A physiologic Comparison of Continuous Neurally Adjusted Ventilation (NeuroPAP) Versus Neurally‐Adjusted Ventilatory Assist (NAVA) in Infants With Respiratory Failure

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Objectives And Hypothesis: Tonic diaphragmatic activity is common during non‐invasive ventilation (NIV), suggesting efforts to increase end‐expiratory lung volume. We assessed the feasibility and physiological impact of NeuroPAP, a novel NIV mode continuously adjusting the delivered positive pressure proportionally to diaphragm electric activity (Edi) during both inspiration and expiration, in infants with respiratory failure. We hypothesized NeuroPAP would enable dynamic control of end‐expiratory pressures (PEEP). Methodology: This prospective crossover study enrolled premature neonates (25–34 weeks) and infants with bronchiolitis supported by NIV‐NAVA. Subjects underwent three ventilation phases: NIV‐NAVA, NeuroPAP, and repeat NIV‐NAVA. Ventilation pressures, Edi, cardio‐respiratory events, neural breathing patterns, and systemic and cerebral oxygenation were assessed. Results: A total of 15 infants with bronchiolitis and 8 premature neonates were included. The overall median PEEP was unchanged between modes, but PEEP was actively adjusted, with increased breath‐to‐breath variability of PEEP in NeuroPAP ( p < 0.001). Compared to pre‐study settings, individual PEEP increased in NeuroPAP in 7, decreased in 9, and was unchanged in 7 patients. In NeuroPAP, the breathing pattern was phasic 74% of the time and tonic 16% of the time, compared to 61% ( p = 0.31) and 23% ( p = 0.36) in NIV‐NAVA. Respiratory rate was lower in NeuroPAP in the neonates ( p = 0.006). The estimated PaO 2 /FiO 2 ratio was higher in the post‐NeuroPAP NIV‐NAVA period in the bronchiolitis group ( p = 0.006). Edi, heart rate, cerebral NIRS, or cardio‐respiratory events were unchanged. Conclusion: In infants with respiratory failure, NeuroPAP allowed dynamic control and personalization of PEEP. The clinical impact of this warrants further evaluation.

This publication has 33 references indexed in Scilit: