Introduction

First trimester screening has a paramount importance to reveal anomalies in the fetus. The combination of maternal age, nuchal translucency (NT), maternal serum free beta-human chorionic gonadotropin (free β-hCG) and pregnancy-associated plasma protein A (PAPP-A) identifies trisomy-21 with an 85% detection rate with a 5% false positivity (1,2). Also, some studies have shown a relationship between low PAPP-A and free β-hCG levels in the 11-13^th weeks of gestation with poor obstetric outcomes such as stillbirth, fetal growth restriction (FGR), preterm delivery and preeclampsia (PE) (3-7).

PAPP-A is an insulin-like growth factor binding protein-4 (IGFBP-4) protease, which is secreted by trophoblasts and decidual cells. Low PAPP-A levels decrease the IGF levels by increasing the IGFBP-4 level. IGF is a stimulant molecule for fetal growth and maturity (8). Meanwhile, free β-hCG is secreted by syncytiotrophoblasts of the placenta in addition to the corpus luteum of the ovary in early pregnancy and it promotes angiogenesis, and its secretion increases under different conditions such as uteroplacental insufficiency (9).

The aim of the presented study was to investigate the predictive value of first trimester PAPP-A and free β-hCG levels for poor obstetric outcomes. Although different cut-off values have been proposed in the literature, we took the 5^th and 10^th percentile ranks of FASTER study to detect the effect of PAPP-A and free β-hCG levels on obstetric outcomes (5).

Methods

The study cohort consisted of pregnant women who underwent first trimester screening in our Perinatology Clinic between January 2016 and March 2018. The medical records were evaluated retrospectively and 549 patients were included in the study. Patients with crown-rump length (CRL) measurement between 45 and 84 mm, patients with singleton pregnancy and patients who delivered at 24 weeks or later were included.

A single operator performed all ultrasonographic evaluations with a 5-8 MHz 3D transducer ultrasound system (General Electric Voluson E8™, GE Healthcare, Chicago, Illinois, United States) via transabdominal route. All maternal variables such as age, height, weight, body mass index, gravidity and parity were recorded. Also, CRL measurements with the concordant gestational week, NT, mean uterine artery pulsatility index, PAPP-A and free β-hCG levels were noted.

PAPP-A and free β-hCG levels were measured by chemiluminescent immunometric assay from blood samples of pregnant women using Immulite 2000 XP device (Siemens A6 Healthcare Sector, Erlangen, Germany). These indicators were converted into multiples of the expected median (MoM) value with PRISCA 5.0 software considering the maternal age, medical history and patient-specific factors. The 5^th and 10^th percentile ranks of PAPP-A and free β-hCG values were determined according to FASTER study and groups were constructed amongst these values to compare the possible differences of preterm delivery, low birth weight (LBW) and PE between groups (5).

According to the World Health Organization, preterm delivery is accepted as labor before 37^th weeks of pregnancy. According to the American College of Gynecologists, LBW is accepted as <2500 gr live birth. PE criteria were as follows: hypertension (systolic pressure ≥140 mmHg/diastolic pressure ≥90 mmHg) starting after 20^th gestational week and accompanied by proteinuria (≥300 mg proteinuria in 24 hours or rate of protein/creatinine >0.3) or end organ dysfunction (platelet count <100000/mL or serum creatinine level ≥1.1mg/dL or two-fold increase in serum transaminase levels). Adana Dr. Turgut Noyan University, Non-interventional Clinical Research Ethics Committee (decision no: KA18/187).

Statistical Analysis

All recorded values such as maternal characteristics, correlation between birth weight and gestational age at delivery between pregnancy groups according to 5^th and 10^th percentile values of PAPP-A (0.42 MoM/0.52 MoM) and free β-hCG (0.35 MoM/0.44MoM) were analyzed with SPSS version 21 software package. Categorical variables were expressed as number and percentage, and constant variables were expressed mean and standard deviation (median and range, where necessary). Student’s t-test for constant variables and Fisher’s Exact test for categorical variables were used to compare baseline characteristics and perinatal outcomes. Spearman correlation analysis was used to describe the relationship between gestational age at delivery, birth weight and PAPP-A and free β-hCG levels. Presence of PE was evaluated with Mann-Whitney U test between groups. A p value of <0.05 was accepted as significant in all statistical analysis.

