Symptoms of infection are common in trekkers at high altitude and more prevalent among those with acute mountain sickness (AMS) (Murdoch, 1995). Since hypoxia results in a given level of work being more strenuous than when performed at sea level (Lawler et al. 1988) these increased infection rates may well be influenced by changes in immune cell distribution and function that are known to be modulated by exercise intensity, duration and fitness (McCarthy & Dale, 1988). Furthermore, recent evidence indicates that anti-oxidant defence mechanisms are overwhelmed on ascent to high altitude and may also be implicated in the aetiology of AMS (Bailey et al. 2001).

With local ethics committee approval, using a randomised single blind cross-over design, eight healthy male subjects (age 21.6 ± 2.3 years,V_O2,peak 3.84 ± 0.59 l min^-1; means ± S.D.) breathed a hypoxic (H; F_I,O2 = 14 %) gas mixture or normoxic (N; F_I,O2 = 20.9 %) room air via a mouthpiece for 150 min. Between 60 and 90 min subjects exercised on a cycle ergometer at a predetermined workload to elicit 50 % normoxicV_O2,peak (70 r.p.m.). Venous blood was taken at baseline (-30 and 0 min), 60, 90 and 150 min. Total leucocyte and subset counts, neutrophil oxidative activity (qualitative nitroblue tetrazolium (NBT) reduction test) and plasma anti-oxidant status (peroxynitrite ABEL¿ Pholasin¿ assay system; Knight Scientific Ltd, Plymouth, UK) were measured. Leucocyte data were corrected for changes in plasma volume (Dill & Costill, 1974). Data were compared between conditions at each time point using ANOVA.

The magnitude of leucocytosis induced by 30 min exercise was greater in H (7.0 ± 1.7 to 10.1 ± 2.8 X 10⁹ l^-1) compared with N (6.9 ± 1.4 to 8.6 ± 2.1 X 10⁹ l^-1; P < 0.005). This was achieved by a relative neutrophilia (H, 4.5 ± 1.4 to 6.0 ± 2.3 X 10⁹ l^-1; N, 4.2 ± 1.3 to 5.0 ± 1.7 X 10⁹ l^-1; P < 0.05) and lymphophilia (H, 1.9 ± 0.4 to 3.3 ± 0.9 X 10⁹ l^-1; N, 2.0 ± 0.5 to 2.7 ± 0.7 X 10⁹ l^-1; P < 0.05) in H. The relative neutrophilia remained evident after 60 min of recovery (P < 0.05). NBT-positive neutrophils were greater in H at 60 min (P = 0.06) and peaked immediately post-exercise (H, 2.8 ± 1.3 X 10⁹ l^-1; N, 2.4 ± 1.0 X 10⁹ l^-1; P = 0.08). This difference became significant at 150 min (P < 0.05). Post-exercise anti-oxidant levels were greater than baseline (H, 398 ± 55 to 452 ± 73; N, 394 ± 45 to 415 ± 37 Vitamin E analogue equivalent units; VEA eq units µmol l^-1) and remained elevated in H at 150 min (H, 421 ± 49; N, 395 ± 43 VEA eq units µmol l^-1). However, these differences were not significant despite n = 6 and n = 7 subjects respectively having greater anti-oxidant values in H.

These data indicate an augmented cellular immune response and a probable paradoxical increase in plasma anti-oxidants after exercise in moderate hypoxia. The significance of these phenomena and the relative contributions of hypoxia, modulating relative work intensity, and hypoxia per se remain to be elucidated.

We would like to thank Dr Jan Knight, Knight Scientific Ltd, Plymouth, UK, for anti-oxidant assay facilities and Walsgrave Hospital, Haematology Department, Coventry, UK, for performing differential blood counts.