PATIENT VENTILATOR ASYNCHRONY IN CRITICALLY ILL CHILDREN - PROSPECTIVE OBSERVATIONAL STUDY

Patient ventilator asynchrony (PVA) is common but infrequently identified in
children. Pre-clinical studies show a clear causal relationship between certain
PVA subtypes and ventilator induced lung injury, diaphragm dysfunction, and
delirium. PVA may exacerbate ventilator induced lung injury (VILI),
particularly if PVA results in a double-cycled (DC) breath. DC breaths imply
that a second breath is delivered by the ventilator before the patient has
fully exhaled. Double cycled breaths (DC) are particularly injurious because
the resultant breath stacking causes unintentionally large tidal volumes and
trans-pulmonary pressures, which exacerbates regional overdistension of the
lung, and is associated with mortality in adults with ARDS.

Specific Aim 1) To identify the frequency and risk factors for PVA and its
subtypes in ventilated children, with a specific focus on DC breaths.
Hypothesis: PVA subtypes related to inadequate ventilator support (flow
undershoot and premature cycling) and reverse triggering will be the most
common causes of DC breaths, with risk factors related to respiratory drive and
ventilator settings.
Specific Aim 2) To develop and test a clinical decision support system using
machine learning techniques to automatically identify common forms of PVA in
children. Hypothesis: (a) Machine learning models using waveforms available on
all mechanical ventilators will accurately identify PVA subtypes compared to
gold standard annotations which include measures of neural drive. (b)
Algorithms can be optimized for high sensitivity and low false alert rates to
identify children with frequent PVA.

Multi-center prospective observational cohort.

Minimal. For this study, no deviation from standard care will occur. Data will
be collected from the mechanical ventilator. sEMG electrodes will be used to
quantify the electrical activity of respiratory muscles.