Preventing heart disease by controlling hypertension: Impact of hypertensive subtype, stage, age, and sex

Transcription

1 Prevention and Rehabilitation Preventing heart disease by controlling hypertension: Impact of hypertensive subtype, stage, age, and sex Nathan D. Wong, PhD, a Gaurav Thakral, BS, a Stanley S. Franklin, MD, a Gil J. L Italien, PhD, b Milagros J. Jacobs, BS, a Joanna L. Whyte, MS, MPH, RD, b and Pablo Lapuerta, MD b Irvine, Calif, and Princeton, NJ Background Hypertension is related to significant morbidity and mortality rates from coronary heart disease (CHD). This report examines the relative and absolute impact on risk for CHD by controlling hypertension to high normal and optimal levels. Methods Among all subjects with untreated or inadequately treated hypertension in the National Health and Nutrition Examination Survey (NHANES) III who were 30 to 74 years of age and without prior CHD, the 10-year risk of CHD was calculated. With the use of sampling weights, the number of CHD events by age group, hypertension subtype (isolated diastolic hypertension [IDH], systolic-diastolic hypertension [SDH], and isolated systolic hypertension [ISH]), and stage of hypertension was estimated. Risk was recalculated and the number of events reestimated, assuming a reduction in blood pressure (BP) to high normal and optimal levels. The number and proportion (population-attributable risk, or PAR%) of events that could be prevented were determined from the differences in events and risk between uncontrolled and controlled BP levels. Derived from this was the number of persons needing treatment per CHD event prevented. Results Control of hypertension to high normal levels could prevent approximately one fifth (PAR 19%) of CHD events in men and one third (PAR 31%) of CHD events in women, whereas control to optimal levels may prevent 37% and 56% of CHD events, respectively (P.01 for differences between men and women). Of CHD events that could be prevented, the greatest proportion occurred from controlling BP among older persons, men, and those with stage 1 hypertension (vs stages 2 and 3) or with ISH (vs IDH or SDH). The number of persons with hypertension needing treatment to prevent one CHD event ranged from 20.5 in men to 38.6 in women when controlled to high normal BP and 10.7 in men and 21.3 in women when controlled to optimal BP. Conclusions The greatest impact from control of hypertension occurs in older persons, men, and those with ISH, whereas the greatest PAR% occurred in women. Optimal control of BP could prevent more than one third of CHD events in men and more than half of events in women. Greater efforts to control hypertension in these populations may have a substantial impact in preventing CHD events. (Am Heart J 2003;145: ) Hypertension is a major risk factor for coronary heart disease (CHD) and is associated with substantial morbidity and mortality rates. Although clinical trials have shown that lowering blood pressure (BP) reduces the risk of cardiovascular events, 1-3 only about half of individuals with hypertension are receiving treatment, and only one fourth are adequately controlled. 3,4 We From the a Heart Disease Prevention Program, University of California, Irvine, Calif, and the b Bristol Meyers-Squibb Pharmaceutical Research Institute, Princeton, NJ. Supported by a grant from the Bristol Meyers-Squibb Pharmaceutical Research Institute. Submitted Jan 3, 2002; accepted Aug 5, Reprint requests: Nathan D. Wong, PhD, Heart Disease Prevention Program, C240 Medical Sciences, University of California, Irvine, CA ndwong@uci.edu. 2003, Mosby, Inc. All rights reserved /2003/$ doi: /s (02) have previously reported that systolic hypertension, and particularly isolated systolic hypertension (ISH), confers the greatest burden on CHD, 5 especially in older persons. 6 We hypothesized that these individuals would show the greatest benefit from controlling hypertension. This report will examine the magnitude of CHD event reduction and associated population-attributable risk (PAR) from control of hypertension to highnormal and optimal BP among adults with hypertension. The extent of this benefit by age group, hypertension subtype, and stage of hypertension will also be determined, including the number needing treatment per CHD event prevented. Methods The Framingham Heart Study risk prediction algorithms7 provide a means of estimating the 10-year CHD risk based on

