Methods development for epidemiologic investigations of the health effects of prolonged ozone exposure. Part I: Variability of pulmonary function measures.

The acute and subacute effects of ambient concentrations of ozone on lung function have been studied extensively in a variety of settings. Such studies generally have focused on measures of function that reflect either lung volumes or flows that are influenced by the physiology of large and small airways (e.g., forced expiratory volume in one second [FEV1). Data from animal studies suggest that the effects of prolonged exposure to elevated ambient concentrations of ozone result in abnormalities in the centriacinar region of the lung; and dosimetry models for humans predict that long-term exposure to ozone could impact the same areas of the human lung. However, alterations in structure at this level of the lung are not well reflected by measuring FEV1 until substantial structural changes have occurred. Measures of the lung function that reflect the functional mechanics of airways smaller than 2 mm in diameter are considered to be more relevant. At least one epidemiologic study has provided evidence that small-airway functions may be relevant to effects of prolonged exposure to environments with high concentrations of oxidants. A considerable body of physiologic data has established that flow rates measured during the terminal portion of a maximum expiratory flow-volume (MEFV) curve are largely governed by airways smaller than 2 mm in diameter A similar interpretation has been given to changes in the slope of phase III (delta N2) of the single-breath nitrogen washout (SBNW) curve. Despite the attractiveness of these measures in relation to airway physiology, some data suggest that measurements of flow via the terminal portions of MEFV and SBNW curves have much greater within-subject variability than forced vital capacity (FVC and FEV1. The present study was undertaken as part of a larger feasibility study to develop methods to study the effects of prolonged exposure to elevated ambient ozone levels on lung function in adolescents. A convenience sample of 239 freshmen (ages 16-20 years) entering the University of California, Berkeley were recruited to participate in this protocol. All were lifelong residents of the San Francisco Bay Area or the Los Angeles Basin. Subjects were studied on two occasions five to seven days apart. At each test session, subjects performed up to eight forced expiratory maneuvers to produce three acceptable and reproducible MEFV curves by modified American Thoracic Society criteria. Tests of SBNW were then performed on the basis of detailed criteria for validity and reproducibility. Eight attempts to generate three curves were allowed. The delta N2 was obtained by a least-squares regression of nitrogen concentrations between the 750-mL and 1750-mL volume points. Instantaneous flow at 75% of expired volume (FEF75%), average flow between the 25% and 75% volume points (FEF25%-75%), and delta N2 were the principal outcomes. Variance components were estimated with a nested random effects model with adjustments for important covariates. The average within-subject coefficients of variation (+/-SD of distribution of means) for male subjects were: FEV1 1.2 (+/-0.8); FEF25%-75% 3.2 (+/-2.3); FEF75% 5.8 (+/-5.0); and delta N2 17.9 (+/-12.3); for female subjects they were: FEV1 1.4 (+/-0.9); FEF25%-75% 3.0 (+/-2.2); FEF75% 6.2 (+/-5.2); and delta N2 19.9 (+/-17.0). The variance attributed to test session was less than 1% for all measures. The percentages of variance due to within-subject variation for each measure (adjusted for sex, area of residence, ethnicity, and height) were: FVC 3.6%; FEV1 3.0%; FEF25%-75% 5.2%; FEF75% 8.9%; and delta N2 23.9%. Of all subjects tested, 234 (97.9%) could provide at least two acceptable MEFV curves, but only 218 (91.2%) could provide at least two acceptable SBNW curves. The results were unchanged by recent history of acute, respiratory illness.(ABSTRACT TRUNCATED)