|Home | About | Journals | Submit | Contact Us | Français|
The prevalence of stroke is increased in individuals with heart failure (HF). Stroke mechanism in HF may be cardiogenic embolism or cerebral hypoperfusion. Stroke risk increases with decreasing ejection fraction and low cardiac output is associated with hypotension and poor survival. We here examine the relationship between blood pressure level, history of stroke/TIA and HF.
We compared the prevalence of self-reported history of stroke or TIA in the REasons for Geographic And Racial Differences in Stroke (REGARDS) participants with HF (as defined by current digoxin use) and without HF. We excluded participants with atrial fibrillation or missing data. We examined the relationship between HF and history of stroke/TIA within tertiles of systolic blood pressure (SBP), adjusting for patient demographic and health characteristics.
Prevalent stroke/TIA were reported by 66 (26.3%) of 251 participants with and 1,805 (8.5%) of 21,202 participants without HF (p<0.0001). Within each tertile of SBP, the unadjusted OR (95%CI) for prior stroke/TIA among those with HF compared to those without HF (the reference group) was, 4.0 (2.8-5.8) for SBP<119.5mmHg, 2.7(1.8-3.9) for SBP≥119.5,<131.5mmHg and 2.3 (1.6-3.2) for SBP ≥131.5mmHg. After adjustment, the relationship between prior stroke/TIA and HF remained significant only within the lowest tertile of SBP (<119.5mmHg) (3.0; 1.5-6.1).
The odds of prevalent self-reported stroke/TIA are increased in participants with HF, and most markedly increased in participants with low SBP. Longitudinal data are needed to determine whether this reflects stroke/TIA secondary to thromboembolism from poor cardiac function or secondary to cerebral hypoperfusion.
Cardiac disease is a major independent risk factor for stroke, ranking third after age and hypertension. Atrial fibrillation (AF) is the cardiac disorder most frequently associated with stroke with about 105,000 strokes per year in the US.1 Heart failure (HF) ranks second as a cause of cardiogenic stroke with about 60,000 strokes per year2, although HF affects about twice the number of individuals in the population in the United States (5 million) than are affected by AF (2.3 million)3. Two community studies have shown the risk of stroke in HF to be 2 to 3 times that of the general population. In the Framingham study, the adjusted risk of stroke associated with HF was 2.1 in women and 2.7 in men.4 In a recent study in Olmsted County Minnesota, the stroke risk among those with HF was 2.9 times the control population risk over 5 years.5
Hypertension, particularly systolic hypertension is a major risk factor for stroke,6, 7 Hypertension doubles the lifetime risk of developing HF,8 but there is inconsistent data on whether hypertension is a risk factor for stroke in patients with HF: A population study did not find hypertension to be a risk factor for stroke in HF in a multivariable analysis.5 Hypertension was not a risk factor for stroke in a study of patients with left ventricular systolic dysfunction,9 and in Studies of Left Ventricular Dysfunction, hypertension imparted a small relative risk of stroke only in men (RR 1.32 95%CI: 1.04-1.69; p=0.02) but no significant risk in women.10 Two recent sub-studies of the Digitalis Investigation Group trial11, 12 however, showed a history of hypertension to be an independent predictor of stroke (adjusted odds ratio 1.46; 95% CI:1.11-1.94)12 and a risk factor for hospitalization for stroke (hazard ratio 1.52; 95%CI:1.11-2.08) among patients with HF.
Reduced ejection fraction has been found to be a risk factor for stroke in several studies,9, 10, 13, 14but there are some inconsistencies and an interaction with cerebrovascular disease may be important for the development of stroke in patients with low ejection fraction.2 Both hypotension15 and low ejection fraction16 have been found to be risk factors for cognitive impairment in patients with HF, and cognitive impairment may signal undiagnosed or “whispering” stroke.17 Decreasing blood pressure is associated with worse survival in HF, as hypotension probably reflects poor cardiac pump function in advanced HF.18 This study was initiated to investigate the association between blood pressure level, prevalent stroke and HF.