2 American Heart Journal Volume 145, Number 5 Wong et al 889 age, systolic and diastolic BP, total and HDL cholesterol, and the presence of diabetes and cigarette smoking. We applied the Framingham risk prediction algorithms 7 to men and women with untreated or inadequately treated hypertension ( 140 mm Hg systolic BP [SBP] or 90 mm Hg diastolic BP [DBP]) but without known CHD from the Third National Health and Nutrition Examination Survey (NHANES III), a United States population sample of noninstitutionalized adults examined with a variety of health measures performed between 1988 and BP was measured on 2 occasions, and those with CHD or who were 30 years of age or 74 years of age were not included because the Framingham algorithms used are not applicable to those groups. 7 These algorithms provided an estimate of the 10-year probability of CHD for each individual, based on Cox proportional hazards regression equations, with coefficients for each risk factor defined according to the risk factor category in which each subject was classified. In short, in this 10-year probability of CHD, P 1 (s[t]) B, where the baseline survival function s[t] using total cholesterol categories (as we have done in our analysis) for men and for women. 7 B e A, where A L G. G, utilizing total cholesterol categories, is calculated as a linear function of the product of each risk factor coefficient and the population mean value by using the equation below used to calculate L for individual values. Coefficients, as shown below, were determined by the Framingham risk algorithms. 7 G is calculated in our NHANES III study cohort to be for men and for women and is then subtracted from individual values of L for each subject, as calculated below, to obtain individualized values for A. L (for men) [ age] (if cholesterol 160) 0 (if cholesterol is ) (if cholesterol is ) (if cholesterol is ) (if cholesterol is 280) (if HDL-C is 35) (if HDL-C is 35-44) 0.0 (if HDL-C is 45-49) (if HDL-C is 50-59) (if HDL-C is 60) (if BP is optimal) 0.0 (if BP is normal) (if BP is high normal) (if BP is stage I hypertension) (if BP is stage II or higher hypertension) (if diabetes is present) 0.0 (if diabetes is not present) (if smoker) 0.0 (if not smoker). L (for women) ( age) ( Age 2 ) (if cholesterol is 160) 0 (if cholesterol is ) (if cholesterol is ) (if cholesterol is ) (if cholesterol is 280) (if HDL-C is 35) (if HDL-C is 35-44) (if HDL-C is 45-49) 0 (if HDL-C is 50-59) (if HDL-C is 60) (if BP is optimal) 0.0 (if BP is normal) (if BP is high normal) (if BP is stage I hypertension) (if BP is stage II or higher hypertension) (if diabetes is present) 0.0 (if diabetes is not present) (if smoker) 0.0 (if not smoker). Analyses were stratified by age group (30-49, years), based on our previous report using age 50 years as a useful cut point to stratify risk. 6 JNC-VI stages of BP as used above included optimal (SBP 120 mm Hg and DBP 80 mm Hg), normal (SBP mm Hg and DBP mm Hg), high normal (SBP mm Hg or DBP mm Hg), stage I hypertension (SBP mm Hg or DBP mm Hg), stage II hypertension (SBP mm Hg or higher or DBP mm Hg or higher), or stage III hypertension (SBP 180 mm Hg or higher or DBP 110 mm Hg or higher). Individuals with hypertension were divided into subtypes: isolated diastolic hypertension (IDH) (SBP 140 mm Hg and DBP 90 mm Hg or higher), systolic-diastolic hypertension (SDH) (SBP 140 mm Hg or higher and DBP 90 mm Hg or higher), and isolated systolic hypertension (ISH) (SBP 140 mm Hg or higher and DBP under 90 mm Hg). The 10-year calculated CHD risk for each group of subjects (age group, hypertension subtype, or JNC-VI stage, among men and women separately) was then multiplied by the NHANES III population size in each respective group. The NHANES III sample frequencies were weighted to the general population on the basis of their examination weights to yield the predicted number of CHD events over 10 years. Because of the stratified multistage probability design of NHANES III, in which certain subgroups were oversampled, weighting must be done according to the probability that one would have been selected from a random sample. This projects estimates to the national population distribution. The CHD risk was then recalculated by setting all hypertensive individuals to control at high normal BP levels. This involved resetting the coefficients in all hypertensive individuals to the coefficients representing high normal as shown in the above equations for L. These coefficients were obtained from the Framingham risk algorithms. 