A method to estimate pulmonary diffusing capacity for O(2) (D(LO2)) during exercise based on routine O(2) and CO(2) transport variables is presented. It is based on the fitting of a mathematical model to gas exchange data. The model includes heterogeneity (described as two exchanging compartments), diffusion limitation and right-to-left shunt. Mass conservation equations and Bohr integration were solved to calculate partial pressures in each compartment. Diffusion was distributed with perfusion. Two-compartment ventilation and perfusion distributions were estimated at rest during conditions of negligible diffusion limitation. These distributions were used during hypoxic and normoxic exercise to obtain the D(LO2) from the model computations (D(LO2)2C) compatible with experimental data. Three normals, four sarcoid patients and four patients after lung resection were studied. An independent technique for carbon monoxide was used to provide experimental estimates of DLo2 (D(LO2)EXP, rebreathing technique for sarcoid patients and single breath for lung resection). D(LO2)2C was highly correlated with D(LO2)EXP (r2 = 0.95, P<0.001) and the slope of the regression line was not statistically different from 1. The mean (D(LO2)EXP - D(LO2)2C) difference was -1.0 +/- 7.4 ml min-1 mmHg-1. The results suggest that use of a refined analytical procedure allows for assessment of D(LO2) from routine O(2) and CO(2) measurements comparable with those obtained from independent carbon monoxide techniques. The method may be an alternative for estimates of D(LO2) during exercise.