Holter Electrocardiographic Approach to Predicting Outcomes of Pediatric Patients With Long QT Syndrome

Abstract

Background: This study was performed to clarify the clinical findings of pediatric patients diagnosed with long QT syndrome (LQTS) through electrocardiographic screening programs and to predict their outcome using Holter electrocardiographic approaches.

Methods and Results: This retrospective study included pediatric patients with a Schwartz score of ≥3.5 who visited the National Hospital Organization Kagoshima Medical Center between April 2005 and March 2019. Resting 12-lead and Holter electrocardiograms were recorded at every visit. The maximum resting QTc and maximum Holter QTc values among all recordings were used for statistical analyses. To test the prognostic value of QTc for the appearance of cardiac events after the first hospital visit, receiver operating characteristic curves were used to calculate the area under the curve (AUC). Among 207 patients, 181 (87%) were diagnosed through screening programs. The prevalence of cardiac events after the first hospital visit was 4% (8/207). Among QTc at diagnosis, maximum resting QTc, and maximum Holter QTc, only maximum Holter QTc value was a predictor (P=0.02) of cardiac events after the hospital visit in multivariate regression analysis. The AUC of the maximum Holter QTc was significantly superior to that of maximum resting QTc.

Conclusions: The maximum Holter QTc value can be used to predict the appearance of symptoms in pediatric patients with LQTS.

Long QT syndrome (LQTS) is a genetic disorder characterized by abnormal ventricular myocardial repolarization and prolongation of the QT interval on an electrocardiogram (ECG).^1–3 LQTS-related cardiac events include syncope or life-threatening arrhythmic events, such as aborted cardiac arrest (ACA) or sudden cardiac death (SCD). Since the first description of 4 patients with deafness and a long QT interval in 1957,⁴ considerable progress has been made in understanding the genetics, pathogenesis, diagnosis, and treatment of LQTS.^1–3 Substantial improvements in patient outcomes have recently been seen.⁵ However, life-threatening arrhythmic events still occur in a certain percentage of patients according to relatively recent reports from both single-center^{5,9,13–15,18} and multicenter studies.^{6–8,10–12,16,17,19}

A nationwide, school-based ECG screening program for heart disease in 1st, 7th, and 10th graders in Japan was established in 1994, and participation is mandatory. This program has identified children and adolescents with prolonged QT intervals. Many patients are asymptomatic at diagnosis in the program,²⁰ partly because they may be screened before the appearance of cardiac events. This means that the development of LQTS-related cardiac events could be prevented if careful examinations are performed and follow-up strategies are implemented from the first visit. However, there are no recent reports on all clinical findings and outcomes of patients who have been diagnosed through the ECG screening program.

QTc values on resting 12-lead ECGs are generally used to predict the appearance of LQTS-related cardiac events.^21,22 A longer QTc value is an independent predictor for cardiac events in patients with Type 1 and 2 LQTS (LQT1 and LQT2, respectively).²³ Holter monitoring of the QT interval showed more frequent QT prolongation during the night-time hours than during daytime hours in patients with LQT2 and LQT3.²⁴ QTc values reached their maximum at night or early in the morning in both pediatric controls and pediatric patients with LQTS.²⁵ In our clinical experience, QTc values at night are quite different from those on a routine ECG recorded during daytime hours in the hospital (Figure 1). These findings suggest that we should consider the maximum QTc values among repeated Holter ECGs; however, the use of the QTc value on Holter ECGs remains controversial.^18,26

Figure 1.

(A,B) Differences in QTc values between the resting electrocardiogram (ECG) (A) and Holter ECG (B) in a 12-year-old girl who was screened by an ECG screening program because of a prolonged QT interval. (A) Resting ECG on the patient’s first hospital visit showed a QT interval of 0.53 s, an RR interval of 1.25 s, and QT intervals corrected by the Bazett formula and Fridericia formula (QTcB and QTcF, respectively) of 0.474 and 0.492, respectively, on the marked beat (○). The Holter recording was started on the same day of the patient’s first hospital visit. (B) Continuous Holter ECG recording during sleep at 00:52 hours the next day. The QTc value was extremely prolonged during the Holter recording. (C) Changes in QTc values in consecutive beats from QT-1 to QT-42 in (B) showing changes in QTcB (Left) and QTcF (Right). Genetic testing using a next-generation sequencer failed to reveal a pathogenic variant.

Therefore, the aims of the present study were to clarify the outcomes of young patients with LQTS in the era of the ECG screening program in a single center where follow-up strategies were unified, and to determine whether QTc values from Holter ECG recordings are applicable and useful for predicting the appearance of cardiac events.

Methods

Study Population

This retrospective study included 207 infants, children, and adolescents who were diagnosed as a high probability of LQTS (i.e., an LQTS [Schwartz] score of ≥3.5) at the National Hospital Organization Kagoshima Medical Center between April 2005 and March 2019 and who were diagnosed before 20 years of age. Kagoshima City has a population of approximately 600,000. In 2005, an outpatient clinic for inherited arrhythmias was established at the National Hospital Organization Kagoshima Medical Center. Children and adolescents who were screened through the screening program in Kagoshima City all attended this center from 2008 to 2013,¹⁹ within the study period. Thereafter, most (but not all) children and adolescents identified through the screening program also attended the center. The end of the study was set to March 2019 to allow for a longer follow-up period for recent patients. Outcomes were finally checked on March 31, 2022.

