It is difficult to establish predictors for diabetic neuropathy because no generally accepted criteria of its diagnosis exist. In previous investigations risk factors profiles for neuropathy differ markedly. The aim of this study was to assess risk predictors for diabetic neuropathy in relation to different criteria of its diagnosis. Ninety-five diabetic patients entered the study. The exclusion criteria included uremia, alcohol abuse and radiculopathy. Control group consisted of 43 healthy volunteers. All patients underwent the clinical assessment, instrumental evaluation of superficial and deep sensation, tests of cardiovascular autonomic function and nerve conduction studies. According to the performed assessment patients were classified into following groups: without neuropathy, suspicion of neuropathy, neuropathy confirmed in clinical examination, neuropathy confirmed in electrophysiological testing, autonomic neuropathy. Analysis showed that the most important predictors in patients with subjective symptoms were type 2 diabetes mellitus, diabetes duration and age of patients. When neuropathy was diagnosed according to the clinical examination, predictors included type 2 diabetesmellitus and duration of the disease. In the cases of neuropathy confirmed by electrophysiological studies and autonomic neuropathy, only diabetes duration appeared to be a significant predictor. Our study demonstrated that predictors for diabetic neuropathy varied in relation to different diagnostic criteria and where the most important predictor for all forms of neuropathy is duration of diabetes. This result indicates the need for frequent screening tests in patients with long duration of the disease, irrespective of its metabolic control, patients' age or type of diabetes.