In chronic air-flow obstruction (CAO), inspiratory mechanics constitute a potential mechanism of compensation for limitation of expiratory air flow. We sought to determine whether assessment of inspiratory function could improve our ability to predict ventilation at maximal exercise (Ve max) in patients with CAO. Resting inspiratory and expiratory pulmonary function studies from 20 patients with ventilatory limitation of exercise due to CAO provided data for development of a new regression model for Ve max. Maximal exercise was quantitated from breath-by-breath analysis of exercise responses at cycle ergometry with work increments of 25 watts each min to tolerance. Multiple regression analysis by 3 methods gave identical results. A 2-variable formula containing peak inspiratory flow rate (PIFR) and the forced expiratory volume in one second (FEV1) correlated strongly with Ve max (r = 0.967) in the formula Ve max (L/min) = 21.34 FEV1 (L) + 6.28 PIFR (L/s) +3.94 (95% Cl = +/- 18 L/min). This model was significantly different from published models containing FEV1 alone (p = 0.0002) and was not improved by additional variables. Similar formulas derived for emphysematous and bronchitic clinical types of CAO did not exhibit significantly different slope and intercept coefficients. Both PIFR and FEV1 correlated strongly with tidal volume at maximal exercise. The PIFR also correlated well with resting peak inspiratory airway pressure (r = 0.775). We conclude that consideration of PIFR in addition to FEV1 can improve our clinical ability to predict Ve max in patients with CAO.