Beyond symptoms: Rhythm surveillance in congenital heart disease

Abstract

Background

Patients with congenital heart disease (CHD) face a lifelong risk of arrhythmias, many of which remain asymptomatic and undetected.

Objective

This study aimed to evaluate the diagnostic yield of different rhythm surveillance strategies, assess symptom–arrhythmia correlation, and determine the clinical value of arrhythmia screening in patients with CHD.

Methods

We retrospectively analyzed 552 patients with CHD (66.1% female; median age 34 years) who underwent 7-day patch rhythm monitoring at a tertiary center between December 2022 and January 2025. Patients were grouped as symptomatic or asymptomatic (screening). The patch device also functioned as an event recorder, with the first 24 hours analyzed as standard Holter monitoring.

Results

The symptomatic group consisted of 349 patients, and the screening group consisted of 203 patients. Any arrhythmias were detected in 64.8% of symptomatic and 68.5% of screening group (P = .410), with significant arrhythmias observed in 16.9% and 19.2%, respectively (P = .494). In total, the detection rate was highest with 7-day monitoring (66.1%) compared with Holter (46.9%) or event recorder (17.3%). Significant arrhythmia prevalence increased with CHD complexity and age, exceeding 15% in the fourth decades of life. Symptom–arrhythmia correlation was weak and reached statistical significance only in the asymptomatic group (P = .05). The treatment change rate was 9.4% with 7-day monitoring and was comparable between symptomatic and screening groups.

Conclusion

Arrhythmias in CHD are common, age dependent, and poorly predicted by symptoms. Prolonged monitoring enhances detection and guides management, supporting risk-stratified surveillance based on age and disease complexity.

Key Findings.

• ▪Arrhythmias are highly prevalent in patients with congenital heart disease (CHD) and are not reliably predicted by symptoms, with comparable rates of clinically significant arrhythmia in symptomatic and asymptomatic individuals.
• ▪Extended 7-day rhythm monitoring improves arrhythmia detection beyond conventional 24-hour Holter or event-based monitoring and frequently results in clinically meaningful management changes.
• ▪Arrhythmia burden increases with age and CHD complexity, supporting a risk-stratified rhythm surveillance strategy guided by age and disease complexity rather than symptom status alone.

Introduction

Congenital heart disease (CHD) affects approximately 0.8% of live births, and survival into adulthood has markedly improved with advances in surgical and medical management.¹ As this population ages, arrhythmias, including both tachyarrhythmias and bradyarrhythmias, represent frequent and clinically significant late complications.² These rhythm disturbances may arise from structural abnormality, progressive myocardial disease, or postoperative sequelae.^3,4 The prevalence and burden of arrhythmias increase with age and are associated with worsening heart failure and adverse hemodynamic consequences.^5,6

Ambulatory electrocardiographic (ECG) monitoring is commonly used for early detection of arrhythmias in this population.7, 8, 9 Although routine rhythm surveillance is recommended for certain high-risk CHD subtypes in adult patients irrespective of symptoms, the optimal application of surveillance strategies for other subtypes of CHD and across different age groups remains less well defined. Therefore, clarifying the diagnostic yield and clinical implications of arrhythmia monitoring across patient groups is of critical importance for optimizing management strategies in CHD.

This study aimed to evaluate the clinical implications of different rhythm surveillance strategies in patients with CHD, with a particular focus on the role of arrhythmia screening in asymptomatic individuals. The primary objective was to compare arrhythmia detection between symptomatic and asymptomatic patients during the same monitoring period. The secondary objective was to assess whether the identification of arrhythmias led to subsequent changes in clinical management.

