PRENATAL DIAGNOSIS, VOL.
11,751-765 (199 1)
FIRST-TRIMESTER ANEUPLOIDY SCREENING USING SERUM HUMAN CHORIONIC GONADOTROPIN (hCG), FREE ahCG, AND PROGESTERONE PAUlL G. KRATZER, MITCHELL S. GOLBUS, SCOTT E. MONROE, DAVID E. FINKELSTEIN AND ROBERT N. TAYLOR R,eproductiveEndocrinology Cenier and Reproductive Genetics Unit,Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, U.S.A.
SUMMARY Jmmunochemical serum assays for human chorionic gonadotropin (hCG), the free ahCG subunit, and progesterone (p) were considered separately and in combination for their ability to screen for chromosomally abnormal pregnancies in the first trimester. Maternal serum was c'ollected from 141 women undergoing chorionic villus sampling at 9-12 menstrual weeks. Trisomy 21 pregnancies had significantly higher hCG levels, while trisomy 18 and 13 pregnancies had markedly lower hCG and progesterone levels than those of chromosomally normal pregnancies. However, the discrimination of normal from aneuploid pregnancies was poor with either hCG alone, progesterone alone, or free ahCG alone. Much improved discrimination was obtained by combining hCG, free ahCG, and P into an aneuploidy index [(P/hCG)(free ahCG/hCG)]. This index distinguished 9 out of 17 (53 per cent) of the trisoniy 21 pregnancies, while only misidentifying 5 out of 112 (4.5 per cent) of the normal pregnancies. The aneuploidy index thus appears promising as a first-trimester biochemical screen for aneuploid pregnancies. KEY woms
First-trimester aneuploidyscreening Aneuploidy index
INTRODUCTION A significant proportion of children are born with birth defects, including 0.4 per cent of newborns who have birth defects resulting from chromosomal abnormalities. The majority of chromosomal abnormalities are due to trisomies, which incre,asein frequency with advancing maternal age. This association with maternal age forms part of the current screening basis for prenatal diagnosis, in which all womlen who will be at least 35 years old at delivery are offered prenatal diagnosis to determine fetal chromosomal constitution (Frigoletto and Little, 1988). Younger wom'en are offered amniocentesis if their risk of a chromosomal abnormality, as determined by maternal serum alpha-fetoprotein (MSAFP) screening, is as great as that of a 35-year-old woman (Palomaki and Haddow, 1987). However, using maternal age to screen is inefficient in two senses. Firstly, screening by maternal age results in offering amniocentesis to only about 20-25 per cent of the women who Addressee for correspondence: Robert N. Taylor, M.D., Ph.D., W. M. Keck Laboratory of Reproductive and Developmental Molecular Biology, Department of Obstetrics, Gynecology and Reproductive Sciences, Room M-1489, University of California, San Francisco, San Francisco, CA 94143-0132, U.S.A.
0197-3851/91/100751-15$07.50 @ 1991 by John Wiley & Sons, Ltd.