Results

A total of 736 records were evaluated and it was realized that 170 patients were lost during follow-up. Also, five pregnancies were terminated due to congenital anomalies, nine patients had stillbirth before 20^th gestational week and one patient delivered at 23^rd gestational week, so these patients were excluded. All patients excluded from the analysis had a PAPP-A and free β-hCG levels above 10^th percentile, but one intrauterine ex fetus (14^th gestational week) had a low PAPPA level (0.26 MoM). Consequently, a total of 549 patients were included in the study. The maternal characteristics and evaluated variables are presented in Table 1. Table 2 shows the patients having PAPP-A or free β-hCG below 5^th and 10^th percentiles.

The indications for preterm delivery were PE, preterm premature rupture of membrane, spontaneous preterm birth, acute fetal distress, chorioamnionitis, abruptio placenta, fetal cardiac anomaly and maternal decompensated cardiac failure.

According to our retrospective calculations, PAPP-A had a power of %95 for preterm delivery/PE comparison and free β-hCG had a power of 100% for PE comparison.

We did not find any significant difference between PAPP-A percentile groups in terms of mean birth weight (5^th and 10^th percentile, p=0.243 and p=0.139, respectively) or gestational age at delivery (5^th and 10^th percentile, p=0.229 and p=0.523 respectively). However, there was a statistically significant difference between free β-hCG percentile groups regarding mean gestational age at delivery (5^th percentile p=0.013 and 10^th percentile p=0.037). There was no difference in birth weights according to free β-hCG levels (5^th percentile p=0.232 and 10^th percentile p=0.778).

The rate of LBW, preterm delivery and PE were 5.5% (n=30), 6.9% (n=38) and 7.8% (n=43) in entire cohort, respectively. Only preterm delivery was significantly higher in PAPP-A below 5^th percentile group (p=0.049). However, no significant cut-off value was obtained by ROC analysis. Also, PE was higher in free β-hCG groups below the 5^th and 10^th percentiles (p=0.032 and p=0.003, respectively). ROC analysis for relationship between free β-hCG levels and PE revealed an AUC of 0.6 (0.5-0.7, 95% CI). A cut-off value of 0.765 MoM for free β-hCG (64% sensitivity, 53.5% specificity, 11.2 % negative predictive value and 63.2% positive predictive value) was determined for the prediction of PE (p=0.03).

SGA and PE were detected at a rate of 50% (n=19) and 29% (n=11) in all preterm deliveries (n=38), respectively. The relationship between maternal serum PAPP-A and free β-hCG values above and below 5^th and10^th percentiles with mean gestational age at delivery, weight and adverse perinatal outcomes are presented in Table 3.

Discussion

According to our results, the rate of preterm delivery was higher in lower percentiles of PAPP-A and the rate of PE was higher in the lower percentiles of free β-hCG. Preterm delivery is considered as a result of the suboptimal intrauterine environment, which causes FGR. The higher prevalence of preterm deliveries with low PAPP-A in our cohort might be affected by the presence of LBW babies in this group (10).

There is a controversy in the literature about the use of first trimester NT, PAPP-A and free β-hCG measurements, alone or in a combination, to estimate birth weight, gestational age at delivery and perinatal outcomes (1,11-13). Poon and Nicolaides stated that PAPP-A, placental growth factor and uterine artery Doppler measurements, in combination with first trimester maternal risk factors, predict early onset PE with a rate of 95% (14). Dugoff et al, in their multicenter FASTER study, defined the percentile ranks of MoM values with the combination of first trimester NT, PAPP-A, and free β-hCG measurements in a cohort of 33.395 pregnant women. Their results showed that a level of PAPP-A ≤10^th percentile indicated almost 2.5-fold higher risk for LBW and correlation with poor obstetric outcomes. For free β-hCG levels, they identified a relationship with spontaneous loss below 24^th weeks of gestation in ≤1^st percentile group (5). Covans and Spencer supported the results of the FASTER study in their study, which confirmed the relationship between low PAPP-A levels and LBW (15). Also, Kaijomaa et al. compared groups with PAPP-A <0.3 MoM and PAPP-A ≥0.3 MoM, and they stated that low PAPP-A values indicated poor obstetric outcomes (16). Our results did not demonstrate any association with mean gestational age at delivery, weight, LBW and PE at low values of PAPP-A, yet we only found that the preterm delivery rate increased with PAPP-A levels below 5^th percentile (<0.42 MoM).