Methods

Study design and patient population

This single-center, retrospective cohort study included patients with CHD who underwent 7-day ECG monitoring at a tertiary referral center between December 2022 and January 2025. Patients were enrolled either for preexisting symptoms suggestive of arrhythmia (eg, syncope, palpitations, chest pain, dyspnea, or dizziness) or through a prospective cohort study designed for routine arrhythmia screening in patients with CHD. The cohort study comprised 313 patients with CHD, including those with septal defects and tetralogy of Fallot, recruited between December 2022 and January 2025. For analysis, patients were categorized into 2 groups according to baseline symptom status before monitoring. Patients with a history of symptoms at enrollment, regardless of enrollment pathway, were assigned to the symptomatic group, whereas patients without previous symptoms who were enrolled for routine screening were assigned to the screening group.

The conduct and reporting of this retrospective cohort study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines and received previous approval from the human research committee of National Taiwan University Hospital. Owing to the retrospective nature of the study, an informed consent was waived. Clinical characteristics, cardiac catheterization reports, echocardiographic and surgical findings, symptoms, and cardiac implantable electronic device data were collected via chart review.

CHD classification

CHD diagnoses were confirmed using International Classification of Diseases, 10th Revision, Clinical Modification codes Q20–Q26. Anatomic complexity was categorized as simple, moderate, or severe according to the Bethesda classification.^10,11

Arrhythmia detection tools

All patients underwent 7-day ambulatory ECG monitoring with the QOCA portable ECG monitoring device (Quanta Computer Inc, Taoyuan City, Taiwan), a wearable, single-lead device capable of continuous recording for up to 14 days without recharging. Patients were instructed to press a device button and complete a symptom diary whenever symptoms occurred; these entries were considered event-recorder events. The first 24 hours of recording were analyzed separately and designated as the 24-hour Holter result.

Using a single continuous 7-day monitoring platform as a reference framework, we simulated and compared commonly applied rhythm surveillance strategies to identify the optimal approach. These strategies included short-duration continuous monitoring (first 24 hours), symptom-triggered event recording, and extended continuous monitoring, all evaluated within the same observation period. This study method was chosen to minimize the impact of day-to-day arrhythmia variability and allow comparison under identical physiological conditions.

Arrhythmia definition

Arrhythmias were defined as follows: (1) ectopic beats, atrial and ventricular ectopic beats; (2) supraventricular tachycardia (SVT), atrial fibrillation or flutter, paroxysmal SVT, and atrial tachycardia (consecutive atrial premature beats for more than 3 beats); (3) ventricular tachycardia (VT), 3 or more consecutive beats of ventricular ectopic beats with a rapid heart rate according to the age; and (4) bradycardia, sinus pause of >3 seconds, Mobitz II second-degree atrioventricular block, and third-degree atrioventricular block.

Arrhythmias were categorized as any arrhythmia (including ectopic beats, SVT, VT, and bradycardia) and significant arrhythmia, defined as sustained SVT, VT, or bradycardia. Sustained SVT was defined as a heart rate of more than 150 beats per minute for more than 30 seconds. Short runs of atrial tachycardia (≥3 consecutive atrial beats) were recorded for descriptive purposes but were not classified as significant arrhythmias unless they met the criteria for sustained SVT. Sinus pauses were considered significant only if they occurred during wakefulness or were associated with symptoms.

Statistical analysis

Descriptive statistics were used to summarize baseline characteristics, reported as counts with percentages, means ± standard deviations, or medians with interquartile ranges, as appropriate. Between-group comparisons were performed using the Student t test for continuous variables and χ² tests for categorical variables. Risk factors for significant arrhythmia were evaluated using χ² tests and logistic regression. A 2-tailed P < .05 was considered statistically significant. Analyses were conducted using SPSS version 30.0 (IBM Corp, Armonk, NY).

Result

Basic demographics between groups

Baseline characteristics of the study cohort are presented in Table 1. A total of 552 patients were enrolled, including 349 in the symptomatic group and 203 in the screening group. The median age was 34 years (range 9–80 years), and 66.1% were female. Patients in the screening group were significantly older than those in the symptomatic group (36.0 [28.0–49.0] vs 30.0 [21.0–43.0] years; P < .001). Female sex predominated in both groups. The distribution of CHD severity differed between groups, with the screening group having a higher proportion of simple CHD and no patient classified as severe CHD. Detailed anatomic classification is presented in Supplemental Table 1.

