Serum Inhibin-A and PAPP-A2 in the prediction of pre-eclampsia during the first and second trimesters in high-risk women

Objectives: Maternal serum inhibin-A, pregnancy associated plasma protein-A (PAPP-A) and PAPP-A2 together with placental growth factor (PlGF), maternal risk factors and uterine artery pulsatility index (UtA PI) were analysed to study their ability to predict pre-eclampsia (PE). Study design: Serial serum samples for the nested case-control study were collected prospectively at 12–14, 18–20 and 26–28 weeks of gestation from 11 women who later developed early-onset PE (EO PE, diagnosis < 34 + 0 weeks of gestation), 34 women who developed late-onset PE (LO PE, diagnosis ≥ 34 + 0 weeks) and 89 controls. Main outcome measures: Gestational age -adjusted multiples of the median (MoM) values were calculated for biomarker concentrations. Multivariate regression analyses were performed to combine first trimester biomarkers, previously reported results on PlGF, maternal risk factors and UtA PI. Area under curve (AUC) values and 95% confidence intervals (CIs) for the prediction of PE and its subtypes were calculated. Results: A high first trimester inhibin-A predicted PE (AUC 0.618, 95%CI, 0.513–0.724), whereas PAPP-A and PlGF predicted only EO PE (0.701, 0.562–0.840 and 0.798, 0.686–0.909, respectively). At 26–28 weeks PAPP-A2 and inhibin-A predicted all PE subtypes. In the multivariate setting inhibin-A combined with maternal prepregnancy body mass index, prior PE and mean UtA PI predicted PE (0.811,0.726–0.896) and LO PE (0.824, 0.733–0.914). Conclusions: At first trimester inhibin-A show potential ability to predict not only EO PE but also LO PE whereas PlGF and PAPP-A predict only EO PE. At late second trimester inhibin-A and PAPP-A2 might be useful for shortterm prediction of PE. Abbreviations: AUC, area under curve; ASA, acetylsalisylic acid; BMI, body mass index; EO PE, early-onset pre-eclampsia (diagnosis < 34 weeks of gestation); IGF, insulin-like growth factor; IGFBP, insulin growth factor binding protein; IUGR, intrauterine growth restriction; LO PE, late-onset pre-eclampsia (diagnosis ≥ 34 weeks of gestation); MoM, multiples of the median; PAPP-A, plasma protein-A; PAPP-A2, plasma protein-A2; PE, pre-eclampsia; PlGF, placental growth factor; UtA PI, uterine artery pulsatility index; ROC, receiver-operating characteristic; SGA, small for gestational age. * Corresponding author at: Department of Obstetrics and Gynaecology, PO Box 23, 90029 OYS, Oulu, Finland. E-mail address: elina.keikkala@oulu.fi (E. Keikkala).


Introduction
Early identification of women destined to develop pre-eclampsia (PE) later in pregnancy would enable preventive interventions such as low-dose acetylsalicylic acid (ASA) and more intensive follow-up of these women [1][2][3]. The most promising predictive model includes a combination of first trimester serum placental growth factor (PlGF) and the maternal clinical risk factors such as pre-pregnancy body mass index (BMI), maternal age and a history of PE. The accuracy could be slightly improved by adding pregnancy-associated plasma protein-A (PAPP-A) and uterine artery pulsatility index (UtA PI) measured by doppler ultrasound [4]. The ability of this model to predict early-onset PE (EO PE) (diagnosis < 34 weeks of gestation) is fairly good. However, its clinical use is limited because its modest value in predicting late-onset PE (LO PE), which is the most common subtype of PE [4].
Inhibin-A is a glycoprotein hormone produced by placental trophoblasts [5,6]. It is used in the second trimester combined screening of chromosomal abnormalities [7]. Several studies have shown that women who developed PE have higher levels of inhibin-A already in the first trimester [6,8,9]. However, its sensitivity has been found to be too low for use as a single marker to predict PE [10].
We studied whether inhibin-A, PAPP-A, and PAPP-A2 analysed at three timepoints during first and second trimesters serve as potential biomarkers to predict PE and its subtypes. The results were compared with the previously reported PlGF results [24]. All first trimester biomarkers were also analysed together with the maternal clinical risk factors and UtA PI.

