Cervical Length and Dilation at Cerclage Track Pregnancy Duration in Ten-Year Cohort

Women whose cervix was still long and closed when a stitch was placed to hold a pregnancy in place tended to carry their babies longer, while those whose cervix had already shortened or dilated delivered earlier, according to a ten-year review of 66 cerclage cases at a Turkish tertiary hospital. But the picture is more complicated than that simple pattern suggests: once the researchers adjusted statistically for other factors, no single characteristic remained a significant independent predictor of gestational age at delivery, and the strongest candidate, cervical length, hovered just at the edge of significance and proved sensitive to how the model was built.

The study, published in Frontiers in Global Women’s Health, retrospectively examined medical records of women who underwent cervical cerclage between January 2013 and January 2023 for cervical insufficiency, a condition in which the cervix painlessly dilates and effaces during the second trimester, leading to preterm birth. Cervical insufficiency is reported to occur in roughly 1 percent of pregnancies and accounts for about 20 percent of spontaneous second-trimester losses. Cerclage, a surgical procedure using suture material to keep the cervix closed, is performed electively based on obstetric history, based on ultrasound findings, or as an emergency when dilation is already present on examination.

The cohort included women operated on between 12 and 25 weeks of gestation, with a median gestational age at cerclage of 14 weeks. Their median age was 28 years, median cervical length 26 millimeters, and median hospital stay after the procedure two days. Four pregnancies, 6.1 percent, were multiple gestations, two twin and two quadruplet. Cerclage was examination-indicated, meaning emergency, in 23 women, 34.8 percent; ultrasound-indicated in 8, 12.1 percent; and history-indicated in 35, 53.0 percent. Because the indication was not recorded as a separate field in all patient files, the researchers classified it from the presenting cervical findings, a decision they later flagged as a potential source of misclassification.

In unadjusted analyses, three baseline characteristics stood out. Later gestational age at cerclage correlated with earlier delivery, with a moderate negative correlation of r = −0.349 (p = 0.004). Shorter cervical length correlated with earlier delivery at r = 0.562 (p < 0.01), and greater cervical dilation at r = −0.540 (p < 0.01). Cervical effacement, which was strongly correlated with dilation (Spearman ρ = 0.88), also tracked with earlier delivery. Maternal age, gravidity, parity, abortion history, and previous cerclage showed no statistically significant correlation with gestational age at delivery.

Stratifying by indication sharpened the contrast. Women who received emergency cerclage, defined as cervical dilation of at least 1 centimeter on physical examination, delivered at a median of 31 weeks, compared with 37 weeks for those with non-emergency cerclage (p = 0.007). They were far more likely to deliver before 34 weeks, 65.2 percent versus 16.3 percent (p < 0.001), and had lower median birth weights, 1,650 grams versus 2,700 grams (p = 0.041). They also had shorter cervical lengths at the time of the procedure, a median of 21 millimeters versus 30 millimeters, and more effacement. Maternal age, parity, previous cerclage, and mode of delivery were similar between the two groups.

Within the emergency stratum alone, longer cervical length (Spearman ρ = 0.62, p = 0.002) and less cervical dilation (ρ = −0.55, p = 0.007) were significantly associated with later delivery. Among women with non-emergency cerclage, neither gestational age at cerclage (ρ = 0.18, p = 0.241) nor cervical length (ρ = 0.15, p = 0.350) showed a significant association. A formal interaction test supported a difference between the strata for cervical length (interaction β = 0.45 weeks per millimeter, 95% CI 0.09 to 0.82, p = 0.016) but not for cerclage timing (p = 0.37). Cervical dilation could not be compared across strata because no woman in the non-emergency group had any dilation.

