Cognitive Status, Stroke Symptom Reports, and Modifiable Risk Factors Among Individuals With No Diagnosis of Stroke or Transient Ischemic Attack in the REasons for Geographic and Racial Differences in Stroke (REGARDS) Study
Background and Purpose— Vascular disorders that increase risk for stroke may be accompanied by decrements in cognitive functioning and by stroke symptoms in the absence of diagnosed stroke or transient ischemic attack. This study evaluates relationships among cognitive status, stroke symptom reports, and cardiovascular and behavioral factors.
Methods— REasons for Geographic and Racial Differences in Stroke (REGARDS), a prospective population study of stroke incidence, assesses stroke risk with telephone interviews and in-home physicals. Excluding subjects with a history of stroke or transient ischemic attack, this analysis includes 14 566 black and white men and women ≥45 years of age. Incremental logistic models examine baseline relationships among cognitive status (Six-item Screener scores), stroke symptom reports, demographics, health behaviors, cardiovascular indices, and depressive symptoms.
Results— A history of stroke symptoms was related to impaired cognitive status after adjusting for age, gender, race, and education but not after adjusting for poor health behaviors, vascular risk factors, and depressive symptoms. Odds of experiencing a stroke symptom increased 35% with each of five modifiable factors (hypertension, diabetes, smoking, lack of exercise, depressive symptoms), and odds of cognitive impairment increased an additional 12% with each modifiable factor. Lifelong abstinence from alcohol, lack of exercise, and depressive symptoms were independently related to impaired cognitive status.
Conclusions— The increased likelihood of cognitive impairment among subjects reporting stroke symptoms in the absence of a diagnosed stroke or transient ischemic attack suggests that such symptoms are not benign and may warrant clinical evaluation that includes a cognitive assessment. Future studies that include brain imaging may clarify the etiology of these symptoms.
In the ongoing REasons for Geographic and Racial Differences in Stroke (REGARDS) study, among adults over age 45 who report never having experienced a clinical stroke or transient ischemic attack (TIA), approximately one in five report having experienced symptoms suggestive of stroke.1 We have speculated that such symptoms might represent undiagnosed small strokes.1 Understanding the relationship of self-reported stroke symptoms to global cognitive status, and whether this relationship is mediated by health behaviors and cardiovascular risk factors, could help delineate the clinical significance of such symptoms and might suggest areas for intervention to improve vascular health and related cognitive functions.
Cognitive impairments and decrements over time are associated with cardiovascular diseases such as hypertension and diabetes and with cerebrovascular changes such as white matter hyperintensities that increase the risk for stroke.2–5 Recent research suggests that prevalence of cognitive impairment is associated with number and severity of vascular risk factors.3,6–8
Difficulties with temporal orientation and verbal recall are among the cognitive deficits associated with vascular risk factors for stroke. These deficits, which can be detected with brief screenings of cognitive function9 such as the Mini Mental State examination10 and the Six-item Screener,11 are related to diagnosed and preclinical vascular disorders6–8,12,13 (Wood FB, et al, unpublished data) and are predictive of vascular, neurodegenerative, and mixed dementia up to 3 years before diagnosis.9,14
This study extends previous research by capitalizing on the large, geographically dispersed, race- and gender-balanced sample of REGARDS subjects. Using stroke symptom reports as a marker of stroke risk and possible undiagnosed stroke, we posited that those reporting stroke symptoms would have worse cognitive status than those without symptoms. We further reasoned that individuals with stroke symptom reports would have more comorbid cardiovascular and behavioral risk factors than their peers without stroke symptoms and that these factors would mediate the odds of cognitive impairment. Thus, our primary aims were to examine differences in baseline cognitive status as indicated by Six-item Screener scores11 among REGARDS participants with and without self-reported stroke symptoms and to determine the relationship of these indices to modifiable risk factors before and after adjusting for demographic variables related to cognitive function and stroke risk.
REGARDS, a population-based study of adults 45 and older in the United States,15 seeks to determine causes of racial and geographic differences in stroke mortality. On completion of recruitment, the cohort will include 30 000 subjects, half from stroke-belt regions and half from other regions of the country, half white and half black, half men and half women, each of whom will have up to 4 years of follow-up data.
