Control of Blood Pressure and Risk of Stroke Among Pharmacologically Treated Hypertensive Patients
Background and Purpose—Despite improved control of blood pressure during the last decades in the United States, a considerable proportion of treated hypertensives have not achieved target blood pressure levels. We estimated the proportion of strokes occurring among treated hypertensive patients that may be attributable to uncontrolled blood pressure.
Methods—A population-based case-control study was conducted among treated hypertensive members of Group Health Cooperative of Puget Sound. Cases were treated hypertensive patients who sustained a first fatal or nonfatal, ischemic (n=460) or hemorrhagic (n=95) stroke during 1989–1996. Controls were a random sample of stroke-free, treated hypertensive Group Health Cooperative enrollees (n=2966), similar in age to the stroke cases. Multiple measurements of blood pressure and other cardiovascular risk factors were collected from medical records. Logistic regression was used to estimate the risk of ischemic stroke and hemorrhagic stroke associated with uncontrolled blood pressure, defined as diastolic blood pressure >90 mm Hg or systolic blood pressure >140 mm Hg. The fraction of strokes attributable to uncontrolled blood pressure among treated hypertensives was calculated.
Results—Blood pressure was uncontrolled in 78% of ischemic stroke cases, 85% of hemorrhagic stroke cases, and 65% of controls. After adjustment for potential confounders, uncontrolled blood pressure among treated hypertensive patients was moderately associated with ischemic stroke (risk ratio=1.5 [95% CI, 1.2 to 1.9]) and strongly related to hemorrhagic stroke (risk ratio=3.0 [95% CI, 1.7 to 5.4]). We estimated that 27% (95% CI, 11% to 39%) of the ischemic strokes and 57% (95% CI, 26% to 75%) of the hemorrhagic strokes among treated hypertensive patients were attributable to uncontrolled blood pressure. Overall, 32% (95% CI, 14% to 45%) of all strokes were attributable to uncontrolled blood pressure.
Conclusions—A considerable proportion of incident strokes among treated hypertensive patients may be prevented by achieving control of blood pressure.
Although the rates of treatment and control of hypertension in the United States have increased over the past 3 decades, these rates appear to have leveled off during the mid 1990s.1 2 The earlier favorable trends in treatment and control of hypertension coincided with a decline in mortality from stroke and coronary heart disease. However, since 1993, the stroke rate has risen slightly, and the decline in coronary heart disease appears to have leveled off.3 Data from the third National Health and Nutrition Examination Survey (NHANES-III) suggest that only approximately half of those taking antihypertensive drugs achieve blood pressure (BP) levels at or below the treatment goal of 140/90 mm Hg.4 Studies in the United Kingdom and the Netherlands have demonstrated that the quality of control of hypertension is strongly related to the occurrence of stroke in the population.5 6 Using data from a population-based case-control study, we estimated the proportion of incident strokes occurring among treated hypertensive patients that may be attributable to uncontrolled BP in the United States.
Subjects and Methods
The setting was the Group Health Cooperative of Puget Sound (GHC) in Seattle, a large-staff model health maintenance organization serving >500 000 members in western Washington.
We used data from a population-based case-control study that was originally conducted to study the associations between antihypertensive therapies and myocardial infarction and stroke. Our methods were similar to those used in previous case-control studies.7
Cases were GHC enrollees, aged 30 to 79 years, who were treated pharmacologically for hypertension and who sustained an incident fatal or nonfatal stroke during July 1989 through December 1996. Potential cases were identified from computerized GHC hospital discharge abstracts, Washington state death files, and the billing records for GHC enrollees who received medical care or services from non-GHC providers. Controls were obtained from a companion study of risk factors for myocardial infarction at GHC.7 Controls were a random sample of GHC enrollees who were treated pharmacologically for hypertension, frequency matched to the myocardial infarction cases by sex, age, and calendar year. Each subject was assigned an index date. For the cases, the index date was the date of the stroke; for controls, the index date was a random date within the calendar year for which they had been selected as controls. In most strata based on sex, 10-year age categories, and index year, the ratio of controls to cases was >3:1. We excluded subjects (1) who were enrollees for <1 year or who had <4 visits before their index dates; (2) who had had a prior stroke; (3) who had a diagnosis of congestive heart failure; (4) whose stroke was a complication of a procedure or surgery; and (5) who had no recorded BP in their medical records within the year before the index date.
