Development of early prediction model for pregnancy-associated hypertension with graph-based semi-supervised learning.

Clinical guidelines recommend several risk factors to identify women in early pregnancy at high risk of developing pregnancy-associated hypertension. However, these variables result in low predictive accuracy. Here, we developed a prediction model for pregnancy-associated hypertension using graph-based semi-supervised learning. This is a secondary analysis of a prospective study of healthy pregnant women. To develop the prediction model, we compared the prediction performances across five machine learning methods (semi-supervised learning with both labeled and unlabeled data, semi-supervised learning with labeled data only, logistic regression, support vector machine, and random forest) using three different variable sets: [a] variables from clinical guidelines, [b] selected important variables from the feature selection, and [c] all routine variables. Additionally, the proposed prediction model was compared with placental growth factor, a predictive biomarker for pregnancy-associated hypertension. The study population consisted of 1404 women, including 1347 women with complete follow-up (labeled data) and 57 women with incomplete follow-up (unlabeled data). Among the 1347 with complete follow-up, 2.4% (33/1347) developed pregnancy-associated HTN. Graph-based semi-supervised learning using top 11 variables achieved the best average prediction performance (mean area under the curve (AUC) of 0.89 in training set and 0.81 in test set), with higher sensitivity (72.7% vs 45.5% in test set) and similar specificity (80.0% vs 80.5% in test set) compared to risk factors from clinical guidelines. In addition, our proposed model with graph-based SSL had a higher performance than that of placental growth factor for total study population (AUC, 0.71 vs. 0.80, p < 0.001). In conclusion, we could accurately predict the development pregnancy-associated hypertension in early pregnancy through the use of routine clinical variables with the help of graph-based SSL.

Nonalcoholic fatty liver disease and early prediction of gestational diabetes mellitus using machine learning methods.

BACKGROUND/AIMS: To develop an early prediction model for gestational diabetes mellitus (GDM) using machine learning and to evaluate whether the inclusion of nonalcoholic fatty liver disease (NAFLD)-associated variables increases the performance of model. METHODS: This prospective cohort study evaluated pregnant women for NAFLD using ultrasound at 10-14 weeks and screened them for GDM at 24-28 weeks of gestation. The clinical variables before 14 weeks were used to develop prediction models for GDM (setting 1, conventional risk factors; setting 2, addition of new risk factors in recent guidelines; setting 3, addition of routine clinical variables; setting 4, addition of NALFD-associated variables, including the presence of NAFLD and laboratory results; and setting 5, top 11 variables identified from a stepwise variable selection method). The predictive models were constructed using machine learning methods, including logistic regression, random forest, support vector machine, and deep neural networks. RESULTS: Among 1,443 women, 86 (6.0%) were diagnosed with GDM. The highest performing prediction model among settings 1-4 was setting 4, which included both clinical and NAFLD-associated variables (area under the receiver operating characteristic curve [AUC] 0.563-0.697 in settings 1-3 vs. 0.740-0.781 in setting 4). Setting 5, with top 11 variables (which included NAFLD and hepatic steatosis index), showed similar predictive power to setting 4 (AUC 0.719-0.819 in setting 5, P=not significant between settings 4 and 5). CONCLUSION: We developed an early prediction model for GDM using machine learning. The inclusion of NAFLDassociated variables significantly improved the performance of GDM prediction. (ClinicalTrials.gov Identifier: NCT02276144).

Metabolic Dysfunction-Associated Fatty Liver Disease and Subsequent Development of Adverse Pregnancy Outcomes.

BACKGROUND & AIMS: Recently, metabolic dysfunction-associated fatty liver disease (MAFLD), rather than nonalcoholic fatty liver disease (NAFLD), was proposed to better describe liver disease associated with metabolic dysfunction (MD). In this study, we attempted to investigate the impact of MAFLD on pregnancy complications. METHODS: The current study is a secondary analysis of a multicenter prospective cohort designed to examine the risk of NAFLD during pregnancy. In the first trimester, enrolled pregnant women were evaluated for hepatic steatosis by liver ultrasonography, and blood samples were collected for biochemical measurements. The study population was divided into 3 groups: no NAFLD, hepatic steatosis but without metabolic dysfunction (non-MD NAFLD), and MAFLD. The primary outcome was the subsequent development of adverse pregnancy outcomes, including gestational diabetes mellitus, pregnancy-associated hypertension, preterm birth, and fetal growth abnormalities. RESULTS: The study population consisted of 1744 pregnant women, including 1523 with no NAFLD, 43 with non-MD NAFLD, and 178 with MAFLD. The risk of subsequent development of adverse pregnancy outcomes was higher in MAFLD than in non-MD NAFLD (adjusted odds ratio, 4.03; 95% CI, 1.68-9.67), whereas the risk was not significantly different between no NAFLD and non-MD NAFLD. Among women with no NAFLD, the presence of MD increased the risk of adverse pregnancy outcomes. However, women with MAFLD were at higher risk for adverse pregnancy outcomes than women with no NAFLD without MD or those with no NAFLD with MD. CONCLUSIONS: In pregnant women, MAFLD may be associated with an increased risk of subsequent adverse pregnancy outcomes.