Table 1.

Baseline characteristics and arrhythmia outcomes between groups

Clinical characteristics	Total (n = 552)	Symptomatic group (n = 349)	Screening group (n = 203)	P value
Age, y	34.0 [24.0–45.0]	30.0 [21.0–43.0]	36.0 [28.0–49.0]	<.001
Women (%)	365 (66.1)	235 (67.3)	130 (64.0)	.437
CHD severity
Simple	305 (55.3)	176 (50.4)	129 (63.5)	.004
Moderate	225 (40.8)	151 (43.3)	74 (36.5)	.1386
Severe	22 (4.0)	22 (6.3)	0 (0)	<.001
Previous therapy	59 (10.7)	42 (12)	17 (8.4)	.196
During the study
Symptoms	213 (38.6)	153 (43.8)	60 (29.6)	.001
Any arrhythmia	365 (66.1)	226 (64.8)	139 (68.5)	.410
Significant arrhythmia	98 (17.8)	59 (16.9)	39 (19.2)	.494
Arrhythmia type
Ectopic beat	225 (57.7)	136 (39.0)	89 (43.8)	.301
SVT	111 (28.5)	74 (21.2)	37 (18.2)	.465
VT	41 (10.5)	23 (6.6)	18 (8.9)	.415
Bradycardia	13 (3.3)	7 (2.0)	7 (3.4)	.448

Values are presented as counts (percentages) and medians [interquartile ranges]. Multiple significant arrhythmias were diagnosed in 14 patients in the symptomatic group and 12 patients in the screening group.

CHD = congenital heart disease; SVT = supraventricular tachycardia; VT = ventricular tachycardia.

Arrhythmia burden

Any arrhythmias were identified in 365 patients (66.1%) using 7-day patch monitoring. The most frequent arrhythmia was ectopic beats (including premature ventricular contractions and premature atrial contractions), detected in 225 patients (40.7%), followed by SVT (111 patients; 20.1%), VT (41 patients; 7.4%), and bradycardia (4 patients; 0.7%). In addition, 26 patients (4.7% of the total cohort) experienced more than 1 arrhythmia. Significant arrhythmias were detected in 98 patients (17.8%), with an age-dependent increase in detection rate. In the fourth decade of life, significant arrhythmias were present in 18.3% of the symptomatic group and 11.9% of the screening group, with atrial arrhythmias in 13.3% and 10.2%, respectively (Figure 1).

Figure 1.

Age-stratified detection rates (%) of significant and atrial arrhythmias across groups. Atrial arrhythmia detection rates (%) (left) and significant arrhythmia detection rates (right) in symptomatic and screening groups. Detection rates are presented as percentages within each age stratum. Rates increased markedly with age in both groups—from 18.3% and 11.9% in the 30–39-year age group to 48.0% and 34.6% in those aged 60 years and older for significant arrhythmias and from 13.3% and 10.2% to 76% and 46.2% for atrial arrhythmias in the symptomatic and screening groups, respectively.

Arrhythmia detection rates varied by CHD severity (Table 2), with higher rates in moderate and severe CHD than in simple CHD. When moderate and severe CHD were analyzed together, ventricular arrhythmias were significantly more frequent than in simple CHD (10.1% vs 5.2%; P = .003).

Table 2.

Arrhythmia across CHD severity

Arrhythmia	Simple (n = 305)	Moderate (n = 225)	Severe (n = 22)	P value
Any arrhythmia	171 (56.1)	178 (79.1)	16 (72.7)	<.001
Significant arrhythmia	45 (14.8)	48 (21.3)	5 (22.7)	.121
Arrhythmia type
Ectopic beat	100 (32.8)	115 (51.1)	10 (45.5)	<.001
SVT	59 (10.3)	47 (20.9)	5 (22.7)	.001
VT	16 (5.2)	23 (10.2)	2 (9.1)	.080
Bradycardia	8 (2.6)	6 (2.7)	0 (0.0)	.615

Values are presented as counts (percentages). Multiple significant arrhythmias were diagnosed in 12 patients in those with mild CHD, 13 in moderate CHD, and 1 in severe CHD.

