Indication and timing of tracheotomy in burn patients

Abstract Background Tracheotomy is the most common and effective strategy to maintain the airway patency of burn patients. However, the indication and timing of tracheotomy in burn patients were mainly experience-relied and sometimes controversial. Here, we developed classification formula to simply determine the indication and timing of tracheotomy in burn patients. Methods This retrospective study was conducted in all the acute burn patients admitted to a Chinese large burn center from January 1, 2019 to December 31, 2021. The following clinical data was collected from medical records: demographic data, burn causes, burn severity [burn area, Burn Index(BI)], inhalation injury, possible tracheotomy indications, tracheotomy complications and clinical outcomes. Logistic regression analyses were performed to screen possible risk factors and construct classification formulas. Receiver operating characteristic (ROC) curves were used to evaluate the performance by areas under curves (AUCs) and calculate the cut-off value of formulas. Results A total of 951 burn patients were finally enrolled in this study, including 114 patients with tracheotomy and 837 patients without tracheotomy. Furthermore, 18 and 96 cases received emergency and prophylactic tracheotomy, respectively. A classification formula was developed by multivariate logistic regression as follows: Logit(p)=-6.854 + 0.026×Age(years)-0.077×Time (hours from burns to admission) + 0.085×BI (total body) + 0.945×BI (head/face/neck) + Inhalation injury (none:0, mild:1.521, moderate:2.637, severe:4.281). AUCs of the classification formula was 0.972 (95% CI = 0.960–0.984), and higher than that of every individual factor. The cut-off value of classification formula, age, hours after injury, BI of total body and head/face/neck was 0.064, 49.50 years, 14.97 hours, 14.50 and 2.83, respectively. Hours after injury was the only indicator for emergency tracheotomy with AUCs of 0.920 (95% CI = 0.819-1.000) and cut-off value of 32.44 hours. Conclusion The developed formulas had good performance. In clinical practice, burn patients with screened risk factors (age > 49.50, later than 14.97 hours, BI of total body > 14.50, BI of head/face/neck > 2.83, and with inhalation injury) were suspected for tracheotomy and could be further evaluated by the classification formula. If the calculated score was larger than 0.064, tracheotomy was very likely needed and planed tracheotomy would better be performed in 32.44 hours after burns. Together, we provided an alternative strategy for determining the indication and timing of tracheotomy in burn patients.