The prognostic value of pretreatment pulse pressure as a predictor of myocardial infarction and the relation of pulse pressure and in-treatment diastolic blood pressure reduction to myocardial infarction were investigated in a union-sponsored systematic hypertension control program. In a prospective study, 2207 hypertensive patients with a pretreatment systolic blood pressure greater than or equal to 160 mm Hg and/or diastolic pressure greater than or equal to 95 mm Hg grouped according to tertile of pulse pressure (PP1, less than or equal to 46; PP2, 47 to 62; PP3, greater than or equal to 63 mm Hg) were further stratified by the degree of diastolic fall: large (L), greater than or equal to 18; moderate (M), 7 to 17; small (S), less than or equal to 6 mm Hg. During an average follow-up of 5 years, 132 cardiovascular events (50 myocardial infarctions, 23 strokes) were observed. Myocardial infarction rates per 1000 person-years were positively related to pulse pressure (PPI, 3.5; PP2, 2.9; PP3, 7.5; PP3 versus PP1, P=.02). Wide pulse pressure was identified as a predictor of myocardial infarction (PP3 versus [PP1+PP2]: relative risk [RR]=2.2, 95% confidence interval [CI]=1.2-4.1), controlling for other known risk factors by Cox regression. A curvilinear relation (resembling a J shape) between diastolic fall and myocardial infarction was observed in patients with the widest pulse pressure, PP3 (L, 9.5; M, 3.9; S, 11.2; L versus M: RR=2.5, 95% CI=1.0-6.2; S versus M: RR=2.9, 95% CI=1.1-8.0). Even after adjusting for age, sex, race, and previous cardiovascular disease using the Mantel-Haenszel method, this relation persisted in PP3 (L versus M: RR=2.6, 95% CI=1.0-6.5; S versus M: RR=3.3, 95% CI=1.2-9.5). A wide pretreatment pulse pressure (greater than or equal to 63 mm Hg) was associated with subsequent cardiovascular complications and identified that subgroup of hypertensive patients at greatest risk of myocardial infarction from either too large or too small a fall in diastolic blood pressure.