Similarly, conflicting results have been published about the relationship between low levels of free β-hCG in the first trimester and poor pregnancy outcomes. Sirikunalai et al. reviewed 13.620 singleton pregnancies and summarized in their study that free β-hCG <0.5 MoM value was associated with FGR, preterm delivery, LBW and low APGAR scores (17). Nevertheless, another study did not show a similar relationship between β-hCG at first or second trimester and FGR and PE, but reported that second trimester intact hCG >3 MoM might predict FGR. The authors concluded that the placental dysfunction might lead to pregnancy complications such as FGR or PE (18). In our cohort, the rate of PE was higher in the lower percentiles of free β-hCG. Extreme levels of free β-hCG may be considered as a marker for prenatal surveillance of pregnancies.

Two studies from our country showed contradictory results. In the first study with 318 singleton pregnancies, no significant relationship was found between low PAPP-A levels and incidence of subsequent pregnancy outcomes (19). In the other study, there was a correlation between low PAPP-A levels (≤0.35 MoM) and preterm delivery, low mean birth weight and low gestational age at delivery. However, low free β-hCG levels were not associated with measured variables (20). In our study, the lower percentiles of free β-hCG were associated with lower gestational age at delivery. On the contrary, such a correlation could not be demonstrated between lower PAPP-A percentiles and birth weight or gestational age at delivery.

Sonek et al. stated that first trimester markers in combination with maternal characteristics predicted early or late onset PE with a high detection rate and a low false positive rate (21). Similarly, we found a relationship between low free β-hCG levels and PE.

On the other hand, there are many studies in the literature stating that PAPP-A levels were not helpful in predicting poor obstetric outcomes (22-24). In our study, however, there was an increase in preterm delivery in low PAPP-A group and the difference was statistically significant. We did not reach any conclusions to accept the null hypothesis since the effect seemed limited. On the other hand, considering literature and our results, pregnancies with a level of PAPP-A below 5^th percentile and free β-hCG do not require exceptional monitoring (25).

Our study has some limitations, including (i) retrospective nature, (ii) lack of interpretation of possible covariates, and (iii) lack of evaluation of comprehensive obstetric outcomes (diabetes mellitus, placental pathologies, etc.).

The main difficulty in making a definitive conclusion on this topic is the heterogeneity of the cohorts and methodological differences between the studies in the literature. In addition, different measurement techniques, experiences, thresholds and analysis methods lead to inconsistent outcomes (11,12, 21, 26-28). Considering the literature, more studies with larger cohorts are needed to establish nomograms in the pregnant population.

Conclusion

Early pregnancy markers such as NT, PAPP-A or free β-hCG, alone or in combination, seem unsatisfactory to predict adverse pregnancy outcomes. The clinical implication of these markers should have better positive predictive values and coherent understanding of the mechanisms of the pathophysiological process of the poor obstetric outcomes. Further studies are needed to explain the relationship between first trimester screening markers and neonatal outcomes while considering possible covariates.

Ethics Committee Approval: Adana Dr. Turgut Noyan University, Non-interventional Clinical Research Ethics Committee (decision no: KA18/187).

Informed Consent: Retrospective study.

Peer-review: Externally peer-reviewed.

Author Contributions: Concept - Ş.Y.B., H.K., G.D.D., S.Y.Ş., S.Y.; Design - Ş.Y.B., H.K., S.Y.Ş., S.Y.; Data Collection and/or Processing - Ş.Y.B., G.D.D., S.Y.; Analysis and/or Interpretation - Ş.Y.B., G.D.D., S.Y.Ş., S.Y., E.Ş.; Literature Search - Ş.Y.B.; Writing Manuscript - Ş.Y.B., H.K., G.D.D.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.