Objective: To identify the association between cervical length (CL) and gestational age at birth. Design: Prospective cohort study. Setting: Seventeen Brazilian reference hospitals. Population: A cohort of 3139 asymptomatic singleton pregnant women who participated in the screening phase of a Brazilian multicenter randomized controlled trial (P5 trial). Methods: Transvaginal ultrasound (TVU) to measure CL was performed from 18 to 22+6 weeks. Women with CL ≤ 30 mm received vaginal progesterone (200 mg/day) until 36 weeks’ gestation. Main Outcome Measures: Area under receive operating characteristic curve (AUC), sensitivity, specificity, Kaplan-Meier curves for preterm birth (PTB), number needed to screen (NNS). Results: CL ≤25mm was associated with extremely severe, severe, moderate and late PTB, whereas a CL 25–30mm was directly associated with late sPTB. The AUC to predict sPTB<28 weeks was 0.82 and for sPTB<34 weeks was 0.67. Almost half of the sPTB occurred in nulliparous women and CL ≤30mm was associated with sPTB <37 weeks (OR = 7.84; 95%CI = 5.5–11.1). The NNS to detect one sPTB <34 weeks in women with CL ≤25mm is 121 and 248 screening tests are necessary to prevent one sPTB <34 weeks using vaginal progesterone prophylaxis. Conclusions: CL measured by TVU is associated with sPTB <34 weeks. Women with CL ≤30mm are at increased risk for late sPTB. Funding: Bill & Melinda Gates Foundation [OPP1107597], the Brazilian Ministry of Health, and the Brazilian National Council for Scientific and Technological Development (CNPq) [401615/20138]. Keywords: cervical length; number needed to screen; preterm birth; short cervix.

Objective This study aims to determine the performance of cervical pessary in singleton pregnancies with a short cervix based on the learning curve. Design, Settings, Population, and Methods Between 2011 and 2018, 128 singleton gestation between 18th to 24th weeks with a short cervix (<25mm) were referred to our quasi-randomized trial. All cases were treated with progesterone, and, when available, cervical pessary was also offered. Three groups were created for statistical analysis: Group 1 (n=33), treated with progesterone-only; Group 2 and Group 3, treated with cervical pessary plus progesterone. Group 2, included the first cases (n=30), defined by the learning curve and cumulative sum analysis (LC-CUSUM), while Group 3, included the subsequent (n=65). Our outcome was delivery before 34 weeks. Main outcome measures and Results LC-CUSUM demonstrated that 30 patients achieved learning. The preterm birth rate before 34 weeks was 27.3% in Group 1, 20% in Group 2, and 4.6% in Group 3. There was no significant difference in the Group 1/Group 2 comparison (OR 1.10, P=0.945); the Group 1/Group 3 comparison, the difference was significant (OR 0.08, P=0.003). Conclusion LC-CUSUM determined 30 pessaries to achieve the best pessary performance. Cervical pessary plus progesterone can reduce the preterm birth before 34 weeks in patients with a short cervix. Funding This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brazil (CAPES) - Finance Code 001 Keywords Preterm birth; learning curve; cervical pessary; vaginal progesterone; singleton pregnancy; short cervix; transvaginal ultrasound.

Objective Does the combined analysis of compounds in vaginal secretions plus vaginal microbiota composition improve the ability to predict risk for spontaneous preterm birth (SPTB) in women with a short cervix. Design Prospective observational study Setting Three hospitals in Brazil Samples Vaginal secretions from 568 women at 21-24 weeks gestation with cervical length measurement. Method Vaginal microbiome composition was determined by analysis of the V1-V3 region of the bacterial 16S ribosomal RNA gene. D- lactic acid and tissue inhibitor of matrix metalloproteinase (TIMP)-1 in vaginal secretions were measured by ELISA. Main outcome measures SPTB in relation to cervical length and dominant vaginal bacteria. Results When Lactobacillus crispatus was the dominant bacterium the concentration of TIMP-1 was lower, and D-lactic acid was higher, than when either L. iners or Gardnerella vaginalis predominated (p < 0.001). By ROC analysis, a D-lactic acid level < 0.7 mM, and a TIMP-1 level > 1.5 ng/ml best predicted the absence of L. crispatus dominance. Vaginal levels of TIMP-1 were highest (p = 0.024) and D-lactic acid levels were lowest (p=0.032) in women with the shortest cervical length. The absence of L. crispatus dominance increased the sensitivity for predicting a SPTB from 22.4% to 81.8% or 78.3% in women with a cervical length <25mm or <30mm, respectively. Conclusions Low vaginal D-lactic acid and high TIMP-1 indicates the loss of L. crispatus dominance and increased risk for SPTB in women with a short cervix.