Computer simulation - a tool for optimisation of ventilator setting in critical lung disease

Abstract: Increasing attention is paid to mechanical ventilation as one cause behind aggravation of lung injury. Lung protective ventilation can be achieved e.g. by minimising tidal lung collapse and re-expansion and by the use of small tidal volume allowing some degree of permissive hypercapnia. The ventilator and the sick lung comprise a very complex system. Identification of a ventilator setting that with respect to respiratory rate, tidal volume, I:E ratio, PEEP, etc is optimal with respect to desired physiological effects is therefore difficult. The main objective behind this thesis was to develop and validate a system for computer simulation of respiratory mechanics and gas exchange that allows prediction of physiological effects of resetting the ventilator. To reach this objective, methods for characterisation of lung physiology, which require little or no disturbance of the breathing pattern, were developed. Complementary studies of how gas exchange is affected by ventilator setting were performed. Longer time for equilibration between inspired and alveolar gas decreased airway dead space. When, in acute lung injury, PaO2 increased in response to PEEP, alveolar dead space decreased and vice versa. The hypothesis that mechanical behaviour and CO2 elimination after resetting respiratory rate and tidal volume could be predicted by simulation in healthy pigs was confirmed. Further, the hypothesis that immediate effects of moderate PEEP increments on mechanics and CO2 elimination in patients with acute lung injury could be predicted by simulation was also confirmed. Future development of the lung model used for simulation was outlined so as to include how time for gas equilibration affects gas exchange and to account for non-linear elastic properties. Iterative simulation may be a tool in future goal-oriented ventilation strategies.