Survival after prolonged length of stay in a trauma intensive care unit

Intensive care unit (ICU) patients comprise a small proportion of patients in the hospital but consume a disproportionate amount of hospital resources. In our cost-conscious environment, it becomes necessary to address the overall performance of our ICUs. This study was designed to analyze survival among trauma ICU (TICU) patients with a length of stay (LOS) >1 month.

We retrospectively reviewed the prospectively collected Trauma Registry Database between January 1, 1995, and January 1, 2005, in an adult TICU from a Level I trauma center. Data on demographics, mechanism of injury, Injury Severity Score (ISS), LOS, and in-hospital survival was collected. Descriptive statistics were calculated and student's t test and comparison of proportions were performed where appropriate. Logistic regression was performed to analyze independent predictors of mortality with significance when p < 0.05.

The initial cohort consisted of 3,556 patients with a mean LOS of 9.8 days (range, 0-274 days). Sixty-nine percent were men, mean age was 44.3 years (range, 0-104 years), and mean ISS was 18 (range, 0-75). The mechanism of injury was blunt trauma in 75%, burns in 15%, and penetrating trauma in 10%. Overall survival was 87%. A total of 339 patients had a LOS >1 month. There was no difference in survival between patients with a LOS <1 month and those with a LOS >1 month (87.1% versus 86.7%). Patients >50 years old (n = 1,251) had a longer LOS (12.5 versus 8.4 days; p < 0.001) and increased mortality (22.1% versus 8.0%; p < 0.001). Age remained an independent predictor of mortality when controlling for ISS.

In our TICU population, extended LOS did not preclude a significant chance of survival. Patients >50 years old had longer LOS and increased mortality. This suggests that the utilization of resources in patients with a prolonged LOS is reasonable and justified.