J. E. Keenan1, B. C. Gulack1, D. P. Nussbaum1, C. Green1, M. L. Shapiro1, S. N. Vaslef1, J. E. Scarborough1 1Duke University Medical Center,Durham, NC, USA
Introduction: Controversy exists over when tracheostomy should be performed in the critically ill trauma patient with prolonged respiratory failure. The purpose of this study was to determine the optimal timing of tracheostomy in this population using a large national data source.
Methods: The research data set of the National Trauma Data Bank (years 2008-2011) was used to identify a cohort of adult patients who required tracheostomy within the first 21 days of hospitalization after blunt trauma (with ISS ≥9) and who had been directly admitted to the ICU or operating room from the emergency department. Patients who may have required tracheostomy for airway stabilization rather than prolonged respiratory failure were excluded, including those who underwent tracheostomy within 4 days from admission, those with head or neck injuries (defined as abbreviated injury scale ≥1), and those who spent less than one day in the ICU. A restricted cubic spline analysis was performed to determine how timing of tracheostomy was associated with in-hospital mortality. The cohort was stratified based on the results of this analysis. Unadjusted baseline demographics, characteristics, and outcomes were compared between groups. Multivariable logistic regression analysis was employed to evaluate the independent association of timing of tracheostomy with mortality after adjustment for age, gender, race, payor status, ACS trauma verification level, ISS, and emergency department GCS.
Results: 9,187 patients were included in the study. Overall in-hospital mortality was 9.9% (n=899). Based on restricted cubic spline analysis, probability of mortality appeared to be higher in patients who underwent tracheostomy prior to hospital day 11 (Figure). The cohort was therefore stratified into early and delayed tracheostomy groups based on whether the patient underwent tracheostomy before or on/after hospital day 11, respectively. There were 5,121 (55.7%) patients in the early tracheostomy group, and 4,066 (44.3%) in the delayed tracheostomy group. Prior to adjustment, the delayed tracheostomy group had a significantly decreased mortality compared to the early tracheostomy group (8.8% vs 10.7%, p=0.004). Following adjustment with multivariable logistic regression, the delayed tracheostomy group continued to have significantly decreased odds of mortality compared to the early tracheostomy group (AOR: 0.83, 95% CI: 0.71-0.96).
Conclusion: This study of a large national cohort supports that delaying tracheostomy up to hospital day 11 provides survival benefit in patients with prolonged respiratory failure after blunt trauma. Trauma surgeons and critical care physicians should therefore consider avoiding tracheostomy placement prior to this time point.
