The purpose of this study was to determine the impact upon classification of patients of the choice of reference equation and the criterion defining the lower limit of the normal range in clinical practice. One thousand consecutive spirometries were checked to calculate the predicted values [forced vital capacity (FVC) and forced expiratory volume in sec (FEV1)] in accordance with the equations by Morris, Cherniack, Crapo, Knudson and the Economic Community for Coal and Steel (ECCS). We quantified the difference between the predicted values obtained for each individual and each equation, determined the percentage of individuals whose classification might have changed from normal to abnormal when using a different equation and defined the lower limit of the normal range in accordance with the determination of 1. the 90% confidence interval or 2. 80% of predicted, comparing their differences. The greatest differences found were between the values given by Morris and Crapo's equations for male FEV1, between Morris and Cherniak for female FEV1 and male FVC and between Morris and Knudson for female FVC. Using 80% of predicted value for female FEV1, up to 35% of tests changed their classification from 'normal' to 'abnormal' upon changing the equation used. A high percentage of tests showed a lower limit of normal defined by the confidence interval under 80% and 70% of predicted value. This study emphasizes the importance of choosing the appropriate reference equation. We do not consider it acceptable to use a fixed percentage of the predicted value as the lower limit of normal because of the great number of patients found to be inappropriately classified.