- Email: [email protected]
First-trimester maternal serum α-fetoprotein screening for chromosome defects
Volume 159 Number 5
Structure-activity relationship of some steroidal hypnotic agents. J Med Chem 1968; 11: 117-25. 17. Backstrom T, Landgren S, Zetterlund B, et al. Effects of ovarian steroid hormones on brain excitability and their relation to epilepsy seizure variation during the menstrual cycle. In: Porter RJ, Mattson RH, Ward AA Jr, Dam M, eds. Advances in epileptology. XVth Epilepsy international symposium. New York: Raven Press, 1984:269-76. 18. H¢gskilde S, NeilsenJW, Carl P, S¢rensen MB. Pregnanolone emulsion. A new steroid preparation for intravenous anesthesia: an experimental study in mice. Anaesthesia I 987;42:586-90. 19. Heuser G. Induction of anesthesia, seizures and sleep by steroid hormones. Anesthesiology 1967;28:173-83.
Hypnotic effect of progesterone and its metabolites
20. Banks P, Peace CB. Enzyme inhibition by steroid anaesthetic agents derived from progesterone. Br J Anaesth 1985;57:512-4. 21. Majewska MD, Harrison NL, Schwartz RD, Barker JL, Paul SM. Steroid hormone metabolites are barbituratelike modulators of the GABA receptor. Science 1986; 232:1004-7. 22. Harrison NL, Majewska MD, Harrington JW, Barker JL. Structure-activity relationships for steroid interaction with the gamma-aminobutyric acid: a receptor complex. J Pharmacol Exp Ther 1987;241:346-53.
First-trimester maternal serum a-fetoprotein screening for chromosome defects Aubrey Milunsky, MB, BCh, DSc,a Jack Wands, MD," Bruno Brambati, MD,c Ida Bonacchi, MD, c and Karen Currie• Boston, Massachusetts, and Milan, Italy Low maternal serum cx-fetoprotein values during the second trimester of pregnancy are associated with an increased risk of Down syndrome in the fetus. In this study a sensitive, monoclonal-based radioimmunoassay for cx-fetoprotein was used to determine whether such an association also applies to the first trimester and if maternal serum cx-fetoprotein screening could successfully detect a significant number of pregnancies in which the fetus had a trisomy or other chromosome disorder. Sera were obtained prospectively from 540 women just before chorionic villus sampling for prenatal diagnosis of chromosome defects (largely because of advanced maternal age) at 8 to 12 weeks' fetal age and assayed for cx-fetoprotein under code without knowledge of the cytogenetic results. Eight of 27 (29.6%) of all serious chromosome defects were associated with low maternal serum a-fetoprotein values (,,;0.6 multiples of the median). Overall, 59 of 540 patients (10.9%) had maternal serum a-fetoprotein values ,,;0.6 multiples of the median, eight of whom had a fetus with a serious chromosome defect. Women whose maternal serum cx-fetoprotein value was .,,0.6 multiples of the median had one in eight odds of carrying a fetus with a trisomy and one in seven odds of the fetus having any serious chromsome defect. From this study of a group of women at higher risk, we conclude that first-trimester maternal serum cx-fetoprotein screening for chromsome defects is feasible. A prospective study to determine detection efficiency is now required of a consecutive routine pregnancy population in whom gestational age is determined by menstrual dates as is usually the case in clinical practice (AM J OesTET GYNECOL 1988;159:1209-13.)
Key words: Chromosome defects, maternal serum a-fetoprotein values, trisomy, multiples of the median
From the Center for Human Genetics and the Department of Pediatrics," Boston University School of Medicine, the Gastrointestinal Unit,' Department of Medicine, Massachusetts General Hospital and Harvard Medical School, and L. Mangiagalli Clinic,' First Institute of Obstetrics and Gynaecology, University of Milan. Supported by U.S. Public Health Service Grant HD20469. Received for publication November 14, 1987; revised Aprill 3, 1988; accepted May 19, 1988. Reprint requests: Aubrey Milunsky, MD, Center for Human Genetics and Department of Pediatrics, Boston University School of Medicine, 80 E. Concord, St., Boston, MA 02118.
Second-trimester maternal serum a-fetop1otein screening is now established as a valuable tool in the prenatal detection of neural tube and other congenital defects and in the recognition of high-risk pregnancies.' Most recently, determination of low maternal serum a-fetoprotein values has shown great promise in the detection of a significant number of pregnancies in which the fetus has Down syndrome. 2 · 3 This approach
1209
1210 Milunsky et al.
November 1988 Am J Obstet Gynecol
3500 3000 2500 2000 ::Ii CL
0
1500 1000
500
00
2
3
4
5
7
6
8
9
10
11
12
AFP Iu/ml
Fig. I. Standard curve for serum a-fetoprotein values <12 IU/ml.
