Predictive value of early human chorionic gonadotrophin serum profiles for fetal growth retardation

Bassam Haddad1, Fady Abirached2, Christine Louis-Sylvestre1, Josiane Le Blond2, Bernard-Jean Paniel1 and Jean-René Zorn2,3

1 Service de Gynécologie-Obstétrique, CHI Créteil, 40 avenue de Verdun, 94010 Créteil and 2 Clinique Universitaire Baudelocque, Hôpital Cochin, 123 BD de Port-Royal, 75014 Paris, France


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The possibility of using first trimester maternal serum human chorionic gonadotrophin (HCG) profiles to predict fetal growth retardation (FGR) was tested in 236 women with singleton pregnancies obtained after in-vitro fertilization (IVF). Pregnancies were monitored by serial analysis (two or more) of serum HCG at at least 48 h intervals. Serum was obtained between the 13th and the 35th day after conception (i.e. on the day of IVF). Early miscarriage occurred in 23.7% and FGR in 10.9% of pregnancies. Serum HCG profiles were higher than the 90th and lower than the 10th percentile in 12.3% and 19.5% of the cases respectively. FGR was significantly more frequent in women with serum HCG profiles lower than the 10th percentile than in women with normal profiles (45.5% versus 7.2%; P < 0.001), with a relative risk of 6.5 (95% confidence interval 2.7–15.6). FGR rates were similar in women with normal and high profiles of serum HCG. Pre-eclampsia and premature delivery rates were similar in women with normal and abnormal profiles of serum HCG. First trimester serum HCG should be further investigated as a potential marker of FGR.

Key words: fetal growth retardation/human chorionic gonadotrophin


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Fetal growth retardation (FGR) is an important cause of fetal morbidity and mortality. Selection of individuals at risk of FGR is a concern as these pregnancies should be closely managed.

Elevated concentrations in second trimester of human chorionic gonadotrophin (HCG) have been associated with Down's syndrome, triploidy or multiple gestation. In several reports studying the relationship between Down's syndrome and second trimester HCG concentrations, high rates of FGR were observed in women with abnormal concentrations of HCG and normally structured fetuses (Gonen et al., 1992Go; Lieppman et al., 1993Go; Wenstrom et al., 1994Go; Morssink et al., 1995Go; Muller et al., 1996Go).

This study was designed to investigate the possible relationship between maternal serum HCG profiles obtained early in the first trimester and FGR in women whose pregnancies were obtained after in-vitro fertilization (IVF). IVF pregnancies were chosen because gestational age is clearly defined and serial blood samples are routinely performed to evaluate early gestation.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Patients
In all, 236 singleton IVF pregnancies obtained between 1994 and 1995 in the Clinique Universitaire Baudelocque, Paris, France were retrospectively analysed (Table IGo).


View this table:
[in this window]
[in a new window]
 
Table I. Patients characteristics (n = 236)
 
All patients had an endovaginal ultrasonography with a 5 MHz probe (Combison 310; Kretz®; Soframed, Sarl, Ittlenheim, France), performed between 5 and 7 weeks gestation to ascertain the presence of a unique intrauterine gestational sac.

All women had serial analysis (two or more) of serum HCG (4.14 ± 1.39, range: 2–9) at 48 h intervals at least. They were evaluated for the risks of the following adverse pregnancy outcomes: FGR without any malformation (birthweight below the 10th percentile of local standard curves), preterm delivery, pre-eclampsia (defined as a diastolic blood pressure >90 mmHg at 4 h apart and 24 h proteinuria >=300 mg or ++ dipstick albuminuria at 4 h apart).

Hormone assays
Serum HCG was assayed by using a commercial reagent kit (Amerlite HCG 60®; Johnson and Johnson Clinical Diagnostics, Chatenay Malabry, France) between the 13th and 35th day after conception (i.e. the day of IVF); 85.9% of all the samples (n = 977) were obtained before the 28th day (Figure 1Go). The values were plotted against the 10th and 90th percentile values determined over the same gestation period in our laboratory for a control series of 333 IVF singleton pregnancies followed during 7 weeks gestation (unpublished data, Figure 2Go). The profile of serum HCG values was considered abnormal when <10th or >90th percentile.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 1. Percentage of serum human chorionic gonadotrophin (HCG) samples obtained on each day between the 13th and 35th day after conception (n = 977). D = day after conception.

 


View larger version (19K):
[in this window]
[in a new window]
 
Figure 2. Reference curves (10th and 90th percentiles) for serum human chorionic gonadotrophin (HCG) profiles in a control series of 333 in-vitro fertilization singleton pregnancies. p = percentile; D = day after conception.

 
Statistical analysis
Statistical analyses were performed by using {chi}2 test with Yates' correction when necessary and Student's t test. P < 0.05 was considered to be significant. Relative risk (RR) and 95% confidence intervals (CI) were calculated when appropriate.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Serum HCG profiles were normal, <10th percentile and >90th percentile in 68.2%, 19.5% and 12.3% of the study cases respectively. In women with serum HCG profiles <10th percentile, the spontaneous early abortion rate was significantly higher than in women with normal profiles ({chi}2 = 76.6, P < 0.001), with a significant increase of the RR at 6.5 (95% CI: 4.1–10.1) (Table IIGo).


