A Population-Based Study of the 22q11.2 Deletion: Phenotype, Incidence, and Contribution to Major Birth Defects in the Population
Objectives. Although several studies describe the 22q11.2 deletion, population-based data are scant. Such data are needed to evaluate properly the impact, distribution, and clinical presentation of the deletion in the population. Our goals were to assess the population-based birth prevalence of the 22q11.2 deletion and its associated phenotype and its impact on the occurrence of heart defects.
Methods. We evaluated data on infants who were born from 1994 through 1999 to women who resided in metropolitan Atlanta. We matched records from the Metropolitan Atlanta Congenital Defects Program (a population-based registry with active case ascertainment), the Sibley Heart Center at Children’s Healthcare of Atlanta, and the Division of Medical Genetics at Emory University. We used birth certificate data for the denominators of the rates.
Results. We identified 43 children with laboratory-confirmed 22q11.2 deletion among 255 849 births. The overall prevalence was 1 in 5950 births (95% confidence interval: 1 in 4417 to 1 in 8224 births). The prevalence was between 1 in 6000 and 1 in 6500 among whites, blacks, and Asians and 1 in 3800 among Hispanics. Most affected children (81%) had a heart defect, and many (1 in 3) had major extracardiac defects (other than velopalatal anomalies), including anomalies of the central nervous system. Overall, the deletion contributed to at least 1 of every 68 cases of major heart defects identified in the total birth cohort and, in particular, to 1 of every 2 cases diagnosed with interrupted aortic arch type B, 1 of every 5 with truncus arteriosus, and 1 of every 8 with tetralogy of Fallot.
Conclusions. The 22q11.2 deletion was common in this birth population. The clinical phenotype included a wide and variable spectrum of major cardiac and extracardiac anomalies. From these population-based data, one can estimate that at least 700 affected infants are born annually in the United States. Population-based estimates such as these should be useful to medical professionals and policy makers in planning for the optimal care of people with the 22q11.2 deletion.
During the 2 decades after its initial identification in some children with the DiGeorge phenotype1,2 and 1 decade after the recognition of submicroscopic lesions in still more patients,3,4 the 22q11.2 deletion has been the subject of numerous studies. These studies have increased our understanding of many clinical aspects of the condition, noting the frequent occurrence of certain heart defects and velopharyngeal anomalies but also underscoring the variability of the overall clinical presentation, which can range from subtle isolated findings to severe multisystem involvement.5–19 More recently, as affected children are followed into adolescence and adulthood, their neurologic, developmental, and psychiatric findings are increasingly appreciated.20–31 This growing body of information, regularly summarized and reviewed,20,21,23,32–39 is beginning to answer some fundamental questions on the relative frequency, variability, and severity of structural anomalies associated with the deletion, the specificity and predictive value of cardiac findings, and the prevalence of the deletion in the population.
Although the existing knowledge base is useful, much of the evidence on 22q11.2 deletion to date has been derived from hospital-based case series, which may be subject to limitations regarding internal and external validity. In particular, case series rely on cases ascertained at selected hospitals or clinics and thus might not be representative of all cases in the general population regarding the severity or spectrum of associated conditions. Also, because the underlying birth population is not defined for many case series, one cannot reliably estimate the population-based prevalence of the deletion and its variation by race and ethnicity and measures of population impact.
Thus, well-designed population-based studies are needed to elucidate the findings from studies of hospital-based series, but their cost and complexity represent a major challenge and, perhaps for this reason, have been rarely attempted for the 22q11.2 deletion. Favorable circumstances made it possible to conduct a population-based study of the clinical, epidemiologic, and genetic aspects of the 22q11.2 deletion using a large, recent, well-defined, and racially diverse birth cohort in metropolitan Atlanta. These circumstances include the collaboration of 3 large centralized programs: a population-based birth defects registry with active case ascertainment from multiple sources; a center of pediatric cardiology and cardiovascular surgery that provides integrated diagnostic, clinical, and surgical services to children with heart defects; and a university-based department of medical genetics that provides comprehensive clinical and laboratory evaluation of genetic conditions.
