Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Cardiovascular disease risk factors and cognitive function
(USC Thesis Other)
Cardiovascular disease risk factors and cognitive function
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
CARDIOVASCULAR DISEASE RISK FACTORS AND COGNITIVE FUNCTION by Nicole M. Gatto _______________________________________________________________ A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (EPIDEMIOLOGY) December 2007 Copyright 2007 Nicole M. Gatto ii Dedication To my father, Joseph, who taught me the importance of hard work and the value of education. iii Acknowledgements I wish to express my deepest gratitude to my advisor, Wendy J. Mack, for her invaluable guidance during the last two years. I would also like to thank my committee members for the generous contribution of their time, as well as my fellow students for their inspiration and support. iv Table of Contents Dedication ii Acknowledgements iii List of Tables vi List of Figures vii Acronyms and Abbreviations viii Abstract x Chapter 1 - Review of the Literature 1 1 Introduction 1 Epidemiology of Cardiovascular Disease and Atherosclerosis 6 Cognitive Function 9 Rationale for Studying Atherosclerosis in Relation to Cognitive Function 11 Biological Basis for an Association between Atherosclerosis and Cognitive Functioning 12 Selection of Studies for Literature Review 22 Literature Review and Critique 56 Summary 59 Questions that Remain Unanswered Chapter 2 - Subclinical Atherosclerosis Contributes to Differences in Cognitive Function in Healthy Adults without Cardiovascular Disease 60 60 Abstract 61 Introduction 62 Methods 62 Study participants 63 Measurement of Subclinical Atherosclerosis 64 Measurement of Cognitive Function 66 Statistical Analysis 70 Results 74 Discussion Chapter 3 - Metabolic Syndrome and Cognitive Function in Healthy Middle- Aged and Older Adults without Diabetes 80 80 Abstract 81 Introduction 82 Methods v 82 Study participants 83 Measurements 85 Statistical Analysis 87 Results 94 Discussion Chapter 4 – Grant Proposal: Abnormalities in the Retinal Microvasculature and Cognitive Function 99 99 Research Plan 99 Study Objectives 101 Specific Aims 102 Background and Significance 102 The US population is aging and age-related health problems including cognitive dysfunction are increasing 103 CVD is pervasive in the Western world 104 CVD effects on cognitive function 106 Rationale for using the retina as a marker for analogous structures in the brain 126 Preliminary Studies/Previous Work 126 Associations between CVD risk factors and cognitive function 131 Associations between retinal microvascular abnormalities and cognitive function 132 Research Design and Methods 132 LALES Study Design 133 Subject Ascertainment 137 Sample sizes 138 Data Collection 149 Data Analysis 153 Statistical analysis plan 158 Timeline and Anticipated Difficulties 158 Timeline 159 Anticipated difficulties 161 Significance of proposal Bibliography 162 vi List of Tables Table 1. Prevalence of Fatty Streaks, Fibrous Plaques and Calcified Lesions in Abdominal Aortic Arteries of White Males aged 15-34 years in US 5 Table 2. Age-Specific Prevalence Rates of Coronary Atherosclerosis by Severity among Fatal General Aviation Accident Victims, 1980-1982 6 Table 3. Summary of 19 Studies Included in Literature Review 14 Table 4. Factor loadings of tests in the neuropsychological battery from principal components analysis 68 Table 5. Baseline Characteristics for BVAIT Subjects with Cognitive Testing 70 Table 6. Associations between Cognitive Factor Scores and Subclinical Atherosclerosis Measures from Regression Models for 504 BVAIT Subjects 73 Table 7. Associations between Subclinical Atherosclerosis and Cognitive Performance from Gender Stratified Adjusted Regression Models 74 Table 8. Baseline Characteristics for Study Subjects with Cognitive Testing by the Presence of the Metabolic Syndrome 88 Table 9. Linear Regression Models Associating the Metabolic Syndrome with Six Measures of Cognition 90 Table 10. Linear Regression Models Associating Components of the Metabolic Syndrome Individually Modeled as Continuous Variables with Cognition 91 Table 11. Linear Regression Models Associating Individual NCEP Metabolic Syndrome Characteristics (Categorical Variables) with Cognitive Function 93 Table 12. Prevalence of retinal microvascular abnormalities in non-diabetic adults from population-based studies 109 Table 13. Associations between Cognitive Factor Scores and Subclinical Atherosclerosis 128 Table 14. Characteristics of LALES Subjects with CASI-S Screening 135 Table 15. SENAS neuropsychological battery scales and abilities measured 145 vii List of Figures Figure 1. Cognitive Function Over Age 9 Figure 2. LALES Cognitive Sub-Samples 136 viii Acronyms and Abbreviations AAC – abdominal aortic calcium ABI - ankle-brachial index AD – Alzheimer’s disease ANOVA – Analysis of Variance ARIC - Atherosclerosis Risk in Communities AVR – arteriole to venule ratio BP – blood pressure BVAIT – B-Vitamin Atherothersclerosis Intervention Trial CAC - coronary artery calcium CAD – coronary artery disease CASI-S – Cognitive Abilities Screening Instrument, Short Version CDC – Centers for Disease and Control CEP - Cognitive Efficiency Profile CHS - Cardiovascular Health Study CIMT – common carotid artery intima-media thickness CVD – cardiovascular disease DR - delayed recall DSM-IV – Diagnostic and Statistical Manual of Mental Disorders, 4 th Edition EAS - Edinburgh Artery Study EBCT - electron beam computed tomography EKG - electrocardiography EVA – Edude de Vieillissement Arteriel ix HDL – high-density lipoprotein cholesterol IMT – intima-media thickness IR – immediate recall LALES - Los Angeles Latino Eye Study LDL - low-density lipoprotein cholesterol MBP – mean arterial blood pressure MCPT - maximal carotid plaque thickness MCI – mild cognitive impairment MMSE – Mini-Mental State Examination NART – National Adult Reading Test NCEP - The National Cholesterol Education Program NIH – National Institutes of Health NHIS – National Health Interview Survey OR – odds ratio PAD – peripheral artery disease PWV - pulse wave velocity RBANS - Repeatable Battery for the Assessment of Neuropsychological Status SENAS – Spanish English Neuropsychological Assessment Scales US – United States VaD – vascular dementia WISH – Women’s Isoflavone Soy Health Trial x Abstract Cardiovascular disease (CVD) is the leading cause of death and a major cause of disability in both men and women in the United States. Atherosclerosis, which is the most common pathological process underlying CVD, may begin in the first and second decades of life and is prevalent to some degree across all age groups thereafter. More severe forms of atherosclerosis tend to be found in middle-aged and older individuals. Major risk factors for atherosclerosis include increasing age, elevated LDL cholesterol, elevated blood pressure, obesity, smoking and diabetes. Cognitive function represents the complex repertoire of abilities reflecting a dynamic interaction between the individual and the social environment throughout life. Determinants of cognition are multifactorial and include a number of demographic, social, biological, physiologic and lifestyle factors. Evidence suggests that several risk factors for CVD may also be risk factors for cognitive dysfunction. A review of the literature summarizes the epidemiologic studies that have examined the association between subclinical