Objective: To describe the epidemiology of noninvasive ventilation therapy for patients admitted to pediatric cardiac ICUs and to assess practice variation across hospitals.
Design: Retrospective cohort study using prospectively collected clinical registry data.
Setting: Pediatric Cardiac Critical Care Consortium clinical registry.
Patients: Patients admitted to cardiac ICUs at PC4 hospitals.
Measurements and main results: We analyzed all cardiac ICU encounters that included any respiratory support from October 2013 to December 2015. Noninvasive ventilation therapy included high flow nasal cannula and positive airway pressure support. We compared patient and, when relevant, perioperative characteristics of those receiving noninvasive ventilation to all others. Subgroup analysis was performed on neonates and infants undergoing major cardiovascular surgery. To examine duration of respiratory support, we created a casemix-adjustment model and calculated adjusted mean durations of total respiratory support (mechanical ventilation + noninvasive ventilation), mechanical ventilation, and noninvasive ventilation. We compared adjusted duration of support across hospitals. The cohort included 8,940 encounters from 15 hospitals: 3,950 (44%) received noninvasive ventilation and 72% were neonates and infants. Medical encounters were more likely to include noninvasive ventilation than surgical. In surgical neonates and infants, 2,032 (55%) received postoperative noninvasive ventilation. Neonates, extracardiac anomalies, single ventricle, procedure complexity, preoperative respiratory support, mechanical ventilation duration, and postoperative disease severity were associated with noninvasive ventilation therapy (p < 0.001 for all). Across hospitals, noninvasive ventilation use ranged from 32% to 65%, and adjusted mean noninvasive ventilation duration ranged from 1 to 4 days (3-d observed mean). Duration of total adjusted respiratory support was more strongly correlated with duration of mechanical ventilation compared with noninvasive ventilation (Pearson r = 0.93 vs 0.71, respectively).
Conclusions: Noninvasive ventilation use is common in cardiac ICUs, especially in patients admitted for medical conditions, infants, and those undergoing high complexity surgery. We observed wide variation in noninvasive ventilation use across hospitals, though the primary driver of total respiratory support time seems to be duration of mechanical ventilation.