Little is known on the long-term validity of reference equations used in the calculation of FEV(1) and FEV(1)/FVC predicted values. This survey assessed the prevalence of chronic airflow obstruction in a population-based sample and how it is influenced by: (i) the definition of airflow obstruction; and (ii) equations used to calculate predicted values. Subjects aged 45 or more were recruited in health prevention centers, performed spirometry and fulfilled a standardized ECRHS-derived questionnaire. Previously diagnosed cases and risk factors were identified. Prevalence of airflow obstruction was calculated using: (i) ATS-GOLD definition (FEV(1)/FVC<0.70); and (ii) ERS definition (FEV(1)/FVC<lower limit of normal) with European Community for Coal and Steel (ECCS) reference equations and with predicted values derived from the presumably normal fraction of the studied population. A total of 5008 subjects (4764 adequate datasets) were studied. Prevalence of airflow obstruction was 8.71% with ATS-GOLD definition and 6.40% with ERS definition and ECCS predicted values. The ERS definition with predicted values derived from the studied population provided a 7.96% prevalence. Severity distribution of airflow obstruction was also influenced by the equation used to calculate predicted values of FEV(1). Prevalence and severity of chronic airflow obstruction are influenced not only by the definition used but also by equations used to calculate predicted FEV(1)/FVC and FEV(1) values. These equations likely need to be periodically revised.