Oxidative stress contributes to the vascular and neuropathic complications of experimental diabetes mellitus. This is evidenced by measures of tissue oxidant damage such as nerve lipid peroxidation and changes in antioxidant protection systems, for example decreased nerve reduced glutathione (GSH) and superoxide dismutase content [1,2]. Functional effects of antioxidant treatment include protection of vascular endothelium function [3±5], and improved nerve blood flow, conduction velocity (NCV) and regenerative capacity [2, 6±10]. a-Lipoic acid (LPA) is a naturally occurring free radical scavenger and transition metal chelator. LPA is also a cofactor for mitochondrial pyruvate dehydrogenase and has been termed a ªmetabolic antioxidantº [11]. Recent studies from Low's group [2,12] showed that treatment with LPA racemate (racLPA) prevented the development of digital sensory NCV deficits, impaired sciatic nerve blood flow, diminished nerve GSH, and deficient glucose uptake by neural tissues in diabetic rats. Short- Diabetologia (1998) Summary Elevated oxidative stress and impaired n-6 essential fatty acid metabolism contribute to defective nerve conduction velocity (NCV) and perfusion in diabetic rats, which may be corrected by free radical scavenger and g-linolenic acid (GLA) treatments. a-Lipoic acid (LPA) has antioxidant actions and both LPA racemate (racLPA) and GLA treatments produced benefits in clinical neuropathy trials. The aims were to study LPA action on neurovascular function in diabetic rats and to investigate potential interactions for co-treatment with GLA and other essential fatty acids. After 6 weeks of diabetes, 2 weeks of racLPA treatment corrected 20 % sciatic motor and 14 % saphenous sensory NCV deficits. The ED 50 for motor NCV restoration was approximately 38 mg kg ±1 day ±1. racLPA also corrected a 49 % diabetic deficit in sciatic endoneurial blood flow. R and S-LPA enantiomers were equipotent in correcting NCV and blood flow deficits. Treatment of diabetic rats with low doses (20 mg kg ±1 day ±1 ) of racLPA and GLA, while having modest effects on their own, showed evidence of marked synergistic action in joint treatment, completely correcting motor NCV and blood flow deficits. This was also noted for the novel compound, SOC0150, which contains equimolar proportions of LPA and GLA (ED 50 9.3 mg kg ±1 day ±1 , containing 3.5 mg LPA). NCV effects also showed marked synergism when racLPA:GLA ratios were varied over a 1:3±3:1 range. In contrast, a compound containing LPA and the n-3 component, docosahexaenoic acid, showed similar activity to LPA alone. Thus, LPA-GLA interactions yield drug combinations and compounds with an order of magnitude increase in efficacy against experimental diabetic neuropathy and are worthy of consideration for clinical trials. [Diabetologia (1998) 41: 390±399]