Subjects
Women in the present nested case-control study participated in the PREDO (Prediction and Prevention of Pre-eclampsia and Intrauterine Growth Restriction) project [25]. Women with clinical risk factors for PE were prospectively recruited between September 2005 and December 2009 at ten participating maternity clinics in Finland. The ethics Committee at Helsinki and Uusimaa Hospital District approved the study and written informed consent was obtained from all participants.
A subcohort of 134 women was included in this study: 45 women diagnosed with PE and 89 controls who did not develop PE (49 with and 40 without known clinical risk factors for PE). In the pre-eclampsia group, 11 had early-onset PE (EO PE, diagnosis < 34 + 0 weeks of gestation) and 34 had late-onset PE (LO PE, diagnosis ≥ 34 + 0 weeks). Eight women who developed PE participated in low-dose ASA trial of the PREDO project. They were treated with low-dose ASA 100 mg/day from 12th to 35th week of gestation. Equal numbers of women who received low-dose ASA and did not develop PE were included as controls. Other controls were chosen based on sample availability.
Inclusion criteria for women with clinical risk factors for PE were as follows: PE, intrauterine growth restriction (IUGR), gestational diabetes or foetal demise in a previous pregnancy; pre-pregnancy obesity (BMI ≥ 30 kg/m2); chronic hypertension; type 1 diabetes; maternal age at childbirth < 20 years or > 40 years; systemic lupus erythematosus (SLE) or Sjögren's syndrome [25].
PE was defined as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg occurring after 20th weeks of gestation and proteinuria (≥300 mg per 24 h or the dipstick equivalent in two consecutive measurements [26]. Women with pre-eclampsia superimposed on chronic hypertension were included in the PE group. SGA was defined as birthweight lower than − 2 SD adjusted for sex and gestational age according to Finnish birth weight charts [27]. Blood samples were obtained at three timepoints during pregnancy (12-14, 18-20 and 26-28 weeks of gestation). Mean UtA PI was measured by doppler ultrasound in all participants at 12-14 weeks of gestation [2].

Laboratory methods
Serum was separated by centrifugation within 1 h and aliquots were stored at − 80 • C until analyses [25]. Serum inhibin-A, PAPP-A and PAPP-A2 concentrations were measured by ELISA Kits according to manufacturer's instructions (Ansh Labs, Webster, USA) from total of 372 serum samples. Inter-assay coefficients of variation (CV) were 10.9% (at control concentration of 110 pg/ml) and 4.9% (at control concentration of 344 pg/ml) for inhibin-A; 2.3% (at control concentration of 310 ng/ ml) and 15.8% (at control concentration of 1150 ng/ml) for PAPP-A, and 8.9% (at control concentration of 0.9 ng/ml) and 19.5% (at control concentration of 3 ng/ml) for PAPP-A2. PlGF concentrations were analysed by a Delfia method Kits as described previously (PerkinElmer, Turku, Finland) [24].

Statistics
Continuous variables between groups were compared using the Mann-Whitney U test and categorical variables were compared using the Chi-Square test. Correlations between continuous variables were analysed using Pearson's correlation. Change of biomarker concentration during advancing pregnancy between two sampling points were analysed using Wilcoxon signed rank test for paired samples. The results were expressed as mean ± SDs, median and 95% confidence intervals (CIs) or interquartile range (IQR) when appropriate. Concentrations were log10 transformed and normality was tested according to the Kolmogorov-Smirnov's test. To adjust the concentrations for gestational age, multiples of the median (MoM) values were calculated for each biomarker: the linear regression equation calculated at three different time points from the samples collected from 40 controls without risk factors from log10 transformed values. For the prediction of PE, clinical background factors (maternal age, pre-pregnancy BMI, prior PE, prior SGA, prior foetus mortus, type 1 diabetes, primiparity, chronic hypertension), biomarkers at 12-14 weeks of gestation and UtA PI were analysed by logistic regression using forward stepwise method where variables with p < 0.10 were included in model. The diagnostic accuracy for separate biomarkers and for the combinations of clinical risk factors and biomarkers were analysed using receiver-operating characteristic (ROC) curve analysis and expressed as the area under the curve (AUC). Analyses were performed using SPSS Statistics version 25 (IBM). Twosided P values < 0.05 were considered statistically significant.
The power of the study was calculated using previous results on PlGF for prediction of EO PE [24]. With these mean MoM values and SDs the power was 98.6% to detect differences between the groups (9 women with EO PE and 67 controls) with two-sided p-value < 0.05.

Clinical characteristics
Clinical characteristics of the study groups are presented in Table 1.
There were no significant differences in gestational age at sampling between the study groups (data not shown). Eight women who developed both subsequent PE and SGA were analysed as a separate subgroup (data not shown).

