OBJECTIVE: To assess the axonopathy and demyelination neuropathy according to the electrophysiologic severity in diabetic neuropathy. METHOD: Electrophysiologic data of 246 patients who had been diagnosed with diabetic neuropathy was obtained and classified into suspected, possible, and definite groups by the criteria of our laboratory. Nerve conduction study was performed in the median, ulnar motor and sensory nerves, peroneal and tibial motor nerves, and sural nerve. Statistics were done with the results from the median motor and sensory, tibial motor and sural nerves. According to the severity of diabetic neuropathy, correlation and Chi-square analysis between amplitudes and latencies were performed. RESULTS: Frequencies of diabetic neuropathy according to electrophysiologic severity were as follows: 24 cases of suspected, 141 cases of possible, and 81 cases of definite neuropathy. The correlation ratios between amplitude and latency were 0.41- 0.79 (p<0.05) in the definite group of all the nerves examined, and below 0.3 in the suspected and possible groups. By Chi-square analysis, amplitude reduction was the predominant finding in the suspected and possible groups.si CONCLUSION: In the early stage of diabetic neuropathy, axonopathy might be the preceding pathogenesis, while with progression of diabetic neuropathy, axonopathy and demyelination may coexist.