Since the late 1950s, new-onset diabetes has been a recognized adverse effect of some antihypertensive drugs. Although diagnostic criteria for diabetes mellitus have changed greatly since then, epidemiologic and physiologic studies suggest that there are differences in glucose tolerance and incident diabetes across antihypertensive drug classes. Data about incident diabetes are now available from 16 long-term, randomized, clinical trials of antihypertensive drugs. From traditional meta-analyses, the rank-ordering of the drug classes is: β-blocker thiazide diuretic > placebo ~ calcium antagonist > angiotensin-converting enzyme (ACE) inhibitor ~ angiotensin receptor blocker (ARB). Earlier studies indicated that most individuals who developed incident diabetes were "prediabetic" before beginning antihypertensive therapy, and ipso facto had increased cardiovascular risk compared with individuals with normal glucose tolerance. An earlier diagnosis of diabetes mellitus generally has little impact (over 5 years) on cardiovascular risk. The clinical importance of differential effects of antihypertensive drugs on incident diabetes is, therefore, much less clear.