7 This process was also repeated for control of hypertension to optimal levels. Both subjects with hypertension currently untreated and those inadequately treated and receiving therapy (where BPs were 140 mm Hg systolic or 90 mm Hg diastolic) were included because control to optimal or even high normal BP levels probably would result in further reduction of CHD risk. Of those with hypertension who fit the study selection criteria, 28% of men and 42% of women were currently reported to be taking medication. The difference between the number of CHD events calculated initially and after control of hypertension represents the number of CHD events that would be prevented. The proportion of the total, original number of CHD events was then defined as the PAR%. The total number of expected CHD events was estimated for all persons with hypertension and then reestimated after control to high normal and optimal levels in separate analyses. This was repeated by age group, JNC-VI stage, and by hypertension subtype as defined above. This provided the absolute number of CHD events within each strata as well as the relative benefit in terms of PAR% associated with control of hypertension. To examine whether results would be similar among untreated persons with hypertension, results were reexamined, excluding those currently treated. Finally, to better understand potential reasons for any differences in benefit between hypertension subtype, means or prevalences of major risk factors were compared between hypertension subtypes. These risk factors included total cholesterol, HDL cholesterol, LDL cholesterol, body mass index (BMI), diabetes, and current cigarette smoking. SAS statistical software (SAS Institute, Cary, NC) was used for all these initial determinations. All calculations estimating CHD events were performed by using NHANES III sample weights. SUDAAN statistical software (Research Triangle Institute, Research Triangle Park, NC) was used to compute weighted variance estimates for

3 890 Wong et al American Heart Journal May 2003 Table I. Means/prevalences of cardiovascular risk factors for men and women with hypertension by age group, hypertension subtype, and JNC-VI stage of hypertension Risk factors (means/prevalence) Sex group Age SBP DBP BMI Total cholesterol HDL LDL Diabetes (%) Smoking (%) 50 y y y y IDH SDH ISH IDH SDH ISH JNC stage I JNC stage II JNC stage III JNC stage I JNC stage II JNC stage III *P.05 across age groups, subtypes, or stage. P.01 across age groups, subtypes, or stage. P.001 across age groups, subtypes, or stage. comparison of PAR% between men and women, across age groups, HBP stages, and HBP subtypes, and for risk factor comparisons. Finally, the number of persons with hypertension needed to treat to high normal and to optimal levels to prevent one CHD event was determined by dividing the number of persons with hypertension in each age group, stage of hypertension, and hypertension subtype for men and women by the number of events that would be prevented if treated to either high normal or to optimal levels of BP. Results Comparison of risk factor prevalence among men and women by age, hypertension subtypes, and JNC-VI staging is shown in Table I. 50 to 74 years of age compared with those 30 to 49 years of age had significantly lower DBP (P.01) and BMI (P.001) but higher total and LDL cholesterol (P.001). Older men 50 to 74 years of age had significantly higher SBP (P.01) than men 30 to 49 years of age, and among both older men and women, more had diabetes but less were cigarette smokers (P.001). Those persons with ISH were significantly older than those with SDH or IDH, had higher total and LDL cholesterol levels (in women only), and had greater likelihood of having diabetes (P.001). Persons with SDH had the highest levels of SBP, and those with IDH had the lowest levels of SBP (P.001). with stage 3 hypertension were significantly older and had higher total cholesterol and more often had diabetes, whereas women with stage 3 hypertension had lower LDL cholesterol but more often had diabetes and smoked cigarettes (P.01) compared with their counterparts with less severe hypertension. Both men and women with stage 3 hypertension had the highest levels of SBP and DBP (P.001). Table II shows the estimated population with hypertension, the 10-year CHD risk, total CHD events expected if levels of BP remain the same (uncontrolled), and number of CHD events that would be prevented by control to high normal and optimal levels of BP, by sex and age group. The corresponding proportion of total CHD events that could be prevented, or PAR%, is also shown. Whereas approximately two thirds of men and three fourths of women with hypertension were 50 to 74 years of age, approximately 80% of CHD events among men and 90% among women with hypertension could occur in this age group. Approximately twice as many CHD events (in thousands) could be prevented by controlling hypertension in men than in women (546 and 243, respectively, for

