A dynamic systems approach to the development and application of new mechanical ventilator technologies.

摘要

Acute respiratory distress syndrome (ARDS) involves heterogeneous lung flooding/collapse often resulting in the need for mechanical ventilation (MV). New modes of MV have been proposed to improve gas exchange. One mode is an enhanced ventilation waveform (EVW) that can track the impact of ventilation setting on heterogeneity of mechanical function. The other mode is variable ventilation (VV) where the volume delivered from breath to breath is varied stochastically. In principle, both can be used to reduce the likelihood of ventilator induced lung injury. The goal of this project was to advance the application of these ventilation modes in a large animal model of ARDS. We applied VV in a sheep saline-lavage lung injury model and showed that VV (n = 7) improved gas exchange, ventilation pressures, mechanical heterogeneity, and resulted in less lung injury. We also examined the time course of recruitment occurring during VV in saline-lavaged excised calf lungs (n = 8) and found that intermittent recruitment is at least partially responsible for improvements seen during VV. In a separate experiment utilizing the same animal model (n = 5), we applied the EVW to examine dynamic lung resistance and elastance (R and E) at various positive end expiratory (PEEP) levels while also obtaining CT scans and gas exchange measures. CT scans revealed an optimal range of PEEPs at which alveolar recruitment was maximized without significant overdistension. This range also corresponded to PEEP levels, which maximized oxygenation and minimized ventilation pressures and indices reflective of the mechanical heterogeneity (e.g., frequency dependence of R and E). We also mapped regional elastance distributions into computational lung models to elucidate PEEP effects on ventilation distribution. Modeling results showed that PEEP also modulates the heterogeneity of ventilation distribution in a way that corresponds with experimental data, particularly the degree and location of alveolar overdistension. We conclude that dynamic mechanics can be used to minimize disease heterogeneity and maximize gas exchange through mechanisms related to the heterogeneity of ventilation distribution. We anticipate that monitoring dynamic R and E during VV will result in maximizing the efficacy of MV while minimizing the occurrence of mechanisms that cause lung injury.