the most common form of PSVT.18 The coexistence of clinical AVNRT and manifest BrS has been previously reported.4,5 There are several important aspects of the relationship between spontaneous AVNRT and BrS.

Hasdemir et al

Spontaneous AVNRT and Brugada Syndrome

5

Table 2 Comparison of 12-lead ECG characteristics before and after ajmaline challenge test in patients with atrioventricular nodal reentrant tachycardia with and without concealed BrS and control subjects without BrS

Baseline P-wave duration (ms) Ajmaline P-wave duration (ms) Change in P-wave duration (ms) Baseline P-wave amplitude (mV) Ajmaline P-wave amplitude (mV) Change in P-wave amplitude (mV) Baseline PR interval (ms) Ajmaline PR interval (ms) Change in PR interval (ms) Baseline QRS duration (ms) Ajmaline QRS duration (ms) Change in QRS duration (ms) Baseline QTc interval (ms) Ajmaline QTc interval (ms) Change in QTc interval (ms) Baseline 12-lead ECG characteristics QRS axis (deg) QRS fragmentation (inferior/lateral/V1–V3) (%) Early repolarization pattern (inferior/lateral) (%) Type of conduction disturbances with ajmaline Right bundle branch block (%) Left anterior fascicular block (%) Left posterior fascicular block (%) Left bundle branch block (%) Nonspecific intraventricular conduction delay (%) QRS fragmentation (inferior/lateral/V1–V3) (%)

Patients with BrS (n = 26)

Patients without BrS (n = 70)

Control subjects (n = 63)

P

115.8 ⫾ 160.9 ⫾ 45.1 ⫾ 0.177 ⫾ 0.224 ⫾ 0.047 ⫾ 144 ⫾ 189.4 ⫾ 45.4 ⫾ 86.4 ⫾ 124.6 ⫾ 38.2 ⫾ 402.5 ⫾ 447.1 ⫾ 44.6 ⫾

120.1 ⫾ 164.1 ⫾ 43.9 ⫾ 0.184 ⫾ 0.225 ⫾ 0.041 ⫾ 150.7 ⫾ 195.2 ⫾ 44.5 ⫾ 87.1 ⫾ 122.9 ⫾ 35.9 ⫾ 409.8 ⫾ 448.2 ⫾ 38.4 ⫾

114.1 ⫾ 148.8 ⫾ 34.7 ⫾ 0.188 ⫾ 0.209 ⫾ 0.020 ⫾ 152.7 ⫾ 194.3 ⫾ 41.6 ⫾ 87.1 ⫾ 116.1 ⫾ 28.9 ⫾ 401.8 ⫾ 438.9 ⫾ 37.1 ⫾

0.188 o.001* .006* 0.788 0.56 .002* 0.569 0.912 0.688 0.178 o.001* o.001* .008† .004* 0.967

24.6 22.5 15.9 0.047 0.061 0.046 21 31.1 17.5 6.6 7.3 5.8 14.4 17.8 13.7

20 (median) 84 12 72 24 0 8 12 100

17.7 23.9 19.9 0.046 0.056 0.048 24.1 33.6 17.8 8.9 11.5 7.9 15.1 16 14

26 (median) 69.6 14.5 18.8 10.1 11.6 1.4 50.7 87

12.1 19.2 16.7 0.042 0.047 0.035 17.9 25.1 14.1 8.9 10.6 8.3 15.9 15.9 13.3

41 (median) 52.4 14.3 20.6 11.1 7.9 0 39.7 73

0.352 .011‡ 0.876 o.0001§ 0.181 0.195 0.560 .003§ .005‡

Data are given as mean ⫾ SD or n (%). P o .05 was considered statistically significant. BrS ¼ Brugada syndrome; ECG ¼ electrocardiographic; QTc ¼ corrected QT Interval. * Control vs patients with and without BrS. † Patients without BrS vs patients with BrS and control subjects. ‡ Control vs patients with BrS. § Patients with BrS vs patients without BrS and control subjects.

First is the prevalence of concealed BrS in patients with clinical spontaneous AVNRT and the epidemiology of BrS; second is the female predominance; third is the clinical and ECG characteristics of these patients, which help clinicians predict the presence of concealed BrS and arrive at decisions regarding management of these patients; fourth is the potential relationship between spontaneous AVNRT and the risk of sudden cardiac arrest; and fifth is the underlying genetic characteristics and the mechanistic link between spontaneous AVNRT and concealed BrS, as discussed below.

