Copyright 2009 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2009
To evaluate the sensitivity of a heart disease screening program for schoolchildren.
Prospective cohort study.
Taitung County, Taiwan.
All first- and fourth-grade elementary school students and first-year junior high school students screened between September 1, 2005, and June 30, 2007.
A 2-level heart disease screening program was undertaken. Level 1 screening included a history questionnaire and phonocardiography and electrocardiography. Level 2 screening consisted of examination by a pediatric cardiologist of all children who had abnormal findings on level 1 screening. A control group comprised children whose level 1 screening results were normal. Children with abnormalities on level 2 screening were referred for a detailed cardiac evaluation.
Main Outcome Measures
Previously unsuspected heart diseases.
Of 25 816 children undergoing the level 1 screening program, 5330 had positive findings, of whom 5235 underwent level 2 screening, along with 1104 children in the control group. The pediatric cardiologist referred 780 children to the hospital for full evaluation, including 18 controls and 114 patients with a previous diagnosis of heart disease. A total of 292 children, including 2 controls, had evidence of heart disease, which was previously unsuspected in 178 (61.0%). Excluding children with a previous diagnosis of heart disease, the sensitivity of the screening program was 82.6%. The specificity and the positive predictive value were 98.3% and 28.6%, respectively.
The 2-level heart disease screening program has good sensitivity and detects asymptomatic but potentially life-threatening heart disease in some children.
Unsuspected heart disease is a major cause of sudden death in schoolchildren. Niimura and Maki1 reported an annual incidence of sudden death in children of approximately 6.4 per 1 000 000. Acute heart failure of unknown etiology was found in 62% of these children, and only 19% had known cardiovascular disease. Of those in whom sudden death was attributed to a cardiac cause, 79% died during sporting activities. Therefore, screening schoolchildren to detect such disorders may be important to reduce the incidence of sudden cardiac death. A further benefit may be the early detection of heart disease that might not become symptomatic until much later, for example, atrial septal defect. This very common congenital heart defect may be unsuspected until adulthood, with some patients already having congestive heart failure or Eisenmenger syndrome at presentation. Muta et al2 had shown that screening in schoolchildren is useful to identify an asymptomatic atrial septal defect, which would allow appropriate follow-up and intervention if indicated, with the goal of preventing later complications.
Several studies3- 5 in various countries have looked at screening to detect congenital or acquired heart diseases in schoolchildren. In Japan, children in their first year of elementary and junior high school are required by law to be screened for heart disease.2 The screening program used6 involves initial evaluation by questionnaire, electrocardiography, and phonocardiography. This is followed, if indicated, by further evaluation. Although seeming to be useful for detecting patients at risk for sudden death, the sensitivity has not been clearly defined.7
Taitung County is a relatively remote county in Taiwan with 16 districts, 1 of which is Taitung City, an urban area with more medical resources than the other, more rural, districts.8 Overall, there are fewer medical resources in the county than elsewhere in Taiwan.9 In 2005 to 2007, a heart disease screening program modified from that used in Japan was supported by the Cardiac Children's Foundation of the Republic of China. All schoolchildren in the first and fourth grades of elementary school and the first year of junior high school in Taitung County were screened. Using data from this mass screening, we estimated the prevalence of heart disease in this area and the sensitivity of the screening program.
Between September 1, 2005, and June 30, 2007, all schoolchildren in the first (aged 6-7 years) and fourth (aged 9-10 years) grades of elementary school and the first year of junior high school (aged 12-13 years) in Taitung County were prospectively screened. The Mackay Memorial Hospital Institutional Review Board approved the study. Parents gave consent for their children to participate in the screening program, which was sponsored by the Taitung County government. Level 1 screening consisted of a questionnaire to be completed by the parents and checked by the teachers; simplified electrocardiography (ECG) using leads I, aVF, V1, and V6; and phonocardiography at the second left intercostal space and apex. The questionnaire included demographic data (sex, age, height, weight, address, and ethnicity), symptoms of heart disease (palpitation, syncope, chest pain, dyspnea, and signs of acute rheumatic fever or Kawasaki disease), a history of heart disease, present cardiopulmonary status, and a family history of heart disease. Both ECG and phonocardiography were performed by trained technicians and were analyzed using a computer, and the results were reviewed by a pediatric cardiologist.
