Population-Based Study of Congenital Heart Defects in Down Syndrome

Transcription

1 American Journal of Medical Genetics 80: (1998) Population-Based Study of Congenital Heart Defects in Down Syndrome Sallie B. Freeman, 1 * Lisa F. Taft, 1 Kenneth J. Dooley, 2 Katherine Allran, 1 Stephanie L. Sherman, 1 Terry J. Hassold, 3 Muin J. Khoury, 4 and Denise M. Saker 3 1 Department of Genetics, Emory University, Atlanta, Georgia 2 Children s Heart Center, Department of Pediatrics, Emory University, Atlanta, Georgia 3 Department of Genetics and the Center for Human Genetics, Case Western Reserve University and University Hospitals of Cleveland, Cleveland, Ohio 4 Office of Genetics and Disease Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia Mental retardation and hypotonia are found in virtually all Down syndrome (DS) individuals, whereas congenital heart defects (CHDs) are only present in a subset of cases. Although there have been numerous reports of the frequency of CHDs in DS, few of the studies have had complete ascertainment of DS in a defined geographic area. The Atlanta Down Syndrome Project, a population-based study of infants born with trisomy 21, provides such a resource. In the first 6.5 years of the study, 243 trisomy 21 livebirths were identified in the five-county Atlanta area (birth prevalence: 9.6/10,000). Cardiac diagnoses were available on 227 (93%) of the cases and 89% of these evaluations were made by echocardiography, cardiac catheterization, surgery, or autopsy. Of the 227 DS infants, 44% had CHDs including 45% atrioventricular septal defect (with or without other CHDs), 35% ventricular septal defect (with or without other CHDs), 8% isolated secundum atrial septal defect, 7% isolated persistent patent ductus arteriosus, 4% isolated tetralogy of Fallot, and 1% other. This report is unique in that it contains the largest number of trisomy 21 infants ascertained in a population-based study where modern techniques for diagnosing cardiac abnormalities predominate. Am. J. Med. Genet. 80: , Wiley-Liss, Inc. Contract grant sponsor: Georgia affiliate of the American Heart Association; Contract grant sponsor: NIH; Contract grant numbers: N01-HD and PO1-HD 32111; Contract grant sponsor: American Academy of Pediatrics Section on Cardiology. *Correspondence to: Sallie B. Freeman, Ph.D., Department of Genetics, 1462 Clifton Road, NE, Emory University, Atlanta, GA sfreeman@genetics.emory.edu Received 9 January 1998; Accepted 23 June Wiley-Liss, Inc. KEY WORDS: trisomy 21; Down syndrome; congenital heart defects; congenital malformations INTRODUCTION Aside from mental retardation and hypotonia, which are present in virtually all Down syndrome (DS) individuals, the frequency of typical findings ranges from about 1% for leukemia [Zipursky 1992] and 12% for various gastrointestinal defects [Knox and ten Bensel, 1972; Scola, 1982] to 40% or more for congenital heart disease (CHD; see Discussion). Pueschel [1982] reviewed the earlier literature in which prevalence figures for CHD in DS ranged from 16% to 62%. Of more recent studies, the largest survey of children with CHD and DS was the Baltimore- Washington Infant Study [Ferencz 1989]. It described 218 children with DS ascertained because of heart defects (Table I). However, despite a wealth of literature on the subject, there is a surprising lack of population-based data on the frequency of CHD in DS. That is, few of the studies have started with complete ascertainment of all DS newborns in a defined geographic area. In one of the largest studies of this type, Stoll et al. [1990] reported the incidence of DS with and without CHD in 118,265 consecutive births in France (Table I). Pradat [1992] described the spectrum of CHD in 167 DS individuals ascertained on a population basis in Sweden. An advantage of these recent studies is that more accurate diagnostic tools have come into widespread use for the early detection of heart defects. The Atlanta Down Syndrome Project (ADSP) was initiated in 1989 to examine correlations between chromosome 21 nondisjunction and environmental, maternal, and other epidemiologic factors. The ADSP works in conjunction with the Centers for Disease Control and Prevention (CDC) to screen all prenatal, clinical, and cytogenetic diagnoses in the five-county metropolitan Atlanta area. Virtually all liveborn and stillborn infants with DS are identified. Cases ascertained through the ADSP-CDC collaboration constitute a valuable resource for determining the incidence of

