Diuretic antihypertensive therapy is recommended as first choice by many guidelines, often in combination with β-blockers. However, such recommendations are based on relatively short-term trials, whereas treatment for hypertension is often a lifetime process. A meta-analysis of seven studies in 58 010 individuals, showed that the 'new' therapies, namely angiotensin-converting enzyme (ACE) inhibitors, angiotensin II type 1 receptor blockers (ARBs) and calcium channel blockers (CCBs) provoke less new diabetes than the conventional 'old' therapies (diuretics and β-blockers). ACE inhibitors/ARBs decreased new diabetes by 20% (P < 0.001), whereas CCBs decreased new diabetes by 16% (P < 0.001). The number needed to treat for approximately 4 years by new rather than old conventional therapy to avoid one case of new diabetes is about 60-70. Other factors contributing to increased coronary risk are increased metabolic syndrome, blood lipid changes and hypokalaemia. It is not certain whether it is the new therapy that provides protection against new diabetes or the conventional therapy that precipitates new diabetes. However, when compared with placebo, ACE inhibition by ramipril or by the ARB, candesartan, both decrease the incidence of new diabetes, raising the hypothesis that these agents actually prevent the changes leading to insulin resistance, possibly by lessening the adverse effects of angiotensin II on the endothelium. Conversely, lipid abnormalities with conventional treatment could exert adverse effects on the endothelium. Therefore endothelial changes could help to explain the benefits of 'modern' treatment compared with the defects of conventional therapy.