This report describes our findings relative to 4 key issues: 1) the population-based prevalence of the 22q11.2 deletion, including its distribution by gender, race, and year of birth; 2) the underlying genetic findings and family history; 3) the associated congenital anomalies, with attention not only to specific congenital heart anomalies but also to extracardiac and multiple congenital anomalies; and 4) the contribution of the deletion to the overall occurrence of certain birth defects.
The study population comprised infants who were born from January 1, 1994, through December 31, 1999, to women who resided in 1 of 5 counties in metropolitan Atlanta (Clayton, Cobb, Dekalb, Fulton, and Gwinnett) at the time of birth of their child. County of residence was ascertained from the child’s birth certificate. During this period, 255 849 infants were born to resident mothers.
Data Sources and Methods
We used data from 3 main sources. One source was the Metropolitan Atlanta Congenital Defects Program (MACDP), managed by the Centers for Disease Control and Prevention and active since 1968. MACDP is a population-based birth defects registry that actively ascertains birth defects among infants who are born to women who reside in the 5-county Atlanta area. Trained abstractors regularly visit birth hospitals, pediatric and specialty wards, and cytogenetic laboratories to identify cases of birth defects from medical records, hospital logs, and laboratory logs. They also obtain and review birth and death certificates from the Georgia Department of Human Resources. The goal of the process is to ascertain all major structural birth defects among live-born children up to 6 years of age, among stillborn infants, and among pregnancies terminated at or after 20 weeks of gestation. The abstracted information is reviewed by MACDP’s medical staff, which includes clinical geneticists and pediatricians. For this study, we also reviewed data from prenatal diagnostic procedures. During the study period, MACDP ascertained 8379 cases of major birth defects and selected genetic conditions.
The second source of data was the Sibley Heart Center at Children’s Healthcare of Atlanta. Sibley Heart Center is the major provider of diagnostic services and the sole provider of surgical services for children with heart defects for metropolitan Atlanta and most of Georgia. In 2001, Sibley Heart Center’s outpatient service performed >30 000 evaluations, and the surgical team performed >880 surgical procedures. Sibley Heart Center participates in MACDP case ascertainment and is visited regularly by MACDP staff. In late 1995, staff at Children’s Healthcare of Atlanta and Sibley Heart Center started an educational campaign to increase knowledge of and testing for the 22q11.2 deletion. The campaign targeted in-hospital and community-based cardiologists, physician’s assistants, and nursing staff, with in-service seminars held approximately twice each year. In 1996, it became standard to screen with the 22q11.2 fluorescence in situ hybridization (FISH) probe all infants and children who have heart defects and are admitted for cardiac surgery and to reevaluate older patients with major heart defects.
The third main source of data was the Division of Medical Genetics at Emory University. The division provides clinical genetic services for individuals within the Atlanta area and much of Georgia. The Emory Genetics Laboratory annually processes approximately 5000 cytogenetic specimens. During the study period, the laboratory performed ∼1400 analyses for 22q11.2 deletion. In the laboratory, the deletion status was determined by FISH using a commercially prepared DNA probe mixture that included the D22S75 locus in chromosome region 22q11.2 (N25 probe) and a distal control locus in chromosome region 22q13 (probes supplied initially by Oncor, Inc, Gaithersburg, MD, and then by Vysis, Inc, Downers Grove, IL). Metaphase chromosomes were prepared by standard cytogenetic methods for cultured peripheral blood lymphocytes. FISH was performed following the manufacturer’s protocols. A positive result, consistent with deletion in 22q11, was determined by absence of the D22S75 hybridization signal on 1 chromosome 22 along with a normal pattern of hybridization for the distal control locus.