Maternal dyslipidemia and altered cholesterol metabolism in early pregnancy as a risk factor for small for gestational age neonates.

We evaluated the relationship between maternal cholesterol levels and its biologically active precursors and metabolites in the first trimester and subsequent risk for small-for-gestational-age birthweight (SGA). This is a secondary analysis of a prospective cohort study which enrolled healthy singleton pregnancies (n = 1337). Maternal fasting blood was taken in the first trimester and followed up till delivery. The lipid parameters were compared between women who delivered SGA neonates (SGA-group, birthweight < 10th percentile, n = 107) and women who did not (non-SGA-group, n = 1230). In addition, metabolic signatures of cholesterol were evaluated in a subset consisting of propensity-score matched SGA (n = 56) and control group (n = 56). Among lipid parameters, maternal high-density lipoprotein cholesterol (HDL-C) levels were significantly lower in SGA-group than in non-SGA-group (p = 0.022). The risk for SGA was negatively correlated with maternal serum HDL-C quartiles (p = 0.003), and this association remained significant after adjustment for confounding variables. In metabolic signatures of cholesterol, the cholesterol/lathosterol ratio in SGA-group was significantly higher than non-SGA-group [(2.7 (1.6-3.7) vs. 2.1 (1.5-2.9), respectively; p = 0.034)], suggesting increased endogenous cholesterol biosynthesis. We demonstrated that dyslipidemia and increased cholesterol biosynthesis led to delivery of SGA neonates even in early pregnancy.

Novel method of real-time PCR-based screening for common fetal trisomies.

BACKGROUND: The non-invasive prenatal test (NIPT) is based on next generation sequencing (NGS) and is used for screening for fetal trisomy. However, it is time-consuming and technically difficult. Recently, peptide nucleic acid (PNA) probe-based real-time polymerase chain reaction (RT-PCR) was developed. This study aimed to examine the performance of the RT-PCR-based NIPT for screening of common fetal trisomies METHODS: From stored maternal plasma, RT-PCR was performed using Patio™ NIPT Detection Kit. In melting curve analysis, the height of melting peaks of target chromosome and reference chromosome was calculated as a peak ratio. The adjusted peak ratio of 8 markers with correction factors in each target chromosome was summated and calculated to z-score. The cut-off value for each target chromosome was established for classification (low risk vs. high risk for trisomy) whose performance was obtained in the validation phase. RESULTS: 330 plasma samples from pregnant women with normal fetus and 22 trisomy cell-line samples were used to establish the optimal cut-off values for z-score of each target chromosome. In the validation phase, 1023 samples from pregnant women including 22 cases with fetal trisomy and 1001 cases of normal control were used. The RT-PCR-based NIPT showed 95.45% sensitivity [95% confidence interval (CI) 77.16-99.88%], 98.60% specificity (95% CI 97.66-99.23%), and 98.53% accuracy (95% CI 97.59-99.18%) for the identification of trisomy 21, 18, or 13. Of 1023 samples, fifteen cases were mismatched for classification [one case as a false negative (false negative rate: 4.5%) and 14 cases as false positives (false positive rate: 1.4%)]. CONCLUSION: The RT-PCR-based NIPT showed high sensitivity and specificity for the detection of common fetal trisomies and it could be a feasible alternative to NGS-based NIPT.

The risk of pregnancy-associated hypertension in women with nonalcoholic fatty liver disease.

BACKGROUND & AIMS: Nonalcoholic fatty liver disease (NAFLD) is an independent predictor of cardiovascular disease (CVD) in non-pregnant adults. Although the biological mechanisms underlying this association are not completely understood, metabolic factors, inflammation, and endothelial dysfunction are likely all involved. The association between NAFLD and pregnancy-associated hypertension (HTN) has not been systematically examined. The aim of this study is to assess the risk of pregnancy-associated HTN in pregnant women with NAFLD. METHODS: This is secondary analysis of a prospective study of healthy pregnant women. Liver ultrasonography was performed at 10-14 weeks of gestation and maternal blood was taken for the measurement of selenoprotein P (SeP), a hepatokine independently associated with both NAFLD and CVD. Pregnancy-associated HTN was defined as the development of gestational HTN, preeclampsia, or eclampsia. RESULTS: Among 877 pregnant women, the risk of developing pregnancy-associated HTN was significantly increased in women with NAFLD compared to those without NAFLD. Grade 2-3 steatosis was a significant predictor of pregnancy-associated HTN, even after adjustment for metabolic risk factors. Circulating levels of SeP were significantly higher in women with versus those without NAFLD (P = .001) and was significantly higher also in women who subsequently developed pregnancy-associated HTN compared with those who did not (P < .005). CONCLUSIONS: Sonographic evidence of NAFLD at 10-14 weeks is an independent predictor of pregnancy-associated HTN. Circulating levels of SeP at that same gestational age are significantly increased in pregnant women with NAFLD who subsequently develop pregnancy-associated HTN.