Epidemiology of BrS The prevalence of concealed BrS in patients presenting primarily with spontaneous AVNRT was 27.1% in our study population. This high prevalence has important implications in terms of epidemiology of BrS. The prevalence of PSVT (AVNRT and atrioventricular [AV] reentrant tachycardia) in the general population in the United States has been reported to be approximately 2.25 cases per 1000 persons.6 If these results could be extrapolated to the entire US population, we estimate that there would be nearly 140,000 new cases per year with PSVT. AVNRT comprises 60% of all PSVT. If high prevalence of concealed BrS in patients with

spontaneous AVNRT is confirmed with further studies, the prevalence of concealed BrS may approach to approximately 21,000 new cases per year. The world-wide prevalence of a Brugada ECG pattern (type 1, 2, and 3) in the general population is estimated to be 0.5–1.6 cases per 1000 persons.14 The prevalence of concealed BrS in the general population is unknown. The prevalence of concealed BrS was surprisingly high in our control subjects. Our results are consistent with a recently published study, showing that the prevalence of feverinduced type 1 Brugada pattern in otherwise asymptomatic patients is 2%, suggesting that asymptomatic BrS is more prevalent than previously thought.19

BrS and female sex Previous studies reported that clinical phenotype of BrS (syncope and sudden cardiac arrest) and presence of type 1 Brugada pattern (spontaneous or drug-induced) is 8–10 times more prevalent in male than in female patients.20 Demographic characteristics of our study population revealed that concealed BrS in patients with spontaneous AVNRT is mainly a disease of female patients. Considering the prevalence of clinical spontaneous AVNRT in the general population and the prevalence of concealed BrS in patients

6

Heart Rhythm, Vol 0, No 0, Month 2015

Figure 1 Twelve-lead electrocardiograms of a 45-year-old female patient who presented with a 4-year history of paroxysmal supraventricular tachycardia (patient 12 in Tables 3 and 4) recorded at baseline and after ajmaline challenge. Electrophysiological study revealed slow/fast atrioventricular nodal reentrant tachycardia. She underwent successful slow pathway ablation. Genetic analysis revealed a missense mutation in SCN10A (p.Val1697IIe). Ajmaline challenge resulted in prolongation of PR interval (190 ms), QRS duration (134 ms) with left bundle branch block, QRS axis of 331, and type 1 Brugada pattern in leads V1 and V2. ICS = intercostal space.

with spontaneous AVNRT, there is potentially a large subgroup of female patients in the population with BrS. This is important in terms of epidemiology, natural history, and prognosis of BrS in female patients.21 The cause(s) of female predominance in patients with AVNRT and concealed BrS is unknown. The female predilection to clinical spontaneous AVNRT is well recognized. Previous studies have shown that female patients with spontaneous AVNRT have shorter slow pathway effective refractory periods than do male patients.22 No sex differences were noted in the fast pathway effective refractory periods. Therefore, a wider “tachycardia window” (ie, the difference between the fast and slow pathway refractory periods) may explain higher prevalence of AVNRT in female patients.22 It has also been shown that the same mutation in SCN5A can cause progressive cardiac conduction defect in female patients, but BrS in male patients, of the same family.23 This together with

the fact that INa is smaller in female vs male ventricular myocytes may explain why AVNRT is more common in female patients in general and in patients with concealed BrS.24

Clinical relevance of concealed BrS As documented in our study, certain clinical and ECG characteristics should raise the possibility of concealed BrS in patients with spontaneous AVNRT. The coexistence of clinical spontaneous AVNRT and concealed BrS is of potential clinical relevance for the appropriate management of those patients in terms of cautious use of certain antiarrhythmic agents known to exacerbate the Brugada phenotype (class IC antiarrhythmic agents, propranolol, and calcium-channel blockers), avoidance of Brugada pattern–inducing noncardiac drugs such as certain selective serotonin reuptake inhibitors and antiepileptic agents,

