BACKGROUND AND OBJECTIVE: Most published reference values for lung function test (LFT) parameters introduce systematic bias. The aim of this study was to compare measured values of FEV(1) and FVC with the corresponding normal predicted values in a Greek population, and to produce reference equations for LFT parameters in this population.METHODS: In a cross-sectional study conducted in Macedonia, Greece, 1080 adult healthy, non-smokers (432 men, 648 women, aged 18-80 years), underwent spirometry. Measured values of FVC and FEV(1) were compared with predicted values determined using three existing sets of reference equations: one recently derived from a European population and two others widely used in Europe (European Coal and Steel Community; ECSC) and the USA (National Health and Nutrition Examination Survey; NHANES III). Height and age were entered into the multivariate regression analysis to produce reference equations for LFT parameters.RESULTS: All three published sets of equations underpredicted FEV(1) in men. FVC was accurately predicted by all equations except NHANES III. The discrepancy was even greater among women; the ECSC equation underpredicted both FEV(1) and FVC, the NHANES III equation overpredicted both FEV(1) and FVC, while the third set of equations accurately predicted FEV(1) but overpredicted FVC. The derived reference equation for FEV(1) in men was -0.28 × age + 0.057 × height - 4.91, and in women -0.021 × age + 0.039 × height - 2.58. The derived reference equation for FVC in men was -0.28 × age + 0.071 × height - 6.763, and in women -0.019 × age + 0.056 × height - 5.018.CONCLUSIONS: Measured FEV(1) and FVC values in a Greek population differed significantly from those predicted using previously published reference equations. The new locally derived spirometry reference equations may be more suitable for evaluation of lung function in everyday practice.