The study was approved by the Ethics Committee of the National Hospital Organization Kagoshima Medical Center (No. 30-69). The procedures in this study were performed in accordance with the Declaration of Helsinki and the ethical standards of the institutional committee on human experimentation.

Patients who had secondary causes of LQTS or who had congenital heart diseases, epilepsy, and attention deficit hyperactivity disorder were excluded. Patients with Jervell and Lange-Nielsen syndrome were also excluded because this disease has an extremely severe clinical course. LQTS-related cardiac events were defined as syncope, seizure, documented torsade de pointes (TdP), ACA, SCD, and appropriate implantable cardioverter defibrillator (ICD) shock; however, there were no patients with either SCD or ICD shock in the present study. Syncope was defined as a sudden loss of consciousness with spontaneous recovery and excluded all events assessed to be likely vasovagal in nature (e.g., emotional reactions, heat or dehydration, or abrupt postural changes).⁵ All patients were primarily followed by a single cardiologist (M.Y.).

Patients were classified by diagnostic events into a screened group or a clinical (not screened) group. The screened group comprised patients who were screened by the program. The clinical group included patients who visited the hospital with LQTS-related cardiac events, those who were diagnosed by a familial study, and those who were diagnosed by chance, namely during the course of examinations for other conditions, such as heart murmur, palpitation, or chest pain.

ECG Recording and Measurement of QT/RR Intervals

The patients and their parents were asked to visit the outpatient clinic at least once a year. A resting 12-lead ECG and a Holter ECG were recorded at each visit in all patients, including infants. A Master 2-step test (if possible) and a treadmill exercise test (if needed) were also performed. Chest radiography, serum biochemistry, and echocardiography were performed at the first visit.

One of the authors (M.Y.) manually measured 3 consecutive QT/RR intervals in lead V5 on resting ECGs using the tangent method. Each QT interval was corrected by the Bazett formula and Fridericia formula (termed QTcB and QTcF, respectively). The mean QTc value of 3 consecutive beats was calculated. After obtaining QTc values from each visit, the maximum QTc value among all repeated resting ECGs was termed the maximum resting QTc and used for analysis.

For measurement of QT intervals on Holter ECGs (SCM-8000 System, V54-11; Fukuda Denshi, Tokyo, Japan), 5 representative periods were selected to exclude selection bias: night-time sleeping, wake up time, and daytime activities in the morning, afternoon, and evening. Each period was arbitrarily defined as 02:00–04:00, 06:00–08:00, 10:00–12:00, 14:00–16:00, and 20:00–22:00 hours, respectively. In each period, ECGs at maximum, mean, and minimum heart rates were chosen. To obtain QTc values near the longest QTc, we tried to find the place where the heart rate increased abruptly from background heart rate (Figure 1B,C), or where notched T waves were present, or where the appearance of peak of T waves was late (Figure 1B). Three consecutive QT/RR intervals with stable RR intervals in Lead CM5 were manually measured in the same way as resting ECGs. Among all QTc values obtained throughout the day and among all Holter ECGs recorded at each visit, the maximum QTc value on Holter ECGs was termed the maximum Holter QTc and used for analysis.

Follow-up Strategies

All patients were asked to adhere to the following at their first visit:

1. Patients are allowed to swim under the supervision of a person or persons able to perform resuscitation. If the patient has a long QTc value, they should not swim.

2. Patients are able to participate in competitive sports if: they have not experienced syncope during exercise; they do not have findings of worsening of QTc during the treadmill exercise test; there is at least 1 instructor who can perform resuscitation present during the sports activity; and if an automated external defibrillator is available on site.

3. Patients must not take QT-prolonging medications. The patients and their parents were asked to bring a list of any QT-prolonging drugs to give to their doctors at their first hospital. The note was handed to the patient and parents at the first visit.

Medication was started if LQTS-related cardiac events occurred in patents with a diagnosis of LQTS.¹⁴ For asymptomatic patients, medication was recommended based on the longest QTcF value on Holter ECGs during follow-up. The mean difference between the longest Holter QTcF and the resting QTcB on the same day in the same patient was around 50 or 60 ms at the beginning of the study (it finally reached a mean [±SD] of 69±45 ms in this study). If the longest Holter QTcF a patient was around or more than 530 ms (470+60 ms), medication was recommended because the established criterion for starting medication is a QTcB of ≥470 ms on the resting ECG.²¹

Genetic Testing

Genetic testing was performed for patients whose maximum Holter QTcF was around or greater than 500 ms or whose parents requested genetic testing. Until 2018, Sanger sequencing was performed in Kagoshima Medical Center for potassium voltage-gated channel subfamily Q member 1 (KCNQ1), potassium voltage-gated channel subfamily H member 2 (KCNH2), sodium voltage-gated channel alpha subunit 5 (SCN5A), potassium voltage-gated channel subfamily E regulatory subunit 1 (KCNE1), and potassium voltage-gated channel subfamily E regulatory subunit 2 (KCNE2). When we failed to detect any pathogenic variants, and after 2018 generally, genetic analyses were performed using HaloPlex HS custom panels (Agilent Technologies, Santa Clara, CA, USA) that included 56 genes related to inherited arrhythmia syndrome and a bench-top next-generation sequencer (MiSeq; Illumina, San Diego, CA, USA). The data obtained were analyzed using SureCall software (Agilent Technologies) in Shiga University of Medical Science.¹² Based on the Clinical Genome Resource (ClinGen),²⁷ LQTS-related genes were limited to the following: KCNQ1, KCNH2, SCN5A, KCNE1, KCNJ2, and calcium voltage-gated channel subunit alpha1 C (CACNA1C). No patients in this study had calmodulin or triadin gene mutations.

