Abstract Background Previous studies show an association between maternal plasma and salivary cortisol and preterm birth but have been primarily conducted in high-income countries. It is unknown whether salivary cortisol is a risk factor for preterm birth in Ghana. Our objective was to determine whether maternal salivary cortisol during pregnancy was associated with pregnancy duration and preterm delivery in Ghana. Methods We conducted a cohort study of 783 pregnant women in Ghana. We measured salivary cortisol at baseline (mean 16 wk), 28 wk., and 36 wk. gestation. Pregnancy duration was determined primarily by ultrasound. We used adjusted linear regression models to examine the association between cortisol and pregnancy duration and Poisson regression models to determine the risk of preterm delivery among women with high cortisol at baseline or 28 wk. gestation. Results Mean pregnancy duration was 39.4 ± 1.8 wk. and 6.6% had a preterm delivery. Mean maternal cortisol increased throughout pregnancy, from 4.9 ± 2.7 nmol/L at baseline (16 wk) to 6.4 ± 3.2 nmol/L at 28 wk. and 7.9 ± 3.0 nmol/L at 36 wk. gestation. In adjusted analyses, higher cortisol concentrations at baseline (β = − 0.39, p = .002) and 28 wk. (β = − 0.49, p = .001), but not 36 wk. (β = − 0.23, p = .084) were associated with a shorter pregnancy duration. Women with high cortisol at baseline (> 6.3 nmol/L) had an increased relative risk of preterm delivery (RR (95% CI): 1.96 (1.13, 3.40)), but the association between high cortisol at 28 wk. and preterm delivery was not significant. There was a significant interaction with fetal sex (p-for-interaction = 0.037): among women carrying male fetuses, high cortisol at baseline increased the risk of preterm delivery threefold (3.18 (1.51, 6.71)) while there was no association (1.17 (0.50, 2.74)) among women carrying female fetuses. Conclusion Higher maternal cortisol is associated with a shorter pregnancy duration and an increased risk of preterm delivery. Subgroup analysis by fetal sex revealed that this association is evident primarily among women carrying male fetuses. Future studies of cortisol and preterm delivery should include consideration of fetal sex as a potential effect modifier.

2022