Any students with abnormalities on the questionnaire, ECG, or phonocardiogram were deemed to have a positive result on level 1 screening. We selected a control group of students whose level 1 screening results were negative by choosing students assigned to the seats numbered 10 and 30 in the same classrooms as children who had a positive result. The control group and those with positive level 1 screening results then underwent level 2 screening, which involved evaluation by a pediatric cardiologist who went to schools to perform further physical examination, including a thorough assessment of the cardiovascular system and inspection for signs of heart disease, and review of the results of level 1 screening. Based on this assessment, children in whom significant heart disease was suspected were referred to Mackay Memorial Hospital for examination using a 12-lead ECG, chest radiography, and color flow Doppler echocardiography, as well as other studies as indicated. The results of this evaluation were then explained to the parents. Children were referred for intervention if appropriate and for follow-up. The schools were informed of the final test results, and advice regarding appropriate physical activity levels for children with heart disease were given based on the recommendations of the 36th Bethesda Conference.10- 16
Taipei City, the capital of the Republic of China, has a larger population and more medical centers than Taitung County. Since 2000, the identical 2-level heart disease screening program described previously herein has been performed in Taipei. We compared the results of screening in Taitung between 2005 and 2007 with the results from Taipei in 2006.17
The results of the 2-level screening program were compared with those of the full evaluation recommended by the pediatric cardiologist. Using the full evaluation as the reference standard, the sensitivity of the screening program was estimated. We also compared the results from Taitung County with those from Taipei using the Fisher exact test. P < .05 was considered statistically significant.
During the study, 26 046 children at the stated grade levels were eligible for screening in 95 elementary schools and 22 junior high schools. Screening in 230 children was incomplete, and they were excluded from the analysis, leaving 25 816 children (a screening rate of 99.1%) who underwent level 1 screening (13 377 boys and 12 439 girls). Approximately three-fifths of the children (n = 15 445) resided in Taitung City, and the remaining (n = 10 371) lived elsewhere in the county. Just more than one-third (n = 9247) were aborigines; 16 569 were nonaborigines.
The flow of participants through the study is illustrated in Figure 1. Possible abnormalities were present in approximately one-fifth of the children at level 1 screening. At level 2 screening by a pediatric cardiologist, most were found to have either nonspecific symptoms on the questionnaire or nonspecific phonocardiography findings (including benign murmurs). The cardiologist recommended that 780 children (762 with positive level 1 results plus 18 control subjects whose level 1 screening results were negative) with suspected or known heart disease undergo full evaluation at the hospital. Excluding 33 children who did not complete further evaluation, 747 underwent full testing; of these, 292 children were diagnosed as having heart disease, including 114 (39.0%) previously known to have a heart condition and 178 (61.0%) in whom it was newly diagnosed. The abnormalities included rheumatic heart disease (1 child), mitral valve prolapse (18 children), Kawasaki disease (19 children), congenital heart disease (75 children) (Table 1), ECG abnormalities (161 children) (Table 2), and miscellaneous disorders (18 children). Of children with previously undiagnosed conditions, 148 (83.1%) had an ECG abnormality. Of the 75 children with congenital heart disease, only 12 (16.0%) had not been previously diagnosed. Of all previously undiagnosed conditions, 26 were potentially lethal, including prolonged corrected QT interval (≥0.46 seconds) in 15 children, second-degree heart block in 4, and serious congenital heart disease in 7. Of the latter, 4 children had an atrial septal defect larger than 1 cm, 1 had a large ventricular septal defect, 1 had a coronary fistula, and 1 had congenitally corrected transposition of the great arteries. As a result of the screening program, 4 children underwent surgery or catheterization, open heart surgery was planned in another 3, and 15 were treated with long-term medication therapy.
Flowchart of participants through the study.
Of the 1104 children from the control group whose level 1 screening results were normal, 18 were referred by the physician for further evaluation, of whom 2 were diagnosed as having a heart defect (1 with mitral valve prolapse and 1 with mild pulmonary stenosis). Assuming that only those 2 children of the 1104 controls had false-negative results on level 1 screening, we then extrapolated to the 20 486 children with normal findings on level 1 screening, estimating that approximately 37 children would have had false-negative results. Including only children in whom a heart condition was not already known, the estimated sensitivity of the 2-level screening program for detecting heart disease was, thus, 82.6% (Table 3). We believe that we detected 100% of the potentially lethal conditions. Mitral valve prolapse and mild pulmonary stenosis are not considered lethal, so if they were missed, the consequences would likely not be very serious.
The prevalence of all heart disease and of congenital heart disease was significantly lower in Taitung than in Taipei (P < .001 for both) (Figure 2). However, the proportion of previously undiagnosed disease in both categories that was discovered only on screening was significantly higher in Taitung than in Taipei (P < .001 and P = .002, respectively) (Figure 3).
Prevalence of heart disease in Taitung County and Taipei City. CHD indicates congenital heart disease; ECG, electrocardiogram; KD, Kawasaki disease; MVP, mitral valve prolapse; and RHD, rheumatic heart disease.
Proportion of previously undiagnosed heart disease in Taitung County and Taipei City. CHD indicates congenital heart disease; ECG, electrocardiogram; KD, Kawasaki disease; MVP, mitral valve prolapse; and RHD, rheumatic heart disease.
This study showed that a 2-level screening program for schoolchildren is useful for detecting asymptomatic heart disease. More than half of the children diagnosed in this series had disorders that had not been diagnosed previously, several of which are potentially lethal.
