Predicting short-term and long-term mortality in elderly emergency patients admitted for intensive care

Objective: The long-term outcomes of intensive care for the growing elderly cohort are not well defined. We explored the predictive factors for 12-month mortality in elderly patients who were admitted to an intensive care unit within 24 hours of emergency department (ED) presentation. Design, setting and participants: A retrospective cohort study of 506 patients aged 80 years and over who were admitted to the Royal Melbourne Hospital ICU within 24 hours of presentation to the ED, between 1 January 2005 and 1 December 2010. Main outcome measures and results: After multivariate regression analysis, independent risk factors for mortality 12 months after hospital discharge were the need for mechanical ventilation (odds ratio [OR], 5.16; 95% CI, 3.00-8.86), presence of acute renal failure (OR, 4.71; 95% CI, 2.04-10.84), age (OR, 1.07; 95% CI, 1.01-1.14), Glasgow coma score (GCS) (OR, 0.89; 95% CI, 0.84-0.93) and serum urea level (OR, 1.05; 95% CI, 1.02-1.07). Independent predictors for mortality in the ICU were the presence of acute renal failure (OR, 14.96; 95% CI, 6.50-34.44), the need for mechanical ventilation (OR, 8.13; 95% CI, 2.77-23.89), and GCS (OR, 0.85; 95% CI, 0.79-0.90). Mortality in the ICU was 16.6%, and 12 months after hospital discharge was 46.3%. Conclusions: Physiological parameters present on admission to the ICU including acute renal failure, the need for mechanical ventilation, a low GCS and high serum urea level, as well as age, have independent predictive value for 12-month mortality, but comorbidities were not predictive. This may help clinicians with decisions about who will benefit most from intensive care treatment.