Introduction: The target of our study is to determine the affiliation of epidemiologic characteristics and death in critical oncologist patients. Methods: It’s a retrospective cohort study in the ICU of “Clinica Las Americas” from January 1, 2004 through December 31, 2006. We included all patients older tan 5 years old with a diagnosed malignancy and treated in the ICU for more than 24 hours. Results: We analyzed 100 patients (57 of which were women), the median age was 60 years old. 50 patients were admitted into the ICU from surgical rooms, 55 had enduring comorbidities, 90 had solid tumors, 44 had received previous oncology surgery, 67 were admitted for postoperative care. The APACHE II average score was 18, while 61 presented organic dysfunction, 53 presented a nosocomial infection, the most common of which was peritonitis (25 cases). We obtained microbiological isolation in 43 cases, the average stay in the ICU was 11 days, 26 patients died in the ICU, the average hospital stay was 18 days and the in-hospital mortality was of 4 cases. In the two-tailed analysis, the relation mortality and APACHE II demonstrated a median of 25 (IQR 12-32) (p=0,004), between renal replacement therapy of renal substitution and death (OR=3,18, 95%CI=0,59-17,14, p=0,10), between transfusion and mortality (OR=2,48, 95%CI=0,89-7, p=0,05), between infection and death (OR=3,19, 95%CI=1,09-9,68, p=0,017), between dysfunction and death (OR=26,23, 95%CI=3,43-559, p=0,00001) and between hematology tumor and death (OR=8,72, 95%CI=2,06-36,9, p=0,003). Conclusions: There are no solid arguments to support the politics of non-admission of a patient with cancer to the ICU: their mortality is high but the number is comparably similar to that of many severe illnesses different from cancer.