First trimester maternal serum free β human chorionic gonadotrophin and pregnancy associated plasma protein A as predictors of pregnancy complications

Authors

  • Charas Y. T. Ong,

    Research Fellow (Fetal Medicine)
    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, London
    Search for more papers by this author
  • Adolfo W. Liao,

    Research Fellow (Fetal Medicine)
    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, London
    Search for more papers by this author
  • Kevin Spencer,

    Consultant (Biochemistry)
    1. Endocrine Unit, Clinical Biochemistry Department, Harold Wood Hospital, Romford, Essex
    Search for more papers by this author
  • Shama Munim,

    Research Fellow (Fetal Medicine)
    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, London
    Search for more papers by this author
  • Kypros H. Nicolaides

    Professor (Fetal Medicine), Corresponding author
    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, London
      Correspondence: Professor K. Nicolaides, Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, Denmark Hill, London SE5 8RX, UK.
    Search for more papers by this author

Correspondence: Professor K. Nicolaides, Harris Birthright Research Centre for Fetal Medicine, King's College Hospital Medical School, Denmark Hill, London SE5 8RX, UK.

Abstract

Objective To examine the value of first trimester maternal serum free β human chorionic gonadotrophin (β hCG) and pregnancy associated plasma protein A (PAPP-A) as predictors of pregnancy complications.

Design Screening study.

Setting Antenatal clinics.

Population Singleton pregnancies at 10–14 weeks of gestation.

Methods Maternal serum free β hCG and PAPP-A were measured at 10–14 weeks of gestation in 5584 singleton pregnancies. In the 5297 (94.9%) pregnancies with complete follow up free β hCG and PAPP-A were compared between those with normal outcome and those resulting in miscarriage, spontaneous preterm delivery, pregnancy induced hypertension or fetal growth restriction and in those with pre-existing or gestational diabetes.

Results Maternal serum PAPP-A increased and β hCG decreased with gestation. The multiple of median maternal serum PAPP-A was significantly lower in those pregnancies resulting in miscarriage, pregnancy induced hypertension, growth restriction and in those with pre-existing or gestational diabetes mellitus, but not in those complicated by spontaneous preterm delivery. The level was < 10th centile of the reference range in about 20% of the pregnancies that subsequently resulted in miscarriage or developed pregnancy induced hypertension or growth restriction, and in 27% of those that developed gestational diabetes. Maternal serum free β hCG was < 10th centile of the reference range in about 15% of the pregnancies that subsequently resulted in miscarriage or developed pregnancy induced hypertension or growth restriction, and in 20% of those that developed gestational diabetes.

Conclusion Low maternal serum PAPP-A or β hCG at 10–14 weeks of gestation are associated with subsequent development of pregnancy complications.

INTRODUCTION

The most effective method of screening for chromosomal abnormalities is by a combination of fetal nuchal translucency thickness and maternal serum free β hCG and PAPP-A at 10–14 weeks of gestation1. It is estimated that the sensitivity of detecting trisomy 21 is about 90% for a false positive rate of 5%1.

In this study, we investigate the potential additional value of early first trimester biochemical screening in the detection of pregnancies at high risk for adverse outcomes. Certainly in the second trimester there is some evidence that elevated levels of free β hCG are associated with the subsequent development of pregnancy complications, such as preterm delivery, pregnancy induced hypertension and fetal growth restriction2–6. Since there is a good correlation in serum free β hCG between the first and second trimesters of pregnancy7,8, it is reasonable to assume that first trimester serum free β hCG may be predictive of adverse pregnancy outcomes. Similarly, there is some evidence that low levels of maternal serum PAPP-A in the first trimester may be associated with miscarriage, low birthweight and preterm delivery9–14.

METHODS

In the Harold Wood Hospital (Essex) and the King's College Hospital (London), all pregnant women are offered screening for trisomy 21 by a co mbination of fetal nuchal translucency and maternal serum free P hCG and PAPP-A in a One Stop Clinic for Assessment of Risk1. Free β hCG and PAPP-A were measured by a KRYPTOR analyser (a random access immunoassay analyser) using time-resolved amplified cryptate emission technology, which provides a result within 30 minutes. Demographic characteristics, ultrasound findings and the results of biochemical testing are entered into a computer database at the time of assessment.