REGARDS is approved by the Institutional Review Boards of all participating institutions. Subjects are recruited from commercially available lists of U.S. residents using mail and telephone contacts. Demographic characteristics, health behaviors, medical history information, stroke symptom history using the Questionnaire for Verifying Stroke-free Status (QVSS),16 and cognitive screening with the Six-item Screener11 are collected through a computer-assisted telephone interview. During a home visit, written informed consent, blood and urine samples, electrocardiogram, and blood pressure and body mass index measures are obtained. Further methodological details are available elsewhere.15 Cross-sectional analyses reported in this article are drawn from baseline data of subjects enrolled between December 18, 2003, and March 1, 2006.
Cognitive Status: Six-Item Screener
Designed for either in-person or telephone administration, the Six-item Screener is a test of global cognitive function derived from the widely used Mini-Mental State examination.10 The Screener has been validated against the Mini-Mental State examination, other cognitive measures, and diagnoses of dementia and cognitive impairment–not demented in two populations: a community-based survey of 344 black adults with a second-stage formal diagnostic evaluation and a clinical sample of 651 adults (16.1% black) with the same diagnostic evaluation.11 Items assess recall and temporal orientation. Scores range from 0 to 6; a score of 4 or fewer correct indicates cognitive impairment.11
Stroke Symptoms: Questionnaire for Verifying Stroke-Free Status
The QVSS16 contains eight items. The first two items elicit history of physician-verified stroke, ministroke, or TIA; a positive response on either of these items excluded people from the current analyses. The remaining items elicit history of stroke symptoms. A positive response on ≥1 of these six stroke symptoms indicates a positive stroke symptom history.
Behaviors included in this analysis were current and past smoking habits, alcohol use, and exercise habits. Pack-years of smoking were calculated, and alcohol use and exercise were dichotomized (previous or current use of alcohol versus lifelong abstinence; any weekly exercise versus none).
Vascular Risk Factors
The following definitions were used. Diabetes—fasting glucose >126 mL/dL, nonfasting glucose >200 mL/dL, or self-reported use of diabetes medications; high cholesterol—total cholesterol >240 mL/dL; hypertension—systolic blood pressure >140 mm Hg or diastolic blood pressure >90 mm Hg (average of two blood pressure measurements), or self-reported use of hypertension medications; obesity—categorized as underweight/normal (body mass index <25 kg/m2) or overweight/obese (body mass index ≥25 kg/m2).
On December 18, 2003, the Six-item Screener was added to REGARDS baseline telephone interviews. As of March 1, 2006, all subjects who were administered the Six-item Screener and QVSS at baseline, who reported no history of stroke, mini-stroke or TIA, and who had completed the baseline home physical were included (n=14 566).
Analyses focused on the relationship of cognitive status (intact versus impaired) to stroke symptoms on the QVSS and to each of the following four variable groups: demographic variables, health behaviors, cardiovascular factors, and depressive symptoms. We predicted that impaired cognitive status would be associated with reports of any stroke symptom both before and after accounting for each variable group, and that higher numbers of stroke symptoms would be associated with increased odds of impaired baseline cognitive status. We hypothesized that three QVSS items (numbness, loss of half your vision, loss of ability to understand what people were saying) would be more weakly related to cognitive function than the remaining items, because these items have been found least sensitive and specific to stroke.18
Differences in presence of any stroke symptom, in reports of each individual stroke symptom, in health characteristics of those with impaired and intact cognitive status and in the pattern of performance on the Six-item Screener were assessed using t tests, analysis of variance, and χ2 statistics. Logistic regression models were used to determine univariate and multivariable relationships between cognitive status and stroke symptom reports (any versus none). Unadjusted and adjusted ORs with 95% CIs were estimated. For multivariable analyses, the four variable groups were entered incrementally so that the mediating effect of each on the relationship between stroke symptom reports and cognitive status could be evaluated.
Secondary analyses focused on five modifiable risk factors (hypertension, diabetes, current smoking, absence of exercise, DEP+). We examined mean number of risk factors by cognitive status, the relationships between cognitive status and total number of risk factors (0–5), and between stroke symptom status and number of risk factors. Finally, we assessed the relationship between stroke symptoms and cognitive status as a function of the number of modifiable risk factors adjusted for demographics.
The sample of 14 566 had a mean age of 65.3 years (SD, 9.3) and was 41% male, 41% black, 54% from the stroke belt, and 12% had less than a high school education. Prevalence of stroke symptoms, health behaviors, vascular risk factors, and DEP+ appears in Table 1 with probability values for differences by cognitive status. Table 2 displays the distribution of QVSS symptoms overall and by cognitive status. Presence of any stroke symptom was significantly related to impaired cognitive status, and each individual symptom but “sudden loss of half your vision” was significantly associated with cognitive impairment.
