Volume 107 (2004), pp. 273–280

The authors of the above paper have now realised that there are problems with the integrity of the data presented in Figures 1, 2 and 3 which compromise the main observations. The Editorial Board of Clinical Science is therefore retracting this paper.