Histologic Predictors of Future Ulcerative Colitis Disease Severity

Abstract

Purpose: Accurate prediction of future disease severity course may optimize management of ulcerative colitis (UC) including tailoring “top-down” versus “step-up” approaches to therapy. We sought to determine if specific histologic variable could predict future clinical UC disease severity. Methods: The initial histopathology parameters from 19 patients with UC that later (≥ 6 months) underwent colectomy for severe disease (colectomy/severe course group) were compared to those from 25 patients who did not require colectomy (no colectomy/mild course group) from 1998–2007. The pathology variables assessed were cryptitis, cryptopenia, crypt branching, crypt abscesses, ulcerations and mucin depletion. The predictive value of disease extent, endoscopic severity, and initial serum albumin were also assessed. Groups were compared using Chi-square, Mann-Whitney or t-test with a significance level of 0.05. Patients with dysplasia were excluded. Results: Patients were followed for a median of 3 years. Median age was 39 years in no colectomy group versus 34 years in colectomy group (NS). Severe cryptitis occurred in 16/19 (84%) in colectomy group versus 13/25 (52%) in no colectomy group (P= 0.026). Cryptopenia was present in 11 (58%) colectomy group versus 10 (40%) no colectomy group (P= 0.239). In colectomy group, crypt abscesses were present in 16 (84%), crypt branching 16 (84%), mucin depletion 18 (96%), ulceration 6 (24%) at the initial histopathology. The prevalence of crypt abscess, crypt branching, cryptopenia, mucin depletion, ulceration were all not significantly different in the colectomy (severe course) versus no colectomy (mild course) group. A Mayo Score ≥2 on initial endoscopy was not significantly more common in colectomy group 13 (69%) versus no colectomy group 18 (75%) (P= 0.228). The initial albumin was significantly lower in the colectomy group 2.9 g/dL (SD ± 0.88) versus noncolectomy group 3.5 g/dL (SD ± 0.57) (P= 0.040). Pancolitis at initial endoscopy was not more prevalent in the colectomy group (32%) versus noncolectomy (32%). Immunosuppressant usage was not statistically different between groups: colectomy group (24%) versus no colectomy (32%). Conclusion: Cryptopenia, mucin depletion, crypt branching and crypt abscesses are common findings in UC but do not predict future clinical disease severity. The initial serum albumin and cryptitis were the only variables predictive of future progression to colectomy. Histopathology with extensive cryptitis was more predictive of future severe disease that eventually required colectomy than was the initial endoscopic severity or disease extent. Future prospective studies are required to confirm the prognostic value of diffuse cryptitis to predict a future colectomy in UC.