Among those with impaired cognitive status, 10 subjects had temporal orientation impairment only, 892 had recall impairment only, and 226 had both deficit types. Compared with those with recall deficits only, subjects with both deficit types were more likely to report a stroke symptom history, to report one-sided numbness, and to have diabetes (P’s<0.01), and these effects remained after accounting for subject age.
Results of logistic regression models estimating ORs for impaired cognitive status appear in Table 3. Unadjusted odds of cognitive impairment were higher for those reporting one or more stroke symptoms than those without symptoms. The OR was attenuated after adjusting for demographics and was no longer significant after adjusting for health behaviors and the remaining variable groups. All demographic variables except region of residence were significantly associated with cognitive status.
Unadjusted analyses also showed that lifetime abstinence from alcohol, more pack-years of smoking, absence of exercise, diabetes, hypertension, and DEP+ predicted worse cognitive status. Abstinence from alcohol, absence of exercise, and DEP+ remained significantly related to cognitive status after accounting for demographics and all other variables in the multivariable model (Table 3). Subjects who were older, male, black, and had less than a high school education were significantly more likely than their reference groups to have impaired cognitive status after adjustment for all other variables.
After assessing the relationship between number of modifiable risk factors (0–5) and presence of stroke symptoms, we evaluated the relationship between number of risk factors and cognitive status with adjustment for stroke symptoms. Adjusted for demographics, the odds of any stroke symptom increased 35% with each additional modifiable risk factor (OR, 1.35; 95% CI, 1.29 to 1.40). After adjusting for demographics and the presence of stroke symptoms, the odds of impaired cognitive status increased an additional 12% with each modifiable risk factor (OR, 1.12; 95% CI, 1.05 to 1.19).
Finally, we assessed the relationship between stroke symptoms and cognitive status as a function of the number of modifiable risk factors. After adjusting for demographics and the number of modifiable risk factors, those with any stroke symptom, relative to those with none, were 20% more likely to have impaired cognitive status (OR, 1.20; 95% CI, 1.03 to 1.39). With each additional risk factor up to four, an ascending proportion of subjects had impaired cognitive status, ranging from 5% of those with no risk factors to 15% of those with four or five risk factors.
This investigation identifies relationships among global cognitive status, stroke symptom reports, and behavioral and cardiovascular risk factors in a race- and gender-balanced national sample of adults without a history of documented stroke or TIA. After controlling for key demographic variables, the higher likelihood of impaired cognitive status among subjects reporting stroke symptoms but no stroke suggests that such symptoms may not be benign but in some cases may identify individuals who have sustained an undiagnosed stroke. Moreover, the fact that known correlates of stroke incidence (poor health behaviors, vascular disease, and depressive symptoms) help account for the relationship between stroke symptoms and cognitive status lends support to the contention that in many cases, stroke symptom reports actually may represent small strokes. Although it has been recognized that stroke symptoms portend a higher likelihood of future stroke, this study suggests that patients presenting with such symptoms warrant not only close monitoring for future stroke, but also evaluation for the possibility of stroke or TIA concurrent with the symptoms, including cognitive assessment.
Table 2 indicates that a fair proportion of cognitively intact subjects also report stroke symptoms. This finding likely reflects limits in the sensitivity of the Six-item Screener as well as the fact that strokes in certain locations (eg, internal capsule, parts of the thalamus, cerebellum) will present with motor or sensory symptoms of the kind tapped by the QVSS but not with cognitive findings.
Among those with cognitive impairment, the presence of both recall and orientation deficits was more strongly associated with stroke symptom reports than recall deficits alone, suggesting that individuals with both deficits might have more advanced vascular cognitive impairment with or without comorbid neurodegenerative disease. Further study of cognitive deficits and discrete stroke symptoms will be possible as we acquire additional data with more sensitive measures of vascular cognitive impairment.
The independent associations of exercise and alcohol use to cognitive status in this large and diverse population are consistent with findings from smaller and more homogenous studies as well as with meta-analytic data.19–21 Not surprisingly, depressive symptomatology had a strong independent association with impaired cognitive status. If some stroke symptom reports do represent undiagnosed small strokes, concomitant depression could be attributable to multiple mechanisms, including shared pathophysiology, and in such a scenario, controlling for depression would likely obfuscate the relationship of stroke symptom reports to cognitive functioning. In this study, the association was no longer significant after accounting for health behaviors. Even so, it is quite likely that depressed subjects had poorer health behaviors.
