The efficacy of the neurotrophic peptide ORG 2766 in diabetic patients with polyneuropathy was evaluated in a double-blind, placebo-controlled, multicentre trial. One hundred and twenty four patients were randomised in five groups to receive 0.1, 0.4, 2 or 5 mg ORG 2766 or placebo, once daily, administered subcutaneously 52 weeks. Thermal discrimination thresholds (TDT) and vibration perception thresholds (VPT), motor and sensory nerve conduction velocity, Hoffmann reflex, heart rate variation during deep breathing and heart rate response after standing up, neurological examination score and neuropathic symptom score were determined at baseline and after 17, 34 and 52 weeks of treatment. Of the nerve function indices studied, at week 52 the TDTwarmth of the hand in the ORG 2766 0.1, 0.4 and 5 mg groups and the TDTcold of the foot in the ORG 2766 0.1 and 0.4 mg groups significantly improved compared with placebo. Further significant improvement as compared with placebo was observed in the paraesthesia score at week 34 and week 52 in the ORG 2766 2 mg group. Only at week 34 had both the heartbeat variation during deep breathing and the VPT of the foot in the ORG 2766 0.1 mg group improved significantly, compared with placebo. No further statistically significant differences were observed at time for the other measures. No adverse reactions were observed. The only recorded drug-induced side effect was pain at the injection site. Taking all measures of efficacy into account, the statistically significant results observed did not show consistency within each measure. Therefore, it is concluded that ORG 2766, in contrast to earlier reports, is not effective in treating diabetic polyneuropathy.