Abstract 11322: Multi-Institutional Analysis of Noninvasive Ventilation Practice After Infant Cardiac Surgery

Background: Hospitals performing infant cardiac surgery commonly use noninvasive ventilation (NIV) for postoperative respiratory support. The epidemiology of NIV use in this population remains unknown and it is unclear how practice varies across centers. Methods: We analyzed all infant surgical encounters (<1 year) in the Pediatric Cardiac Critical Care Consortium (PC 4 ) registry from 10/2013--12/2015. NIV included high flow nasal cannula (HFNC) or positive airway pressure (PAP) support. We described patient, operative and postoperative characteristics of patients who did and did not received NIV. To examine casemix-adjusted duration of total respiratory support and the contribution of mechanical ventilation (MV) vs. NIV we created a model incorporating age, prematurity, extracardiac anomalies, surgical complexity and MV at preoperative admission. We calculated and compared adjusted mean duration of total support (MV + NIV), MV, and NIV across hospitals using observed-to-expected duration. Results: The cohort included 3708 encounters from 15 institutions: 2032 (55%) received NIV postoperatively, 40% were neonates, 36% were high complexity. Neonatal age group, extracardiac anomalies, single ventricle, procedure complexity, preoperative respiratory support, longer MV duration, and greater postoperative disease severity were all associated with NIV therapy vs. none (p<0.001 for all). Of those receiving NIV, 93% and 43% received HFNC and PAP, respectively. Across hospitals, NIV use ranged from 32%-65%, and adjusted mean duration ranged from 1-4 days (3 days observed mean). Differences in total adjusted respiratory support were largely driven by duration of MV (Figure 1). Conclusions: NIV therapy is common after infant cardiac surgery, though use varies across hospitals. The analysis suggests that initiatives to reduce differences across hospitals in postoperative respiratory support time should initially focus on MV practices, not NIV.