Current issues of ACP Journal Club are published in Annals of Internal Medicine


High blood pressure heightened risk for death in elderly persons

ACP J Club. 1995 Nov-Dec;123:82. doi:10.7326/ACPJC-1995-123-3-082

Source Citation

Glynn RJ, Field TS, Rosner B, et al. Evidence for a positive linear relation between blood pressure and mortality in elderly people. Lancet. 1995 Apr 1;345:825-9.



To investigate the association between blood pressure (BP) and mortality in elderly persons over 10 years of follow-up.


10-year cohort analytic study.


Community-based study in Boston.


3657 participants (81% of those identified as eligible) aged ≥ 65 years (mean age, 74 y; 62% women) who agreed to a baseline BP measurement in their homes between 1982 and 1983 and a follow-up assessment every 3 years thereafter.

Assessment of Risk Factors

Participants were classified into 5 groups according to baseline BP values. Reference groups were systolic BP (SBP), 140 to 149 mm Hg and diastolic BP (DBP), 80 to 89 mm Hg. History of myocardial infarction, stroke, or cancer; height and weight; cigarette or alcohol use; prescription and nonprescription drug use; disability; and physical activity were recorded.

Main Outcome Measures

Relative risks (RR) for all-cause mortality and cardiovascular death associated with SBP and DBP.

Main Results

During the 10-year follow-up period, 1709 people died. Among the 1223 deaths that occurred after 3 years of follow-up, lower SBP was associated with lower death rates after controlling for other risk factors (P for trend < 0.001). Compared with the reference group, participants with SBP ≥ 160 mm Hg had a 21% higher death rate (RR, 1.21; 95% CI, 1.01 to 1.44). DBP and total mortality were not associated. 813 (48%) deaths were caused by cardiovascular disease. Among the 567 cardiovascular deaths that occurred after 3 years of follow-up, linear trends in risk for death were associated with increased SBP (P for trend < 0.001) and increased DBP (P for trend = 0.006). During the first 3 years of follow-up, increased death rates were associated with low and high SBP (RR for death adjusted for other confounders, 1.48; CI, 1.07 to 2.04 and 1.40; CI, 1.02 to 1.91, respectively). DBP < 70 mm Hg was associated with increased death rates (RR, 1.34; CI, 1.02 to 1.76).


During a 10-year follow-up period, elevated systolic blood pressure was associated with increased risk for all-cause mortality occurring after 3 years in elderly persons. High death rates associated with low systolic and diastolic blood pressure occurred in the first 3 years of follow-up.

Source of funding: National Institute on Aging.

For article reprint: Dr. R.J. Glynn, Division of Preventive Medicine, Brigham and Women's Hospital, Boston, MA 02215, USA. FAX 617-734-1437.


This ambitious study by Glynn and colleagues addresses 3 issues. It confirms an increase in mortality in elderly persons with high BP, investigates whether a protective benefit is derived from low BP, and proposes a research design in which data from the initial years are dropped as potentially misleading.

In analyzing their data, the authors separated out the 486 deaths occurring in the first 3 years from the 1709 deaths that occurred in all 10 years. Their reasoning was that these early deaths might bias any eventual relation between low DBP and protection from mortality because early deaths are influenced by pre-terminal disease (and this group had more diseases and medications and greater frailty).

By eliminating deaths in the first 3 years (the authors do not explain why they chose 3 rather than 2 or 4 years), a new cohort is developed, unsullied by association with those targeted for early demise. But this is, in essence, a new cohort, and it has to be analyzed only from the 3-year start point (for a total of 7 years, not 10 years). Other authorities (1) who study hypertension in elderly persons favor the results that are obtained from a relatively short follow-up of 3 to 5 years. Glynn and colleagues propose that persons without comorbidity or frailty be stratified according to BP; this is the design used by most large intervention trial groups (e.g., the European Working Party on Hypertension in the Elderly and the Systolic Hypertension in the Elderly Program).

This study confirms what several other studies have found about BP in elderly persons: An association exists between elevated BP and increased mortality and is stronger for SBP than DBP. This supports the treatment of hypertension in elderly persons. The authors also conclude that lower BP in elderly persons is associated with higher short-term mortality, but this conclusion is difficult to translate into a clinical strategy (i.e., should BP be raised somehow to alleviate this short-term mortality?), and the value or danger of low BP in an older age group remains unresolved.

Bruce E. Johnson, MD
Michael Lawson, MD University of Kansas Medical Center Kansas City, Kansas


1. Gifford RW Jr. Cleve Clin J Med. 1995; 62:29-35.

Author's Response

Most previous short-term studies in elderly persons found U-shaped or J-shaped associations of SBP or DBP with mortality. We confirmed that these short-term relations exist, but we found strikingly different long-term relations. We presented these results separately because of the evolving associations between BP and mortality over time.