Hyperinsulinaemia and hypertension commonly coexist, and a large body of evidence points to a common pathogenesis based on the presence of underlying insulin resistance (the "insulin hypothesis" of hypertension). Metformin improves insulin sensitivity in liver and muscle as its primary antihyperglycaemic mechanism of action, and intensive glycaemic management with metformin significantly reduced the risk of macrovascular diabetic complications in the UK Prospective Diabetes Study. The clinical outcome benefits in the metformin group included a significant reduction in the risk of stroke (- 41% vs + 14% with sulphonylurea or insulin treatment, p=0.032), which is well known to be highly sensitive to changes in blood pressure. Furthermore, a placebo-controlled study has shown that metformin significantly improved endothelial function, a key regulator of vascular tone and blood pressure, in type 2 diabetic patients. However, clinical studies have shown that metformin treatment is not associated with clinically relevant reductions in blood pressure in man. These apparently conflicting observations are difficult to reconcile. Either the beneficial vascular actions of metformin involve physiological systems not involved in the control of blood pressure, or counter-regulatory mechanisms prevent beneficial effects of metformin on the vasculature being translated into a clinically meaningful antihypertensive effect. Further research will be required to resolve this paradox.