The ability to assess endothelial function non-invasively with B-mode ultrasound has lead to its widespread application in a variety of studies. However, the absolute values obtained using this approach vary considerably across studies. We studied whether technical aspects of the methodology can explain the wide variety in absolute values across studies.A literature search was performed to identify published reports on flow-mediated vasodilatation (FMD) of the brachial artery published between 1992 and 2001. Information on type of equipment (wall track/B-mode), location of the measurement (antecubital fossa/upper arm), occlusion site (lower/upper arm), occlusion duration (min), and occlusion pressure was extracted. Patient characteristics were also extracted. For the healthy populations, mean FMD varied from 0.20 to 19.2%; for the coronary heart disease (CHD) patients FMD varied from -1.3 to 14%; for subjects with diabetes mellitus FMD varied from 0.75 to 12%. Compared with occlusion at the upper arm, lower arm occlusion was related to decreased FMD (mean difference in FMD -2.47%; 95% CI 0.55-4.39). An occlusion duration of > or =4.5 min was related to an increased FMD compared with an occlusion time of < or =4 min (mean difference 1.30%; 95% CI 0.35-2.46). These findings were adjusted for other technical aspects of the methodology and for differences in risk factors between populations.Mean FMD differs widely between studies. There is a great overlap between populations (healthy, CHD, diabetics). Our findings suggest that the technical aspects of the measurements, the location, and the duration of the occlusion may explain some of these differences, whereas type of equipment, location of the measurement, and occlusion pressure do not.