Human Reproduction vol.13 no.3 pp.744–748, 1998
Second trimester maternal dimeric inhibin-A in the
multiple-marker screening test for Down’s syndrome
Martin A.Renier1,3, Annie Vereecken2,
Erik Van Herck2, Danny Straetmans2,
Paul Ramaekers2 and Philippe Buytaert1
1University
Hospital of Antwerp, Department of Obstetrics and
Gynecology, University of Antwerp and 2Laboratory of Clinical
Pathology, 2000 Antwerp, Belgium.
3To
whom correspondence should be addressed
The aim of this study was to evaluate the additional
value of dimeric inhibin-A serum concentration in second
trimester multiple-marker screening tests for pregnancies
affected by Down’s syndrome. We anticipated that second
trimester maternal serum dimeric inhibin-A concentrations
would be altered in pregnancies complicated by fetal
Down’s syndrome and that dimeric inhibin-A would perform better than one of the three substances analysed in
the multiple-marker screening test currently in use. A total
of 1156 serum samples were screened for dimeric inhibin-A
in parallel with the routine classic triple test screening
programme performed on a random obstetric population.
Classic triple test performance was compared with detection rates obtained after substitution of unconjugated
oestriol by inhibin-A and with the performance of inhibin-A
and α-fetoprotein alone. Absolute dimeric inhibin-A
maternal serum concentrations of Down’s syndrome pregnancies were indeed significantly higher than those of
normal pregnancies in our screened population. The performance of dimeric inhibin-A in combination with the
multiple-marker screening test, however, is limited because
of its strong correlation with intact human chorionic
gonadotrophin.
Key words: dimeric inhibin-A/Down’s syndrome/screening/
second trimester
Introduction
Merkatz et al. (1984) reported that pregnancies complicated
by fetal autosomal trisomies are characterized by low maternal
serum α-fetoprotein (MSAFP) concentrations. In 1987, Cuckle
et al. published the first results on screening for Down’s
syndrome using AFP and maternal age. Numerous other
pregnancy-associated maternal serum markers for fetal trisomy
21 (Down’s syndrome) have been evaluated since. It has
been established that human chorionic gonadotrophin (HCG)
concentrations are higher and oestriol concentrations are lower
in Down’s syndrome pregnancies. These three parameters
(AFP, HCG and free oestriol), in addition to maternal age, are
used routinely in risk calculation programmes for second
744
trimester Down’s syndrome screening. Multiple-marker screening tests for fetal Down’s syndrome detect ~60% of all Down’s
syndrome cases in women aged ,35 years (screen-positive
rate 5 3.8%) and 75–89% in women aged ù35 years (screenpositive rate 5 25%) (Haddow et al., 1992; Cheng et al.,
1993). It is not surprising, therefore, that prenatal screening
for Down’s syndrome has become an important and established
part of modern prenatal care in Belgium.
Efforts to improve biochemical screening have centred on
the search for better markers in order to either improve the
Down’s syndrome detection rate or reduce the false-positive
rate. Numerous pregnancy-associated substances analysed in
serum have been evaluated. Additional markers, such as
progesterone (Kratzer et al., 1991), pregnancy-associated
plasma protein A (Brambati et al., 1993), schwangerschaftsprotein 1 (Brock et al., 1990) and placental alkaline phosphatase
(Ind et al., 1993), have been evaluated but without success.
Inhibin, a heterodimeric protein with a molecular weight of
32 000, has emerged recently as a promising candidate. It is
composed of one α-subunit and one of two related β-subunits
(β-A or β-B), and is produced by the syncytiotrophoblast of
human placenta (Petraglia et al., 1987). Of the bioactive
dimeric inhibins, only inhibin-A is present in maternal serum
during pregnancy (Riley et al., 1996). Three early studies,
depending on assays using antibodies directed against epitopes
only on the α-subunit, reported conflicting immunoreactive
inhibin concentrations in maternal serum from second trimester
as well as first trimester pregnancies affected by Down’s
syndrome (Van Lith et al., 1992; Spencer et al., 1993; Cuckle
et al., 1994a). However, an estimation of dimeric inhibin-A
concentrations using a two-site immunoassay with monoclonal
antibodies has reported an additional value in the detection of
Down’s syndrome in two recent studies (Aitken et al., 1996;
Wallace et al., 1996).