Even before formal matching, the 3 data sources (the registry, the Sibley Heart Center, and the laboratory) partially overlapped. For example, abstractors from the registry regularly visit the Sibley Heart Center and the laboratory, and physicians at the Sibley Heart Center primarily use the laboratory of the Division of Medical Genetics for genetic testing. Although the reason for genetic testing was difficult to evaluate in individual cases, common reasons would have included the presence of heart defects that required surgery, regardless of whether they were associated with extracardiac anomalies or dysmorphic features; features of the DiGeorge or velocardiofacial syndrome; velopharyngeal insufficiency; and an affected relative in the presence of even minimal clinical features.
Data Matching and Analysis
Institutional review boards at the 3 institutions approved the study, which included matching the data sets from Centers for Disease Control and Prevention’s MACDP registry, Emory University’s Division of Medical Genetics, and Sibley Heart Center at Children’s Healthcare of Atlanta. To ensure confidentiality protections, data went from the Division of Medical Genetics (laboratory records of people tested for the deletion) and Children’s Healthcare of Atlanta (records of children with 22q11.2 deletion and heart defects) to the MACDP data set. Only MACDP staff who were directly connected to data matching had access to the merged data set with personal identifiers. The analytic files did not include personal identifiers.
The 3 databases were merged, and positive and ambiguous matches were verified manually. For children who had 22q11.2 deletion and were not ascertained by MACDP but were ascertained by the other 2 databases, MACDP staff obtained the birth certificates to determine mothers’ residence at time of birth and confirm their study eligibility. We used the number of liveborn infants from birth certificate data, stratified by year of birth and race of the mothers, as the denominator for the computation of rates.
Some of our tables group together the following major heart defects: interrupted aortic arch type B, truncus arteriosus, tetralogy of Fallot and its anatomic variants (with pulmonary atresia or with absent pulmonary valve), and D-transposition of the great arteries. In past studies, this group has been variably labeled as conotruncal heart defects (referring to their putative embryologic origin) or outflow tract defects (referring to their anatomic location). In this report, we use the term conotruncal heart defects.
We initially identified 45 children for whom a deletion of the 22q11.2 region was documented. Of these children, 2 had chromosomal rearrangements. In 1 of the 2 children, the rearrangement was described as 45,XY,der(5)t(5;22)(p15.2;q11.2),−22. The rearrangement involved chromosomes 5 and 22 and resulted in monosomy of both the DiGeorge and cri-du-chat critical regions. Clinical findings included a small ventricular septal defect, bifida uvula, micrognathia, microphthalmia, hypotelorism, and absent fingernails. Several other minor anomalies and cat-like cry were also noted. The child’s mother carried the balanced form of the translocation. The other child carried a complex rearrangement involving a translocation of chromosomes 21 and 22 in which material was deleted from the proximal long arm of chromosome 22. The father carried a balanced translocation of chromosomes 21 and 22. The child presented with typical findings associated with the DiGeorge phenotype, including truncus arteriosus and absent thymus, as well as severe bilateral cleft lip and palate. In both of these cases, the deleted region of 22 encompassed not only the DiGeorge critical region but also a more extensive region of proximal 22q. The first case also included monosomy of 5p. Because their 22q deletions were larger than the common form of 22q1.2 deletion, we chose not to include these 2 cases in the remainder of this report. The remaining 43 cases covered in this report had FISH-confirmed 22q11.2 deletion without chromosomal rearrangements.
