American Journal of Medical Genetics 37:366-370 (1990)

Malformations and Minor Anomalies in Children Whose Mothers Had Prenatal Diagnosis: Comparison

-

Between CVS and Amniocentesis

Paige Kaplan, Jean Jr. Normandin, Golder N. Wilson, Henri Plauchu, Abby Lippman, and Michel Vekemans Divisions of Medical Genetics, Departments of Medicine (P.K.) and Pediatrics (P.K., J.N., G.N. W., HP. , M.V.) and Department of Epidemiology and Biostatistics, (A.L.), McGill Uniuersity, Montreal, Quebec, Canada

The frequency of abortion following chorionic villus sampling (CVS) is similar to that follow- ing amniocentesis. However, there is no infor- mation on long-term effects, such as malfor- mations in liveborn children exposed to CVS. We evaluated 189 infants whose mothers had either CVS or amniocentesis as participants in the Canadian Collaborative Randomized Trial, a prospective assessment of the safety of CVS compared with amniocentesis. The par- ticipation rate of children who could be con- tacted was 95%. Ninety-five of the 189 infants (50.2%) had been exposed to CVS, 87 (46%) to amniocentesis, and 7 (3.8%) to both. (The latter group was excluded from calculations.) One hundred twenty-eight (128) children had z one minor anomalies but no major abnormal- ities: 58 of 95 (60%) in the CVS and 70 of 87 (80%) in the amniocentesis group. Twenty-six children had malformations: 17 (17.8%) in the CVS and 9 (10.3%) in the amniocentesis group. Only one anomaly, Sturge-Weber dysplasia (amniocentesis group), was potentially severe and none were life-threatening. Superficial cavernous hemangiomas (strawberry nevi) were noted more frequently in children in the CVS group (12.6%) than in the amniocentesis group (3.4%), but only slightly higher than in the general public.

We conclude that exposure to CVS is not associated with an increased frequency of malformations or minor anomalies in infants compared with amniocentesis although we observed a higher frequency of superficial cavernous (strawberry) hemangiomas in the children in the CVS group. Received for publication November 16, 1989; revision received

March 23, 1990. Address reprint requests to Paige Kaplan, M.B., B.Ch.,

FRCP(C), Division of Genetics, Royal Victoria Hospital, 687 Pine Avenue West, Montreal, Canada H3A 1Al. Dr. Kaplan is presently at the Children’s Hospital of Philadelphia, Biochemical Develop- ment and Molecular Diseases, 34th & Civic Center Boulevard, Philadelphia, PA 19104.

0 1990 Wiley-Liss, Inc.

KEY WORDS: malformations following CVS, abortion, liveborn children

INTRODUCTION The prenatal diagnosis of several hundred genetic and

hereditary disorders has become possible in the past 2 decades with the refinement of methods such as ultra- sonography and amniocentesis. Although the accuracy of the test results and the low frequency of procedure- related, non-induced abortion following amniocentesis are well established, a major concern for pregnant women and their families has been the relatively late gestational age at which the tests are performed and results obtained.

Chorionic villus sampling (CVS), performed a t ap- proximately 10 weeks of gestation, can circumvent some of these problems. It has the advantages of being done before a pregnant woman feels fetal movements, main- taining privacy of the pregnancy and having results of tests available within several days. In addition, bio- chemical and DNA testing may be done rapidly on small masses of tissue by using amplification techniques. However, many obstetricians are loathe to have their pregnant patients who are a t risk for genetic disorders be subjected to a procedure whose risks are not yet fully known [Fahy and Lippman, 19881. There is also wide- spread anecdotal belief in the community that CVS causes a higher rate of spontaneous abortions than does amniocentesis, as well as inflicting physical and neuro- logical damage to the fetus. Because CVS entails remov- ing trophoblastic tissue, there is the potential for delete- rious effects on the fetus, either by direct contact or indirectly, such as infection or hemorrhage.