Table I. Representative sample of first-trimester maternal serum a-fetoprotein cases assayed in duplicate Case identification No.
3836 3897 4072 3843 3844 3920 3842 4015 3959 38~6
3891 4014 3921 4002 4044
Weeks' gestation
8 8 8 9 9 9"
IO IO IO II II II
12 12 12
CPM 1
CPM 2
Concentration 1 (IV/ml)
694 720 684 778 831 757 822 832 874 1423 1697 1523 2027 2136 2080
714 736 698 788 799 791 841 787 930 1295 1649 1414 1932 1848 2148
1.4 1.6 1.3 1.9 2.3 1.8 2.2 2.3 2.6 6.5 8.7 7.3 11.2 12. l 11.6
is especially important since 75% to 80% of all Down syndrome offspring are delivered unexpectedly by women <35 years of age.4 Hence any screening method that recognizes women who were unaware of being at increased risk of bearing a child with a chromosome defect would represent an important advance. Once recognized, such women could be offered definitive chorionic villus sampling or amniocentesis for prenatal chromosomal studies. To this end, we have established a sensitive radioimmunoass~y based on monoclonal antibodies to a-fetoprotein. 5 We now report on the use of maternal serum a-fetoprotein screening of maternal sera obtained prospectively just before chorionic villus sampling, presented briefly elsewhere, 6 - 8 for the diagnosis of chromosome defects at 8 to 12 weeks' fetal age. The
Concentration 2 (IV /ml)
Average concentration (IV/ml)
1.5 1.7 1.4 2.0 2.1 2.0 2.3 2.0 2.9 5.6 8.3 6.5 10.5 9.8 12.2
1.5 1.6 1.4 2.0 2.2 1.9 2.3 2.1 2.8 6.1 8.5 6.9 10.8 10.9 11.9
± SEM 0.05 0.05 0.05 0.05
O.IO 0.10 0.05 0.15 0.15 0.45 0.20 0.40 0.35 l.15 0.30
primary goal of our study was to determine the detection efficiency of prospective first-trimester maternal serum a-fetoprotein screening for chromosome defects in a group of women at increased risk of having a child with a chromosome defect. Methods
a-Fetoprotein assay. After systematic testing of different combinations of anti-a-fetoprotein monoclonal antibodies, Bellet et al. 5 demonstrated that the most sensitive assay for serum a-fetoprotein was a "simultaneous-sandwich" monoclonal-based radioimmunoassay based on a mixture of "capture" a-fetoprotein monoclonal antibodies (AF 01 and AF 03) on solid phase support, with AF 01 as the radiolabeled indicator antibody. For this assay, polystyrene beads (outer diameter 0.65 cm; Precision Plastic Ball,
First-trimester a-fetoprotein screening
Volume 159 Number 5
1211
i 2. 5
33
I
First Trimester MSAFP-Pre•CVS
I
!Not Weight-Adjusted]
30
I
I Chromosome Disorders
25
•
Trisomy 21
0
Trisomy 18
e
Trisomy 13
6.
Other
I
I I I I
/ I I /
20
/
.....
/
e.....
/
/
/
:::J
/ Cl. LI..
·'
15
/
•
/
/MED
< Cf)
-·-
~
10
--
/
/
/
•
0
"
•
•A11111
/
/
,,,,,,,,,,
.
0.6 ,,,.0.5
11
12
./
;I,11ir;,,,, /
5
-·-·
----
f f' o~---
4_.-·-
•
8
9
/
./
,,,,,,,,,
- .-. 10
WEEKS GESTATION Fig.%. First-trimester maternal serum a-fetoprotein (MSAFP) screening shows 0.5, 0.6, median, and 2.5 multiples of the median derived without maternal weight adjustment, and plotted values of 27 serious chromosome defects. One patient with a trisomy 22 fetus underwent assay twice (line joining points at IO and 12 weeks). CVS, Chorionic villus sampling.
Chicago, Ill.) coated with AF 01 and AF 03 antibodies were used. Serum samples of a-fetoprotein-positive standards (200 µ.l) and 180,000 cpm of iodine 125labeled AF 0 l (specific activity 10 to 16 µCi/ µ.g; l Ci = 3. 7 x 10 becquerels) in 100 ml of buffer (50% fetal calf serum in phosphate-buffered saline solution, pH 7.2) were added simultaneously with the anti-afetoprotein-coated beads. After incubation at 45° C for l hour, the solid phase support was then washed three times with distilled water, and the radioactivity that was bound (CPM) was measured in a Beckman -y-well counter. Further details of this assay have been reported elsewhere.5 The sensitivty of th~ monoclonal-based radioimmunoassay sensitivity is approximately 0.5 ng/ml (0.4 IU/ml) and is particularly accurate between 2.0 and 10.0 ng/ml (l.5 and 7.7 IU/ml). The standard curve
is linear between 1.0 and 60.0 ng/ml (0.8 and 46.5 IU/ml) (critical part of curve shown in Fig. 1). It is important that our available data (not presented) also show that a-fetoprotein levels could be measured reliably over a wide range of concentrations in both the first and second trimesters. Serum samples were measured in duplicate (Table I). We noted an interassay variation between 4% and 7%. We observed that a wide range of a-fetoprotein values can be accurately measured with a coefficient of variation of generally less than 5%. To determine the normal nonpregnant female maternal serum a-fetoprotein values, we measured a-fetoprotein in duplicate in serum derived from 640 female blood donors (ages 18 to 45 years). The mean (±SD) normal value for serum was 0.3 ± 0.09 IU/ml (0.36 ± 0.11 ng/ml). Patients. The study population consisted of 540