Hasdemir et al

Spontaneous AVNRT and Brugada Syndrome

7

Figure 2 Twelve-lead electrocardiograms of a 40-year-old female patient who presented with an 18-year history of paroxysmal supraventricular tachycardia (patient 15 in Tables 3 and 4) recorded at baseline and after ajmaline challenge. Electrophysiological study revealed slow/fast atrioventricular nodal reentrant tachycardia. She underwent successful slow pathway ablation. Genetic analysis revealed a missense mutation in SCN1B (p.Val168Ala). Ajmaline challenge resulted in prolongation of PR interval (156 ms), QRS duration (132 ms), left anterior fascicular block with a QRS axis of 491, and type 1 Brugada pattern in leads V1 and V2. ICS = intercostal space.

consideration of standard preventive measures such as use of antipyretics during fever, and long-term follow-up of these patients.25

be more prevalent in those harboring mutations in SCN10A, which was the gene we found to be most often associated with BrS and AVNRT.15

Clinical presentation of concealed BrS

Twelve-lead ECG characteristics during AVNRT

There are limited data in terms of relationship between chest pain and migraine headaches and BrS. One important cause of chest pain in patients with BrS is coronary artery spasm associated with ST-segment elevation on a 12-lead ECG.26 None of our patients with concealed BrS displayed ST-T wave changes during their chest pain episodes. Further studies are required in order to determine the prevalence and underlying mechanism(s) of chest pain and migraine headaches in patients with concealed BrS. It is noteworthy that among patients with BrS, chest pain has been reported to

Pseudo-r0 deflection in lead V1 during PSVT is a specific finding for slow/fast type AVNRT, present in 40%–60% of patients.27 Our study population with concealed BrS showed a higher (92.3%) prevalence of pseudo-r0 deflection in lead V1 during AVNRT as compared to previous reports. Analysis of the atrial activation sequence during AVNRT in humans revealed that the pseudo-r0 deflection in lead V1 is determined by activation of the superolateral aspect of the right atrium.10 Our patient population with AVNRT and BrS had longer V-Aright atrial appendage time interval compared to

8 Table 3

Heart Rhythm, Vol 0, No 0, Month 2015 Clinical and electrophysiological characteristics of genetically screened patients

Patient no.

Age (y)

Age of onset of AVNRT (y)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

54 59 21 30 65 61 57 38 50 18 52 45 29 60 40 50 28

44 18 20 30 60 27 21 16 42 16 25 41 28 54 22 28 21

Sex

Chest pain

M M F F M F F M F F F F F F F F M

No No Yes No No No No Yes No Yes Yes Yes No No Yes No No

Syncope

Migraine headache

Type of BrS

Type of AVNRT

Initiation of AVNRT

Spontaneous AVNRT cycle length (ms)

Associated arrhythmias

Yes No No No No Yes No Yes No Yes No No No Yes No No Yes

No No No Yes No No No No No No Yes No No Yes No No Yes

Manifest Suspicious Suspicious Suspicious Suspicious Suspicious Suspicious Concealed Concealed Concealed Concealed Concealed Concealed Concealed Concealed Concealed Concealed

S/F S/F S/F S/F S/F S/S S/F S/F S/F S/F S/F S/F S/F S/F S/F S/F S/F

HST HST ⫾ HST ⫾ HST HST HST ⫾ HST HST HST ⫾ HST HST HST ⫾ HST HST ⫾ HST ⫾ HST ⫾ HST

300 410 360 330 290 260 310 330 320 350 360 310 350 345 320 345 275

PVCs ⫾ VT None None None None None None None PVCs None None None None None None None None

sleep sleep

sleep

sleep

sleep sleep sleep sleep

AVNRT ¼ atrioventricular nodal reentrant tachycardia; BrS ¼ Brugada syndrome; F ¼ female; HST ¼ high sympathetic tone; M ¼ male; PVC ¼ premature ventricular contraction; S/F ¼ slow/fast; S/S ¼ slow/slow; VT ¼ ventricular tachycardia.

patients with AVNRT without BrS (Table 1). As a result of this finding, we hypothesize that a higher prevalence of pseudo-r0 deflection may result from the delayed activation of the superolateral aspect of the right atrium as a result of underlying genetic mutations.

Concealed BrS: The link between spontaneous AVNRT and sudden cardiac arrest? Clinical spontaneous AVNRTs may not always be as benign as generally perceived. A previous study reported 3 female patients with structurally normal heart and spontaneous AVNRT, who developed polymorphic ventricular tachycardia and/or cardiac arrest.28 On the basis of our findings, underlying concealed BrS can be the myocardial substrate for malignant ventricular arrhythmias in those patients.

