The Oxygen Uptake Efficiency Slope (OUES), a new parameter derived from respiratory gas analysis, has been suggested as a submaximal index of cardiopulmonary functional reserve. We evaluated the clinical application and the effect of physical training on the OUES in patients with coronary artery disease (CAD). Maximal cycle-ergometer testing with respiratory gas analysis (breath-by-breath) was performed in 590 patients with CAD and again after three months of physical training in 425 patients. OUES was determined from the linear relation of oxygen uptake (VO2) vs. the logarithm of pulmonary ventilation (VE) during exercise, i.e. VO2 = a log(10) V-E + b, where a is the OUES. The ventilatory anaerobic threshold (VAT) and the slope of the relation of VE V carbon dioxide production (VCO2) (V-E-VCO2 slope) were also determined. Correlation coefficients of the relation from which OUES was derived in individuals averaged 0.975 +/- 0.024 (mean SD) when calculated from data up to a respiratory gas exchange ratio of 1.0. Submaximal OUES was marginally lower (5.4 +/- 7.9%, p < 0.05) than the OUES calculated from 100% of respiratory exercise data. Of all submaximal parameters, submaximal OUES (r = 0.837, p < 0.001) and VAT (r = 0.860, p < 0.001) correlated best with peak VO2, followed by V-E-VCO2 slope (r = -0.469, p < 0.001). OUES was lower in patients who underwent coronary artery bypass grafting as compared with patients after coronary angioplasty (p < 0.05). Peak VO2 and OUES increased significantly (p < 0.001) after training with 24 +/- 19.2 % and 20.9 +/- 19.3 %, respectively. Changes in peak V02 correlated better with changes in OUES and in VAT (r = 0.61 and r = 0.55, p < 0.001, respectively) than with changes in VE-VC02 slope (r = -0.171, p < 0.001). The submaximal OUES is clinically useful for the quantification of exercise performance and is sensitive to physical training in patients with CAD.