How comparable is the clinical grading of obstetric anal sphincter injury with that determined by four-dimensional translabial ultrasound?

Objectives To evaluate the agreement between grading of obstetric anal sphincter injuries (OASI) on translabial ultrasound (TLUS), using a newly developed algorithm and on examination, using the current clinical classification. A secondary aim was to assess correlation of tear severity, as defined on ultrasound, with symptoms of anal incontinence and/or fecal urgency. Methods A retrospective study on 260 patients seen at a perineal clinic between 2012 and 2018 after recent primary repair of OASI. All patients underwent a standardized interview including the St. Mark's anal incontinence score (SMIS) and four‐dimensional (4D) TLUS. Post processing of ultrasound volume data was performed blinded against all other data. Using tomographic ultrasound imaging (TUI) a set of eight slices was obtained and the central six slices were evaluated for sphincter abnormalities. Slices with distortion, thinning or defects were rated abnormal. The following algorithm was used to grade OASI: 3a tear was diagnosed if <4/6 slices were abnormal at the external anal sphincter (EAS), 3b if EAS was abnormal in ≥4/6 slices and 3c/4th grade tears if both EAS and internal anal sphincter (IAS) were abnormal in ≥4/6 slices. Clinical grading of OASI was determined as per the Royal College of Obstetricians and Gynecologists (RCOG) guideline. Agreement between clinical and TLUS diagnosis of OASI was evaluated by weighted kappa. Results Forty‐five patients (17%) were excluded due to missing data or a repeated OASI, leaving 215 complete datasets for analysis. The average follow up interval was 2.4 months (range 1‐11) after OASI and the mean age was 29 years (range 17‐42). One hundred and seventy five (81%) were vaginally primiparous. OASI were graded clinically as 3a in 87 women, 3b in 80, 3c in 29 and 4 in 19. On imaging, full agreement between clinical and TLUS grading was noted in 107 (49.7%) with a weighted kappa of 0.398, in 96 there was disagreement by one category (44.6%) with a weighted kappa of 0.74, and in 12 (5.6%) there was disagreement by two categories. Twenty‐four women (11%) were found to have a normal sphincter on imaging. Overall, potential clinical over‐diagnosis was noted in 72 women (33%) and potential under‐diagnosis In 36 (17%). Seniority of the diagnosing obstetrician did not significantly alter agreement between clinical and sonographic OASI grading (kappa 0.44, 0.43, and 0.34, for specialists, senior and junior residents, respectively). The association between symptoms of anal incontinence and/or fecal urgency and TLUS grading did not quite reach significance (P= 0.052). Conclusions Clinical and TLUS‐based grading of OASI showed fair agreement. Clinical over‐diagnosis may be increasingly common in our population, although under‐ diagnosis may still occur in a significant minority.