Validity and reproducibility of the digital detection of levator trauma

Abstract

Levator ani muscle trauma is a common consequence of vaginal childbirth and detectable on digital vaginal palpation. To ascertain validity and reproducibility of this test, we saw 110 women for an interview, prolapse staging, digital vaginal palpation by two blinded examiners and four-dimensional translabial ultrasound. The mean age was 55.5 years (range 17-85) and the median parity was 2 (range 0-8). Three patients could not be assessed, leaving 107 datasets representing 214 assessments of a right or left puborectalis muscle. Levator defects were found in 21 women (20%) with nine bilateral defects. There was agreement between assessors in 173/214 (81%), k=0.411, signifying moderate agreement. Agreement with an independent blinded review of tomographic ultrasound data was k=0.495. Even after substantial training, the agreement between assessors using digital palpation for the diagnosis of levator trauma remains only moderate. There seems to be a substantial learning curve. Palpatory detection of major levator trauma is less repeatable than identification by ultrasound.