The Reliability and Accuracy of Endoscopic Items and Scores Used in the Assessment of the Ileoanal Pouch and Cuff.

Abstract

Currently used endoscopic items for the assessment of pouchitis and cuffitis have deficiencies in reliability and validation. We assessed the reliability and accuracy of new endoscopic items for pouchitis and of the Ulcerative Colitis Endoscopic Index of Severity [UCEIS] for cuffitis. Three new endoscopic items were assessed and included in the Monash pouchitis endoscopic subscore: bleeding [absent/contact/spontaneous]; erosions [absent/<10/≥10]; and ulceration [absent/<10%/≥10%]. Three raters evaluated 44 pouchoscopy videos in duplicates, in random order. Intra- and inter-rater reliability of all endoscopic items and UCEIS were assessed. Clinical and histological pouchitis disease activity index [PDAI] subscores were also assessed and faecal calprotectin was measured. All three Monash endoscopic items had substantial intra-rater reliability with intraclass correlation coefficients [ICCs] >0.61 [95% CI >0.61], compared with only ulcers from the currently used PDAI endoscopic subscore, but inter-rater reliability was only substantial for ulceration and no better than those of the currently used endoscopic items. The Monash endoscopic subscore had a strong positive correlation with the reference standard global endoscopic lesion severity r = 0.80 [95% CI 0.80-0.80] and the reference standard PDAI endoscopic subscore r = 0.70 [95% CI 0.67-0.73], which was higher than the correlation observed for the currently used PDAI endoscopic subscore. The UCEIS had substantial intra-rater reliability, but only fair inter-rater reliability and poor diagnostic performance for cuffitis. The Monash endoscopic items, and endoscopic subscore they generate, have enhanced overall performance compared with the currently used PDAI items and subscore. Further validation and responsiveness to change in disease state are indicated.