The primary multivariable model, a multiple linear regression with gestational age at delivery in weeks as the outcome, included gestational age at cerclage, cervical length, cervical dilation, parity, and maternal age as covariates. It explained 38 percent of the variance in delivery timing (adjusted R² = 0.33), yet no covariate reached statistical significance. Cervical length showed a borderline positive association (β = 0.21 weeks per millimeter, 95% CI −0.01 to 0.43, p = 0.055), and cervical dilation a borderline negative one (β = −1.36 weeks per centimeter, 95% CI −2.92 to 0.21, p = 0.088). Gestational age at cerclage (β = −0.05, p = 0.849), parity (p = 0.518), and maternal age (p = 0.154) were not associated. Because cervical length and dilation were strongly correlated (r = −0.71), their separate contributions could not be reliably distinguished in this sample.

For the binary outcome of delivery before 34 weeks, with 22 events, logistic regression limited to two covariates found neither cervical dilation (OR 1.83 per centimeter, 95% CI 0.97 to 3.46, p = 0.062) nor cervical length (OR 0.90 per millimeter, 95% CI 0.80 to 1.01, p = 0.064) significant. An exploratory model adding gestational age at cerclage, roughly seven events per variable, also found none of the three significant. Residual diagnostics revealed that the linear model’s residuals departed from normality (Shapiro–Wilk p = 0.003) without evidence of heteroskedasticity (Breusch–Pagan p = 0.68), and no observation was highly influential (maximum Cook’s distance 0.12). When the model was refitted with heteroskedasticity-consistent HC3 standard errors, cervical length reached significance (p = 0.029), while cervical dilation did not (p = 0.137), underscoring that the cervical length result depends on model specification.

A sensitivity analysis excluding the four multiple pregnancies (n = 62) produced similar unadjusted correlations and similar regression estimates, with no covariate reaching significance, suggesting the multiple gestations did not drive the findings. In an alternative linear model in which an indicator for emergency cerclage replaced cervical dilation, cervical length became independently significant (β = 0.31 weeks per millimeter, 95% CI 0.10 to 0.52, p = 0.005), while the emergency indicator itself was not (β = −0.92 weeks, p = 0.646). Taken together, the authors interpret the crude association of later cerclage with earlier delivery as largely reflecting differences in cervical status at presentation, because later cerclage was performed in women with more advanced dilation (Spearman ρ = 0.41 between cerclage week and dilation).

Descriptively, 29 women (43.9 percent) delivered at term, 41 (62.1 percent) delivered by cesarean section, and birth weight was at least 2,500 grams in 41 pregnancies. Apgar scores, presented as descriptive secondary outcomes, were interpreted as delivery-related rather than as measures of cerclage effect. The authors emphasize that without an untreated control group these outcomes cannot quantify the effectiveness of the procedure itself. They position the clinical value of the work in risk stratification among women who have already undergone cerclage, where cervical length and dilation at presentation appear most informative for identifying those at higher risk of earlier delivery, though neither was independently significant after adjustment. Limitations include the retrospective design, small sample size, heterogeneous indications, potential confounding by indication and baseline disease severity, telephone follow-up for some delivery data, and possible underpowering of the multivariable models. The authors call for prospective controlled studies with larger samples, clearly defined indications, and appropriate comparison groups to determine the independent contributions of cervical characteristics and procedure timing to pregnancy outcomes.

Subject of Research: Associations of maternal and cervical characteristics with gestational age at delivery in women undergoing cervical cerclage

Article Title: Clinical and cervical factors associated with obstetric outcomes following cervical cerclage: a ten-year retrospective cohort study

Article References: Sengi, A., Gündüz, R., Özmaskan, F., Çalışır, U., Kaydal, M., & Duran, M. N. (2026). Clinical and cervical factors associated with obstetric outcomes following cervical cerclage: a ten-year retrospective cohort study. Frontiers in Global Women's Health, 7, Article 1986338. https://doi.org/10.3389/fgwh.2026.1986338

Image Credits: AI Generated

DOI: 10.3389/fgwh.2026.1986338

Keywords: cervical cerclage, cervical insufficiency, preterm birth, gestational age at delivery, cervical length, cervical dilation, retrospective cohort, multivariable regression, emergency cerclage, obstetric outcomes, neonatal outcome, risk stratification