The effects of atenolol and nifedipine on gas exchange were studied in 27 patients with effort angina in a randomized cross-over trial. Semi-supine bicycle exercise tests (ramp program, 20 W/min) with measurement of gas exchange were carried out after consecutive 2-week treatment periods with atenolol (50 mg b.i.d.) and slow-release nifedipine (20 mg b.i.d.). In the range of subthreshold exercise, the slope of the VO2 workload line was lower with atenolol than with nifedipine (8.64 +/- 1.59 vs 10.28 +/- 1.74 ml.min-1.W-1, p < 0.005) as determined by linear regression analysis excluding the initial 30 W. The intercept of the curve on the VO2 axis was higher with atenolol than with nifedipine (366 +/- 111 vs 299 +/- 113 ml.min-1, p < 0.05). VO2 was higher (p < 0.05) with nifedipine than with atenolol for workloads above 65 W. A similar pattern was seen if the drug effects on the slope of the VCO2-workload relation were analyzed (7.11 +/- 1.92 vs 8.54 +/- 1.85 ml.min-1.W-1, p < 0.02). The intercept on the VCO2 axis was not different among the treatments. VCO2 was higher (p < 0.05) with nifedipine than with atenolol for workloads above 55 W. Minute ventilation was higher (p < 0.05) with nifedipine than with atenolol at all points of the analysis. The data suggest that the ventilatory requirements and the energy cost for aerobic exercise are higher with nifedipine than with atenolol. This may become relevant in cardiovascular patients with concomitant pulmonary disease.(ABSTRACT TRUNCATED AT 250 WORDS)