Aims/hypothesis. Dysfunction of the vascular endothelium is commonly observed in Type 2 diabetes, and endothelial function may be an important outcome for clinical trials in diabetic samples. However, the most commonly used non-invasive test of endothelial function (flow-mediated dilation [FMD]) is technically challenging to perform, and no previous studies have carefully examined the reproducibility of FMD measurements in individuals with Type 2 diabetes. In this study, we tested the hypothesis that larger day-to-day changes in insulin and glucose are associated with larger fluctuations in FMD. Methods. Ultrasound was used to measure the FMD (% change from baseline diameter) of the brachial artery in 18 healthy adults with Type 2 diabetes on three separate occasions, in the absence of changes to diet, activity level or medications. The CV and mean deviations between pairs of FMD scores in the same individual were used as the primary outcome variables.Results. The CV for FMD (29.7%) was higher than the level traditionally accepted for biochemical assays. However, this CV estimate is within the low range of published values for FMD in healthy individuals. FMD scores were not significantly correlated with glucose or insulin levels. However, subjects with the largest variability in FMD also showed the largest fluctuations in glucose (r=0.52), insulin (r=0.47) and heart rate (r=0.48) (p≤0.05). Conclusions/interpretation. FMD can be reliably measured in individuals with Type 2 diabetes, and population-specific data on reliability is critical for the design of adequately powered studies of endothelial function.Keywords Coefficient of variation · Endothelium-dependent vasodilation · Flow-mediated dilation · Glucose · Heart rate · Insulin resistance · Test-retest reliability · Type 2 diabetes mellitus