Statistical Analysis

Statistical analyses were performed using statistical software (IBM^® SPSS^® Statistics v23.0; IBM Japan, Ltd., Tokyo, Japan). Data are expressed as the median and interquartile range (IQR). Statistical analyses were performed using the Mann-Whitney test or Fisher’s exact probability test. To predict the presence or absence of cardiac events, logistic regression analysis was performed using the following as independent variables: age at diagnosis, sex, presence or absence of the pathogenic variant, family history of LQTS, screened or clinical (not screened) diagnosis, cardiac events before the first visit, resting QTc at diagnosis, maximum resting QTc value, maximum Holter QTc value, and follow-up period. To test the prognostic value of QTc for the presence of cardiac events, receiver operating characteristic (ROC) curves were used to calculate the area under the curve (AUC). To determine the cut-off QTc, Youden’s index and the distance from the point (0,1) and the ROC plots were used. Two-sided P<0.05 was considered statistically significant.

Results

The characteristics of patients classified by the presence or absence of cardiac events after the hospital visit are presented in Table 1. The mean age and mean QTc values at diagnosis did not differ between the 2 groups; however, maximum resting QTcF, maximum Holter QTcB, and maximum Holter QTcF were significantly longer in the group with than without cardiac events after the first hospital visit (P=0.01, P<0.001, and P<0.001, respectively). In addition, the median Schwartz score was higher in the group with than without cardiac events (P<0.001), primarily because the Schwartz score awards 1 or 2 points for the presence of cardiac events (Supplementary Table 1). The prevalence of cardiac events before the first hospital visit was also higher in the group with than without cardiac events (P=0.008; Table 1).

Table 1.

Characteristics of Patients With LQTS Classified by the Appearance of a Cardiac Event After the First Hospital Visit

	Event (−)	Event (+)	P value	Total
No. (%) subjects	199 (96)	8 (4)		207 (100)
Female sex	84 (42)	7 (88)	0.02	91 (44)
No. (%) screened by the program	124 (87)	7 (88)	0.26	181 (87)
At diagnosis
Age (years)	12.5 [7.7–13.2]	12.8 [8.9–13.2]	0.57	12.5 [7.8–13.2]
Heart rate (beats/min)	75 [67–85]	65 [62–74]	0.047	74 [66–85]
QTcB (resting) (ms)	483 [472–497]	481 [468–490]	0.81	483 [472–497]
Schwartz score	4.5 [4.0–5.0]	6.0 [5.6–6.8]	<0.001	4.5 [4.0–5.0]
Genotype tested	116 (58)	8 (100)	0.02	124 (60)
Genotype determinedA	36 (31)	5 (63)	0.11	41 (33)
LQT1	13	2		15
LQT2	15	3		18
LQT3	2	0		2
Other	6	0		6
Family history of LQTS	45 (23)	1 (13)	0.69	46 (22)
Family history of SCD	2 (1)	0 (0)	>0.99	2 (1)
Overall cardiac events	2 (1)	8 (100)	<0.001	10 (5)
Cardiac events before visit	2 (1)	2 (25)	0.008	4 (2)
Syncope/seizure	2	8		10
Documented TdP	0	0		0
Aborted cardiac arrest	0	0		0
Therapy
Oral	46 (23)	8 (100)	<0.001	54 (26)
β-blocker	24	3		27
Na channel blocker	15	1		16
β-blocker and Na channel blocker	7	4		11
ICD/LCSD	0/0	0/0		0/0
Follow-up period (years)	5.7 [3.3–7.8]	10.4 [9.2–10.9]	<0.001	6.0 [3.4–8.2]
QTc values at various times
QTcF at diagnosis (resting) (ms)	467 [451–479]	472 [457–482]	0.28	468 [452–481]
Maximum QTcB (resting) (ms)	484 [476–499]	499 [476–524]	0.16	486 [477–503]
Maximum QTcF (resting) (ms)	469 [454–486]	487 [473–521]	0.01	471 [456–490]
Maximum QTcB (Holter) (ms)	549 [533–574]	614 [599–631]	<0.001	553 [534–580]
Maximum QTcF (Holter) (ms)	512 [491–532]	582 [548–598]	<0.001	514 [493–540]

Unless indicated otherwise, values are expressed as n (%) or the median [interquartile range]. ^AThe percentages were calculated as (number determined)/(number tested)×100. Statistical analyses were performed using the Mann-Whitney test or Fisher’s exact probability test. ICD, implantable cardioverter defibrillator; LCSD, left cardiac sympathetic denervation; LQT1, LQT2, LQT3, long QT syndrome types 1, 2, and 3, respectively; LQTS, long QT syndrome; QTcB, QT interval corrected by Bazett’s formula; QTcF, QT interval corrected by Fridericia’s formula; SCD, sudden cardiac arrest; TdP, torsade de pointes.

