Results

In a population of 69 129 livebirths in 1994 and 1995 there were 478 cases of congenital heart disease diagnosed in life in infancy (6.9 per 1000 births), of which 207 were eligible for inclusion in this study. A further seven cases were born alive but diagnosed only after postmortem examination, giving a total of 485 liveborn cases (7 per 1000 livebirths). Eight infants died without having a sample taken for chromosome analysis. In five further cases the child had not had an invasive procedure, so there had been no opportunity to obtain a sample, and in 18 cases staff omitted to take the sample at the time of the procedure. Samples from three children were sent for chromosome analysis but not marked for inclusion in the study. In these three cases chromosome analysis was normal but FISH studies were not carried out. In three cases FISH analysis failed. Thus results of FISH analysis were obtained from 170 of the 207 cases.

Five submicroscopic deletions within chromosome band 22q11 were detected among these 170 cases, four of which had been clinically suspected, and two submicroscopic deletions within chromosome band 7q11–23 in children with Williams’ syndrome. Twenty one structural chromosome abnormalities were identified by standard metaphase analysis: 15 cases of trisomy 21, one trisomy 18, two cases of Turner’s syndrome, one apparently balanced de novo translocation 46,XY,t(5;21) (q11.2;q21), one unbalanced translocation 46,XX, −10,+der(10) t(3;10)(p23;p13)mat, and a mosaic 45,X/47, XYY/48,XYYY.

CHROMOSOME 22q11 DELETIONS IN INFANTS WITH SIGNIFICANT CONGENITAL HEART DISEASE

Four of the five infants with congenital heart disease and deletions in chromosome band 22q11, who were identified through the cardiothoracic unit, were suspected to have the deletion on clinical grounds, either because of their cardiac anatomy or because of additional features. One infant had type B interruption of the aortic arch, a ventricular septal defect (VSD), and neonatal hypocalcaemia. Another child with a large VSD and hypocalcaemia died with unexplained interstitial pulmonary fibrosis. One infant with pulmonary atresia and VSD had an aberrant subclavian artery raising suspicion of the diagnosis. The fourth infant had a VSD with left ventricular outflow tract obstruction and an anomalous right subclavian artery. The fifth infant, in whom the diagnosis had not been suspected, had pulmonary atresia and a VSD. Two paediatric cardiologists had felt that the baby was not dysmorphic; language and development problems have subsequently become apparent.

One baby with a cardiovascular malformation was diagnosed at necropsy after sudden death at 11 days old. She had type B interruption of the aortic arch, VSD, and absent thymus. The deletion was detected on cultured fibroblasts from a skin biopsy obtained at necropsy. Her mother and grandmother were later shown to have a deletion within chromosome 22q11.

Among the 207 babies eligible for inclusion in the study there were three with interruption of the aortic arch, 10 with pulmonary atresia with VSD, 26 with tetralogy of Fallot, five with truncus arteriosus, and 31 with VSD. In addition one case of truncus arteriosus and one case of interrupted arch presented with death and were diagnosed at necropsy. Two of four cases of interruption of the aortic arch had deletions, two of 10 cases with pulmonary atresia and VSD had deletions, and two of 31 cases of VSD had deletions within chromosome band 22q11.

CHROMOSOME 22q11 DELETIONS IN INFANTS WITHOUT SIGNIFICANT HEART DISEASE

In the same population of livebirths, three further infants were diagnosed as having deletions within chromosome 22q11. One child presented at 1 week of age with a hoarse voice secondary to a laryngeal web, and subsequently had hypocalcaemic seizures. He also had an umbilical hernia, right inguinal hernia, left undescended testis, and an atrial septal defect. He was not dysmorphic other than having a round ear shape. One child who did not have any structural abnormalities was referred because of his dysmorphic appearance; his mother was also shown to have a deletion. The third child had been seen by clinical geneticists several times before the diagnosis was made. She was dysmorphic with brachycephaly, simple ears, small nostrils, small carp shaped mouth, and a bifid right thumb. Her weight was on the third centile and she had mild developmental delay. FISH analysis was carried out when she began to vocalise and the sounds were noted to have a nasal quality.

