Pulse pressure has been more strongly associated with cardiovascular outcomes, especially myocardial infarction and heart failure, than has systolic, diastolic, or mean arterial pressure in a variety of populations. Little is known, however, of the comparative effects of various classes of antihypertensive agents on pulse pressure. In retrospective analyses of the Veterans Affairs Single-Drug Therapy for Hypertension Study, we compared changes in pulse pressure with 6 classes of antihypertensive agents: 1292 men with diastolic blood pressure of 95 to 109 mm Hg on placebo were randomized to receive hydrochlorothiazide, atenolol, captopril, clonidine, diltiazem, prazosin, or placebo. Drug doses were titrated to achieve a goal diastolic blood pressure of <90 mm Hg during a 4- to 8-week medication titration phase. Pulse pressure change (placebo subtracted) was assessed from baseline to the end of the 3-month titration and 1-year maintenance. Mean baseline systolic, diastolic, and pulse pressures were 152, 99, and 53 mm Hg, respectively. Reductions in pulse pressure during titration were greater (P<0.001) with clonidine (6.7 mm Hg) and hydrochlorothiazide (6.2 mm Hg) than with captopril (2.5 mm Hg), diltiazem (1.6 mm Hg), and atenolol (1.4 mm Hg); reduction with prazosin (3.9 mm Hg) was similar to all but clonidine. After 1 year, pulse pressure was reduced significantly more (P<0.001) with hydrochlorothiazide (8.6 mm Hg) than with captopril and atenolol (4.1 mm Hg with both); clonidine (6.3 mm Hg), diltiazem (5.5 mm Hg), and prazosin (5.0 mm Hg) were intermediate. These data show that classes of antihypertensive agents differ in their ability to reduce pulse pressure. Whether these differences affect rates of cardiovascular events remains to be determined.