atherosclerosis and cognitive function. Some studies reported associations between atherosclerosis and reduced cognitive function, while others reported weak or null associations. The studies tended to measure atherosclerosis in medium to large-sized arteries, and focused on older populations with more advanced atherosclerosis, with CVD, or who may have not been cognitively intact. This research includes two analyses aimed at further investigating the relationship between subclinical atherosclerosis and risk factors for atherosclerosis and cognitive function specifically in healthy, cognitively intact xi middle-aged and older adults without CVD or diabetes. The first analysis examines the cross-sectional association between carotid artery intima-media thickness (IMT) and cognitive function, and the second examines the cross-sectional association between the metabolic syndrome, a risk factor for CVD, and cognitive function. Results from the analyses provide some support for a role of IMT and the metabolic syndrome in reduced cognitive function, specifically in the area of verbal memory and conceptual abilities. A third analysis is proposed in which the retinal vessels are used as markers of the cerebral microvasculature to more directly examine microvascular disease and cognition. 1 Chapter 1 - Review of the Literature Introduction This chapter begins with a brief summary of the descriptive epidemiology of cardiovascular disease and atherosclerosis, and a description of areas of human cognitive function. The rationale for studying atherosclerosis in relation to cognitive function, as well as proposed biological bases for an association between atherosclerosis and cognition are provided. The criteria used to select studies for inclusion in the literature review are explained. Following the actual review and critique of the literature is a discussion of questions remaining with respect to the relationship between atherosclerosis and cognitive function. Epidemiology of Cardiovascular Disease and Atherosclerosis Cardiovascular diseases, including coronary heart disease, coronary artery disease, hypertension and stroke, are highly prevalent in the United States (US) population. In the year 2003, an estimated 1 in 4 or over 71 million Americans had cardiovascular disease (CVD) [3], and each year, almost 40% of all deaths are attributed to CVD [28]. Coronary heart disease (CHD), which is the leading cause of mortality in the US, was responsible for one in every five deaths in the US in the year 2003 or over 685,000 deaths. Stroke is the third leading cause of death in the US and accounted for over 157,000 deaths in the year 2003. Taken together, deaths from heart disease and stroke were more numerous than the eight other leading 2 causes of death combined [28]. Direct and indirect costs associated with CVD, including health expenditures and loss of productivity, is estimated at over $403 billion for the year 2006 [111]. As the US population ages, the incidence of CVD is likely to increase and the burdens associated with it will continue to grow [16]. Atherosclerosis is the most common pathological process underlying CVD [8,15,23,167,170]. Atherosclerosis is a condition in which fat, cholesterol, fibrous tissue and other substances accumulate in the innermost layer of large- and medium- sized arteries, resulting in a narrowing and hardening of blood vessels. The atherosclerotic process occurs on a continuum, beginning with damage to arterial endothelium and progressing to vascular remodeling, plaque formation and stenosis [15]. The distinction between subclinical and clinical atherosclerosis is made when consequences of the presence of plaques (severe occlusion of the artery or plaque rupture resulting in thrombus) are symptomatically manifested. Estimating the incidence, prevalence and mortality for atherosclerosis is not straightforward given that the biological history of the disease exists on a continuum, and the chronic nature of the condition is such that the disease is likely to exist in the subclinical stage for decades. Descriptive statistics for atherosclerosis depend on how atherosclerosis is defined, and thus are likely to vary over time and region as well as with the technology used to measure disease. For example, prevalence estimates derived from the National Health Interview Survey (NHIS) are based on a questionnaire item asking whether survey participants “ever had” atherosclerosis, implying that they had experienced symptoms or received a diagnosis of 3 atherosclerosis (therefore, clinical disease) [109]. Thus, these estimates would not capture the true number of people with all stages of atherosclerosis in the US since they are unlikely to include subclinical disease. Alternatively, human autopsy studies yield valuable information about the natural history of atherosclerosis, and allow for an examination of the prevalence of atherosclerosis in various stages of the disease and at different ages. The use of autopsy studies in atherosclerosis epidemiology dates back at least to 1953 when a study reported prevalent atherosclerosis among young American soldiers killed in the Korean War [44]. Recent autopsy studies among Caucasians in the US and European, African-Americans and Asians in Europe and Asia ranging in age from 2- 90 years provide estimates of the extent of atherosclerosis among diverse populations and over a wide age range of ages [12,123,135,152,156,190]. An autopsy study of 216 individuals aged 2-39 in the US [12] found fatty streaks of atherosclerosis in at least one blood vessel (either aortic or coronary artery) in over 90% of autopsied individuals. Prevalent fibrous plaques, defined as raised, collagen capped lesions were observed in the aortic arteries of US children as young as 2-10 years, and in the coronary arteries in children as young as 11-15 years. The prevalence of plaques in the aortic and coronary arteries increased with increasing age. In the aortic arteries, the prevalence of fibrous plaques in males aged 2-10 years was approximately 20%, 35% in ages 21-25 and 70% in ages 31-39. In the coronary arteries, the prevalence of fibrous plaques was approximately 20% in males aged 11- 4 15 years, 50% in ages 21-25 years and 100% in ages 31-39. Females tended to develop plaques in both the aortic and coronary arteries at older ages than males. Another autopsy study of 2,876 15-34 year old individuals in the US (24.3% women and 51.8% African-American) confirmed that atherosclerosis begins in youth [152] and increases in prevalence and extent with increasing age. Intimal lesions appeared in all the aortic arteries and in more than half of the right coronary arteries of the youngest age group (15-19 years). The existence of any lesion in the right coronary arteries increased from about 60% in individuals aged 15-19 years, to 65% in individuals aged 20-24, 70% in individuals aged 25-29, and almost 80% in individuals aged 30-34. Fatty streaks were present in the abdominal aortic arteries of all (100%) of autopsied individuals studied (Table 1). Fibrous plaques in the aortic arteries were less prevalent, ranging from about 3-10% among individuals aged 15- 19 years and increasing to about 55-65% among individuals aged 30-34 years. Calcified lesions in both the aortic and coronary arteries were rare among individuals aged 15-34, appearing in the coronary arteries roughly in the third decade of life. For a given age range, calcified lesions in the aortic arteries were more prevalent than in the coronary arteries. 