Data Collection and Definitions
Information on BP and other cardiovascular disease risk factors was abstracted from medical records and obtained from a telephone interview of consenting survivors. Abstraction of the information from the medical records was done by trained research assistants who were not blinded to case-control status but were unaware of the study hypothesis. Subjects were considered pharmacologically treated for hypertension when a recording of antihypertensive drug use for the indication of hypertension was present in the medical records and indicated that the subject was treated at the index date. Control of BP was defined according to the mean of the last 3 treated BP readings recorded in the medical records during the year before the index date. If <3 readings were available, we used either the mean of 2 readings (n=563) or a single reading (n=473) to assess the level of BP achieved by treatment.
The level of BP achieved by treatment was categorized according to the classification scheme of hypertension of the sixth report of the Joint National Committee (JNC-VI) on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.4 According to this classification, a diastolic blood pressure (DBP) <90 mm Hg and a systolic blood pressure (SBP) <140 mm Hg are considered controlled. This classification would provide the most informative results for clinicians and policy makers in the United States to assess the implications of managing hypertension according to these national practice guidelines.
Complete data were uniformly available from the medical records for case-control status and medical conditions such as pharmacological treatment of hypertension, angina, and diabetes.
In preliminary analyses of demographic and behavioral risk factor data such as smoking, physical activity, race, marital status, and education, the agreement between medical record and self-reported measures was good to excellent. Self-reported data, available for 52.4% of the subjects, were used for these variables; if not available, then data from the medical record were used. After information from these 2 sources was combined, data were missing on smoking (1.0%), physical activity (8.0%), race (2.4%), education (28.1%), cholesterol (5.0%), duration of treatment for hypertension (10.2%), and pretreatment BP (29.7%). We used an approximate Bayesian bootstrap method to impute these missing values. This multiple imputation method is a modification of the hot-deck method and takes account of the imputation variability.8 In sensitivity analyses, the results using multiply-imputed data were similar to those seen in the analysis of subjects with complete data.
All statistical tests were 2-tailed. We used stratification and logistic regression to control for potential confounding factors of the association between BP control and stroke. We used odds ratios to estimate the risk of stroke associated with the various JNC-VI stages of hypertension,4 compared with the risk in subjects with controlled BP (DBP <90 mm Hg and SBP <140 mm Hg). The population attributable fractions and their 95% CIs were calculated according to Whittemore.9 We also examined the relation between tertile of DBP (bottom tertile, DBP <82 mm Hg; middle tertile, DBP 82 to 90 mm Hg; top tertile, DBP >90 mm Hg) and SBP (bottom tertile, SBP <146 mm Hg; middle tertile, SBP 146 to 159 mm Hg; top tertile, SBP ≥159 mm Hg) and risk of stroke. All analyses were performed separately for ischemic and hemorrhagic strokes.
During 1989–1996, 791 treated hypertensive patients were hospitalized for or died out of hospital from an incident stroke. We also identified 3654 controls who were eligible. We excluded 38 cases and 149 controls who were enrolled for <1 year or had <4 visits before their index date, 110 cases and 199 controls with chronic heart failure, 17 cases and 64 controls who had no treated BP recorded in their medical records, 60 cases and 289 controls whose most recent BP reading was taken >1 year before the index date, and 23 cases for whom no determination of the type of stroke (ischemic or hemorrhagic) could be made on the basis of the available information. This analysis included 460 ischemic stroke cases, 95 hemorrhagic stroke cases, and 2966 controls.
Compared with controls, ischemic stroke cases had higher mean levels of treated SBP, pretreatment SBP, and total cholesterol (Table 1⇓). A number of risk factors, including diabetes and history of cardiovascular disease, were more prevalent among ischemic stroke cases than among controls. Compared with controls, hemorrhagic stroke cases had higher mean levels of treated SBP and DBP, were more often current smokers, and more often had a history of transient ischemic attack, but not of cardiovascular disease (Table 1⇓).
The mean BP during the year before the index date was more often uncontrolled among ischemic (77.6%) and hemorrhagic (85.3%) stroke cases than among controls (65.3%) (Table 2⇓). Uncontrolled BP was moderately associated with ischemic stroke (risk ratio [RR]=1.52 [95% CI, 1.19 to 1.94]) and strongly related to hemorrhagic stroke (RR=3.03 [95% CI, 1.69 to 5.41]). The relative risks progressively increased with the level of uncontrolled BP defined according to the JNC-VI classification of hypertension, more so for hemorrhagic stroke than for ischemic stroke (Table 2⇓).