2 214 Freeman et al. Current study (Atlanta) Ferencz et al., 1989 (Baltimore- Washington, DC) TABLE I. Reports of CHDs in DS Martin 1989 (Washington, DC) Stoll 1990 (France) Khoury and Erickson, 1992 (Atlanta) Pradat, 1992 (Sweden) Wells 1994 (Alabama) Rowe and Uchida, 1961 (Canada) Study period Unknown Population based Yes No No Yes Yes Yes No No Number of cases 227 b b 174 b Percentage with CHD 44% 47% 44.6% 33% 48% 40.2% Types of CHD (% of total with CHD) a AVSD 45.0% 60.1% 49% 41.9% 53% 59% 38.8% 35.7% VSD 35.0% 15.6% 22% 29.0% 17% 30.6% 32.8% 2 ASD 26.0% { 9.6% 14% N/A d 20% 31.7%e 28.6% 8.6% Coarctation of the aorta 1.0% 2.3% N/A d 4.8% N/A d 3.6% 4.1% N/A d TOF 5.0% f 7.3% 3% 3.2% N/A d N/A d 8.2% 1% Other 9.0% 1.8% 8% d 14.5% N/A d N/A d 14.3% 11% a Total may be >100% when more than one defect per infant is present. Total may be <100% because PDAs were omitted from this table. b Number of cases with records available for review (total number of cases may be greater). c 8% isolated. d N/A, not available or grouped with other cardiac defects. e ASD and/or VSD. f 4% isolated. CHD in DS. Moreover, this study is the largest reported to date where modern cardiac diagnostic techniques predominate. Thus it provides a highly accurate accounting of CHD in DS. METHODS Case Ascertainment The ADSP population was ascertained through the ongoing Metropolitan Atlanta Congenital Defects Program (MACDP) of the CDC. The MACDP uses multiple sources to identify infants born with birth defects including DS. MACDP personnel monitor births in the five-county metropolitan Atlanta area by regularly reviewing records from birth hospitals, pediatricians, specialty clinics, and cytogenetic laboratories. Details of the system are described elsewhere [Edmonds 1981]. Additionally, for the present study, identification of DS infants within the five-county Atlanta area was facilitated because many were referred to Emory University for karyotyping, genetic consultation, or cardiac evaluation. Included in the current report are all liveborn infants with trisomy 21 born between January 1, 1989 and June 30, 1995 to women living in the five-county Atlanta area at the time of the birth. The number of stillbirths ( 20 weeks gestation) is reported here, but, because no cardiac information was available, these cases are not included in the tabulation of CHD. Record Review MACDP abstractors reviewed records of infants with trisomy 21 and collected data on phenotypic manifestations. They noted the methods of diagnosis for each case [e.g., cardiac exam, electrocardiogram (ECG), chest X-ray, echocardiography]. However, most detailed cardiac information in this study was obtained by reviewing records of the pediatric cardiologists who evaluated the children referred to their practice. Furthermore, it was customary for all ECGs and echocardiograms to be interpreted by pediatric cardiologists. Reports of cardiac catheterizations, surgeries, and autopsies were reviewed by ADSP study personnel. The most definitive diagnostic procedure was used to classify the cases and resolve any discrepancies. For example, a ventricular septal defect (VSD) noted on physical examination was classified as a partial atrioventricular septal defect (AVSD) if the echocardiogram showed the VSD to be of an inlet type. Because of referral patterns in the Atlanta area, most cardiac evaluations in this study were done through a single large university-based pediatric cardiology practice, thus providing consistency in