5 Table 1. Prevalence of Fatty Streaks, Fibrous Plaques and Calcified Lesions in Abdominal Aortic Arteries of White Males aged 15-34 years in US (PDAY study) Prevalence (%) Age Group (years) Fatty Streak Fibrous Plaque Calcified Lesion 15-19 100 13.1 1.4 20-24 100 18.3 0.4 25-29 100 37.3 3.8 30-34 100 65.2 12.6 Reproduced from Strong et al.[152] Reports of 2,562 autopsies (46.8% women) performed during the years 1979-1994 on residents of Olmstead County, Minnesota were reviewed for the presence of coronary disease[135]. All autopsy examinations in the county were performed in a single department using a uniform and comprehensive system of autopsy techniques. Investigators considered the right, left main, left anterior descending and left circumflex coronary arteries, which were graded for the presence of disease by visual inspection. Significant coronary artery disease was defined as the presence of grade IV (76-99%) or greater reduction in cross-sectional luminal area of any of the major epicardial vessels (right, left main, left anterior descending or left circumflex coronary arteries), accompanied by grade III (51-75%) or greater reduction in the left main coronary artery. For the period of 1990-1994, 32% of men and 16% of women aged 20-59 years had significant coronary artery disease. The prevalence of significant coronary artery disease among those aged 60 years or older was 68% in men and 51% in women. Finally, an autopsy study of 710 male pilots involved in fatal general aviation accidents in the US during the years 1980-82 provides age-specific prevalence rates 6 of coronary atherosclerosis at different levels of severity [17]. Investigators classified coronary atherosclerosis in one of three severity categories based on the percent occlusion prevalent in at least one coronary artery, Grade I: <33% luminal occlusion, Grade II: 33-66% luminal occlusion, or Grade III: >66% luminal occlusion. Table 2 summarizes prevalence rates of atherosclerosis of different levels of severity in the population of pilots by age group. Table 2. Age-Specific Prevalence Rates of Coronary Atherosclerosis by Severity among Fatal General Aviation Accident Victims, 1980-1982 Age Group (years) Minimal to Moderate Atherosclerosis (Grades I and II) (per 10,000) Severe Atherosclerosis (Grade III) (per 10,000) ≤ 29 36.2 0.0 30-39 63.9 0.6 40-49 79.8 1.5 ≥ 50 78.4 7.4 Reproduced from Booze & Staggs[17] Results of this and other autopsy studies suggest that atherosclerosis appears in the first and second decades of life, is prevalent to some degree across all age groups thereafter, and more severe atherosclerotic disease becomes apparent in middle aged and older individuals. Cognitive Function Cognitive function represents the complex repertoire of abilities reflecting a dynamic interaction between the individual and the social environment throughout life [22,98]. Cognition or cognitive function are conceptualized as four major classes, 1) receptive functions (abilities involved in selecting, acquiring, classifying and 7 integrating information), 2) memory and learning (information storage and retrieval), 3) thinking (mental organization and reorganization of information), and 4) expressive functions (the means through which information is communicated or acted upon). Determinants of cognition are multifactorial and include a number of demographic, social, biological, physiologic, genetic, neurological and lifestyle factors. Factors related to psychological and emotional health including depression, psychological stress, anxiety and quality of life also contribute to cognitive function. Cardiovascular and metabolic risk factors including high blood pressure, overweight and obesity, diabetes and stroke [138] have all been shown to be associated with cognitive function [112]. This research is based on a biological model in which the cumulative effects of such determining factors contribute to an individual’s level of cognitive functioning at a given time. Cognitive function follows an approximate trajectory over the human lifespan with cognitive growth apparent during childhood, a plateau in cognition with relative stability and little variation during middle and older ages, and cognitive decline in the latter years of life [22]. Cognitive function in individuals and populations over the lifespan exists on a continuum, and both intra- and interindividual variation is expected. “Normal aging” is accompanied by expected changes in cognition (Figure 1). Cognitive impairment and dementia are conditions in which cognitive functioning is reduced beyond that expected by aging alone. Thus while some decline in cognition is expected with normal aging, cognitive impairment and dementia are not necessary consequences of aging. An understanding what factors 8 contribute to reducing cognitive function would be useful in predicting the development of cognitive impairment or dementia. The goal of this research is to contribute to the understanding of determinants of cognitive functioning other than demographic factors (age, socioeconomic status) or comorbid conditions including neurological diseases, brain injuries or psychological conditions such as depressed mood. The focus of this research – cognitive function as a health outcome, is continuously distributed in the population. This research will examine cognitive dysfunction as exists in the normal population distribution of cognition. The two analyses included in this dissertation are focused on populations that were selected to be healthy, and thus it is unlikely that cognitive impairment or dementia will be present in the sample. Thus, the cognitive outcome is not the presence or absence of disease that can be defined using some criteria, but rather the range of cognitive abilities existing in the selected population. Figure 1. Cognitive Function Over Age Rationale for Studying Atherosclerosis in Relation to Cognitive Function Epidemiologic studies have shown that atherosclerosis, CVD and cognitive dysfunction share many of the same risk factors. Established risk factors for atherosclerosis and CVD including hypertension, diabetes and cigarette smoking, as well as novel, more recently identified factors, such as elevated homocysteine levels have also been shown to be associated with reduced cognitive function, cognitive impairment and dementia [38]. 