The recurrent failure of apparently promising neuroprotective drugs to improve outcome in trials of patients with acute ischaemic stroke may partially be explained by over-optimistic conclusions about efficacy as a result of methodological shortcomings in preclinical studies. We assessed the methodological quality of animal studies of five different neuroprotective agents that have been tested in 21 clinical trials including a total of more than 12,000 patients with acute ischaemic stroke.We performed a literature search restricted to full publications on the effects of clomethiazole, gavestinel, lubeluzole, selfotel, or tirilazad mesylate on infarct volume or functional outcome in animal models of acute focal cerebral ischaemia. We used a rating scale to assess the methodological quality of the included studies. One point was attributed to each of 10 items. A score of 4 to 6 points was considered "medium" and a score above 7 "high."A total of 45 articles were included. The median score on the methodological quality index was 3; 18 studies had a medium score and one a high score. Randomised treatment allocation was mentioned in 19 studies (42 %), blinded administration of study medication in 10 (22 %), and blinded outcome assessment in 18 (40 %). The study drug was administered at a median of 10 min (range, -60 to 360 min) after the onset of ischaemia.The evidence for neuroprotective efficacy that formed the basis for initiating the 21 trials was obtained in animal studies with a methodological quality that would, in retrospect, not justify such a decision. More rigorous preclinical study methodology may lead to more reliable and reproducible results.