To examine whether sea-level hypoxic ventilatory responses (HVR) predict acute mountain sickness (AMS) and document temporal changes in ventilation, HVR, gas exchange, and fluid balance, we measured these parameters at low altitude (100 m) and daily during 3 days at high altitude (4559 m). At low altitude, there were no significant differences in rest or exercise isocapnic HVR, poikilocapnic HVR at rest, and hypercapnic ventilatory response between 12 subjects without significant AMS and 11 subjects who fell sick. No low altitude ventilatory responses correlated with AMS or fluid balance at high altitude. On day 1, isocapnic HVR was significantly lower in the AMS group [0.86 +/- 0.43 (SD) vs. 1.43 +/- 0.63 L/min/% Sa(O2), p < 0.05). AMS was associated with higher AaD(O2), lower Pa(O2), and Sa(O2), while Pa(CO2) was not different between subjects with and without AMS. Both groups showed equivalent reductions in urine volume, sodium output, and gain in body weight on day 1 while climbing to 4559 m, but on day 2 only subjects without AMS had diuresis, natriuresis, and weight loss. We conclude that (1) susceptibility to AMS, fluid balance, and ventilation at high altitude cannot be predicted by low altitude HVR testing and (2) that the failure to increase HVR on arrival at high altitude and impaired gas exchange, possibly due to interstitial edema, may account for the more severe hypoxemia in AMS.