We estimated the extent of the inflammatory mucosal lesion by colonoscopy and biopsy in 60 patients with their first attack of ulcerative colitis. Proctitis was found in 12, proctosig-moiditis in 19, left-sided colitis in 10, and extensive colitis in 19. Fourteen clinical variables and laboratory measurements (bowel frequency, stool consistency, rectal bleeding, fecal mucus/pus, temperature, pulse rate, white blood cell count, hematocrit, sedimentation rate, serum iron, serum albumin, serum α2-globulin, serum C-reactive protein, and seromucoids) were determined. All the variables except rectal bleeding and hematocrit were correlated (p > 0.001) with the extent of colitis. On stepwise discriminant analysis, only C-reactive protein distinguished proctosigmoiditis from more extensive disease at a level of p > 0.001; no other variables improved the discrimination. Cross-validation by the “jack-knife method” showed that 86.7% of patients were correctly classified, the errors consisting in underestimation of disease in 8/29 patients with extensive colitis.
© Lippincott-Raven Publishers.