9 10 Hypertension has been associated with reduced cognitive function in several large population-based studies [38]. A number of these studies have provided evidence of a dose-response relationship between increasing blood pressure level (BP) and reduced cognitive function and others have shown that elevated BP clearly precedes lower cognitive test performance by one to three decades [42,88,134,148,154,155]. Type 1 and Type 2 diabetes have also been shown to be associated with poorer cognitive function in several studies, although some studies have reported null associations [38]. Current cigarette smoking has been associated with reduced cognitive function in at least one recent study [68]. Research regarding the effect of cholesterol, homocysteine, obesity and alcohol intake on cognition are more mixed [38]. At least some of the studies that failed to detect an association between CVD risk factors and cognitive function were limited by misclassification of exposure. Specifically, in the case of diabetes and cigarette smoking, where “dose” is probably critical for an effect on cognition, classifying participants into “yes/no” categories will fail to accurately represent duration of diabetes (duration of exposure), levels of glucose control, or quantities and duration of smoking. A “yes/no” diabetes category lumps together individuals with recent diabetes diagnoses and those who have had the condition for a number of years. Similarly, for smoking, while pack-year variables are an attempt to quantify amount of smoking, they don’t capture recently or intensity of smoking. Furthermore, a larger problem exists in many of the studies in that the cross-sectional designs cannot adequately address timing of exposure and 11 onset of outcome. Overall, however, the evidence linking atherosclerosis risk factors to cognition is at least sufficient to merit further investigation into the association. Biological Basis for an Association between Atherosclerosis and Cognitive Functioning The diminished delivery of oxygen resulting from reductions in blood flow to the brain is another possible mechanism by which atherosclerosis may affect cognition. Lower cerebral blood flow can be due to heart disease or stenosis of cerebral and carotid arteries [34]. Ruitenberg et al. (2005) showed that greater cerebral blood flow velocity was related to a lower prevalence of cognitive decline and dementia. Results of this study suggest that cerebral hypoperfusion may precede and contribute to the onset of cognitive dysfunction and dementia [140]. Deficits in certain cognitive abilite is may be explained by their association with areas of the brain that are more vulnerable to the effects of cerebral hypoxemia associated with vascular disease. In populations that are cognitively normal, small decreases in oxygen supply to the brain associated with subclinical atherosclerosis may explain some of the individual differences in cognitive performance and the cumulative effect of small decrements in brain oxygen over time may contribute to declines in cognitive function. 12 Selection of Studies for Literature Review A number of previous epidemiologic studies have investigated the relationship between cardiovascular disease risk factors, including atherosclerosis and cognitive function. I conducted a systematic review of the published literature to identify and then summarize epidemiologic studies that evaluated the association between atherosclerosis and cognitive function. My aim was to identify studies that were similar to the research in this dissertation in evaluating the range of cognitive function in “healthy” populations, drawn from population-based or community-based samples. Using the PubMed database, I conducted a comprehensive search of the published literature to identify articles reporting on atherosclerosis (at any stage) and cognitive dysfunction, which were published between January 1950, and July 2006 in English. I initially searched for studies using the keywords “atherosclerosis” and (“cognition” or “cognitive function”). More specific search terms were then used including “subclinical”, “IMT”, “carotid artery”, “carotid ultrasound”, “ankle brachial index”, “arterial stiffness”, “plaque”, “pulse wave velocity”, “vasoreactivity”, “arterial calcium” with “cognitive”, “cognition” and “cognitive function” to identify additional articles. Approximately 175 articles were identified from the various search combinations. I compiled and reviewed abstracts of all studies to determine whether they were indeed epidemiologic studies and whether I should include them in the literature review. I obtained full-length articles of any study that appeared to be relevant based on the abstract in order to examine the study in more detail. 13 Studies initially identified by the literature search were retained for review if they focused on non-cognitively impaired, non-demented populations, as well as included populations without clinical cardiovascular disease. The main reasons that studies were excluded from consideration were 1) they were not epidemiologic studies (i.e., review articles, case reports, clinical trials, animal studies); 2) they did not analyze or report on an atherosclerosis-cognition association; 3) atherosclerosis was not a primary independent variable or cognitive function was not the dependent variable; 4) the cognitive outcomes assessed were only dementia or cognitive impairment; 5) the populations of focus included individuals with clinical CVD (eg. CABG patients) or with comorbid neurological conditions (eg. lupus). In total, 19 studies identified from the published literature were included in this review. Table 3 summarizes the studies by lead author, location, cohort (if applicable), study population, cognitive outcome, model information and results. Table 3. Summary of 19 Studies Included in Literature Review 14 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Auperin 1996 France EVA 1,279 French adults aged 59-71 IMT Carotid artery plaques Cognitive performance: Battery of 7 tests analyzed individually Cross-sectional; ANOVA Mean IMT and presence/absence of plaques by low, medium and high performance Covariates: age, education, depression, smoking, alcohol, BP, BMI Significant trend for increased IMT and lower performance Men: Digit Symbol and Word Fluency; Women: no association; Men: OR of poor vs high performance for presence of carotid plaques: for Digit Symbol = 2.10 (1.06-4.16); OR of middle vs high performance = 2.16 (1.22- 3.83); Women: no association; Hofman 1997 Netherlands Rotterdam 284 dementia cases AD=207, VaD = 50, other dementia = 27 1,698 controls IMT Carotid plaques ABI AD, VaD, other dementia Cross-sectional; Logistic regression OR of dementia per 0.20mm increase in IMT, presence of plaques, ABI Covariates: age, sex; also BMI, BP, total cholesterol IMT: OR of AD = 1.3 (1.0- 1.6); VaD = 1.9 (1.3-2.8); Other dementia = 0.8 (0.4- 1.5). Presence of carotid plaques: OR of AD: 1.8 (1.2-2.7); VaD: 3.2 (1.6- 6.8); Other dementia: 1.6 (0.6-4.3). PAD: OR of AD: 1.3 (0.9-1.8) VaD: 2.5 (1.3- 4.8); Other dementia: 1.0 (0.4-2.4) Cerhan 1998 USA ARIC 13,913 adults aged 45-64 w/o history of stroke or TIA IMT Cognitive function: Delayed Word Recall Digit Symbol Word Fluency Cross-sectional; ANCOVA Covariates: age, education, occupation, race and field center Significant differences in mean scores: 1) across CIMT quintiles, 2) comparing thickest quintile to thinnest; for Digit Symbol; Women only: Delayed Word Recall Table 3, continued 15 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Knopman 2001 USA ARIC 10,963 adults 76% of original cohort IMT Cognitive decline: Change in three tests over 6 years Longitudinal; Linear regression Covariates: age, gender, race, education, site, CNS meds No significant differences in mean scores on any test across tertiles of CIMT; Changes in scores over follow-up were small Johnston 2004 USA CHS 4,006 adults ≥65 years w/o history of stroke, TIA or carotid endarectomy IMT Cross-sectional: cognitive impairment (<80/100); Longitudinal (5 year f/u): cognitive decline (decrease in 1 point/year) Modified-MMSE Digit Symbol Cross-sectional; Logistic regression CIMT in 4 th quartile vs.1 st Covariates: age, sex, ethnicity, education, HTN, LDL, HDL, diabetes, smoking, heart failure OR cognitive impairment left side = 1.5 (0.98-2.3); right side = 0.95 (0.64-1.4) OR cognitive decline left side = 1.4 (1.0-1.8); right side = 1.0 (0.78-1.3) Muller 2006 Netherlands Community based 396 Dutch men aged 40-80 IMT PWV ABI Battery of 