Because antihypertensive therapy is effective in elderly patients with isolated systolic hypertension, attention has been focused on the systolic blood pressure as a predictor of cardiovascular risk. However, it is a normal diastolic pressure that separates patients with isolated systolic hypertension from those with essential hypertension. The normal diastolic and elevated systolic pressures are largely due to age-related stiffening of the aorta. An indistensible aorta causes the pressure pulse to travel faster than normal, where it is quickly reflected off the peripheral resistance. The reflected wave then returns to the central aorta in systole rather than diastole. This augments the systolic pressure further, increasing cardiac work while reducing the diastolic pressure, on which coronary flow is dependent. The potential harm of further reducing the diastolic pressure with antihypertensive therapy, especially in patients with coronary heart disease, underlies the controversial "J curve." By decreasing the blood pressure, all antihypertensive agents improve aortic distensibility, but no agents do so directly; the nitrates come the closest. Such an agent would be useful because any therapeutic increase in aortic distensibility would decrease systolic pressure without greatly reducing diastolic pressure. The problem is complicated by the suspected inaccuracy of the cuff technique in predicting the aortic diastolic pressure. New noninvasive methods to predict the aortic diastolic pressure may help in the future. At present, the combination of a high systolic and normal diastolic pressure-a widened pulse pressure-seems to be the best predictor of cardiovascular risk in patients with hypertension or heart disease. Patients with isolated systolic hypertension should be treated, but marked diastolic hypotension should be avoided.