Although CVS is now performed widely, only recently have its safety and accuracy been studied in a prospec- tive manner [Canadian Collaborative CVS-Amniocen- tesis Clinical Trial Group, 19891. The major objective of the Canadian Collaborative Randomized Trial was to compare the effects of transcervical chorionic villus sampling with amniocentesis. It was a controlled study in which women were randomly assigned by computer to

Malformations Following CVS and Amniocentesis 367

each test group. Preliminary data analyses, from this study as well as from the non-randomized study in the United States [Rhoads et al., 19891, suggest that total abortion rates do not differ significantly between the 2 groups. The present study examined the frequencies of malformations, minor anomalies, and birthmarks in liveborn children whose mothers had CVS or amniocen- tesis a t the McGill University Hospitals as part of the Canadian Trial.

MATERIALS AND METHODS The aim was to enroll all the infants whose mothers

had undergone prenatal diagnosis a t McGill University hospitals as participants in the Canadian Collaborative CVS-Amniocentesis Clinical Trial. All women having prenatal diagnosis at McGill University were eligible to participate in the Trial. At the time of consent and enrollment, pregnant women were requested to allow their infants to be evaluated after birth. Women were randomly assigned to have either CVS or amniocen- tesis. The infants whose mothers had undergone CVS were the study subjects. The infants exposed to amnio- centesis in the trial acted as controls.

The children were examined, whenever possible, in the first year of life. The study was carried out between OctoLcr 19% and September 1987. At the time of the assessment, the examiners were not aware of the reason for prenatal diagnosis nor the procedure used in each case. Initially, the 3 examiners (P.K., G.W., H.P.), all of whom are clinical geneticists and pediatricians, individ- ually examined each of a small subset of children and their evaluations of minor anomalies and malformations were compared. Because these were identical, each child thereafter was examined by only one geneticist (P.K. 75%, G.W. 22'37, H.P. 3%).

A standardized evaluation form, based on that of Leppig et al. [19871, was used to document minor anom- alies and malformations, as well as birthmarks and growth parameters (length, occipito-frontal circum- ference [OFC], weight, intercanthal distances and inter- nipp1e:chest ratios). The frequency and types of anoma- lies in our 2 groups were compared with each other as well as with the minor anomalies and malformation rates established by Leppig et al. [19871 and with the birthmark incidence noted by Jacobs [ 1957; Jacobs and Walton, 19761. Minor anomalies were defined as physi- cal defects of minimal or no medical importance [Kalter and Warkany, 19831. Malformations were defined as

congenital structural abnormalities with surgical, med- ical, or cosmetic importance [Leppig et al., 1987; Kalter and Warkany, 19831 and they were graded as severe, moderate, or mild. Length was measured with a rigid board; OFC and other measurements were made with a new cloth tape.

Data were obtained by non-examiners about compli- cations during pregnancy and delivery, exposure to pos- sible teratogens during pregnancy, growth parameters a t birth, and postnatal problems.

RESULTS During the study period, 343 babies were born alive;

16 spontaneous abortions had occurred and 8 abortions had been induced for genetic abnormalities. The sponta- neous abortuses had not been examined. One hundred eighty-nine of the liveborn children were evaluated; 144 families could not be contacted; and the mothers of 10 refused to participate (Table I). Participation rate of those children who could be contacted was 95% and of the total group was 55% (1891343). The loss of contact often occurred with women whose infants were already 1-2 years old. Refusal to participate was usually due to the inconvenience of coming to the medical center be- cause of distance or the mother's work schedule; none of these mothers reported any malformations and all gave us permission to contact their child's pediatrician for confirmation of normality, although this was not done.

Of the 189 children evaluated, 95 (50.2%) had been exposed to CVS, 87 (46%) to amniocentesis, and 7 (3.8%) to both [Table 11. The latter group will be described, but not included, in the calculation of the frequency of mal- formations.

