Density dependence of maximal expiratory air flow (DD) has been used in adults as a test of early obstructive airway disease (OAD). Whether DD is useful as an epidemiologic tool to identify childhood risk factors for OAD is not known. In a population-based sample of 133 children 8 to 23 yr of age, we calculated density dependence at 50 and 25% of vital capacity (DD50 and DD25) (the ratios between maximal expiratory flow rates breathing helium-oxygen and air gas mixtures at each of these lung volumes), and the volume of isoflow (VisoV) (the lung volume, expressed as a percentage of vital capacity, at which maximal flow rates when breathing each gas mixture are equal), measured airway responsiveness using eucapnic hyperventilation with cold air, and obtained health and household information with questionnaires. Mean levels (+/- SD) of DD50, DD25, and VisoV were: 1.49 +/- 0.14, 1.37 +/- 0.18, and 10.7 +/- 10%. The DD50 significantly increased with age in these growing children (p less than 0.05), but DD50 was found to be significantly lower (1.42 +/- 0.14 versus 1.52 +/- 0.13; p less than 0.01) among children with nonspecific bronchial hyperresponsiveness. The DD50 also was significantly reduced among children with a history of a recent upper respiratory tract illness (URI) (p less than 0.01). There were no significant associations of DD with history of asthma, personal smoking, parental smoking, or respiratory illness during infancy. The reproducibility of DD50 was assessed on a subsample of 90 subjects in whom DD was measured during 2 surveys 1 yr apart.(ABSTRACT TRUNCATED AT 250 WORDS)