Treatment and Control of Hypertension in the Community
A Prospective Analysis
Cross-sectional national data indicate poor levels of treatment and control of hypertension. We identified factors that prospectively predict initiation of antihypertensive therapy and attainment of blood pressure control in the community. We included all Framingham Heart Study subjects examined between 1987 and 1999 who had untreated or uncontrolled hypertension (systolic ≥140 or diastolic ≥90 mm Hg) at a baseline examination and presented for follow-up examination 4 years later. Clinical covariates were examined for their association with initiation or control at follow-up. Among 1103 hypertensive participants who were untreated at baseline, 350 (31.7%) subjects were receiving therapy at follow-up, including 25.7% of subjects with stage 1 and 51.2% of those with stage ≥2 hypertension at baseline. Multivariate predictors of initiation of therapy included higher systolic and diastolic pressure, prevalent and interim cardiovascular disease, and left ventricular hypertrophy. Other cardiovascular risk factors did not predict initiation of treatment. Among 2475 hypertensive participants who were uncontrolled (treated or untreated) at baseline, 988 (39.9%) were controlled at follow-up. Prevalent cardiovascular disease and interim initiation of therapy predicted control; older age and higher baseline systolic levels predicted lack of control. These data provide estimates of longitudinal rates of treatment and control of hypertension in the community. It appears that global risk was not taken into consideration when making decisions for initiation of therapy. Greater emphasis is needed on achieving blood pressure control in all patients but particularly among older subjects and those with systolic hypertension.
More than 50 million Americans currently have hypertension, a major risk factor for the development of cardiovascular disease (CVD) and end-organ damage. However, cross-sectional national surveillance data indicate that among hypertensive subjects, 68% are aware that they have hypertension, 54% are receiving treatment, and only 27% are controlled to goal blood pressure levels of systolic blood pressure (SBP) <140 mm Hg and diastolic blood pressure (DBP) <90 mm Hg.1 Given these disappointing statistics, it is important to understand barriers to the initiation of antihypertensive therapy and control of hypertension in the community.
In a previous cross-sectional analysis2 of treatment and control of hypertension, we found that older subjects and those with left ventricular hypertrophy (LVH) or obesity were more likely to be receiving antihypertensive therapy. In that study, these same three covariates were significantly associated with lack of overall control to goal blood pressure levels. The cross-sectional nature of the previous analysis did not allow us to determine the effect of pretreatment blood pressure levels on initiation of treatment or on attainment of blood pressure control or to draw conclusions about prospective predictors of treatment and control. In the present analysis, we sought to determine in a prospective fashion those covariates associated with (1) the initiation of antihypertensive therapy among untreated hypertensive subjects and (2) attainment of goal blood pressure levels among uncontrolled hypertensive subjects over a period of 4 years.
Study design and entry criteria for the Framingham Heart Study have been detailed elsewhere.3,4 All examinations and procedures were approved by the Institutional Review Board of Boston Medical Center, and all participants provided informed consent. For the present analysis, we included all subjects who were examined between 1987 and 1999 and were seen at a baseline and a follow-up examination 4 years later. Subjects may have presented for more than one 4-year examination cycle during the study period; if so, each visit with its corresponding 4-year follow-up was used and the subject was defined as a “participant.” This method of pooling person-examinations has been shown to provide valid estimates of effect and approximates time-dependent Cox analyses.5
At each examination, blood pressure was measured twice in the left arm by a physician as described previously.2 For the analysis of initiation of antihypertensive therapy (the “initiation cohort”), we included all participants with hypertension (defined as SBP ≥140 mm Hg or DBP ≥90 mm Hg) who were not receiving antihypertensive therapy at baseline. For the analysis of blood pressure control (the “control cohort”), we included all hypertensive participants (treated or untreated) whose baseline blood pressure was uncontrolled (SBP ≥140 or DBP ≥90 mm Hg). Participants were eligible for inclusion in both the initiation and control cohorts, if appropriate. Our definitions of hypertension and blood pressure control reflect the definitions used in the Third National Health and Nutrition Examination Survey (NHANES III)6 and the Fifth and Sixth Reports of the Joint National Committee on the Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC-V7 and JNC-VI1).
