Investigation of Association with First-Trimester Free Human Chorionic Gonadotropin-ß, Pregnancy-Associated Plasma Protein-A, and Adverse Pregnancy Outcomes

OBJECTIVES: This study aimed to investigate the association between first-trimester screening maternal serum markers (free human chorionic gonadotropin-beta (β-hCG), pregnancy-associated plasma protein-A, and adverse pregnancy outcomes (gestational hypertension, preeclampsia, preterm delivery, intrauterine growth restriction, oligohydramnios, intrauterine ex-fetus, abruptio placentae, and gestational diabetes mellitus). STUDY DESIGN: This was a retrospective cohort study including 1516 women who delivered a singleton pregnancy at GATA Haydarpasa Education Hospital from 2006 to 2009. Patients with multiple pregnancies, chromosome aberrations, or fetal anomalies were excluded. Extreme values of correctedpregnancy-associated plasma protein-A and free β-hCG, and their association with adverse pregnancy outcomes were analyzed. RESULTS: Adverse pregnancy outcomes at the cutoff level of ≤0.25 c-pregnancy-associated plasma protein-A MOM values positive likelihood ratio (+LR) was 7.5 (95%CI 2.426-19.836) and had a significant correlation (r=0.108, p<0.01). It was also a significant correlation with adverse pregnancy outcomes (r=0.189, p<0.01) at the cut-off level of ≤0.50 corrected-pregnancy-associated plasma protein-A MOM values and the +LR was 2.388 (95%CI 1.889-2.991). Association between low corrected-pregnancy-associated plasma protein-A MOM values and adverse pregnancy outcomes were statistically significant and had a poor association (AUC, 0.630 95%CI 0.596-0.663, p<0.01). Preeclampsia was statistically significant, however had a fair association (AUC, 0.74 95% CI 0.690-0.802, p<0.01). Preterm birth was statistically significant but had a poor association (AUC, 0.568 95% CI 0.512-0.624, p<0.05). CONCLUSION: First-trimester maternal serum low pregnancy-associated plasma protein-A values are significantly associated with adverse pregnancy outcomes. It may be a useful tool to predictive not only chromosome anomalies but also adverse pregnancy outcomes.


Introduction
First-trimester screening tests were found to be effective for trisomy 21 screening (1)(2)(3)(4)(5). This screening technique is used to evaluate fetal nuchal translucency (NT) ultrasound and portant role of PAPPA in modulating ovarian function, placentation, placental function, and female fertility by control of the bioavailability of ovarian IGF (8,10). If the level of PAPPA is insufficient to successfully cleave IGF, IGF remains in its bound, inactive form, possibly leading to diminished fetal and placental growth (11).
While several other publications have shown statistically significant associations between low PAPPA levels and preeclampsia, fetal loss, preterm birth, and (IUGR), they have failed to demonstrate high sensitivity and high positive predictive values to deem a patient high or low risk for the adverse outcomes (12)(13)(14)(15)(16)(17)(18)(19)(20). Free β-hCG is a syncytiotrophoblastderived promoter of cell growth and differentiation in the embryo (21). Evidence is lacking to support an association between free β-hCG and non-chromosomal adverse pregnancy outcomes (19,22,23).
In this retrospective cohort study, we aimed to investigate the relationship between the first-trimester screening maternal serum markers especially corrected PAPPA levels, and adverse pregnancy outcomes such as gestational hypertension, preeclampsia, IUGR, oligohydramnios, preterm birth, intrauterine ex fetus (IUEF), gestational diabetes mellitus (GDM), placental abruption.