Genetics and the mechanistic link between spontaneous AVNRT and BrS Our findings point to a mechanistic link between spontaneous AVNRT and BrS secondary to a loss of function of INa. Reduced levels of INa are known to suppress INadependent parameters such as excitability and conduction, leading to heterogeneous prolongation of refractoriness; thus, this facilitates the development of unidirectional block and reentry, giving rise to AVNRT. Loss of function of INa is also known to cause an outward shift in the balance of currents in right ventricular epicardium. This outward shift in the balance of current can accentuate the epicardial action potential notch, thus giving rise to repolarization and depolarization abnormalities that give rise to the BrS phenotype, including the development of phase 2 reentry and polymorphic ventricular tachycardia.29 Atrial vulnerability and structural remodeling have been shown to be enhanced in patients with BrS.2 Most forms of

AVNRT are created by reentry between 2 (or more) atrial connections to the AV node. The fast AV nodal pathway (shortest conduction time) is formed by transitional cells crossing the tendon of Todaro superiorly. Two slow AV nodal pathways are formed by the rightward and leftward inferior extensions of the AV node.12 Atrial tissue surrounding Koch’s triangle is involved in all types of AVNRT. Optical mapping data obtained from isolated rabbit and adult mongrel dog AV nodal preparations suggested that AV nodal pathways are located outside the compact AV node and atrial and transitional cells are involved in the reentrant circuit of AVNRT.30 In light of these findings, we hypothesize that a loss of function of INa secondary to a mutation in sodium channel–related genes (eg, SCN5A, SCN1B, and SCN10A) may cause reduced excitability, thus leading to block in one of the AV nodal pathways and the development of reentrant re-excitation. The high yield of presumed pathogenic variants (76.5%) is likely due to the fact that patients with BrS were selected on the basis of their presentation with AVNRT. Pathogenicity of the variants was presumed on the basis of relatively low 1000 genome minor allele frequency, in silico pathogenicity prediction tools, previous associations with BrS, as well as previously conducted functional expression studies, suggesting pathogenicity. In addition, INa or depolarization reserve has been shown to be smaller in female patients than in male patients; therefore, the impact of these mutations may be greater, accounting for the higher incidence of AVNRT in female patients.24 Interestingly, lower repolarization reserve in female patients owing to lower levels of transient outward potassium current (Ito) account for a lower incidence of BrS in female patients.31

Diabetes mellitus and systemic hypertension: The common links for the AVNRT substrate? Spontaneous AVNRT may represent an acquired atrial remodeling process caused by diabetes mellitus and/or

Hasdemir et al

Table 4

Summary of mutations and rare variants in genetically screened patients Type of Reported ID Exon mutation

Change in nucleotide

Change in amino acid

1

PKP2

Missense

c.548G4A

S183N

2 3 4

SCN10A – 18 HEY2 rs138413872 5 Unknown CaCNA1c – 19

Missense Missense

c.3233T4C c.896C4G

V1078A A299G

0 0

0 0.04

Deletion

E850del

0

0.31

SCN1B

rs201209882

5

Missense

c.2548_2550 delGAG c.638G4A

G213D

0.05

0

0

Damaging

SCN10A KCNJ11 SCN1B

rs77804526 27 rs41282930 5 rs121434627 3

Missense Missense Missense

c.5089G4A c.1154C4G c.259G4C

V1697I S385C E87Q

0.41 1.14 0

1.01 1.02 0

0.8 .05 .64

Tolerated Damaging Tolerated

rs34270799

38

Missense

c.9900C4A

S3300R

1.6

2.28

.04

Damaging

Possibly damaging

1

Missense

c.22A4G

T8A

0.19

0.49

0

Damaging

5 6 7

ANK2

Global MAF (1000 genome)

Global MAF (ESP)

