Introduction: High-frequency oscillatory ventilation (HFOV) is routinely used in the care of neonates and small children who fail ventilatory management by conventional measures. Despite more than 10 years of use in pediatrics, clinical challenges exist in transitioning patients between conventional ventilatory (CV) modes and HFOV. In particular, the relationship of mean airway pressure (MAP) and lung volume in both CV states and HFOV requires definition, an issue made even more pertinent following the recent government approval of an oscillatory ventilator for patients greater than 35 kg. We propose to study the static and dynamic effects of HFOV on regional air distribution, MAP, and regional lung recruitment in healthy and injured lungs in a canine model of acute lung injury (ALI), and to correlate these values with CV settings under the same conditions.

Background: HFOV provides a means of ventilating critically ill patients in a manner believed to minimize ventilation associated lung injury by recruiting the lung while minimizing the extremes of over-distention and airway collapse and re-opening. The relationship of MAP and mean lung volume during CV of injured lungs is poorly understood. Often the clinical goal is not matching lung volume or MAP when transitioning between HFOV and CV, but instead oxygenation optimization while minimizing over-distention on plain radiographs. However, severity of lung disease and difficulties in ventilation and/or oxygenation often mandate high MAP settings during HFOV. Therefore, comparing CV and HFOV effect on MAP, mean lung volume and regional lung distribution may improve our understanding how HFOV limits excessive volume (volutrauma) and/or pressure (barotrauma) while improving oxygenation.

In addition, there are many dynamic factors involved in air delivery during HFOV. The three primary issues are inertia (flow tends to go in a straight line), asymmetry and non-linearity of inspiratory and expiratory impedances (the extreme results in expiratory flow limitation, outflow obstruction and air trapping) and heterogeneity of regional mechanics (resistance and compliance). The high instantaneous airflow rate during HFOV interacting with these factors, may result in a change in the distribution of ventilation and mean lung volume. In disease states such as ALI, regional changes in lung mechanics combined with alteration in compliance and expiratory impedance can affect the dynamic airflow and lung volume distribution during HFOV. The correlation between MAP, potential alterations in lung volume and dynamic air distribution during oscillatory and traditional ventilation can now be quantitatively measured in the injured and un-injured lung by computed tomagraphic (CT) imaging.

Advances over the past decade in CT imaging have produced scanners that can image the entire lung at high resolution in 10 to 15 seconds. This new technology enables CT imaging of the entire lung, and by applying accepted models for partitioning volume into air and tissue compartments, permits calculation of lung air volume as well as the distribution of regional aeration. As we have demonstrated in prior studies, such analysis can be performed before and after lung injury and identical regions of lung can be correlated by reproducible anatomic markings such as bronchi and blood vessels. Also, we have successfully imaged the entire lung during steady state HFOV at frequencies of 5 to 15 Hz and have found that the aeration patterns are insignificantly affected by motion artifact, thus permitting quantitative analysis. We will use these advanced imaging techniques and novel image analysis to non-invasively determine the distribution of lung volume in the injured and un-injured lung during HFOV and CV.

Hypothesis: Mean lung volume is increased during HFOV when compared to static lung volume at the same MAP (measured at the carina). We postulate this increase in aeration will be greater towards the base of the un-injured lung during HFOV when compared to continuous positive airway pressures (CPAP), and this difference will be increased with increases in frequency ( f ), increases in tidal volume (Vt), but will decrease with increases in MAP. In the setting of ALI, this dynamic distribution of lung volume will become more pronounced and even greater regional variation will be demonstrated. Improved recruitment during HFOV will be confirmed not only by increases in regional lung aeration in injured and un-injured lungs, but by improvement in measured blood gas values of PaO2.

Specific Aims: The specific aims of our study will be addressed by comparing oscillatory and traditional modes of ventilation through measurements of lung air volume distribution made with high-resolution CT imaging in injured and un-injured lungs. The aims are as follows:

1) Determine the relationship between MAP and mean lung volume during HFOV and CV. The mean lung volume will be determined by volumetric CT imaging over a range of CPAP settings equal in value to MAP during HFOV. Differences, if any, will be demonstrated by changes in mean lung volume and dynamic distribution from apex to base.

2) Determine the effect alterations of f, Vt, and MAP has on mean lung volume and regional air distribution during HFOV. The distribution of mean lung volume will change with alterations in f, Vt , and MAP during HFOV. Volumetric CT imaging will be performed during steady state HFOV and following systematic alterations in MAP, Vt and f (covering settings above and below eucapnia). Thereby, regional lung volume distribution will be determined and comparisons will be made to CPAP. Data will be analyzed in terms of the distribution of air content from apex to base, as well as, the distribution from dependent to non-dependent regions (vertical gradient).

3) Quantify and compare differences in regional distribution of air and mean lung volume in the injured lung during HFOV and CV. ALI changes global and regional lung mechanics and thus, potentially, the static and dynamic distribution of air volume during HFOV. To better understand the impact of ALI on these phenomena, we will repeat the studies described in Aims 1 and 2 in the same set of animals (each one week later) post-injury by saline lung lavage. This approach will allow us to compare and contrast pre- and post-injury volumetric CT imaging to detect any patterns and/or trends that may exist, as dynamic air volume distribution will most likely be altered. In addition, comparisons of recruitment behaviors between HFOV and CV and their effect on PaO2 will be made.