November 1988 Am J Obstet Gynecol
1212 Milunsky et al.
four cases (Table II) were later found to be normal after subsequent amniocentesi s studies.
Table II. Cytogenetic and other results after CVS in 540 screened patients Normal karyotypes Balanced translocations Spontaneous abortions Single gene disorder Encephalocele Twins Chromosome disorders Autosomal trisomies Trisomy 21 (10) Trisomy 18 (6) Trisomy 13 (2) Trisomy 14 (1) Trisomy 22 (l) Double trisomy (2)§ Sex chromosome abnormality (XXY) Mosaics Miscellaneous abnormalities
500* 3
6t 1 l:j:
2 27 22
l
2
211
*Cytogenetic discrepancies (four).
t All had normal karyotypes. *Normal maternal serum a-fetoprotein. §Trisomies 18/11 and 18/21. !!Unbalanced YI 13 and 13/ 14 translocations.
women planning to undergo chorionic villus sampling, primarily for advanced maternal age or because of the previous birth of a child with a chromosome defect. The basic indications and selection criteria for chorionic villus sampling were identical to those one of us (B. B.) published elsewhere. 9 All patients accepted for chorionic villus sampling had a detailed ultrasound study performed just before the procedure and 87% had also been subjected to an ultrasound study 1 to 2 weeks before chorionic villus sampling. Therefore fetal age was accurately established. Blood samples were obtained shortly before the chorionic villus sampling procedure, the serum removed, and stored at - 20° C. These samples were coded and sent by airmail from Milan to Boston. This cohort also included eight cases with chromosome defects that two of us (B. B. and 0 I. B.) had previously studied.' Samples were then assayed and results were returned to Milan, and thereafter the karyotypes of all 540 patients were released to us. Some 75% of the women in this study were ~35 years of age, and 81 % weighed between 100 and 149 pounds. Fifteen percent weighed between 150 and 200 pounds, with no patients exceeding 200 pounds. The cytogenetic and other results after chorionic villus sampling (Table II) show that 27 had chromosome disorders, of which 22 had the various trisomies as shown. Initial cytogenetic results thought to be abnormal on chromosome study of the chorionic villus sampling in
Results
Detection efficiency. Median values and multiples of the median for maternal serum a-fetoprotein were calculated for the 540 patients (Fig. 2). Assay results for the 27 cases in which fetal chromosome disorders were detected are shown in Fig. 2. Two lower limits were arbitrarily selected at 0.5 and 0.6 multiples of the median, respectively. Fig. 1 shows that three of 22 (13.6%) trisomies fell below 0.5 multiples of the median. When the ~0.6 multiples of the median limit was used, seven of 22 (31.8%) fetal trisomies were reflected by maternal serum a-fetoprotein values at or below that cutoff. Of interest is the single case of trisomy 22 found with a normal maternal serum a-fetoprotein value at 10 weeks but a decidedly low value at 12 weeks, when a second chorionic villus sampling procedure was performed (Fig. 2). Maternal serum a-fetoprotein values were assessed to determine whether adjustment for maternal weight would add to detection efficiency. Although no difference was observed for patients whose values fell below 0.5 multiples of the median, there was an approximate 4% loss of efficiency with the 0.6 multiples of the median cutoff and considering trisomies only. When detection efficiency for all chromosome disorders was considered and no maternal weight adjustment was performed, three of 27 ( 11. l %) had maternal serum a-fetoprotein values ~0.5 multiples of the median, whereas a striking eight of 27 (29.6%) had values ~0.6 multiples of the median. It is important to note that our investigation was neither a consecutive nor a random population study. Nevertheless, to determine the efficiency of detection in a blindly studied cohort, we calculated the odds of a chromosome defect being present when the two lower limits for maternal serum a-fetoprotein were used (Table III). In 540 cases, 6% and 11 %, respectively, had maternal serum a-fetoprotein values ~0.5 and ~0.6 multiples of the median. As noted for the lowest cutoff, the odds of any chromosome disorder being present was 1: 11, which was not statistically significant. In contrast, the odds of any serious chromosome disorder being present when the ~0.6 multiples of the median cutoff was used was 1: 7, a highly significant (p < 0.005) finding in a group of patients studied prospectively and blinded to the karyotype analysis. Comment The primary aim of our study was to determine whether it would be possible during the first trimester
Volume 159 Number 5
First-trimester a-fetoprotein screening
1213
Table III. Odds of a chromosome defect in 540 cases according to which lower limit cutoff is used for maternal serum u-fetoprotein values
32 (6)
3
7
0 3 1:16 (6.3) I: 11 (9.4) I: 11 (9.4) (NS)
I
8 1:20 (5) 1:8(11.9) 1:7(13.6)(p < 0.005)
MOM, Multiples of the median.