These studies are all retrospective and have been performed
on small series. In a preliminary trial we analysed 260 serum
samples for dimeric inhibin-A and 1156 serum samples in a
second trial, in parallel with our routine second trimester
Down’s syndrome screening programme. We attempted to
confirm the promising preliminary inhibin-A data for second
trimester Down’s syndrome screening in a larger series and in
routine day-to-day screening procedures.
Materials and methods
The Down’s syndrome screening programme in our laboratory
includes AFP, HCG and free oestriol as biochemical markers. Second
trimester serum samples were analysed for AFP, β-HCG and unconjugated oestriol using isotopic assays (I125). Our primary goal was to
© European Society for Human Reproduction and Embryology
Dimeric inhibin-A in screening for Down’s syndrome
use routine daily assays rather than specially designed ones in the
screening programme for Down’s syndrome. AFP and HCG were
measured with an immunoradiometric assay respectively from
Diagnostic Products Corporation (Los Angeles, CA, USA) and
BioSource Europe SA (formerly Medgenix-Belgium, Fleurus,
Belgium). Results were expressed in multiple of the median (MoM)
for the appropriate gestation and the log 10 value was used for risk
calculations. Unconjugated oestriol was measured with a sensitive
radioimmunoassay from Diagnostic Systems Laboratories Inc.
(Webster, TX, USA); results were also expressed in MoM without
logarithmic transformation.
Screening for Down’s syndrome is performed twice a week (~80
samples each run) and all assays are performed in duplicate. For risk
calculations we used the algorithm published by Palomaki et al. (1992;
Foundation for Blood Research, Box 190, Scarborough, ME 04074,
USA), which can be used freely if acknowledged as such. After each
run, a quality control program calculated the means and SD for each
parameter. MoM values were improved and controlled each month
with an exponential regression analysis. Correction was performed for
weight and smoking. In cases of multiple pregnancy (twins), we used
the correction factors published by Wald (1991). The obstetric population from an ethnic point of view was considered to be homogeneous
Caucasian because most other ethnic groups refused triple testing for
religious reasons. Mean 6 SD maternal age distribution was 32 6 1
years. In the affected pregnancies group (n 5 18) all mothers were
multiparous. In the unaffected (control) group the mean 6 SD maternal
age was 28 6 1 years. The group consisted of 20% nullipara and 80%
multipara. In this control group 2% of newborn babies suffered from
major or minor neonatal complications. Therefore it was highly unlikely
that the control group was biased for factors that would raise inhibin-A
or HCG concentrations at 16 weeks of pregnancy, as was suggested
recently for hypertensive pregnancies at 29 weeks (Muttukrishna
et al., 1997).
To evaluate the discriminative power of dimeric inhibin-A, we analysed 260 serum samples stored at –20°C, including 14 samples from
Down’s syndrome pregnancies, in the first retrospective pilot trial.
Reagents for dimeric inhibin-A were commercially available from
Serotec Ltd (Kidlington, UK; product code MCA 127 3KZZ). Dimeric
inhibin-A was measured by a two-site enzyme-linked immunosorbent
assay which utilised an immobilized anti-βA inhibin subunit monoclonal antibody as the capture antibody. The second or detection
antibody was a monoclonal antibody specific for the α-subunit of
inhibin and was coupled to alkaline phosphatase. The assay was
rather cumbersome and time-consuming (2.5 days), and therefore not
ideal for routine day-to-day large-scale population screening purposes.
In the second trial on 1156 serum samples we modified the original
assay (with the support of N.Groome) so that results were obtained
on the same working day (duration time of the assay 66 h). In this
partially prospective trial, the former affected serum samples, as well
as four new affected ones, were reintroduced and reanalysed at random,
along with incoming routine screening samples. The sensitivity of
the modified assay was 0.3 pg/tube (4 pg/ml); the variation within
and between assays was 3.9 and 18.7% respectively.