Of the 207 patients, 181 (87%) were diagnosed through the screening program (Supplementary Table 2). Of the 26 patients in the clinical group, 1 was diagnosed by the presence of cardiac events, 8 were diagnosed by a familial study, and 17 were diagnosed by chance. The mean age of patients was lower in the clinical than screened group (P<0.001). The risk for the appearance of cardiac events after the first hospital visit did not differ between the 2 groups (Supplementary Table 2).

The diagnostic yield of genetic testing was low (33%; Table 1), even though patients who underwent such testing were analyzed using next-generation sequencers. The prevalence of the overall presence of cardiac events was significantly higher in patients with than without pathogenic variants (7/41 [17.1%] vs. 3/83 [3.6%], respectively; P=0.01).

Predictive factors for the appearance of cardiac events after the first hospital visit were determined using QTcB values as markers of the QT interval (Table 2). Univariate regression analysis showed that female sex, Schwartz score, cardiac events before the first hospital visit, maximum resting QTc value, maximum Holter QTc values, and the follow-up period were significant predictors of the appearance of cardiac events. Multivariate regression analysis showed that longer maximum Holter QTc was the sole predictive factor among all variables (P=0.01). When QTcF values were used instead of QTcB values, the predictor in the multivariate logistic regression analysis was the same (Supplementary Table 3).

Table 2.

Predictive Factors for the Appearance of Cardiac Events After the First Hospital Visit (Using QTcB Values as Markers of the QT Interval)

Variables	Reference	Univariate regression				Multivariate regression (n=207)
		n	OR	95% CI	P value	OR	95% CI	P value
Age	Per age	207	1.05	0.85–1.30	0.64
Female sexA	Male	207	9.58	1.16–79.4	0.04	8.72	0.49–156	0.14
Schwartz score	Per point	207	9.13	2.67–31.2	<0.001	6.64	0.74–59.8	0.09
Pathogenic variants	Absent	124	3.70	0.84–16.3	0.08
Family history of LQTS	Absent	207	0.49	0.06–4.08	0.51
Clinical groupB	Screened group	207	0.99	0.12–8.42	>0.99
Cardiac events before visit	Absent	207	32.8	3.94–273	0.001	1.04	0.009–121	0.99
QTcB at diagnosis	Per QTcB	207	1,799	2.4E-10–1.4E+16	0.62
Maximum QTcB (resting)	Per maximum QTcB	207	6.0E+7	1.59–2.3E+15	0.04	1.2E-13	1.2E-37–1.3E+11	0.29
Maximum QTcB (Holter)	Per maximum Holter QTcB	207	2.8E+35	1.8E+17–4.5E+53	<0.001	9.3E+28	1.4E+6–6.3E+51	0.01
Follow-up period	Per year	207	1.47	1.16–1.86	0.002	1.29	0.92–1.80	0.14

^ASex was dichotomized (boys=1, girls=2). ^BDiagnosis by screening or clinically was dichotomized (screened group=1, clinical group=2). CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.

Predictive factors for the overall presence of cardiac events were determined using QTcB values as markers of the QT interval (Table 3). Univariate regression analysis showed that female sex, Schwartz score, the presence of pathogenic variants, maximum resting QTcB, maximum Holter QTcB, and follow-up period were significant predictors of the appearance of cardiac events. Multivariate regression analysis showed that a longer maximum Holter QTcB and Schwarz score were independent predictors of the appearance of cardiac events. When QTcF values were used instead of QTcB values, the predictors in the multivariate logistic regression analysis were the same (Supplementary Table 4).

Table 3.

Predictive Factors for the Overall Presence of Cardiac Events by Logistic Regression Analysis (Using QTcB Values as Markers of the QT Interval)

Variables	Reference	Univariate regression				Multivariate regression (n=207)
		n	OR	95% CI	P value	OR	95% CI	P value
Age	Per age	207	1.07	0.88–1.29	0.52
Female sexA	Male sex	207	5.49	1.14–26.5	0.03	3.85	0.62–8.90	0.19
Schwartz score	Per point	207	11.6	3.28–40.7	<0.001	5.08	1.20–21.6	0.03
Pathogenic variantsB	Absent	124	5.49	1.34–22.5	0.02
Family history of LQTS	Absent	207	0.38	0.05–3.04	0.36	1.20	0.31–4.67	0.79
Clinical groupC	Screened group	207	1.80	0.36–8.99	0.47
Cardiac events before visit	Absent	207	5.3E+10	0.000–	0.999
QTcB at diagnosis	Per QTcB	207	3.2E+10	0.001–1.1E+18	0.16
Maximum QTcB (rest)	Per maximum QTcB	207	5.7E+8	45.1–7.3E+15	0.02	0.002	8.1E-19–8-5.7E+12	0.73
Maximum QTcB (Holter)	Per maximum Holter QTcB	207	1.6E+32	7.1E+16–3.4E+47	<0.001	5.0E+26	3.4E+7–7.3E+45	0.006
Follow-up period	Per year	207	1.30	1.07–1.57	0.009	1.04	0.81–1.34	0.77

^ASex was dichotomized (boys=1, girls=2). ^BThe presence of a pathogenic variant was not used as an independent variable in the multivariate regression analysis because the number of patients who were tested was limited to 124. In the multivariate regression analysis with 124 patients, the presence of a pathogenic variant was not a significant predictor of the overall appearance of cardiac events (data not shown). ^CDiagnosis by screening or clinically was dichotomized (screened group=1, clinical group=2). Other abbreviations as in Tables 1,2.

