Intraobserver and interobserver measurements of the anorectal angle and perineal descent in defecography

PURPOSE: Anorectal angle and perineal descent can be measured either by dra wing a line defined by the impression of the puborectalis muscle and the ta ngential of the posterior rectal wall(Method A) or by drawing a straight li ne at the level of the posterior rectal wall parallel to the central longit udinal axis of the rectum (Method B). The aim of this study was to assess t he reproducibility of measuring anorectal angle and perineal descent by two different methods according to intraobserver and interobserver measurement and to evaluate which method yields more consistent results. METHODS: Five physicians who have had an average of 1.3 pears (range, 6 months to 1.5 ye ars) experience in defecographic measurement drew both lines on 63 randomly selected defecographic films and measured anorectal angle and perineal des cent by the two methods. The defecographic parameters were measured twice b y each observer during a three-week interval. To avoid potential bias, one physician who did not participate in either measure-ment of perineal descen t or anorectal angle performed all data collection. Intraobserver and inter observer agreement was quantified using Shrout and Fleiss intraclass correl ation coefficients. RESULTS: The mean and range of intraclass correlation c oefficients for intraobserver agreement of measuring anorectal angle and pe rineal descent by Method A were 0.71 (0.6-0.78) and 0.89 (0.74-0.97), respe ctively, whereas with Method B the coefficients were 0.81 (0.73-0.89) and 0 .93 (0.89-0.99), respectively. Regarding the interobserver agreement of the five observers, the mean coefficients for measurement of both anorectal an gle and perineal descent by both methods showed similar agreement levels (0 .88 and 0.98 by Method A and 0.89 and 0.97 by Method B), The mean (+/- stan dard deviation) values of anorectal angle and perineal descent found by Met hod B were significantly larger than those found by Method A (103.3 degrees +/- 19.6 and 6.56 +/- 3.20 cm and 91.1 degrees +/- 25.6 and 5.64 +/- 3.42 cm, respectively; P < 0.001). CONCLUSION: Intraobserver and interobserver i ntraclass correlation coefficients of anorectal angle and perineal descent, which were measured by both methods, were more than 0.60, indicating that both methods are reliable and consistent for measurement of anorectal angle and perineal descent. However, centers should consistently use the same li ne for measurement of anorectal angle and perineal descent because of the s tatistically significant differences between the two methods and the possib ility of inconsistent results.