A search was made of the database to identify all singleton pregnancies, which had first trimester biochemical testing from the onset of the service in May 1998 until 30th July 1999. The hospital notes and delivery suite records were then searched for each one of these patients to identify any pregnancy complications and obtain delivery details.

The following outcome measures were examined: miscarriage, spontaneous preterm delivery, gesta-tional diabetes, pregnancy induced hypertension and intrauterine growth restriction. Miscarriage was defined as spontaneous delivery < 24 completed weeks of pregnancy. Preterm and early preterm delivery were defined by spontaneous delivery < 37 and < 34 completed weeks, respectively. Gestational diabetes mellitus was diagnosed by glucose tolerance test according to the World Health Organisation criteria15. A 75 g dose of glucose was taken orally after an overnight fast and the diagnosis of diabetes was made if the fasting blood glucose level was 8.0 mmol/L and/or the level two hours after the glucose dose was ≥ 11.0 mmol/L. At the Harold Wood Hospital the policy was to offer glucose tolerance testing at about 24 weeks to women with a family history of diabetes, those who had previously delivered a baby with a birthweight of ≥ 4.5 kg, those with a previous unexplained fetal loss, women weighing > 100 kg, those with a history of gestational or latent diabetes, and women with glycosuria or a large for gestation fetus in their current pregnancy. At the King's College Hospital glucose tolerance testing was carried out at about 30 weeks in women with a random blood glucose level > 6.7 mmol/L at 28 weeks. Pregnancy induced hypertension was defined by a diastolic blood pressure of 110 mmHg or more on any one occasion or a diastolic blood pressure of 90 mmHg or more on two consecutive occasions four hours apart in women with no pre-existing hypertensive or renal disease, pregnancy induced hypertension was subdivided into pro-teinuric and nonproteinuric, depending on the presence or absence of either > 300 mg of total protein in a 24-hour urine collection or an 1+ albumin on a reagent strip16. Intrauterine growth restriction was defined by a birthweight below the 10th, 5th or 3rd centiles for gestational age17.

Statistical analysis

Median levels of free P hCG and PAPP-A for the interval between 73 and 97 gestational days were determined from the study population and multiple of the median values were calculated for each subject. After decadic logarythmical transformation of free β hCG and PAPP-A concentrations the distribution for each gestational age (calculated from the fetal crow-rump length obtained at ultrasound examination) was demonstrated to be normal using Kolmogorov-Smirnov goodness of fit test. The multiple of median values for each clinical group were calculated. Decadic logarythmi-cally transformed multiple of median values were compared between the different outcome groups and the control group using t test for independent samples with unequal variance.

RESULTS

During the study period, 5584 singleton pregnancies were examined at 10–14 weeks of gestation. Complete follow up was obtained from 5297 (94.9%) of the patients. In 4551 (85.9%) cases the women were Caucasian, the mean maternal age was 29.2 years (15–45 years) and 1741 (32.9%) were primigravidae. In 4297 cases (81.1%) the pregnancies were unaffected by chromosomal or structural anomalies or adverse pregnancy outcome and this group formed the control population. Maternal serum PAPP-A increased significantly with gestation (median PAPP-A = 10[(0.179 × gestational age in weeks) − 1.74], r2= 0.238) (Fig. 1). The multiple of median maternal serum PAPP-A was significantly lower in those pregnancies resulting in miscarriage, proteinuric and non proteinuric pregnancy induced hypertension, growth restriction and in those with pre-existing or gestational diabetes mellitus, but not in those complicated by spontaneous preterm delivery (Table 1). Maternal serum PAPP-A was < 10th centile of the reference range in about 20% of the pregnancies that subsequently resulted in miscarriage or developed pregnancy induced hypertension or growth restriction, and in 27% of those that developed gestational diabetes (Table 2).

Figure 1.

Maternal serum PAPP-A with gestation showing median, 5th and 95th centiles.

