ABSTRACT
BACKGROUND: We evaluated the clinical characteristics and factors associated with mortality in very elderly patients ≥ 90 y of age admitted to the ICU. METHODS: We evaluated age-specific rates of admission and mortality in 16,935 subjects ≥ 18 y old and retrospectively analyzed the clinical data of 155 (0.92%) subjects ≥ 90 y old admitted to the ICU from January 2003 to July 2012. The clinical mortality index was defined as the ICU mortality rate associated with clinical risk factors including poor nutrition, do not resuscitate (DNR) order, pneumonia, chronic renal failure, cancer, mechanical ventilation, use of a vasopressor, and admission from a ward. RESULTS: The mortality rate of ICU subjects ≥ 90 y of age was 32.3%. A Cox's regression hazard model revealed that high glucose (P = .006), poor nutrition (P = .001), high Simplified Acute Physiology Scoring II scores (P < .001), DNR order (P = .002), and vasopressor treatment (P = .03) were independent predictive factors of mortality in subjects ≥ 90 y of age admitted to the ICU. An increasing number of clinical risk factors was associated with progressively higher mortality rates. All subjects with more than 5 risk factors died. CONCLUSIONS: The very elderly subjects (≥ 90 y) admitted to the ICU had a higher mortality rate compared with subjects of other ages. High Simplified Acute Physiology Scoring II scores, poor nutritional status, high glucose, use of vasopressors, and DNR orders should be considered as important predictors of mortality in very elderly ICU patients. The level of ICU treatment should be carefully considered in very elderly patients presenting with 5 or more risk factors.