To determine which QTc value more effectively predicted the appearance of cardiac events, ROC curves were created for maximum Holter QTcB and maximum Holter QTcF values. For this analysis, the maximum resting QTc values were also included (Figure 2) because Holter ECGs were not routinely tested. The ROC curves to predict the appearance of cardiac events after the visit to the National Hospital Organization Kagoshima Medical Center are shown in Figure 2A. The AUCs of the maximum Holter QTc values were significantly superior to those of the maximum resting QTc values. The ROC curves to predict the overall presence of cardiac events are shown in Figure 2B. The AUCs of the maximum Holter QTc values were superior to those of the maximum resting QTc values, except between the maximum Holter QTcF and the maximum resting QTcF (P=0.055).

Figure 2.

Receiver operating characteristic curves of 4 QTc values for predicting (A) the appearance of cardiac events after the first hospital visit and (B) the presence of overall cardiac events. Lines represent the maximum (Max) Holter QT interval corrected by the Fridericia formula (QTcF; a), Max Holter QT interval corrected by the Bazett formula (QTcB; b), Max resting QTcF (c), and Max resting QTcB (d). (A) The areas under the curve (AUCs) and 95% confidence intervals (CIs) for (a), (b), (c), and (d) were 0.942 (0.888–0.996), 0.938 (0.858–1.017), 0.770 (0.635–0.905), and 0.646 (0.426–0.866), respectively. P values for differences in AUCs were as follows: P=0.86 for (a) vs. (b); P=0.009 for (a) vs. (c); P=0.003 for (a) vs. (d); P=0.01 for (b) vs. (c); P=0.002 for (b) vs. (d); and P=0.049 for (c) vs. (d). (B) The AUCs and 95% CIs for (a), (b), (c), and (d) were 0.910 (0.829–0.990), 0.928 (0.856–0.999), 0.790 (0.675–0.905), and 0.653 (0.455–0.851), respectively. P values for differences in AUCs were as follows: P=0.52 for (a) vs. (b); P=0.06 for (a) vs. (c); P=0.003 for (a) vs. (d); P=0.01 for (b) vs. (c); P=0.001 for (b) vs. (d); and P=0.02 for (c) vs. (d).

The best cut-off values for maximum resting QTcB, maximum resting QTcF, maximum Holter QTcB, and maximum Holter QTcF to predict the appearance of events after the first hospital visit as assessed by Youden’s index and by the shortest distance (in parentheses) were ≥500 (≥490), ≥470 (≥480), ≥590 (≥590), and ≥540 (≥540) ms, respectively (Table 4). The sensitivity of the best cut-offs values for maximum resting QTcB, maximum resting QTcF, maximum Holter QTcB, and maximum Holter QTcF by Youden’s index and by the shortest distance (in parentheses) was 0.500 (0.625), 1.000 (0.625), 0.875 (0.875), and 1.000 (1.000), respectively. The best cut-off values for the overall presence of cardiac events were the same as those for the appearance of events after the first hospital visit (Supplementary Table 5).

Table 4.

Statistical Analysis for Cut-Off Values to Predict the Appearance of Cardiac Events After the First Hospital Visit

Cut-off (ms)	Sensitivity	Specificity	PPV	NPV	Youden’s index	Distance
a. Maximum resting QTcB value
≥470	1.000	0.176	0.047	1.000	1.176	0.824
≥480	0.750	0.317	0.042	0.969	1.067	0.728
≥490	0.625	0.628	0.063	0.977	1.253	0.528*
≥500	0.500	0.759	0.077	0.974	1.259*	0.555
≥510	0.375	0.819	0.077	0.970	1.194	0.651
≥520	0.250	0.915	0.105	0.968	1.165	0.755
≥530	0.125	0.930	0.067	0.964	1.055	0.878
b. Maximum resting QTcF value
≥450	1.000	0.181	0.047	1.000	1.181	0.819
≥460	1.000	0.322	0.056	1.000	1.322	0.678
≥470	1.000	0.518	0.077	1.000	1.518*	0.482
≥480	0.625	0.698	0.077	0.979	1.323	0.481*
≥490	0.500	0.779	0.083	0.975	1.279	0.547
≥500	0.375	0.879	0.111	0.972	1.254	0.637
≥510	0.375	0.920	0.158	0.973	1.295	0.630
c. Maximum Holter QTcB value
≥550	1.000	0.508	0.075	1.000	1.508	0.492
≥560	1.000	0.618	0.095	1.000	1.618	0.382
≥570	0.875	0.724	0.113	0.993	1.599	0.303
≥580	0.875	0.799	0.149	0.994	1.674	0.237
≥590	0.875	0.879	0.226	0.994	1.754*	0.174*
≥600	0.750	0.945	0.353	0.989	1.695	0.256
≥610	0.500	0.980	0.500	0.980	1.480	0.500
d. Maximum Holter QTcF value
≥510	1.000	0.467	0.070	1.000	1.467	0.533
≥520	1.000	0.613	0.094	1.000	1.613	0.387
≥530	1.000	0.739	0.133	1.000	1.739	0.261
≥540	1.000	0.814	0.178	1.000	1.814*	0.186*
≥550	0.750	0.859	0.176	0.988	1.609	0.287
≥560	0.625	0.925	0.250	0.984	1.550	0.383
≥570	0.625	0.970	0.455	0.985	1.595	0.376

