When assessing the quality of clinical trials in intermittent claudication, on principle the same strict biometric and medical standards need to be applied for studies on the effect of walking exercise as for conservative drug treatment studies. However, as demonstrated by a recent review of the Cochrane Collaboration, this is apparently not always the case. The review included ten randomized controlled studies published between 1966 and 1997 with a total of 258 patients. The exercise regimes varied considerably between studies with regard to mode, intensity, frequency, duration and supervision, just as the treadmill tests employed to measure walking distance and walking time, respectively (primary endpoint). None of these studies used a confirmatory approach, and they all have a number of further methodological flaws when taking current biometric and medical guidelines into account. Also unmentioned in the abstract remains the important fact that, owing to missing data, finally only 53 patients from three studies could be pooled for the actual meta-analysis. All in all, the Cochrane review under discussion here is hardly suitable to scientifically prove the generally accepted efficacy of walking exercise in intermittent claudication. At the same time, it becomes clear that even publications of well-reputed institutions such as the Cochrane Collaboration should always be critically scrutinized. On the other hand, applying modern guidance for investigation under the aspect of evidence based medicine must not lead to a retrospective discrimination of accepted historical scientific data and thereby call into question proven therapeutic strategies.