Hum Genet (1982) 61:318-324 © Springer-Verlag 1982 Original Investigations Rates of Trisomies 21, 18, 13 and Other Chromosome Abnormalities in About 20 000 Prenatal Studies Compared with Estimated Rates in Live Births Dina M. Schreinemachers ~*, Philip K. Cross11 *, and Ernest B. Hook 1.2• Birth Defects Institute*, Center for Laboratories and Research, New York State Department of Health, Albany, NY 12201, USA 2Department of Pediatrics, Albany Medical College, Albany, NY 12208, USA Sqmmary. Data were analyzed on the results of 19 675 prenatal cytogenetic diagnoses reported to two chromosome registries on women aged 35 or over for whom there was no known cytogenetic risk for a chromosome abnormality except parental age. The expected rates at amniocentesis of 47,+21; 47,+18; 47,+13; XXX; XXY; XYY; and other clinically significant cytogenetic defects by maternal age were obtained from a regression analysis on the observed rates, using a first degree exponential model. After an adjustment for maternal age, these rates were compared with previously estimated rates by maternal age in live births. The rates of 47,+21 at amniocentesis and live birth are approximately parallel, with the latter about 80% of the amniocentesis rates. The rates of 47,+18 at amniocentesis and live birth are approximately parallel, with the live birth rates about 30% of the amniocentesis rates, consistent with high fetal mortality of 47,+ 18 after amniocentesis. The rates of 47,+ 13 at amniocentesis indicate an increase in maternal age that is not as marked as that previously estimated in live births. The rates at amniocentesis for XXX and XXY increase with maternal age, with the rates of XXY almost identical to those estimated previously in live births, suggesting no late fetal mortality of XXY. The rates of XYY show a slight decrease with maternal age also consistent with little late fetal mortality of XYY. No consistent trend with age is seen for the pooled group of other clinically significant defects. Introduction While it has been noted that the rates of 47,+21; 47,+18; and 47,+13 at amniocentesis are higher than the rates in live births (Hook 1978, 1980; Hook et al. 1979), only recently have sufficient data become available to enable characterization of changes with maternal age, in rates at amniocentesis, as has been done in live births. We report here regression models for maternal age-specific rates for chromosome abnormalities diag- nosed prenatally. The rates by 1-year intervals from age 35 and up were calculated from data on 19675 prenatal diagnoses, obtained by pooling the results reported to two chromosome registries and compared with previously published estimated rates in live births. For those abormalities whose rate at amniocentesis is uniformly higher than the rate at live birth, the data provide * Current State affilitations of the authors are: Bureau of Maternal and Child Health, New York State Department of Health, 2509 Tower Mall, Albany, NY 12237, USA evidence that is at least consistent with the occurrence of spontaneous fetal deaths after that time in gestation when amniocentesis is usually performed. Materials and Methods The amniocentesis data were obtained by pooling the results from two series: the Interregional Cytogenetic Register System (Prescott et al. 1978), 3477 cases in 1977-1980, and the New York State Chromosome Registry (Hook et al. 1981), 16 198 cases in 1977-1980, the data of which were contributed by 21 New York State and 7 New England centers (see Acknowledgements). Only those cases were included in this study where the maternal age was 35 or over at the time of amniocentesis and the reason for study was elevated parental age or a reason other than a known cytogenetic risk, such as cytogenetic evaluation of amniotic fluid obtained for study of alpha fetoprotein or inborn errors of metabolism (see also Addendum). The rates per 1000 amniocenteses by maternal age were calculated for 47,+21; 47,+18; 47,+13; XXX; XXY; XYY; "all other" genotypes that were regarded as clinically significant, and the total number of these cytogenetic abnormalities pooled together. To adjust for the fact that amniocentesis is usually at 16-20 weeks and birth around 40 weeks of gestation, 0.4 years was added to the maternal ages at amniocentesis. This enables fair graphic comparison with estimated rates in live births. The data on live births for this comparison were those estimated by Hook (1981). If a range of rates was given in this reference, the midpoint of each range was selected for the live birth rate for comparison here. These live birth rates were derived from various studies in the late 1960s and early 1970s, and adjustments in these were for rates of 47,+21 and 47,+18 for a temporal increase in Down's syndrome rate in 1970-1974 compared with 1965-1969. [See Hook (1981) for further details.] The rate for 47,+13 in live births used in the analysis in this paper was obtained by subtracting the rate of translocation Patau syndrome in live births, 0.04 per 1000 (Hook 1980), from the midpoint of the ranges of the maternal age-specific rates given for Patau syndrome (trisomies + translocation) by Hook (1981). Rates were plotted on semilogarithmic paper. Inspection of the graphed rates of 47,+21; 47,+18; XXX; and XXY diagnosed at amniocentesis suggested an exponential increase with mater- nal age. A regression analysis was undertaken using the model y = exp (a +bx), where x is the maternal age at time of expected live birth and y, the rate of the cytogenetic defect at age x. The 0340-6717/82/0061/0318/$ 01.40