Thirty patients with diabetic polyneuropathy were recruited from the diabetic clinic in Hospital Universiti Sains Malaysia from 1996 to 1998. They were randomly assigned either sulbutiamine (Arcalion(®)) (15 patients) or no treatment (control group; 15 patients). Glycaemic control was assessed by blood glucose and HbA1. Severity of neuropathy was assessed by symptom and sign score, and electrophysiological parameters (nerve conduction velocity and compound muscle action potential) at entry to the study and after 6 weeks. There were improvements in the electrophysiological parameters in the treatment group when compared to the controls with significant improvement in the median nerve conduction velocity (p<0.001), median compound muscle action potential (p<0.001), peroneal nerve conduction velocity (p<0.001), and peroneal compound muscle action potential (p<0.001). No significant improvement in symptom and sign scores were noted between the groups but a significant improvement compared to base line was noted for the sulbutiamine treated group. (p< 0.05). The glycaemic control in both groups was not significantly different at base line and was stable throughout the study. Sulbutiamine objectively improved peripheral nerve function in diabetic polyneuropathy although the symptom score did not improve, possibly due to the short duration of the study.