diagnosis. Classification We classified CHD according to standard anatomic nomenclature, which assigns priority to the morphologically dominant traits. The classification scheme used in this study is presented in Table II. The total cases were divided into two groups: those with CHD and those without. The CHD cases were then classified by assigning them first to a broad category and then by refining the categories to include any associated cardiac defects [Park 1977]. The CHD cases were divided into six groups: AVSD, VSD, secundum atrial septal defect (2 ASD), tetralogy of Fallot (TOF), patent ductus arteriosus (PDA), and other. For the purposes of this study, patent foramen ovale (PFO) versus 2 ASD and a PDA associated with prematurity or which closed spontaneously were not considered forms of CHD. RESULTS The birth population in the five-county Atlanta area from January 1, 1989 through June 30, 1995 was 255,401 (252,374 livebirths, 3,027 stillbirths) and 250 infants with trisomy 21 were identified (243 livebirths, 7 stillbirths). This gives a birth prevalence of 9.8 per 10,000 births for trisomy 21 (9.6 per 10,000 livebirths). Table III indicates the total trisomy 21 livebirths in the geographic area, the number of cases with and

3 Congenital Heart Defects in Down Syndrome 215 TABLE II. CHDs in DS CHD Number (%) AVSD Total 45 (45%) Complete Isolated 19 With 2 ASD 5 Only And arch abnormality a 1 And TOF 1 With persistent PDA only 7 With arch abnormality b 2 Partial 1 ASD 3 Inlet VSD 7 VSD Total 35 (35%) Isolated 18 With 2 ASD (±PDA) only 7 With 2 ASD and other c 4 With persistent PDA only 4 With other d 2 2 ASD (isolated) 8 (8%) TOF (without AVSD) 4 (4%) PDA (persistent) 7 (7%) Other heart defects e 1 (1%) Total 100 (100%) a Aberrant left subclavian artery. b Right aortic arch, coarctation. c Double outlet right ventricle (2), pulmonic stenosis (2). d Coarctation, double aortic arch, and pulmonic stenosis. e Vascular ring (aberrant left subclavian artery and right aortic arch). without cardiac information, and the methods of diagnosis for cases with and without CHD. Of the 243 cases ascertained, there were 227 cases (93%) with cardiac records available for review. Congenital heart disease was present in 100 cases (44%), and absent in 127 cases (56%). For those with CHD, all diagnoses were made by echocardiogram, cardiac catheterization, surgery, or autopsy. Eighty percent of those without CHD were evaluated by echocardiography, but, in 17%, CHD was ruled out by ECG or physical examination. In the three remaining cases, the method used to exclude CHD was not documented. Table II presents the types of heart defects found in this DS population. AVSD was the most common type, representing 45% of CHDs. Complete AVSDs were seen in 35 cases and were usually isolated (19 cases). Defects associated with a complete AVSD included 2 ASD, PDA, arch abnormalities (coarctation, right aortic arch, aberrant right subclavian artery), and TOF. Partial AVSDs included three ostium primum ASDs and seven inlet VSDs. VSD was the second most common type of CHD (35% of CHDs). Isolated VSDs were frequent; however, VSDs were also commonly associated with a variety of other defects: 2 ASD, PDA, arch abnormalities, right ventricular outflow tract obstruction (pulmonic stenosis and pulmonary atresia), and double-outlet right ventricle. Isolated secundum ASD (8%), TOF (4%), persistent PDA (7%), and one case with a vascular ring (aberrant left subclavian artery and right aortic arch) made up the remainder. Table IV presents the racial and maternal age distribution of cases with and without CHD. There were no significant differences in either the mother s race or age between those cases with and without CHD or AVSD. DISCUSSION Table III demonstrates that, in the present study, diagnoses were chiefly made on the basis of echocardiography, cardiac catheterization, surgery, or autopsy. This was true of infants with CHD (100%) and without CHD (80%). Tubman et al. [1991] compared the accuracy of