4 American Heart Journal Volume 145, Number 5 Wong et al 891 Table II. Coronary heart disease events prevented by control of blood pressure to high normal and optimal levels of blood pressure, US population year of age by sex and age group Controlled to high normal Controlled to optimal No. with hypertension (1000s) 10-y CHD risk Uncontrolled Events (1000s) (1000s) PAR% (1000s) PAR% y y Total 11, y * * y * * Total * * Figures may be approximate (rounding errors). *P.01 compared with men. Table III. Coronary heart disease events prevented by control of blood pressure to high normal and optimal levels of blood pressure, US population years of age by sex and stage of hypertension Controlled to high normal Controlled to optimal No. with hypertension (1000s) 10-y CHD risk Uncontrolled Events (1000s) (1000s) PAR% (1000s) PAR% Stage I Stage II Stage III Stage I * * Stage II * * Stage III * * Figures may be approximate (rounding errors). *P.01 compared with men. control to high normal BP and 1042 and 440, respectively, for control to optimal BP). A greater proportion (PAR%) of CHD events in women (31% for control to high normal and 56% for control to optimal BP) than in men (19% for control to high normal and 37% for control to optimal BP) (P.01 between men and women) could be prevented. This proportion remains approximately the same regardless of age in both men and women. Table III presents CHD risk, total events expected, events that could be prevented, and PAR% for men and women with hypertension controlled to high normal and to optimal, subclassified by JNC-VI stage of hypertension. The preponderance of CHD events, as well as the greatest absolute number of events (in thousands) that could be prevented by control of hypertension, occurred among persons with stage 1 hypertension (371 in men and 145 in women for control to high normal and 743 in men and 288 in women for control to optimal), despite PAR% being somewhat lower for those with stage 1 hypertension compared with stages 2 and 3 hypertension. Table IV presents CHD risk, total events expected, events that could be prevented, and PAR% for men and women with hypertension controlled to high normal and optimal, subclassified by subtype IDH, SDH, and ISH. The greatest number of CHD events (in thousands) prevented by control of hypertension occurred

5 892 Wong et al American Heart Journal May 2003 Table IV. Coronary heart disease events prevented by control of blood pressure to high normal and optimal levels of blood pressure, US Population years of age by sex and hypertensive subtype Controlled to high normal Controlled to optimal No. with hypertension (1000s) 10-y CHD risk Uncontrolled Events (1000s) (1000s) PAR% (1000s) PAR% IDH SDH ISH IDH * * SDH * * ISH * * Figures may be approximate (rounding errors). *P.01 compared with men. Table V. Comparison of population-attributable risks among entire and untreated population with hypertension for treatment to high normal and optimal levels of blood pressure Treatment to high normal Treatment to optimal Entire Untreated only Entire Untreated only Overall Age y Age y Stage I hypertension Stage II hypertension Stage III hypertension IDH SDH ISH Overall Age y Age y Stage I hypertension Stage II hypertension Stage III hypertension IDH SDH ISH in those with ISH (283 for men and 186 for women controlled to high normal levels and 553 for men and 342 for women controlled to optimal levels). More than half of the events in men that could be prevented by control of hypertension were in those with ISH, whereas three quarters of events that could be prevented in women by the control of hypertension were in those with ISH. Table V shows a comparison of PAR% for treatment to high normal and optimal BP levels by age group, stage of hypertension, and hypertension subtype for men and for women between the entire cohort with hypertension and those with hypertension not reporting treatment with hypertension medication. Little difference is seen in PAR% for either age group, stage of hypertension, and hypertension subtype for both men and for women between the entire cohort with hypertension and those not reporting treatment with hypertension medication. Table VI presents the number of persons with hyper-

