Numerous population studies have reported that pulmonary function following the adolescent growth phase appears to be in a steady-state, where there is little or no growth occurring up to 40 yrs of age. We examined longitudinal forced expiratory volume in one second (FEV1), changes using three different statistical approaches to determine which subjects actually have significant trends during this period. Participants, who were employees at a metal processing plant, underwent quarterly spirometry for up to 10 yrs. Test results up to 33 yrs of age were included in the analysis. Each subject's FEV1 data was first analysed using simple linear regression (SLR) to test for a statistically significant linear slope. Next, each subject's data were fitted using bootstrap sampling (BSS) of their original data, to yield reduced estimates of the slope variances and increase the power of detecting a significant trend. And thirdly, we fitted a regression breakpoint (BKPT) model to the data to find those subjects who may have piecewise linear growth or decline in function. All analyses were stratified, based on smoking status. Subjects included 111 nonsmokers and 110 smokers. Among the nonsmokers, 34 subjects had significant slopes using SLR, an additional three using BSS, and only two with BKPT. Among the smokers, 36 had a significant trend using SLR, 7 were added using BSS, and no additional subjects with BKPT. We conclude that in young adult males lung function is not in a steady-state and that as many as 40% have a significant slope, either positive or negative.(ABSTRACT TRUNCATED AT 250 WORDS)