Atrial septal defect (ASD) is a common congenital heart disease with an estimated incidence of 56 per 100,000 live births (1). Children with isolated ASD are generally asymptomatic and frequently do not present with significant heart murmurs (1–3). As a result, diagnosis or referral for ASD is occasionally delayed until adulthood, when patients may present with comorbidities such as arrhythmias, heart failure, and pulmonary hypertension (4–6). Early shunt closure is safe and effective, and studies have shown that performing the procedure before the age of 25 provides a normal life expectancy (7). Therefore, it is important to develop a strategy for detecting asymptomatic ASD patients during childhood.

The 12-lead electrocardiogram (ECG) is a simple and inexpensive tool commonly used in the cardiology clinic. ECG findings in ASD include right bundle branch block (iRBBB), rsR' type of iRBBB, right axis deviation, isolated negative T waves, and notches in the lower leads (II, III, aVF) (3, 8–10). In Japan, school ECG mass screening for cardiovascular diseases has been conducted since 1995, when this system was legislated (8). With the use of the guidelines for systematically interpreting ECG, this system has an impact on the early detection of substrates for sudden cardiac death, long QT syndrome (11), myocardial diseases (12) and idiopathic pulmonary arterial hypertension (13) at the community level. Although ASD cases are occasionally found by school ECG screening (14, 15), the impact of this healthcare system on detecting ASD in childhood has not been investigated.

In the present study, we investigated the impact of school ECG-based screening system in detecting children with hemodynamically significant ASD at a tertiary referral center in Japan.

A total of 116 patients (53 boys and 63 girls) with ASD were included in the study, in which the median age at diagnosis and at catheterization was 3.0 [interquartile range (IQR), 0.12–6.0] years and 8.9 (IQR, 5.9–12.8) years, respectively. Forty-three patients (37%) were detected through ECG screening (Screening group), while the remaining 73 patients were diagnosed through other means (Non-screening group) (Figure 1). The Non-screening group included 31 patients (27%) diagnosed in infant or preschool health check-ups, 21 patients (18%) diagnosed by cardiovascular findings in outpatient clinics, and 21 patients (18%) diagnosed incidentally during echocardiograms obtained for other reasons. Forty-one (95%) of the 43 patients in the Screening group were picked up by ASD-related ECG abnormalities. The other 2 patients in the Screening group were diagnosed with ASD after a heart murmur was noted during school examinations. In one of them, an ECG at the time of catheterization showed IRBBB. The other demographic and clinical characteristics were shown in Figure 1 and Table 2.

The age distribution at diagnosis showed that the largest number of patients were diagnosed before one year of age, and most of patients aged six or older were diagnosed at the age of school ECG screening (Figure 2). Out of a total of 116 patients, 67 (58%) were diagnosed during preschool age and 49 (42%) were diagnosed at six years of age or older. Among the patients diagnosed at six years of age or older, 43 out of 49 (88%) were identified through school ECG screening. In the Non-screening group, only two out of the 6 patients diagnosed at six years of age or older were prompted by heart murmur (Supplementary Table S2). Events prompting cardiovascular examination by age range is shown in Supplementary Table S3.

