Endoneurial hypoxia has been suggested as a mechanism of human and experimental diabetic neuropathy (EDN). We found that rats rendered diabetic for 4 months had reduced nerve blood flow (NBF) and nerve oxygen tension (Pno2). The NBF was reduced by at least 33% in EDN and 60% of the oxygen tensions in the endoneurial 02 histogram were less than 25 mm Hg (3.3 kPa) in EDN compared with only 19% in the controls. To test the hypothesis that EDN may in part be due to hypoxia, we studied the effectiveness of oxygen supplementation in preventing some electrophysiologic and biochemical abnormalities. Rats with EDN had reduced caudal nerve conduction velocity and had a resistance to ischemic conduction block. When a matched groups of rats with EDN were 02 supplemented for 4 weeks, the time to 50% block of nerve conduction and nerve conduction velocity was no longer statistically different from controls. Endoneurial free sugars (glucose, fructose, sorbitol) were markedly increased in EDN. Oxygen supplementation resulted in no change in plasma glucose; by contrast, these increased endoneurial free sugars were significantly reduced (towards normal) by 60%, 33%, and 34%, respectively. myo-Inositol, however, was further decreased by oxygen supplementation. These findings of a partial prevention of electrophysiologic and biochemical abnormalities support a role of hypoxia in the pathogenesis of EDN.While the pathogenesis of diabetic neuropathy remains obscure (1), certain clues have been provided by some recent studies. These include reduced energy utilization (2), increased sorbitol and decreased nerve free myo-inositol concentrations (2, 3), increased intra-axonal sodium levels (4), and a reduced rate of incorporation of lipid and amino acids into myelin (5). Axonal transport rates are also reduced in experimental diabetic neuropathy (6, 4.5 (50 mg-ml-1; dose 1 ml-kg-1). The control groups received an intraperitoneal injection of citrate buffer (1 ml-kg-') alone. The rats were housed in cages with plastic floors covered with sawdust and wood shavings and were fed Purina Rat Chow (no. 5001) with an unrestricted supply of water. Rats were accepted as diabetic if fasting glucose exceeded 300 mg/dl 2 days after streptozotocin and remained >300 mg/dl at the time of sacrifice. NBF and Oxygen Tension Measurements. NBF was measured by using the hydrogen (H2) clearance method (8, 9); H2 and 02 were sensed in nerve with platinum microelectrodes (3-to 5-gm tip diameter) polarized +0.25 V and -0.65 V, respectively, using the principle of polarography (9-11). NBF was calculated from the time constant of the H2 washout curve (9-11). Sciatic Pno2 was sampled as for H2 by using identical microelectrodes but polarized -0.65 V. After in vitro calibration, the electrode was inserted into the nerve fascicle until the tip was at the center of the fascicle. Recordings were repeated at steps of 0.1 x radius as the micropipet was withdrawn. The electrode was recalibrated and the process was repeated at a different site three to five times...