Prevalence by Race, Gender, and Year of Birth
These 43 cases were ascertained among 255 849 children who were born from 1994 through 1999 to women who resided in the study area. The overall prevalence was 1.7 per 10 000 births or 1 in 5950 births (95% confidence interval [CI]: 1 in 4417 to 1 in 8224; Table 1). The prevalence among whites, blacks, and Asians was similar and varied from ∼1 in 6000 to 1 in 6500, whereas the prevalence among Hispanics was ∼1 in 3800. Although the point estimate of the prevalence among Hispanics was nearly twice that of other ethnic groups, the 95% CI around such estimates were wide and overlapped. The prevalence was similar for male and female patients, in the group as a whole (Table 1), and within racial groups (data not shown). The occurrence rate of the deletion did not seem to increase over time. In fact, the rate in 1994–1996 was 1 in 5041 births (95% CI: 1 in 3389 to 1 in 7868 births), whereas the rate in 1997–1999 was 1 in 7098 (95% CI: 1 in 4545 to 1 in 11 792 births).
The major structural anomalies diagnosed among the 43 infants are summarized in Table 2. In addition, dysmorphic features and minor anomalies were documented for 80% of the 43 children; however, children without documentation of such dysmorphic features lacked reports of a genetic consultation and may have had undocumented dysmorphic features.
The most common structural anomaly was a heart defect, reported in 35 (81%) of the affected children (Table 3). The largest group of such anomalies was the conotruncal defects (interrupted aortic arch type B, truncus arteriosus, tetralogy of Fallot and its variants, and D-transposition of the great arteries), reported in 27 (63%) children with the deletion. Among the conotruncal defects, tetralogy of Fallot and its variants were commonly diagnosed (15 cases) and often (9 in 15 cases) presented at the more severe end of the anatomic spectrum (eg, with absent pulmonary valve or with pulmonary atresia and ventricular septal defect). Vascular anomalies frequently accompanied the heart defects. The most common types were a right aortic arch, with or without mirror image branching of the arch arteries, and an aberrant origin of the subclavian artery.
Contribution to the Occurrence of Heart Defects in the Population
We used these data to evaluate the contribution of the 22q11.2 deletion to the occurrence of congenital heart defects in the underlying birth population. During the study period, 2396 infants received a diagnosis of a major heart defect, for an occurrence rate of heart defects of 9.4 per 1000 births. The deletion was ascertained in 1.5%, or 1 in 68 cases of heart defects in the population.
The contribution of the deletion was higher for certain heart defects (Fig 1). Approximately 1 of every 2 cases of interrupted aortic arch, 1 of every 5 cases of truncus arteriosus, and 1 of every 8 cases of tetralogy of Fallot in the population were attributable to the deletion. The contribution of the deletion to D-transposition of the great arteries and hypoplastic left heart was minimal.
Extracardiac Defects and Multiple Congenital Anomalies
Of the 43 affected children, 14 (33%) had anomalies other than those commonly associated with the deletion (heart defects, oral clefts, thymus and parathyroid anomalies, and dysmorphic features). The most common such anomalies, summarized in Table 2, were central nervous system anomalies (12%), including myelomeningocele (2 cases) and cerebral heterotopias (1 case); renal anomalies (12%), including renal hydronephrosis and atrophy; skeletal anomalies (7%); and gastrointestinal anomalies (7%), including 2 instances of anteriorly displaced anus, in 1 case with perineal fistula.
Two affected children had affected family members with laboratory-confirmed 22q11.2 deletion. In 1 family, the carrier mother had a cleft palate. In the other family, the carrier mother and a carrier sibling both had a cleft palate. We were able to document laboratory testing for the deletion in parents of 4 additional families, with negative results.
These population-based findings help to clarify and expand certain key aspects of the 22q11.2 deletion. Clinically, the findings further detail the spectrum of cardiac anomalies among children with the deletion. Notably, the most common heart anomalies among children with the 22q11.2 deletion were not necessarily the most specific or predictive. For example, tetralogy of Fallot was the most common heart anomaly, followed by interrupted aortic arch type B and truncus arteriosus. However, when viewed in relation to heart defect occurrence in the general pediatric population, the presence of tetralogy of Fallot was less predictive of the deletion (present in only 1 of every 8 cases of tetralogy of Fallot in the population), compared, for example, with interrupted aortic arch type B (1 of every 2 such cases had the 22q11.2 deletion), a rarer but more specific defect. Our findings also indicate that the deletion probably is a minor contributor to major heart anomalies such as transposition of the great arteries and hypoplastic left heart, a suggestion also made by other reports.6,15,17,40–42
These findings can be helpful not only in clinical practice but also in etiologic studies because they underscore the etiologic heterogeneity of these cardiac anomalies. For example, researchers involved in etiologic studies should consider screening for the 22q11.2 deletion before examining the role of environmental factors and, if possible, assess risk factors separately for each type of conotruncal heart anomaly.
