We hypothesized that maximal bronchoconstriction can be predicted from the bronchomoter effect of deep inhalation (DI) and the degree of airway sensitivity to methacholine (MCh). We studied 26 healthy or mildly asthmatic subjects with limited response to MCh (maximal FEV1 decrease, 23 +/- 9 SD%; Group 1) and 26 subjects with moderate to severe asthma with exaggerated response (maximal FEV1 decrease > 40%, Group 2). The effect of DI was quantified as the linear regression coefficient of the percent decrements of maximal (Vm50) versus partial (Vp50) forced expiratory flow at 50% of FVC over the initial steps of challenge (MP slope). Airway sensitivity was inferred from the MCh doses (PDs) causing Vm50 or Vp50 to decrease by 40% or FEV1 by 15%. The absence of limit to bronchonstriction was predicted by either MP slope or any PD with accuracies between 71 and 81%, but with an accuracy of 87% by a discriminant function including MP slope and PD40Vp50. Within Group 1, the maximal FEV1 decrease correlated linearly with MP slope (r2 = 0.41); but it was better predicted by a multiple regression including MP slope and PD40Vp50 (In mg) (r2 = 0.54). We conclude that the magnitude of the bronchodilator effect of DI during induced bronchoconstriction and airway sensitivity predict the level of maximal bronchoconstriction in vivo. We speculate that these parameters reflect some of the mechanisms modulating the response to bronchoconstrictor stimuli such as airway wall structure, airway-to-parenchymal interdependence, and contractile properties of airway smooth muscle.