of pregnancy to recognize a group of women who unknowingly had an increased risk of carrying a fetus with a serious chromosome defect. The strengths of this study are its prospective nature, the fact that u-fetoprotein assays were performed under code to any cytogenetic diagnosis, that all cases were karyotyped, and that fetal age assessment was accurately determined. Overall, we demonstrated that eight of 27 (29.6%) of all chromosome defects had maternal serum u-fetoprotein values :;;;0.6 multiples of the median. Furthermore, three of 10 pregnancies having fetuses with Down syndrome used the same :;;;0.6 multiples of the median cutoff. Thus this monoclonal antibody-based assay was sensitive, reliable, and reproducible. Merkatz et al. 11 had originally noted the association between low maternal serum u-fetoprotein values and Down syndrome as measured in the second trimester. Cuckle et al.2 had calculated from accumulated secondtrimester maternal serum u-fetoprotein screening data that with a :;;;0.6 multiples of the median cutoff, 33% of pregnancies with a Down syndrome fetus had values below this limit. As previously noted, our prospective findings in the first trimester are remarkably consonant with their calculated expectations for the second trimester. We emphasize that this was not a consecutive random pregnancy population study, and therefore the results should not be extrapolated for all pregnancies. We sought to determine whether maternal serum u-fetoprotein values could be reliably detected during the first trimester and if such measurements could be used to select out women at an increased risk of carrying a fetus with a serious chromo.some defect. Selection of a lower limit cutoff of :;;;0.6 multiples of the median is purely arbitrary and may change with further experience. We have demonstrated that first-trimester maternal serum u-fetoprotein screening can be very successful in detecting serious fetal chromosome defects, as long
as it is preceded by careful ultrasound study to accurately establish fetal age. Currently, the vast majority of women do not have an early first-trimester ultrasound study. Hence we are now in the process of determining whether maternal serum u-fetoprotein screening can be successful in the first trimester depending solely on last menstrual period dates for fetal age assessment and without prior ultrasound study. We thank Dr. Tim Heeren for his help with the statistics. REFERENCES I. Milunsky A. The prenatal diagnosis of neural tube and other congenital defects. In: Milunsky A, ed. Genetic disorders and the fetus: diagnosis, prevention and treatment. New York: Plenum Press, 1986:453-519. 2. Cuckle HS, Wald NJ, Lindenbaum RH. Maternal serum alpha-fetoprotein measurement: a screening test for Down syndrome. Lancet 1984; I :926-9. 3. Palomaki GE. Collaborative study of Down syndrome screening using maternal serum alpha-fetoprotein and maternal age. Lancet 1986;2:1460. 4. Holmes LB. Genetic counselling for the older pregnant woman: new data and questions. N Engl J Med 1978; 298:1419-21. 5. Bellet DH, Wands JR, Isselbacher KJ, et al. Serum alphafetoprotein levels in human disease: perspective from a highly specific monoclonal radioimmunoassay. Proc Natl Acad Sci USA. 1984;81:3869-73. 6. Milunsky A, Wands JR, Bellet D. First trimester maternal serum alpha-fetoprotein (MSAFP) screening. Pediatr Res 1984; I 9:328A. 7. Milunsky A, Wands JR, Brambati B. First trimester maternal serum alpha-fetoprotein (MSAFP) screening for chromosome defects. Pediatr Res 1987;21:292A. 8. Milunsky A, Wands JR, Brambati B. First trimester maternal serum alpha-fetoprotein screening for chromosome defects. AmJ Hum Genet 1987;41:28IA. 9. Brambati B, Oldrini EF, Lanzani A. Chorionic villus sampling: an analysis of the obstetric experience of I 000 cases. Prenat Diagn 1987;7: 157-69. 10. Brambati B, Simoni G, Bonacchi I, et al. Fetal chromosomal aneuploidies and maternal serum alphafetoprotein levels in first trimester. Lancet 1986;2: 165-6. 11. Merkatz IK, Nitowsky HM, Macri JN, Johnson WE. An association between low maternal serum a-fetoprotein and fetal chromosomal abnormalities. AM J OBSTET GvNECOL 1984; 148:886-94.
