Arteriovenous O-2 content (a-vCO(2)) differences increase during exercise in normal subjects through several mechanisms including PO2, O-2 pressure at which hemoglobin (Hb) is half saturated with O-2 (P-50), and Hb concentration changes. The present study was undertaken to evaluate how much these biochemical changes are relevant to a-vCO(2) difference through exercise in patients with heart failure. Twenty-seven patients with congestive heart failure [10 patients in functional class A (peak exercise O-2 uptake >20 ml . kg(-1). min(-1)), 9 in class B (20-15 ml . kg(-1). min(-1)), and 8 in class C (15-10 ml . kg(-1). min(-1))] underwent a cardiopulmonary exercise test with once-per-minute simultaneous blood sampling from the pulmonary and systemic arteries for determination of Hb, PO2, PCO2, pH, O-2 content (CO2), Hb saturation and lactic acid (pulmonary artery only), and calculation of P-50. Analysis of data was done at six exercise stages: the first at rest, the last at peak exercise, and the second to the fifth at one-, two-, three-, and four-fifths of O-2 consumption increase. a-vCO(2) difference at peak. exercise was 14.3 +/- 2.1, 16.9 +/- 2.4, and 14.7 +/- 2.1 (SD) ml/dl in class A, B, and C patients, respectively. The contribution of Hb, P-50, and PO2 changes to the increments of a-vCO(2) difference during exercise was 21, 17, and 63%, respectively; the only interclass difference observed was for P-50, which plays a greater role in a-vCO(2) difference in class A. Hb changes act mainly at the arterial site, whereas P-50 and PO2 act at the venous site. Hb increase was constant through the test, venous P-50 increase was greater above anaerobic threshold, and venous PO2 reduction was most remarkable at the onset of exercise; in class C patients, no venous PO2 change was recorded in the second half of exercise. Thus a-vCO(2) difference increase during exercise is notable in patients with heart failure but unrelated to the severity of the syndrome. Hb, P-50, and, to the greatest degree, PO2 changes participate in the increment of a-vCO(2) difference. In class C patients, the lack of PO2 reduction in the second half of exercise suggests the achievement of a ''whole body critical venous PO2.''