Despite considerable research we still do not have a comprehensive explanation for the pathogenesis of diabetic neuropathy. Although chronic hyperglycaemia is almost certainly involved, it is not known whether the primary pathology is metabolic, microvascular, or an interaction between the two. Hyperglycaemia-induced polyol pathway hyperactivity associated with nerve sorbitol accumulation and myo-inositol depletion may play a part in the genesis of diabetic neuropathy. The case for microvascular disease in diabetic neuropathy is now strong. Fibre loss in human sural nerve is multifocal, suggesting ischaemia. The degree of vessel disease has been related to the severity of neuropathy. People with chronic obstructive pulmonary disease develop the so called "hypoxic neuropathy" in which similar microvascular changes occur as in diabetic neuropathy. In rats with experimental diabetic neuropathy nerve blood flow is reduced and oxygen supplementation or vasodilator treatment improved the deterioration in conduction velocity and nerve blood flow. Similarly, in human diabetic neuropathy, there is impaired nerve blood flow, epineurial arterio-venous shunting and a reduction in sural nerve oxygen tension. At what stage during the development of nerve damage these changes occur is yet to be determined.