Structure-activity relationship of some steroidal hypnotic agents. J Med Chem 1968; 11: 117-25. 17. Backstrom T, Landgren S, Zetterlund B, et al. Effects of ovarian steroid hormones on brain excitability and their relation to epilepsy seizure variation during the menstrual cycle. In: Porter RJ, Mattson RH, Ward AA Jr, Dam M, eds. Advances in epileptology. XVth Epilepsy international symposium. New York: Raven Press, 1984:269-76. 18. H¢gskilde S, NeilsenJW, Carl P, S¢rensen MB. Pregnanolone emulsion. A new steroid preparation for intravenous anesthesia: an experimental study in mice. Anaesthesia I 987;42:586-90. 19. Heuser G. Induction of anesthesia, seizures and sleep by steroid hormones. Anesthesiology 1967;28:173-83.
Hypnotic effect of progesterone and its metabolites
20. Banks P, Peace CB. Enzyme inhibition by steroid anaesthetic agents derived from progesterone. Br J Anaesth 1985;57:512-4. 21. Majewska MD, Harrison NL, Schwartz RD, Barker JL, Paul SM. Steroid hormone metabolites are barbituratelike modulators of the GABA receptor. Science 1986; 232:1004-7. 22. Harrison NL, Majewska MD, Harrington JW, Barker JL. Structure-activity relationships for steroid interaction with the gamma-aminobutyric acid: a receptor complex. J Pharmacol Exp Ther 1987;241:346-53.
First-trimester maternal serum a-fetoprotein screening for chromosome defects Aubrey Milunsky, MB, BCh, DSc,a Jack Wands, MD," Bruno Brambati, MD,c Ida Bonacchi, MD, c and Karen Currie• Boston, Massachusetts, and Milan, Italy Low maternal serum cx-fetoprotein values during the second trimester of pregnancy are associated with an increased risk of Down syndrome in the fetus. In this study a sensitive, monoclonal-based radioimmunoassay for cx-fetoprotein was used to determine whether such an association also applies to the first trimester and if maternal serum cx-fetoprotein screening could successfully detect a significant number of pregnancies in which the fetus had a trisomy or other chromosome disorder. Sera were obtained prospectively from 540 women just before chorionic villus sampling for prenatal diagnosis of chromosome defects (largely because of advanced maternal age) at 8 to 12 weeks' fetal age and assayed for cx-fetoprotein under code without knowledge of the cytogenetic results. Eight of 27 (29.6%) of all serious chromosome defects were associated with low maternal serum a-fetoprotein values (,,;0.6 multiples of the median). Overall, 59 of 540 patients (10.9%) had maternal serum a-fetoprotein values ,,;0.6 multiples of the median, eight of whom had a fetus with a serious chromosome defect. Women whose maternal serum cx-fetoprotein value was .,,0.6 multiples of the median had one in eight odds of carrying a fetus with a trisomy and one in seven odds of the fetus having any serious chromsome defect. From this study of a group of women at higher risk, we conclude that first-trimester maternal serum cx-fetoprotein screening for chromsome defects is feasible. A prospective study to determine detection efficiency is now required of a consecutive routine pregnancy population in whom gestational age is determined by menstrual dates as is usually the case in clinical practice (AM J OesTET GYNECOL 1988;159:1209-13.)
Key words: Chromosome defects, maternal serum a-fetoprotein values, trisomy, multiples of the median
From the Center for Human Genetics and the Department of Pediatrics," Boston University School of Medicine, the Gastrointestinal Unit,' Department of Medicine, Massachusetts General Hospital and Harvard Medical School, and L. Mangiagalli Clinic,' First Institute of Obstetrics and Gynaecology, University of Milan. Supported by U.S. Public Health Service Grant HD20469. Received for publication November 14, 1987; revised Aprill 3, 1988; accepted May 19, 1988. Reprint requests: Aubrey Milunsky, MD, Center for Human Genetics and Department of Pediatrics, Boston University School of Medicine, 80 E. Concord, St., Boston, MA 02118.
Second-trimester maternal serum a-fetop1otein screening is now established as a valuable tool in the prenatal detection of neural tube and other congenital defects and in the recognition of high-risk pregnancies.' Most recently, determination of low maternal serum a-fetoprotein values has shown great promise in the detection of a significant number of pregnancies in which the fetus has Down syndrome. 2 · 3 This approach
1209
1210 Milunsky et al.
November 1988 Am J Obstet Gynecol
3500 3000 2500 2000 ::Ii CL
0
1500 1000
500
00
2
3
4
5
7
6
8
9
10
11
12
AFP Iu/ml
Fig. I. Standard curve for serum a-fetoprotein values <12 IU/ml.
Table I. Representative sample of first-trimester maternal serum a-fetoprotein cases assayed in duplicate Case identification No.