6 American Heart Journal Volume 145, Number 5 Wong et al 893 Table VI. Number of persons with hypertension needed to treat to high normal levels of blood pressure per coronary heart disease event prevented by age group, stage of hypertension, and hypertensive subtype (men and women) Control to high normal blood pressure Control to optimal blood pressure All persons with hypertension Age y Age y JNC-VI stage I JNC-VI stage II JNC-VI stage III IDH SDH ISH Figures may be approximate (rounding errors). tension who need to be treated to either high normal or to optimal levels to prevent one CHD event by age group, stage of hypertension, and hypertension subtype for men and for women. The number of persons needing treatment (NNT) to high normal levels of BP to prevent one CHD event is 20.5 for men and 38.6 for women, whereas NNT to optimal levels of BP is 10.7 for men and 21.3 for women. NNT is substantially greater for younger versus older men (36.5 for years of age and 15.9 for years of age for control to high normal and 19.2 and 8.3, respectively, for control to optimal BP) as well as for younger versus older women (102 for years of age and 33 for years of age for control to high normal and 55.9 and 18.2, respectively, for control to optimal BP). The NNT is greatest for persons with stage I hypertension compared with those with stages II and III hypertension, as well as for persons with IDH compared with SDH or ISH, because those with stage I hypertension as well as those with IDH are at lower overall CHD risk than those with stages II and III hypertension or SDH or ISH. Discussion Our data suggest that among those with hypertension, one fifth of CHD events in men and nearly one third of CHD events in women could be prevented by controlling hypertension to high normal levels, and more than one third of CHD events in men and onehalf of CHD events in women could be prevented by controlling hypertension to optimal levels. With the increasing burden of CHD by age group in both men and women, the absolute number of preventable CHD events also increases with age. Whereas the absolute benefit could be greatest in men with stage 1 hypertension or ISH, the relative benefit of controlling hypertension to high normal levels, measured by PAR%, was greater among women. Clustering of risk factors Risk of cardiovascular disease does not depend solely on BP but is markedly influenced by other cardiovascular risk factors that tend to cluster with hypertension Less than 20% of persons with hypertension in the Framingham Heart Study had hypertension alone, whereas about 20% of CHD events in men and 48% in women showed clustering of 3 risk factors in a given individual. 10 In our NHANES sample of persons with hypertension but without CHD, 25% of persons had at least 2 additional risk factors (dyslipidemia, diabetes, and/or obesity) in addition to hypertension or advanced age. It is possible that the relative benefit (eg, PAR%) of treatment of hypertension would be different, perhaps greater, in those with accompanying additional risk factors such as dyslipidemia, diabetes, smoking, or obesity; however, it was beyond the scope of the current report to examine this further. Sex differences in CHD risk Although the Framingham Heart Study originally observed a consistently greater number of CHD events caused by hypertension in men compared with women at any age, 12,13 others have noted the attributable risk percent to be as great or greater in women, 14 which parallels our findings. In part, this was due to women having high ratios of total/hdl cholesterol, greater risk from diabetes, increased propensity for left ventricular hypertrophy, and an overall increased longevity compared with men. 13 Risk ratios for CHD were 2-fold greater for women with diabetes than for the men without diabetes and almost 4-fold greater than for women without diabetes. 13 Although left ventricu-

7 894 Wong et al American Heart Journal May 2003 lar hypertrophy (LVH) and increased left ventricular mass were more common in men and may in part explain their greater absolute cardiac event rates, 15 the Framingham Study has shown that women with LVH were at greater risk of cardiovascular events 16 than were men with LVH. This could partially explain why hypertension control in women might result in a greater proportion of CHD events prevented. Staging of hypertension by severity and subtype The greatest absolute benefit in terms of preventable CHD events would occur by control of persons with stage 1 hypertension to high normal BP, although more advanced stages (2 and 3) could yield a greater proportional benefit; for example, a greater PAR%. Because the greatest number of CHD events occur among persons with stage 1 ISH, the greatest number of preventable CHD events could occur from treating this population. 