Our reference values in the second trimester were in agreement
with those in the leaflet of the kit (Table I).
Statistical analyses were performed using a software program
(NCCS 6.0; NCSS, Kaysville, UT, USA). The results were presented
as means 6 SD. An unpaired t-test was used to compare pairs of
means; when necessary, the unequal variance of data was taken
into account.
Results
In the preliminary pilot trial, 260 samples from second trimester
pregnancies (12–20 weeks), including 14 samples from trisomy
Table I. Dimeric inhibin-A reference values (pg/ml)
Weeks of pregnancy
Groome et al.a
Our laboratory
13
14
15
16
17
18
328
253
175
174
185
167
167
203
154
146
148
132
aCited
(193–612)
(172–543)
(67–642)
(82–424)
(96–391)
(50–324)
(11–365)
(44–662)
(60–267)
(59–260)
(52–259)
(48–239)
in Aitken et al. (1996).
Table II. Dimeric inhibin-A multiple of the median (MoM) concentrations
MoM value 6 SD
Trial 1
Unaffected (n 5 246)
Trisomy 21 affected (n 5 14)
Trial 2
Unaffected (n 5 1138)
Trisomy 21 affected (n 5 18)
0.93
2.15
P5
0.97
1.78
P,
6 0.64
6 0.75
0.0154
6 0.65
6 0.84
0.0001
Range
0.29–1.57
1.40–2.90
0.03–5.39
0.53–3.79
Table III. Comparative detection rates for Down’s syndrome
Cut-off 1/270 (trial 1)
Trisomy 21 detected
Sensitivity (%)
Specificity (%)
Cut-off 1/300 (trial 1)
Trisomy 21 detected
Sensitivity (%)
Specificity (%)
False-positive rate 5% (trial 2)
Trisomy 21 detected
Sensitivity (%)
Specificity (%)
Triple test
AFP/
inhibin-A
AFP/
inhibin-A/HCG
9/14
64.3
89.9
12/14
85.7
90
11/14
78.6
93.7
11/14
78.6
87.7
1/250
11/18
61.5
95.2
13/14
93
91
1/160
7/18
38.9
94.6
11/14
78.6
93.7
1/16
12/18
66.7
94.7
AFP 5 α-fetoprotein; HCG 5 human chorionic gonadotrophin.
21 pregnancies, were analysed for dimeric inhibin-A. There
was only a minor fluctuation in the absolute dimeric inhibin-A
concentrations in the second trimester of pregnancy. The mean
concentration in unaffected pregnancies (n 5 246) was 167
pg/ml, while in affected pregnancies (n 5 14) the value was
375 pg/ml. In Table II, dimeric MoM values (mean 6 SD)
for affected and unaffected pregnancies are shown for both
trials. Dimeric inhibin-A MoM concentrations were higher in
affected pregnancies at a significance of P 5 0.01 for the
pilot trial.
There was only a small overlap between affected and
unaffected pregnancies. In Table III detection results are shown
for the two trials using three different methods. The second
column shows the results for AFP 1 inhibin-A, while in the
last column free oestriol is substituted by the dimeric inhibinA results in the algorithm for the triple-risk calculation. Using
a cut-off value of 1/270, we missed five trisomy 21 pregnancies
in our routine classic triple-test screening programme; if we
used the cut-off value of 1/300 we missed only three pregnancies (first column). With AFP and dimeric inhibin-A as
screening parameters (second column), only two affected
745
M.A.Renier et al.
Figure 1. Multiple of the median (MoM) inhibin-A concentrations in normal (u) and affected (d) pregnancies.
pregnancies were missed with the cut-off value of 1/270 and
only one with a cut-off value of 1/300. On inclusion of HCG
and with a cut-off value of 1/270, three affected pregnancies
were missed. The results we obtained in this small series were
promising and encouraging, but there was still a high falsepositive rate (13%) and therefore the associated risk of an
unnecessary amniocentesis was high.
