Endothelial dysfunction is an early atherosclerotic event that precedes clinical symptoms and may also render established plaque vulnerable to rupture. Noninvasive assessment of endothelial function is commonly undertaken using the flow-mediated dilation (FMD) technique. Some studies indicate that FMD possesses independent prognostic value to predict future cardiovascular events that may exceed that associated with traditional risk factor assessment. It has been assumed that this association is related to the proposal that FMD provides an index of endothelium-derived nitric oxide (NO) function. Interestingly, placement of the occlusion cuff during the FMD procedure alters the shear stress stimulus and NO dependency of the resulting dilation: cuff placement distal to the imaged artery leads to a largely NO-mediated response, whereas proximal cuff placement leads to dilation which is less NO dependent. We used this physiological observation and the knowledge that prognostic studies have used both approaches to examine whether the prognostic capacity of FMD is related to its role as a putative index of NO function. In a meta-analysis of 14 studies (>8300 subjects), we found that FMD derived using a proximal cuff was at least as predictive as that derived using distal cuff placement, despite the latter being more NO dependent. This suggests that, whilst FMD is strongly predictive of future cardiovascular events, this may not solely be related to its assumed NO dependency. Although this finding should be confirmed with more and larger studies, we suggest that any direct measure of vascular (endothelial) function may provide independent prognostic information in humans.