The REasons for Geographic And Racial Differences in Stroke (REGARDS) study is a population-based study of adults aged 45 years and older in the United States19, seeking to determine causes of racial and geographic differences in stroke. Enrollment began in January, 2003, and was completed in October, 2007. The cohort consists of 30,228 participants, 45% men, 55% women (50/50 planned); 58% white, 42% black (50/50 planned); 56% residing in the Stroke Belt region of southeastern US, and 44% in the remaining 40 contiguous US (50/50 planned), each of whom is being followed biannually by telephone to ascertain cognitive assessment, stroke and cardiovascular outcomes.19
REGARDS is approved by the Institutional Review Boards of all participating institutions. Participants were recruited from commercially-available lists of U.S. residents using mail and telephone contact methods. Those who agreed to participate answered demographic, quality of life, health behavior, and medical history information; reported HF symptoms (two questions); and reported stroke history and symptoms using the Questionnaire for Verifying Stroke-free Status (QVSFS) 20 all via a computer-assisted telephone interview. During a subsequent home visit, written informed consent was obtained, as well as blood and urine samples, electrocardiogram, and blood pressure and body mass index measures. Medication audits also were performed during the home visit. Further methodological details are available elsewhere19.
We used current use of digoxin at baseline as the proxy for HF diagnosis. The only current clinical indications for digoxin are AF and HF and digoxin use has a specificity of about 99% and a sensitivity of about 28% for the diagnosis of HF.21
The QVSFS20 contains eight items. The first two items elicit history of physician-verified stroke, mini-stroke, or TIA; a positive response on either of these items indicates a positive stroke/TIA history.
Age, sex, race, and region were ascertained by self report.
(SES) was represented by income and education levels. Annual income was categorized into three levels: less than $25K, $25K and above, and unreported. Education was categorized into two levels: high school or below, and greater than high school.
Alcohol and smoking histories were defined as follows: alcohol consumption—dichotomized as current vs. previous or no (lifelong abstinence) use; smoking—current vs. never or past smoker.
The following definitions were used: diabetes—fasting glucose ≥126 ml/dL, non-fasting glucose ≥200 ml/dL, or self-reported use of diabetes medications; hypertension—systolic blood pressure (SBP) ≥140 mmHg or diastolic blood pressure≥90 mmHg (average of two BP measurements), or self-reported use of hypertension medications; atrial fibrillation (AF)—based on electrocardiographic diagnosis, or a positive response to the question “Has a doctor or other health professional ever told you that you had atrial fibrillation?”; ischemic heart disease, based on self-reported history of myocardial infarction (MI), coronary bypass, angioplasty, or stenting, or based on electrocardiographic evidence of MI; hemoglobin level was determined from blood assays.
Descriptive statistics were computed overall and for those with and without HF. Logistic regression models were fitted to examine the relationship between HF and history of stroke/TIA. Incremental multivariable models were assessed, with covariates entered into the models in groups. Groups of variables are defined as follows: demographics – age, sex, race, region; SES factors – education, income; health behaviors – alcohol use, smoker and comorbidities – diabetes, hypertension, heart disease and hemoglobin. Frequency of stroke/TIA history among those with and without HF was compared within each tertile of SBP.
The relationship between HF and prior stroke/TIA was examined as a function of SBP tertile. The tertiles were defined as: t1, SBP < 119.5 (n=7222); t2, 119.5 < SBP < 131.5 (n=7896); t3, 131.5 < SBP (n=8192). After assessing the interaction between SBP tertile and HF, analyses were stratified by SBP tertile, and the relationship was examined prior to and after adjustment for all factors in the final model above. The SBP tertile analysis was repeated using only subjects with a history of hypertension.
Table 1 presents the description of the population, both overall and by HF status. Of the 30,228 participants recruited to REGARDS, 223 were excluded due to no information regarding prior stroke, 3691 were excluded due to no reported income information, an additional 2920 participants were excluded due to missing one or more covariates, and 1942 participants with atrial fibrillation were excluded. The final sample size for analysis was 21,453.
Among this community-dwelling cohort, prevalent stroke/TIA was present in 26.3% and 8.5% of participants with and without HF, respectively (p<0.0001). Table 2 presents the results of the incremental logistic regression modelling. The results indicate that the likelihood of prior stroke is higher among those with HF. This association was attenuated after multivariable adjustment, but remained significant with an odds ratio of 2.2; 95%CI: 1.5-3.4).
