The temporal dynamics of ozone-induced FEV1 changes in humans: an exposure-response model

Although ozone is known to induce reversible decrements in forced expiratory volume in 1 s (FEV1), no exposure-response model has been identified that accurately describes the dynamics of response to the changing concentrations and activity patterns of normal ambient human exposure. The purpose of the current analysis was to identify and evaluate a dynamic model of FEV1 response using a large existing data set (541 volunteers, 864 exposures, 3485 FEV1 measures) with a wide range of exposure conditions (ozone = 0.0 to 0.4 ppm, activity level = rest to heavy exercise, duration = 1 to 7.6 h), including recovery in clean air. A previously described model containing a differential equation and a logistic function was modified to include a new between-subjects variance structure and was fitted to the data. The model described well the mean observed response data across the range of exposure conditions, including the periods of recovery in clean air. Predicted values of individual responses were distributed lognormally and appeared to accurately describe the distribution of observed responses. We observed that responsiveness to ozone decreased with age, that response was weakly related to body size, and that response was marginally more sensitive to changes in ozone concentration than to changes in minute ventilation. In summary, we have identified a dynamic ozone exposure-response model that accurately describes the temporal pattern of FEV1 response to a wide range of changing exposure conditions and that may have utility for predicting population responses to ambient exposures.