In the past 40 years, debate on the aetiology of diabetic neuropathy has oscillated between propositions based on vascular and neurochemical events. The original work of Fagerberg,'suggesting a vascular aetiology, fell out of favou? and latterly the field has been dominated by a neurochemical approach highlighting the effects of hyperglycaemia on the neuronal polyol pathway and phosphoinositide metab~ lisrn.~,~ More recently, the vascular hypothesis enjoyed a revival resulting from demonstrations of reduced nerve blood flow and endoneurial hypoxia in diabetic rats; 5 the latter was also established in neuropathic patients. 6 A period of vigorous advocation of apparently contrasting hypotheses polarized the debate into metabolic and vascular camps. Recent advances in our understanding of metabolic changes in diabetes have identified processes with a wide-ranging pathogenetic potential, such as advanced glycation7 and alterations in essential fatty acid metabolism.* Coupled with a great increase in our knowledge of vascular biology, particularly the autocrine/paracrine role of the endothelium9-" and contemporary research on the effects of diabetes and hypergly~ aemia,'~-'~ powerful arguments can be made linking metabolic processes with