Hypertension and Borderline Isolated Systolic Hypertension Increase Risks of Cardiovascular Disease and Mortality in Male Physicians
Background The objective of this study was to examine whether definite hypertension and borderline isolated systolic hypertension predict subsequent cardiovascular disease and mortality.
Methods and Results This was a prospective cohort study with a mean follow-up of 11.7 years. The subjects were a group of 18 682 apparently healthy US men, aged 40 to 84 years, participating in the Physicians' Health Study, a randomized trial of low-dose aspirin and β-carotene. The main outcome measures were total cardiovascular disease, myocardial infarction, stroke, cardiovascular death, and all-cause mortality. Hypertension was associated with substantially increased risks of total cardiovascular disease (relative risk [RR] 1.92; 95% confidence interval [CI], 1.70 to 2.18), myocardial infarction (RR, 1.78; 95% CI, 1.49 to 2.13), stroke (RR, 2.19; 95% CI, 1.78 to 2.69), and cardiovascular death (RR, 2.10; 95% CI, 1.68 to 2.63). Borderline isolated systolic hypertension was associated with significantly increased risks of cardiovascular disease (RR, 1.32; 95% CI, 1.09 to 1.59), stroke (RR, 1.42; 95% CI, 1.04 to 1.93), and cardiovascular death (RR, 1.56; 95% CI, 1.13 to 2.15), as well as a possible but nonsignificant increased risk of myocardial infarction (RR, 1.26; 95% CI, 0.95 to 1.67). Hypertension and borderline isolated systolic hypertension were associated with significantly increased risks of 41% and 22%, respectively, for all-cause mortality.
Conclusions Hypertension as well as borderline isolated systolic hypertension are associated with elevated risks of cardiovascular diseases, especially stroke and cardiovascular death. Hypertension is associated with an increased risk of myocardial infarction, and borderline isolated systolic hypertension predicts a possible but more modest increase in risk. These data add to the existing evidence that hypertension is a major cardiovascular risk factor and extend the findings to borderline isolated systolic hypertension.
Hypertension is a major risk factor for cardiovascular disease, and elevated systolic as well as diastolic blood pressures are important determinants of risk.1 2 Furthermore, treatment of isolated systolic hypertension, defined as a systolic blood pressure ≥160 mm Hg and diastolic blood pressure <90 mm Hg, significantly decreases the risks of stroke as well as of major cardiovascular disease in the elderly.3 Borderline isolated systolic hypertension, defined as a systolic blood pressure between 140 and 159 mm Hg and a diastolic blood pressure <90 mm Hg, increases with age and is present in >10% of the adult population.4 5 The available data are sparse from prospective studies of borderline isolated systolic hypertension and risks of cardiovascular disease,4 and therefore guidelines for therapeutic intervention remain inconclusive.6 The Physicians' Health Study, a large cohort of apparently healthy men, provided an opportunity to evaluate the relationship of hypertension and of borderline isolated systolic hypertension with subsequent risks of cardiovascular diseases, including myocardial infarction (MI), stroke, and cardiovascular death, as well as all-cause mortality.
Physicians' Health Study
The subjects and methods of the Physicians' Health Study have been described previously.7 Briefly, 22 071 US male physicians, aged 40 to 84 years at entry and free from prior MI, stroke, transient cerebral ischemia, and malignancy (other than nonmelanoma skin cancer), were randomly assigned, with the use of a 2×2 factorial design, to one of four treatment groups: aspirin (325 mg every other day, provided by Bristol-Myers Products), β-carotene (50 mg on alternate days, provided by BASF), both active agents, or both placebos.
At baseline, participating physicians reported current systolic and diastolic blood pressures and whether they were currently taking drug treatment for hypertension. Among the 22 071 randomized men, 3389 had missing data (2407 on blood pressure and 982 on treatment status); therefore, 18 682 formed the cohort at baseline for this investigation.
In validation studies of physicians, measured blood pressure is highly correlated with self-report of systolic (r=.72) and diastolic (r=.60) blood pressures (both P<.0001).8 Additional studies of the agreement of measured and self-reported blood pressure found a correlation similar to that for two measurements of blood pressure within a year.9
Data were also collected on a variety of other cardiovascular risk factors, including weight, cigarette smoking, diabetes mellitus, elevated cholesterol, angina pectoris, parental history of MI before age 60, use of alcohol, and exercise frequency.