Following the results from our first trial, in the second trial
we increased the number of samples from 260 to 1156,
including 18 affected pregnancies. Figure 1 depicts the distribution of dimeric inhibin-A MoM concentrations between weeks
12 and 20 of pregnancy in the second trial for normal and
affected pregnancies. All the dimeric inhibin-A results (Figure
1) were obtained using the modified enzyme-linked immunosorbent assay test (short incubation times) described above.
Inhibin-A MoM concentrations were significantly higher (P ,
0.001) in trisomy pregnancies than in unaffected pregnancies
(Table II), but absolute concentrations were lower in our study
than in other published research (Wallace et al., 1996).
In analogy with the then current publications on dimeric
inhibin-A and Down’s syndrome screening, in our second trial
we focused on a 5% false-positive ratio rather than a fixed
cut-off risk value (Table III).
The incorporation of dimeric inhibin-A in the algorithm
(columns 2 and 3) resulted in totally different combined risk
values in comparison with the classic triple test. In contradiction
to the first trial, the combination of AFP 1 intact HCG 1
dimeric inhibin-A scored better than the combination of
AFP 1 inhibin-A alone, reaching a sensitivity of 66.7% and
a specificity of 94.7% (Table III; column 3).
746
Discussion
Results of the second trial are highly representative of results
from day to day practice. From a statistical viewpoint, however,
ideally a cohort of at least 18 000 samples is needed when 18
Down’s syndrome pregnancies are included (a risk of 1/1000
in the second trimester of pregnancy). Compared with the
literature, our number of controls is high in comparison with
the number of affected pregnancies included (Table IV).
Most studies in the literature are retrospective and could
tend to overestimate performance because the ideal conditions
of a study may not apply in practice. One study (Wallace
et al., 1996) obtained a higher prediction rate (75% at a 5%
false-positive rate) by adding inhibin-A to the AFP and β-HCG
combination in the first trimester screening. They found a
strong correlation between inhibin-A and intact HCG, which
excluded the use of intact HCG 1 inhibin-A in the algorithm
(two non-independent variables). We confirmed this correlation
(r 5 0.507) in the second trimester and thus were forced to
omit HCG from the algorithm, which decreased the detection
ratio (P 5 0.095) significantly in our risk calculation.
Wald et al. (1997) recently reduced their estimated detection
rate from 78 to 75% for a 5% false-positive rate using
the quadruple test (AFP, unconjugated oestriol, free β-HCG,
inhibin-A with maternal age, maternal weight adjustment and
an ultrasound scan examination to estimate gestational age).
They revised their underestimated SD of inhibin-A in pregnancies with Down’s syndrome. This highlights the importance
of a laboratory’s own assay standards for inhibin and explains
why our inhibin serum concentrations may have slightly
different values especially for our modified ‘short’ assay.
Dimeric inhibin-A in screening for Down’s syndrome
Table IV. Published results
Reference
No. of
trisomy 21
No. of
controls
MoM
inhibin-A
% False-positive
rate (%)
% Sensitivity
Wallace et al. (1996)
Aitken et al. (1996)
Wallace et al. (1995)
Cuckle et al. (1995)
Cuckle (1996)
21
14
23
19
44
150
206
89
95
202
2.20
1.79
2.46
1.60
2.24
5.3
–
4.0
–
5.0
62
–
65
–
75
MoM 5 multiple of the median.
Table V. Influencing factors on inhibin-A multiple of the median (MoM)
concentrations
Number
Smoking
Yes
No
69
627
Vaginal bleeding
Yes
No
67
1088
Ovulation induction
Yes
No
142
1014
Inhibin-A MoM
1.442 6 0.908
0.958 6 0.647
P , 0.0001
0.994 6 0.506
0.986 6 0.674
NS
0.935 6 0.532
0.993 6 0.681
NS
NS 5 not significant.
We confirmed that dimeric inhibin-A is not predictive for
other chromosomal anomalies, and the inclusion of inhibin-A
in our multi-marker screening programme did not contribute
to the detection of trisomy 18 (n 5 4; inhibin-A MoM 5
0.61 6 0.16). We also confirmed the significantly higher
inhibin-A MoM concentrations in trisomy 21 pregnancies.
