To evaluate the effects of fidarestat (SNK-860) on vibration perception threshold, as measured by C64 quantitative tuning fork (64Hz) analysis, as well as its effects on subjective symptoms in patients with diabetic polyneuro-pathy. Open-label, prospective study conducted at 12 hospitals in the central area of Honshu, Japan. Fidarestat was administered at a dosage of 1mg once daily after breakfast for 28 weeks. Vibration perception threshold of upper and lower extremities was determined using a C64 quantitative tuning fork, and measured at baseline and after 12 and 28 weeks of treatment. Subjective symptoms, including numbness, spontaneous pain and hypoaesthesia, were evaluated every 4 weeks. Subjective symptoms were evaluated in 22 patients, and vibration perception threshold data were available for 19 patients. Vibration perception threshold at baseline was negatively correlated with the severity of the following subjective symptoms: numbness in the upper limbs, and numbness, coldness and hot flushes, smarting pain causing difficulty walking and hypoaesthesia in the lower limbs. During treatment with fidarestat, vibration perception threshold increased significantly in the upper (p = 0.0017) and lower (p = 0.0001) limbs. The following symptoms were also significantly improved: severity of numbness in the lower limbs, heaviness in the foot, coldness and hot flushes in the lower limbs, smarting pain causing difficulty walking, sensation as if walking on sand, sensation as if walking on an uneven road, spontaneous pain in the lower limbs, and dizziness. Adverse events occurred in four patients. Administration of fidarestat after breakfast was effective in significantly alleviating some symptoms of diabetic polyneuropathy. The C64 quantitative tuning fork analysis is useful in the diagnosis of diabetic polyneuropathy, and as a measure of the severity of the neuropathological symptoms of this condition.