The association between both types of stroke and uncontrolled BP did not significantly differ between men and women, between older and younger subjects (above and below the median age), or between subjects with or without a history of cardiovascular disease, with or without diabetes, or with high or low pretreatment DBP or SBP (above and below the median pretreatment BP). The results remained virtually unchanged after adjustment for pretreatment DBP, duration of treated hypertension, body mass index, educational level, marital status, race, physical activity, and current use of aspirin. Exclusion of subjects who had <3 BP readings in their medical records before the index date did not substantially change the results.
Estimation of the population attributable fractions suggests that 27% (95% CI, 11% to 39%) of ischemic and 57% (95% CI, 26% to 75%) of hemorrhagic strokes may have been attributable to uncontrolled BP. One third of the ischemic strokes and one half of the hemorrhagic strokes that were attributable to uncontrolled BP occurred among those with mildly elevated BP levels (JNC-VI stage I). Overall, 32% (95% CI, 14% to 45%) of all strokes may have been attributable to uncontrolled BP.
The association between ischemic stroke risk and tertile of BP was similar for DBP and SBP, whereas the risk of hemorrhagic stroke was more strongly associated with SBP than with DBP (Table 3⇓).
In this study we demonstrated that a large proportion of pharmacologically treated hypertensive patients in this population-based study had elevated BP. During the year before the index date, approximately 65% of the control subjects, 78% of the ischemic stroke cases, and 85% of the hemorrhagic stroke cases had a BP >90 mm Hg diastolic or 140 mm Hg systolic. Approximately 27% of incident ischemic strokes and 57% of incident hemorrhagic strokes among pharmacologically treated hypertensive patients were attributable to uncontrolled BP. Patients with mild elevations of BP accounted for approximately one third of the excess incident ischemic strokes and approximately one half of the excess incident hemorrhagic strokes.
The strengths of this observational study include the use of population-based case and control subjects, the comparable ascertainment of potential confounding factors, and the use of multiple measurements of BP to assess control of BP. Although we adjusted for potential confounding factors, residual confounding cannot be excluded because of the possibility of incomplete or inaccurate measurement of covariates and unknown or unmeasured confounding factors for which adjustment was not possible. Furthermore, the medical records abstractors were not blinded to case-control status and may have classified potential confounding factors differently for cases and controls. In approximately 30% of subjects there were no data on pretreatment BP values. Although the confounding effects of pretreatment BP levels were virtually the same in the analysis of subjects with complete data as in the analysis of subjects with missing pretreatment BP values, we cannot completely rule out the possibility that subjects with uncontrolled BP had higher pretreatment BP levels and therefore a higher risk of stroke.
While stroke risk is clearly reduced by BP treatment in clinical trials,10 11 12 the population attributable fractions that we estimated may not directly apply to other settings in which population characteristics and levels of BP control differ. The general US population includes more blacks, Hispanics, and uninsured people who are more likely to have poor BP control,1 2 13 14 and this difference may limit the generalizability of our findings. The rate of control of hypertension in the United States according to the NHANES-III survey (1988–1991)2 was, however, higher (45%) than the rate of control of hypertension among the controls in our study population (34.7%). This difference may have been due to the smaller number of BP measurements in our study (3 during 1 year) compared with the NHANES-III survey (6 during 2 visits), which would tend to overestimate the prevalence of hypertension.14a
This study not only confirms the strong association between achieved BP and risk of stroke among treated hypertensive patients5 15 16 17 but also provides separate estimates for ischemic and hemorrhagic stroke and the proportion of these strokes that may have been attributable to inadequate control of BP. SBP was more strongly associated with the risk of hemorrhagic stroke than DBP, whereas DBP and SBP were similarly associated with the risk of ischemic stroke. Several studies have found a stronger association between hypertension and hemorrhagic stroke than with ischemic stroke.18 19 20 However, several studies in Chinese populations have demonstrated a similar risk of ischemic and hemorrhagic stroke in relation to hypertension.21
Previously it was demonstrated in this same population that 15% of the incident myocardial infarctions among treated hypertensive patients were attributable to uncontrolled BP.22 Possible causes of uncontrolled BP include treatment-resistant hypertension,23 recent start of antihypertensive drug therapy, lack of access to medical care,14 suboptimal treatment by physicians, and patient noncompliance. Lack of compliance with antihypertensive drugs24 25 26 27 28 and lack of appropriate treatment29 30 31 have been identified as the most important barriers to hypertension control. Interventions directed at improvement of compliance32 and optimization of antihypertensive drug treatment33 may lead to improved BP control.