several types of diagnostic procedures in a study of DS and CHD. They found that 29% of infants with a normal ECG were subsequently found by echocardiography to have a significant heart defect. There were 16 infants in our study for whom CHD was ruled out solely by an ECG plus six with only a physical examination. If 29% of these actually had a CHD, this would add six cases to our list of those with heart defects, but the overall rate would not increase appreciably (47% versus 44%). Little can be said about the 16 for whom we have no cardiac information. If we assume that none of them had a heart defect, the prevalence of CHD would change from 44% to a minimum possible rate of 41%. In order to obtain an accurate estimate of the prevalence and types of heart defects in DS, it is essential to have a population-based sample and to use the most reliable diagnostic methods currently available. The difficulty has been finding a population where both of these criteria can be met. Over the past decade there have been a number of studies reporting the occurrence of CHD in DS. Table I summarizes several of these recent reports and indicates whether or not they are population based. For comparison, the table also includes an earlier study conducted in the 1950s [Rowe and Uchida, 1961]. Many of the studies have enrolled only those cases presenting to a cardiologist or to a geneticist and, therefore, they do not necessarily include all DS infants born in a defined population. For example, in the Baltimore-Washington Infant Study [Ferencz 1989], infants were ascertained through pediatric cardiology clinics; thus, infants without CHD were not included and an overall assessment of the prevalence of CHD in DS was not possible. However, the authors of that study used echocardiography, cardiac catheterization, surgery, and autopsy to obtain an accurate accounting of the types of heart defects in all DS infants with CHD who were born in the geographic area covered by their study. They reported the highest rate of AVSD of any of the studies (60.1%), but a generally lower rate of non-avsd VSD (15.6%) and 2 ASD (9.6%). These differences may reflect the types of cases most often referred to pediatric cardiologists. Khoury and Erickson [1992] reported on DS and CHD using the same geographic area and similar ascertainment methods as in the present study, but their sample was collected from 1968 to 1989 whereas our collection began in Therefore, there is only a oneyear overlap between the two study populations. In general, over the 22-year period covered in their survey, the rate of reported CHD in DS rose from 20% to over 50%, an increase they attribute to improved ascertainment of these defects. For those cases with

4 216 Freeman et al. TABLE III. Trisomy 21 Case Ascertainment and Method of Cardiac Evaluation Method of cardiac evaluation Echocardiogram Cases Number (%) Unknown Physical exam ECG (and/or other) b Trisomy 21 livebirths 243 No cardiac information 16 (7%) Cardiac information 227 (93%) CHD present 100 (44%) CHD absent 127 (56%) 3 a a CHD ruled out in newborn period but type of cardiac evaluation not recorded on abstracted record. b Echocardiography and/or cardiac catheterization, surgery, or autopsy. CHD, they reported similar rates of AVSDs (53% versus 45% of CHD in our study) and 2 ASD (20% versus our 26% isolated or with other CHD), but their proportion of VSDs was smaller (17% versus our 35%). In general, all of the recent reports listed in Table I confirmed that an AVSD is the most common heart defect in DS (38.8% to 60.1%). Results from the present study fell in the midrange (45%). Among the several reports, non-avsd VSDs occurred in 15.6% to 35% of DS infants, the latter proportion being from the present study. Not included in Table I are the rates for PDA, which ranged widely from 3% to 41.9%. This reflects the difficulty in determining when a PDA is persistent and clinically significant and when it is related to prematurity or to having a cardiac evaluation