On follow-up questionnaires every 6 months for the first year and annually thereafter, participants were asked whether they had experienced any cardiovascular disease event since the return of the last questionnaire. For all reported cardiovascular disease events, relevant medical records were obtained from >95% of the participants on receipt of their consent or, in the case of death, consent from the next of kin. Confirmation of reported diagnoses was made after examination of medical records and other available information by an end-points committee of physicians that included two internists, a cardiologist, and a neurologist, all blinded to treatment assignments. The diagnosis of nonfatal MI was confirmed by use of World Health Organization criteria.10 Nonfatal stroke was defined as a typical neurological deficit, sudden or rapid in onset, lasting >24 hours and attributable to a cerebrovascular event. Death due to a cardiovascular cause was documented by convincing evidence of a cardiovascular mechanism from all available sources, including death certificates and hospital records.7 Reports of MI, stroke, or cardiovascular death were considered confirmed only when corroborating information could be obtained, and all analyses are based on confirmed events.
The blinded aspirin component of the Physicians' Health Study was terminated early due primarily to the emergence of a statistically extreme 44% reduction in risk of a first MI.7 The β-carotene component of the trial continued uninterrupted and ended as planned on December 31, 1995. Data for the current analysis include all confirmed cardiovascular disease events since the start of the trial through August 1994. Morbidity follow-up was 99.2% complete, mortality follow-up was 100% complete, and the mean follow-up time was 11.7 years (range, 10.4 to 13.0 years).
Blood Pressure Categories
Participants were defined on the basis of guidelines from the fifth report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure as having either hypertension (systolic blood pressure ≥160 mm Hg, diastolic blood pressure ≥90 mm Hg, or current treatment for hypertension), borderline isolated systolic hypertension (systolic blood pressure of 140 to 159 mm Hg, diastolic blood pressure <90 mm Hg, and no treatment), or normal blood pressure (systolic blood pressure <140 mm Hg, diastolic blood pressure <90 mm Hg, and no treatment).6
The mean systolic blood pressure in this population was 126.2 mm Hg with a standard deviation of 11.8 mm Hg. Untreated participants (n=16 307) were further categorized by systolic blood pressure levels of <120 mm Hg (making up 23.5% of the total), 120 to 129 mm Hg (37.0%), 130 to 139 mm Hg (27.0%), and ≥140 mm Hg (12.5%). These categories correspond to the most recently published classification, which defines systolic blood pressure of <130 mm Hg as normal, 130 to 139 mm Hg as high normal, and ≥140 mm Hg as hypertensive.6
Age-adjusted means and prevalence rates of baseline characteristics were compared among the blood pressure categories. Incidence rates were computed for total first MI (fatal plus nonfatal), total first stroke (fatal plus nonfatal), cardiovascular death, total important cardiovascular events (incident nonfatal MI, nonfatal stroke, or cardiovascular death), and all-cause mortality. Incidence rates were calculated by dividing the number of events by person-time of follow-up for men in each category of blood pressure. Relative risks (RRs) for hypertension or for borderline isolated systolic hypertension were calculated as the ratio of the incidence rate in either hypertension category divided by the corresponding rate in the normal blood pressure category, which served as the referent. To evaluate whether there was a linear relation in categories of increasing systolic blood pressure in untreated participants, RRs for each category were similarly calculated by comparison with the lowest category (<120 mm Hg), which served as the referent. Multivariate RRs were derived from proportional hazards models.11 We tested the assumption of proportional hazards over time for participants who reported hypertension or borderline isolated systolic hypertension at baseline, and we found no significant change in RRs over time. Therefore, analyses include outcomes over the entire duration of follow-up.
For all RRs, two-sided probability values and 95% CIs were calculated. We calculated the probability values for trend by using blood pressure categories as an ordinal variable in the proportional hazards model. All models controlled for age (in years) and randomized treatment assignment. The multivariate analyses were also controlled for quartiles of body mass index, diagnosis of diabetes (yes, no), smoking status (smokes ≥20 cigarettes/d, smokes <20 cigarettes/d, smoked in the past, never smoked), parental history of MI before age 60 (yes, no), alcohol intake (five categories ranging from daily or greater to none), physical exercise (four categories ranging from five times per week or greater to less than weekly), and the presence of angina (yes, no) at baseline. We did not control for baseline lipid levels because data were missing in >10% of participants. To estimate the effect of hyperlipidemia on the risks of cardiovascular diseases, an additional multivariate model was constructed that added a history of high cholesterol (self-reported high cholesterol or measured cholesterol >260 mg/dL).
