BACKGROUND: Patients with stage IV cancer and bowel obstruction (BO) present a complicated management problem. We sought to determine if specific parameters could predict outcome after surgery. METHODS: Records of patients with stage IV cancer and BO treated from 1991 to 2008 were reviewed. For surgical patients, 30-day morbidity and 90-day mortality were assessed using exact multivariable logistic regression methods. RESULTS: Of 198 patients, 132 (66.7%) underwent surgery, 66 medical treatment alone, and demographics were similar. A total of 41 patients (20.7%) were diagnosed with stage IV cancer and BO synchronously, all treated surgically; the remaining presented metachronously. Medically managed patients were more likely to have received chemotherapy in the 30 days prior to BO (45 of 66 [68.2%] vs 40 of 132 [30.3%], p < .01). In the surgical group, 30-day morbidity was 35.6%, while 90-day mortality was 42.3%. Median overall survival for synchronous patients was 14.1 months (95% confidence interval [95% CI] 7.6-23.2), and 3.7 months (95% CI 2.5-5.2) and 3.6 months (95% CI 1.5-5.2) for metachronous patients treated surgically and medically, respectively. A multivariate model for 90-day surgical mortality identified low serum albumin, metachronous presentation, and ECOG > 1 as predictors of death (p < .05). A model for 30-day surgical morbidity yielded low hematocrit as a predictive factor (p < .05). CONCLUSIONS: This cohort identifies characteristics indicative of morbidity and mortality in stage IV cancer and BO. Low serum albumin, ECOG > 1, and metachronous presentation predicted for 90-day surgical mortality. These data suggest factors that can be used to frame treatment discussion plans with patients.