SIFT— score

SIFT— prediction

Polyphen-2— prediction

Reference 17

0.73 0.33

Tolerated Tolerated

Benign Benign Burashnikov E et al (2010) Possibly damaging Benign Benign Benign

8 9 10

Unknown Unknown KCNE2 rs2234916

11 12 13

SCN10A SCN10A SCN10A SCN5A

rs73062575 rs77804526 rs138832868 rs41311127

17 27 21 22

Missense Missense Missense Missense

c.3133C4A c.5089G4A c.3803G4A c.3878T4C

P1045T V1697I R1268Q F1293S

0.5 0.41 0.1 0

2.27 1.01 .2 0

0.88 0.8 0.01 0.11

Tolerated Tolerated Damaging Tolerated

Probably damaging Benign Benign Benign Benign

14

GPD1L

rs72552293

4

Missense

c.370A4G

I124V

0.1

0.1

0.76

Tolerated

Benign

KCNJ8

rs117808169

1

Missense

c.263C4G

A88G

0.1

0.0

1

Tolerated

Benign

15

SCN1B RyR2

– –

4 90

Missense Insertion

0 0

0.13

Tolerated

Benign

KCNE1 rs17857111 Unknown

1

Missense

V168A 4316_4317 insR R32H

0 0

16 17

c.503T4C c.12946_12947 insGAA c.95G4A

0

0

0.04

Damaging

Benign

Hu D et al (2014) Watanabe H et al (2008) Mank-Seymour AR et al (2006)

Spontaneous AVNRT and Brugada Syndrome

Patient no. Gene

Sesti F et al (2000) Brody JA et al (2012) Hu D et al (2014) 15 Ackerman MJ et al (2004) Van Norstrand DW et al (2007) Barajas-Martinez H et al (2011)

Splawski I et al (2000)

1000 Genome ¼ The 1000 Human Genome Project Database; ESP ¼ Exome Sequencing Project; MAF ¼ minor allele frequency; SIFT ¼ Sorting Intolerant From Tolerant.

9

10 systemic hypertension in some patients on the basis of higher prevalence of diabetes mellitus and systemic hypertension in patients with AVNRT with or without BrS as compared with the control group in our study population. An atrial electrical and/or structural remodeling process in diabetic and/or spontaneous hypertension animal models is characterized by conduction slowing and heterogeneity, prolongation of action potential duration, and increased interstitial fibrosis, and results in atrial arrhythmias.32,33 In terms of relationship between diabetes mellitus and AVNRT, voltage-gated INa have been identified in human pancreatic β cells. Genetic defects in genes affecting expression of INa in the pancreatic β cells may increase susceptibility to the development of diabetes mellitus via β-cell injury.34

Study limitations The follow-up period was relatively short; thus, we are not able to report the long-term prognosis of these patients. ECG readings were not blinded. We did not perform genetic testing on patients with AVNRT not associated with the BrS phenotype and are therefore unable to gauge to what extent mutations in INa-related genes contribute to this phenotype. Additional functional expression studies are clearly required to ascertain the degree to which the novel mutations uncovered lead to loss of function of INa so as to be able to more definitively ascribe causality of the variants to both AVNRT and the BrS ECG.

Conclusion Our results suggest that spontaneous AVNRT could be the first clinical manifestation of concealed BrS, particularly in female patients, and that genetic variants that reduce INa may provide a mechanistic link between AVNRT and BrS and predispose to expression of both phenotypes.

Acknowledgments We are grateful to Sue Bartkowiak, for assistance with the administration of our genetic database and to Dan Hu, MD, PhD, for helpful discussions.

Appendix Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.hrthm. 2015.03.015.

References 1. Brugada P, Brugada J. Right bundle branch block, persistent ST segment elevation and sudden cardiac death: a distinct clinical and electrocardiographic syndrome. A multicenter report. J Am Coll Cardiol 1992;20:1391–1396. 2. Francis J, Antzelevitch C. Atrial fibrillation and Brugada syndrome. J Am Coll Cardiol 2008;51:1149–1153. 3. Antzelevitch C, Brugada P, Borggrefe M, et al. Brugada syndrome: report of the second consensus conference: endorsed by the Heart Rhythm Society and the European Heart Rhythm Association. Circulation 2005;111:659–670. 4. Eckardt L, Kirchhof P, Loh P, Schulze-Bahr E, Johna R, Wichter T, Breithardt G, Haverkamp W, Borggrefe M. Brugada syndrome and supraventricular tachyarrhythmias: a novel association? J Cardiovasc Electrophysiol 2001;12:680–685.