*Best values of Youden’s index and the distance in each QTc values. Youden’s index denotes (sensitivity+specificity−1); the largest value is the best value. Distance denotes the shortest distance between the point (0,1) and the receiver operating characteristic (ROC) plot, that minimizes ([1−sensitivity]²+[1−specificity]²)^0.5; the smallest value is the best value. NPV, negative predictive value; PPV, positive predictive value. Other abbreviations as in Table 1.

Discussion

The present study showed that the prevalence of cardiac events after the first hospital visit was 4%. Multiple logistic regression analysis revealed that the maximum Holter QTc value was the sole predictor of the appearance of cardiac events after the first hospital visit (P=0.02). Using the maximum Holter QTc value is appropriate to predict the appearance of symptoms in pediatric patients with LQTS.

Risk of Cardiac Events After the First Hospital Visit

Relatively recent reports have shown that the risk for LQTS-related cardiac events is high at about or more than a few dozen percent.^{6–14,16–18} A possible reason for the high risk in these studies may be that the data included cardiac events before the first hospital visit in areas where the screening program was not present. Rohatgi et al reported that 27.4% of patients of their entire cohort were symptomatic at the first visit and that the risk of the appearance of cardiac events after this visit in the symptomatic patients was 25.2% (42/166 patients), compared with a risk of 1.8% (8/440 patients) for patients without symptoms at their first visit (P<0.001).⁵ In the present study, the prevalence of cardiac events after the first hospital visit was 50% (2/4) in previously symptomatic patients and 3.0% (6/203) in asymptomatic patients (P=0.008).These data suggest that if patients begin follow-up while asymptomatic, as in the era of ECG screening programs, if they are given enough information about strategies such as sophisticated preventive procedures¹⁵ or precision medicine,²⁸ and if information is sufficiently conveyed to patients and their families, we would be able to decrease the risk for cardiac events to a reasonably low level. We should then consider using QTc values based on Holter recordings to predict the appearance of cardiac events after the first hospital visit.

QTc Values on Holter ECG Recordings

QTc values on resting 12-lead ECGs are generally used to predict the appearance of LQTS-related cardiac events.²¹ QTc values on Holter recordings are well known to be quite longer than those on resting 12-lead ECGs.^23–25,29 Holter recordings provide around 100,000 QT/RR pairs per day. Therefore, automated or semiautomated measurements of QT intervals have been adopted to differentiate the data of patients with LQTS from those of normal or control individuals.^18,23,29,30 However, data obtained by automated QT measurement do not seem to be applicable for clinical use because QT intervals at higher or lower heart rates^18,24 or too-long or too-short QT intervals²⁴ are automatically discarded by the software. In our experience, automatic measurement of too-long or too-short QTc values must be extensively checked, particularly when the isoelectronic baseline fluctuates or biphasic T waves, low-voltage T waves, and artifacts are present. In addition, many studies used 1 Holter recording per patient to evaluate the QTc values of patients with LQTS. We should use the longest QTc values among repeated Holter ECGs, as for repeated resting 12-lead ECGs.²¹ Therefore, in the present study we: (1) manually measured Holter recordings;²⁴ (2) chose 5 representative periods in a day and obtained 3 ECGs (maximum, average, and minimum heart rates) from each period; and (3) recorded Holter ECGs at every patient’s visit. Finally, we used the maximum Holter QTc value throughout the day and among repeated Holter recordings.

Of the 3 QTc values (i.e., resting QTc at diagnosis, maximum resting QTc, and maximum Holter QTc), the maximum Holter QTc value obtained using our methodology was the sole predictor of the appearance of cardiac events after the first hospital visit (Table 2; Supplementary Table 3). The maximum Holter QTc was one of the independent predictors of the overall presence of cardiac events (Table 3; Supplementary Table 4). We believe that the maximum QTc on Holter ECGs should be considered for the prediction of cardiac events in patients with LQTS; however, no reports have addressed the maximum Holter QTc values among repeated recordings of Holter ECGs. Future studies should be performed to validate the present data to determine whether the maximum Holter QTc is independently predictive of the appearance of new cardiac symptoms for all symptomatic patients at diagnosis or for all patients with pathogenic variants. The cut-off values in this study should also be validated.