A Comparison of Predictive Performances between Old versus New Criteria in a Risk-Based Screening Strategy for Gestational Diabetes Mellitus.

Background: The definition of the high-risk group for gestational diabetes mellitus (GDM) defined by the American College of Obstetricians and Gynecologists was changed from the criteria composed of five historic/demographic factors (old criteria) to the criteria consisting of 11 factors (new criteria) in 2017. To compare the predictive performances between these two sets of criteria. Methods: This is a secondary analysis of a large prospective cohort study of non-diabetic Korean women with singleton pregnancies designed to examine the risk of GDM in women with nonalcoholic fatty liver disease. Maternal fasting blood was taken at 10 to 14 weeks of gestation and measured for glucose and lipid parameters. GDM was diagnosed by the two-step approach. Results: Among 820 women, 42 (5.1%) were diagnosed with GDM. Using the old criteria, 29.8% (n=244) of women would have been identified as high risk versus 16.0% (n=131) using the new criteria. Of the 42 women who developed GDM, 45.2% (n=19) would have been mislabeled as not high risk by the old criteria versus 50.0% (n=21) using the new criteria (1-sensitivity, 45.2% vs. 50.0%, P>0.05). Among the 778 patients who did not develop GDM, 28.4% (n=221) would have been identified as high risk using the old criteria versus 14.1% (n=110) using the new criteria (1-specificity, 28.4% vs. 14.1%, P<0.001). Conclusion: Compared with the old criteria, use of the new criteria would have decreased the number of patients identified as high risk and thus requiring early GDM screening by half (from 244 [29.8%] to 131 [16.0%]).

Correction: Nonalcoholic fatty liver disease is a risk factor for large-for-gestational-age birthweight.

[This corrects the article DOI: 10.1371/journal.pone.0221400.].

Nonalcoholic fatty liver disease is a risk factor for large-for-gestational-age birthweight.

OBJECTIVE: Nonalcoholic fatty liver disease (NAFLD) is a well-recognized hepatic manifestation of metabolic disease in adults and has been associated with the development of gestational diabetes (GDM). Hepatic insulin resistance can result in increased release of glucose (from gluconeogenesis) and free fatty acids (due to enhanced lipolysis), which can lead in turn to fetal overgrowth. However, the relationship between maternal metabolic factors (such as circulating levels of triglycerides, free fatty acids [FFA], or adipokines) and excessive fetal birthweight in NAFLD has not been carefully examined. In this study, we evaluated the relationship between NAFLD and the subsequent risk of large-for-gestational-age (LGA) birthweight. METHOD: Singleton nondiabetic pregnant women were evaluated for the presence of fatty liver at 10-14 weeks of gestation by abdominal ultrasound. The degree of fatty liver was classified as Grade 0-3 steatosis. At the time of liver ultrasound, maternal blood was taken after fasting and measured for adiponectin and FFA. LGA was defined as birthweight >90th percentile for gestational age. RESULTS: A total of 623 women were included in the analysis. The frequency of LGA was 10.9% (68/623), and the frequency of NAFLD was 18.9%. The risk of LGA increased significantly in patients with Grade 2-3 steatosis in the first trimester. The relationship between Grade 2-3 steatosis and LGA remained significant after adjustment for maternal age, pre-pregnancy BMI, GDM, and maternal serum triglyceride levels. The concentration of maternal blood adiponectin at 10-14 weeks was significantly lower in cases with LGA than non-LGA, but the maternal blood FFA concentrations were not different between the groups. CONCLUSION: The presence of Grade 2-3 steatosis on ultrasound in early pregnancy was associated with the increased risk of delivering an LGA infant, even after adjustment for multiple confounding factors including GDM. Adiponectin may be the linking biomarker between NAFLD and LGA.

Non-alcoholic fatty liver disease in the first trimester and subsequent development of gestational diabetes mellitus.

