Lung function is a strong predictor of overall mortality in asthma and chronic obstructive pulmonary disease (COPD). FEV1 is considered to be the "gold standard," whereas peak expiratory flow (PEF) is mostly used in absence of FEV1 measurements. We compared the predictive power of PEF and FEV1, measured after maximal bronchodilation, which included a short course of oral corticosteroids. The study population comprised 491 asthmatics and 1,095 subjects with COPD. Pulmonary function tests were performed between 1983 and 1988, and survival data were obtained by September 1997, when 127 asthmatics and 723 subjects with COPD had died. Predictors of survival were examined by Cox proportional hazards analyses. After controlling for age, smoking, sex, and body mass index, we found best PEF to be at least equal to best FEV1 as predictor of overall mortality in subjects with COPD. The predictive power of best PEF was in part maintained after controlling for best FEV1. In asthma, best FEV1 seemed to be a better predictor of mortality than best PEF. Despite close correlation to FEV1, PEF apparently provides independent prognostic information in patients with COPD. This may be due to PEF and FEV1 reflecting different components of COPD, i.e., chronic bronchitis, small airways disease, and emphysema. Furthermore, extrapulmonary components such as muscle mass and general "vigour" probably affect PEF to a greater extent than they affect FEV1.