|
|
| |||||||||||||||||||||||||||||||||||||||
Background As maternal age advances, the risk of fetal Down's syndrome increases. Pregnant women 35 years of age or older are routinely offered amniocentesis because of this risk. Recently, maternal serum markers have been reported to be useful in screening for Down's syndrome, primarily in younger women. We therefore investigated whether offering amniocentesis only to selected women 35 years of age or older who were identified by screening measurements in serum might prove a useful alternative to the current practice.
Methods We studied 5385 women with singleton pregnancies who were 35 years of age or older and were undergoing routine amniocentesis. Along with information about the pregnancy, we obtained a serum sample for measurement of alpha-fetoprotein, unconjugated estriol, and human chorionic gonadotropin. Individual estimates of the risk of Down's syndrome in the fetus were calculated for each pregnancy before the karyotype was known.
Results If amniocentesis had been reserved for the women calculated to have a risk greater than 1 in 200 of having a fetus with Down's syndrome, 48 of the 54 cases of Down's syndrome (89 percent) would have been identified; 25 percent of the unaffected pregnancies would also have been identified as being at high risk for Down's syndrome (false positives). Seven of 15 fetuses (47 percent) with other trisomies, 11 of 25 (44 percent) with sex aneuploidy, and 1 of 9 (11 percent) with miscellaneous chromosomal abnormalities would also have been detected. In practice, such screening would have made 75 percent of the amniocenteses unnecessary, along with a proportion of the amniocentesis-associated fetal losses. If the cutoff for the risk of Down's syndrome were set higher than 1 in 200, both the rate of detection and the false positive rate would be lower. Conversely, these rates would be higher if the cutoff were set lower.
Conclusions Prenatal screening of serum to generate individual estimates of the risk of Down's syndrome in the fetus can provide a basis for decision making in the cases of women 35 years of age or older, as it does in younger pregnant women, and is an alternative to current testing practices. .
In 1984, maternal serum alpha-fetoprotein concentrations were found to be lower, on average, in the presence of Down's syndrome,3 and it was demonstrated that this biochemical measurement could be used as a screening test in the second trimester4. Screening for Down's syndrome could thus be made available to younger women. The discriminatory power of this screening test in women younger than 35 years of age was estimated to be comparable to the use of maternal age alone as a screening test. Subsequent prospective intervention trials in the United States confirmed the reliability of that estimate5,6.
In 1987 and 1988, two additional serum markers, the levels of human chorionic gonadotropin7 and unconjugated estriol,8 were identified as significantly increasing the sensitivity of screening for Down's syndrome when combined with the measurement of alpha-fetoprotein. The risk of fetal Down's syndrome becomes greater as serum concentrations of human chorionic gonadotropin increase and serum concentrations of unconjugated estriol decrease. Screening protocols using these three tests and maternal age were subsequently evaluated in the United States and Europe9,10,11,12,13. The higher rate of detection of Down's syndrome that was achieved with this approach made a compelling case that biochemical screening might prove useful for pregnant women 35 years of age or older. We undertook this prospective study to evaluate biochemical screening in a group of women undergoing amniocentesis and chromosome analysis because of their age.
Methods
Recruitment of Subjects
The study protocol was approved by the institutional review committees at the study and recruitment centers and by the institutional review committee of the state of California. Between December 1990 and October 1992, 5425 women at least 35 years of age with singleton pregnancies were enrolled in the study at 14 prenatal diagnostic centers (see the Appendix). All the women gave informed consent before entry into the study. In the cases of 40 of the women, the indication for amniocentesis was a positive serum alpha-fetoprotein screening test; those women were excluded from the analysis. The remaining 5385 women provided information about their date of birth, race or ethnic background, weight, and other demographic and pregnancy-related factors. Estimates of gestational age were available for over 99 percent of the pregnancies, based on both the date of the last menstrual period and results of ultrasonography. Among these women, 71 percent were non-Hispanic whites, 13 percent were Hispanic whites, 10 percent were Asian, and 6 percent were black.