Material and Method
This was a retrospective cohort study including the final number of 1516 women delivered of a singleton pregnancy at GATA Haydarpasa Education Hospital from 2006 to 2009 for whom first-trimester serum markers analysis, as well as pregnancy outcome data, were available. Multiple pregnancies confirmed fetal chromosomal abnormalities and fetal demise before 20 weeks of gestation were excluded. Records with incomplete pregnancy outcome data were excluded. Clinical records and hospital computer automation systems were used as data sources. The three first-trimester markers including PAPPA level, free β-hCG, and NT were evaluated. Information on maternal medical and obstetrical history, gestational age, maternal weight, fetal crown-rump length (CRL), and NT measures at the examination, type of delivery, time of weeks at delivery, and pregnancy outcomes were collected. The adverse pregnancy outcomes examined including gestational hypertension (GHT), preeclampsia, preterm birth, IUGR, oligohydramnios, low birth weight, IUEF, abruptio placentae, and GDM.
Pregnancy-associated plasma protein-A values and free β-hCG values measured in maternal serum were obtained by the chemiluminescence method using immulite kits in Immulite 1000 analyzer of the Hospital Biochemistry Laboratory. Multiples of median values (MOM) of the corrected median were calculated by using the first-trimester screening test form which was filled in perinatology outpatient clinic with PRISCA package software program. Demographic distribu-tions of patients, ultrasonographic findings in first-trimester between 10 weeks 0 days and 13 weeks 6 days of gestation (CRL between 38 and 83 mm), and biochemical test data were collected in the Office 2007 Excel program.
Preterm birth was defined as a delivery occurring before 37 weeks of pregnancy. IUGR was defined as a fetus whose estimated weight was below the 10th percentile for gestational age. GHT was defined as a new onset of hypertension (systolic blood pressure 140 mmHg and/or diastolic blood pressure 90 mmHg) were documented on at least two occasions at least 6 hours apart at 20 weeks of gestation, in the absence of proteinuria. Preeclampsia was defined as the presence of systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg after 20 weeks of pregnancy detected on two or more occasions separated by 6 h with the presence of proteinuria randomly detection of ≥2+ on dipstick or 0.3 g in a 24 h urine specimen. Intrauterine fetal demise, ex fetus, or stillbirth was defined as intrauterine fetal deaths after 20 weeks of pregnancy. Oligohydramnios was defined as an amniotic fluid index <5 cm. GDM was diagnosed if fasting glucose level 92 and 126 mg/dL at first-trimester or at least one value of plasma glucose concentration 92, 180, and 153 mg/dL (for fasting, 1and 2-h post-glucose load glucose values, respectively), after performing a 75 g oral glucose tolerance test or at least two values of plasma glucose concentration 95, 180, 155, and 140 mg/dL (for fasting, 1., 2., and 3. hour post-glucose load glucose values, respectively), after performing a 100 g oral glucose tolerance test if 50 g oral glucose tolerance test first hour glucose level was >140 mg/dL at 24-28 weeks of gestation. Small for gestational age (SGA) was defined as a birth weight of less than the 10 th percentile for gestational age.
In the primary analysis, relations between ≤0.25 MOM and ≤0.50 MOM levels of PAPPA and adverse pregnancy outcomes were evaluated. Descriptive analysis of adverse pregnancy outcomes under <1.0 MOM level with 0.25 MOM intervals of PAPPA was performed ( Figure 1).
Statistical analyses were performed with the SPSS package (SPSS Inc., Windows, version 17.0). Categorical variables were compared by the χ² test. Continuous variables were described using median and interquartile ranges and categorical variables using frequencies. Spearman correlation and Receiver operating characteristic (ROC) curve analysis were performed to evaluate the relationship between independent variables. The traditional academic point system was used to classify the area under the curve (AUC). AUC from 0.90 to 1.0, 0.80 to 0.90, 0.70 to 0.80, 0.60 to 0.70, and 0.50 to 0.60 indicates excellent, good, fair, poor, and fail association ability. Likelihood ratio, sensitivity, specificity, predictive values, and 95%CI were calculated by using the diagnostic test calculator program of evidence-based medicine toolbox in (https://ebmtools.knowledgetranslation.net/calculator/diagnostic/) webpage. Statistical significance was defined as p-value 0.05.

