We have recently demonstrated that changes in the work of breathing during maximal exercise affect leg blood flow and leg vascular conductance (C. A. Harms, M. A. Babcock, S. R. McClaran, D. F. Pegelow, G. A. Nickele, W. B. Nelson, and J. A. Dempsey. J. Appl. Physiol. 82: 1573-1583, 1997). Our present study examined the effects of changes in the work of breathing on cardiac output (CO) during maximal exercise. Eight male cyclists [maximal O2 consumption (VO2 max): 62 +/- 5 ml . kg-1 . min-1] performed repeated 2.5-min bouts of cycle exercise at VO2 max. Inspiratory muscle work was either 1) at control levels [inspiratory esophageal pressure (Pes): -27.8 +/- 0.6 cmH2O], 2) reduced via a proportional-assist ventilator (Pes: -16.3 +/- 0.5 cmH2O), or 3) increased via resistive loads (Pes: -35.6 +/- 0.8 cmH2O). O2 contents measured in arterial and mixed venous blood were used to calculate CO via the direct Fick method. Stroke volume, CO, and pulmonary O2 consumption (VO2) were not different (P > 0.05) between control and loaded trials at VO2 max but were lower (-8, -9, and -7%, respectively) than control with inspiratory muscle unloading at VO2 max. The arterial-mixed venous O2 difference was unchanged with unloading or loading. We combined these findings with our recent study to show that the respiratory muscle work normally expended during maximal exercise has two significant effects on the cardiovascular system: 1) up to 14-16% of the CO is directed to the respiratory muscles; and 2) local reflex vasoconstriction significantly compromises blood flow to leg locomotor muscles.