A blood sample was obtained and centrifuged, and the serum was refrigerated until just before shipment. The sample was shipped overnight within one to five days after its collection to a central laboratory in Scarborough, Maine, for analysis of alpha-fetoprotein, unconjugated estriol, and human chorionic gonadotropin (collectively referred to hereafter as the biochemical markers). The analyses were performed within two days after the arrival of the sample, but the results were not made available for clinical purposes. The coordinating center in Berkeley, California, subsequently obtained the results of the chromosomal studies for each pregnancy and transmitted the information to Scarborough for storage and analysis. Among the 5385 women who met the eligibility criteria, 54 cases of Down's syndrome and 49 other chromosomal abnormalities in the fetuses were identified on the basis of amniotic-fluid karyotype. Balanced translocations, pseudo-mosaicisms, normal variants, and common inversions were included among the 5282 pregnancies with a normal karyotype.
Serum Measurements
Alpha-fetoprotein was measured in serum with an in-house assay14. Unconjugated estriol was measured with the Amerlex-M radioimmunoassay (RIA) kit specific for unconjugated estriol (modified to increase sensitivity15), or the Amerlex-M high-sensitivity unconjugated estriol RIA kit (Eastman Kodak, Rochester, N.Y.). Human chorionic gonadotropin was measured with the Amerlex-M extended-range HCG RIA kit or the Amerlex-M 
-hCG RIA kit (Eastman Kodak). The coefficient of variation between assays was 8 percent or less for all three assays.
Statistical Analysis
At the end of the study, median values were calculated for each of the three analytes, based on estimates of gestational age derived from the date of the last menstrual period. Multiples of these median values were then calculated from the results obtained for each woman. The values for all three analytes were adjusted for maternal weight,16 and serum alpha-fetoprotein values were adjusted for race17 and the presence of insulin-dependent diabetes mellitus18. The serum unconjugated estriol values were adjusted for Hispanic and Asian women by dividing the values by 1.12 and 1.17, respectively (the median concentrations found in those groups). Individual risks of fetal Down's syndrome were then calculated from each woman's estimated age at delivery, in combination with the biochemical markers, with published values and algorithms9,19,20. A more complete description of the calculation of risk is available elsewhere10.
Results
Efficiency of Screening for Down's Syndrome
The rate of detection of Down's syndrome (n = 54) and the corresponding false positive rate in the 5282 pregnancies without chromosomal abnormalities are shown in Figure 1. At 5 percent and 15 percent false positive rates, the rates of detection of Down's syndrome were 59 and 78 percent, respectively. At a false positive rate of 25 percent, the rate of detection was 89 percent; beyond that point, relatively large increases in the false positive rate yielded only small gains in detection. With the current practice of performing amniocentesis routinely in all women at least 35 years of age, all the cases of Down's syndrome are identified, but all the women with unaffected pregnancies also undergo amniocentesis.
|
It is routine practice to convert each woman's biochemical values, in combination with her age, into a composite estimate of the risk of Down's syndrome in her fetus. These estimates form the basis for decisions about what constitutes the high-risk population. A risk cutoff of 1 in 200 detected 89 percent of the 54 cases of Down's syndrome with a 25 percent rate of false positive results (the same rates shown in Figure 1). At a risk level of 1 in 100 the detection and false positive rates were 78 and 13 percent, respectively; at a risk of 1 in 300 they were 89 and 34 percent.
Efficiency of Biochemical Screening for Other Chromosomal Abnormalities
Table 1 shows the detection rate and the false positive rate when three different cutoff levels were used for the risk of chromosomal abnormalities other than Down's syndrome. Among the 15 fetuses with other autosomal trisomies, 7 (47 percent) were identified at a risk cutoff of 1 in 200. In addition, 25 fetuses had sex aneuploidy, 44 percent of which were also detected at that cutoff level. Thus, the screening protocol for Down's syndrome identified other important chromosomal abnormalities at a rate higher than the 25 percent that would be expected if there was no relation between those disorders and the assigned risk of Down's syndrome. Nine fetuses had miscellaneous chromosomal abnormalities, of which one (11 percent) was detected.