Results
The final number of the study population was 1516 after excluding multiple pregnancies, fetal chromosomal abnormalities, and fetal demise before 20 weeks of gestation, records with incomplete pregnancy outcome data during the study period, from the beginning of January 2006 to at the end of December 2008. Descriptive characteristics of the study group of 1516 pregnant women were shown in table I.
Adverse pregnancy outcomes among 1516 pregnant women were as follows 33 (2.2%) GHT, 68 (4.5%) preeclampsia, 107 (7.1%) preterm birth, 29 (1.9%) IUGR, 14 (0.9%) IUEF, 33 (2.2%) oligohydramnios, 69 (4.6%) GDM, 12 (0.8%) abruptio placentae. Adverse pregnancy outcomes distribution showed among 0.25 MOM intervals of under <1.0 c-PAPPA MOM values (Table II). It has been used maternal weight correction of maternal serum PAPPA values to reduces the population variability of the markers and the impact on detection rate and false-positive rate is very small (3)(4)(5).    There was no relationship between high c-PAPPA or free corrected β-hCG MOM levels and adverse pregnancy outcomes except it was a statistically significant association with the cut-off level of corrected free β-hCG ≥2.9 MOM values and fetal macrosomia (>4500g) (r=0.068, p<0.01). Also, we analyzed that below the 5t h percentile of corrected free β-hCG (≤0.33 MOM) cutoff level, it was evidence of statistically significant association with small for gestational age (10 th percentile birth weight) (r=0.088), p<0.01). Corrected PAPPA MOM values frequency distribution and percentiles of adverse pregnancy outcomes were shown in figure 1 and Table II.
When the number of adverse pregnancy outcomes analyzed by One-Sample Kolmogorov-Smirnov test whether it is appropriate to a normal distribution, groups have been found to follow a normal distribution (p<0.05). Nonparametric tests χ² and Spearman correlation were used for correlation analysis for data analysis.
The likelihood ratio was preferred to sensitivity, specificity, and predictive values. Because they have advantages over sensitivity, specificity, and positive/negative predictive values because they are less likely to change with the prevalence of the disorder, they can be calculated for several levels of the symptom/sign or test, they can be used to combine the results of multiple diagnostic tests and they can be used to calculate the post-test probability for a target disorder.

