There is an increased risk of developing colorectal neoplasia in ulcerative colitis (UC) and surveillance colonoscopy is recommended for early detection. We investigated the precise features of UC retrospectively to identify a subgroup with longstanding extensive UC at increased risk of neoplasia. From 1985 to August 1997, we experienced eight UC patients with colorectal cancer and eight with definite dysplasia. All 16 had extensive disease of seven years or more in duration. During the same period, 61 of 334 UC patients without colorectal neoplasia were available for detailed study, allowing evaluation of non-surgical patients with extensive colitis of seven years or more in duration. Basic clinical factors including family history of cancer, expressions of disease activity and durations of pharmacotherapy were investigated. Univariate analysis revealed four significant factors: intractability (P=0.001), periods of inflammation persisting for 3 months or more (P<0.01) and total durations of diarrhea (P<0.01) and hematochezia (P<0.05). The number of admissions and the duration of systemic steroid administration were higher in the neoplasia group but without statistical significance. Multivariate analysis revealed two significant factors: duration of diarrhea (P<0.001) and age at onset (P<0.01). Chronic active disease is a risk factor for colorectal cancer or dysplasia in extensive and longstanding UC.