The effects of the aldose reductase inhibitor ponalrestat (600 mg day-1) on sensory, electrophysiological, and autonomic function were examined in 50 patients with chronic symptomatic, distal symmetrical diabetic neuropathy in a 52-week randomized, double-blind, parallel-group, placebo-controlled, single-centre study. In an endeavour to identify patients with a degree of neuropathy potentially amenable to pharmacological intervention, a minimum conduction velocity of 30 m s-1 was set for the peroneal motor nerve. At 52 weeks, no significant differences were observed between the ponalrestat and placebo groups in motor (ulnar, median, and peroneal) or sensory (ulnar and radial) nerve conduction velocities, vibration perception thresholds, adjectival symptom scores or tests of autonomic function (mean electrocardiographic R-R interval variability on deep breathing and orthostatic blood pressure response). Ponalrestat was clinically well tolerated and had no significant effect on glycaemic control. The lack of beneficial effects of ponalrestat may in part reflect the advanced stage of the neuropathic process in patients with established symptomatic disease, and the poor reproducibility of current neurophysiological techniques. Firmer knowledge of clinico-pathological correlates allied to improved non-invasive neurophysiological measurement techniques should facilitate the selection of patients for future therapeutic trials in diabetic neuropathy.