Adults (age > 50 years) admitted to the surgical intensive care unit (SICU) are at high risk for delirium. Little is known about the role traumatic injury plays in the development of delirium because these patients have often been excluded from studies. Identification of specific risk factors for delirium among older adults following injury would be useful to guide prevention strategies. We attempted to identify modifiable factors that would predict delirium in an older trauma population admitted to the SICU.
Data were collected prospectively from July 2012 to August 2013 at a Level I trauma center on consecutive trauma patients, older than 50 years, admitted to the SICU. Patients who died in the SICU were excluded. Delirium was assessed every 12 hours using the Confusion Assessment Method for the ICU scale. Demographic, injury, social, and clinical variables were reviewed. Bivariate analysis determined significant factors associated with delirium. A multivariate logistic regression model was used to predict delirium risk. After preliminary results, additional analysis compared patients with chest injury (defined as chest Abbreviated Injury Scale [AIS] score ≥ 3) with those without.
A total of 115 patients met criteria, with a mean age of 67 years, Injury Severity Score (ISS) of 19, and Glasgow Coma Scale (GCS) score of 14. The incidence of delirium was 61%. Variables present on admission, which were positive predictors of delirium, were as follows: age, ISS greater than 17, GCS score less than 15, substance abuse, and traumatic brain injury (defined as head AIS score ≥ 3). Chest injury (defined as chest AIS score ≥ 3) was a negative predictor of delirium. Significant risk factors influenced by clinical treatment included doses of opioids and propofol, restraint use, and hours deeply sedated (Richmond Agitation Sedation Scale [RASS] score ≤ -3). Clinical treatments with negative predictability were ventilator-free days/30 (vent-free), benzodiazepine-free days/30 (benzo-free), and restraint-free days/30. In a regression model considering age, vent-free days, chest injury, traumatic brain injury, GCS score, benzo-free days, and hours sedated, only age (odds ratio [OR], 1.1; 95% confidence interval [CI], 1.01-1.1; p = 0.03) was a predictor of delirium, while vent-free days (OR, 0.79; 95% CI, 0.65-0.96; p = 0.02) and chest injury (OR, 0.3; 95% CI, 0.09-0.83; p = 0.02) were significant negative predictors of delirium. Patients with chest injury had lower delirium incidence (44%) versus those without (75%) (p = 0.002) despite similar GCS score, ISS, and clinical variables.
Delirium is common in older trauma patients admitted to the SICU, and for every year for those older than 50 years, the chance of delirium increases by 10%. While higher ISS increases delirium risk, we identified several modifiable treatment variables including days patients were deeply sedated, mechanically ventilated, and physically restrained. Interestingly, patients with chest injury experienced less delirium, despite similar injury severity and clinical variables, perhaps owing to frequent health care provider interactions.
Level of evidence:
Prognostic/epidemiologic study, level III.