View this table:
[in this window]
[in a new window]
 
Table II. Early maternal serum human chorionic gonadotrophin (HCG) profiles and early miscarriage
 
Nineteen cases of FGR were observed in the study group. Rate of FGR in women with serum HCG profiles <10th percentile was significantly higher than in women with normal profile (45.5% versus 7.2%; P < 0.001, Table IIIGo), with a significant increase of the RR at 6.5 (95% CI: 2.7–15.6) (Table IVGo). Rates of FGR in women with normal and high profiles of serum HCG were not statistically different (Table IIIGo). Analysis of the incidence of FGR was also performed after exclusion of the three women with pre-eclampsia (one having a FGR and an HCG profile <10th percentile, and two having appropriate growth fetuses and normal HCG profiles). This showed that rate of FGR in women with serum HCG profiles <10th percentile was significantly higher than in women with normal profiles (40% versus 7.6%; {chi}2 = 8.34, P < 0.01), with a significantly increased RR of 5.6 (95% CI: 2.1–14.7). Rates of FGR in women with normal and high profiles of serum HCG were not statistically different after the exclusion of women with pre-eclampsia (14.8% versus 7.6%, respectively).


View this table:
[in this window]
[in a new window]
 
Table III. Pregnancy outcomes after the first trimester, and early maternal serum human chorionic gonadotrophin (HCG) profiles (n = 180)
 

View this table:
[in this window]
[in a new window]
 
Table IV. Early maternal serum human chorionic gonadotrophin (HCG) profiles and fetal growth retardation (FGR) (n = 180)
 
No differences were found in rates of pre-eclampsia, late fetal loss and premature delivery between women with normal and abnormal profiles of serum HCG (Table IIIGo). Concerning the 10 women who delivered prematurely, three of them had hypotrophic fetuses (two with low and one with high serum HCG profiles).

Sensitivity, specificity, positive predictive value and negative predictive value of low serum HCG profile early in the first trimester for FGR were 26, 96, 45 and 92%, respectively (Table VGo).


View this table:
[in this window]
[in a new window]
 
Table V. Predictive value of early serum human chorionic gonadotrophin (HCG) profiles for fetal growth retardation (expressed in %)
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
A significant association between elevated concentrations of maternal serum HCG observed in the second trimester and FGR has been previously described (Gonen et al., 1992Go; Morssink et al., 1995Go; Benn et al., 1996Go; Muller et al., 1996Go). Therefore, this study was designed to investigate the possible relationship between maternal serum HCG obtained earlier in the first trimester and FGR. The study was performed in women undergoing IVF because gestational age was clearly defined and serial blood samples routinely performed to evaluate early gestation. A retrospective longitudinal analysis was chosen in order to evaluate the predictive values of early measurements of plasma HCG. Serial HCG assays (4.14 ± 1.39 per patient, range: 2–9) were routinely performed by our team to evaluate early gestation at 48 h intervals at least. This methodology reduces the risk of misinterpretation due to the possibly pulsatile secretion of HCG as previously suggested (Owens et al., 1981Go).

The proportions of elevated and low profiles of early first trimester serum HCG in the study population were at 12.3% and 19.5% respectively. The high frequency of low profiles is partially explained by the significant rate of spontaneous abortion and this association has already been described (Stabile et al., 1989Go; Henderson et al., 1992Go). The high rate of early miscarriage in this population is in accordance with that observed after IVF in North America (Assisted Reproductive Technology in the United States and Canada, 1998Go).

High serum HCG profiles (>90th percentile) obtained early in pregnancy in our series were not significantly associated with FGR, in agreement with the prior observation (Morssink et al., 1998Go) later in the first trimester. We have observed a RR of 2.1, which is in accordance with RR values obtained in the second trimester (Gonen et al., 1992Go; Lieppman et al., 1993Go; Benn et al., 1996Go; Muller et al., 1996Go) but it was not significant. It may be that our series was too short to allow a sufficient statistical power (0.23).

Low serum HCG profiles (<10th percentile) obtained early in the first trimester were strongly associated with FGR (RR 6.5). This association remained significant after controlling for pre-eclampsia (RR 5.6). In the second trimester, low HCG concentrations were also described as having a weak relationship with FGR (RR 1.2) (Morssink et al., 1995Go). However, in another study (Santolaya-Forgas et al., 1994Go), a low second trimester maternal serum HCG had no influence on pregnancy outcome. Recently, no relationship was demonstrated between serum free ßHCG concentrations, obtained at 73 ± 4 days of gestation, just before chorionic villus sampling, and FGR (Morssink et al., 1998Go). Our findings, which do not agree with those of Morssink et al. (1998), could be explained by the fact that gestational age in our study population was precisely known by the conception date whereas in Morssink et al. (1998) it was defined by last menstrual period and ultrasound examination, with a possible discrepancy of 7 days.