AIMS/HYPOTHESIS: Although there is substantial evidence that non-alcoholic fatty liver disease (NAFLD) is associated with impaired glucose homeostasis, the clinical significance of NAFLD in pregnant women has not been well determined. This study investigates the relationship between NAFLD in the first trimester and the subsequent development of gestational diabetes mellitus (GDM). METHODS: A multicentre, prospective cohort study was conducted in which singleton pregnant Korean women were assessed for NAFLD at 10-14 weeks using liver ultrasound, fatty liver index (FLI) and hepatic steatosis index (HSI). Maternal plasma adiponectin and selenoprotein P concentrations were measured. Participants were screened for GDM using the two-step approach at 24-28 weeks. RESULTS: Six hundred and eight women were included in the final analysis. The prevalence of NAFLD was 18.4% (112/608) and 5.9% (36/608) developed GDM. Participants who developed GDM had a higher prevalence of radiological steatosis (55.6% vs 16.1%; p < 0.001) and higher FLI (40.0 vs 10.7; p < 0.001) and HSI (35.5 vs 29.0; p < 0.001). The risk of developing GDM was significantly increased in participants with NAFLD and was positively correlated with the severity of steatosis. This relationship between NAFLD and GDM remained significant after adjustment for metabolic risk factors, including measures of insulin resistance. Maternal plasma adiponectin and selenoprotein P levels were also correlated with both NAFLD severity and the risk of developing GDM. CONCLUSIONS/INTERPRETATION: NAFLD in early pregnancy is an independent risk factor for GDM. Adiponectin may be a useful biomarker for predicting GDM in pregnant women.

Genotypic prevalence of human papillomavirus infection during normal pregnancy: a cross-sectional study.

AIM: Genital human papillomavirus (HPV) infection is a necessary factor in most cases of cervical cancer, but malignant transformation requires the presence of additional cofactors such as pregnancy. Little is known about the effect of pregnancy on genital HPV carriage. We therefore analyzed the prevalence and genotypic patterns of genital HPV infections in normal pregnancies. METHODS: The prevalence of HPV infection was measured in 960 consecutive normal pregnant or post-partum women by HPV-DNA chip analysis of cervical swabs. Data were analyzed by trimester and adjusted for sociodemographic, reproductive and reported sexual history. RESULTS: The overall prevalence of HPV infection in the population was 24.3%. High-risk HPV genotypes were detected in 68.2% of infected subjects, including HPV 16 (18.7%), 39 (16.4%), 53 (10.1%), and 56 (9.4%). High-risk HPV genotypes were significantly more prevalent in the second trimester (23.8%) compared with the other periods (first trimester, 13.2%; third trimester, 17.4%; post-partum, 15.1%; P = 0.010). However, the high-risk HPV genotypes 16 or 18 were detected most frequently in the third trimester (7.2%) as compared to the other periods (first trimester, 2.9%; second trimester, 5.2%; post-partum, 2.1%; P = 0.03). After adjusting for confounding variables, overall HPV infection (odds ratio = 1.84, 95% confidence interval = 1.24-2.75) and high-risk HPV genotypes (odds ratio = 1.94, 95% confidence interval = 1.23-3.05) were significantly more common in the second trimester. CONCLUSION: The second trimester may be the most vulnerable period in high-risk HPV infections, which necessitates future investigations.

Risk of vertical transmission of human papillomavirus throughout pregnancy: a prospective study.

OBJECTIVE: Much controversy still exists about maternal-to-infant transmission of human papillomavirus (HPV) infection, specifically about the magnitude of the risk and the route and timing of such vertical transmission. This prospective cohort study examines the risk of vertical transmission of maternal HPV in each trimester of pregnancy. STUDY DESIGN: One hundred fifty three healthy pregnant women were followed longitudinally throughout pregnancy and cervical swabs obtained in each trimester and postpartum for HPV detection. Cord blood, neonatal nasopharyngeal aspirates, and placental biopsies were collected at delivery. DNA isolation, polymerase chain reaction, and hybridization were performed using the GG HPV Genotyping Chip Kit (Goodgene Inc., Seoul, Korea). Detection of HPV in neonates was defined as the presence of HPV DNA in either cord blood or neonatal nasopharyngeal aspirate. RESULTS: HPV DNA was detected in 14%(22/153) of healthy women in the first trimester, 18%(22/124) in the second trimester, and 10%(15/153) in the third trimester; 24%(37/153) were positive for HPV DNA on at least one occasion in pregnancy. At birth, 5.2%(8/153) of neonates were HPV DNA positive. Seven of these eight infants were born to HPV-positive mothers. Placental HPV DNA was positive in 3.3%(5/152) of cases, and all five cases were from mothers with at least one HPV-positive test. Detection of HPV DNA in neonates was associated with detection of HPV in mothers during any of the three trimesters of pregnancy. CONCLUSION: HPV DNA was detected at birth in 5.2%(8/153) of neonates born to healthy women, and was associated with the detection of HPV in mothers during any of the three trimesters of pregnancy.