Intra- and interobserver reliability of levator ani muscle biometry and avulsion using three-dimensional endovaginal ultrasonography.

OBJECTIVES

To test intra- and interobserver reliability of assessment of levator ani muscle (LAM) biometry and avulsion using antenatal and postnatal three-dimensional (3D) endovaginal ultrasonography (EVUS), and to determine levator-urethra gap (LUG) values on EVUS.

METHODS

Primigravid women were scanned prior to delivery, early postpartum and 3 months postpartum, with EVUS performed at rest using a standardized protocol. During post-processing, measurements were taken in the plane of minimal hiatal dimensions by two independent investigators blinded to the clinical information and each other's results. LAM attachment to the pubic bone was assessed at the pubococcygeus and puborectalis levels using a score system: (1) intact; (2) partial avulsion (< 50%); (3) partial avulsion (≥ 50%); and (4) complete avulsion. Intraclass correlation coefficients (ICCs) and limits of agreement (LOAs) were calculated for each time point, with intraobserver analysis conducted in a random sample of 20 women scanned 3 months following delivery.

RESULTS

One hundred and sixty-nine antenatal scans, 83 early postpartum scans and 75 scans at 3 months postpartum were performed. The intra- and interobserver ICCs, respectively, were 0.95 and 0.86-0.88 for hiatal area, 0.90 and 0.16-0.74 for hiatal transverse diameter, 0.91 and 0.73-0.80 for hiatal anteroposterior diameter, 0.50 and 0.32-0.52 for LAM thickness at the '9 o'clock position' and 0.55 and 0.33-0.45 for LAM thickness at the '3 o'clock position'. Both intra- and interobserver analysis revealed acceptable LOAs for hiatal measurements, but the LOAs were wide for thickness measurements. The correlation of LAM avulsion score was excellent on intra- and interobserver analysis. Antenatal mean ± SD LUGs were 18.8 ± 2.4 mm and 19.2 ± 2.3 mm on right and left sides, respectively; the intraobserver ICC was 0.82-0.91 but LOAs were wide, while interobserver ICC was 0.13-0.68 and also had wide LOAs.

CONCLUSIONS

3D-EVUS is a reliable tool for the assessment of hiatal measurements and LAM avulsion in women during pregnancy and after delivery, but performs less well for measurements of LAM thickness and LUG. EVUS can therefore be used in research studies involving childbirth and recurrent prolapse.