CHD = congenital heart disease; SVT = supraventricular tachycardia; VT = ventricular tachycardia.

In subgroup analyses, any arrhythmia detection rates did not differ significantly between symptomatic and screening groups (64.8% vs 68.5%; P = .374), nor did the detection rates of significant arrhythmias (16.9% vs 19.2%; P = .494). The distribution of arrhythmia types was also similar, with ectopic beats being the most common in both groups (39.0% vs 43.8%), followed by SVT (21.2% vs 18.2%), VT (6.6% vs 8.9%), and bradycardia (2.0% vs 3.4%).

Risk factors for significant arrhythmia

To define the risk of significant arrhythmia in this cohort, we performed logistic regression analysis. In univariate analysis, age, CHD severity, and previous treatment were significantly associated with significant arrhythmias (P < .001, P = .048, and P < .001, respectively). In multivariate analysis, age, CHD severity, and previous treatment remained significantly associated with significant arrhythmia (P < .001, P = .003, and P < .001, respectively) (Table 3). The indication of 7-day patch monitoring (symptomatic vs screening group) was not associated with significant arrhythmia in either univariate or multivariate regression analyses.

Table 3.

Univariate and multivariate analysis of risk factors for significant arrhythmia

Variables	Univariate analysis		Multivariate analysis
	Odds ratio (95% CI)	P value	Odds ratio (95% CI)	P value
Age, y	1.034 (1.020–1.048)	<.001	1.040 (1.024–1.056)	<.001
Gender	1.517 (0.970–2.373)	.068
Group	0.856 (0.547–1.338)	.494
CHD grade	1.450 (1.003–2.095)	.048		.003
Moderate: simple			2.263 (1.345–3.805)	.002
Severe: simple			4.045 (1.286–12.731)	.017
Previous treatment	4.230 (2.385–7.504)	<.001	3.096 (1.672–5.732)	<.001

CHD = congenital heart disease; CI = confidence interval.

Comparison of arrhythmia detection tools and their impact on management

Arrhythmia detection rates differed significantly across modalities: 66.1% with 7-day patch monitoring, 46.9% with 24-hour Holter, and 17.3% with event recording. The patch device detected 90% of arrhythmias by day 5. Among 259 patients with any arrhythmias identified on Holter, an additional arrhythmia type was detected in 60 (23.2%) during extended monitoring, leading to treatment changes in 24 patients (40%) (Figure 2).

Figure 2.

Diagnostic yield and management impact of different monitoring modalities. Among 552 patients with congenital heart disease, 7-day patch monitoring achieved the highest arrhythmia detection and treatment modification rates compared with 24-hour Holter and event-recorder monitoring. Extended monitoring after Holter identified additional arrhythmias that led to further management changes, including medication adjustment, ablation, or device implantation. CIED = cardiac implantable electronic device.

Overall, management changes occurred in 9.4% of cases after any arrhythmia detection with the 7-day patch compared with 7.6% after Holter and 3.9% after event recording (Figure 2). The treatment change rates were higher in the symptomatic group than the screening group without reaching statistical significance (11.5% vs 6.4%; P = .052). Importantly, management changes were predominantly driven by clinically significant arrhythmias, particularly sustained SVT and VT, and the proportion of patients requiring treatment modification among those with significant arrhythmias was also not significantly different between symptomatic and screening groups (39.0% vs 28.2%; P = .273). Patients with VT and SVT were more likely to undergo treatment modification than those with ectopic beats or bradycardia (53.7% and 27.9% vs 4.4% and 7.1%, respectively) (Figure 3). Among 41 patients with VT, 31 had clinically significant episodes and 8 were identified solely with ventricular triplets and 2 with slow ventricular runs and were therefore excluded, of whom 20 (64.5%) underwent management changes: 18 received new medications, 3 underwent ablation, and 1 received implantable cardioverter-defibrillator implantation.