10 tests Three cognitive domains: Memory, processing, executive functioning Cross-sectional; Linear regression Change per 1 mm increase in IMT Covariates: age, education IMT: Memory: B= -1.32 (- 2.63, -0.01); Processing: B = 0.75 (-1.60, 3.10); Executive: B = -0.89 (-2.16, 0.38); Subjects with prevalent CVD or sub- clinical CVD had lower memory scores; Greater PWV and ABI (<0.9) was associated with lower performance in all three cognitive domains (but ns) Table 3, continued 16 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Wright 2006 USA Northern Manhattan Study 2,215 adults ≥40 years (mean age 67.9) without history of stroke maximum carotid plaque thickness (MCPT) cognitive status: MMSE Cross-sectional; MCPT was examined as a mediator of primary association MCPT was not a mediator of alcohol consumption- cognition association Breteler 1994 Netherlands Rotterdam Study 4,971 adults aged 44-94 atherosclerotic plaques/lesions in internal carotid arteries; generalized atherosclerosis (ABI <0.90) MMSE; two categorizations of cognitive impairment (<24 & <26) Cross-sectional; Multivariate logistic regression models PAD and presence of lesions were associated with a shift in MMSE toward lower values; Significant differences in MMSE <24 vs. >=24 & <26 vs >= 26 for PAD, presence of lesions Covariates: age, gender, education, smoking Fukuhara 2006 Japan Data Bank Survey 203 adults ≥85 years Brachial-ankle PWVMMSE; subjects divided in two groups: ≥24 = normal, <24 = cognitive impairment Cross-sectional; Multiple logistic regression comparing cognitive impairment vs. normal cognition; Covariates: education, gender, BMI, BP, cholesterol, alcohol, smoking, medication use B=-0.157 (p=0.003) decrease in MMSE score per m/s PWV; Excluding subjects with ABI <0.9, B=- 0.196 (p=0.001) decrease in MMSE per m/s PWV Fujiwara 2005 Japan Community- dwelling volunteers 352 elderly adults ≥70 years w/o CVD brachial femoral PWV (cm/s) MMSE; subjects divided in two groups: ≥24 = normal, 24 = cognitive impairment Cross-sectional; Multiple logistic regression comparing highest PWV tertile to lowest. Covariates: age, sex, education, pulse pressure, ABI, iron, albumin, total cholesterol, histories of hypertension, diabetes, hyperlipidemia Compared to PWV in lowest tertile, OR of MMSE <24 for PWV in middle tertile = 9.66 (1.15-80.93) (p=0.037); for PWV in highest tertile = 3.42 (0.37 - 31.96) (p=0.282) Table 3, continued 17 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Scuteri 2005 Italy Hospital-based 84 elderly subjects (mean age 78) with memory deficits, without history of stroke, atrial defibrillation carotid femoral PWV normalized for BP MMSE Cross-sectional; Multiple linear regression Covariates: age, sex, education, prevalent CVD, LDL&HDL, anti- hypertension medications, nitrates B = -0.28 p<0.01 Hanon 2005 France Hospital-based 308 elderly subjects >60 years (mean age 78) with memory impairment carotid femoral PWV MMSE, neuropsychological battery (CEP) of 4 tests Subjects were divided into four groups: AD, VaD, MCI and normal cognition Cross-sectional; ANCOVA: used to look at differences in PWV between groups and Multivariate linear and logistic regression Covariates: age, gender, SBP, education, anti- hypertension treatment, presence of CVD ANCOVA: PWV was significantly higher in VaD or AD, higher in MCI compared to normal cognition; Logistic regression: per 2 m/s PWV, OR of AD = 1.73 (1.27- 2.47); VaD = 3.52 (1.87- 8.05); Linear regression: per 2 m/s PWV MMSE: -0.091, p<0.001 CEP: -0.029, p<0.001 Rosano 2005 USA CHS 409 adults mean age 78 years EBCT CAC as measure of total coronary plaque burden and marker of generalized athero cognitive status based on neuropsychological evaluation; subjects divided into three groups: normal cognition, MCI and dementia Cross-sectional; ANOVA, logistic regression Covariates: age, sex, race and prevalent CVD No difference in cognitive impairment or dementia comparing CAC quartiles 2,3 & 4 to quartile 1 (p=0.23); OR of cognitive impairment/dementia for CAC in Q ≥2 versus Q1 = 1.76 (0.9, 3.3) Table 3, continued 18 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Van Exel 2002 Netherlands Leiden-85 599 adults all 85 years atherosclerosis “burden” variable created = total number of CVD pathologies from medical history including 1.) arterial surgery, 2.) stroke, 3.) intermittent claudication, 4.) MI, 5.) angina or myocardial ischemia MMSE: ≥28 = normal cognition, <28 = cognitive impairment; Three cognitive tests Measuring processing speed, attention, memory. Cognitive impairment was defined as scores below median value. Cross-sectional; ANOVA groups of number of CVD pathologies (0, 1, >=2) Multiple logistic regression comparing cognitive impairment vs. normal cognition Covariates: gender, education, depression, CVD medications, NSAID use Significant difference in MMSE scores by CVD pathologies; OR of cognitive impairment compared to 0 CVD pathologies for 1 CVD pathology = 1.2 (0.8,1.8); ≥ 2 CVD pathologies = 2.0 (1.3, 3.0) defined by processing speed for 1 CVD pathology = 1.2 (0.7-1.8) ≥2 CVD pathologies = 2.1 (1.2- 3.5). No significant differences for memory. Saxton 2000 USA CHS 989 adults >66 years, mean age 74 years predominantly white Subjects divided into three groups: clinical disease (CVD events), sub- clinical disease (evidence from exam), no disease Neuropsychological battery of 13 tests Cross-sectional; ANCOVA for differences in test scores between groups Covariates: age, education, gender clinical disease group scored lower on Trails A, Block Design and Digit Symbol compared to no disease group ; sub-clinical disease group scored significantly lower on Fruit Fluency & Digit Symbol Moser 2004 USA Volunteers 14 elderly adults mean age 73 years vessel function (forearm blood flow measured by plethysmography) in response to 3 agents to assess vasodilation, smooth muscle function Neuropsychological battery (RBANS, Form A) assessing 5 cognitive domains, global cognitive function Cross-sectional; spearman and partial correlation coefficients Covariates: age, education, depression (individually, not jointly) partial correlation coefficients adjusted for age - for endothelium-dependent vasodilation: r=0.39, p=0.42; for endothelium-independent vasodilation r=0.42, p=0.158 for smooth muscle function r=0.73 p=0.005 Table 3, continued 19 Lead Author, Year Location Cohort Study Population Atherosclerosis Measure Cognitive Measure Model Information Results Woo 2006 China Community- dwelling volunteers 3,999 adults ≥65 years (mean age 72.5 years) ABI *analyzed as the dependent variable MMSE; cog impairment = MMSE <24 Cross-sectional; ANCOVA, Logistic regression Covariates: age, gender, diabetes, hypertension, MI, alcohol intake, smoking, vitamin C intake, ability to walk 6 m, history of stroke cognitively impaired had significantly lower ABI; OR of ABI <0.9 for MMSE <24 = 1.58 (1.19, 2.09) compared to ABI ≥ 0.09 Price 2006 Scotland EAS 700 adults aged 55-74 ABI Low = <0.9 & <0.95 Normal = ≥0.9 & ≥0.95 Neuropsychological battery of 6 tests Multiple linear (test scores continuously) and logistic regression (scores in tertiles) comparing low versus normal ABI Covariates: age, gender, smoking, education, presence of CVD at baseline Low ABI (<0.95) associated with significantly lower Digit Symbol and verbal fluency scores, and lower performance on all other tests (ns). Low ABI (<0.90 or <0.95) was associated with significantly with cognitive performance in the lowest