For both the initiation and control cohorts, we examined the prevalence at the baseline examination of categoric clinical covariates, defined a priori. Prevalent CVD was defined as a history of angina, coronary insufficiency, myocardial infarction, stroke, or congestive heart failure. In addition, we determined the occurrence of new CVD events (“interim CVD”) in the 4-year period between the baseline and follow-up examinations. For the control cohort, we also examined the prevalence of baseline antihypertensive therapy and the incidence of initiation of antihypertensive therapy. The control cohort was composed of participants who were uncontrolled at baseline, whether treated or not. Statistical analyses were performed separately for these two subgroups and for all uncontrolled participants together.
Comparisons between rates of initiation of therapy and control to goal levels across covariate strata were performed with χ2 or Fisher’s exact tests, as appropriate. During the study period, in 1993, recommendations for the classification, treatment and control of hypertension changed with the publication of the JNC-V report.7 We therefore checked for differences in rates of initiation and control across the study period by examining these rates among participants with follow-up before 1993 compared with after 1993. A 2-tailed probability value of <0.05 was defined as statistically significant. Stepwise logistic regression was used to identify the multivariate predictors of initiation of therapy and control to goal levels. Age group was forced into the models first. The criterion for entry into the models was a probability value <0.05. These analyses were repeated after limiting the study sample to participants with blood pressure ≥160/95 at baseline to reflect the earlier definition of controlled blood pressure. Statistical analyses were performed using SAS statistical software.8
There were 1103 participants (46.1% women) in the initiation cohort and 2475 participants (52.4% women) in the control cohort. The characteristics of the two cohorts are shown in Table 1. The mean ages of the initiation and control cohorts were 57±10 years (range, 28 to 91 years) and 63±12 years (range, 28 to 95 years), respectively.
Initiation of Antihypertensive Therapy
Among the 1103 participants who were uncontrolled and untreated at baseline, 350 (31.7%) initiated antihypertensive therapy during follow-up (Table 2), including 25.7% with JNC stage 1 hypertension and 51.2% with stage ≥2 hypertension at baseline. Men and women initiated therapy at similar rates, and there were no significant differences between smokers and nonsmokers, obese and nonobese participants, or participants with elevated versus desirable cholesterol levels. Diabetics were more likely to initiate therapy than nondiabetics. Two thirds of participants with LVH, more than half of participants with prevalent CVD, and nearly 80% of participants with an interim CVD event initiated therapy.
After forcing age into the model, significant multivariate predictors of initiation of antihypertensive therapy (shown in Table 3) included stage 2 or higher SBP and DBP levels. The presence of target organ damage (LVH) and prevalent CVD also predicted initiation, and the occurrence of an interim CVD event during follow-up carried the highest odds ratio for initiation of therapy.
Blood Pressure Control
Among 2475 uncontrolled hypertensive participants at baseline, 988 (39.9%) had blood pressure levels at goal (<140/90 mm Hg) at the follow-up examination (Table 4). Overall, 44.4% (n=641) of untreated/uncontrolled participants and 33.7% (n=347) of treated/uncontrolled participants at baseline were controlled at follow-up (P<0.001). In univariate analysis, interim initiation of antihypertensive therapy predicted better control at follow-up; in contrast, the use of antihypertensive therapy at baseline was associated with lower rates of blood pressure control at follow-up. Younger age and lower levels of SBP at baseline predicted higher blood pressure control rates. Among treated uncontrolled participants, men (compared with women) and participants with LVH (compared with those without LVH) were significantly more likely to be controlled at follow-up. Current smoking, diabetes, total cholesterol level, and body mass index were not significantly associated with blood pressure control at follow-up. Neither prevalent CVD nor the occurrence of interim CVD events was associated with blood pressure control at follow-up in the two subgroups.
The significant multivariate predictors of control to goal blood pressure levels among all uncontrolled participants at baseline are shown in Table 5. There were similar findings among untreated and treated participants. Overall, participants ≥55 years of age at the baseline examination were only two-thirds as likely to be controlled at follow-up as participants <55 years of age. Older age was a stronger predictor of lack of control among treated than among untreated participants. Participants with SBP 140 to 159 mm Hg at baseline were half as likely to be controlled at follow-up as participants with SBP <140 mm Hg, whereas participants with SBP ≥160 mm Hg were only one-quarter as likely to be controlled, regardless of baseline treatment status. Among untreated hypertensive participants at baseline, those who had antihypertensive therapy initiated during follow-up were almost twice as likely to be controlled as those who did not.