3836 3897 4072 3843 3844 3920 3842 4015 3959 38~6
3891 4014 3921 4002 4044
Weeks' gestation
8 8 8 9 9 9"
IO IO IO II II II
12 12 12
CPM 1
CPM 2
Concentration 1 (IV/ml)
694 720 684 778 831 757 822 832 874 1423 1697 1523 2027 2136 2080
714 736 698 788 799 791 841 787 930 1295 1649 1414 1932 1848 2148
1.4 1.6 1.3 1.9 2.3 1.8 2.2 2.3 2.6 6.5 8.7 7.3 11.2 12. l 11.6
is especially important since 75% to 80% of all Down syndrome offspring are delivered unexpectedly by women <35 years of age.4 Hence any screening method that recognizes women who were unaware of being at increased risk of bearing a child with a chromosome defect would represent an important advance. Once recognized, such women could be offered definitive chorionic villus sampling or amniocentesis for prenatal chromosomal studies. To this end, we have established a sensitive radioimmunoass~y based on monoclonal antibodies to a-fetoprotein. 5 We now report on the use of maternal serum a-fetoprotein screening of maternal sera obtained prospectively just before chorionic villus sampling, presented briefly elsewhere, 6 - 8 for the diagnosis of chromosome defects at 8 to 12 weeks' fetal age. The
Concentration 2 (IV /ml)
Average concentration (IV/ml)
1.5 1.7 1.4 2.0 2.1 2.0 2.3 2.0 2.9 5.6 8.3 6.5 10.5 9.8 12.2
1.5 1.6 1.4 2.0 2.2 1.9 2.3 2.1 2.8 6.1 8.5 6.9 10.8 10.9 11.9
± SEM 0.05 0.05 0.05 0.05
O.IO 0.10 0.05 0.15 0.15 0.45 0.20 0.40 0.35 l.15 0.30
primary goal of our study was to determine the detection efficiency of prospective first-trimester maternal serum a-fetoprotein screening for chromosome defects in a group of women at increased risk of having a child with a chromosome defect. Methods
a-Fetoprotein assay. After systematic testing of different combinations of anti-a-fetoprotein monoclonal antibodies, Bellet et al. 5 demonstrated that the most sensitive assay for serum a-fetoprotein was a "simultaneous-sandwich" monoclonal-based radioimmunoassay based on a mixture of "capture" a-fetoprotein monoclonal antibodies (AF 01 and AF 03) on solid phase support, with AF 01 as the radiolabeled indicator antibody. For this assay, polystyrene beads (outer diameter 0.65 cm; Precision Plastic Ball,
First-trimester a-fetoprotein screening
Volume 159 Number 5
1211
i 2. 5
33
I
First Trimester MSAFP-Pre•CVS
I
!Not Weight-Adjusted]
30
I
I Chromosome Disorders
25
•
Trisomy 21
0
Trisomy 18
e
Trisomy 13
6.
Other
I
I I I I
/ I I /
20
/
.....
/
e.....
/
/
/
:::J
/ Cl. LI..
·'
15
/
•
/
/MED
< Cf)
-·-
~
10
--
/
/
/
•
0
"
•
•A11111
/
/
,,,,,,,,,,
.
0.6 ,,,.0.5
11
12
./
;I,11ir;,,,, /
5
-·-·
----
f f' o~---
4_.-·-
•
8
9
/
./
,,,,,,,,,
- .-. 10
WEEKS GESTATION Fig.%. First-trimester maternal serum a-fetoprotein (MSAFP) screening shows 0.5, 0.6, median, and 2.5 multiples of the median derived without maternal weight adjustment, and plotted values of 27 serious chromosome defects. One patient with a trisomy 22 fetus underwent assay twice (line joining points at IO and 12 weeks). CVS, Chorionic villus sampling.
Chicago, Ill.) coated with AF 01 and AF 03 antibodies were used. Serum samples of a-fetoprotein-positive standards (200 µ.l) and 180,000 cpm of iodine 125labeled AF 0 l (specific activity 10 to 16 µCi/ µ.g; l Ci = 3. 7 x 10 becquerels) in 100 ml of buffer (50% fetal calf serum in phosphate-buffered saline solution, pH 7.2) were added simultaneously with the anti-afetoprotein-coated beads. After incubation at 45° C for l hour, the solid phase support was then washed three times with distilled water, and the radioactivity that was bound (CPM) was measured in a Beckman -y-well counter. Further details of this assay have been reported elsewhere.5 The sensitivty of th~ monoclonal-based radioimmunoassay sensitivity is approximately 0.5 ng/ml (0.4 IU/ml) and is particularly accurate between 2.0 and 10.0 ng/ml (l.5 and 7.7 IU/ml). The standard curve
is linear between 1.0 and 60.0 ng/ml (0.8 and 46.5 IU/ml) (critical part of curve shown in Fig. 1). It is important that our available data (not presented) also show that a-fetoprotein levels could be measured reliably over a wide range of concentrations in both the first and second trimesters. Serum samples were measured in duplicate (Table I). We noted an interassay variation between 4% and 7%. We observed that a wide range of a-fetoprotein values can be accurately measured with a coefficient of variation of generally less than 5%. To determine the normal nonpregnant female maternal serum a-fetoprotein values, we measured a-fetoprotein in duplicate in serum derived from 640 female blood donors (ages 18 to 45 years). The mean (±SD) normal value for serum was 0.3 ± 0.09 IU/ml (0.36 ± 0.11 ng/ml). Patients. The study population consisted of 540
