In a study of 615 new patients with Crohn's disease consecutively diagnosed at the Cleveland Clinic between 1966 and 1969, 592 patients were observed (mean greater than 13 yr, minimum 7 yr), giving a follow-up rate of 96.3%. The original hypothesis was that initial anatomic involvement (the clinical pattern) bears directly on clinical course and prognosis. Disease sites were as follows: 246 ileocolic, 165 small intestine, and 181 colon/anorectal. Among patients with ileocolic disease, 225 (91.5%) had surgery. For the small intestine pattern, the operative incidence was 65.5%; for the colon/anorectal pattern, it was 58%. Operations were for specific reasons: internal fistula with abscess or intestinal obstruction for ileocolic pattern; intestinal obstruction for small intestine pattern; and severe perianal disease or toxic megacolon for colon/anorectal pattern. Complications among nonoperated patients included perianal fistulas and extraintestinal manifestations. No statistical correlation existed between type and duration of medical treatment and prognosis. Seventy-five deaths occurred (12.8%), 36 of which related directly to Crohn's disease. Even after many years, symptoms continued and quality of life tended to be suboptimal among operated patients. For nonoperated patients, the most favorable quality of life was experienced by those with segmental involvement of the colon or ileum. Poor prognosis correlated with ileocolic disease and presence of sepsis because of an internal fistula.