Trauma in the very elderly: a community-based study of outcomes at trauma and nontrauma centers.

BACKGROUND: Little research has examined trauma outcomes in the very elderly (>80 years), the fastest growing subset of our geriatric population. Our objective was to describe demographics, mechanism of injury and injury severity of very elderly trauma patients and examine the association between trauma center (TC) verification and hospital mortality in this age group.
METHODS: Retrospective cohort study. Database consisted of a 1996 countywide trauma registry. Subjects consisted of patients > 80 years of age. The setting consisted of Level I (TCI) and Level II (TCII) trauma centers, and acute care (AC) hospitals. The z score analysis was performed using the Major Trauma Outcome Study and a county-specific risk/outcome equation. In addition, a logistic regression model examined hospital mortality (outcome variable) using age, ISS, arrival GCS, and TC verification as predictor variables. Statistical analysis included descriptive statistics; ANOVA; and forward stepwise logistic regression model (OR; 95% CI).
RESULTS: Four hundred fifty-five patients with a mean age of 85.9 (+/-4.8) years (range 80-101). Overall mortality was 9.9%. Using z score analysis, survival at TCII performed as predicted (-1.59), while AC performed less than predicted (-3.41). In the regression model, GCS (OR 0.68; CI 0.57-0.79), ISS (OR 1.1; CI 1.05-1.2) and AC setting (OR 3.2; CI 1.1-9.5) predicted hospital mortality. TCs had significantly better outcomes than AC hospitals in a subset of severely injured patients (ISS 21-45) (56% v 8% survival; p < 0.01).
CONCLUSION: Risk-adjusted outcomes, in this population, differed between TC and AC settings. Head injury, injury severity, and lack of TC verification are associated with hospital mortality in very elderly trauma patients.