In 93% of all cases, prenatal testing had been done because of advanced maternal age (Table 11): 90.5% in

TABLE I . Children in Study

Male Female Total %

No. evaluated (1) cvs 57 38 95 50.2 i i j Amniocentesis 41 46 87 46.0 (3) CVS and 3 4 7 3.8

101 88 189 100.0 amniocentesis

No. refusing evaluation (1) cvs 6 60 (2) Amniocentesis 4 40

10 100

TABLE 11. Reasons for Prenatal Diagnosis

cvs Amniocentesis Total I Family history (FH) genetic disorder

Father-balanced translocation 1 1 FH trisomy 21 3" 2 FH trisomy 18 1" Both parents P-thalassemia carriers 2 1

Tay-Sachs carriers 1 Sickle cell carriers 1

9(9.5%) 4(4.6%) 13(7.1%) I1 Advanced maternal age (AMA) 86(90.5%) 83(95.4%) 169(92.9%) "Also AMA in 1 mother in each group.

368 Kaplan et al.

the CVS group and 95.4% in the amniocentesis group. The remaining pregnancies were a t risk for genetic dis- orders because of previous trisomies, paternal balanced translocations or both parents being carriers for severe autosomal recessive disorders. In general the mothers were middle class, well nourished, and denied using illegal drugs.

Ninety-four percent of the children were white, 3% black, and 3% Oriental or Asian, comparable with the ethnic composition of the Montreal population. At the time of evaluation, the children ranged in age from 2 weeks to 28 months, with the distribution in both groups being similar (Table 111). Four women in the CVS group delivered prematurely between 33 and 35 weeks. The one premature delivery in the amniocentesis group oc- curred a t 27 weeks, presumably due to placenta previa.

TABLE 111. Ages (Months) of Children Evaluated

cvs Amniocentesis Males 0.5-26.0 0.25-27.0 Females 0.5-28.0 1.0-28.0

Mean 7.4 8 . 3 Median 13.75 13.88

There were 128 children who had one or more minor anomalies without major abnormalities: 58 of 95 (60%) in the CVS and 70 of 87 (80%) in the amniocentesis groups (Table IV). Half the children in each group had only one minor anomaly. The most frequent minor anomaly was epicanthic fold(s), regardless of age. The children with epicanthic folds were between 2 and 14 months with the exception of a boy 24 months old who was exposed to CVS and who also was the only one of the 6 Oriental children to have true epicanthic folds. Abnor- mal configuration of the external helices, small chin, and clinodactyly were the other minor anomalies seen fairly often (Table IV).

Malformations and dysplasias were noted in 26 chil- dren, involving 17.8% (17195) who had been exposed to CVS and 10.3% (9187) to amniocentesis (Table V). In the CVS group, none of the malformations or dysplasias were severe, 4 were moderate (ventricular septa1 defect [VSD] which closed spontaneously [one child], hydrocele [21, Duane anomaly [l l) , and 13 were mild. Twelve were dysplasias [strawberry hemangiomal, and one was post- axial polydactyly in a white child. One boy in the CVS group had an unclassified multiple congenital anomaly/ mental retardation (MCAIMR) syndrome, with a nor- mal karyotype confirmed after birth. At age one year he

TABLE IV. Infants With Minor Anomalies

1 Minor 2 Minor

>3 Minor

Total

cvs ~

Male Female Male (Nos.) (Nos.) (Nos.)

Amniocentesis Female (Nos.)

11 13 6

30 ~

58/95 [60%1

13 9 6

28 -

16 7

11 34

~

13 14

36 3

70187 [SO%] __ -~ ~-

CVS Amniocentesis (Nos.) (Nos.)

~~ ~~~

Types of minor malformations Epicanthic folds 30 31 Abnormal configuration of helices 16 18 Small chin 7 10 Wide alveolar ridge 15 11 Capillary hemangiomas: scalp & face 8 10

Clinodactyly 5 9 Single palmar crease 2 5

elsewhere 8 6

Cafe-au-lait spots (<5) 13 6

TABLE V. Infants with Malformations CVS Amniocentesis ~ ~ _ _ _ _ _

3 a . b Hemangioma (cavernous “strawberry”) 12 Sturge Weber dysplasia 0 1 Craniosynostosis 0 2 Polydactyly (postaxial) 1 0 Cardiac: VSD 1 1“ Hydrocele 2 2 Blocked tear duct 0 1 Duane anomaly 1 0 Total Excluding cavernous hemangiomas 5 (5.2%) 7 (8.0%) Including cavernous hemangiomas 17 (17.8%) 10 (10.3%)

a 1 child had 2 malformations. bP<0.024 by chi-square test.