MoMs for biomarker concentrations
Inhibin-A concentration decreased from 12 to 14 to 18-20 weeks of gestation and increased after that (p < 0.001 for controls and p = 0.001 for PE) in all groups (p < 0.001 for both) ( Supplementary Fig. 1A). Inhibin-A concentration was higher than in controls at 18-20 and 26-28 weeks (Table 2A) and the MoM values were higher than in controls at all studied gestational weeks in women with subsequent PE (Table 2B). Inhibin-A MoM values were higher at 18-20 and 26-28 weeks in women with subsequent EO PE and at 26-28 weeks in women with LO PE compared to controls (Table 2B) (Table 2A and 2B), PAPP-A2 concentrations increased during advancing pregnancy in all groups (p < 0.001 for all) ( Supplementary Fig. 1C). PAPP-A2 concentrations and MoM values were higher in women with subsequent PE as compared to controls at 26-28 weeks. Higher PAPP-A2 MoM values were observed already at 18-20 weeks in women with subsequent EO PE (Table 2A and 2B).
Inhibin-A concentrations and MoM values were higher in women with subsequent PE with SGA already at 12-14 weeks of gestation. Higher PAPP-A2 MoM values were observed already at 18-20 weeks in women with subsequent PE with SGA (p < 0.001) in addition to 26-28 weeks of gestation (data not shown).

Prediction of PE, EO PE and LO PE
The AUC values for all studied biomarkers including previous data on PlGF [24) are shown in Table 3. The AUC value for the prediction of all subtypes of PE at 12-14 weeks of gestation was 0.618 for inhibin-A and it improved during advancing pregnancy (0.643 at 18-20 weeks and 0.717 at 26-28 weeks of gestation). PAPP-A2 predicted PE at 26-28 weeks (AUC 0.700) and PlGF at 18-20 and 26-28 weeks (AUC 0.648 and 0.733, respectively) ( Table 3).

Multivariate analyses at 12-14 weeks of gestation
The combinations of the studied biomarkers, previous data on PlGF, maternal clinical risk factors and UtA PI were studied using stepwise forward logistic regression analysis ( Table 4). The MoM for Inhibin-A and PAPP-A showed capacity to predict PE with an AUC value for the combination of 0.694. AUC value for BMI, prior PE and UtA PI was 0.796. Combination of Inhibin-A MoM with pre-pregnancy BMI, prior PE, and UtA PI gave an AUC value of 0.811 (Table 4, Fig. 1).
Of the studied biomarkers, PAPP-A MoM and PlGF MoM predicted EO PE (AUC 0.701 and 0.798, respectively). AUC value for BMI, prior SGA and UtA PI was 0.825.Combining PlGF MoM with information of prior SGA, pre-pregnancy BMI, and UtA PI gave an AUC value 0.906 for prediction of EO PE. PlGF MoM and clinical risk factors were superior to PAPP-A MoM to predict EO PE in the multivariate setting (Table 4, Supplementary Fig. 2).
None of the studied biomarkers alone predicted LO PE. Combining

Table 2B
Multiples of the median (MoM) values of serum Inhibin-A, pregnancy associated plasma protein-A (PAPP-A) and PAPP-A2.  The AUC and P values are based on receiver operating characteristics (ROC) curve compared to controls (n = 89). PE = pre-eclampsia, EO PE = early-onset preeclampsia (diagnosis < 34 weeks of gestation), LO PE = late-onset pre-eclampsia (diagnosis ≥ 34 weeks of gestation), MoM = multiples of median, PAPP-A = pregnancy associated plasma protein-A, PlGF = placental growth factor.
inhibin-A MoM with pre-pregnancy BMI, information of prior PE, and UtA PI the AUC value to predict LO PE was 0.824. Combination of clinical risk factors and UtA PI without inhibin-A gave an AUC value of 0.807 and when UtA PI was removed the AUC value for the combination of inhibin-A and clinical risk factors was 0.815 (Table 4, Supplementary  Fig. 3).
In women who developed PE with SGA Inhibin-A MoM combined with information about chronic hypertension and UtA PI gave an AUC value of 0.857 (95% CI 0.724-0.990; p = 0.001). In a multivariate setting with clinical risk factors Inhibin-A performed better than PlGF in predicting PE with SGA (data not shown).

Correlations
At 12-14 weeks, PAPP-A and PAPP-A2 correlated negatively with maternal pre-pregnancy BMI and PAPP-A2 correlated positively with chronic hypertension status. Otherwise, the markers studied were independent of maternal age, parity, pre-pregnancy BMI and chronic hypertension. When comparing inhibin-A, PAPP-A2 and PAPP-A concentrations between ASA and placebo users in women with subsequent PE, EO PE or LO PE, among women with subsequent PE, inhibin-A at 18-20 weeks of gestation was slightly lower in women having ASA than in women receiving placebo (MoM values: 0.94, 0.76 -1.26; median, IQR; vs. 1.58, 1.17-2.86; p = 0.006). Other biomarker concentrations did not differ between ASA and placebo users (data not shown).

