Colocutaneous fistulas complicating diverticulitis.

The records of 93 patients with colocutaneous fistulas associated with diverticulitis treated at the Cleveland Clinic between 1965 and 1983 were reviewed. There were 56 males and 37 females with an age range of 19 to 80 years (median, 57 years). Eighty-eight fistulas followed surgery for diverticulitis while five developed spontaneously. The presence of a diverting stoma in 34 patients did not prevent fistula formation but did decrease morbidity (x2 = 12.75, P less than 0.001). Initial investigations showed a high incidence of recent weight loss (in 40 percent) and hypoalbuminemia (47 percent), although these factors did not influence outcome. Patients with high output (greater than 200 cc/day) fistulas (n = 9) fared significantly worse than those with low outputs. There were 28 patients with fistulas to other organs, 20 involving small bowel. Factors leading to persistence of the fistulas included sepsis (42 cases) and sigmoid colon distal to an intended colorectal anastomosis (38 cases). Ninety-two patients underwent surgery, 80 percent having a one- or two-stage resection and anastomosis. There was one postoperative death and complications occurred in 44 patients (48 percent). Surgery was successful in producing patients without stoma or fistula in 71 cases (77 percent). There were five recurrent fistulas, 14 new fistulas, and 13 patients retained their stomas. A diagnosis of Crohn's disease was made in ten patients who had a high rate of complicated fistulas, recurrent fistulas, and retained stomas. Patients with carcinomas (n = 5) also did poorly, but those on systemic steroids (n = 7) fared no worse than patients not receiving them. This study emphasizes the role of diversion of the fecal stream in reducing the morbidity of colonic fistulas. It is clearly important to carry out a true colorectal anastomosis after resection for diverticulitis, and in patients with unusually complicated clinical courses, the diagnosis of Crohn's disease should be entertained.