Prediction of spontaneous preterm birth: salivary progesterone assay and transvaginal cervical length assessment after 24 weeks of gestation, another critical window of opportunity.

Measurement of salivary progesterone (SP4) levels and cervical length (CL) after 24 weeks to assess their potential predictive value among asymptomatic women at high risk of spontaneous preterm birth (PTB). This prospective observational (noninterventional) study consecutively recruited asymptomatic women at high risk of spontaneous PTB. SP4 and CL were measured at recruitment (24-28 weeks of gestation) then repeated after 3-4 weeks. All recruited women were followed up regularly till delivery. The primary outcome measure was the occurrence of spontaneous PTB. One hundred and thirty four women completed the study, 22 (16.4%) and 32 (23.9%) women had early (<34 weeks) and late (≥34 weeks) PTB, respectively. Initially, the mean CL was 3.2 ± 0.6 cm and the mean SP4 was 4062.8 ± 814.6 pg/ml; with follow up, the mean CL became 3.0 ± 0.6 cm and the mean SP4 became 3871.6 ± 1136.9. Women with early PTB had significantly lower initial and follow up CL and SP4 measures when compared to women with late PTB and those who had birth at term. The rate of drop in SP4 and CL measurements between the two visits was also significantly higher among women with early PTB than those with late PTB and term birth. Receiver-operating characteristic (ROC) curves showed that, CL was a good predictor but SP4 was a better predictor of PTB as the area under the curve (AUC) for CL was less than that for SP4 at both visits (i.e. 0.858 and 0.868 versus 0.986 and 0.990 at the initial and follow up visits, respectively). There was a statistically significant correlation between CL and SP4 measurements. Multivariable binary logistic regression analysis revealed that follow up SP4 measurement was the only independent predictor of spontaneous PTB, and neither BMI, maternal age, SP4 nor CL were independent predictors of early spontaneous PTB. After 24 gestational weeks, SP4 assessment is a simple and reliable promising tool to predict spontaneous PTB among asymptomatic high-risk women, with a little superior performance than CL measurement.