In terms of overall phenotype, we noted that most children who have heart anomalies and have the 22q11.2 deletion also have multiple minor anomalies and dysmorphic features.42–45 Conversely, the deletion was unlikely in those children who have a conotruncal heart defect without these features.14,46
The frequency and patterns of extracardiac anomalies were also notable. For example, although anomalies of the central nervous system have been reported sporadically in the literature,30,47,48 the presence of neural tube defects, brainstem anomalies, and brain heterotopias in 1 of every 9 children in this study indicates that such conditions are a common part of the clinical phenotype. Indeed, their true frequency might be even higher because some of these anomalies, such as those of neuronal migration, are identified only through special imaging studies.
The occurrence of other anomalies, including anterior displacement of the anus, malrotation of the colon, eventration of the diaphragm, and vertebral and renal anomalies, adds to the complexity of the clinical presentation and management of such cases. From a developmental perspective, these patterns of cardiac and extracardiac defects raise questions about what genetic effects or interactions with environmental factors might account for such variable outcomes. Insights on these issues might lead to strategies for mitigating or preventing the more severe outcomes.
From a population perspective, the findings indicate that, at least in this community, the 22q11.2 deletion was relatively common, occurring in 1 in 5950 newborns (Table 1). As summarized in Table 4, this point estimate was higher than that observed in the few other studies that examined well-defined populations from northern Belgium (1 in 6400),19 northern United Kingdom (1 in 7700),49 and southern France (1 in 9700).50 The higher prevalence that we report might be caused by random fluctuation, as reflected by the wide and overlapping CIs of these studies. However, it might also be explained by increased genetic testing and more complete ascertainment. For example, we used active case ascertainment from multiple sources, as well as additional data from referral centers of pediatric cardiology and medical genetics. In addition, the more recent time of the study allowed for a fuller implementation of genetic testing. Nevertheless, our estimate should be considered a minimum prevalence, one that will likely increase as clinicians expand the range of conditions for which testing is requested and as children are observed longer. In fact, although the difference was not statistically significant, the prevalence for children who were born during the first 3 years of the study and therefore with a longer time under observation was 1 in 5041 births (lower 95% CI: 1 in 3389 births), whereas for children who were born during the last 3 years of the study, the prevalence was 1 in 7098 births (lower 95% CI: 1 in 4545 births).
The population-based approach of this study provides insight on the impact of the deletion among racial groups. The prevalence of the deletion was similar among white, black, and Asian children, suggesting that selective underascertainment by race is unlikely, at least in this community. The prevalence among children of Hispanic origin seemed to be higher than in other groups. However, given the small number of children in this group, such variation is consistent with random fluctuation. Its significance, if any, will need to be assessed in future, larger studies.
This study had a number of potential limitations. First, some cases may have been missed. Because genetic testing for the deletion still depends on clinical referral, incomplete ascertainment of cases is a reasonable possibility, particularly for those cases whose phenotypic findings are minimal, of late onset, or sufficiently uncommon not to raise clinical suspicion. Thus, our prevalence rates and proportion of cases with major malformations should be regarded as minimum estimates. Likewise, the proportion of familial cases likely represents a minimum estimate because genetic testing was not done or known for all parents. Why some parents might be reluctant to undergo testing is an issue that, in our view, deserves to be studied further. Finally, this study was not designed to assess the spectrum of developmental and psychiatric findings associated with the deletion. A different approach will be needed to examine this important issue.