The baseline prevalences were 19.0% for hypertension (n=3561), 6.8% for borderline isolated systolic hypertension (n=1266), and 74.2% for normal blood pressure (n=13 855). None of the 1266 subjects with borderline isolated systolic hypertension were receiving drug treatment, whereas 66.7% (n=2376) of those with hypertension were receiving treatment. The mean systolic blood pressure at baseline was 137.9 mm Hg for those with hypertension, 142.3 mm Hg for those with borderline isolated systolic hypertension, and 121.7 mm Hg for those with normal blood pressure (Table 1⇓).
Men with hypertension were older than men with normal blood pressure (mean age, 57.3 versus 51.9 years, respectively). In age-adjusted analyses, those with hypertension were more likely to have a parental history of MI or to be overweight, diabetic, daily alcohol users, and physically inactive. Those with borderline isolated systolic hypertension were older than normotensive subjects (mean age, 59.0 versus 51.9 years, respectively) and more likely to be overweight, diabetic, current cigarette smokers, daily alcohol users, and physically inactive (Table 1⇑).
Hypertension and Borderline Isolated Systolic Hypertension and Risks of Cardiovascular Diseases
During an average of 138 months of follow-up, there were 657 total first MIs (fatal and nonfatal), 473 total first strokes (fatal and nonfatal), 432 total cardiovascular deaths, and 1255 total deaths.
As shown in Table 2⇓, men with hypertension were significantly more likely (age-adjusted RR, 2.17; 95% CI, 1.93 to 2.44) to develop cardiovascular disease (ie, nonfatal stroke, nonfatal MI, or cardiovascular death). In multivariate analysis, the RR for incident cardiovascular disease was 1.92 (95% CI, 1.70 to 2.18). Men with hypertension had significantly increased risks for MI (RR, 1.78; 95% CI, 1.49 to 2.13), stroke (RR, 2.19; 95% CI, 1.78 to 2.69), and cardiovascular death (RR, 2.10; 95% CI, 1.68 to 2.63), adjusting for all covariates in the full multivariate model.
Men with borderline isolated systolic hypertension were significantly more likely to develop cardiovascular disease than those with normal blood pressure (age-adjusted RR, 1.47; 95% CI, 1.22 to 1.77). In multivariate analyses, the RR for cardiovascular disease was 1.32 (95% CI, 1.09 to 1.59). Borderline isolated systolic hypertension was associated with a significantly increased risk of MI (age-adjusted RR, 1.38; 95% CI, 1.05 to 1.81), although this finding was no longer statistically significant in multivariate analysis. Those with borderline isolated systolic hypertension had a statistically significant increase in risk of stroke (RR, 1.42; 95% CI, 1.04 to 1.93) and of cardiovascular death (RR, 1.56; 95% CI, 1.13 to 2.15), after adjustment for all covariates (Table 2⇑). Adding a history of high cholesterol to the multivariate model did not substantially alter the magnitude or significance of the associations of hypertension or of borderline isolated systolic hypertension with incident cardiovascular disease.
The RRs for all-cause mortality were significantly elevated by 41% and 22% for those with hypertension and borderline isolated systolic hypertension, respectively (Table 2⇑). Although cardiovascular death was associated with blood pressure categories, there was no significant association with noncardiovascular death for hypertension (multivariate RR, 1.14; 95% CI, 0.96 to 1.35) or borderline isolated systolic hypertension (multivariate RR, 1.10; 95% CI, 0.87 to 1.38).
In the 3389 men with missing data regarding blood pressure or blood pressure treatment, the risk of incident cardiovascular disease, MI, stroke, cardiovascular death, and all-cause mortality was not significantly different from the risk for men with normal blood pressure, after adjusting for covariates in the multivariate model.
Increasing Systolic Blood Pressure and Risks of Cardiovascular Diseases
In untreated men, the risk of incident cardiovascular disease, MI, stroke, and cardiovascular death increased progressively across clinical categories of systolic blood pressure (<120, 120 to 129, 130 to 139, and ≥140 mm Hg), without evidence of a threshold or a J-shaped association (Figure⇓). Compared with those with blood pressure <120 mm Hg, the RRs of cardiovascular death among participants with blood pressure 120 to 129 mm Hg, 130 to 139 mm Hg, and ≥140 mm Hg were 1.20 (95% CI, 0.76 to 1.89), 1.82 (95% CI, 1.18 to 2.82), and 2.39 (95% CI, 1.52 to 3.75), respectively (P for trend <.0001), after adjusting for covariates in the full model. There were also positive linear relations between clinical categories of systolic blood pressure and the adjusted RR of incident cardiovascular disease, MI, and stroke (Figure⇓) as well as mortality. Similar progressive and significant increases in the adjusted RR of incident cardiovascular disease, MI, stroke, cardiovascular death, and all-cause mortality were seen across categories of diastolic blood pressure ranging from <75 to ≥90 mm Hg.
