The results from several nonexperimental studies suggest that drug treatment of hypertension in daily clinical practice may not be as effective as in randomized controlled trials. These nonexperimental studies had limitations with regard to the selection of an appropriate control group. The objective of our study was to assess the effect of drug treatment of hypertension under circumstances of everyday medical practice on the incidence of stroke by using a prognostically comparable untreated reference group. Within two prospective, population-based cohort studies among 45,000 men and women 20 years of age and older in the Netherlands, we selected a cohort of 2,301 hypertensive subjects who either received drug treatment for hypertension (N = 1,318) or were untreated for hypertension but were determined to be "candidates" for drug treatment on the basis of their level of blood pressure and the presence of other cardiovascular risk factors (N = 983). Follow-up averaged 4.6 years and was complete for 91% of the hypertensives. Compared with untreated hypertensive subjects who were "candidates" for drug treatment, subjects who received drug treatment for hypertension had, after adjustment for potential confounders, a 39% [95% confidence interval (CI) = 3-61%] reduced risk of stroke. About 46 (95% CI = 29-599) hypertensive patients need to be treated with antihypertensive drugs for 5 years to prevent one stroke in the general Dutch population. When a prognostically comparable reference group is used, the drug treatment of hypertension under circumstances of everyday medical practice appears to be effective in the reduction of the incidence of stroke. The relative risk reduction that we found was similar to those found in randomized controlled trials.