Abstract
Preterm birth is a major cause of perinatal morbidity and mortality. In France, about 7% of births occur before 37 weeks of gestation, nearly two-thirds of which are spontaneous. Preterm labor (PTL), defined as the combination of uterine contractions and cervical changes between 22 and 37 weeks, represents the most frequent condition leading to prematurity. However, half of the women hospitalized for PTL ultimately deliver at term, highlighting the need for reliable tools to identify those truly at risk. Transvaginal ultrasound measurement of cervical length is currently the reference exam. Its negative predictive value is excellent, but its positive predictive value remains limited and depends on the threshold used. Other ultrasound parameters, such as internal os funneling or the presence of intra-amniotic sludge, have been investigated, though with variable results. Biomarkers identified in cervicovaginal secretions represent an alternative or complementary approach. Fetal fibronectin, initially assessed qualitatively, is now studied quantitatively, allowing sensitivity or specificity to be adjusted depending on the chosen threshold. PAMG-1 shows higher specificity, but its use is not recommended by most professional societies. Maternal serum biomarkers, particularly those related to inflammation or infection, are also being studied, though only on an exploratory basis. Finally, several algorithms combining clinical data, ultrasound, and biomarkers have shown promising results in predicting the risk of short-term delivery. The challenge remains to optimize the management of PTL by identifying all women at high risk, while limiting the iatrogenic effects of unnecessary hospitalizations.