High-intensity exercise (> or =90% of maximal O(2) uptake) sustained to the limit of tolerance elicits expiratory muscle fatigue (EMF). We asked whether prior EMF affects subsequent exercise tolerance. Eight male subjects (means +/- SD; maximal O(2) uptake = 53.5 +/- 5.2 ml.kg(-1).min(-1)) cycled at 90% of peak power output to the limit of tolerance with (EMF-EX) and without (CON-EX) prior induction of EMF and for a time equal to that achieved in EMF-EX but without prior induction of EMF (ISO-EX). To induce EMF, subjects breathed against an expiratory flow resistor until task failure (15 breaths/min, 0.7 expiratory duty cycle, 40% of maximal expiratory gastric pressure). Fatigue of abdominal and quadriceps muscles was assessed by measuring the reduction relative to prior baseline values in magnetically evoked gastric twitch pressure (Pga(tw)) and quadriceps twitch force (Q(tw)), respectively. The reduction in Pga(tw) was not different after resistive breathing vs. after CON-EX (-27 +/- 5 vs. -26 +/- 6%; P = 0.127). Exercise time was reduced by 33 +/- 10% in EMF-EX vs. CON-EX (6.85 +/- 2.88 vs. 9.90 +/- 2.94 min; P < 0.001). Exercise-induced abdominal and quadriceps muscle fatigue was greater after EMF-EX than after ISO-EX (-28 +/- 9 vs. -12 +/- 5% for Pga(tw), P = 0.001; -28 +/- 7 vs. -14 +/- 6% for Q(tw), P = 0.015). Perceptual ratings of dyspnea and leg discomfort (Borg CR10) were higher at 1 and 3 min and at end exercise during EMF-EX vs. during ISO-EX (P < 0.05). Percent changes in limb fatigue and leg discomfort (EMF-EX vs. ISO-EX) correlated significantly with the change in exercise time. We propose that EMF impaired subsequent exercise tolerance primarily through an increased severity of limb locomotor muscle fatigue and a heightened perception of leg discomfort.