The number of capillaries per mm2, minimum intercapillary distance, number of endothelial nuclei per capillary section, and percentage of capillaries closed were evaluated in transverse sections of fascicles of 45 control and 36 diabetic sural nerves. All controls and patients were prospectively studied to ascertain their diabetic and neuropathic status. An index of pathology was introduced and it was found to provide a sensitive and reliable measurement of the presence and severity of neuropathy. The number of capillaries and minimum intercapillary distance of diabetic nerves were not significantly different from those of controls (P > 0.05). Diabetic nerves exhibited a small but statistically significant increase in the number of endothelial nuclei per capillary that was positively correlated with the severity of neuropathy. The most striking abnormality was the statistically significant increase in the percentage of capillaries closed in patients with neuropathy as compared to those without neuropathy and controls. Among diabetics, this percentage increased with the severity of neuropathy (P = 0.008). The two capillary abnormalities that have been demonstrated may play a role in the development of diabetic polyneuropathy.Diabetic polyneuropathy, a common complication of diabetes mellitus, is typically expressed as sensory loss, pain, and autonomic dysfunction in the feet and legs (1). Although the mechanisms underlying diabetic neuropathy remain unknown, chronic hyperglycemia may lead to metabolic derangement that directly affects neurons (axons) or Schwann cells or indirectly affects them by first altering another tissue-e.g., vessels. Since a higher rate of atherosclerotic heart disease and peripheral vascular disease occurs among diabetics (2, 3) and since arteriosclerosis with vessel occlusion is reported for vasa nervorum of nerve (4-6), arteriosclerosis has been postulated as a cause of diabetic neuropathy. This hypothesis, however, may not explain the common occurrence of an abnormality of nerve conduction and development of a diffuse neuropathy among diabetic patients who do not manifest peripheral vascular disease (7-10). Attention, therefore, has been focused on metabolic derangements that might affect neurons (or their peripheral axons) or Schwann cells. Among the metabolic mechanisms that have been considered are the following: (i) lipid alterations (11-13); (ii) accumulation of sorbitol and fructose (14, 15); (iii) decreased nerve myo-inositol (16); (iv) increased nonenzymatic glycosylation of protein (17, 18); (v) decrease of Na+,K+-ATPase (19); (vi) increased intraaxonal sodium (20); (vii) alterations in axonal flow (21) and axonal attenuation (22); and (viii) tissue dehydration (23). Since the nerve conduction abnormality is not readily reversed after near normalization of blood glucose for periods of up to 8 months, an irreversible nerve alteration or an intervening pathologic alteration interposed between metabolic derangement and neuropathic dysfunction is inferred (24).A functional...