The effects of the aldose reductase inhibitor, sorbinil, on symptomatic symmetrical diabetic polyneuropathy were studied during a 6-month period in a double-blind parallel group placebo-controlled trial. Twenty-seven patients received sorbinil and 28 placebo. The patients were assessed by clinical examination, neurophysiological measurements, sensory threshold determinations and tests of autonomic nerve function. No major clinical benefit was seen in the sorbinil-treated patients and no differences in sensory thresholds were observed. In three out of nine neurophysiological tests (motor nerve conduction velocity of the posterior tibial nerve, F-wave latency and sensory distal latency of the ulnar nerve) and one out of five tests of autonomic nerve function (heart rate variation during deep breathing) significant differences between the patient groups evolved in favour of sorbinil treatment. An overall evaluation of the temporal development of these and remaining neurophysiological and autonomic variables suggested a small but significant benefit from sorbinil treatment. There was no evidence of continuing improvement throughout the treatment period and beneficial effects observed were no greater than those seen in previous trials of considerably shorter treatment periods. It is concluded that sorbinil treatment results in some improvement in peripheral nerve function in symptomatic diabetic polyneuropathy, but that the long-term effect may be of limited value.