Figure 3.

Treatment changes by arrhythmia types. Left: Treatment change rates (%) for patients with any arrhythmia and significant arrhythmia in the symptomatic and screening groups. Treatment change rates were higher in the symptomatic group but did not reach statistical significance. Right: Proportion (%) of patients who underwent medication adjustment, catheter ablation, or CIED implantation according to arrhythmia types. Treatment modifications were most frequent in patients with VT, followed by SVT, whereas changes were uncommon in those with ectopic beats or bradycardia. CIED = cardiac implantable electronic devices; SVT = supraventricular tachycardia; VT = ventricular tachycardia.

Symptoms and arrhythmia correlation

The detection rate of any arrhythmia was comparable between groups (64.8% symptomatic vs 68.5% screening; P = .374), as were the detection rates of significant arrhythmias (16.9% vs 19.2%; P = .494). Ectopic beats and SVT were most frequently associated with palpitations (32.1% and 35.1%, respectively). VT and bradycardia were less commonly symptomatic. Notably, chest pain (27.9%) and palpitations (22.9%) were also reported by patients without documented arrhythmia.

Patients in the symptomatic group reported more symptoms than those in the screening group (43.8% vs 29.6%; P < .001). However, symptom–arrhythmia correlation was not significant in the symptomatic group (P = .266) but reached borderline significance in the screening group (P = .050). In those with significant arrhythmia, the correlation with symptoms was not significant in either group (P = .234 in the symptomatic group and P = .551 in the screening group).

Discussion

Main findings

This study demonstrates 3 key observations: (1) arrhythmias were common in patients with CHD regardless of symptoms, with comparable arrhythmia rates detected in symptomatic and asymptomatic patients; (2) prolonged monitoring improved arrhythmia detection and influenced management decisions; and (3) symptoms in patients with CHD were heterogeneous and showed limited correlation with objectively documented arrhythmias, highlighting the importance of systematic screening even in the absence of symptoms.

Arrhythmias in CHD

The detection rate of any arrhythmia was 66.1%, with significant arrhythmias in 17.8% of patients—slightly higher than previous reports.³ This may reflect the higher complexity of CHD at our tertiary center and broader definitions of tachyarrhythmias (≥3 consecutive beats). Consistent with previous reports, the arrhythmia burden increases with age,⁷ and our study shows that the burden of significant arrhythmia exceeds 15% after the age of 30 years.

Among different severities of CHD, any arrhythmia rates were higher in moderate and severe CHD than in simple CHD (79.1% and 72.7% vs 56.1%; P < .001). Although rates of significant arrhythmia did not differ, likely owing to the limited number of severe cases, the detection rates of ventricular arrhythmia were also higher in moderate and severe CHD than in simple CHD. This result was also comparable with previous studies, which showed elevated arrhythmia risk in moderate and severe CHD, such as single-ventricle physiology, tricuspid atresia, transposition of great arteries, and tetralogy of Fallot.12, 13, 14

Diagnostic yield of monitoring modalities

7-day patch monitoring provided the highest diagnostic yield (66.1%) compared with 24-hour Holter (46.9%) and event recording (17.3%). Importantly, 9.4% of patients underwent treatment modification after arrhythmia detection by 7-day patch monitoring. Extended monitoring not only improved detection but also revealed additional arrhythmias in 23.2% of patients after the first 24 hours, leading to treatment changes in 40% of these cases. These findings are consistent with previous evidence that prolonged monitoring increases diagnostic yield.^7,15, 16, 17

Although 14-day monitoring has shown even higher yields in previous studies (cumulative arrhythmia detection of 66% vs 47%),^15,17,18 our study used a 7-day protocol owing to safety and compliance considerations, particularly in pediatric patients. Previous adult studies have reported substantial dermatologic complications with longer monitoring durations, particularly beyond the first several days of patch attachment.¹⁶ Considering a considerable proportion of pediatric patients (n = 76; 13.7%) in our study, the skin irritation and compliance may be worse than in adult patients. Our study conducted a 7-day protocol, with a median wearing duration of 6 days.

