We evaluated the sensitivity and specificity of different indices drawn from the forced expiratory curve in order to detect the best one for evaluating bronchial reactivity in the general population, and particularly, for distinguishing between normal and asthmatic subjects. 654 subjects, between 15 and 64 years of age, who were representative for age and sex of the general population of a small Lombardy town (Italy), were studied. Of the total sample, 448 subjects were clinically normal and asymptomatic, 87 were symptomatic or with acute upper respiratory disease within 30 days before the challenge with methacholine, 43 with allergic rhinitis, 26 asthmatics and 50 chronic bronchitics. All the subjects who had a baseline FEV1 greater than 85% predicted underwent methacholine challenge. Provocative LnDoses of a 6, 10, 15% fall in FEV1 (LnPD6, LnPD10, LnPD15) and provocative LnDose causing FEV1 to fall more than 2 SD below the mean of 5 FEV1 measurements following buffer inhalation (LnTD) were evaluated. Furthermore we calculated the "Dose-Response slope" (SL) proposed by O'Connor. The cut-off value of each index to define a responder was calculated by discriminant analysis of the response to the challenge in normal and in the asthmatic group. Sensitivity, specificity and predictive value of LnPD6, LnTD and SL were 79, 74, 14%; 71, 74, 13% and 46, 97, 48% respectively. LnPD10 had the best sensitivity (83%) with a high specificity (83%), but, as compared to LnPD15, whose sensitivity was 79% and specificity 89%, a lower predictive value (21% vs 27%).(ABSTRACT TRUNCATED AT 250 WORDS)