Levator ani muscle avulsion during childbirth: a risk prediction model

Objective To establish the incidence of levator ani muscle (LAM) avulsion in primiparous women and to develop a clinically applicable risk prediction model. Design Observational longitudinal cohort study. Setting District General University Hospital, United Kingdom. Sample Nulliparous women at 36 weeks of gestation and 3 months postpartum. Methods Four-dimensional transperineal ultrasound was performed during both visits. Tomographic ultrasound imaging at maximum contraction was used to diagnose no, minor or major LAM avulsion. A risk model was developed using multivariable ordinal logistic regression. Main outcome measures Incidence of LAM avulsion and its risk factors. Results Of 269 women with no antenatal LAM avulsion 71% (n = 191) returned postpartum. No LAM avulsion was found after caesarean section (n = 48). Following vaginal delivery the overall incidence of LAM avulsion was 21.0% (n = 30, 95% confidence interval [95% CI] 15.1-28.4). Minor and major LAM avulsion were diagnosed in 4.9% (n = 7, 95% CI 2.2-9.9) and 16.1% (n = 23, 95% CI 10.9-23.0), respectively. Risk factors were obstetric anal sphincter injuries (odds ratio [OR] 4.4, 95% CI 1.6-12.1), prolonged active second stage of labour per hour (OR 2.2, 95% CI 1.4-3.3) and forceps delivery (OR 6.6, 95% CI 2.5-17.2). A risk model and nomogram were developed to estimate a woman's individual risk: three risk factors combined revealed a 75% chance of LAM avulsion. Conclusions Twenty-one percent of women sustain LAM avulsion during their first vaginal delivery. Our risk model and nomogram are novel tools to estimate individual chances of LAM avulsion. We can now target postnatal women at risk of sustaining a LAM avulsion.