versus tertile for Digit symbol and verbal fluency but not memory. Wong 2002 USA ARIC 7,526 adults aged 45-64 w/o history of stroke, CNS medications Retinal microvascular abnormalities as surrogate for atherosclerosis Delayed Word recall, Digit Symbol, word fluency ANCOVA. Covariates: age, sex, race, center, education, diabetes, occupation, glucose, hypertension, BP, IMT, cholesterol, HDL-C, triglycerides, smoking, alcohol Significantly lower mean scores on 3 tests in persons with any retinopathy. Significant elevation in risk of cognitive impairment (based on 3 tests separately) for any retinopathy 20 A major source of heterogeneity among the studies selected for review was the diversity of methods used to measure atherosclerosis. The selected studies employed the following methods to assess atherosclerosis: • ultrasonography of carotid arteries: assessment of carotid artery intima-media thickness (IMT) and quantification of carotid plaques; • electron beam tomography (EBT): measurement of coronary artery calcium (CAC) as an indicator of total coronary plaque burden and generalized atherosclerosis of the cerebral vascular bed and other peripheral vascular beds; • blood pressure measurement to derive an ankle-brachial index (ABI) as a measure of generalized atherosclerosis and peripheral artery disease; • measurement of pulse volume waveforms to calculate carotid-femoral or carotid-brachial pulse wave velocity (PWV) as measure of arterial stiffness (greater distance to time traveled indicates poorer prognosis) ; • plethysmographic measure of blood flow as an indicator of vessel function; • retinal photography to identify retinal microvascular abnormalities and measure diameters of the microvasculature; 21 • quantification of CVD pathologies such as myocardial infarction, angina and stroke to derive a surrogate measure for atherosclerosis and atherosclerotic burden; In addition, several studies used a combination of measures of atherosclerosis to quantify the presence of subclinical or clinical atherosclerosis at multiple vascular sites and to estimate “burden” of disease. Among the studies that examined the effect of subclinical atherosclerosis on cognitive function using measures of IMT, four were population-based [on three cohorts, Rotterdam, Cardiovascular Health Study (CHS) and Atherosclerosis Risk in Communities (ARIC)] and two were community-based [Edude de Vieillissement Arteriel (EVA) and the study of Dutch independently living men]. Both the ARIC and CHS cohorts were studied cross-sectionally (baseline associations between IMT and cognitive function), as well as longitudinally for cognitive decline. While there was some overlap in the neuropsychological measures used to assess cognitive function in the studies selected for review, there was also a great degree of heterogeneity in the tests employed as well as the cognitive abilities assessed. Details of the cognitive abilities and measures will be provided in the discussion of the individual studies. However, one main distinction between studies were those that used one or only a few cognitive measures to assess one or a few aspects of cognition (Rotterdam, CHS and ARIC cohorts) versus those that employed an extensive battery of measures to assess a range of cognitive abilities or domains 22 (studies of French [7] and Dutch individuals [103]). Studies that assessed cognitive function with a single test usually used the Mini-Mental State Exam (MMSE). Limitations of the MMSE will be discussed later in the chapter. The earliest studies were from the Rotterdam cohort [21,61] and the EVA cohort in France [7]. Later studies included the ARIC cohort [29,75], the CHS cohort [66,139,142] and a study of Dutch independently living men [103]. Studies ranged in size from the hundreds [103], to <1,500 [7,61] to several thousand (CHS, n=4000; & ARIC, n=13,000). Most studies selected included middle-aged, older and elderly adults, and excluded subjects who had a positive history of stroke. As mentioned above, selected studies varied in the assessment and categorization of atherosclerosis, in the definition of the cognitive outcome, in analytic methods as well as in the covariates included in analyses. Literature Review and Critique The first population-based study to examine the association between atherosclerosis and cognition was that of Breteler et al. (1994) who studied 4,971 participants aged 55-94 in the Rotterdam study for whom cognitive test data were available [21]. The Rotterdam study was a single-center prospective study of randomly selected residents aged 55 and over of a suburb in the Netherlands, including institutionalized individuals. Of the 7,100 residents invited to participate, 5,670 (80%) agreed to participate, with a final study sample of 5,530. Two measures of atherosclerosis were included. Ultrasonography of both carotid arteries was performed to identify 23 atherosclerotic lesions, defined as a focal widening relative to adjacent segments with protrusion into the lumen, in the internal carotid arteries. Plaques in the carotid arteries were classified as either present or absent. ABI was calculated and peripheral arterial disease (PAD) was considered present if the ABI was <0.9. Cognitive function was assessed with the MMSE and two cutoff values (<24 and <26) were used to identify participants with suspected dementia or cognitive dysfunction. The population distribution of MMSE scores were shifted toward lower values in participants with atherosclerosis, measured either as lesions localized to the carotid arteries or in the lower extremities compared to participants without atherosclerotic disease. In separate multiple logistic regression models adjusted for age, gender and education, a significantly greater proportion of participants with atherosclerosis measured either as PAD or prevalent plaques had MMSE scores below 24 or 26 (all p-values <0.003). Adding smoking status to the models resulted in minor change to the effect estimates. The MMSE is a widely used, standardized method for assessing global mental status. In this respect, the use of this instrument in the Rotterdam and other studies allows for some comparability between the cohort and other populations as well as for interpretability with published population normative data. It is important to note, however, that the MMSE was originally developed as a clinical instrument to evaluate the cognitive status of elderly people and to screen for cognitive impairment and dementia [46,160]. Furthermore, while the MMSE is brief, easy to administer, has good reliability and has been translated into a number of languages, the lack of 24 an administration manual means that scoring and interpretation vary between users [101]. Conventional cutoff scores exist and are widely used. Out of a total of 30, a score of ≥ 28 is generally considered normal cognition [53] and a score of < 24 is used to identify individuals with dementia [79]. In order to compensate for the varying accuracy of the MMSE in individuals of different ages, education levels and ethnicities, cut points may be changed [20,79] as was done in the Rotterdam study. There are problems with the MMSE scores and its cut points, both conventional and modified. When administering the instrument to cognitively “normal” individuals who are likely to score above such cut-offs, there is an expected “ceiling effect”. This makes the MMSE limited in its ability to discriminate between cognitive performance at the upper range of functioning. As applied to epidemiologic studies of cognitively intact populations, the MMSE provides a very small range of scores with which to assess variations in “normal” cognitive function. Furthermore, while the MMSE assesses orientation, attention, immediate and short-term recall, language, and the ability to follow simple verbal and written commands, it was not designed to examine specific domains of cognition. Thus none of the studies included in this review that used this instrument were able to evaluate specific cognitive abilities. Therefore, if a study’s aim is to examine the range of normal cognition among populations who are likely to be cognitively intact, or to be able to evaluate specific cognitive abilities within the range of cognition, the MMSE may not be an appropriate measure. 