|
Table 2 shows the detection rates associated with three false positive rates for three combinations of markers used to screen for Down's syndrome. At a 25 percent false positive rate, serum alpha-fetoprotein measurements alone detect 70 percent of the cases of Down's syndrome. When serum human chorionic gonadotropin levels are added, 85 percent of the cases are identified. When all three measures are used, 89 percent are identified.
|
Figure 2 shows the estimated number of amniocenteses and the proportional number of amniocentesis-related fetal losses among a hypothetical group of 10,000 women 35 years of age or older, expressed as functions of the rate of detection of Down's syndrome. In this calculation, the prevalence of Down's syndrome was 1 per 100 (as found in this study) and the rate of amniocentesis-related fetal loss was assumed to be 5 per 1000 procedures1,21. In this example, 50 fetal losses occurred, and all 100 cases of Down's syndrome were identified, when all 10,000 women in this hypothetical group underwent amniocentesis as currently recommended (solid circle, Figure 2). If the multiple-marker screening protocol was used with the risk cutoff set at 1 in 200, 2500 amniocenteses would be performed, 13 estimated amniocentesis-related fetal losses would occur, and 89 of the cases of Down's syndrome would be detected (open circle, Figure 2). Detecting the 11 cases of Down's syndrome that were not identified by biochemical screening would require 7500 additional amniocenteses and the loss of 37 additional unaffected fetuses.
|
Assuming that the current average cost of genetic counseling, amniocentesis, and karyotyping in the United States is $1,000, universal amniocentesis in the hypothetical group of 10,000 women described above would cost $10 million and would detect all 100 cases of Down's syndrome. Assuming the average total cost of the three serum measurements to be $75, the total cost of biochemical screening of the same 10,000 women would be $750,000, and 2500 of those women would be identified as candidates for amniocentesis. The costs of diagnostic testing and counseling for those 2500 women would be $2.5 million. The combined costs of screening and diagnosis would therefore be $3.25 million. This second approach would identify 89 of the 100 cases of Down's syndrome. The cost per case detected in the second trimester with current practices would be $100,000, as compared with $36,500 for biochemical screening. The corresponding estimated costs per live-born infant with Down's syndrome would be $130,000 and $47,000, respectively (approximately 23 percent of fetuses with Down's syndrome are spontaneously aborted in the third trimester22). This analysis does not include the additional direct health care and social costs associated with caring for the estimated eight live-born children with Down's syndrome who would not be detected by biochemical screening; the remaining three would be spontaneously lost in the third trimester. Detecting these eight cases would require an additional expenditure of $6.75 million ($840,000 per case). When these figures are extrapolated to the estimated 380,000 pregnancies occurring each year in women 35 years of age or older in the United States, the reduction in diagnostic costs resulting from the use of multiple-marker biochemical screening would be approximately $250 million (assuming total participation).
Discussion
In this study, the rate of detection of Down's syndrome when we used multiple biochemical markers was 89 percent, with a 25 percent false positive rate, among women 35 years of age or older. Gestational ages were estimated on the basis of the dates of the last menstrual period. The use of gestational ages determined by ultrasonography did not substantially reduce the false positive rate, because incorrectly dated pregnancies account for a relatively small proportion of the high false positive rate in this age group (data not shown). In addition, adding serum measurements of unconjugated estriol to the screening protocol contributes less to its efficiency in this age group than it does among younger women. The rate of detection of Down's syndrome is lower in these women, providing greater opportunity to identify additional cases by measuring unconjugated estriol (Table 3).
|
One argument against replacing current practice with a policy based on biochemical screening is that the latter will detect only 9 of every 10 cases of Down's syndrome, whereas routine amniocentesis detects all 10 cases. This fact needs to be weighed, however, against the large reduction in the number of women requiring amniocentesis that can be achieved with biochemical screening and, as a consequence, the reduction in fetal losses related to the procedure.
