Abstract Objective: To investigate gestational age–specific variations in cervical length (CL) and evaluate the predictive ability of CL for preterm delivery (PTD) across different trimesters. Methods: This retrospective longitudinal study included a total of 8659 women who delivered singleton live births between June 1, 2022, and December 31, 2024, at Peking University First Hospital. Associations between CL measured at ≤ 19 +6 , 20–24 +6 , and 28–33 +6 gestational weeks and PTD were evaluated using receiver operating characteristic curve analysis and log-rank tests. Results: CL showed substantial variability across gestational ages, with pronounced shortening observed during 28–33 +6 weeks. Significant differences in CL distribution between women with PTD and those with term delivery emerged as early as 20–24 +6 weeks. The predictive value of CL for PTD increased with advancing gestational age. Optimal CL thresholds for predicting PTD ≤ 34 +6 weeks were 28 mm, 15 mm, and 10 mm at ≤ 19 +6 , 20–24 +6 , and 28–33 +6 weeks, respectively, while corresponding thresholds for PTD < 37 weeks were 28 mm, 26 mm, and 10 mm. Women with CL below these cut-offs had significantly lower probabilities of term delivery. The 26-mm threshold at mid-gestation (20–24 +6 weeks) closely aligned with the conventional 25-mm definition of short cervix, supporting the validity of the minimum P approach. The consistent 10-mm threshold in late gestation suggests that lower diagnostic cut-offs after 28 weeks may be required to guide clinical management. Conclusion: CL progressively decreases with increasing gestational age and has emerged as a more potent predictor of PTD after 20 weeks of gestation. A short cervix in late pregnancy may be more appropriately defined as < 10 mm at 28–33 +6 gestational weeks.
Zhang et al. (Fri,) studied this question.