In these prospective cohort data in male physicians, both hypertension and borderline isolated systolic hypertension were associated with significantly increased risks of incident cardiovascular disease, stroke, and cardiovascular death, as well as all-cause mortality. The association with MI was significant for hypertension and of marginal significance for borderline isolated systolic hypertension, after control for cardiovascular risk factors.
Our findings for participants with hypertension at baseline are consistent with the large body of prospective data showing that elevations of systolic or of diastolic blood pressure increase risks of cardiovascular disease morbidity and mortality.1 However, our data also demonstrate a strong progressive increase in risk of MI, stroke, and cardiovascular death across strata of “normal” levels of systolic blood pressure even after controlling for confounding factors. For example, “high-normal” (130 to 139 mm Hg) systolic blood pressure in untreated participants was associated with substantially increased risks of MI (RR, 1.59; 95% CI, 1.19 to 2.12), stroke (RR, 1.87; 95% CI, 1.25 to 2.79), and cardiovascular death (RR, 1.82; 95% CI, 1.18 to 2.82) compared with systolic blood pressure <120 mm Hg (Figure⇑). Data from the Multiple Risk Factor Intervention Trial and the Chicago Heart Association Detection Project in Industry also demonstrated increases in cardiovascular mortality across normal levels of systolic as well as diastolic blood pressure.1 Similar increases in risk of stroke were reported across normal levels of systolic blood pressure in a cohort of 30 681 health professionals.9
The study findings for the association of borderline isolated systolic hypertension with all-cause mortality have not been described previously. Similar findings of a significant association between borderline isolated systolic hypertension and both incident cardiovascular disease and cardiovascular death were reported in the Framingham Study.4 Our larger physician cohort has a higher educational and socioeconomic status than the population cohort of Framingham.12 Thus, data from the Physicians' Health Study add to the small body of available evidence suggesting strong and significant associations of borderline isolated systolic hypertension with risks of stroke and cardiovascular death, and weaker but possibly significant associations with risk of MI.
Several limitations of the present study merit consideration. Self-reported measures of blood pressure and blood pressure treatment may have led to some degree of misclassification. However, existing data suggest that healthcare professionals are excellent self-reporters of blood pressure and other health measures.8 9 Furthermore, any such misclassification would be random and therefore tend to underestimate any true association between blood pressure level and disease incidence.13 Another potential source of measurement error is the possibility of “regression dilution” bias. This phenomenon in prospective observational studies is well described for measured baseline diastolic blood pressure, whereby uncorrected use of a single baseline measurement can result in systematic and substantial underestimation of the magnitude of the true association of “usual” blood pressure with subsequent risk of disease.14 No correction for potential regression dilution has been made in the current analyses with self-reported measures. The expected effect of such bias, if present, would be to underestimate any true association between blood pressure level and risks of disease.
Despite control for multiple cardiovascular risk factors, the possibility of residual confounding cannot be excluded. After adjustment for age, there were several differences in baseline risks controlled for in the analyses comparing those with normal blood pressure and either hypertension or borderline isolated systolic hypertension. Although unmeasured confounding is plausible, the likelihood in a cohort of physicians homogeneous in educational attainment and socioeconomic status is less than in the general population.
Borderline isolated systolic hypertension has been described as a discrete blood pressure category by the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure in reports dating back to 1984.5 6 In the fifth Joint National Committee report, borderline isolated systolic hypertension is classified as “stage 1 (mild) hypertension,” for which it is recommended that “antihypertensive medications should be started, especially in individuals with target organ disease and/or other known risk factors for cardiovascular disease.”6 However, the report acknowledges that “physicians may elect to withhold antihypertensive drug therapy from patients with diastolic blood pressure in the 90- to 94-mm Hg range and systolic blood pressure in the 140- to 149-mm Hg range.”6
Despite guideline recommendations that support the initiation of drug therapy for uncontrolled mild diastolic and isolated systolic hypertension, many physicians appear to have a high threshold for initiation of drug treatment. For example, British general practitioners in a recent survey were far less likely to treat older (>65 years) than younger (<45 years) patients for the same elevated levels of blood pressures. The diastolic treatment cutoff points were significantly greater than 95 (diastolic) and 160 mm Hg (systolic), and only 51% of practitioners treated isolated systolic hypertension.15 In the United States, where treatment guidelines are less conservative, additional research is required to better describe actual thresholds for treatment and the magnitude of risk in frankly hypertensive patients.