November 1988 Am J Obstet Gynecol
1212 Milunsky et al.
four cases (Table II) were later found to be normal after subsequent amniocentesi s studies.
Table II. Cytogenetic and other results after CVS in 540 screened patients Normal karyotypes Balanced translocations Spontaneous abortions Single gene disorder Encephalocele Twins Chromosome disorders Autosomal trisomies Trisomy 21 (10) Trisomy 18 (6) Trisomy 13 (2) Trisomy 14 (1) Trisomy 22 (l) Double trisomy (2)§ Sex chromosome abnormality (XXY) Mosaics Miscellaneous abnormalities
500* 3
6t 1 l:j:
2 27 22
l
2
211
*Cytogenetic discrepancies (four).
t All had normal karyotypes. *Normal maternal serum a-fetoprotein. §Trisomies 18/11 and 18/21. !!Unbalanced YI 13 and 13/ 14 translocations.
women planning to undergo chorionic villus sampling, primarily for advanced maternal age or because of the previous birth of a child with a chromosome defect. The basic indications and selection criteria for chorionic villus sampling were identical to those one of us (B. B.) published elsewhere. 9 All patients accepted for chorionic villus sampling had a detailed ultrasound study performed just before the procedure and 87% had also been subjected to an ultrasound study 1 to 2 weeks before chorionic villus sampling. Therefore fetal age was accurately established. Blood samples were obtained shortly before the chorionic villus sampling procedure, the serum removed, and stored at - 20° C. These samples were coded and sent by airmail from Milan to Boston. This cohort also included eight cases with chromosome defects that two of us (B. B. and 0 I. B.) had previously studied.' Samples were then assayed and results were returned to Milan, and thereafter the karyotypes of all 540 patients were released to us. Some 75% of the women in this study were ~35 years of age, and 81 % weighed between 100 and 149 pounds. Fifteen percent weighed between 150 and 200 pounds, with no patients exceeding 200 pounds. The cytogenetic and other results after chorionic villus sampling (Table II) show that 27 had chromosome disorders, of which 22 had the various trisomies as shown. Initial cytogenetic results thought to be abnormal on chromosome study of the chorionic villus sampling in
Results
Detection efficiency. Median values and multiples of the median for maternal serum a-fetoprotein were calculated for the 540 patients (Fig. 2). Assay results for the 27 cases in which fetal chromosome disorders were detected are shown in Fig. 2. Two lower limits were arbitrarily selected at 0.5 and 0.6 multiples of the median, respectively. Fig. 1 shows that three of 22 (13.6%) trisomies fell below 0.5 multiples of the median. When the ~0.6 multiples of the median limit was used, seven of 22 (31.8%) fetal trisomies were reflected by maternal serum a-fetoprotein values at or below that cutoff. Of interest is the single case of trisomy 22 found with a normal maternal serum a-fetoprotein value at 10 weeks but a decidedly low value at 12 weeks, when a second chorionic villus sampling procedure was performed (Fig. 2). Maternal serum a-fetoprotein values were assessed to determine whether adjustment for maternal weight would add to detection efficiency. Although no difference was observed for patients whose values fell below 0.5 multiples of the median, there was an approximate 4% loss of efficiency with the 0.6 multiples of the median cutoff and considering trisomies only. When detection efficiency for all chromosome disorders was considered and no maternal weight adjustment was performed, three of 27 ( 11. l %) had maternal serum a-fetoprotein values ~0.5 multiples of the median, whereas a striking eight of 27 (29.6%) had values ~0.6 multiples of the median. It is important to note that our investigation was neither a consecutive nor a random population study. Nevertheless, to determine the efficiency of detection in a blindly studied cohort, we calculated the odds of a chromosome defect being present when the two lower limits for maternal serum a-fetoprotein were used (Table III). In 540 cases, 6% and 11 %, respectively, had maternal serum a-fetoprotein values ~0.5 and ~0.6 multiples of the median. As noted for the lowest cutoff, the odds of any chromosome disorder being present was 1: 11, which was not statistically significant. In contrast, the odds of any serious chromosome disorder being present when the ~0.6 multiples of the median cutoff was used was 1: 7, a highly significant (p < 0.005) finding in a group of patients studied prospectively and blinded to the karyotype analysis. Comment The primary aim of our study was to determine whether it would be possible during the first trimester
Volume 159 Number 5
First-trimester a-fetoprotein screening
1213
Table III. Odds of a chromosome defect in 540 cases according to which lower limit cutoff is used for maternal serum u-fetoprotein values
32 (6)
3
7
0 3 1:16 (6.3) I: 11 (9.4) I: 11 (9.4) (NS)
I
8 1:20 (5) 1:8(11.9) 1:7(13.6)(p < 0.005)
MOM, Multiples of the median.
