Diabetic neuropathy is a serious complication affecting as many as 60 % of diabetic patients [1]. Its clinical manifestations can vary from reduced nerve conduction velocity and sensory deficits to autonomic disturbances associated with increased mortality [2,3]. Despite numerous research efforts, the pathogenesis of diabetic neuropathy remains incompletely understood. Proposed mechanisms include microvascular abnormalities, altered Schwann cell or neuronal metabolism, deficiencies of neurotrophic substances and slowing of axonal transport [4±7]. In diabetic patients, changes in function of peripheral and autonomic nerves can be detected very early in the dis- Abstract Aims/hypothesis. To improve understanding of the pathophysiology of diabetic neuropathy and to establish a primate model for experimental studies, we examined nerve changes in baboons with Type I (insulin-dependent) diabetes mellitus. We also examined the effect of aminoguanidine (an inhibitor of the formation of advanced glycation end products) on nerve function. Methods. Male baboons (Papio hamadryas) were assigned to four groups; control, diabetic, control and diabetic treated with aminoguanidine. Diabetes was induced with streptozotocin (60 mg/kg, intravenous). Insulin and aminoguanidine (10 mg/kg) were injected subcutaneously daily. Motor and sensory nerve conduction velocity was measured using standard techniques. Autonomic function was examined by measuring heart rate response to positional change. Sural nerve morphometry was analysed in the diabetic group (mean duration 5.5 years) along with their age-matched controls. Results. The diabetic groups were smaller in size with a mean HbA 1 c of 8.9 ± 1.2 %. The nerve conduction velocity and heart rate response was reduced in the diabetic groups. Morphometric analysis of the diabetic sural nerve showed smaller axon diameter (2.99 ± 0.06 mm vs 3.29 ± 0.06 mm; p < 0.01) accompanied by thinner myelin (1.02 ± 0.02 mm vs 1.15 ± 0.02 mm, p < 0.01) with no change in the axon density. Treatment with aminoguanidine for 3 years had no effect on glycaemic control and did not restore conduction velocity or autonomic dysfunction in the diabetic animals, contrary to the studies in rats. Conclusions/interpretation. These results show that the primate is a good model to study diabetic neuropathy and suggest that the accumulation of advanced glycation end products are not an early mechanism of nerve damage in this disorder. [Diabetologia (2000) 43: 110±116]