Nerve conduction abnormalities in peripheral nerves from diabetic patients may be early indicators for the future development of symptomatic neuropathy. In this study, three weeks of experimental diabetes in the rat caused a significant decrease in motor nerve conduction velocity measured in vivo (45.3 +/- 3.6 m/s; mean +/- S.D.) compared to controls (57.7 +/- 4.5 m/s). myo-Inositol administration to diabetic rats (500 mg/rat per day) for the duration of the study, partially prevented this decrease (50 +/- 4.4 m/s). An analogue of myo-inositol, PP56 (D-myo-inositol-1,2,6-trisphosphate), at a dose of 24 mg/rat per day completely prevented this reduction in diabetic rats (53.4 +/- 5.8 m/s). Resistance to hypoxic conduction block was determined in vitro in endoneurial preparations and was assessed as the decline in compound action potential amplitude over a 40 min period of hypoxia. Compound action potential amplitude (as % of initial value +/- S.D.) was significantly greater in diabetic preparations compared with controls at 40 min of hypoxia (76.1 +/- 9.1 vs. 54.8 +/- 14.7 respectively). Treatment to diabetic rats with myo-inositol did not significantly affect this value (79.9 +/- 16.6) but PP56 treatment partially prevented the increased resistance to hypoxic conduction block (69.4 +/- 16.0). This study demonstrates that these acute abnormalities of nerve function in early experimental diabetes may be attenuated by the administration of PP56, possibly acting via a vascular mechanism.