An additional argument for maintaining current practice is that other chromosomal abnormalities, some of them medically important, are also detected by amniocentesis. In our study, 47 percent of the fetuses with other autosomal trisomies were identified when amniocentesis was offered to 25 percent of the women. A supplementary protocol specifically designed to determine the risk of trisomy 18 further increases this rate of detection, with a negligible increase in the number of amniocenteses24. These other autosomal trisomies are not compatible with life, and affected fetuses are usually lost spontaneously before term22. Forty-four percent of the fetuses with a sex aneuploidy were identified with multiple-marker biochemical screening. The detection of sex aneuploidy is generally viewed as a byproduct of prenatal chromosome analysis; the consequences of these abnormalities for health are not substantial and, in and of themselves, they do not justify amniocentesis25. When the diagnosis of a sex-chromosome aneuploidy is made, however, 50 percent or more of the pregnancies are electively terminated26,27. Among the nine fetuses with miscellaneous chromosomal abnormalities, an estimated 80 to 90 percent of those born alive would not have a mental or physical disability28.
A third argument in support of maintaining current practice is that the medical and social costs necessary to maintain patients with Down's syndrome are high. Although accurate measurement of these costs is difficult, few take issue with the validity of this claim. The considerable savings in diagnostic costs that can be achieved with multiple-marker biochemical screening will, as a consequence, be at least partially offset by the costs of maintaining the 10 percent of fetuses with Down's syndrome who are not identified prenatally.
On the basis of this study, three health care policy options might be considered for pregnant women 35 years of age or older: continue the current policy of universal amniocentesis, replace this policy with a policy of using multiple-marker biochemical screening, or allow women a choice between these two options. Deciding among these options involves weighing the advantages of screening (more accurate information about individual risks of Down's syndrome, fewer procedure-related fetal losses, and lower overall diagnostic costs) against the disadvantages (failure to detect 10 percent of the cases of Down's syndrome, failure to detect a proportion of other chromosomal disorders, and added costs for maintaining the children born in the missed cases).
Supported by a grant (MCJ-237014-02-2) from the Maternal and Child Health Bureau, Department of Health and Human Services, and from the National Institute of Child Health and Human Development.
Source Information
From the Foundation for Blood Research, Scarborough, Me. (J.E.H., G.E.P., G.J.K.); the California Department of Health Services, Genetic Disease Branch, Berkeley (G.C.C., L.S.L.); and the California Public Health Foundation, Berkeley (P.A.B.).
Address reprint requests to Dr. Haddow at the Foundation for Blood Research, Scarborough, ME 04074.
References
The centers and investigators that collaborated in this study are as follows: Alfigen-The Genetics Institute, Pasadena, Calif. -- J.C. Kelly and G.R. DeVore; University of Iowa Hospitals and Clinics, Iowa City -- S. Grant and K. Wenstrom; University of California, San Francisco -- M.S. Golbus and J.D. Goldberg; Pennsylvania Hospital, Philadelphia -- A.E. Donnenfeld; Cedars-Sinai Medical Center, Los Angeles -- D.E. Carlson and L.D. Platt; Loma Linda University Medical Center, Loma Linda, Calif. -- H. Brar and D. Abad; Children's Hospital Oakland, Oakland, Calif. -- A. Bastian and S. Sherman; Children's Hospital Los Angeles, Los Angeles -- R.E. Falk and K.A. Wendt; Kaiser Permanente of Southern California, Pasadena -- H.N. Bass and C. Francisco; Fetal Diagnostic Center, Laguna Hills, Calif. -- D.C. Lagrew, Jr., and C. Roquemore; Prenatal Diagnostic Center of Southern California, Los Angeles -- J. Williams III and C.H. Rubin; Jefferson Medical College, Philadelphia -- A. Johnson and R.J. Wapner; University of Tennessee, Memphis -- S. Elias and J.L. Simpson; and University of Connecticut Health Center, Storrs -- L. Ciarleglio and A. Craffey.
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Triple-Marker Screening of Serum for Down's Syndrome
Herrmann J., Haddow J. E., Palomaki G. E.
Extract |
Full Text
N Engl J Med 1994;
331:681-682, Sep 8, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
