Tracheostomy timing in traumatic brain injury: a propensity-matched cohort study.

BACKGROUND The optimal timing of tracheostomy in patients with severe traumatic brain injury (TBI) is controversial; observational studies have been challenged through confounding by indication, and interventional studies have rarely enrolled patients with isolated TBI. METHODS We included a cohort of adults with isolated TBI who underwent tracheostomy within 1 of 135 participating centers in the American College of Surgeons' Trauma Quality Improvement Program, during 2009 to 2011. Patients were classified as having undergone early tracheostomy (ET, ≤8 days) versus late tracheostomy (>8 days). Outcomes were compared between propensity score-matched groups to reduce confounding by indication. In sensitivity analyses, we used time-dependent proportional hazard regression to address immortal time bias and assessed the association between hospital ET rate and patients' outcome at the hospital level. RESULTS From 1,811 patients, a well-balanced propensity-matched cohort of 1,154 patients was defined. After matching, ET was associated with fewer mechanical ventilation days (median, 10 days vs. 16 days; rate ratio [RR], 0.70; 95% confidence interval [CI], 0.66-0.75), shorter intensive care unit stay (median, 13 days vs. 19 days; RR, 0.70; 95% CI, 0.66-0.75), shorter hospital length of stay (median, 20 days vs. 27 days; RR, 0.80; 95% CI, 0.74-0.86), and lower odds of pneumonia (41.7% vs. 52.7%; odds ratio [OR], 0.64; 95% CI, 0.51-0.80), deep venous thrombosis (8.2% vs. 14.4%; OR, 0.53; 95% CI, 0.37-0.78), and decubitus ulcer (4.0% vs. 8.9%; OR, 0.43; 95% CI, 0.26-0.71) but no significant difference in pulmonary embolism (1.8% vs. 3.3%; OR, 0.52; 95% CI, 0.24-1.10). Hospital mortality was similar between both groups (8.4% vs. 6.8%; OR, 1.25; 95% CI, 0.80-1.96). Results were consistent using several alternate analytic methods. CONCLUSION In this observational study, ET was associated with a shorter duration of mechanical ventilation, intensive care unit stay, and hospital stay but not hospital mortality. ET may represent a mechanism to reduce in-hospital morbidity for patients with TBI. LEVEL OF EVIDENCE Therapeutic study, level II.