The primary aim of this study was to define predictors of computed tomography (CT)-guided percutaneous abscess drainage treatment failure in complicated diverticulitis. A 10-year retrospective analysis of inpatients seen in surgical consultation for diverticular abscess management subsequently referred for CT-guided percutaneous drainage (PD) was conducted. The clinical courses of patients undergoing a technically successful PD were categorized into three groups: 1) no colectomy; 2) elective colectomy; and 3) nonelective colectomy. Forty study patients were identified. Thirteen (33%) of the 40 patients required a nonelective colectomy, 20 patients (50%) underwent elective resection, and seven patients (18%) have been managed nonoperatively with no recurrent diverticulitis for a median of 46.8 months (range, 3.2 to 84.3 months). Forward logistic regression identified the presence of immunosuppression or renal insufficiency (creatinine 1.5 mg/dL or greater) as factors independently associated with failure of PD and need for nonelective colectomy. No clinical, laboratory, or radiologic variables were predictive of long-term nonoperative success. Although PD allows for the resolution of intra-abdominal sepsis for most cases of diverticulitis complicated by an abscess, a substantial proportion progress to nonelective colectomy, emphasizing the need for clinical vigilance in follow-up.