Genetic Testing

The yield of genetic testing in the present study (33%) was much lower than in previous studies (around 70%¹² or 80%³¹). The genetically determined prevalence of LQTS is at least 1 in 2,534 among apparently healthy live births (i.e., in the general population).³² The probability of diagnosing LQTS through ECGs in the general population is 1 in 988 among 7th graders (aged 12 years) according to the criteria of the HRS/EHRA/APHRS expert consensus statement.²⁰ Therefore, the estimated yield of genetic testing in the general population with ECG-determined LQTS is 39% (988 divided by 2,534). In a large study involving 44,596 otherwise healthy infants in Italy, the prevalence of disease-causing variants in infants with a QTcB of >470 ms was 43% (12/28 infants).³² These data suggest that the yield of genetic testing in patients from the general population who were diagnosed through an ECG is around 40%.

Study Limitations

This study had 3 main limitations. First, many patients were asymptomatic at their first visit and more than half did not have pathogenic variants, although one of the aims of this study was to clarify the outcome of those patients. The statistical analysis of a resting QTc of ≥470 ms, which is the recommended value at which medication should be started,²¹ revealed that the sensitivity was 100% but the specificity was 17.6% (Table 4). Additional studies are needed for patients in areas where proactive cardiogenetic counseling programs² or school-based ECG screening programs^19,20 exist. Second, the number of patients with cardiac events after the first hospital visit was also very small (8/207 patients [4%]). This may be an advantage of the screening program but a disadvantage for statistical robustness. An increase in the number of patients with cardiac events after the visit is not desired; however, a large number of patients with symptoms should be obtained. Finally, a gene-specific statistical analysis revealed that the maximum Holter QTc value was not a significant predictor of the appearance of cardiac events in patients with LQT1 or LQT2 (Supplementary Table 5), probably because of a small number of patients with pathogenic mutations and those with cardiac events. Further studies are also needed to investigate this issue after increasing the number of patients with each pathogenic mutation and those with cardiac events.

Conclusions

The present study showed that the prevalence of cardiac events after the first hospital visit was 4%. Multivariate logistic regression analysis revealed that the maximum Holter QTc was the sole predictive factor for the appearance of cardiac events after the first hospital visit. These data indicate that the maximum Holter QTc should be considered when predicting the appearance of cardiac events, although the data should be validated in areas where there are no screening programs.

Acknowledgment

The authors thank Angela Morben, DVM, ELS, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Sources of Funding

This work was supported, in part, by Health and Labour Sciences Research Grants from the Japanese Ministry of Health, Labour and Welfare (Comprehensive Research on Cardiovascular Diseases; H22-032, H26-002, H29-055, 21FC1004).

Disclosures

The authors declare that there is no conflict of interest.

IRB Information

This study was approved by the Ethics Committee of the National Hospital Organization Kagoshima Medical Center (No. 30-69).

