RESUMO
BACKGROUND: Vaginal progesterone therapy significantly reduces preterm birth (PTB) rates in those high-risk pregnancies with a sonographic short cervix (≤25 mm) and/or a history of spontaneous PTB. Cervical length (CL) is routinely measured at the midtrimester morphology scan; however, CL surveillance thereafter is not currently recommended. Progesterone's precise mechanism of action remains unknown, though if it indeed influences CL, shortening after treatment initiation could indicate therapeutic failure and risk of PTB. AIMS: The aim was to explore the utility of serial transvaginal ultrasound (TVU) measurement of CL at 16, 19 and 22 weeks for predicting PTB in high-risk pregnancies prescribed progesterone therapy. METHODS: A retrospective cohort study was conducted involving women who attended the King Edward Memorial Hospital PTB Prevention Clinic from 2015 to 2019 and were prescribed progesterone therapy. CL was measured at 16, 19 and 22 weeks by TVU. CL change across three time points was assessed using linear mixed models; then relationships between CL change between 16-19 and 19-22 weeks and PTB were analysed using logistic regression models. RESULTS: Term birth was most likely when CL did not decrease across both time periods. The addition of 16-19 week decrease in CL to a model, including CL at 19 weeks alone, for predicting PTB increased sensitivity from 43.2 to 56.3%, specificity from 73.2 to 77.4%, and overall accuracy from 61.7 to 70.2%. CONCLUSION: For high-risk women prescribed vaginal progesterone therapy, serial measurement of the cervix at 16 and 19 weeks improves clinical ability to predict PTB from current recommendations of 19-week measurement alone.