Performance of intense dynamic exercise in highly trained athletes is associated with a reduced arterial haemoglobin saturation for O2 (SaO2) and lower arterial PO2 (PaO2). We hypothesized that compared with upright exercise, supine exercise would be accompanied by a smaller reduction in SaO2 because of a lower maximal O2 uptake (VO2max) and/or a more even ventilation-perfusion distribution. Eight elite bicyclists completed progressive cycle ergometry to exhaustion in both positions with concomitant determinations of ventilatory data, arterial blood gases and pH. During upright cycling VO2max averaged 75 +/- 1.6 mL O2 min-1 kg-1 (+/-SEM) and it was 10.6 +/- 1.7% lower during supine cycling (P < 0.001). Also the maximal pulmonary and alveolar ventilation were lower during supine cycling (by 15 +/- 2% and 21 +/- 3%, respectively; P < 0.001) which related to a 0.8 +/- 0.1 L lower tidal volume (P < 0.001). In all subjects and independent of work posture PaO2 and SaO2 decreased from rest to exhaustion (from 99 +/- 3 to 82 +/- 2 Torr and 98.1 +/- 0.2 to 95.2 +/- 0.4%, respectively; P < 0.001); alveolar-arterial PO2 difference increased from 6 +/- 2 to 37 +/- 3 Torr in both body positions. At exhaustion arterial PCO2 was lower in upright than in supine (33.4 +/- 0.6 vs. 35.9 +/- 0.9 Torr; P < 0.01), suggesting a greater relative hyperventilation in upright. Arterial pH was similar in upright and supine at rest (both 7.41 +/- 0.01) and at exhaustion (7.31 +/- 0.01 vs. 7.32 +/- 0.01, respectively). We conclude that despite a lower Vo2max and supposedly an improved ventilation-perfusion distribution, altering body position from upright to supine does not influence arterial O2 desaturation during intense exercise.