In overall patients, heart murmur (Levine 2/6 or higher) was found in 58.4% of patients at diagnosis and 40.4% at catheterization. The percentage of patients with significant heart murmurs was lower in the Screening group than in the Non-screening group (36.3% vs. 69.1% at diagnosis, p < 0.01, Figure 3A; 25.6% vs. 50% at catheterization, p < 0.05, Figure 3B). The percentage of patients with comorbidities (e.g., genetic disorders) tended to be higher in the Non-screening group than in the Screening group, although this was not statistically significant (Figure 3C). The auscultatory findings by age group were shown in Supplementary Table S4. The mean pulmonary artery pressure (mPAP), pulmonary-to-systemic flow ratio (Qp/Qs), and pulmonary vascular resistance index (PVRi) were comparable between the Screening and Non-screening groups in overall patients (Figures 3D–F) and in patients without comorbidities such as Down syndrome (Supplementary Figures S1A–C). The proportion of patients with clinically relevant ECG findings at catheterization, including rsR' type of iRBBB (lead V1 or V2), inverted T in V4 (invTV4) and ST depression in aVF, as well as right axis deviation, iRBBB and crochetage pattern, were comparable in both groups (Table 3). The patients with a composite ECG parameter (rsR' type of IRBBB, invT4 or ST depression in aVF) accounted for 79% in Screening group, which was higher than in Non-screening group (62%, p = 0.02) (Table 3). Eight patients (21%) with the negative composite ECG parameter at catheterization in the screening group were prompted by heart murmur (n = 1), PAC (n = 1), non-rsR’ type of iRBBB (n = 4), and 2 patients with negative invTV4 at catheterization but with positive invTV4 at school ECG. Of note, although ECG abnormality was typically rsR' type of iRBBB, some patients had invTV4 or ST depression in aVF without IRBBB (Supplementary Figures 2A,B). These ECG findings were correlated with hemodynamic parameters. Patients with rsR' type of iRBBB had higher Qp/Qs than those without iRBBB (rsR') (2.3 vs. 1.7, p < 0.01) and patients with invTV4 had higher Qp/Qs than those without invTV4 (2.3 vs. 1.8, p < 0.01) (Figure 4). Patients with ST depression in the lead aVF had higher mPAP (18.5 mmHg vs. 15.0, p < 0.02, Figure 4).

This is the first study to investigate the impact of school ECG screening on the diagnosis of hemodynamically significant ASD in a single tertiary medical service area in Japan. Firstly, the school ECG screening group accounted for 88% of patients diagnosed at school age and 37% of overall pediatric patients in Japan, with the 3 corresponding peaks of the number of patients at diagnosis. Secondly, compared with the non-ECG screening group which were mainly detected by the health care and health checkups in infancy or preschool period, the ECG screening group had a lower proportion of audible heart murmur at catheterization, but exhibited similar levels of hemodynamic parameters. Thirdly, patients who were positive for a composite ECG parameter (the presence of rsR' type of iRBBB, inverted T in V4, or ST depression in the aVF lead) accounted for 79% of the ECG screening group at catheterization, and each of these parameters was correlated with hemodynamic parameters in the overall patient population. The present study suggests that the inclusion of school ECG screening, using the present composite ECG parameter, in the pediatric healthcare system may help detect otherwise undetectable ASD patients with clinically relevant hemodynamics (Figure 5).

The coverage of this study in the tertiary medical service area may be reasonably high considering the estimated incidence of ASD patients. This study was conducted at the only tertiary referral center in a prefecture in Japan, which covers approximately 310,000 of the population under 18 years of age in the prefecture, with an annual number of live births of approximately 10,000 (16). Based on the current study enrollment of 116 pediatric ASD patients over 11 years from 2009 to 2019, the estimated incidence of pediatric ASD patients was 28 per 100,000 live births. Previous reports on ASD diagnosis using echocardiography have reported the incidence ranging from 100 to 390 per 100,000 live births (2, 17, 18), with spontaneous closure rates of 62%–89% (17, 19). Calculating patients with indications for closure from these reports yields 11–148 per 100,000 live births. Despite being a retrospective study, the coverage rate in this study may be reasonably high considering the estimated incidence of ASD patients for shunt closure in the tertiary medical service area.

ECG screening has been performed for all school children in Japan (8) and for individuals participating in competitive sports in parts of Italy (20). Early studies have shown that ASD was included in the list of diseases prompted to be diagnosed by ECG screening in addition to the substrate for sudden cardiac death (8, 14, 15, 21). However, it was unclear how significantly this healthcare system impacts the detection of ASD in the clinical settings. In the present study, the school ECG screening group accounted for 88% of patients diagnosed at school age and 37% of overall pediatric patients in a tertiary medical service area, with the three peaks of the age distribution at diagnosis that coincided with the age of ECG screening. In contrast, a study in Boston identified ASD in school-age patients predominantly through incidental echocardiography or the presence of a heart murmur. In preschool children, however, a similar proportion of detection through heart murmurs, cardiovascular findings, or incidental echocardiography were confirmed in both studies. These findings suggest that the implementation of school ECG screening has an impact on the detection of ASD in overall children and in school students. Furthermore, in the present study, compared with the non-screening group, the screening group exhibited a lower proportion of audible heart murmur despite comparable levels of hemodynamic parameters. These findings suggest that ECG screening program could be an effective strategy for detecting hemodynamically significant ASD in students, who are asymptomatic and murmurless.

