Prediction of neonatal morbidity and very preterm delivery using maternal steroid biomarkers in early gestation

• Published in: PloS one Vol. 16; no. 1; p. e0243585
• Main Authors: Patil, Avinash S, Grotegut, Chad A, Gaikwad, Nilesh W, Dowden, Shelley D, Haas, David M
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 06.01.2021; Public Library of Science (PLoS)
• Subjects: Adult; Amniotic fluid; Biobanks; Biological markers; Biology and Life Sciences; Biomarkers; Biomarkers - blood; Correlation analysis; Corticosteroids; Demographics; Female; Fetuses; Gestation; Gestational age; Gynecology; Health aspects; Health risks; Humans; Identification and classification; Infant mortality; Infant, Newborn; Infants; Laboratories; Lung diseases; Mass spectroscopy; Medicine; Medicine and Health Sciences; Metabolites; Morbidity; Neonates; Newborn babies; Obstetrics; Patient outcomes; Phenotype; Physical Sciences; Plasma; Plasma levels; Pregnancy; Premature birth; Premature Birth - diagnosis; Premature Birth - epidemiology; Premature labor; Progesterone; Regression Analysis; Regression models; Risk; Risk factors; ROC Curve; Sex; Spectroscopy; Steroids; Steroids - blood; Variables; Young Adult; United States; Indiana; United States > US; Arizona; Indianapolis Indiana
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0243585

Preterm delivery is a common pregnancy complication that can result in significant neonatal morbidity and mortality. Limited tools exist to predict preterm birth, and none to predict neonatal morbidity, from early in pregnancy. The objective of this study was to determine if the progesterone metabolites 11-deoxycorticosterone (DOC) and 16-alpha hydroxyprogesterone (16α-OHP), when combined with patient demographic and obstetric history known during the pregnancy, are predictive of preterm delivery-associated neonatal morbidity, neonatal length of stay, and risk for spontaneous preterm delivery prior to 32 weeks' gestation. We conducted a cohort study of pregnant women with plasma samples collected as part of Building Blocks of Pregnancy Biobank at the Indiana University School of Medicine. The progesterone metabolites, DOC and 16α-OHP, were quantified by mass spectroscopy from the plasma of 58 pregnant women collected in the late first trimester/early second trimester. Steroid levels were combined with patient demographic and obstetric history data in multivariable logistic regression models. The primary outcome was composite neonatal morbidity as measured by the Hassan scale. Secondary outcomes included neonatal length of stay and spontaneous preterm delivery prior to 32 weeks' gestation. The final neonatal morbidity model, which incorporated antenatal corticosteroid exposure and fetal sex, was able to predict high morbidity (Hassan score ≥ 2) with an area under the ROC curve (AUROC) of 0.975 (95% CI 0.932, 1.00), while the model without corticosteroid and fetal sex predictors demonstrated an AUROC of 0.927 (95% CI 0.824, 1.00). The Hassan score was highly correlated with neonatal length of stay (p<0.001), allowing the neonatal morbidity model to also predict increased neonatal length of stay (53 [IQR 22, 76] days vs. 4.5 [2, 31] days, above and below the model cut point, respectively; p = 0.0017). Spontaneous preterm delivery prior to 32 weeks' gestation was also predicted with an AUROC of 0.94 (95% CI 0.869, 1.00). Plasma levels of DOC and 16α-OHP in early gestation can be combined with patient demographic and clinical data to predict significant neonatal morbidity, neonatal length of stay, and risk for very preterm delivery, though validation studies are needed to verify these findings. Early identification of pregnancies at risk for preterm delivery and neonatal morbidity allows for timely implementation of multidisciplinary care to improve perinatal outcomes.