Symptom–arrhythmia correlation in patients with CHD

The arrhythmia detection rate of the event recorder was 17.3% in our cohort, the lowest among the 3 diagnostic modalities evaluated. Overall, 38.6% of patients reported symptoms, with a higher frequency in the symptomatic group than the screening group (43.8% vs 29.6%; P < .001). Despite this difference, arrhythmia detection rates were comparable between the 2 groups. This finding highlights the inherent limitations of symptom-guided monitoring, given that not all symptomatic patients experienced symptoms during the monitoring period and intermittent arrhythmias may have been missed. To mitigate this limitation, continuous 7-day monitoring was used to maximize diagnostic yield. Importantly, even with this potential underdetection, arrhythmia prevalence remained similar between groups, underscoring the substantial arrhythmia burden in asymptomatic patients. Similarly, a previous pediatric study reported no significant difference in tachyarrhythmia detection between patients with and without palpitations (13% vs 12%).¹⁹ Collectively, these findings indicate a substantial mismatch between symptoms and objectively documented arrhythmias in patients with CHD.

In our analysis, symptoms were not significantly correlated with arrhythmia diagnosis. In subgroup analysis, the correlation between symptoms and arrhythmia was only significant in the screening group (P = .05) but not in the symptomatic group (P = .266). This pattern persisted after excluding pediatric patients (P = .05 in the screening group vs P = .146 in the symptomatic group). Notably, some patients enrolled for routine screening reported symptoms during the monitoring period despite the absence of previous symptoms at baseline. 1 potential explanation is that heightened symptom awareness during continuous monitoring may lead to symptom reporting that is more temporally aligned with documented arrhythmic events in the screening group. In contrast, symptoms in patients with preexisting complaints may be influenced by nonarrhythmic factors, including psychological comorbidities commonly reported in CHD populations, thereby attenuating the observed symptom–arrhythmia correlation.

Previous work has shown that symptom profiles, including palpitations, chest pain, dyspnea, and dizziness, do not differ significantly across CHD severity.^20,21 Moreover, the high prevalence of anxiety and depression in patients with CHD^22,23 may contribute to nonspecific symptoms, further weakening the correlation between subjective complaints and objective arrhythmia detection.

Asymptomatic arrhythmia and screening in CHD

A key observation is the high detection rate of arrhythmias in patients with asymptomatic CHD. Significant arrhythmias occurred in 19.2% of asymptomatic patients (screening group), similar to the symptomatic group (16.9%). Atrial arrhythmias were also frequent in asymptomatic patients (16.7%), in line with previous studies.^14,24 The treatment change rates after a significant arrhythmia diagnosis were comparable between groups. These results suggest a substantial burden of arrhythmia even in asymptomatic patients. Although earlier reports suggested that 30%–80% of arrhythmias in CHD may be asymptomatic,^8,25,26 few studies have quantified their prevalence specifically in an asymptomatic CHD cohort. Our study demonstrates that asymptomatic patients carry a burden of arrhythmia comparable with symptomatic patients, particularly in adulthood.

Age-stratified analyses showed a relatively low detection rate of significant arrhythmias in younger patients (0–19 years), consistent with contemporary data from healthy pediatric populations,²⁷ whereas the burden of significant arrhythmia increased with advancing age. These data support a risk-based approach to rhythm surveillance that incorporates age and disease complexity rather than symptom status alone.

Conclusion

Arrhythmias in CHD are common, age dependent, and not reliably predicted by symptoms. Rhythm surveillance should be risk stratified according to age and disease complexity rather than symptom status alone, with prolonged monitoring serving as a key strategy to optimize detection and guide management.

Disclosures

The authors have no conflicts of interest to disclose.