BACKGROUND Diabetic peripheral neuropathy is a common complication of diabetes mellitus. OBJECTIVE To assess the efficacy of aldose reductase inhibitors in the prevention, reversal or delay in the progression of diabetic peripheral neuropathy. METHODS The Cochrane Diabetes Group's database was searched and the citation lists of identified trials and previous reviews checked. Investigators identified as active in the field were approached for overlooked studies. METHODS Randomised controlled trials of aldose reductase inhibitors versus placebo, no treatment or other treatment in diabetic patients with or without clinical neuropathy. METHODS Nerve conduction velocity was the only end point measured in all trials. Treatment effect was evaluated in terms of nerve conduction velocity mean difference in median and peroneal motor and median and sural sensory nerves. RESULTS 19 trials, testing 4 different aldose reductase inhibitors for between 4 to 208 weeks duration (median 24 weeks), met the inclusion criteria for the meta-analysis. A small but statistically significant reduction in decline of median and peroneal motor nerve conduction velocities was present in the treated group when compared to the control group (weighted mean 0.66 m/s 95% CI 0.18-1.14 m/s and 0.53 m/s 95% CI 0.02-1.04m/s respectively). No clear benefit of aldose reductase inhibitor treatment was observed in either of the sensory nerves. CONCLUSIONS Although aldose reductase inhibitor treatment has been demonstrated to diminish the reduction in motor nerve conduction velocity, the clinical relevance of such a change in this outcome measure is uncertain. There was no effect in terms of this outcome measure in the smaller sensory fibres, degeneration of which is primarily responsible for the most common neuropathic syndrome associated with diabetes, that of severe pain and loss of sensation in the extremity leading in some cases to ulceration and eventual amputation.