Basis Function Modelling of Respiratory Patients with High or Low Auto-PEEP

摘要

Pulmonary mechanics of critically ill patients with Chronic Obstructive Pulmonary Disease (COPD) are affected by bronchial pathway blockages, leading to insufficient gas exchange and poor outcomes. The blockages can result in incomplete expiration, and consequently many COPD patients have a higher than normal auto-PEEP. While mechanical ventilation (MV) is an important therapy for these patients, there is no consensus on the optimal MV settings. The nonlinear autoregressive (NARX) model of pulmonary mechanics was adapted to capture changes noted to occur in high auto-PEEP patients when pressure is increased. Specifically, resistance basis functions were added to the model to capture linear resistance changes with pressure. The result was a statistically significant increase in resistance at low pressure for high auto-PEEP patients, compared to low auto-PEEP patients (p < 0.01). Additionally, the pressure dependent elastance curve of the high auto-PEEP group indicated that there was a greater proportion of un-recruited pulmonary tissue at low pressure compared to the low auto-PEEP group. The high auto-PEEP group also tended to have a lower rate of pulmonary distension at higher pressures. The modified NARX model was able to identify physiological differences between the two groups, and the outcomes suggest that high auto-PEEP patients benefit from higher positive end expiratory pressure (PEEP) than low auto-PEEP patients.