Malformations Following CVS and Amniocentesis 369

had moderate developmental delay; Duane anomaly; coarse facies with wide alveolar ridge, protruding tongue, epicanthic fold, anteverted nares, prominent maxilla, protruding ears with abnormal antihelix; um- bilical hernia; sacral dimple; and 3 hemangiomas.

In the subgroup of 9 children exposed to CVS for rea- sons other than advanced maternal age, 3 had dyspla- sias (strawberry hemangiomas) and minor anomalies (included in the previous rates.) In 2 cases, the parents of these children were heterozygotes for p-thalassemia and the third couple was heterozygous for sickle cell disease. One boy, whose father carried a balanced trans- location, had a normal karyotype, 46XY , and epicanthic folds.

Of the malformations and dysplasias identified in the 9 (10.3%) children exposed to amniocentesis, one was potentially severe (Sturge-Weber dysplasia), 5 were moderate (VSD, hydroceles, blocked tear duct, and cra- niosynostosis), and the others were mild dysplasias (hemangiomas). In the subgroup of 4 children exposed to amniocentesis for genetic reasons and not advanced ma- ternal age, there were no malformations; 3 each had a minor anomaly. One boy carried the same balanced translocation as his father and had 3 minor anomalies when examined a t 11 months.

A third group of 7 children who had been exposed to both CVS and amniocentesis was excluded from these calculations. None of the 3 boys and 4 girls had any malformations. Three children had 2 or more minor anomalies.

DISCUSSION CVS is a recently developed technique for obtaining

fetal tissue for prenatal diagnosis. The safety for the mother and fetus has been tested in the randomized Canadian Collaborative Trial and the total fetal loss rate following the procedure appears to be similar to that following amniocentesis. Similar results have been reported in a non-randomized trial by Rhoads et al. [1989]. Many obstetricians and pregnant women have been reluctant to employ CVS because of a fear that disturbing the chorion in early pregnancy may perturb fetal development and cause malformations [Fahy and Lippman, 19881. The possibility of litigation blaming CVS for congenital malformations exists. To determine whether these were valid concerns, we evaluated the infants born to women having CVS andlor amniocen- tesis as part of a controlled randomized trial. The pro- portion of children with minor anomalies but no malfor- mations in the CVS group (60%) was similar to that in the amniocentesis group (80%). The rates appear high but it should be noted that 50% of the anomalies were trivial, e.g., epicanthic folds (30% children in CVS group, 31% in amniocentesis group). The frequency of minor anomalies was similar to, although somewhat higher than, that in Leppig’s study of 4,305 children. One reason could be that, unlike Leppig et al. [19871, we included, as minor anomalies, some hemangiomas (which are dysplasias), and pigmentary changes, such as cafe-au-lait spots. Another factor was the age of the children studied. The children in our study were exam- ined at older ages (2 weeks to 28 months) than those in

Leppig’s group who were examined in the first week of life. It is possible that more minor anomalies could have become apparent in our older group. In addition, most cavernous (strawberry) hemangiomas would appear only several weeks or months after birth [Jacobs and Walton, 19761.

However, the differences in age should not alter com- parisons of the malformation rate between Leppig’s and our study. The frequency of total malformations in the CVS group appears higher than in the amniocentesis group, but the difference is not significantly different. Furthermore, there was no recurrent pattern of abnor- malities in either group. If the presence of hemangiomas (which are actually dysplasias) is excluded from the analysis, the overall frequencies of malformations de- crease in both the CVS and amniocentesis groups, with the frequency in the CVS group (5.2%) being lower than in the amniocentesis group (8.0%). Although none of the malformations was life-threatening, a few children re- quired surgery for repairs of hydroceles and hernias, including one in a prematurely born infant. The ven- tricular septa1 defects closed spontaneously in both chil- dren. The Duane anomaly, considered a moderately severe malformation, occurred in a child with an un- classified MCA/MR syndrome.

