Gretchen Bandoli, Laura L Jelliffe-Pawlowski, Sky K Feuer, Liang Liang, Scott P Oltman, Randi Paynter, Kharah M Ross, Christine Dunkel Schetter, Kelli K Ryckman, Christina D Chambers

We hypothesized second trimester serum cortisol would be higher in spontaneous preterm births compared to provider-initiated (previously termed 'medically indicated') preterm births. We used a nested case-control design with a sample of 993 women with live births. Cortisol was measured from serum samples collected as part of routine prenatal screening. We tested whether mean-adjusted cortisol fold-change differed by gestational age at delivery or preterm birth subtype using multivariable linear regression. An inverse association between cortisol and gestational age category (trend p = 0.09) was observed. Among deliveries prior to 37 weeks, the mean-adjusted cortisol fold-change values were highest for preterm premature rupture of the membranes (1.10), followed by premature labor (1.03) and provider-initiated preterm birth (1.01), although they did not differ statistically. Cortisol continues to be of interest as a marker of future preterm birth. Augmentation with additional biomarkers should be explored.