Table 1.  Median MoM of maternal serum pregnancy associated plasma protein A (PAPP–A) and free βbT human chorionic gonadotrophin (βhCG) in each adverse outcome group. Values are given as n or median (95% CI), unless otherwise indicated. MOM = multiple of the median.
  PAPP–AFree β hCG
 nMedian MoMPMedian MoMP
Control42971.049 (1.028,1.070) 1.010 (0.984,1.036) 
Miscarriage540.755 (0.547, 0.963)0.0330.900 (0.709, 1.090)0.083
Spontaneous preterm delivery     
 < 37 weeks1920.978 (0.872,1.084)0.0550.922 (0.767,1.076)0.406
 < 34 weeks470.933 (0.743, 1.123)0.0880.920 (0.611,1.229)0.660
Pregnancy induced hypertension     
 Proteinuric1350.903 (0.802, 1.005)0.0010.879 (0.740,1.019)0.012
 Nonproteinuric800.837 (0.703, 0.972)0.0010.924 (0.655,1.194)0.163
Birthweight     
 < 10th centile3950.967 (0.895, 1.040)0.0021.033 (0.931,1.135)0.993
 < 5th centile1710.900 (0.798,1.002)0.0011.017 (0.877,1.157)0.565
 < 3rd centile1030.819 (0.694, 0.943)0.0011.038 (0.851,1.215)0.500
Diabetes mellitus     
 Pre–existing350.839 (0.674, 1.004)0.0100.748 (0.000, 2.039)0.068
 Gestational490.848 (0.691,1.006)0.0020.783 (0.587, 0.979)0.004
Table 2.  Sensitivity of maternal serum pregnancy associated plasma protein A (PAPP–A) and free βbT human chorionic gonadotrophin (β hCG) below the 5th centile, 10th centile and the median of the appropriate reference range for gestation in the prediction of pregnancy complications. Values are given as n or n (%).
  PAPP–AFree β hCG
 n<5th<10th<median<5th<10th<median
TOTAL5297279 (5.3)540 (10.2)2566 (48.4)238 (4.5)472 (8.9)2659 (50.2)
Miscarriage544 (7.4)11 (20.4)32 (59.3)5 (9.3)9 (16.7)29 (53.7)
Spontaneous preterm delivery       
 <37 weeks19215 (7.8)26 (13.5)102 (53.1)11 (5.7)23 (12.0)104 (54.2)
 <34 weeks477 (14–9)9 (19–1)28 (59–6)4 (8–5)6 (12–8)25 (53–2)
Pregnancy induced hypertension       
 Proteinuric13515 (11–1)26 (19–3)79 (58–5)10 (7–4)22 (16–3)79 (58–5)
 Nonproteinuric8011 (13–8)20 (25–0)51 (63–8)7 (8–8)12 (15–0)43 (53–8)
Birthweight       
 <10th centile39531 (7–8)49 (12–4)203 (51–4)24 (6–1)48 (12–2)193 (48–9)
 <5th centile17122 (12–9)31 (18–1)97 (56–7)12 (7–0)27 (15–8)84 (49–1)
 <3rd centile10315 (14–6)20 (19–4)63 (61–2)8 (7–8)18 (17–5)50 (48–5)
Diabetes mellitus       
 Pre–existing352 (5–7)5 (14–3)22 (62–9)3 (8–6)7 (20–0)25 (71–4)
 Gestational499 (18–4)13 (26–5)30 (61–2)6 (12–2)9 (18–4)33 (67–3)

Maternal serum free β hCG decreased significantly with gestation (median free β hCG = 10[(−0.079 × gestational age in weeks) + 2.53], r2= 0.047) (Fig. 2). The multiple of median maternal serum free β hCG was significantly lower in those pregnancies resulting in proteinuric pregnancy induced hypertension and gestational diabetes mellitus (Table 1). Maternal serum free β hCG was < 10th centile of the reference range in about 15% of the pregnancies that subsequently resulted in miscarriage or developed pregnancy induced hypertension or growth restriction, and in 20% of those that developed gestational diabetes (Table 2).

Figure 2.

Maternal serum free β hCG with gestation showing median, 5th and 95th centiles.

