In order to determine the local prevalence of polyneuropathy among adult outpatients with type II (non-insulin-dependent) diabetes mellitus, we applied a series of standardised measures to patients attending a multidisciplinary diabetes clinic. The study group comprised 94 men and 15 women; mean age, 70.6 +/- 7.8 years; mean duration of diabetes, 11.7 +/- 10.1 years; and mean HbA 1 C , 8.3% +/- 1.7%. Neuropathy Symptom Scores >= 1 were present in 97% of patients (mean, 3 +/- 2; range, 0-12), and 95% had Neuropathy Disability Scores >= 2 (mean, 27 +/- 19; range, 0-87). 52% of men reported impotence. Autonomic dysfunction on cardiovascular reflex testing was present in 46% of patients (39/84). Finger and toe vibration perception thresholds were greater than 3SD higher than mean thresholds measured in control subjects without diabetes in 43% and 58% of patients, respectively. Polyneuropathy, defined as lower limb sensory and motor nerve conduction velocity or latency outside mean +/- 2 SD of that measured in age-matched controls, was present in 49% of patients (53/109). These results suggest that there is a high prevalence of polyneuropathy in Australian out-patients with type II diabetes mellitus. In this study, clinical assessment using Neuropathy Disability Scores was not diagnostically useful since only five patients had a normal score. Using nerve-conduction studies as the ''gold standard'' diagnostic criteria, the best alternative test for the presence of polyneuropathy was toe vibration perception threshold (sensitivity 74%, specificity 56%). In view of the emerging evidence that intensive glycaemic control reduces the rate of progression of polyneuropathy, we recommend that patients with type II diabetes mellitus have nerve-conduction studies performed for early detection of this important complication.