How reliable is measurement of the anorectal angle by videoproctography?

The anorectal angle can be determined either by constructing a straight line along the lower border of the rectum (Method A) or by using the central longitudinal axis of the lower rectum (Method B). We have used a computer program to derive the centroid of the rectum for Method B. The coefficients of variation for angles measured at rest, during maximum pelvic floor contraction, and during attempted defecation were 0.616, 0.351, and 0.358, respectively, compared with 0.993, 0.972, and 0.968 for Method B. The presence of a rectocele had no influence on the measurement of the anorectal angle in incontinence, but there was a significant difference in assessment of the angle between constipated patients (P less than 0.05) and controls (P less than 0.05). Posterior indentation of the rectum had no significant influence on measurement of the angle in any group. These data indicate that a computer-derived centroid is more reliable for measurement of angles, but a correction factor for anterior rectocele is needed in constipated patients and controls.