in the first day or two of life. For example, the higher rate of PDA reported by Khoury and Erickson [1992], 38% versus our 7%, probably reflects the fact that those authors relied predominantly on newborn records whereas we were able to exclude the transient PDAs seen in some newborns by reviewing the cardiac reevaluations performed after initial hospital discharge. Finally, in the majority of studies, coarctation of the aorta and TOF were found in less than 10% of individuals. Although only the more recent studies of CHD in DS are discussed here and emphasis is placed on the importance of modern methods of cardiac diagnosis, it is remarkable that the four-decade-old study by Rowe and Uchida [1961] found results very similar to current studies that use more sophisticated diagnostic methods. Previous studies on CHD in DS have not always included information on demographic factors such as maternal age and race. Of the three population-based studies of DS cited in this article, the French population enrolled by Stoll et al. [1990] was over 90% white and the authors did not report CHD information separately by race or maternal age. In the study by Pradat [1992], subjects were ascertained from two Swedish birth defects registries. The authors did not include racial or maternal age information, although, presumably, the population was predominantly white. The Atlanta study reported by Khoury and Erickson [1992] included 346 white and 174 black infants with DS. The authors reported a significantly greater incidence of CHD in black infants (38.5% versus 28.6% for whites; odds ratio (OR) 1.6, 95% confidence interval (CI) 1.0 to 2.3). Our cases, from the same geographical area but a different time period, showed a trend toward increased CHD in black infants with DS (49% in blacks versus 40% in whites), but the difference was not significant (OR 1.48, 95% CI 0.8 to 2.7). Comparing the proportion of DS infants with CHD between women <35 years old and 35 years old, Khoury and Erickson [1992] did not find a significant difference. Similarly, we report no relationship between maternal age and CHD in our cases. Rowe and Uchida [1961] found no apparent difference in the maternal age distribution in cases with and without CHD. Their population was from Ontario, Canada, and again, presumably, was predominantly white, although racial and ethnic data were not included. Since the ADSP was originally designed to investigate the influence of maternal and environmental factors on chromosome nondisjunction, we have collected extensive epidemiological data on our population. A detailed analysis of the influence of these TABLE IV. Mother s Race and Age in DS Cases With and Without CHD Total trisomy 21 livebirths a CHD (%) No CHD (%) AVSD (%) No cardiac information All cases (44%) 127 (56%) 45 (20%) 16 Race Caucasian (40%) 76 (60%) 24 (19%) 6 Black (49%) 42 (51%) 18 (22%) 9 Hispanic 6 4 (67%) 2 (33%) 0 0 Asian 13 5 (42%) 7 (58%) 3 (25%) 1 Maternal age < (45%) 89 (55%) 33 (20%) (42%) 38 (58%) 12 (18%) 6 a In five-county Atlanta area from 1/1/89 6/30/95. Numbers in parentheses are percentages of all cases with cardiac information. None of the differences in this table reached significance (i.e., all P > 0.05).

5 Congenital Heart Defects in Down Syndrome 217 factors on CHD in DS will be the subject of a later paper. It is generally known that both the frequency and the types of CHD vary between the DS and non-ds population. For example, in the Baltimore-Washington Infant Study, AVSD was found in only 2.8% of non-ds cases but in 60.1% of DS cases [Ferencz 1989]. As pointed out by others [Van Praagh 1984; Ferencz 1989], several cardiac lesions seen in the non-ds population are rarely if ever found in individuals with DS. This was true in the present study where there were no cases of heterotaxy or transposition of the great arteries. These differences between the DS and non-ds populations suggest that a variety of spatially and temporally distinct processes are occurring during