Our data suggest that untreated borderline isolated systolic hypertension, as well as high-normal levels of systolic blood pressure, confers clinically important elevations in risks of all-cause mortality, cardiovascular death, and stroke. The potential public health impact of an effective intervention requires an estimation of the actual number of affected persons who would require treatment to achieve a clinical benefit.16 Borderline isolated systolic hypertension occurred in 6.8% of our study participants, but 20% of persons in a general community develop borderline isolated systolic hypertension by the age of 70 years.4 In pooled estimates from randomized trials of elderly persons with either diastolic or systolic hypertension, treatment of small numbers of patients prevents cardiovascular events, and absolute benefits are greater in the elderly than in the young.17 However, no trials have focused on treatment of borderline isolated systolic hypertension. In our prospective cohort study, male physicians were free from important cardiovascular events at baseline, and the absolute risk of cardiovascular disease among individuals with borderline isolated systolic hypertension was ≈11.5% over the duration of the study. Therefore, a hypothetical intervention able to achieve a modest 25% risk reduction would require treatment of 35 affected participants for ≈11 years to prevent one cardiovascular event. Because the absolute risk of incident cardiovascular disease is greater in a general population containing more elderly persons, intervention with a modestly effective therapy in a general population would likely require treatment of substantially fewer than 35 affected persons to prevent one event. Such an effective treatment for borderline isolated systolic hypertension might yield absolute benefits as great as treatment of mild diastolic hypertension in younger, lower-risk persons.
In this regard, recommendations for interventions to reduce blood pressure in patients with borderline isolated systolic hypertension seem warranted. Although treatment of mild to moderate hypertension and isolated systolic hypertension confers net clinical benefits, with an acceptable side-effect profile,3 18 19 20 21 there is continued debate regarding how far to lower blood pressure with antihypertensive drugs.22 In particular, it is not yet clear whether the benefits outweigh the risks for treatment of borderline isolated systolic hypertension or of high-normal levels of systolic blood pressure. Our data support the need for randomized trials in persons with borderline isolated systolic hypertension. These data add to the existing evidence that hypertension is a major cardiovascular risk factor and extend the finding to borderline isolated systolic hypertension.
This study was supported by National Institutes of Health grant CA40360. Dr O'Donnell was supported by Institutional National Research Service Award AG00158.
- Received June 24, 1996.
- Revision received October 2, 1996.
- Accepted October 28, 1996.
- Copyright © 1997 by American Heart Association
Ascherio A, Rimm EB, Giovannucci EL, Colditz GA, Rosner B, Willett WC, Sacks F, Stampfer MJ. A prospective study of nutritional factors and hypertension among US men. Circulation. 1992;86:1475-1484.
Rose GA, Blackburn H. Cardiovascular Survey Methods. Geneva, Switzerland: World Health Organization; 1982:58. Monograph Series.
Cox DR. Regression models and life tables. J Roy Stat Soc (B). 1972;34:187-200.
Dawber TR, Meadors GF, Moore FE Jr. Epidemiological approaches to heart disease: the Framingham Study. Am J Public Health. 1951;41:279-286.
Hennekens CH, Buring JE. Epidemiology in Medicine. Boston, Mass: Little, Brown & Co; 1987:153-177.
MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease, I: prolonged differences in blood pressure—prospective observational studies corrected for the regression dilution bias. Lancet. 1990;335:765-774.
Dickerson JE, Brown MJ. Influence of age on general practitioners' definition and treatment of hypertension. BMJ. 1995;310:574.
Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA, Godwin J, Nawab Q, Taylor JO, Hennekens CH. Blood pressure, stroke, and coronary heart disease, II: short-term reductions in blood pressure—overview of randomised drug trials in their epidemiological context. Lancet. 1990;335:827-838.
Hebert PR, Fiebach NH, Eberlein KA, Taylor JO, Hennekens CH. The community-based randomized trials of pharmacologic treatment of mild-to-moderate hypertension. Am J Epidemiol. 1988;127:581-590.
Applegate WB, Pressel S, Wittes J, Luhr J, Shekelle RB, Camel GH, Greenlick MR, Hadley E, Moye L, Perry HM, Schron E, Wegener V. Impact of the treatment of isolated systolic hypertension on behavioral variables: results from the Systolic Hypertension in the Elderly Program. Arch Intern Med. 1994;154:2154-2160.