The main strengths of our study are that it is population based, recent, and based on a relatively large birth population compared with other published population-based studies.49,50 Another key strength is that the study relied on active case ascertainment from multiple clinical, genetic, and epidemiologic sources. Of particular importance, in our view, was the involvement of institutions that provide clinical, surgical, and genetic services to the population. Integrating these data with those from a population-based registry of birth defects also allowed us also to evaluate the contribution of the deletion to the occurrence of congenital heart defects in the population.
This population-based study estimates the prevalence of the deletion and its impact in the community and provides information on the spectrum of associated anomalies. These findings lend themselves to some final considerations. First, because of the wide phenotypic spectrum associated with the deletion, it would be reasonable to increase the range of clinical manifestations that would prompt genetic testing. Second, the barriers to genetic testing of family members should be evaluated carefully to ensure that families who wish to undergo such testing can do so without being penalized, financially or otherwise. Third, the growing list of health conditions associated with the deletion and the increasing estimates of prevalence suggest the need for greater attention to the clinical and public health impact of the 22q11.2 deletion.
Even if underascertained, the prevalence of the deletion seems to be high, twice that of phenylketonuria. Extrapolating these data on a national scale, we estimate that 700 children or more are born with the deletion every year in the United States alone. Many of these children will present with a wide spectrum of cardiac, neurodevelopmental, gastrointestinal, and other anomalies. Along with their families, they will face complex clinical, surgical, and developmental challenges. Population-based data such as these can help pediatricians, health care planners, and other health professionals to plan and manage better the acute and long-term care of these individuals.
We are grateful to the staff of MACDP, Sibley Heart Center, and the Division of Medical Genetics for contributing data and for supporting the study. We thank Don Gambrell for assistance with generating birth population data, Janet Cragan for assistance with the prenatal diagnosis data, and David Erickson and Cynthia Moore for helpful comments.
- Received June 19, 2002.
- Accepted October 22, 2002.
- Reprint requests to (L.D.B.) National Center on Birth Defects and Developmental Disabilities, Mailstop F-45, Centers for Disease Control and Prevention, 4770 Buford Hwy NE, Atlanta, GA 30341. E-mail:
- Liling J, Cross I, Burn J, Daniel CP, Tawn EJ, Parker L. Frequency and predictive value of 22q11 deletion. J Med Genet.1999;36 :794– 795
- ↵Ryan AK, Goodship JA, Wilson DI, et al. Spectrum of clinical features associated with interstitial chromosome 22q11 deletions: a European collaborative study. J Med Genet.1997;34 :798– 804
- ↵Devriendt K, Fryns JP, Mortier G, van Thienen MN, Keymolen K. The annual incidence of DiGeorge/velocardiofacial syndrome. J Med Genet.1998;35 :789– 790
- Lu JH, Chung MY, Hwang B, Chien HP. Prevalence and parental origin in tetralogy of Fallot associated with chromosome 22q11 microdeletion. Pediatrics.1999;104 :87– 90
- ↵Borgmann S, Luhmer I, Arslan-Kirchner M, Kallfelz HC, Schmidtke J. A search for chromosome 22q11.2 deletions in a series of 176 consecutively catheterized patients with congenital heart disease: no evidence for deletions in non-syndromic patients. Eur J Pediatr.1999;158 :958– 963
- ↵Goodship J, Cross I, LiLing J, Wren C. A population study of chromosome 22q11 deletions in infancy. Arch Dis Child.1998;79 :348– 351
- ↵Tezenas Du Montcel S, Mendizabai H, Ayme S, Levy A, Philip N. Prevalence of 22q11 microdeletion [letter]. J Med Genet.1996;33 :719
- Copyright © 2003 by the American Academy of Pediatrics