Histologic findings at diagnosis as predictive markers of clinical outcome in pediatric ulcerative colitis

Background: The role of histological inflammation at diagnosis as a possible prognostic factor for disease course has not been investigated. Aims: To assess whether histologic findings at diagnosis could predict clinical outcomes and evaluate the association between clinical, biochemical, endoscopic, and histological findings. Methods: Prospective single-center study including pediatric UC patients with a minimum follow-up of 12 months. The association between histological activity (Nancy Index, Robarts Histopathology Index, and Geboes Score) and 12-month clinical outcomes was evaluated. Secondarily, we assessed the correlation between histological scores and endoscopic and inflammatory markers at the diagnosis. Inter-observer agreement for histologic and endoscopic scores was also evaluated. Results: Forty-nine UC patients were included. No association was found between 1-year clinical relapse and the three histological indices at diagnosis (p > 0.05). Good concordance was found among the three histological scores (p < 0.001), and between all histological and endoscopic indices (p < 0.05). No correlation was found between histologic scores and serum inflammatory markers. Inter-observer agreement was good for eMayo, Nancy and Robarts score (k = 0.71, k = 0.74 and k = 0.68, respectively) and moderate for Geboes (k = 0.46). Conclusions: Histological findings at diagnosis cannot be used as a predictor of the disease course. The three histological scores used in routine clinical practice show an overall good correlation and reliability.