Discussion
This study aimed to investigate the relationship between the first-trimester screening maternal serum markers especially corrected PAPPA levels and adverse pregnancy outcomes such as GHT, preeclampsia, IUGR, oligohydramnios, preterm birth, IUEF, GDM, and abruptio placentae (placenta mediated obstetric complications). This study suggests that first-trimester maternal serum low c-PAPPA MOM values were significantly associated with adverse pregnancy outcomes. These markers alone do not have high sensitivity and low false positivity rates for clinical use. However, except for Down syndrome risk assessment in the first-trimester screening test, there was a significantly increased correlation and the risk was identified between adverse pregnancy outcomes and especially low c-PAPPA (≤0.5 MOM) values. Our study results agreed with previous studies that have found a low c-PAPPA level to be associated with an increased risk of IUGR, preeclampsia, preterm birth, and IUEF (12)(13)(14)(15)(16)(17)(18)22,24,25). Similar to previous studies, it was not evidence of statistically significant association with any cut-off levels of corrected free β-hCG MOM values and other adverse pregnancy outcomes. However, we found that there was a statistically significant association between extreme values of corrected free β-hCG MOM values and fetal birth weight.
Gomes et al.'s study suggest that an unexplained low level of PAPPA (<10th percentile) during the first trimester can be associated with adverse outcomes and placenta-mediated obstetric complications namely preeclampsia, fetal growth restriction, and abruptio placentae (26). Low levels of free β-hCG (<1st, 5th, and 10th percentiles) were not associated with adverse outcomes. Dugoff et al. and Smith et al. studies demonstrated an association between low levels of free β-hCG (<0.4 or 0.5 MOM) and birth weight below the 5th or 10th percentile (OR 1.6-1.7) (17,23). In our study, we analyzed that below the 5th percentile of corrected free β-hCG (≤0.33 MOM) cut-off level, it was the evidence of statistically significant association with small for gestational age (10th percentile birth weight) (r=0.088), p<0.01). And also, we found that it was also a statistically significant association with the cutoff level of corrected free β-hCG ≥2.9 MOM values and fetal macrosomia (>4500 gram) (r=0.068, p<0.01).
Smith et al. found that pregnant women who had values below the 5 th percentile as a result of their PAPPA measurements between the gestational weeks had an increased risk of IUGR, preterm birth, preeclampsia, and stillbirth (11,23). Based on their findings, they suggested that the control of the IGF system in trophoblasts in the first and early second trimester may play a key role in determining the prognosis of pregnancy (23). In animal experiments, it has been shown that blocking of the PAPPA gene results in a 40% reduction in fetal development (27). It is thought that PAPPA has either autocrine and paracrine effects on local activities of IGF (28). These data demonstrate an important role of PAPPA in modulating ovarian function, placentation, placental function, and female fertility by control of the bioavailability of ovarian IGF (8,29). If the level of PAPPA is insufficient to successfully cleave IGF, IGF remains in its bound, inactive form, possibly leading to diminished fetal and placental growth and also placental inflammation (11). In our study, we had a statistically significant correlation with IUEF, IUGR, preeclampsia, and low level of PAPPA.
Canini et al. found that PAPPA values were significantly lower in babies with SGA, and PAPPA was significantly higher in babies with LGA (30). They found a significant correlation between birth weight percentiles and c-PAPPA values (r=0.192, p<0.001). In this case, high PAPPA levels may be expected to be associated with high IGF and thus fetal macrosomia (30). In the study of Dugoff et al., there was no significant association with fetal macrosomia (over 4500 g) at levels above the 99.9th percentile of PAPPA level (p<0.05) (17). However, in our study we found that there was a significant correlation between maternal serum free β-hCG corrected MOM ≥2.9 MOM values with fetal macrosomia infants (r=0.068, p<0.01). There was no significant association with the cut-off level of free β-hCG≤0.5 MOM value and GDM (p>0.05). However, it was a significant association between GDM and the cutoff level of corrected PAPPA ≤0.5 MOM values (r=0.064, p<0.05).
The traditional approach to screening for preeclampsia, which is based on maternal demographic characteristics and medical history, identifies approximately 30% of cases destined to develop early PE for a false-positive rate of 5% (26). It is necessary to find additional predictive parameters to classify the high-risk pregnancy group. Poon et al used the combination of mean arterial pressure (MAP), placental growth factor (PIGF), and PAPPA MOM values to increase the likelihood ratio or post-test probability. In this way, they were aimed at the care of low-risk pregnancies is decentralized to local, 1-stop settings or at home, whereas high-risk pregnancies are cared for in specialized centers (31).
We were referred to analyze the predictive values of adverse pregnancy outcomes by likelihood ratio because it can be used to calculate the post-test probability and it is less affected by the prevalence of the adverse pregnancy outcome (Table III and IV). This situation provided us that if we detected increased post-test probability, we would have guided the patient to another test and the first test's post-probability value could be used as the next test pre-test probability. In this way, calculating the final likelihood ratio of certain cutoff levels obtained from a meta-analysis may be a useful tool with additional tests for conditions and diseases that have low prevalence. In our study, we have reported the likelihood ratio results of PAPPA ≤0.25 and ≤0.50 MOM cutoff values to be able to use in future meta-analysis. And we saw that the positive likelihood ratio changes by the cut of level value are certain these changes will reflect the post-test probability. Posttest probability calculating is very important nowadays because we are living in the artificial intelligence century. In artificial intelligence, one of the suitable and calculable data that can be used in artificial intelligence algorithms may be the post-test probability value. According to Bayes theorem, the formula may be as shown (32).
Post-test probability of total adverse pregnancy outcome for first pregnancy = pre-test probability (i.e. prevalence) × positive likelihood ratio (i.e. first-trimester screening test) × positive likelihood ratio (Doppler ultrasonography) × positive likelihood ratio (i.e. second-trimester screening test) × positive likelihood ratio (second-trimester ultrasonography) Taking into account the cutoff level of the test and considering the patient in a high-risk group or not is insufficient to artificial intelligence algorithms. Such a formula as shown may help the calculation of the pre-test probability of the firsttrimester screening test of the second pregnancy and may be useful data for artificial intelligence algorithms to identify and follow the high-risk pregnant women in the National health care program.
The most important limitation of our study is the need for more cases in the evaluation of the prediction of adverse obstetric outcomes with low prevalence. It may not be appropriate to exclude miscarriage cases because of the significant association between the low levels of PAPPA and adverse pregnancy outcomes such as IUGR and IUEF. While we were preparing this article for an academic publication, the military coup incident took place. We could not access the data after 2009 due to the hospitals' handover and system changes. It may be appropriate to evaluate the association between the low level of PAPPA and miscarriages in further studies in the future.

Conclusion
In our study, it was observed that first-trimester maternal serum low PAPPA MOM values were significantly associated with adverse pregnancy outcomes. These markers alone do not have high sensitivity and low false positivity rates for clinical use. However, except for Down syndrome risk assessment in the first-trimester screening test, there was a significantly increased correlation, and the risk was identified between adverse pregnancy outcomes and especially low corrected PAPPA (≤0.5 MOM) values. It is necessary to be careful when planning the follow-up of pregnancies of first-trimester markers by increasing the sensitivity and decreasing the false positivity with additional tests to be performed in the following gestational weeks.

Declarations
Ethics approval and consent to participate All participants signed informed written consent before being enrolled in the study.