In this study, the prevalence of a composite ECG parameter [the presence of iRBBB (rsR'), invTV4, or ST depression of aVF] was 63% in the non-screening group as well as 79% in the screening group, suggesting a reasonably high sensitivity of this parameter for detecting hemodynamically significant ASD. The sensitivity of rsR' QRS pattern in detecting hemodynamically significant ASD (55% in the non-screening group and 70% in the screening group) was consistent with previous reports (54%–79%) (2, 15, 18, 22). Although the ECG finding invTV4 in ASD patients was reported previously (10, 23), the sensitivity of invTV4 or ST depression in aVF in detecting ASD was unknown. These findings suggest that the present composite parameter may help in detecting ASD. Furthermore, these three parameters had hemodynamic relevance. The present study also showed that iRBBB (rsR') that has been regarded as reflecting right ventricular overload and invTV4 were associated with higher Qp/Qs, which is consistent with the previous studies (18, 23, 24). Patients with ST depression in aVF, an ECG index suggestive of right ventricular pressure load (25), was associated with the higher mPAP, not Qp/Qs. Furtheremore, these parameters have high specificity: the prevalence of iRBBB (rsR') and invTV4 were reported in general Japanese school children (IRBBB: 0.983% in male and 0.410% in female; InvTV4: 0.12% in male and 0.15% in female in 1st grader, 0.10% in male and 0.13% in female in the 7th graders, 0% in male and 0.04% in female in the 10th graders) (26, 27).

In addition to the present three ECG parameters, we also evaluated right-axis deviation and “crochetage” pattern in the present study. Although these parameters increased the sensitivity in detecting ASD, the specificity of such parameters was reported to be low. The sensitivity and specificity of “crochetage” pattern for ASD detection has been shown to be lower in children than in adults (sensitivity: 31.7 vs. 57; specificity: 86 vs. 92) (28, 29), and a relatively high prevalence (2.74%) has been reported in children with normal echocardiographic findings (30). Collectively, the proposed composite ECG parameter may be useful for the screening of hemodynamically significant ASD in this age group.

There are four limitations in this study. Firstly, although patients were enrolled using a prospectively constructed patient database, the retrospective nature of this study should be considered in interpreting the results. Auscultatory findings are a subjective indicator to promote the diagnosis of ASD. Although board-certified pediatric cardiologists perform auscultation at the time of catheterization, there may be inter-observer variability. Secondly, the specificity or the false positive rate of the ECG index for ASD diagnosis was not examined in the present study or in the literature, although the low prevalence of relevant ECG parameters, including 0.41%–0.98% for iRBBB (rsR') and 0.04%–0.15% for invTV4, were reported in general Japanese school children (8, 26, 27). Thirdly, the limitations of the school ECG screening in extracting ASD: the present composite parameter was positive in only 79% of patients. In addition, given the natural history of ASD during life, patients who did not have significant shunting in childhood may have increased shunt blood flow in adulthood, due to age-related changes in cardiac function (i.e., decreased left ventricular compliance). Fourth, cost-effectiveness cannot be determined in this study. Tanaka et al. proved the cost-effectiveness of school ECG screening from the viewpoint of sudden death prevention (31). Since the detection of ASD is a byproduct of this program, it may be cost-effective as well. In the future, automated ECG recording and application of deep learning systems to ECG interpretation may further improve cost-effectiveness.

Firstly, patients treated for otherwise undetectable ASD, who were prompted by the present ECG screening program, may be free from comorbidities associated in adulthood. Secondly, on top of the detection of substrates for sudden cardiac death, ASD should be included in the target heart diseases for pediatric ECG screening. Thirdly, the present composite parameter (presence of rsR' type of iRBBB, inverted T in V4 or ST depression) may be recommended for the ECG screening of ASD.

The present study shows that school ECG screening detects otherwise unrecognized ASD, which prompted the diagnosis of the majority of patients at school age and >one-third of overall patients in Japan. These findings suggest that ECG screening program could be an effective strategy for detecting hemodynamically significant ASD in students, who are asymptomatic and murmurless.