25 Hofman et al., 1997 conducted a nested case-control study of 284 cases of dementia with onset less than three years before study entry and 1,700 controls participating in the Rotterdam study [61]. Among the dementia cases, 72.8% had Alzheimer’s disease (AD), 17.6% had vascular dementia (VaD) and 9.5% had dementia classified as “other”. Ultrasonography of the near and far walls of the common carotid arteries was performed. Three measures of atherosclerosis were obtained: mean IMT of the far walls of the distal common carotid arteries [(left + right)/2], presence of plaques in the common and internal carotid arteries and the carotid artery bifurcation, and ABI. PAD was considered present if ABI was <0.9. A composite score indicating burden of atherosclerotic disease and ranging from 0-3 was calculated for each subject by giving a point each for having CIMT > 75 th percentile, presence of plaques in one carotid artery and ABI <0.9. Multiple logistic regression models were used to estimate the association between atherosclerosis (four separate models for each measure of atherosclerosis or burden of atherosclerosis) and dementia subtypes while adjusting for age and sex. All indicators of atherosclerosis were associated with an increased risk of any dementia as well as specific dementia subtypes. The presence of plaques was most strongly associated with dementias (OR’s ranging from 1.6-3.2) and CIMT was least strongly associated (OR’s ranging from 0.8-1.9). A 0.20 mm increase in CIMT was associated with a 30% greater risk of AD (OR = 1.3, 95% CI = 1.0-1.6), a 90% greater risk of VaD (OR = 1.9, 95% CI = 1.3-2.8) and a 30% greater risk of all dementia (OR = 1.3, 95% CI = 1.1-1.6). The presence of plaques was associated with an 80% increase in risk of AD (OR = 1.2-2.7), a 220% increase in risk of VaD (OR = 3.2, 95% CI = 1.6-6.8) and a 90% increase in risk of 26 all dementias (OR = 1.9, 95% CI = 1.3-2.7). PAD was associated with a 30% increase in risk of AD (OR = 1.3, 95% CI = 0.9-1.8), a 150% increase in risk of VaD (OR = 2.5, 95% CI = 1.3-4.8) and a 50% increase in risk of all dementias (OR = 1.5, 95% CI = 1.1-2.0). A greater burden of atherosclerosis was highly significantly associated with AD, VaD and all dementia (all p-trends <0.001) but not associated with other dementias (p = 0.392). Additional adjustment for blood pressure, total cholesterol and body-mass index (BMI) essentially did not alter the observed associations. It is important to note the observation of stronger relationships between atherosclerosis and VaD than AD. The Rotterdam studies were focused on more severe cognitive impairment and dementia among a population that included prevalent CVD. From the first study, using the entire range of MMSE scores, we know that the presence of atherosclerosis was associated with a shift in MMSE scores towards lower values. However, the studies don’t provide results quantifying the effect of atherosclerosis on cognitive function among the cognitively intact portion of the population. Auperin et al. (1996) studied a cohort of 1, 279 adults aged 59-71 in France (EVA cohort) excluding those with a history of stroke. Four ultrasonographic measurements of the bilateral common carotid arteries at plaque-free sites were used to calculate CIMT. Plaques from all segments of the carotid arteries (common carotid arteries, carotid bifurcations and the first 2 cm of the internal carotid arteries) were quantified. Plaques were prevalent in 28% of men and 17% of women, 85% of participants had stenosis <20%; mean CIMT was 0.69 ± 0.14 mm in men and 0.65 ± 27 0.11 mm in women. Cognitive function was assessed with a battery of seven tests including Trails B and Digit Symbol (to assess visual attention and psychomotor speed), Word Fluency (verbal fluency), Paced Auditory Serial Addition Test (auditory attention), Benton Visual Retention Test (visuospatial abilities), Auditory Verbal Learning Test (memory – immediate recall and learning) and Raven Matrices (logic and reasoning). Scores on the individual tests were divided into quartiles and then the two middle percentiles were combined, resulting in three categories of cognitive abilities for each test, low, intermediate and high. The MMSE was also administered to study participants as a measure of global cognition; a score of 24 was used as a cut-off between normal and poor cognitive skills. MMSE scores of <24 occurred in 3.4% of men and 5.6% of women. The investigators used polychotomous logistic regression to assess the association between increasing CIMT (mm) or the presence of plaques and cognition comparing low and intermediate levels of cognitive performance versus high performance. Separate models were run for men and women and were adjusted for age, education, smoking, alcohol consumption, blood pressure (systolic and diastolic), BMI and depressed mood. The presence of carotid plaques was significantly positively associated with poor (OR=2.10, 95% CI = 1.06, 4.16) and intermediate (OR=2.16, 95% CI = 1.22, 3.83) levels of performance compared to high performance on the Digit Symbol Test among men [7]. There was evidence for a significant trend of increasing number of plaques and decreasing performance on the Digit Symbol and Paced Auditory Serial Addition Test (both p<0.05). Odds ratios for the other tests were also elevated, but non-significantly. There was a statistically significant trend in decreasing categories 28 of performance on the Digit Symbol Test (p=0.05) and the Word Fluency Test (p=0.03) with increasing thickness of mean CIMT (mm) in men (unadjusted model). A 0.1 mm increase in CIMT was associated with a 75% greater odds of poor performance on the Word Fluency test compared to high performance (OR=1.73, 95% CI = 1.08, 2.77) in men. In women, however, there was no evidence of an association between cognitive performance and presence or number of plaques or CIMT. The authors’ findings of an association between carotid plaques and increased CIMT with cognitive performance in men but not women could be explained by a threshold effect of atherosclerosis. Since more women than men were included in the study, the authors had as much if not more power to detect an association in women than in men. Compared to women, men had a significantly greater prevalence of carotid plaques and significantly thicker CIMT. It is possible that in order to detect an effect of atherosclerosis on cognition, a certain level of atherosclerosis must be reached. In the EVA study population, it is possible that this threshold was reached in men, but not in women who had less evidence of atherosclerosis. The EVA study and a number of other studies included in