Objective: This study aimed to evaluate III and IV degree tears rates and related risk factors in a single Italian centre. The secondary goal was to build a predictive model based on identified risk factors. Study design: This was a retrospective cohort study. All vaginal deliveries from 2011 to 2015 in a single Italian University Hospital were analysed. Univariate analysis was applied to evaluate the overall association between each factor and severe tear. Multivariate logistic regression was used to build a predictive model for the absolute risk of severe tear. We computed a resampling validated measure (AUC) of the predictive accuracy of the model and we provided a nomogram for the risk calculation in clinical practice. Results: 62 out of 10133 patients (0.61%) had a severe perineal tear. Univariate analysis identified gestational age >40 weeks, nulliparity, moderate/severe obesity, oxytocin use in pushing stage, sinciput presentation, instrumental delivery, shoulder dystocia, pushing stage >= 90 min, lithotomy position, birth weight >4 kg, head circumference at birth >34 cm and length at birth >50 cm as risk factors. Multivariate analysis identify moderate/severe obesity (OR = 2.8), instrumental delivery (OR = 2.6) and birth weight (OR = 1.1/hg) as independent risk factors. Using the predicted risk score from the final model (bootstrap-validated AUC 70%), we designed a nomogram for severe perineal tears absolute risk calculation. Conclusion: Moderate/severe obesity, instrumental delivery and foetal weight resulted as independent risk factors for severe obstetrical tears. (C) 2017 Elsevier B.V. All rights reserved.
