Women with missed obstetric anal sphincter injuries (OASIs) are at an increased risk of anal incontinence. Our aim was to assess the accuracy of three-dimensional (3D) transperineal ultrasound (TPUS) compared with clinical examination for detecting OASIs. A cross-sectional study of women undergoing their first vaginal delivery. Perineal trauma was initially assessed by the doctor or midwife performing the delivery (accoucheur) and women were then re-examined by the trained research fellow (KW). A 3D TPUS was performed immediately after delivery before suturing to identify OASIs. The research fellow's clinical diagnosis was used as the reference standard. A power calculation determined that 216 women would be required for the study. Two hundred and sixty-four women participated and 226 (86%) delivered vaginally. Twenty-one (9%) sustained OASIs. Six (29%) of these tears were missed by the accoucheur but were identified by the research fellow. TPUS identified 19 of the 21 (90.5%) OASIs. One percent (n = 2) had sonographic appearances of an anal sphincter defect that was not seen clinically. The positive and negative predictive value of TPUS to detect OASIs was 91% and 99% respectively. TPUS identified 91% of OASIs compared with 71% detected by the accoucheur, which was not statistically significant. The detection rate of OASIs with TPUS and with the clinical findings of the accoucheur was similar. Given the training and financial implications needed for TPUS, attention needs to be focused on the training of midwives and doctors to identify anal sphincter injuries by clinical examination.