17,18 However, persons with SDH have more advanced hypertension (stages 2 and 3) and could show the greatest relative benefit. Similarly, Ogden et al 19 have shown that the NNT to prevent a cardiovascular event was smaller in persons with at least one additional major risk factor. Deficiencies in awareness, treatment, and control of hypertension To achieve control of hypertension in a greater percentage of individuals, intensified efforts must be made to identify persons with undetected hypertension, to treat those currently untreated, and to enhance treatment in those who are inadequately treated. A recent report indicated that nearly half of physicians had not heard of the JNC-VI guidelines and that those familiar with the guidelines tended to treat to lower thresholds. 20 However, even physicians aware of these guidelines do not treat SBP aggressively, resulting in inadequate control of persons with stage 1 hypertension or ISH. 6,21,22 These challenges will need to be surmounted before the year 2010 target goals for BP control rates of 50% to 55% can be reached. 23 This will require intensified efforts at fully implementing the JNC-VI guidelines, which include initiating lifestyle modification and pharmacologic intervention as well as greater acceptance of SBP as the predominant standard on which to base the diagnosis and treatment of hypertension. 3,24,25 Limitations There are several important considerations that may affect the validity of our findings. Persons with hypertension controlled to high normal or optimal were assumed to represent the same level of risk as those with these levels who were untreated. Those with hypertension may actually be at a higher risk, even when treated, which could underestimate the number of CHD events in this group. Therefore, the PAR% could be overestimated from our analyses. We have examined the potential impact of controlling hypertension to both high normal levels as well as to optimal levels of BP as defined by the JNC-VI 3 ; the risk of CHD death begins to rise at SBP levels 120 mm Hg and levels of DBP 80 mm Hg. 2 Whereas control to optimal levels represents an ideal situation that currently has not been achievable in the majority of persons with hypertension, control to high normal levels represents a more practical population target for BP control than the goal of achieving optimal BP levels. Furthermore, we have examined the potential impact of hypertension control in isolation of reducing other risk factors but realized that lifestyle measures normally undertaken in conjunction with pharmacologic therapy probably will benefit other risk factors as well, which could provide greater risk reduction than that accomplished solely by BP lowering. Last, although the Framingham risk equations are applicable for stages of hypertension as determined by JNC-VI, 3 projected risk reductions for each of these stages cannot differentiate hypertensive subtypes (ISH, IDH, and SDH) within the same stage, as the equations used did not use BPs as continuous variables; categoric indicators for JNC-VI stages were used instead. In addition, the Framingham equations may not be fully applicable across all ethnic groups included in our study, although a recent report demonstrates validation of the Framingham risk scores among major ethnic groups in several large, prospective studies. 26 Conclusions The largest number of projected CHD events that could be prevented from control of hypertension to high normal or optimal levels of BP occurred in older persons, men, and those with ISH. Nearly double the number of CHD events could potentially be prevented by control of hypertension to optimal compared with high normal levels. Whereas more than twice as many CHD events could be prevented from hypertension control in men compared with women, a greater proportion of CHD events would be prevented in women (30.9% for control to high normal and 55.9% for control to optimal BP) compared with men (19.2% and 36.5%, respectively). The preponderance of CHD events that would be prevented from control of hypertension would be among those with stage I hypertension as well as those with isolated systolic hypertension, which comprise the majority of persons with hypertension. Intensified efforts to identify and more effectively treat these populations will be crucial to achieve year 2010 goals for hypertension control.