Treatment and Control of Participants With Blood Pressure ≥160/95 mm Hg
Among 447 participants with untreated blood pressure ≥160/95 mm Hg at baseline, 206 (46.1%) initiated therapy by follow-up. Multivariate predictors of initiation of therapy by follow-up were prevalent CVD (OR, 2.32; 95% CI, 1.16 to 4.63) and interim CVD (OR, 7.69; 95% CI, 2.23 to 26.5). Among 851 participants (treated or untreated) with blood pressure ≥160/95 mm Hg at baseline, 565 (66.4%) had blood pressure <160/95 mm Hg at follow-up and 269 (31.6%) had blood pressure <140/90 mm Hg at follow-up. Interim initiation of therapy was the sole predictor of blood pressure <160/95 at follow-up (OR, 2.21; 95% CI, 1.52 to 3.23). Predictors of blood pressure <140/90 mm Hg at follow-up included interim initiation of therapy (OR, 2.24; 95% CI, 1.59 to 3.15), prevalent CVD (OR, 1.63; 95% CI, 1.08 to 2.44), and interim CVD (OR, 1.88; 95% CI, 1.14 to 3.08). Participants with age ≥65 years (OR, 0.57; 95% CI, 0.36 to 0.91), baseline SBP 140 to 159 mm Hg (OR, 0.46; 95% CI, 0.24 to 0.91), or baseline SBP ≥160 mm Hg (OR, 0.30; 95% CI, 0.15 to 0.58) were significantly less likely to have blood pressure <140/90 mm Hg at follow-up.
Changes in Initiation and Control During the Study Period
Overall rates of initiation of antihypertensive therapy increased significantly from the early (before 1993) to the later (1993 and after) part of the study period, from 28.9% to 43.3% (P<0.001). Among stage ≥2 hypertensives at baseline, the rate of initiation increased from 48.0% to 62.1% (P=0.06), and among stage 1 hypertensives it increased from 23.3% to 36.5% (P<0.001). However, overall rates of control to levels <140/90 mm Hg decreased from 41.3% to 36.3% (P=0.03). In the early versus later periods, rates of control were 45.8% versus 39.7% (P=0.045) among untreated hypertensives at baseline and 34.1% versus 32.7% (P=0.65) among treated hypertensives. After adjustment for age, there was no significant difference in rates of control to <140/90 mm Hg between the early and later periods.
Among untreated hypertensive subjects in the community, stage 2 hypertension (SBP ≥160 or DBP ≥100 mm Hg), the presence of LVH or prevalent CVD, or an interim CVD event predicted the initiation of antihypertensive therapy. Among subjects with uncontrolled hypertension, younger age, lower baseline SBP, LVH, prevalent CVD, and interim initiation of antihypertensive therapy predicted control to goal blood pressure (SBP <140 and DBP <90 mm Hg) at follow-up. Smoking, diabetes, total cholesterol, and body mass index did not predict initiation of antihypertensive therapy or control to goal at follow-up. The higher the SBP at baseline, the greater the likelihood of initiation of antihypertensive therapy, but the lower the likelihood of control at follow-up. During the 12 years of the study period, rates of initiation of therapy increased, but overall control of blood pressure to levels <140/90 mm Hg did not improve.
The current study identifies clinical predictors of initiation of therapy and attainment of blood pressure control over time. Recent cross-sectional studies2,9 have only been able to examine factors associated with being on antihypertensive therapy or being at goal blood pressure at a single visit. As such, these studies examined subjects who were preselected by their clinicians for therapy (or no therapy). In our previous cross-sectional analysis2 of Framingham Heart Study participants examined between 1990 and 1995, we found that clinical covariates associated with being on antihypertensive therapy included older age, LVH, and obesity. However, the same three covariates were also associated with having uncontrolled hypertension, probably because subjects with these characteristics had higher blood pressures before treatment.
Using data from NHANES III, Hyman and Pavlik9 recently identified cross-sectional covariates associated with lack of hypertension control: age ≥65 years, male sex, and no reported visits to a physician in the preceding 12 months. However, the authors examined only data related to demographics and access to health care. Our study benefited from knowledge of clinical covariates, such as antecedent blood pressure levels and the presence of target-organ damage or other risk factors, which may have influenced decisions regarding initiation of therapy or affected the ease with which goal blood pressure could be attained. Thus, the current study provides a longitudinal perspective on patterns of antihypertensive therapy that is not available from cross-sectional studies.
