A low proportion of dietary calories as carbohydrate has been suggested for patients with chronic obstructive pulmonary disease, because oxidation of carbohydrate (CHO) compared to fat results in greater CO₂ production (co₂) and, at the same arterial Pco₂ (Pa_CO2), higher alveolar and minute ventilation (e) and increased dyspnea. We hypothesized that a low CHO-high fat diet, although reducing co₂ and e at rest, might result in only a small change in co₂ and e during exercise. Eight healthy volunteers were randomized to receive for 24 h either isocaloric diets containing 10% or 70% of total calories from CHO (remainder of nonprotein calories from fat). Measurements of co₂, e, and respiratory gas exchange ratio (R) were made at rest and during constant work rate cycle exercise below the anaerobic threshold. Five to seven days later, the alternate diet was given and the studies were repeated. At rest, mean co₂ and R were significantly lower after the low CHO diet compared to the high CHO diet. Mean resting e was less but not significantly (high CHO 9.6 [0.7] versus low CHO 8.7 [0.8] L/min, mean [SEM]). During exercise, mean co₂ and R were significantly less after the low CHO diet, but mean e was only slightly smaller and not significantly different between diets (high CHO 25.4 [1.1] versus low CHO 24.0 [1.0] L/min). The increase in co₂ from rest to exercise was relatively independent of the substrate mix recently consumed, suggesting that the exercising muscles use stored muscle glycogen as substrate during short bouts of low-intensity exercise despite changes in substrate utilization by nonmuscle tissues at rest. From this study in normal subjects, we conclude that altering the proportion of dietary carbohydrate and fat is unlikely to reduce greatly exertional dyspnea in patients with lung disease.