OBJECTIVES: To evaluate outcome and risk factors, particularly the Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system, for in-hospital mortality in very elderly patients after admission to an intensive care unit (ICU). METHODS: Retrospective chart review of patients > or =85 years admitted to the ICU. We recorded age, sex, previous medical history, primary diagnosis, date of admission and discharge or death, APACHE II score on admission, use of mechanical ventilation and inotropics, and complications during ICU admission. RESULTS: 104 patients > or =85 years (1.3% of all ICU admissions) were studied. The ICU and in-hospital mortality rates for these patients were 22 and 36% respectively. Factors correlated with a greater in-hospital mortality were: an admission diagnosis of acute respiratory failure (chi2; P = 0.007), the use of mechanical ventilation (chi2; P = 0.00005) and inotropes (chi2; P = 0.00001), complications during ICU admission (chi2; P = 0.004), in particular acute renal failure (chi2; P = 0.005), and an APACHE II score > or =25 (chi2; P = 0.001). The APACHE II scoring system and the use of inotropes were independently correlated with mortality. CONCLUSION: ICU and in-hospital mortality are higher in very elderly patients, particularly in those with an APACHE II score > or =25. The most important predictors of mortality are the use of inotropes and the severity of the acute illness.