Our findings probably represent a “best case” scenario from the community. Framingham Heart Study participants represent a volunteer cohort who are highly compliant with their follow-up examinations. In addition, at each clinic visit, findings are reviewed with the participant and a letter is sent to the participant’s primary care physician, with specific mention made of the blood pressure measurement, presence or absence of hypertension, and electrocardiographic findings. Given this added surveillance for untreated and uncontrolled hypertension, one might expect that community physicians in Framingham would treat hypertension more readily and more aggressively. However, we found longitudinal rates of initiation of therapy and control to goal blood pressure that were suboptimal. Thus, in our cohort as in the nation, uncontrolled hypertension continues to occur even in individuals who have access to health care.9 Further longitudinal studies, unbiased by limitations of cross-sectional data, would be useful to confirm our findings in other settings and to provide targets for clinicians and policymakers to identify hypertensive patients most likely to require aggressive treatment to achieve blood pressure control.
Predictors of Initiation and Control
Among untreated subjects, stage 2 or higher SBP and DBP levels at baseline predicted initiation of therapy, consistent with the current JNC-VI guidelines. JNC-IV,10 which was published in 1988 during the initial part of our study period, used a blood pressure classification scheme and recommended treatment based largely on DBP levels and recommended initiating treatment for subjects with isolated systolic hypertension only at levels of SBP ≥160 mm Hg. In 1993 and 1997, JNC-V7 and JNC-VI1 emphasized SBP and DBP levels equally in determining blood pressure stage and need for treatment. This may have contributed to our finding that only stage 2 or higher hypertension predicted initiation of therapy, whereas stage 1 hypertension did not. Alternatively, subjects with stage 1 hypertension at a Framingham Heart Study visit may not have had consistently elevated blood pressure when seen by their physicians, making them less likely to initiate treatment.
That a higher SBP level was a strong predictor of lack of blood pressure control is not surprising. Subjects with higher levels of SBP at baseline will require more intensive therapy than those with lower SBP to attain goal levels. Subjects with higher SBP also tend to be older and have a longer duration of hypertension and more target organ damage, which may perpetuate systolic hypertension and make it refractory to treatment.
Older subjects initiated treatment more often than younger subjects, but older age predicted lack of blood pressure control at follow-up, particularly among treated subjects at baseline (Table 5). Thus, these data support the notion that physicians may be willing to treat hypertension but are less aggressive in attempting to attain goal blood pressure levels in older patients,11 despite their being at the highest risk for developing the sequelae of uncontrolled hypertension and despite proven efficacy of drug therapy.12–14
Both prevalent and interim CVD events were associated with higher rates of initiation of therapy during follow-up, consistent with clinical guideline recommendations on secondary prevention.15,16 LVH, a marker of target organ damage from hypertension that confers high risk for cardiovascular events,17 also predicted initiation of therapy as recommended by JNC-VI.1 Uncontrolled hypertensive subjects with prevalent CVD, an interim CVD event, or LVH initiated therapy >50% of the time during follow-up. However, each of these covariates was present in only a small minority of hypertensive subjects. Prevalent CVD and LVH also predicted control to goal, suggesting that clinicians were attempting to treat such patients more aggressively. In contrast with the results of our prior cross-sectional analysis,2 in this study LVH at baseline predicted improved control at follow-up, particularly among hypertensive subjects who were treated at baseline. The disparity between the two studies may reflect our ability to account for baseline blood pressure levels with the current study design.
Our study sample was composed of middle-aged and older white individuals. SBP tends to rise with advancing age, whereas DBP rises until the sixth decade, after which it tends to fall.6,18 Therefore, a younger, more ethnically diverse sample than we studied might include more subjects with poor diastolic control. Our findings are based on two blood pressure readings obtained at a single examination. It is therefore possible that some subjects were misclassified with regard to hypertension status. Misclassification probably would have affected subjects we classified into stage 1 more than those classified into higher stages. Furthermore, it may have caused us to underestimate rates of initiation of therapy among truly hypertensive subjects and to overestimate rates of control at follow-up. If anything, this would bias our estimates of the effect size of covariates on blood pressure control toward the null.
Our data suggest that clinicians in the community were initiating antihypertensive therapy more frequently in recent years, and they were attempting to focus efforts on initiation and control in patients with CVD or target-organ damage, who would benefit from secondary prevention. In addition, therapy was initiated in 51.2% of subjects with JNC stage 2 or higher hypertension. However, it should be recognized that whereas the absolute and relative risks of CVD are greater in stage 2 and higher hypertension, stage 1 contains the vast majority of hypertensive individuals, so the population-attributable risk for CVD is higher for stage 1.19,20 Thus, a suboptimal number of events would be prevented when stage 1 hypertension is undertreated and inadequately controlled.
