The present study examined if high-altitude (HA) exposure altered the vascular generation of free radicals across the human lungs and subsequent implications for the pathophysiology of high-altitude pulmonary oedema (HAPE). Since HAPE is preceded by a marked elevation in pulmonary artery pressure (PAP), we hypothesised that HAPE would be characterised by a net outflow or release of free radicals that would directly correlate with the rise in PAP. Thirty four subjects were examined at sea-level (SL) and 20h following active ascent to HA (4559m) following an overnight stay. Resting plasma samples were obtained from a central venous (superior vena cava) and radial arterial catheter for direct detection of the ascorbate free radical (A^●-) by EPR spectroscopy. Haemoglobin and haematocrit were also measured to correct for plasma volume shifts. Pulmonary blood flow (PBF) was determined by a re-breathing technique and pulmonary A^●- exchange kinetics calculated via the Fick method [arterio-venous concentration difference (a-v_diff) x (plasma) PBF]. Systolic PAP (PASP) was estimated by Doppler echocardiography and chest radiography confirmed HAPE. Clinical acute mountain sickness (AMS) was diagnosed as previously described (Bailey et al., 2006). Due primarily to an increase in the arterial concentration, the a-v_diff of A^●- increased from 18 (mean) ± 195 (SD) at SL to 206 ± 215 arbitrary units (AU)√Gauss (G) at HA (P < 0.05, Wilcoxon Matched Pairs Signed Ranks Test) resulting in a marked increase in A^●- outflow (SL: 56 ± 736 vs. HA: 750 ± 834 AU√G/min, P< 0.05). Outflow correlated (r = 0.35, P < 0.05, Pearson Product Moment Correlation) with the increase in PASP also observed at HA (SL: 23 ± 4 vs. HA: 38 ± 9 mmHg, P < 0.05). Four subjects developed HAPE and twenty subjects developed mild to severe AMS. Compared to healthy controls (n = 10), HAPE subjects presented with a greater increase in PASP (+27 ± 9 vs. +13 ± 6 mmHg, P < 0.05) and corresponding outflow of A^●- (+2379 ± 652 vs. 0 ± 1020 AU√G/min, P < 0.05) that were linearly related (r = 0.80, P = 0.05). These findings provide the first direct evidence for a net outflow or release of free radicals across the lungs at HA that may have implications for the pathophysiology of HAPE. A superoxide-mediated reduction in endothelium-derived nitric oxide bioavailability may prove the underlying mechanism responsible for the hypoxic pulmonary vasoconstriction observed in HAPE.