Measurements of oxygen deficit during submaximal exercise were correlated with the anaerobic threshold (as measured by gas exchange analysis), peak work rate on a ramp protocol, and the ability to perform constant work rate exercise in 10 male patients with New York Heart Association class 2 congestive heart failure and 12 age- and gender-matched normal controls. All subjects performed a maximal ramp exercise test for measurement of the anaerobic threshold. In addition, several 15-min constant work rate exercise sessions were conducted to evaluate oxygen deficit, measured as the area between the "ideal" square curve of oxygen consumption at the onset of constant work rate exercise and the actual exponentially shaped curve. Since the oxygen deficit significantly correlated with the plateau oxygen consumption during the 25-W constant work rate exercise (r = 0.61, p = 0.002), the oxygen deficit was normalized by the rectangular area of 15-min oxygen consumption above baseline. This normalized value significantly correlated with the inverse of the anaerobic threshold (r = 0.81, p < 0.0001). The logarithm of the normalized oxygen deficit significantly correlated with the maximum ramp work rate (r = -0.86, p < 0.0001) and the highest constant work rate sustained for 15 min (r = -0.82, p < 0.0001). In addition, the time to reach plateau oxygen consumption for the 25-W exercise significantly correlated with the inverse of the anaerobic threshold (r = -0.78, p < 0.0001), the maximum ramp work rate (r = -0.76, p < 0.0001), and the highest constant work rate sustained for 15 min (r = -0.74, p < 0.0001). Thus, the oxygen deficit seen in patients with heart failure during constant work rate exercise results from abnormally slow oxygen uptake kinetics and correlates with exercise capacity as measured by anaerobic threshold (via gas exchange analysis) and maximal and submaximal exercise tolerance. Oxygen deficit warrants further evaluation as a submaximal index of functional capacity in patients with heart failure.