The historical preoccupation with DBP may have created the misperception that treatment of hypertension requires only control of the diastolic component. Consequently, older patients with elevated SBP, in whom the DBP often is normal,6,18 may be undertreated because they are perceived to be adequately controlled. Our data add further weight to the growing appreciation that in older patients SBP is the key blood pressure component on which risk stratification and treatment decisions should be made. Isolated systolic hypertension is the most common form of hypertension;21 SBP correctly classifies JNC stage far better than DBP among high-normal and hypertensive subjects;21,22 SBP is at least as strong a risk factor as DBP, and often stronger, in prediction of adverse CVD outcomes;20,23,24 and large clinical trials12,14 have demonstrated substantial benefit with treatment of isolated systolic hypertension in patients ≥60 years old. The Coordinating Committee of the National High Blood Pressure Education Program recently published a clinical advisory statement recommending that SBP become the major criterion for the diagnosis, staging, and management of hypertension in middle-aged and older patients, who represent the vast majority of hypertensives.11
Although prevalent CVD and LVH predicted both initiation and control at follow-up, other CVD risk factors (smoking, diabetes, cholesterol level, and body mass index) did not. Therefore, in contrast to JNC-V and JNC-VI recommendations,1,7 it appears that global CVD risk was not fully considered when making treatment decisions for patients with hypertension. To assist clinicians and improve cost-effectiveness of treatment, greater emphasis should be placed on global risk assessment in public health efforts and in future clinical practice guidelines on hypertension. Furthermore, we observed no improvement in rates of control during the study period, indicating that greater emphasis is needed on attaining goal blood pressure, not just initiating therapy.
Dr Lloyd-Jones was supported by grant K23 HL04253 from the National Institutes of Health, Bethesda, Md.
- Received May 3, 2002.
- Revision received May 22, 2002.
- Accepted August 22, 2002.
Lloyd-Jones DM, Evans JC, Larson MG, O’Donnell CJ, Roccella EJ, Levy D. Differential control of systolic and diastolic blood pressure: factors associated with lack of blood pressure control in the community. Hypertension. 2000; 36: 594–599.
Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families: the Framingham offspring study. Am J Epidemiol. 1979; 110: 281–290.
Burt VL, Whelton P, Roccella EJ, Brown C, Cutler JA, Higgins M, Horan MJ, Labarthe D. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988–1991. Hypertension. 1995; 25: 305–313.
SAS Institute. SAS/STAT Software: Changes and Enhancements Through Release 6.11. Cary, NC: SAS Institute Inc; 1996.
The 1988 Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med. 1988; 1988: 148: 1023–1038.
Izzo JL, Levy D, Black HR. Importance of systolic blood pressure in older Americans. Hypertension. 2000; 35: 1021–1024.
Staessen JA, Fagard R, Thijs L, Celis H, Arabidze GG, Birkenhager WH, Bulpitt CJ, de Leeuw PW, Dollery CT, Fletcher AE, Forette F, Leonetti G, Nachev C, O’Brien ET, Rosenfeld J, Rodicio JL, Tuomilehto J, Zanchetti A. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. Lancet. 1997; 350: 757–764.
Franklin SS, Gustin W, Wong ND, Larson MG, Weber MA, Kannel WB, Levy D. Hemodynamic patterns of age-related changes in blood pressure: the Framingham Heart Study. Circulation. 1997; 96: 308–315.
Franklin SS, Jacobs MJ, Wong ND, L’Italien GJ, Lapuerta P. Predominance of isolated systolic hypertension among middle-aged and elderly US hypertensives. Hypertension. 2001; 37: 869–874.
Lloyd-Jones DM, Evans JC, Larson MG, O’Donnell CJ, Levy D. Differential impact of systolic and diastolic blood pressure level on JNC-VI staging. Hypertension. 1999; 34: 381–385.
Neaton JD, Kuller L, Stamler J, Wentworth DN. Impact of systolic and diastolic blood pressure on cardiovascular mortality. In: Laragh JH, Brenner BM, eds. Hypertension: Pathophysiology, Diagnosis, and Management. 2nd ed. New York: Raven Press; 1995: 127–144.