The tertile analysis (Table 3) shows that the odds of stroke/TIA are higher for participants with HF than without HF within all SBP tertiles, and the association was strongest among those in the lowest SBP tertile. Participants with HF were equally distributed between the three tertiles. In the unadjusted model, the p-value for the interaction between HF and tertile of SBP was marginally significant (0.067). Relative to those without HF, the unadjusted OR (95%CI) for prior stroke/TIA among those with HF within ascending tertiles of SBP were respectively 4.0 (2.8-5.8), 2.7(1.8-3.9) and 2.3 (1.6-3.2) which remained significant after multivariable adjustment in the lowest tertile (<119.5mmHg) (3.3; 1.6-6.5). On repeating the analysis using only subjects with a history of hypertension, the unadjusted OR (95%CI) for prior stroke/TIA among those with HF within ascending tertiles of SBP were respectively 3.9 (2.0-7.4), 2.0(1.0-3.9) and 3.2 (2.0-5.1). After multivariable adjustment, the observed associations were attenuated, although the relationship remained significant among participants in the lowest tertile: (OR:1.27; 95%CI:1.1-6.6).
Our findings confirm that after multivariable adjustment, HF (defined as patients without AF using digoxin), is associated with an increase in prevalent stroke/TIA. In HF participants, prior stroke/TIA was present 2.2 times the prevalence in participants without HF. The frequency of HF in patients with stroke (14%)22 is higher than in the general population,18 and patients with HF constitute an important subgroup of stroke patients, as they have a poor outcome and high rates of mortality and stroke recurrence.22 The increased prevalence of stroke/TIA history in participants with HF in our study (26.3%) may partly be due to the high enrollment rate of African Americans (who constitute 42% of the REGARDS population). African Americans have a much higher rate of HF than whites.23 Another factor is that transient ischemic attacks were included in our study in addition to stroke.
The association of stroke/TIA history with lower systolic blood pressure raises the possibility that cerebral hypoperfusion could be an important pathogenetic mechanism for stroke in HF patients. Vulnerability of the brain to hypoperfusion in HF is also supported by a report of increased risk of cognitive impairment15 with lower blood pressures in patients with HF. Increasing severity of HF and decreasing ejection fraction are associated with decreased cerebrovascular reactivity24 and decreased global cerebral blood flow.25 Cerebral infarcts are larger in patients with left ventricular systolic dysfunction than in controls with normal ejection fraction26 particularly distal to a high-grade stenosis27,and focal cerebral hypoperfusion distal to a stenosis could be a mechanism for an increased prevalence of stroke in patients with HF. The second SBP tertile analysis we performed, using only participants with a history of hypertension, is against cerebral hypoperfusion being an important pathogenetic mechanism in stroke in HF however. This analysis showed a lower odds ratio for stroke in the lowest SBP tertile compared with first SBP tertile analysis of participants with both normotensives and hypertensives. Hypertensives have altered cerebrovascular reactivity and require a higher systolic blood pressure to maintain cerebral blood flow than normotensives,28 and this makes them more vulnerable to hypotension. We would have expected there would be an higher odds ratio of stroke/TIA in the lowest SBP tertile in hypertensives than in the population with both normotensives and hypertensives, if the pathogenesis of stroke in HF was due to cerebral hypoperfusion.
An alternative explanation is that low blood pressure selects out HF patients with poor cardiac function and low cardiac output. These patients may be at increased risk of intraventricular thrombus formation and cerebral embolism due to poor function of the left ventricle. This is supported by the finding that decreasing blood pressure is associated with increasing mortality in an analysis of pooled HF studies.18 Low blood pressure in these patients is most likely due to poor cardiac output as its association with decreased survival is independent of pharmacological treatment.18 This association of low blood pressure and poor survival in HF patients is the opposite to what is seen in the general population.29 If patients with HF and low SBP are at increased risk of cardiogenic embolism, anticoagulation treatment has the potential of reducing the risk of stroke in these patients. This is currently being investigated in the Warfarin versus Aspirin in Reduced Cardiac Ejection Fraction (WARCEF) study.30
This is the first cohort study to show a connection between lower SBP levels and risk of stroke/TIA in HF. These cross-sectional results may have potential implications for management of blood pressure in patients with HF. Pharmacological treatments for HF including angiotensin converting enzyme inhibitors and beta blockers, lower the blood pressure. Reducing systolic blood pressure to the “lowest level tolerated” has been recommended in patients with HF.31 If HF patients are at risk for hypoperfusion-related cerebral injury, excessive lowering of SBP could increase the risk of stroke, particularly in patients with prior stroke in whom the risk of stroke recurrence is higher. Further research is needed to determine whether SBP lowering in HF patients may carry a risk of stroke and if so, to attempt to determine safe lower limits for SBP in these patients.