Received 28 January 1991 Accepted 15 April 1991
752
P. G. KRATZER ETAL.
deliver aneuploid infants (Adams et al., 1981). This low yield is due to the much larger number of births to women younger than 35 compared with the number of births to women older than 35, even though the frequency of aneuploidy in younger women is less. Secondly, after performing a screening test, a definitive test needs to be perfonned. Preferably, the definitive tests, amniocentesis and chorionic villus sampling (CVS), should be offkred to only a select group of patients. It would be desirable, therefore, to have a screening test that is more efficient than maternal age alone in predicting aneuploidy, and that could be offered to women of all ages. As fetal and placental development are abnormal in aneuploid conceptuses (Philippe and Boue, 1969; Honore et al., 1976; Kuhlmann et al., 1990), one might expect abnormal expression of fetoplacental products. Several recent papers have suggested that biochemical screening of second-trimester pregnancies may offer an improvement over using maternal age for screening. A low concentration of MSAFP was the first biochemical parameter to be associated with aneuploidy (Merkatz ef al., 1984). Subsequently, differences in the maternal serum concentrations of human chorionic gonadotropin (hCG) and the free a-subunit of hCG (Bogart et al., 1987) and uncalnjugated oestriol (Canick et al., 1988) have been reported between aneuploid and euploid gestations. Combining assays for MSAFP, unconjugated oestriol, hCG, and maternal age identified 60 per cent of aneuploids in one study, as opposed to only 2!0-25 per cent identified by maternal age, using the same 5 per cent cut-off level of false positives (Wald et al., 1988b).Other studies have confirmed the usefulness of MSAFP, hCG, and unconjugated oestriol in screening for aneuploidy in the second trimester (Osathanondh et al., 1989; Hey1 et al., 1990; MacDonald et al., 1991). However, the value of including unconjugated oestriol has been questioned (Macri et a!., 1990). Performing a biochemical screen in the first trimester would have the additional benefit of allowing a confirmation to be made by CVS, and the decision whether to terminate the pregnancy at an earlier stage of pregnancy when this decision is emotionally less traumatic and termination is technically safer. Unfortunately, previous reports on biochemical testing in the first trimester have been less encouraging than those in the second trimester. Potentially useful differences have been reported for MSAFP (Milunsky et al., 1988; Cuckle et at., 1988; Mantingh et al., 1989) and unconjugated oestriol (Cuckle c’t al., 1988), although the overlap between normal and aneuploid pregnancies appeared greater than in the second trimester. Furthermore, it has been reported that hCG (Cuckle et al., 1988;Bogart er a].,1989;Brock et al., 1990) and free ahCG (Bogart et al., 1989) do not differ between normal and aneuploid pregnancies in the first trimester. We have assayed first-trimester serum samples for the placental products hCG, free ahCG, and progesterone (P) and present evidence that a combination of these assays appears useful as an early, non-invasive screen for aneuploidy. MATERIALS AND METHODS Pa tien ts The retrospective study sample consisted of a primary group of 141 patients at 9-12 menstrual weeks (10.7k0.1, mean+_SEM)at the University of California, San Francisco (UCSF) who desired a CVS procedure to determine the conceptus’s
FIRST-TRIMESTERANEUPLOJDY SCREENING
753
chrornosome constitution because of advanced maternal age, and of a secondary group of 18 patients who had an elective abortion at 16-20 (1 8.4 & 0.3) menstrual weeks. Menstrual age was calculated from the last normal menstrual period and was confirmed by ultrasonography. All the pregnancies were singleton pregnancies. Among the CVS patients, chromosome studies identified 17 fetuses with trisomy 21, 7 witlh trisomy 18, 3 with trisomy 13, and 1 each with trisomy 16 and trisomy 22. Control samples from pregnancies with chromosomally normal fetuses were then identified that followed each trisomy and were matched with respect to menstrual age. The second-trimester samples were obtained from patients undergoing elective therapeutic abortions for the following indications: (1) 12 patients desired a therapeutic abortion and carried a presumed normal fetus; (2) 3 patients had a trisomy 21 fetus detected by genetic amniocentesis; and (3) 3 patients had fetuses with neural tube defects (2 with anencephaly and 1 with spina bifida). Serum samples were drawn after obtaining informed consent before each procedure in accordance with the guidelines of the UCSF and Children's Hospital of San Francisco Human Subjects Committees. Assays
The sera were stored frozen at - 70°C before being assayed blindly. No weight or race corrections'were made. The sera were evaluated using the following assays: (1) hCG was assayed by a double antibody radio-immunoassay that detected both intact hCG and free D C G subunits. The hCG levels were expressed as international units per millilitre (IU/ml) using the Second International Standard for hCG. The intra- and inter-assay coefficients of variation were 8 and 14 per cent, respectively. (2) The free a-subunit of hCG was assayed by a radio-immunoassay which employed a monoclonal antibody to the a-subunit of hCG obtained from Biomierica, Inc. (Newport Beach, CA). This assay had
11,751-765 (199 1)
FIRST-TRIMESTER ANEUPLOIDY SCREENING USING SERUM HUMAN CHORIONIC GONADOTROPIN (hCG), FREE ahCG, AND PROGESTERONE PAUlL G. KRATZER, MITCHELL S. GOLBUS, SCOTT E. MONROE, DAVID E. FINKELSTEIN AND ROBERT N. TAYLOR R,eproductiveEndocrinology Cenier and Reproductive Genetics Unit,Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, U.S.A.
