Our objective was to evaluate pretreatment predictors of longevity, particu
larly blood pressure, in a large cohort of hypertensive men. During 1974 to
1976, 10,367 male hypertensive veterans (47% black) were identified at scr
eening and subsequently characterized in 32 special Veterans Administration
(VA) hypertension clinics. Their mean age was 52 years and mean blood pres
sure (BP) 154/100 mm Hg. During an average of 21 years of follow-up, 61% di
ed. Risk ratios for all-cause mortality as functions of BP and other risk f
actors are presented for each variable alone; for each variable controlling
for age, race, and BP; and for a multivariate model. We observed that when
the entire cohort was divided into deciles by systolic blood pressure (SBP
) and by diastolic blood pressure (DBP), the risk ratios for 21-year mortal
ity increased from lowest to highest decile by 178% for SBP and 16% for DBP
. When the deciles were computed separately by age group, increases from lo
west to highest decile for those less than 40 years of age were 138% for SB
P and 263% for DBP. For those over 60 years, the increases were 154% and -1
0%, respectively. Although blacks were younger and had more severe diastoli
c hypertension than whites, the risk ratios were similar within each race g
roup. Risk patterns for mean arterial pressure and pulse pressure resembled
those for SBP but had smaller gradients. Survival curves for BP groups sug
gested constant mortality rates during follow-up. Other significant observa
tions included decreasing mortality with increasing body mass index and inc
reased mortality in the Stroke Belt. We concluded that pretreatment SBP str
ongly predicted all-cause mortality during 21-year followup. For the young,
both SBP and DBP were strong predictors; for the elderly, only SBF was pre
dictive. Am J Hypertens 2000;13:724-733 (C) 2000 American Journal of Hypert
ension, Ltd.