DISCUSSION

This screening study has demonstrated that maternal serum PAPP-A at 10–14 weeks of gestation is lower in those pregnancies resulting in miscarriage, pregnancy induced hypertension, growth restriction and in those with pre-existing diabetes mellitus or those that develop gestational diabetes. Although maternal serum free β hCG was also decreased in pregnancies that subsequently developed these complications, significant differences were shown only for those resulting in pro-teinuric pregnancy induced hypertension and gestational diabetes mellitus

The association between diabetes mellitus and decreased levels of maternal serum human chorionic gonadotrophin is well described and in second trimester serum screening for trisomy 21 the appropriate corrections are made for such pregnancies18. This study has demonstrated that both free β hCG and PAPP-A are decreased in the first trimester both in pregnancies with pre-existing diabetes and in those that subsequently develop gestational diabetes. Although this observation suggests that the metabolic derangements responsible for the altered levels of free β hCG and PAPP-A predate the diagnosis of gestational diabetes, the measurement

of these placental proteins at 10–14 weeks does not provide sensitive prediction of gestational diabetes. The finding that PAPP-A is decreased in pregnancies with maternal diabetes mellitus is compatible with the results of a previous study that reported decreased levels of both human placental lactogen and PAPP-A in such pregnancies during the first trimester19. Our findings have implications on first trimester biochemical screening for trisomy 21 by a combination of free β hCG and PAPP-A and this requires further investigation.

The finding that in pregnancies resulting in miscarriage maternal serum PAPP-A is decreased is compatible with the results of previous studies9–12. It is possible that low maternal serum levels are the consequence of impaired placentation and/or placental function which can also cause miscarriage. Certainly in chromosomal abnormalities, such as trisomies 18 and 13 and maternally derived triploidy, which are associated with a very high rate of early pregnancy loss, the maternal serum concentration of PAPP-A is very low20–22. However, these chromosomal abnormalities are also associated with very low levels of maternal serum free β hCG but the concentration of this protein is not altered in pregnancies resulting in miscarriage.

The development of pregnancy induced hypertension and fetal growth restriction are thought to be the consequence of impaired placentation due to inadequate tro-phoblastic invasion of the maternal spiral arteries23. Low maternal serum PAPP-A at 10–14 weeks of gestation may be a marker of inadequate placentation and this may be the explanation for the association of low PAPP-A and subsequent development of pregnancy induced hypertension and fetal growth restriction. In contrast to hypertension and growth restriction, spontaneous preterm delivery was not associated with altered levels of maternal serum PAPP-A.

This study has also found an association between low levels of maternal serum free β hCG and subsequent development of proteinuric pregnancy induced hypertension. Several previous studies reported that increased levels of free β hCG during the second trimester are associated with adverse pregnancy outcome, including pregnancy induced hypertension, fetal growth restriction, preterm delivery, miscarriage and perinatal death2–6. A possible explanation for the apparent difference in serum concentrations of free β hCG during the first and second trimesters in pregnancies developing complications is that the low levels at 10–14 weeks are the consequence of impaired placentation and smaller placental mass, whereas the high levels in the second trimester may be the result of hypoperfusion-related stimulation of production of this hormone.

Previous retrospective studies examining the relationship between maternal serum human chorionic gonadotrophin and PAPP-A in the first trimester and the subsequent development of pregnancy complications provided conflicting results. Johnson et al. 13 measured maternal serum PAPP-A and intact human chorionic gonadotrophin levels at 7–13 weeks of gestation in 62 pregnancies conceived by in vitro fertilisation and embryo transfer. They found significantly lower PAPP-A levels in the eight women that subsequently delivered preterm and a weak association with low birthweight. Pedersen et al.14, measured maternal serum PAPP-A at 8–14 weeks in 93 pregnancies and found an inverse correlation with duration of pregnancy and positive correlation with birthweight. In contrast, a case control study involving 800 patients reported the absence of significant association between first trimester free β hCG and PAPP-A levels and subsequent preterm delivery (n= 87) or small for gestational age fetuses (n= 73)24.

At 10–14 weeks of gestation low maternal serum PAPP-A and free β hCG are associated with the development of pregnancy complications, including miscarriage, hypertensive disease, gestational diabetes and fetal growth restriction. However, the sensitivity for these pregnancy complications is only about 20% for a screen positive rate of 10%. This is similar to findings with abnormal serum biochemistry in the second trimester. In the future it is likely that a combination of biochemical and sonographic markers would be used to assess individual risk for pregnancy complications in a similar way to that of screening for chromosomal defects.

Ancillary