Impact of NAVA Ventilation on Brain Oxygenation and Perfusion in Children With Congenital Heart Disease

1. Scientific context In post-operative cardiac surgery, invasive ventilation is often necessary, but the induced positive intra thoracic pressures can be detrimental to hemodynamics via several mechanisms: decrease in venous return, increase in pulmonary vascular resistance and increase in the afterload of the right ventricle. This effect is particularly problematic in patients with a restrictive right ventricle, who are preload dependent, and in patients with cavo pulmonary connections for single ventricle palliation. In these patients, decreasing ventilator positive pressure improves cardiac output and cerebral oxygenation. However, very early extubation is not always possible or safe, and it is therefore essential to optimize mechanical ventilation in order to minimize its hemodynamic consequences. In adults, ventilation in NAVA (Neurally Adjusted Ventilatory Assist) mode, regardless of the level of assistance, allows the preservation of intra thoracic pressure variations characteristic of spontaneous ventilation and limits the impact of ventilation on right ventricular ejection. In children, inspiratory and mean pressures are also lower in NAVA compared to conventional ventilation. Assessment of brain oxygenation and perfusion in cardiovascular resuscitation is important both as a hemodynamic parameter and as a neurological prognostic factor. An innovative non-invasive tool, Diffuse Correlation Spectroscopy coupled with Near Infrared Spectroscopy (DCS-NIRS), allows the non-invasive evaluation of both cerebral oxygenation and variations in cerebral blood flow.
2. Hypothesis and objectives The hypothesis is that the NAVA ventilation mode, which generates lower intra thoracic pressures than conventional ventilation modes, will improve cerebral hemodynamics, cardiac output, and oxygen transport of at-risk patients in the post-operative cardiac surgery setting compared with conventional ventilation. Main objective: In pediatric patients following cardiac surgery with risk of right ventricular diastolic dysfunction or passive pulmonary venous return (including Glenn, hemi-Fontan, Fontan, and Tetralogy of Fallot surgery), investigators will evaluate the impact of NAVA ventilation on:

• cerebral perfusion: cerebral blood flow (mm2/s) measured by DCS.
• cerebral oxygenation: cerebral tissue saturation (%) and regional cerebral O2 extraction (OEF, measured by NIRS-DCS). Secondary objective: In pediatric post-operative cardiac surgery patients with risk of restrictive right ventricle or passive pulmonary venous return (including Glenn, hemi-Fontan, Fontan, and Tetralogy of Fallot type surgery), investigators will compare the impact of conventional ventilation to NAVA mode on:
• Brain regional O2 consumption extracted from NIRS-DCS measurements
• Systemic hemodynamics, including cardiac index, lactate level, central venous oxygen saturation (ScvO2), and oxygen transport
• Ventilation parameters
• Comfort as measured by the COMFORT behavioural scale

1. Methods

This is a prospective cross-over, single-center, physiological study to be conducted in the Pediatric Intensive Care Unit of CHU Sainte Justine (Montreal, Canada).

Two periods of 30 to 60 minutes in each of 2 ventilation modes will be compared: conventional ventilation mode (as prescribed by the treating team) and NAVA mode.

On admission to the PICU postoperatively, study patients will be fitted with a naso-gastric NAVA probe.

Before performing any study measurements, investigators will wait for patient stabilization, which is defined by:

• absence of hemodynamically significant bleeding
• stabilization of the ventilation parameters
• stable inotrope dose

Study period ventilation will be set as follows:

• conventional ventilation: ventilation as prescribed by the treating team will be continued. Usually, ventilation is in PRVC (pressure regulated volume control) mode with a tidal volume between 5 and 8 mL/kg, a peep between 4 and 5, a respiratory rate to aim for normocapnia, and FiO2 to aim for PaO2 between 100 and 150 mmHg.
• NAVA Mode: NAVA level setting to target an Edi level between 3 and 20 V, tidal volume between 5 and 7 mL/kg, and no respiratory distress on physical examination. The trigger will be set at 0.5 V. The maximum pressure will be limited to 30 cmH2O. The PEEP level will not be changed from that prescribed by the treating physician.

To facilitate comparability of brain flow and oxygenation measurements, FiO2, NO concentration, inotrope dose, and sedation will be kept as stable as possible during these periods, at the discretion of the treating physician. Similarly, if possible, blood transfusions will not be administered during the study period.

During the two periods of ventilation, the following parameters will be recorded continuously:

• ventilatory parameters: ventilator parameters according to the set mode: peak inspiratory pressure, mean airway pressure, PEEP, tidal volume, total and spontaneous minute ventilation, Edi level, FiO2, exhaled CO2 via volumetric capnography.
• hemodynamic parameters: heart rate (HR), mean arterial pressure, systolic, diastolic, central venous pressure (CVP), left atrial pressure (LAP), type and dose of vasoactive medications received with calculation of the vasoactive-inotropic score.

In the second half of each period, investigators will perform:

• a central arterial and venous blood gas and an arterial lactate measurement.
• a measurement of cerebral blood flow by diffuse correlation spectroscopy and a measurement of cerebral regional tissue saturation and oxygen extraction. These measurements are done non-invasively and painlessly by applying a probe to the child's forehead for approximately 20 minutes.
• a limited cardiac ultrasound for the calculation of cardiac output (CO).
• calculation of urine output during the ventilation period.

In addition, for each patient, the following information will be collected:

• age at the time of surgery
• cardiac diagnosis
• existence and quantification of aorto-pulmonary collaterals
• preoperative medications
• operative details: surgical procedure, duration of cardiopulmonary bypass, aortic cross-clamp duration, inotropes administered (and maximum dose).
• result of the postoperative trans-esophageal cardiac echo (cardiac function, evaluation of repair, presence of residual anatomic lesions)
• arterial and venous blood gases, lactate and haemoglobin levels prior to study ventilation periods
• time interval between the end of cardiopulmonary bypass and the start of study measurements

The variables will be expressed in terms of mean standard deviation or median (interquartile), depending on the nature of their distribution. Because of the sample size, analysis of the differences between t variables during the 2 modes of ventilation will be done using the non-parametric Wilcoxon test (all the variables studied are of the continuous type). A value of p <0.05 will be considered significant.

For the calculation of sample size, investigators estimated, using the study by Huang et al, that the standard deviation of brain saturation would be 6%, and that the expected difference between the 2 ventilation modes would be at least 6%. To reach a power of 90% with an alpha risk set at 0.05, a minimum of 24 patients is required. Given the risk of loss of sight (e.g. technical difficulty in obtaining the main variable) and due to the heterogeneity of the patient population, we decided to aim to include 30 patients.

Note that no power calculation based on the variations in cerebral blood flow measured in spectroscopy was performed, because there was no preliminary data on the subject.

Approval from the Research Ethics Committee of Sainte Justine University Hospital and the University Medical Affairs Department was obtained. Written consent will be required from the parents.