Background and objective: Pulmonary function and height may be regarded as adult indices of exposures accumulated across the entire life course and in early life, respectively. As such, we hypothesised that pulmonary function would be more strongly related to mortality than height. Studies of the association of height and lung function with mortality--which are currently modest in number--will clarify the relative utility of these risk indices and the mechanisms underlying observed patterns of disease risk.
Design, setting, and participants: Data were drawn from the Whitehall study, a prospective cohort study of 18,403 middle-aged non-industrial London-based male government employees conducted in the late 1960s. Data were collected on stature, spirometry measures (including forced expiratory volume in one second [FEV1]) and a range of covariates. These analyses are based on the 3083 non-smoking men with complete data.
Main outcome measures: Mortality ascribed to all-causes, cardiovascular disease, respiratory disease and site-specific cancers.
Results: Both height and FEV1 were associated with a range of physiological, behavioural and socio-economic risk factors. Relations with these risk factors were seen more frequently for FEV1 and, where they occurred, were of somewhat higher magnitude. During a maximum of 35 years follow-up, half the non-smokers had died (n = 1545). FEV1 (HR(per one SD increase); 0.89; 0.84, 0.95) was somewhat more strongly related to total mortality than height (0.96; 0.91, 1.01) in a fully adjusted model, but this difference did not attain statistical significance at conventional levels (p-value for difference = 0.15). Of the eight independent disease-specific outcomes examined, the only convincing evidence of a differential effect was for deaths from respiratory causes which was unsurprisingly more strongly related to FEV1 than height (p-value for difference = 0.03).
Conclusions: In the present study, height and FEV1 were essentially similarly related to both risk factors and mortality outcomes, thus not providing support for our hypothesis. Both factors would appear to have some utility as markers of early life exposures.