The effects of prolonged differences in diastolic blood pressure (DBP) on the risks of stroke and of coronary heart disease (CHD) were estimated from nine major prospective observational studies involving about 420,000 men and women who were followed up for intervals of 6-25 years. The results indicate that a prolonged difference of about 6 mm Hg in DBP was associated with approximately 37% fewer strokes and 23% fewer CHD deaths and nonfatal myocardial infarctions. The effects of equivalent reductions in DBP produced by antihypertensive drug treatment but maintained for only a few years have been estimated in several overviews of randomized trials involving a total of 30,000-40,000 patients. The results of the overviews indicate that treatment reduced the risk of stroke by about 40%, suggesting that most or all the long-term potential benefits for stroke due to lower DBP were achieved within about 3 years of beginning treatment. The risks of nonfatal myocardial infarction and CHD death may have been reduced by about 10% among patients allocated to active treatment; the 95% confidence limits for the difference ranged from about zero to about 20%. Whatever the true effect of treatment on CHD, it would appear somewhat less than the difference in risk estimated from the observational studies for a prolonged difference in DBP of the same size. This apparent shortfall in benefit may reflect a long time-course for changes in DBP to have their full effects on CHD, possible adverse side effects of the principal trial treatments, or both.