Noninvasive Ventilation and Outcomes in Bronchiolitis

Abstract

Objectives: Evaluation of potential benefits of noninvasive ventilation for bronchiolitis has been precluded in part by the absence of large, adequately powered studies. The objectives of this study were to characterize temporal trends in and associations between the use of noninvasive ventilation in bronchiolitis and two clinical outcomes, invasive ventilation, and cardiac arrest.

Design: Multicenter retrospective cross-sectional study.

Setting: Forty-nine U.S. children's hospitals participating in the Pediatric Health Information System database.

Patients: Infants under 12 months old who were admitted from the emergency department with bronchiolitis between January 1, 2010, and December 31, 2018.

Measurements and main results: Primary outcomes were rates of noninvasive ventilation, invasive ventilation, and cardiac arrest. Trends over time were assessed with univariate logistic regression. In the main analysis, hospital-level multivariable logistic regression evaluated rates of outcomes including invasive ventilation and cardiac arrest among hospitals with high and low utilization of noninvasive ventilation. The study included 147,288 hospitalizations of infants with bronchiolitis. Across the entire study population, noninvasive and invasive ventilation increased between 2010 and 2018 (2.9-8.7%, 2.1-4.0%, respectively; p < 0·001). After adjustment for markers of severity of illness, hospital-level noninvasive ventilation (high vs low utilization) was not associated with differences in invasive ventilation (5.0%, 1.8%, respectively, adjusted odds ratio, 1.8; 95% CI, 0·7-4·6) but was associated with increased cardiac arrest (0.36%, 0.02%, respectively, adjusted odds ratio, 25.4; 95% CI, 4.9-131.0).

Conclusions: In a large cohort of infants at children's hospitals, noninvasive and invasive ventilation increased significantly from 2010 to 2018. Hospital-level noninvasive ventilation utilization was not associated with a reduction in invasive ventilation but was associated with higher rates of cardiac arrest even after controlling for severity. Noninvasive ventilation in bronchiolitis may incur an unintended higher risk of cardiac arrest, and this requires further investigation.