This study investigates the role of central vs. peripheral factors in the limitation of maximal oxygen uptake (Vo(2max)) with moderate hypoxia [inspired fraction (Fi(O(2))) =14.5%]. Fifteen endurance-trained athletes performed maximal cycle incremental tests to assess Vo(2max), maximal cardiac output (Q(max)), and maximal arteriovenous oxygen (a-vO(2)) difference in normoxia and hypoxia. Muscle biopsies of vastus lateralis were taken 1 wk before the cycling tests to evaluate maximal muscle oxidative capacity (V(max)) and sensitivity of mitochondrial respiration to ADP (K(m)) on permeabilized muscle fibers in situ. Those athletes exhibiting the largest reduction of Vo(2max) in moderate hypoxia (Severe Loss group: -18 +/- 2%) suffered from significant reductions in Q(max) (-4 +/- 1%) and maximal a-vO(2) difference (-14 +/- 2%). Athletes who well tolerated hypoxia, as attested by a significantly smaller drop of Vo(2max) with hypoxia (Moderate Loss group: -7 +/- 1%), also display a blunted Q(max) (-9 +/- 2%) but, conversely, were able to maintain maximal a-vO(2) difference (+1 +/- 2%). Though V(max) was similar in the two experimental groups, the smallest reduction of Vo(2max) with moderate hypoxia was observed in those athletes presenting the lowest apparent K(m) for ADP in the presence of creatine (K(m+Cr)). In already-trained athletes with high muscular oxidative capacities, the qualitative, rather than quantitative, aspects of the mitochondrial function may constitute a limiting factor to aerobic ATP turnover when exercising at low Fi(O(2)), presumably through the functional coupling between the mitochondrial creatine kinase and ATP production. This study suggests a potential role for peripheral factors, including the alteration of cellular homeostasis in active muscles, in determining the tolerance to hypoxia in maximally exercising endurance-trained athletes.