Screening tools that have been suggested in this population include a focused history and physical examination, ECG, phonocardiography, and even echocardiography.6,18,19 Fuller5 compared the cost-effectiveness of a focused cardiovascular history and physical examination, 12-lead ECG, and echocardiography and concluded that 12-lead ECG was the most cost-effective. As noted previously herein, the nationwide heart disease screening conducted in Japan has added phonocardiography to the history, physical examination, and simplified ECG.2 Children with abnormal findings on screening are then immediately referred for full evaluation. The 2-level screening program was designed to reduce the number of children who require a full evaluation by sending a pediatric cardiologist to the schools to evaluate those with abnormalities on level 1 screening. Although 20.6% of the children initially screened were seen by the cardiologist, only 3.0% of the total study population actually required the full evaluation. The addition of level 2 evaluation thus greatly reduced the false-positive rate for the screening program overall. Formal cost-effectiveness assessment is beyond the scope of this study, but it is likely that this 2-level screening would reduce the cost of finding children with significant heart disease.
A major difficulty in evaluating large-scale screening programs such as this is assessing the ability of the program to minimize false negatives, that is, demonstrating adequate sensitivity. We are not aware of many studies that have used a control group as we did. In the present control group of 1104 students whose level 1 screening results were negative, on level 2 screening, only 18 were thought by the pediatric cardiologist to require further evaluation, of whom only 2 had abnormalities found on the full examination. The mitral valve prolapse in 1 control and the pulmonary stenosis in the other were considered mild and not life-threatening. Level 1 screening was, thus, very sensitive (97.7%) in identifying defects that a physician would also detect.
Because it was not feasible to perform the full hospital-based evaluation on all 1104 controls, we could only estimate the overall sensitivity of the 2-level screening program by extrapolating the number of false negatives in the control subjects to the entire screened population, which may have introduced an error in the calculations. The 18 controls were sent for further evaluation because the pediatric cardiologist heard a heart murmur that was not detected on level 1 phonocardiography screening. The sensitivity and specificity of phonocardiography thus may not be adequate for such a screening program, in which it is important to differentiate benign from potentially pathologic heart murmurs. If we perform a similar study in the future, we may have general pediatricians perform auscultation in the level 1 screening in place of phonocardiography.
The importance of broad-based screening of schoolchildren is highlighted by the higher prevalence of previously undiagnosed disease in Taitung County compared with Taipei City. Because the same screening program was used for children in the same age groups in both areas, it is reasonable to compare the results. The prevalence of all heart disease and of congenital heart disease was lower in Taitung County than in Taipei City. The reason for this is not clear, although 1 possibility is that infant mortality is higher in rural areas in Taiwan than in urban areas such as Taipei, so that fewer infants with congenital heart disease may survive to school age. On the other hand, of children with congenital or any heart disease found on screening in Taitung, a smaller proportion than in Taipei had been previously diagnosed. This may indicate that medical resources are indeed less available in Taitung, although we cannot rule out the possibility that families there do not choose to access care even if available. Whatever the explanation, we believe that these results highlight the need for periodic heart disease screening for schoolchildren in Taitung County.
The main limitation is noted previously herein, that is, that we can estimate the sensitivity of the screening program only by extrapolating from the 2 control subjects who were diagnosed as having heart disease even though their level 1 screening results were negative. This study also was not designed to explore the reasons for differences in the prevalence of certain types of heart disease, whether by geographic area or ethnicity, although it provides useful baseline data should such research be undertaken.
In conclusion, this 2-level heart disease screening program is sensitive for detecting asymptomatic high-risk heart disease in schoolchildren, especially in an area where medical resources may be scarce or underused, as in Taitung County.
Correspondence: Ming-Ren Chen, MD, Division of Cardiology, Department of Pediatrics, Mackay Memorial Hospital, 92, Section 2, Chun-San N Rd, Taipei, Taiwan 104 (email@example.com).
Author Contributions: All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lue and Chen. Acquisition of data: Yu. Analysis and interpretation of data: Yu, Wu, and Chen. Drafting of the manuscript: Yu. Critical revision of the manuscript for important intellectual content: Lue, Wu, and Chen. Statistical analysis: Wu. Administrative, technical, or material support: Yu and Chen. Study supervision: Lue and Chen.
Accepted for Publication: August 5, 2008.
Financial Disclosure: None reported.
Funding/Support: The screening program was sponsored by the Taitung County government and the Cardiac Children's Foundation of the Republic of China.
Role of the Sponsors: The Taitung County government conducted this study. The Cardiac Children's Foundation of the Republic of China designed the study and collected the data.
Yu C, Lue H, Wu S, Chen M. Heart Disease Screening of Schoolchildren in Taiwan. Arch Pediatr Adolesc Med. 2009;163(3):233-237. doi:10.1001/archpediatrics.2008.562