The aim of the study was to estimate the relative importance of the Bohr effect and redistribution of blood from the non-exercising tissues on the arterial-venous oxygen content differences across the exercising extremities and the central circulation in patients with chronic heart failure; the relationship among femoral vein, systemic and pulmonary artery oxygen partial pressure and hemoglobin saturation was determined. It has been reported that the maximal reduction in femoral vein pO2 precedes peak oxygen consumption and lactic acidosis threshold in patients with chronic heart failure and normal subjects during exercise. The increase in oxygen consumption at work rates above lactic acidosis threshold, therefore, must be accounted for by increase in blood flow in the exercising muscles and right-ward shift on the oxyhemoglobin dissociation curve. Since the total cardiac output increase is blunted in patients with chronic heart failure, diversion of blood flow from non-exercising to exercising tissues may account for some of the increase in muscle blood flow. Ten patients with chronic heart failure performed a progressively increasing leg cycle ergometer exercise test up to maximal effort while measuring ventilation and gas concentration for computation of oxygen uptake and carbon dioxide production, breath-by-breath. Blood samples were obtained, simultaneously, from systemic and pulmonary arteries and femoral vein at rest and every minute during exercise to peak oxygen consumption. At comparable levels of exercise, femoral vein pO2, hemoglobin saturation and oxygen content were lower than in the pulmonary artery. PCO2 and lactate concentration increased steeply in femoral vein and pulmonary artery blood above lactic acidosis threshold (due to lactic acid build-up and buffering), but more steeply in femoral vein blood. These increases allowed femoral vein oxyhemoglobin to dissociate without a further decrease in femoral vein pO2 (Bohr effect). The lowest femoral vein pO2 (16.6 +/- 3.9 mmHg) was measured at 66 +/- 22% of peak VO2 and before the lowest oxyhemoglobin saturation was reached. Artero-venous oxygen content difference was higher in the femoral vein than in the pulmonary artery; this difference became progressively smaller as oxygen consumption increased. "Ideal" oxygen consumption for a given cardiac output (oxygen consumption expected if all body tissues had maximized oxygen extraction) was always higher than the measured oxygen consumption; however the difference between the two was lost at peak exercise. This difference positively correlated with peak oxygen consumption and cardiac output increments at submaximal but not at maximal exercise. In conclusion, femoral vein pO2 reached its lowest value at a level of exercise at or below the lactic acidosis threshold. Further extraction of oxygen above the lactic acidosis threshold was accounted for by a right shift of the oxyhemoglobin dissociation curve. The positive correlation between increments of cardiac output vs "ideal" and measured oxygen consumption suggests a redistribution of blood flow from non-exercising to exercising regions of the body. Furthermore the positive correlation between exercise capacity and the difference between "ideal" and measured oxygen consumption suggests that patients with the poorer function have the greater capability to optimize blood flow redistribution during exercise.