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-23-0409

References

• 1.   Schwartz PJ, Crotti L, Insolia R. Long QT syndrome: From genetics to management. Circ Arrhythm Electrophysiol 2012; 5: 868–877.
• 2.   Schwartz PJ, Ackerman MJ, Antzelevitch C, Bezzina CR, Borggrefe M, Cuneo BF, et al. Inherited cardiac arrhythmias. Nat Rev Dis Primers 2020; 6: 58.
• 3.   Wilde AAM, Amin AS, Postema PG. Diagnosis, management and therapeutic strategies for congenital long QT syndrome. Heart 2022; 108: 332–338.
• 4.   Jervell A, Lange-Nielsen F. Congenital deaf-mutism, functional heart disease with prolongation of the Q-T interval and sudden death. Am Heart J 1957; 54: 59–68.
• 5.   Rohatgi RK, Sugrue A, Bos JM, Cannon BC, Asirvatham SJ, Moir C, et al. Contemporary outcomes in patients with long QT syndrome. J Am Coll Cardiol 2017; 70: 453–462.
• 6.   Ozawa J, Ohno S, Hisamatsu T, Itoh H, Makiyama T, Suzuki H, et al. Pediatric cohort with long QT syndrome: KCNH2 mutation carriers present late onset but severe symptoms. Circ J 2016; 80: 696–702.
• 7.   Mazzanti A, Maragna R, Vacanti G, Monteforte N, Bloise R, Marino M, et al. Interplay between genetic substrate, QTc duration, and arrhythmia risk in patients with long QT syndrome. J Am Coll Cardiol 2018; 71: 1663–1671.
• 8.   Kutyifa V, Daimee UA, McNitt S, Polonsky B, Lowenstein C, Cutter K, et al. Clinical aspects of the three major genetic forms of long QT syndrome (LQT1, LQT2, LQT3). Ann Noninvasive Electrocardiol 2018; 23: e12537.
• 9.   Kwok SY, Liu AP, Chan CY, Lun KS, Fung JL, Mak CC, et al. Clinical and genetic profile of congenital long QT syndrome in Hong Kong: A 20-year experience in paediatrics. Hong Kong Med J 2018; 24: 561–570.
• 10.   Lahrouchi N, Tadros R, Crotti L, Mizusawa Y, Postema PG, Beekman L, et al. Transethnic genome-wide association study provides insights in the genetic architecture and heritability of long QT syndrome. Circulation 2020; 142: 324–338.
• 11.   Younis A, Aktas MK, Rosero S, Kutyifa V, Polonsky B, McNitt S, et al. Outcome by sex in patients with long QT syndrome with an implantable cardioverter defibrillator. J Am Heart Assoc 2020; 9: e016398.
• 12.   Ohno S, Ozawa J, Fukuyama M, Makiyama T, Horie M. An NGS-based genotyping in LQTS; minor genes are no longer minor. J Hum Genet 2020; 65: 1083–1091.
• 13.   Tse G, Lee S, Zhou J, Liu T, Wong ICK, Mak C, et al. Territory-wide Chinese cohort of long QT syndrome: Random survival forest and Cox analyses. Front Cardiovasc Med 2021; 8: 608592.
• 14.   Ergül Y, Tunca Şahin G, Kafalı HC, Öztürk E, Özgür S, Haydin S, et al. Clinical and genetic characteristics and course of congenital long QT syndrome in children: A nine-year single-center experience. Anatol J Cardiol 2021; 25: 250–257.
• 15.   Tobert KE, Bos JM, Garmany R, Ackerman MJ. Return-to-play for athletes with long QT syndrome or genetic heart diseases predisposing to sudden death. J Am Coll Cardiol 2021; 78: 594–604.
• 16.   Goldenberg I, Bos JM, Yoruk A, Chen AY, Lopes C, Huang DT, et al. Risk prediction in women with congenital long QT syndrome. J Am Heart Assoc 2021; 10: e021088.
• 17.   Mazzanti A, Trancuccio A, Kukavica D, Pagan E, Wang M, Mohsin M, et al. Independent validation and clinical implications of the risk prediction model for long QT syndrome (1-2-3-LQTS-Risk). Europace 2022; 24: 614–619.
• 18.   Waddell-Smith KE, Chaptynova AA, Li J, Crawford JR, Hinds H, Skinner JR. Holter recordings at initial assessment for long QT syndrome: Relationship to genotype status and cardiac events. J Cardiovasc Dev Dis 2022; 9: 164.
• 19.   Fukuyama M, Horie M, Aoki H, Ozawa J, Kato K, Sawayama Y, et al. School-based routine screenings of electrocardiograms for the diagnosis of long QT syndrome. Europace 2022; 24: 1496–1503.
• 20.   Yoshinaga M, Kucho Y, Nishibatake M, Ogata H, Nomura Y. Probability of diagnosing long QT syndrome in children and adolescents according to the criteria of the HRS/EHRA/APHRS expert consensus statement. Eur Heart J 2016; 37: 2490–2497.
• 21.   Priori SG, Wilde AA, Horie M, Cho Y, Behr ER, Berul C, et al. Executive summary: HRS/EHRA/APHRS expert consensus statement on the diagnosis and management of patients with inherited primary arrhythmia syndromes. Europace 2013; 15: 1389–1406.
• 22.   Zeppenfeld K, Tfelt-Hansen J, de Riva M, Winkel BG, Behr ER, Blom NA, et al. 2022 ESC guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Eur Heart J 2022; 43: 3997–4126.
• 23.   Priori SG, Schwartz PJ, Napolitano C, Bloise R, Ronchetti E, Grillo M, et al. Risk stratification in the long-QT syndrome. N Engl J Med 2003; 348: 1866–1874.
• 24.   Page A, Aktas MK, Soyata T, Zareba W, Couderc JP. “QT clock” to improve detection of QT prolongation in long QT syndrome patients. Heart Rhythm 2016; 13: 190–198.
• 25.   Ogawa Y, Tanaka T, Kido S. Maximum QTc on Holter electrocardiography in children. Pediatr Int 2018; 60: 507–512.
• 26.   Tamr Agha MK, Fakhri G, Ahmed M, El Sedawy O, Abi Saleh B, Bitar F, et al. QTc interval on 24-hour holter monitor: To trust or not to trust? Ann Noninvasive Electrocardiol 2022; 27: e12899.
• 27.   Adler A, Novelli V, Amin AS, Abiusi E, Care M, Nannenberg EA, et al. An international, multicentered, evidence-based reappraisal of genes reported to cause congenital long QT syndrome. Circulation 2020; 141: 418–428.
• 28.   Gnecchi M, Sala L, Schwartz PJ. Precision medicine and cardiac channelopathies: When dreams meet reality. Eur Heart J 2021; 42: 1661–1675.
• 29.   Viitasalo M, Oikarinen L, Väänänen H, Swan H, Piippo K, Kontula K, et al. Differentiation between LQT1 and LQT2 patients and unaffected subjects using 24-hour electrocardiographic recordings. Am J Cardiol 2002; 89: 679–685.
• 30.   Follansbee CW, Beerman L, Arora G. Automated QT analysis on Holter monitors in pediatric patients can differentiate long QT syndrome from controls. Pacing Clin Electrophysiol 2018; 41: 50–56.
• 31.   Taggart NW, Haglund CM, Tester DJ, Ackerman MJ. Diagnostic miscues in congenital long-QT syndrome. Circulation 2007; 115: 2613–2620.
• 32.   Schwartz PJ, Stramba-Badiale M, Crotti L, Pedrazzini M, Besana A, Bosi G, et al. Prevalence of the congenital long-QT syndrome. Circulation 2009; 120: 1761–1767.