The present study examined the effects of caffeine on blood pressure (BP) regulation in hypertensive men during exercise. Twenty unmedicated, mild hypertensives (HT, BP = 140/90 to 160/105 mm Hg) and 12 age-matched, normotensives (NT, BP < 130/80 mm Hg) performed 30 min of extended bicycle exercise following a single dose of caffeine (3.3 mg/kg, equivalent to 2 to 3 cups of coffee) and placebo in a double-blind, cross-over design. Hemodynamic measurements were made at predrug, 40-min postdrug and during exercise. At predrug baseline, HT had significantly higher HR (67 v 57 beats/min) and BP (141/96 v 118/72 mm Hg) than NT. At postdrug baseline, caffeine increased systolic and diastolic BP, and peripheral vascular resistance (P < .01 in all cases), decreased HR (P < .05) and did not significantly change stroke volume and cardiac output for both groups. During exercise, HR response was greater on caffeine day than placebo day in HT (P < 0.05) only. Systolic BP was consistently elevated on caffeine day compared to placebo day in both groups (P < .001). Diastolic BP was elevated in HT for 30 min of exercise on caffeine day, but this pressor effect disappeared at 15 min of exercise in NT. As a result, rate-pressure products were significantly higher on caffeine days in HT at postdrug and during exercise. On caffeine day, 7 (39%) HT and 1 (8%) NT showed an excessive BP response (> 230 for systolic or > 120 for diastolic) during exercise. In conclusion, caffeine has significant hemodynamic effects on mild hypertensives at rest and during exercise. The increased rate-pressure products following caffeine during exercise place a greater workload on the heart, and abstinence from caffeine, especially before exercise, may be beneficial for persons with hypertension.