To target areas in which interventions might be undertaken to reduce stroke risk, we selected risk factors amenable to modification (hypertension, diabetes, current smoking, absence of exercise, DEP+) and explored whether these factors have an additive effect on the likelihood of stroke symptom reports and impaired cognitive status. After adjusting for age, gender, race, education, and region of residence, the presence of more modifiable risk factors was associated with increased odds of any stroke symptom and of impaired cognitive status. These findings demonstrate additive detrimental effects and underscore the need for interventions targeting vascular disease, health behaviors, and depression.
Limitations of this study include the cross-sectional design, which does not permit us to determine whether cognitive impairment itself—whether preceding the experience of stroke symptoms, sustained from the ostensible stroke event, or both—contributes to a lower likelihood of reporting stroke symptoms or of recognizing their significance and subsequently presenting clinically for a diagnosis of stroke or TIA. If cognitive impairment indeed lowers the likelihood of presenting with stroke symptoms, detection of clinically meaningful events may prove more challenging, but its importance is not diminished. A second study limitation is the use of a global cognitive screener that is less sensitive and specific to vascular cognitive impairment than more fine-grained measures. Furthermore, the community and clinical validation samples11 for the Six-item Screener were smaller, more homogenous with respect to region and race, and older on average than REGARDS subjects. The proportion of REGARDS subjects meeting the threshold for impaired cognitive status is lower than was found in these validation samples, and those who met this threshold in REGARDS were significantly older (mean age, 69 years) than those with intact status (mean age, 65 years). A final caveat is that statistically significant findings in a study of this magnitude are likely and must be carefully considered in terms of clinical relevance.
Despite these limitations, the study makes a substantial contribution by identifying, in a large and balanced national sample of adults, potentially modifiable correlates of cognitive impairment and stroke symptoms that together may represent undiagnosed small strokes. Longitudinal and expanded cognitive assessments in REGARDS are underway and will provide prospective data that will help to elucidate the relationship of cognitive performance trajectories to stroke risk, incidence, and mortality. Future studies that include detailed neurologic examinations and imaging may add to our understanding of the relationship among stroke risk factors, undiagnosed stroke, and cognitive function.
Sources of Funding
This research was supported by U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health.
- Received September 1, 2006.
- Revision received October 18, 2006.
- Accepted November 8, 2006.
Howard VJ, McClure LA, Meschia JF, Pulley L, Orr SC, Friday GH. High prevalence of stroke symptoms among persons without a diagnosis of stroke or transient ischemic attack in a general population: The REasons for Geographic And Racial Differences in Stroke (REGARDS) Study. Arch Intern Med. 2006; 166: 1952–1958.
Elias MF, Sullivan LM, D’Agostino RB, Elias PL, Beiser A, Au R, Seshadri S, DeCarli C, Wolf PA. Framingham Stroke Risk Profile and lowered cognitive performance. Stroke. 2004; 35: 404–409.
Laukka EJ, Jones S, Fratiglioni L, Backman L. Cognitive functioning in preclinical vascular dementia: a 6-year follow-up. Stroke. 2004; 35: 1805–1809.
Meschia JF, Brott TG, Chukwudelunzu FE, Hardy J, Brown RD, Meissner I, Hall LJ, Atkinson EJ, O’Brien PC. Verifying the stroke-free phenotype by structured telephone interview. Stroke. 2000; 31: 1076–1080.
Melchior LA, Huba GJ, Brown VB, Reback CJ. A short depression index for women. Educational and Psychological Measurement. 1993; 53: 1117–1125.
Jones WJ, Williams LS, Meschia JF. Validating the Questionnaire for Verifying Stroke-Free Status (QVSS) by neurological history and examination. Stroke. 2001; 32: 2232–2236.
Colcombe S, Kramer AF. Fitness effects on the cognitive function of older adults: a meta-analytic study. Psychol Sci. 2003; 14: 125–130.
Espeland MA, Gu L, Masaki KH, Langer RD, Coker LH, Stefanick ML, Ockene J, Rapp SR. Association between reported alcohol intake and cognition: results from the women’s health initiative memory study. Am J Epidemiol. 2005; 161: 228–238.