Background Progesterone is being used for the prevention of spontaneous preterm birth(SPTB). It is proposed to promote maintenance of pregnancy via its anti-inflammatory properties and actions on prostaglandin synthesis and uterine smooth muscle.

Objective To determine if women at high risk of cervical shortening have lower serum concentrations of progesterone and if vaginal supplementation influences concentrations.

Methods Serum progesterone was measured using ELISA immunoassay in longitudinal samples (n=226) in 64 women (14-28 weeks) with a history of SPTB. As part of a prospective study, women had cervical length assessment every two weeks and if the cervix was short (<25 mm), they were randomly assigned to cerclage or progesterone. Concentrations of progesterone were measured in longitudinal serum samples. Data were log-transformed, analysed using STATA, and results expressed as geometric means and ratios (95% confidence intervals; CI).

Results Thirty-six percent delivered <37 weeks. Baseline progesterone concentrations (14-18 weeks) were similar (36.1 versus 39.7ng/ml) in preterm and term deliveries (ratio 0.90,95%CI0.75-1.08;p=0.27). Concentrations in women destined to develop a short cervix were similar to those who did not (ratio 0.98,95%CI0.84-1.15;p=0.83). Randomisation to vaginal progesterone supplementation had little effect on serum concentrations (15% increase, 95%CI0.98-1.35;p=0.09). There was no significant difference between women who delivered <34 weeks (effect 0.89,95%CI0.74-1.06;p=0.19) and <37 weeks (effect 0.92,95%CI0.79-1.09;p=0.34) to those who delivered at term.

Conclusions Serum progesterone concentrations were similar in women regardless of shortening, gestation at delivery or treatment. Any beneficial effect of vaginal progesterone supplementation, therefore, is more likely to be mediated via local effects on the cervix.