8 American Heart Journal Volume 145, Number 5 Wong et al 895 References 1. Keys A. Seven Countries: a multivariate analysis of death and coronary heart disease. Cambridge: Harvard University Press; Stamler J, Stamler R, Neaton JD. Blood pressure, systolic and diastolic, and cardiovascular risks: US population data. Arch Intern Med 1993;153: Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997;157: Burt VL, Cutler JA, Higgins M, et al. Trends in the prevalence, awareness, treatment, and control of hypertension in the adult US population. Hypertension 1995;26: Franklin SS, Khan SA, Wong ND, et al. Is pulse pressure useful in predicting risk for coronary heart disease? The Framingham Heart Study. Circulation 1999;100: Franklin SS, Jacobs MJ, Wong ND, et al. Predominance of isolated systolic hypertension among middle-aged and elderly US hypertensives: analysis based on National Health and Nutrition Examination Survey (NHANES) III. Hypertension 2001;37: Wilson PWF, D Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97: Third Report on Nutrition Monitoring in the United States. Vol I. Washington, DC: US Government Printing Office; Alderman MH. Blood pressure management: individualized treatment based on absolute risk and the potential for benefit. Ann Intern Med 1993;119: Wilson PWF, Kannel WB, Silbershatz H, et al. Clustering of metabolic factors and coronary heart disease. Arch Intern Med 1999; 159: Kannel WB. Elevated systolic blood pressure as a cardiovascular risk factor. Am J Cardiol 2000;85: Kannel WB. Metabolic risk factors for coronary heart disease in women: perspective from the Framingham study. Am Heart J 1987;114: Kannel WB, Wilson PWF. Risk factors that attenuate the female coronary disease advantage. Arch Intern Med 1995;155: Anastos K, Charney P, Charon RA, et al. Hypertension in women: what is really known? s Caucus Working Group on s Health of the Society of General Internal Medicine. Ann Intern Med 1991;115: Kromholz HM, Larson M, Levy D. Sex differences in cardiac adaption to isolated systolic hypertension. Am J Cardiol 1993;72: Levy D, Larson MG, Vasan RS, et al. The progression from hypertension to congestive heart failure. JAMA 1996;275: Sagie A, Larson MG, Levy D. The natural history of borderline isolated systolic hypertension. N Engl J Med 1993;329: O Donnell CJ, Ridker PM, Glynn RJ, et al. Hypertension and borderline isolated systolic hypertension increase risk of cardiovascular disease and mortality in male physicians. Circulation 1997;95: Ogden LG, He J, Lydick E, et al. Long-term absolute benefit of lowering blood pressure in hypertensive patients according to the JNC-VI risk stratification. Hypertension 2000;35: Hyman DJ, Pavlik VN. Self-reported hypertension treatment practices among primary care physicians. Arch Intern Med 2000;160: Hyman DJ, Pavlik VN. Characteristics of patients with uncontrolled hypertension in the United States. N Engl J Med 2001;345: Chobanian AV. Control of hypertension: an important national priority. N Engl J Med 2001;345: Wright JT, Hammonds VC. Hypertension: epidemiology and contemporary management strategies. In: Wong ND, Black HR, Gardin JM, editors. Preventive cardiology. New York: McGraw Hill; Black HR, Kuller LH, O Rourke MF, et al. The first report of the Systolic and Pulse Pressure (SYPP) Working Group. J Hypertens 1999;17(5 Suppl):S Izzo JL, Levy D, Black HR. Clinical advisory statement: importance of systolic blood pressure in older Americans. Hypertension 2000; 35: D Agostino RB Sr, Grundy S, Sullivan LM, et al. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA 2001;286:180-7.