Abstract
BACKGROUND: Tracheostomy is commonly performed in critically ill patients because of its clinical advantages over prolonged translaryngeal endotracheal intubation. Early tracheostomy has been demonstrated to reduce the duration of mechanical ventilation and length of stay. However, its association with mortality remains ambiguous. This study aimed to evaluate the association between the timing of tracheostomy and mortality in patients receiving mechanical ventilation. METHODS: We performed a retrospective cohort analysis of adult patients who underwent tracheostomy during their intensive care unit (ICU) admission between April 2015 and March 2019. Patients who underwent tracheostomy before or after 29 days of ICU admission were excluded. Data were collected from the nationwide Japanese Intensive Care Patient Database. The primary outcome was hospital mortality. The timing of tracheostomy was stratified by quartile, and the association between patient outcomes was evaluated using regression analysis. RESULTS: Among the 85558 patients admitted to 46 ICUs during the study period, 1538 patients were included in the analysis. The quartiles for tracheostomy were as follows: quartile 1, ≤ 6 days; quartile 2, 7-10 days; quartile 3, 11-14 days; and quartile 4, > 14 days. Hospital mortality was significantly higher in quartile 2 (adjusted odds ratio [aOR]: 1.52, 95% confidence interval [CI]: 1.08-2.13), quartile 3 (aOR: 1.82, 95% CI: 1.28-2.59), and quartile 4 (aOR: 2.26, 95% CI: 1.61-3.16) (p for trend < 0.001) than in quartile 1. A similar trend was observed in the subgroup analyses of patients with impaired consciousness (Glasgow Coma Scale score < 8) and respiratory failure (PaO(2):FiO(2) ≤ 300) at ICU admission (p for trend = 0.081 and 0.001, respectively). CONCLUSIONS: This multi-institutional observational study demonstrated that the timing of tracheostomy was significantly and independently associated with hospital mortality in a stepwise manner. Thus, early tracheostomy may be beneficial for patient outcomes, including mortality, and warrants further investigation.