Noninvasive Positive Pressure Ventilation Use and In-Hospital Cardiac Arrest in Bronchiolitis

Abstract

IMPORTANCE: A recent study showed an association between high hospital-level noninvasive positive pressure ventilation (NIPPV) use and in-hospital cardiac arrest (IHCA) in children with bronchiolitis. OBJECTIVES: We aimed to determine if patient-level exposure to NIPPV in children with bronchiolitis was associated with IHCA. DESIGN, SETTING AND PARTICIPANTS: Retrospective cohort study at a single-center quaternary PICU in North America including children with International Classification of Diseases primary or secondary diagnoses of bronchiolitis in the Virtual Pediatric Systems database. MAIN OUTCOMES AND MEASURES: The primary exposure was NIPPV and the primary outcome was IHCA. MEASUREMENTS AND MAIN RESULTS: Of 4698 eligible ICU admissions with bronchiolitis diagnoses, IHCA occurred in 1.2% (57/4698). At IHCA onset, invasive mechanical ventilation (IMV) was the most frequent level of respiratory support (65%, 37/57), with 12% (7/57) receiving NIPPV. Patients with IHCA had higher Pediatric Risk of Mortality-III scores (3 [0–8] vs. 0 [0–2]; p < 0.001), more frequently had a complex chronic condition (94.7% vs. 46.2%; p < 0.001), and had higher mortality (21.1% vs. 1.0%; p < 0.001) compared with patients without IHCA. Return of spontaneous circulation (ROSC) was achieved in 93% (53/57) of IHCAs; 79% (45/57) survived to hospital discharge. All seven children without chronic medical conditions and with active bronchiolitis symptoms at the time of IHCA achieved ROSC, and 86% (6/7) survived to discharge. In multivariable analysis restricted to patients receiving NIPPV or IMV, NIPPV exposure was associated with lower odds of IHCA (adjusted odds ratio [aOR], 0.07; 95% CI, 0.03–0.18) compared with IMV. In secondary analysis evaluating categorical respiratory support in all patients, compared with IMV, NIPPV was associated with lower odds of IHCA (aOR, 0.35; 95% CI, 0.14–0.87), whereas no difference was found for minimal respiratory support (none/nasal cannula/humidified high-flow nasal cannula [aOR, 0.56; 95% CI, 0.23–1.36]). CONCLUSIONS AND RELEVANCE: Cardiac arrest in children with bronchiolitis is uncommon, occurring in 1.2% of bronchiolitis ICU admissions. NIPPV use in children with bronchiolitis was associated with lower odds of IHCA.