Heart Rhythm, Vol 0, No 0, Month 2015 5. Schimpf R, Giustetto C, Eckardt L, Veltmann C, Wolpert C, Gaita F, Breithardt G, Borggrefe M. Prevalence of supraventricular tachyarrhythmias in a cohort of 115 patients with Brugada syndrome. Ann Noninvasive Electrocardiol 2008;13: 266–269. 6. Orejarena LA, Vidaillet H Jr, DeStefano F, Nordstrom DL, Vierkant RA, Smith PN, Hayes JJ. Paroxysmal supraventricular tachycardia in the general population. J Am Coll Cardiol 1998;31:150–157. 7. Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders: 2nd edition. Cephalalgia 2004;24:9–160. 8. Priori SG, Wilde AA, Horie M, et al. Executive summary: HRS/EHRA/APHRS expert consensus statement on the diagnosis and management of patients with inherited primary arrhythmia syndromes. Heart Rhythm 2013;10:e85–e108. 9. Take Y, Morita H. Fragmented QRS: what is the meaning? Indian Pacing Electrophysiol J 2012;12:213–225. 10. Huikuri HV, Marcus F, Krahn AD. Early repolarization: an epidemiologist’s and a clinician’s view. J Electrocardiol 2013;46:466–469. 11. Kuga K, Li A, Endo M, Niho B, Suzuki A, Enomoto M, Kanemoto M, Yamaguchi I. Atrial components contributing to pseudo r0 deflection in lead V1 in slow/fast atrioventricular nodal reentrant tachycardia: analysis of the atrial activation sequence by basket catheter isochronal mapping. Circ J 2002;66: 236–240. 12. Nakagawa H, Jackman WM. Catheter ablation of paroxysmal supraventricular tachycardia. Circulation 2007;116:2465–2478. 13. Maury P, Rollin A, Sacher F, et al. Prevalence and prognostic role of various conduction disturbances in patients with the Brugada syndrome. Am J Cardiol 2013;112:1384–1389. 14. Mizusawa Y, Wilde AA. Brugada syndrome. Circ Arrhythm Electrophysiol 2012;5:606–616. 15. Hu D, Barajas-Martinez H, Kahlig K, Rajamani S, Belardinelli L, Pfeiffer R, Dezi F, Antzelevitch C. Genetic variants in SCN10A associated with Brugada syndrome, right bundle branch block and atrioventricular block [abstract]. Heart Rhythm 2012;9:S395. 16. Bezzina CR, Barc J, Mizusawa Y, et al. Common variants at SCN5A-SCN10A and HEY2 are associated with Brugada syndrome, a rare disease with high risk of sudden cardiac death. Nat Genet 2013;45:1044–1049. 17. Cerrone M, Lin X, Zhang M, et al. Missense mutations in plakophilin-2 cause sodium current deficit and associate with a Brugada syndrome phenotype. Circulation 2014;129:1092–1103. 18. Delacrétaz E. Clinical practice. Supraventricular tachycardia. N Engl J Med 2006;354:1039–1051. 19. Adler A, Topaz G, Heller K, Zeltser D, Ohayon T, Rozovski U, Halkin A, Rosso R, Ben-Shachar S, Antzelevitch C, Viskin S. Fever-induced Brugada pattern: how common is it and what does it mean? Heart Rhythm 2013;10: 1375–1382. 20. Benito B, Sarkozy A, Mont L, Henkens S, Berruezo A, Tamborero D, Arzamendi D, Berne P, Brugada R, Brugada P, Brugada J. Gender differences in clinical manifestations of Brugada syndrome. J Am Coll Cardiol 2008;52:1567–1573. 21. Sacher F, Meregalli P, Veltmann C, et al. Are women with severely symptomatic Brugada syndrome different from men? J Cardiovasc Electrophysiol 2008;19: 1181–1185. 22. Liuba I, Jönsson A, Säfström K, Walfridsson H. Gender-related differences in patients with atrioventricular nodal reentry tachycardia. Am J Cardiol 2006;97: 384–388. 23. Kyndt F, Probst V, Potet F, Demolombe S, Chevallier JC, Baro I, Moisan JP, Boisseau P, Schott JJ, Escande D, Le Marec H. Novel SCN5A mutation leading either to isolated cardiac conduction defect or Brugada syndrome in a large French family. Circulation 2001;104:3081–3086. 24. Barajas-Martinez H, Haufe V, Chamberland C, Roy MJ, Fecteau MH, Cordeiro JM, Dumaine R. Larger dispersion of INa in female dog ventricle as a mechanism for gender-specific incidence of cardiac arrhythmias. Cardiovasc Res 2009;81:82–89. 25. Postema PG, Wolpert C, Amin AS, Probst V, Borggrefe M, Roden DM, Priori SG, Tan HL, Hiraoka M, Brugada J, Wilde AA. Drugs and Brugada syndrome patients: review of the literature, recommendations, and an up-to-date website (www.brugadadrugs.org). Heart Rhythm 2009;6:1335–1341. 26. Chinushi M, Kuroe Y, Ito E, Tagawa M, Aizawa Y. Vasospastic angina accompanied by Brugada-type electrocardiographic abnormalities. J Cardiovasc Electrophysiol 2001;12:108–111. 27. Kalbfleisch SJ, el-Atassi R, Calkins H, Langberg JJ, Morady F. Differentiation of paroxysmal narrow QRS complex tachycardias using the 12-lead electrocardiogram. J Am Coll Cardiol 1993;21:85–89. 28. Wang YS, Scheinman MM, Chien WW, Cohen TJ, Lesh MD, Griffin JC. Patients with supraventricular tachycardia presenting with aborted sudden death: incidence, mechanism and long-term follow-up. J Am Coll Cardiol 1991;18: 1711–1719.