SUMMARY Jmmunochemical serum assays for human chorionic gonadotropin (hCG), the free ahCG subunit, and progesterone (p) were considered separately and in combination for their ability to screen for chromosomally abnormal pregnancies in the first trimester. Maternal serum was c'ollected from 141 women undergoing chorionic villus sampling at 9-12 menstrual weeks. Trisomy 21 pregnancies had significantly higher hCG levels, while trisomy 18 and 13 pregnancies had markedly lower hCG and progesterone levels than those of chromosomally normal pregnancies. However, the discrimination of normal from aneuploid pregnancies was poor with either hCG alone, progesterone alone, or free ahCG alone. Much improved discrimination was obtained by combining hCG, free ahCG, and P into an aneuploidy index [(P/hCG)(free ahCG/hCG)]. This index distinguished 9 out of 17 (53 per cent) of the trisoniy 21 pregnancies, while only misidentifying 5 out of 112 (4.5 per cent) of the normal pregnancies. The aneuploidy index thus appears promising as a first-trimester biochemical screen for aneuploid pregnancies. KEY woms
First-trimester aneuploidyscreening Aneuploidy index
INTRODUCTION A significant proportion of children are born with birth defects, including 0.4 per cent of newborns who have birth defects resulting from chromosomal abnormalities. The majority of chromosomal abnormalities are due to trisomies, which incre,asein frequency with advancing maternal age. This association with maternal age forms part of the current screening basis for prenatal diagnosis, in which all womlen who will be at least 35 years old at delivery are offered prenatal diagnosis to determine fetal chromosomal constitution (Frigoletto and Little, 1988). Younger wom'en are offered amniocentesis if their risk of a chromosomal abnormality, as determined by maternal serum alpha-fetoprotein (MSAFP) screening, is as great as that of a 35-year-old woman (Palomaki and Haddow, 1987). However, using maternal age to screen is inefficient in two senses. Firstly, screening by maternal age results in offering amniocentesis to only about 20-25 per cent of the women who Addressee for correspondence: Robert N. Taylor, M.D., Ph.D., W. M. Keck Laboratory of Reproductive and Developmental Molecular Biology, Department of Obstetrics, Gynecology and Reproductive Sciences, Room M-1489, University of California, San Francisco, San Francisco, CA 94143-0132, U.S.A.
0197-3851/91/100751-15$07.50 @ 1991 by John Wiley & Sons, Ltd.
