Uterocervical angle and cervical length for prediction of preterm birth: a retrospective cohort study

Abstract Background Preterm birth (PTB) remains a leading cause of neonatal morbidity and mortality worldwide. Although transvaginal cervical length (CL) is an established predictor, its discriminatory performance is limited. The uterocervical angle (UCA) has been proposed as an additional sonographic parameter reflecting biomechanical properties of the cervix and lower uterine segment, but evidence regarding its clinical utility remains limited. This study aimed to evaluate the predictive value of UCA and CL for PTB and to assess whether their combined use improves risk stratification. Methods This retrospective cohort study included 548 pregnancies at the University Hospital for Gynecology and Obstetrics Innsbruck (2013–2023). Transvaginal ultrasound measurements of CL and UCA were analyzed. Associations with PTB were assessed using multivariable logistic regression, and discriminatory performance was evaluated via receiver operating characteristic (ROC) curve analysis. Optimal cut-off values were determined using Youden index. In a predefined subgroup of 366 pregnancies, models incorporating CL, UCA, and PartoSure were analyzed. Results PTB occurred in 278 (50.7%) pregnancies, while 270 (49.3%) delivered at term. No significant differences were observed for measurements taken between 24 + 0 and 27 + 6 weeks. In contrast, between 28 + 0 and 34 + 0 weeks, CL was significantly shorter and UCA significantly wider in pregnancies resulting in PTB. Both parameters remained independently associated with PTB in multivariable analysis. ROC analysis demonstrated moderate discriminatory ability for CL (AUC = 0.614) and weaker performance for UCA (AUC = 0.561), while their combination improved discrimination (AUC 0.644). Using Youden index–derived cut-offs (CL < 15.5 mm; UCA ≥ 101°), the observed proportion of PTB was highest when both markers were present. In the PartoSure subgroup ( n = 366), short CL (OR 2.54, 95% CI 1.52–4.25, p < 0.001) and wide UCA (OR 2.10, 95% CI 1.21–3.64, p = 0.009) remained independently associated with PTB, whereas PartoSure positivity was not (OR 1.66, 95% CI 0.74–3.70, p = 0.217). The model integrating CL, UCA, and PartoSure yielded the numerically highest discriminatory performance among the models evaluated in this subgroup (AUC = 0.609). Conclusions CL and UCA measured between 28 + 0 and 34 + 0 weeks are independently associated with PTB and provide complementary information for risk stratification. While overall predictive performance remains moderate, combining ultrasound parameters with PartoSure improves discrimination and may enhance clinical decision-making. Trial registration Retrospectively registered.