Discussion
We found that inhibin-A together with clinical risk factors and UtA PI performed well in the prediction of PE, especially LO PE was predicted already at 12-14 weeks of gestation. Previously established biomarkers, Table 4 The area under the curve (AUC) values from receiver operating characteristic (ROC) curve for combinations of maternal clinical risk factors, studied biomarkers and uterine artery pulsatility index (UtA PI) at 12-14 weeks of pregnancy in predicting of pre-eclampsia (PE), early-onset pre-eclampsia (EO PE) and late-onset preeclampsia (LO PE). The  PlGF and PAPP-A, showed advantage only for the prediction of EO PE.
In line with earlier studies, inhibin-A concentration decreased after the first trimester and increased after midgestation [5] and was higher in women with subsequent PE compared to controls [6,9,28]. Our study revealed that inhibin-A was better than PlGF and PAPP-A for prediction of LO PE at 12-14 weeks of gestation but it did not predict EO PE. In a previous meta-analysis, the sensitivity of inhibin-A to predict EO PE was only 32% (95% CI, 25-39%) with a specificity of 90% [10] and to the best of our knowledge its utility to predict LO PE has not been examined in earlier studies. Inhibin-A is an endocrine regulator of the gonadal and placental function [5,6,29]. In pre-eclamptic pregnancies the production of inhibin-A in placental trophoblasts have been shown to be increased compared to normal placentas with unknown mechanism [30,31].
In the present study concentrations of PAPP-A and PAPP-A2 increased during advancing gestation in women with or without subsequent PE. This is in agreement with the results of earlier studies [32,33]. Our study demonstrates that low levels of PAPP-A at 12-14 weeks of gestation are associated with increased risk of EO PE. This has also been seen in earlier studies [9,14,34,35]. In a multivariate analysis PAPP-A did not improve the prediction of all PE or EO PE when using the combination of maternal serum PlGF or inhibin-A, maternal characteristics, obstetric history and uterine artery PI. This is in line with several earlier studies [4,36,37]. Other studies have shown association of early pregnancy low PAPP-A levels with other placental disease such as SGA and preterm delivery [14,15,34]. Low PAPP-A levels also correlate more strongly to EO PE than LO PE [35]. These studies support the notion of PAPP-A rather being a serum marker for placental disease in general than being specific for PE because most of the women with EO PE have placental dysfunction and SGA foetus. PAPP-A is an IGFBP-4 proteinase and the presence of IGF is obligatory for its proteolytic function [38]. IGFs have an important role in the regulation of trophoblast invasion [34]. Low PAPP-A levels may rather be a cause than a consequence in early placental dysfunction.
In the present study increased concentration of PAPP-A2 was associated with all studied PE subtypes at 18-20 and 26-28 weeks of gestation. Increased concentrations have been shown previously in subsequent [22] and established PE [19,32]. In contrast to PAPP-A, the cleavage of IGFBP-5 by PAPP-A2 does not require the presence of IGF [16]. Hypoxia, which is known to occur in the PE placenta, has been shown to upregulate the expression of PAPP-A2 mRNA in placental explants [21] and placental cell lines (BeWo cells) in cell culture [39]. These findings suggest that increased PAPP-A2 serum levels are more likely to be a consequence of placental pathology than vice versa. Thus, PAPP-A2 may be more sensitive and specific for PE than PAPP-A.
In the prediction of all PE and EO PE, Inhibin-A [40] and PAPP-A2 showed similar ability with PlGF at 26 to 28 weeks of gestation. PlGF can be used in short-term prediction of PE with or without soluble fmslike tyrosine kinase-1 (sFlt-1) [41,42]. Our study suggests that inhibin-A and PAPP-A2 may, as well, have potential for short-term prediction of PE among women having suspected symptoms of PE.
The strength of our study is the prospectively recruited, wellcharacterised cohort with three consecutive serum samples from the same individuals. Its main weakness is the relatively small sample size and the phenomenon that ASA might have had an effect on biomarker concentrations at 18-20 and 26-28 weeks of gestation and development of PE. However, the difference between ASA and placebo users was observed only in concentrations of inhibin-A at 18-20 weeks in women with subsequent PE with unknown mechanism.
In conclusion, this study showed that inhibin-A with maternal clinical risk factors and UtA PI show potential ability to predict all PE and LO PE at 12-14 weeks of gestation whereas PlGF was superior for predicting EO PE. PAPP-A2 and inhibin-A show potential capacity for short-term prediction of PE at 26-28 weeks of gestation.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.