Wallace et al. (1996) reported that inhibin-A alone offered a
detection rate of 62% and could detect cases previously
undetected by routine screening in the second trimester. We
counted two patients with a false-negative result on the
inhibin-A triple test who were, however, detected with the
classic triple test; in these cases there was no advantage to
using inhibin-A in triple testing. However we could confirm
the theoretical advantage of the use of dimeric inhibin-A above
other current parameters (especially unconjugated oestriol)
because of its very minor fluctuations between 14 and 18
weeks of pregnancy (Table I). Therefore inaccuracies in
estimating the gestational stage of pregnancy would have
minor consequences in the calculation of inhibin-A MoM
concentrations.
Several authors have suggested that other factors which
could influence maternal serum MoM concentration, such as
vaginal bleeding (Cuckle et al., 1994b), ovulation induction
(Heinonen et al., 1996; Ribbert et al., 1996), increasing
gravidity and parity (Mooney et al., 1995) and smoking
(Cuckle et al., 1992), should be taken into account.
We rigorously recorded these factors (Table V) and noticed
a highly significant influence of smoking, which therefore
constitutes a possible correction factor in future protocols. Our
results are consistent with previous retrospective studies with
respect to some but not all results. Therefore an enlarged
multicentre study over a longer period of time is necessary
before making firm conclusions about the use of dimeric
inhibin-A in second trimester Down’s syndrome screening
programmes. We believe that the data are not yet sufficiently
consistent to incorporate inhibin-A in day-to-day laboratory
screening procedures.
Acknowledgements
The advice of Nigel Groome while adapting his inhibin assays is
gratefully acknowledged.
References
Aitken, D.A., Wallace, E.M., Crossley, J.A. et al. (1996) Dimeric inhibin A
as a marker for Down’s syndrome in early pregnancy. N. Engl. J. Med.,
334, 1231–1236.
Brambati, B., MacIntosh, M.C.M., Teisner, B. et al. (1993) Low maternal
serum levels of pregnancy associated plasma protein A (PAPP-A) in the
first trimester in association with abnormal fetal karyotype. Br. J. Obstet.
Gynaecol., 100, 324–326.
Brock, D.J.H., Barron, L., Holloway, S. et al. (1990) First trimester maternal
serum biochemical indicators in Down’s syndrome. Prenat. Diagn., 10,
245–251.
Cheng, E.Y., Luthy, D.A., Zebelman, A.M. et al. (1993) A prospective
evaluation of a second-trimester screening test for fetal Down’s syndrome
using maternal serum α-fetoprotein, hCG, and unconjugated estriol. Obstet.
Gynecol., 81, 72–77.
Cuckle, H.S., Wald, N.J. and Thompson, S. (1987) Estimating a woman’s risk
of having a pregnancy associated with Down’s syndrome using her age and
serum α-fetoprotein level. Br. J. Obstet. Gynaecol., 94, 387–402.
Cuckle, H.S., Wald, H.J. and Densem, J.W. (1992) The effect of smoking in
pregnancy on maternal serum α-foetoprotein, unconjugated estriol, HCG,
progesterone and DHEA-sulphate levels. Br. J. Obstet. Gynaecol., 92, 272.
Cuckle, H.S., Holding, S. and Jones, R. (1994a) Maternal serum inhibin levels
in second-trimester Down’s syndrome pregnancies. Prenat. Diagn., 14,
387–390.
Cuckle, H., van Oudgaarden, E.D., Mason, G. et al. (1994b) Taking account
of vaginal bleeding in screening for Down’s syndrome. Br. J. Obstet.
Gynaecol., 101, 948–953.
Cuckle, H.S., Holding, S., Jones, R. et al. (1995) Maternal serum dimeric
inhibin A in second trimester Down’s syndrome pregnancies. Prenat.
Diagn., 15, 385–386.
Cuckle, H. (1996) Evaluation of screening programmes. Baillières Clin.
