The preventive effects of combined or separate treatment for 10 weeks with an aldose reductase inhibitor, epalrestat (50 mg/kg/day), and a vasodilator, cilostazol (30 mg/kg/day), on nerve conduction deficits and morphometric alterations were examined in streptozotocin-induced diabetic rats. The average motor nerve conduction velocities (MNCV) in the tail nerve of the untreated diabetic (DM) group, the group treated with epalrestat (ES), the group treated with cilostazol (CZ), the group with both agents together (ES&CZ), and the normal control group were 34.7, 37.7, 39.3, 39.0 and 42.1 m/s, respectively. All treatments partially but significantly prevented a reduction in MNCV. The MNCV in the ES&CZ group was almost the same as in the CZ group. In a morphometric study of the sural nerve, the DM group showed a reduction in the average diameter of myelinated fiber and in occupancy (percentage of the fascicular area occupied by myelinated fibers), and a shift in the diameter-frequency histogram to smaller diameters. Only the CZ group showed evidence of a partial but significant preventive effect on the decrease in occupancy. In the CS and ES&CZ groups, there was a significant tendency away from the shift of histograms to smaller diameters. The ES&CZ group did not show any fewer morphometric changes than the CZ group. Thus, there was no synergism between the effects of epalrestat and cilostazol on the development of experimental diabetic neuropathy. This finding may provide a useful clue to the mechanisms of action of ES and CZ in diabetic neuropathy.