the formation of the heart and the presence of a third copy of a gene or genes on chromosome 21 has an impact on only specific developmental points. This provides incentive to search chromosome 21 for genes that, when present in three copies, disrupt specific steps in the embryological development of the heart [Duff 1990; Zittergruen 1995; Davies 1995]. Our laboratory is currently searching for these genes by identifying areas of disomic homozygosity on chromosome 21 [Feingold 1995] and by studying allelic variation in candidate genes for CHD. ACKNOWLEDGMENTS We acknowledge the cooperation of the Children s Heart Center at Emory University and thank the children with Down syndrome and their families for their participation. The work was supported by a grant from the Georgia affiliate of the American Heart Association, NIH contract N01-HD 92907, and NIH grant PO1-HD D.M.S. was supported by the American Academy of Pediatrics Section on Cardiology Research Fellowship Award. REFERENCES Davies GE, Howard CM, Farrer MJ, Coleman MM, Bennett LB, Cullen LM, Wyse RKH, Burn J, Williamson R, Kessling AM (1995): Genetic variation in the COL6A1 region is associated with congenital heart defects in trisomy 21 (Down s syndrome). Ann Hum Genet 59: Duff K, Williamson R, Richards SJ (1990): Expression of genes encoding two chains of the collagen type VI molecule during human fetal heart development. Int J Cardiol 27: Edmonds LD, Layde PM, Levy MJ, Flynt JW, Erickson JD, Oakley GP Jr (1981): Congenital malformations surveillance: Two American systems. Int J Epidemiol 10: Feingold E, Lamb NE, Sherman SL (1995): Methods for genetic linkage analysis using trisomies. Am J Hum Genet 56: Ferencz C, Neill CA, Boughman JA, Rubin JD, Brenner JI, Perry LW (1989): Congenital cardiovascular malformations associated with chromosome abnormalities: An epidemiologic study. J Pediatr 114: Khoury MJ, Erickson JD (1992): Improved ascertainment of cardiovascular malformations in infants with Down s syndrome, Atlanta, Am J Epidemiol 136: Knox GE, ten Bensel RW (1972): Gastrointestinal malformations in Down s syndrome. Minn Med 55: Martin GR, Rosenbaum KN, Sardegna KM (1989): Prevalence of heart disease in trisomy 21: An unbiased population. Pediatr Res 25:77A. Park SC, Matthews RA, Zuberbuhler JR, Rowe RD, Neches WH, Lenox CC (1977): Down syndrome with congenital heart malformation. Am J Dis Child 131: Pradat P (1992): Epidemiology of major congenital heart defects in Sweden, J Epidemiol Community Health 46: Pueschel SM (1982): Biomedical aspects in Down syndrome: Cardiology. In Pueschel S, Rynders J (eds): Down Syndrome: Advances in Biomedicine and the Behavioral Sciences. Cambridge, Massachusetts: Ware Press, pp Rowe RD, Uchida IA (1961): Cardiac malformation in mongolism. Am J Med 31: Scola PS (1982): Biomedical aspects in Down syndrome: Gastroenterology. In Pueschel S, Rynders J (eds): Down Syndrome: Advances in Biomedicine and the Behavioral Sciences. Cambridge, Massachusetts: Ware Press, pp Stoll C, Alembik Y, Dott B, Roth M (1990): Epidemiology of Down syndrome in 118,265 consecutive births. Am J Med Genet Suppl 7: Tubman TRJ, Shields MD, Craig BG, Mulholland HC, Nevin NC (1991): Congenital heart disease in Down s syndrome: Two year prospective early screening study. BMJ 302: Van Praagh S, Antoniadis S, Otero-Coto E, Leidenfrost RD, Van Praagh R (1984): Common atrioventricular canal with and without conotruncal malformations: An anatomic study of 251 postmortem cases. In Nora JJ, Takao A (eds): Congenital Heart Disease: Causes and Processes. New York: Futura, pp Wells G, Barker SE, Finley SC, Colvin EV, Finley WH (1994): Congenital heart disease in infants with Down s syndrome. South Med J 87: Zipursky A, Poon A, Doyle J (1992): Leukemia in Down syndrome: A review. Pediatr Hematol Oncol 9: Zittergruen MM, Murray JC, Lauer RM, Burns TL, Sheffield VC (1995): Molecular analysis of nondisjunction in Down syndrome patients with and without atrioventricular septal defects. Circulation 92: