Model-based identification of flow-limited and non-flow-limited COPD patients in plethysmographic data.

Abstract

Chronic Obstructive Pulmonary Disease (COPD) patients, unlike healthy subjects, develop an increase in resistance in expiration. Model-based methods have previously assumed resistance to change linearly with flow. This study explores nonlinear, dynamic expiratory resistance in COPD subjects with plethysmographic data of non-flow-limited (NFL) and flow-limited (FL) subjects. A nonlinear model that incorporates a time varying, dynamic resistance in expiration only is presented. Alveolar pressure and airflow measurements for subjects were separated into NFL (n = 25), and FL (n = 35). The median inspiratory resistance identified is 3 cmH2O for NFL and 5.9 cmH2O for FL, and the linear expiratory resistance is 4.2 (RMSE = 0.6) and 11.4 (RMSE = 1.7) cmH2O. The nonlinear expiratory mean resistance is 4.9 (RMSE = 0.1) for NFL, and 13.5 (RMSE = 0.2). This study presents a method for examining nonlinear expiratory resistance in COPD subjects. The FL subjects showed significant nonlinearity in expiration. The model fit for the nonlinear expiratory resistance was good. A novel method for examining nonlinear, dynamic, expiratory resistance in COPD patients using plethysmographic P-Q loops is presented with an identifiable model, capturing disease state. The model captures static and dynamic changes in resistance as model-based markers for COPD diagnosis and monitoring, with possible future clinical use.