Received 28 January 1991 Accepted 15 April 1991
752
P. G. KRATZER ETAL.
deliver aneuploid infants (Adams et al., 1981). This low yield is due to the much larger number of births to women younger than 35 compared with the number of births to women older than 35, even though the frequency of aneuploidy in younger women is less. Secondly, after performing a screening test, a definitive test needs to be perfonned. Preferably, the definitive tests, amniocentesis and chorionic villus sampling (CVS), should be offkred to only a select group of patients. It would be desirable, therefore, to have a screening test that is more efficient than maternal age alone in predicting aneuploidy, and that could be offered to women of all ages. As fetal and placental development are abnormal in aneuploid conceptuses (Philippe and Boue, 1969; Honore et al., 1976; Kuhlmann et al., 1990), one might expect abnormal expression of fetoplacental products. Several recent papers have suggested that biochemical screening of second-trimester pregnancies may offer an improvement over using maternal age for screening. A low concentration of MSAFP was the first biochemical parameter to be associated with aneuploidy (Merkatz ef al., 1984). Subsequently, differences in the maternal serum concentrations of human chorionic gonadotropin (hCG) and the free a-subunit of hCG (Bogart et al., 1987) and uncalnjugated oestriol (Canick et al., 1988) have been reported between aneuploid and euploid gestations. Combining assays for MSAFP, unconjugated oestriol, hCG, and maternal age identified 60 per cent of aneuploids in one study, as opposed to only 2!0-25 per cent identified by maternal age, using the same 5 per cent cut-off level of false positives (Wald et al., 1988b).Other studies have confirmed the usefulness of MSAFP, hCG, and unconjugated oestriol in screening for aneuploidy in the second trimester (Osathanondh et al., 1989; Hey1 et al., 1990; MacDonald et al., 1991). However, the value of including unconjugated oestriol has been questioned (Macri et a!., 1990). Performing a biochemical screen in the first trimester would have the additional benefit of allowing a confirmation to be made by CVS, and the decision whether to terminate the pregnancy at an earlier stage of pregnancy when this decision is emotionally less traumatic and termination is technically safer. Unfortunately, previous reports on biochemical testing in the first trimester have been less encouraging than those in the second trimester. Potentially useful differences have been reported for MSAFP (Milunsky et al., 1988; Cuckle et at., 1988; Mantingh et al., 1989) and unconjugated oestriol (Cuckle c’t al., 1988), although the overlap between normal and aneuploid pregnancies appeared greater than in the second trimester. Furthermore, it has been reported that hCG (Cuckle et al., 1988;Bogart er a].,1989;Brock et al., 1990) and free ahCG (Bogart et al., 1989) do not differ between normal and aneuploid pregnancies in the first trimester. We have assayed first-trimester serum samples for the placental products hCG, free ahCG, and progesterone (P) and present evidence that a combination of these assays appears useful as an early, non-invasive screen for aneuploidy. MATERIALS AND METHODS Pa tien ts The retrospective study sample consisted of a primary group of 141 patients at 9-12 menstrual weeks (10.7k0.1, mean+_SEM)at the University of California, San Francisco (UCSF) who desired a CVS procedure to determine the conceptus’s
FIRST-TRIMESTERANEUPLOJDY SCREENING
753
chrornosome constitution because of advanced maternal age, and of a secondary group of 18 patients who had an elective abortion at 16-20 (1 8.4 & 0.3) menstrual weeks. Menstrual age was calculated from the last normal menstrual period and was confirmed by ultrasonography. All the pregnancies were singleton pregnancies. Among the CVS patients, chromosome studies identified 17 fetuses with trisomy 21, 7 witlh trisomy 18, 3 with trisomy 13, and 1 each with trisomy 16 and trisomy 22. Control samples from pregnancies with chromosomally normal fetuses were then identified that followed each trisomy and were matched with respect to menstrual age. The second-trimester samples were obtained from patients undergoing elective therapeutic abortions for the following indications: (1) 12 patients desired a therapeutic abortion and carried a presumed normal fetus; (2) 3 patients had a trisomy 21 fetus detected by genetic amniocentesis; and (3) 3 patients had fetuses with neural tube defects (2 with anencephaly and 1 with spina bifida). Serum samples were drawn after obtaining informed consent before each procedure in accordance with the guidelines of the UCSF and Children's Hospital of San Francisco Human Subjects Committees. Assays
The sera were stored frozen at - 70°C before being assayed blindly. No weight or race corrections'were made. The sera were evaluated using the following assays: (1) hCG was assayed by a double antibody radio-immunoassay that detected both intact hCG and free D C G subunits. The hCG levels were expressed as international units per millilitre (IU/ml) using the Second International Standard for hCG. The intra- and inter-assay coefficients of variation were 8 and 14 per cent, respectively. (2) The free a-subunit of hCG was assayed by a radio-immunoassay which employed a monoclonal antibody to the a-subunit of hCG obtained from Biomierica, Inc. (Newport Beach, CA). This assay had