Hasdemir et al

Spontaneous AVNRT and Brugada Syndrome

29. Antzelevitch C. Genetic, molecular and cellular mechanisms underlying the J wave syndromes. Circ J 2012;76:1054–1065. 30. Wu J, Zipes DP. Mechanisms underlying atrioventricular nodal conduction and the reentrant circuit of atrioventricular nodal reentrant tachycardia using optical mapping. J Cardiovasc Electrophysiol 2002;13:831–834. 31. Gaborit N, Varro A, Le Bouter S, Szuts V, Escande D, Nattel S, Demolombe S. Gender-related differences in ion-channel and transporter subunit expression in non-diseased human hearts. J Mol Cell Cardiol 2010;49: 639–646.

11 32. Watanabe M, Yokoshiki H, Mitsuyama H, Mizukami K, Ono T, Tsutsui H. Conduction and refractory disorders in the diabetic atrium. Am J Physiol Heart Circ Physiol 2012;303:H86–H95. 33. Lau DH, Shipp NJ, Kelly DJ, et al. Atrial arrhythmia in ageing spontaneously hypertensive rats: unraveling the substrate in hypertension and ageing. PLoS One 2013;8:e72416. 34. Hoeijmakers JG, Faber CG, Merkies IS, Waxman SG. Channelopathies, painful neuropathy, and diabetes: which way does the causal arrow point? Trends Mol Med 2014;20:544–550.

CLINICAL PERSPECTIVES Atrioventricular nodal reentrant tachycardia (AVNRT) is the most common form of paroxysmal supraventricular tachycardia. The causes of AVNRT are unresolved. Spontaneous or inducible AVNRT has been previously shown to coexist with manifest Brugada syndrome (BrS). A concealed BrS electrocardiogram was uncovered in 27.1% of patients with AVNRT in our study population. Female sex, presence of chest pain and migraine headaches, worsening of duration and/or frequency of AVNRT with Brugada pattern–inducing drugs, and presence of the pseudo-r0 pattern in lead V1 during AVNRT were predictors of concealed BrS. Genetic screening identified 19 mutations or rare variants in 13 genes in 13 of 17 patients with both AVNRT and BrS (yield ¼ 76.5%). Ten of these 13 genotype-positive patients (76.9%) harbored genetic variants known or suspected to cause a loss of function of cardiac sodium channel current (SCN5A, SCN10A, SCN1B, GPD1L, PKP2, and HEY2). Our findings expand the phenotypic overlap of loss of function mutations in sodium channel current to include AVNRT. Current treatment options for AVNRT include atrioventricular nodal–blocking agents (verapamil, diltiazem, propranolol, and metoprolol), class IC antiarrhythmic agents (propafenone and flecainide), and catheter ablation. The identification of frequent coexistence of clinical spontaneous AVNRT and concealed BrS calls for greater vigilance in the use of certain antiarrhythmic agents that are known to exacerbate the Brugada phenotype (verapamil, diltiazem, propranolol, and class IC antiarrhythmic agents), avoidance of Brugada pattern–inducing noncardiac drugs, and possible need for standard preventive measures such as use of antipyretics during fever, as well as long-term follow-up.