The association between early low serum HCG profile and subsequent FGR may reflect an impaired trophoblast function which might later lead to placental insufficiency. Recently, a relationship between FGR and hepatocyte growth factor (HGF), which is mainly synthesized and released in pregnant women by the placenta and trophoblast (Wolf et al., 1991Go), has been analysed (Aoki et al., 1998Go). Decreased maternal HGF concentrations were found in those women having a small for gestational age infant (Aoki et al., 1998Go). This may indicate a decreased release of HGF from the placenta as a result of decreased production reflecting placental insufficiency.

To our knowledge, this study is the first time the association between FGR and early low serum HCG has been described so early during gestation. The low sensitivity of low serum HCG profiles for prediction of FGR (26%, Table VGo) does not support the systematic use of such a marker as a screening test. In contrast, in women undergoing IVF, the good positive predictive value of early HCG profiles performed to monitor early gestation could be used to predict women at risk of FGR.


    Notes
 
3 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Aoki, S., Hata, T., Manabe, A. and Miyazaki, K. (1998) Decreased maternal circulating hepatocyte growth factor (HGF) concentrations in pregnancies with small for gestational age infants. Hum. Reprod., 13, 2950–2953.[Abstract/Free Full Text]

Assisted reproductive technology in the United states and Canada: 1995 results generated from the American Society for Reproductive Medecine/Society for Assisted Reproductive Technology Registry (1998) Society for Assisted Reproductive Technology and the American Society for Reproductive Medicine, Birmingham, Alabama. Fertil. Steril., 69, 389–398.[ISI][Medline]

Benn, P.A., Horn, D., Briganti, S. et al. (1996) Elevated second-trimester maternal serum hCG alone or in combination with elevated alpha-fetoprotein. Obstet. Gynecol., 87, 217–222.[Abstract/Free Full Text]

Gonen, R., Perez, R., David, M. et al. (1992) The association between unexplained second-trimester maternal serum hCG elevation and pregnancy complications. Obstet. Gynecol., 80, 83–86.[Abstract]

Henderson, D.J., Bennett, P.R. and Moore, G.E. (1992) Expression of human chorionic gonadotropin alpha and beta subunits is depressed in trophoblast from pregnancies with early embryonic failure. Hum. Reprod., 7, 1474–1478.[Abstract]

Lieppman, R.E., Williams, M.A., Cheng, E.Y. et al. (1993) An association between elevated levels of human chorionic gonadotropin in the midtrimester and adverse pregnancy outcome. Am. J. Obstet. Gynecol., 168, 1852–1856.[ISI][Medline]

Morssink, L.P., Kornman, L.H., Beekhuis, J.R. et al. (1995) Abnormal levels of maternal serum human chorionic gonadropin and alpha-fetoprotein in the second trimester: relation to fetal weight and preterm delivery. Prenat. Diagn., 15, 1041–1046.[ISI][Medline]

Morssink, L.P., Kornman, L.H., Hallahan, T.W. et al. (1998) Maternal serum levels of free beta-hCG and PAPP-A in the first trimester of pregnancy are not associated with subsequent fetal growth retardation or preterm delivery. Prenat. Diagn., 18, 147–152.[ISI][Medline]

Muller, F., Savey, L., Le Fiblec, B. et al. (1996) Maternal serum human chorionic gonadotropin level at fifteen weeks is a predictor for preeclampsia. Am. J. Obstet. Gynecol., 175, 37–40.[ISI][Medline]

Owens, O.M., Ryan, K.J. and Tulchinsky, D. (1981) Episodic secretion of human chorionic gonadotropin in early pregnancy. J. Clin. Endocrinol. Metab., 53, 1307–1309.[Abstract]

Santolaya-Forgas, J., Burd, L.I. and Burton, B.K. (1994) Clinical significance if low levels of second-trimester maternal serum human chorionic gonadotropin. Fetal Diagn. Ther., 9, 362–366.[ISI][Medline]

Stabile, I., Campbell, S. and Grudzinskas, J.G. (1989) Ultrasound and circulating placental protein measurements in complications of early pregnancy. Br. J. Obstet. Gynaecol., 96, 1182–1191.[ISI][Medline]

Vaillant, P., David, E., Constant, I. et al. (1996) Validity in nulliparas of increased beta-human chorionic gonadotropin at mid-term for predicting pregnancy-induced hypertension complicated with proteinuria and intrauterine growth retardation. Nephron, 72, 557–563.[ISI][Medline]

Wenstrom, K.D., Owen, J., Boots, L.R. et al. (1994) Elevated second-trimester human chorionic gonadotropin levels in association with poor pregnancy outcome. Am. J. Obstet. Gynecol., 171, 1038–1041.[ISI][Medline]

Wolf, H.K., Zarnegar, R., Oliver, L. et al. (1991) Hepatocyte growth factor in human placenta and trophoblast disease. Am. J. Pathol., 138, 1035–1043.[Abstract]

Submitted on March 12, 1999; accepted on August 4, 1999.