A 40-year-old male who was found to be human immunodeficiency virus-positive when he presented with bloody diarrhea in 1986 is described. Clinical, laboratory, endoscopic and histological findings were all compatible with ulcerative colitis, and stool cultures were repeatedly negative for pathogens. Colitis was initially mild and controlled with intermittent oral aminosalicylic acid products. Since 1993 he has had more significant symptoms requiring prednisone up to 40 mg/day. Repeat colonoscopy disclosed pancolitis and biopsies did not show evidence of cytomegalovirus infection. He has not had an acquired immune deficiency syndrome-defining illness. CD4 cells fell below normal as his colitis worsened. This case raises questions about immune regulation in ulcerative colitis because the patient has active disease in addition to a reduced number of T helper cells. It also presents a difficult management problem because the patient has a limited life expectancy and is reluctant to accept colectomy, and further immunosuppressive therapy may be dangerous.