The findings of our study suggest that achieving control of BP among pharmacologically treated hypertensive patients might prevent 57% of hemorrhagic and 27% of ischemic strokes.
This study was supported in part by grants HL40628, HL43201, and HL60739 from the National Heart, Lung, and Blood Institute and grant AG09556 from the National Institute on Aging. Dr Psaty is a Merck/SER Clinical Epidemiology Fellow (sponsored by the Merck Co Foundation, Rahway, NJ, and the Society for Epidemiologic Research, Baltimore, Md). Dr Kaplan is a Howard Hughes Medical Institute Predoctoral Fellow.
- Received August 18, 1999.
- Revision received November 10, 1999.
- Accepted November 10, 1999.
- Copyright © 2000 by American Heart Association
Burt VL, Cutler JA, Higgins M, Horan MJ, Labaharthe D, Whelton P, Brown C, Roccella EJ. Trends in the prevalence, awareness, treatment, and control of hypertension in the adult US population: data from the Health Examination Surveys, 1960 to 1991. Hypertension. 1995;26:60–69.
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.
National Heart, Lung, and Blood Institute. Fact Book Fiscal Year 1996. Bethesda, Md: US Dept of Health and Human Services, National Institutes of Health; 1997.
Du X, Cruickshank K, McNamee R, Saraee M, Sourbutts J, Summers A, et al. Case-control study of stroke and the quality of hypertension control in north west England. BMJ. 1997;314:272–276.
Klungel OH, Stricker BHC, Paes AHP, Seidell JC, Bakker A, Voko Z, Breteler MB, de Boer A. Excess stroke among hypertensive men and women attributable to undertreatment of hypertension. Stroke. 1999;30:1312–1318.
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, for the Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. Lancet. 1997;350:757–764.
Manolio TA, Kronmal RA, Burke GL, O’Leary DH, Price TR. Short-term predictors of incident stroke in older adults: the Cardiovascular Health Study. Stroke. 1996;27:1479–1486.
Bulpitt CJ, Palmer AJ, Fletcher AE, Beevers DG, Coles EC, Ledingham JG, O’Riordan PW, Petrie JC, Rajagopalan BE, Webster J. Optimal blood pressure control in treated hypertensive patients: report from the Department of Health Hypertension Care Computing Project (DHCCP). Circulation. 1994;90:225–233.
Trafford JA, Horn CR, O’Neal H, McGonigle R, Halford-Maw L, Evans R. Five year follow-up of effects of treatment of mild and moderate hypertension. BMJ (Clin Res Ed).. 1981;282:1111–1113.
How Lin C, Shimizu Y, Kato H, Robertson TL, Furonaka H, Kodama K, Fukunaga Y. Cerebrovascular disease in a fixed population of Hiroshima and Nagasaki, with special reference to relationship between type and risk factors. Stroke. 1984;15:653–660.
Stemmermann GN, Hayashi T, Resch JA, Chung CS, Reed DM, Rhoads GG. Risk factors related to ischemic and hemorrhagic cerebrovascular disease at autopsy: the Honolulu Heart Study. Stroke. 1984;15:23–28.
He J, Klag MJ, Wu Z, Whelton PK. Stroke in the People’s Republic of China, II: meta-analysis of hypertension and risk of stroke. Stroke. 1995;26:2228–2232.
Alderman MH, Budner N, Cohen H, Lamport B, Ooi WL. Prevalence of drug resistant hypertension. Hypertension. 1988;11(suppl II):II-71–II-75.
Stockwell DH, Madhavan S, Cohen H, Gibson G, Alderman MH. The determinants of hypertension awareness, treatment, and control in an insured population. Am J Public Health. 1994;84:1768–1774.
Klein LE. Compliance and blood pressure control. Hypertension. 1988;11(suppl II):II-61–II-64.
Miller NH, Hill M, Kottke T, Ockene IS. The multilevel compliance challenge: recommendations for a call to action: a statement for healthcare professionals. Circulation. 1997;95:1085–1090.