The Predictive Applicability of Liberal vs Restrictive Intubation Criteria in Adult Patients With Suspected Inhalation Injury—A Retrospective Cohort Study

Abstract

Abstract This study compares the ability of liberal vs restrictive intubation criteria to detect prolonged intubation and inhalation injury in burn patients with suspected inhalation injury. Emerging evidence suggests that using liberal criteria may lead to unnecessary intubation in some patients. A single-center retrospective cohort study was conducted in adult patients with suspected inhalation injury admitted to intensive care at Queen Elizabeth Hospital, Birmingham between April 2016 and July 2019. Liberal intubation criteria, as reflected in local guidelines, were compared to restrictive intubation criteria, as outlined in the American Burn Association guidelines. The number of patients displaying positive characteristics from either guideline was compared to the number of patients who had prolonged intubation (more than 48 hours) and inhalation injury. In detecting a need for prolonged intubation (n = 85), the liberal criteria had greater sensitivity (liberal = 0.98 [0.94–1.00] vs restrictive = 0.84 [0.75–0.93]; P = .013). However, the restrictive criteria had greater specificity (restrictive = 0.96 [0.89–1.00] vs liberal = 0.48 [0.29–0.67]; P < .001). In detecting inhalation injury (n = 72), the restrictive criteria were equally sensitive (restrictive = 0.94 [0.87–1.00] vs liberal = 0.98 [0.84–1.00]; P = .48) and had greater specificity (restrictive = 0.86 [0.72–1.00] vs liberal = 0.04 [0.00–0.13]; P < .001). In patients who met liberal but not restrictive criteria, 65% were extubated within 48 hours and 90% did not have inhalation injury. Liberal intubation criteria were more sensitive at detecting a need for prolonged intubation, while restrictive criteria were more specific. Most patients intubated based on liberal criteria alone were extubated within 48 hours. Restrictive criteria were highly sensitive and specific at detecting inhalation injury.