A large number (1441343) of children could not be located for inclusion in the study despite attempts through various routes. The loss of contact often oc- curred with children already 1-2 years old; a high pro- portion of Montrealers live in rental dwellings and it is not unusual for them to move fairly frequently. Sixty- eight of 144 (47%) of these children had been exposed to CVS, 701144 (49%) to amniocentesis, and 3/144 (2%) to both procedures; 3 patient files were unretrievable. Six of the 10 children whose mothers refused participation in the study had been exposed to CVS and 4 to amniocen- tesis, all because of advanced maternal age. While none were acknowledged to have malformations, it is possible that minor anomalies and hemangiomas could have been overlooked. Although these losses to followup may decrease the power of this study, they appear to be non- differential and should not bias comparisons between the CVS and amniocentesis groups.

If any significant congenital abnormalities were to occur due to CVS they would reflect interference in development after the 10th week. By that time, the critical periods of development for most organs, except the central nervous system, have been passed. However, the CVS might have caused disruptions or deformations which would have altered previously normal organ de- velopment. None of the abnormalities in the children studied could be attributed to these pathological pro- cesses, particularly disruptions. We did observe a higher frequency of dysplasia (cavernous strawberry hemangi- omas) in the children in the CVS group, but the small numbers preclude us from determining whether this is significant. Neither group had a higher frequency than in the general population (8.7%-10.1%) [Jacobs and Walton, 19761.

In conclusion, evaluation of a group of liveborn chil- dren, whose mothers participated in a randomized con- trolled trial, suggests that CVS is not associated with an

370 Kaplan et al.

increased frequency of major andlor minor congenital abnormalities compared with amniocentesis. Unfor- tunately, the limited size of the population which we could examine precluded identification, as statistically significant, of small differences in anomaly and malfor- mation rates between the groups. However, this study had sufficient power (80%) to detect a relative risk of 2.5 for total abnormalities in the CVS-exposed children, had the frequency of problems in the amniocentesis group approximated the rate of 40% observed by Leppig et al. [19871. Although these data should be confirmed in a larger group of children, they may be taken into account usefully when the decision is being made whether to use CVS rather than amniocentesis as a method of prenatal diagnosis.

ACKNOWLEDGMENTS We thank all the families who participated so enthusi-

astically in the trial; Dr. David Rosenblatt gave invalu- able support; Ms. Lola Cartier, Laura Arbour, Marlene Beauchamp, Susan Conacher, Mellisa Fahy, Orna Golan, Claude PrBvost, and Thea Weisdorf assisted in the evaluations; Drs. Mary Fujiwara and Ken Morgan gave important advice; Ms. Ruth King and Ms. Marie-

France Williams prepared the manuscript. The Cana- dian Collaborative Trial was supported by the Medical Research Council of Canada.

REFERENCES Canadian Collaborative CVS-Amniocentesis Clinical Trial Group

(1989): Multicentre randomized clinical trial of chorion villus sam- pling and amniocentesis. Lancet 1:l-6.

Fahy M, Lippman A (1988): Prenatal diagnosis and the Canadian Collaborative Randomized Trial of chorionic villi sampling: The physicians’ view. Am J Med Genet 31:953-961.

Jacobs AH (1957): Strawberry hemangiomas. Calif Med 86:8-10. Jacobs AH, Walton RG (1976): The incidence of birthmarks in the

neonate. Pediatrics 58:218-222. Kalter H, Warkany J (1983): Congenital malformations. N Engl J Med

308:424-431. Leppig KA, Werlerm Cann C, Cook CA, Holmes LB (1987): Predictive

value of minor anomalies. I. Association with major malformations. J Pediatr 1110:530-537.

Rhoads GG, Jackson LG, Schlesselman SE, de la Cruz FF, Desnick RJ, Golbus MS, Ledbetter DH, Lubs HA, Mahoney MJ, Pergament E, Simpson JL, Carpenter FLJ, Elias S, Ginsberg NA, Goldberg JD, Hobbins JC, Lynch L, Shiono PH, Wapner R J , Zachary JM (1989): The safety and efficacy of chorion villus sampling for early prenatal diagnosis of cytogenetic abnormalities. N Engl J Med 320:609-617.