of pregnancy to recognize a group of women who unknowingly had an increased risk of carrying a fetus with a serious chromosome defect. The strengths of this study are its prospective nature, the fact that u-fetoprotein assays were performed under code to any cytogenetic diagnosis, that all cases were karyotyped, and that fetal age assessment was accurately determined. Overall, we demonstrated that eight of 27 (29.6%) of all chromosome defects had maternal serum u-fetoprotein values :;;;0.6 multiples of the median. Furthermore, three of 10 pregnancies having fetuses with Down syndrome used the same :;;;0.6 multiples of the median cutoff. Thus this monoclonal antibody-based assay was sensitive, reliable, and reproducible. Merkatz et al. 11 had originally noted the association between low maternal serum u-fetoprotein values and Down syndrome as measured in the second trimester. Cuckle et al.2 had calculated from accumulated secondtrimester maternal serum u-fetoprotein screening data that with a :;;;0.6 multiples of the median cutoff, 33% of pregnancies with a Down syndrome fetus had values below this limit. As previously noted, our prospective findings in the first trimester are remarkably consonant with their calculated expectations for the second trimester. We emphasize that this was not a consecutive random pregnancy population study, and therefore the results should not be extrapolated for all pregnancies. We sought to determine whether maternal serum u-fetoprotein values could be reliably detected during the first trimester and if such measurements could be used to select out women at an increased risk of carrying a fetus with a serious chromo.some defect. Selection of a lower limit cutoff of :;;;0.6 multiples of the median is purely arbitrary and may change with further experience. We have demonstrated that first-trimester maternal serum u-fetoprotein screening can be very successful in detecting serious fetal chromosome defects, as long
as it is preceded by careful ultrasound study to accurately establish fetal age. Currently, the vast majority of women do not have an early first-trimester ultrasound study. Hence we are now in the process of determining whether maternal serum u-fetoprotein screening can be successful in the first trimester depending solely on last menstrual period dates for fetal age assessment and without prior ultrasound study. We thank Dr. Tim Heeren for his help with the statistics. REFERENCES I. Milunsky A. The prenatal diagnosis of neural tube and other congenital defects. In: Milunsky A, ed. Genetic disorders and the fetus: diagnosis, prevention and treatment. New York: Plenum Press, 1986:453-519. 2. Cuckle HS, Wald NJ, Lindenbaum RH. Maternal serum alpha-fetoprotein measurement: a screening test for Down syndrome. Lancet 1984; I :926-9. 3. Palomaki GE. Collaborative study of Down syndrome screening using maternal serum alpha-fetoprotein and maternal age. Lancet 1986;2:1460. 4. Holmes LB. Genetic counselling for the older pregnant woman: new data and questions. N Engl J Med 1978; 298:1419-21. 5. Bellet DH, Wands JR, Isselbacher KJ, et al. Serum alphafetoprotein levels in human disease: perspective from a highly specific monoclonal radioimmunoassay. Proc Natl Acad Sci USA. 1984;81:3869-73. 6. Milunsky A, Wands JR, Bellet D. First trimester maternal serum alpha-fetoprotein (MSAFP) screening. Pediatr Res 1984; I 9:328A. 7. Milunsky A, Wands JR, Brambati B. First trimester maternal serum alpha-fetoprotein (MSAFP) screening for chromosome defects. Pediatr Res 1987;21:292A. 8. Milunsky A, Wands JR, Brambati B. First trimester maternal serum alpha-fetoprotein screening for chromosome defects. AmJ Hum Genet 1987;41:28IA. 9. Brambati B, Oldrini EF, Lanzani A. Chorionic villus sampling: an analysis of the obstetric experience of I 000 cases. Prenat Diagn 1987;7: 157-69. 10. Brambati B, Simoni G, Bonacchi I, et al. Fetal chromosomal aneuploidies and maternal serum alphafetoprotein levels in first trimester. Lancet 1986;2: 165-6. 11. Merkatz IK, Nitowsky HM, Macri JN, Johnson WE. An association between low maternal serum a-fetoprotein and fetal chromosomal abnormalities. AM J OBSTET GvNECOL 1984; 148:886-94.