Oxidative stress is suggested to be central to the ageing process, with endogenous antioxidant defence and repair mechanisms in place to minimize damage. Theoretically, supplementation with exogenous antioxidants might support the endogenous antioxidant system, thereby reducing oxidative damage, ageing-related functional decline and prolonging life- and health-span. Yet supplementation trials with antioxidants in animal models have had minimal success. Human epidemiological data are similarly unimpressive, leading some to question whether vitamin C, for example, might have pro-oxidant properties in vivo. We supplemented cold exposed (7+/-2 degrees C) female C57BL/6 mice over their lifespan with vitamin C (ascorbyl-2-polyphosphate), widely advocated and self administered to reduce oxidative stress, retard ageing and increase healthy lifespan. No effect on mean or maximum lifespan following vitamin C treatment or any significant impact on body mass, or on parameters of energy metabolism was observed. Moreover, no differences in hepatocyte and lymphocyte DNA oxidative damage or hepatic lipid peroxidation was seen between supplemented and control mice. Using a DNA macroarray specific for oxidative stress-related genes, we found that after 18 months of supplementation, mice exhibited a significantly reduced expression of several genes in the liver linked to free-radical scavenging, including Mn-superoxide dismutase. We confirmed these effects by Northern blotting and found additional down-regulation of glutathione peroxidase (not present on macroarray) in the vitamin C treated group. We suggest that high dietary doses of vitamin C are ineffective at prolonging lifespan in mice because any positive benefits derived as an antioxidant are offset by compensatory reductions in endogenous protection mechanisms, leading to no net reduction in accumulated oxidative damage.