The limitations of this study include its cross-sectional nature. Since our study was a cross-sectional analysis of a cohort study, we are unable to establish a time directionality between onset of HF and of stroke/TIA.It is possible that the association of lower SBP and prevalent stroke could reflect more aggressive blood pressure control in participants with HF following a stroke compared with HF patients without stroke. This is unlikely however since hypertension is a risk factor for recurrent stroke and remains elevated following stroke, despite antihypertensive treatment.32 Definitive determination of the directionality of the relationship between HF or hypotension and stroke will require a longitudinal study and the ongoing REGARDS and REGARDS-MI studies, which are tracking incident stroke and cardiac events and hospitalizations, will provide an appropriate context for such a study. The basing of diagnosis of HF on digoxin usage is also a limitation of the study. Use of digoxin to diagnose HF may have resulted in patients with diastolic HF being excluded from our HF group. Diastolic HF is not an established risk factor for stroke/TIA and studies reporting HF as a risk factor for stroke4, 5 have not separated HF into systolic and diastolic groups. Diagnosis of HF by digoxin usage has been found to have a very high specificity but modest sensitivity.21 The low sensitivity may have biased our results towards the null, but is unlikely to have qualitatively affected our results, because the prevalence of HF in the general population is only approximately 5% at the mean age of our study participants33 and misclassified participants with HF would only constitute a small minority of our “no HF” group. Exclusion of patients who were not using digoxin from our HF group is unlikely to have altered our study results since the rate of stroke is almost identical in patients with HF using digoxin or not using digoxin.34 A major strength of this study is the large cohort size, which has allowed for a statistically robust analysis in subgroups of SBP tertiles. The large proportion of African American participants in this study is also a strength.
The REGARDS research project is supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Department of Health and Human Service. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Neurological Disorders and Stroke or the National Institutes of Health. Representatives of the funding agency have been involved in the review of the manuscript but not directly involved in the collection, management, analysis or interpretation of the data. The authors acknowledge the participating investigators and institutions for their valuable contributions: The University of Alabama at Birmingham, Birmingham, Alabama (Study PI, Statistical and Data Coordinating Center, Survey Research Unit): George Howard DrPH, Leslie McClure PhD, Virginia Howard PhD, Libby Wagner MA, Virginia Wadley PhD, Rodney Go PhD, Monika Safford MD, Ella Temple PhD, Margaret Stewart MSPH, David Rhodes RN; University of Vermont (Central Laboratory): Mary Cushman MD; Wake Forest University (ECG Reading Center): Ron Prineas MD, PhD; Alabama Neurological Institute (Stroke Validation Center, Medical Monitoring): Camilo Gomez MD, Susana Bowling MD; University of Arkansas for Medical Sciences (Survey Methodology): LeaVonne Pulley PhD; University of Cincinnati (Clinical Neuroepidemiology): Brett Kissela MD, Dawn Kleindorfer MD; Examination Management Services, Incorporated (In-Person Visits): Andra Graham; Medical University of South Carolina (Migration Analysis Center): Daniel Lackland, DrPH; Indiana University School of Medicine (Neuropsychology Center): Frederick Unverzagt, PhD; National Institute of Neurological Disorders and Stroke, National Institutes of Health (funding agency): Claudia Moy, Ph.D.
Additional partial funding was provided by an investigator-initiated grant-in-aid from Amgen Corporation. Amgen did not have any role in the design and conduct of the study, the collection, management, analysis, and interpretation of the data, or the preparation or approval of the manuscript.