COMPARISON OF POPULATION PREVALENCE OF CHROMOSOME 22q11 DELETION AND TRISOMY 21

Nine cases of submicroscopic deletion with chromosome band 22q11 were identified in infancy in a population of 69 129 livebirths, giving a birth prevalence of 13 per 100 000 (95% confidence interval (CI) 4.5 to 21.5).

In the same birth cohort there were 53 liveborn infants with trisomy 21 (77 per 100 000 livebirths), 15 of whom had significant heart disease. There were 48 terminations of pregnancy because of trisomy 21 and three terminations because of abnormalities resulting from chromosome 22q11 deletions in the same population in 1994 and 1995.

Discussion

The prevalence of deletions within chromosome band 22q11 in this study was 13 per 100 000 births (95% CI 4.5 to 21.5). In 1993, before the prospective study began, eight infants were born who had deletions diagnosed in infancy in the same population; seven of these had significant heart defects and one did not have congenital heart disease. Two further children born in 1993 were not diagnosed as having a deletion until the age of 4 years. One had tetralogy of Fallot and was later noted to have nasal speech. The second presented with developmental delay, ulnar deviation of the lateral three digits on each hand, and a history of frequent infections. He did not have a heart defect. He was of mixed racial origin and initially his subtle dysmorphic features were thought to be in keeping with the parental origins. This gives a prevalence for children born in 1993 of 25.7 per 100 000 births (95% CI 17.6 to 41.7).

Our figures are in keeping with a five year retrospective population study based on the birth defects registry of the Bouches-du-Rhône area in Southern France11 in which 12 cases of 22q11 deletion were identified in 116 452 births, giving a prevalence of 10.3 per 100 000 (95% CI 5.3 to 22.3). Ten of these 12 children had a heart defect. The authors point out that their figure is likely to be an underestimate of the true prevalence of the deletion because it only includes cases presenting with symptoms in infancy; this applies equally to our data. The phenotype associated with deletions within chromosome band 22q11 is very variable,12 as demonstrated by the fact that two of the mothers of the infants in this case series had previously undiagnosed deletions. One of these mothers had no overt clinical problems and the other had had a VSD repaired in childhood. Another adult with a deletion, whom we ascertained through a child with truncus arteriosus, had never come to medical attention; he was not dysmorphic, had no cardiac abnormality, and did not have nasal speech.

In this study four of the five children with congenital heart disease would have been detected when they presented if chromosomes had only been tested on the basis of clinical suspicion. In the fifth child there would have been a delay before diagnosis. Early diagnosis is important both for management of the patient and for assessing risk of recurrence in future pregnancies. Some authors have implied that a clinical diagnosis can be made in all deletion cases.13-16 However, although the diagnosis may become obvious in time it is not always obvious when the infant presents. The presentation of some of the children without heart defects in this study was atypical and one of the deletions in those with heart disease was not suspected clinically at initial presentation.

It is difficult to define criteria for undertaking deletion studies. There have been anecdotal reports of almost every type of heart defect in association with the chromosome 22q11 deletion.8 17 18 On the other hand, a deletion frequency of 3% does not warrant testing of all children with congenital heart disease. We suggest that FISH analysis is indicated in all children with type B interruption of the aortic arch, truncus arteriosus or pulmonary atresia with VSD. It should be performed when there is an aberrant subclavian artery or right sided aortic arch in association with a heart defect. It is also indicated when there are features associated with the condition such as hypocalcaemia, decreased T cell numbers, dysmorphic features, or renal abnormalities with or without a congenital heart defect. Finally, if a child who has not fallen within these criteria is noted to have a nasal tone when they start vocalising the test should be performed.

One child in this series was only identified as having a cardiovascular malformation, interruption of the aortic arch, after death. The child, her mother, and grandmother all had the deletion and there was a significant risk of recurrence. Interruption of the aortic arch is known to present with sudden death in infancy19 20 and the potential genetic implications of this should perhaps be more widely recognised.

In conclusion, the most common chromosomal cause of significant congenital heart disease remains trisomy 21, while the second most common is chromosome band 22q11 deletion. Diagnosis of deletion within chromosome band 22q11 is important for patient management and because of the possibility that one of the parents may also have the deletion. Early diagnosis enables tests to be offered to parents before they start another pregnancy. However, if the diagnosis is to be made early, the clinician must be aware of the wide range of problems associated with deletions within chromosome band 22q11.