Obstet. Gynecol., 10, 631–645.
Haddow, J.E., Palomaki, G.E., Knight, G.J. et al. (1992) Prenatal screening
for Down’s syndrome with use of maternal serum markers. N. Engl. J.
Med., 327, 588–593.
Heinonen, S., Ryynänen, M., Kirkinem, P. et al. (1996) Effect of in vitro
fertilization on human chorionic gonadotropin serum concentrations and
Down’s syndrome screening. Fertil. Steril., 66, 398–403.
Ind, T.E., Iles, R.K., Wathen, N.C. et al. (1993) Low levels of amniotic fluid
placental alkaline phosphatase in Down’s syndrome. Br. J. Obstet. Gynaecol.,
100, 847–849.
Kratzer, P.G., Golbus, M.S., Monroe, S.E. et al. (1991) First-trimester
aneuploidy screening using serum human chorionic gonadotrophin (hCG),
free αhCG, and progesterone. Prenat. Diagn., 11, 751–765.
Merkatz, I.R., Nitowsky, H.M. and Macri, J.N. (1984) An association between
low maternal serum α-fetoprotein and fetal chromosomal abnormalities.
Am. J. Obstet. Gynecol., 148, 886–894.
747
M.A.Renier et al.
Mooney, R.A., Arvan, D.A., Saller, D.N. et al. (1995) Decreased maternal
serum hCG levels with increasing gravidity and parity. Obstet. Gynecol.,
86, 900–905.
Muttukrishna, S., Knight, P.G., Groome, N.P. et al. (1997) Activin A and
inhibin A as possible endocrine markers for pre-eclampsia. Lancet, 349,
1285–1288.
Palomaki, G.E., Knight, G.J., Haddow, J.E. et al. (1992) Prospective
intervention trial of a screening protocol to identify fetal trisomy 18 using
maternal serum α-fetoprotein, unconjugated oestriol, and human chorionic
gonadotropin. Prenat. Diagn., 12, 925–930.
Petraglia, F., Sawchenko, P., Lim, A.T. et al. (1987) Localization, secretion,
and action of inhibin in human placenta. Science, 237, 187–189.
Ribbert, L.S.M., Kornman, L.H., De Wolf, B.T.H. et al. (1996) Maternal
serum screening for fetal Down’s syndrome in IVF pregnancies. Prenat.
Diagn., 16, 35–38.
Riley, S.C., Wathen, N.C., Chard, T. et al. (1996) Inhibin in extra-embryonic
coelomic and amniotic fluids and maternal serum in early pregnancy. Hum.
Reprod., 11, 2772–2776.
Spencer, K., Wood, P.J. and Anthony, F.W. (1993) Elevated levels of maternal
serum inhibin immunoreactivity in second trimester pregnancies affected
by Down’s syndrome. Ann. Clin. Biochem., 30, 219–220.
Van Lith, J.M.M., Pratt, J.J., Beekhuis, J.R. et al. (1992) Second trimester
maternal serum immunoreactive inhibin as a marker for fetal Down’s
syndrome. Prenat. Diagn., 12, 801–806.
Wald, N.J. (1991) Maternal serum unconjugated oestriol and human chorionic
gonadotrophin levels in twin pregnancies: implications for screening for
Down’s syndrome. Br. J. Obstet. Gynaecol., 98, 905–909.
Wald, N.J., Densem, J.W. and George, L. (1997) Inhibin-A in Down’s
syndrome pregnancies: revised estimate of standard deviation. Prenat.
Diagn., 17, 285–290.
Wallace, E.M., Grant, V.E., Swanston, I.A. et al. (1995) Evaluation of maternal
serum dimeric inhibin A as a first-trimester marker of Down’s syndrome.
Prenat. Diagn., 15, 359–362.
Wallace, E.M., Swanston, I.A., McNeilly, A.S. et al. (1996) Second trimester
screening for Down’s syndrome using maternal serum dimeric inhibin A.
Clin. Endocrinol., 44, 17–21.
Received on May 28, 1997; accepted on November 19, 1997
748