this review used the Digit Symbol test to assess cognitive functioning. While this test is extremely sensitive to dementia [90], an alternative test, the Symbol Digit Modalities Test (“Symbol Digit”) [150] was developed to improve some of the difficulties associated with the Digit Symbol. The Digit Symbol requires that the test taker “learn” the symbol responses in order to complete the test. The Symbol Digit reverses the presentation 29 of the test material so that the task requires digit responses rather than symbol responses. The Symbol Digit test is thus thought to be “easier” to take, and may be administered orally or as a written test (personal communication with V. Henderson). Cerhan et al. (1998) examined the cross-sectional association between IMT and cognitive function in the ARIC study of 13,913 US adults aged 45-64, and Knopman et al. (2001) tested the association between IMT and cognitive decline after a six- year follow-up of 10,963 members of the same cohort. The ARIC study calculated IMT as the average of measurements taken over twelve sites bilaterally along near and far walls of the carotid artery and assessed cognition using the Delayed Word Recall, Digit Symbol and Word Fluency tests. Cerhan et al. reported significant differences in age-, education-, occupation-, race- and study site-adjusted mean scores in the Digit Symbol test (p<0.005) across quintiles of IMT for both men and women and in the Delayed Word Recall test (p<0.05) for women. Subjects with the thickest IMT (in the highest quintile) had significantly lower adjusted scores on the Digit Symbol test compared to subjects with the thinnest IMT (the lowest quintile) (p<0.005) [29]. Women with IMT in the highest quintile had significantly lower adjusted scores on the Delayed Word Recall compared to women with IMT in the lowest quintile (p<0.05). Absolute differences in mean scores between the highest and lowest quintiles were small, on the order of 0.1-0.2 standard deviations (SD). Knopman et al. reported no significant differences in mean change in any of the three cognitive function tests over the 6-year follow-up between tertiles of IMT measured at baseline. Overall declines in test scores were relatively small (Delayed Word 30 Recall: 0.1 SD, Digit Symbol Test: 0.2 SD, Word Fluency: 0.05 SD) [75]. The ARIC study measured CIMT at three bilateral sites along the carotid artery, including portions of the common carotid and internal carotid arteries as well as at the carotid bifurcation, which differs from the methods used to assess CIMT in the present study. At the date of publication, the ARIC cohort was the youngest to have been studied. While greater numbers of women then men in the study had IMT in the lower quintiles (thinner IMT), the authors detected an effect of atherosclerosis on cognition among women as well as men. With respect to the threshold hypothesis, these results could be interpreted that compared to women in the EVA study, women in ARIC had levels of atherosclerosis sufficient to affect cognition. The methods used to quantify IMT in the ARIC study differ from those used in studies at the Atherosclerosis Research Unit at USC (“USC methods”) where data was obtained for this research. ARIC methods measure the maximum IMT as determined by visual inspection at twelve locations along the near and far wall of the carotid artery bilaterally: the carotid bifurcation, the distal 1.0 cm straight portion of the common carotid artery, and the proximal 1.0 cm of the internal carotid artery. The “mean IMT” variable used in their analyses was calculated by taking the average of maximum IMT measurements taken at the twelve arterial segments (thus the overall mean IMT measure is an average of 12 individual measurements of IMT taken at 6 locations). Maximum measurements are likely to include atherosclerotic disease that has progressed beyond artery wall thickening to lesions. Thus this 31 method probably counts more advanced atherosclerotic disease (lesions) in the measure of IMT, which should reflect pathophysiologically earlier stages of disease. This could have the effect of overestimating the extent of atherosclerosis in its earlier stages [145]. In contrast, USC methods, measure IMT at one standardized location along a 1.0 cm length of the right common carotid artery distal to the carotid artery bulb. The CIMT variable derived by USC investigators and that was used in analyses is the average of 70-100 individual measurements of IMT taken along this segment length. Thus, the ARIC IMT measure and the USC CIMT measure are derived from measurements taken at different anatomical locations, limiting the comparability of the two measures [145]. The ARIC method may affect the associations observed in their two studies of cognitive function. Their method results in some degree of missing data for IMT for which the investigators impute values. In addition, given that the mean IMT measure is derived from a smaller number of measurements suggests that there may greater variability in their data. Furthermore, the reliability of near wall IMT has been questioned [172,173]. Moreover, the ARIC method involves subjectivity from readers who identify locations of maximum IMT along the artery. All scenarios could result in greater “noise” in ARIC IMT data and effect estimates that would be biased towards the null. Furthermore, it is known that the extent of atherosclerosis is greater at arterial bifurcations than at sites along the artery. Thus, the ARIC method of quantifying mean IMT probably results in greater estimates of subclinical atherosclerosis 32 compared to the USC method. All subjects in the ARIC cohort have IMT measured using this method, and there is no reason to believe that the ARIC approach would result in different measurements among subjects with poorer cognitive functioning. Therefore, potential misclassification of exposure (mean IMT) is not expected to be differential among ARIC study participants. However, if the same individual were evaluated separately by the ARIC and USC approaches, a given level of cognitive performance on the same cognitive test administered in both studies would be associated with a greater absolute thickness of IMT in ARIC compared to that in USC. A series of three papers examined associations between different measures of atherosclerosis and cognitive function in the Cardiovascular Health Study (CHS) population. The CHS study is a population-based study of risk factors for CVD and stroke in a sample of over 5,800 community-dwelling adults ≥65 years recruited from Medicare Part A lists starting in 1989 from four US communities in Maryland, Pennsylvania, California and North Carolina. The first study published by Saxton et al. (2000) focused on 989 of the CHS participants who had been administered a battery of cognitive tests. The subjects were predominantly white (93.9%) and had a mean age of 73.6 ± 4.5 years. Participants were classified into three groups: 1.) clinical disease (n=286, 28.9%), if they had a positive history of a CVD event or complications, atrial fibrillation detected by electrocardiography (EKG) or use of a pacemaker ; 2.) subclinical disease (n=345, 34.9%) if they had had an abnormal EKG, low ejection fraction, IMT > 80 th percentile, stenosis > 25%, ABI ≤ 0.9 or a 33 positive responses on the Rose questionnaire for angina or intermittent claudication; 3.) no disease (n=358, 36.2%) if they were free from all forms of disease. The extensive battery of neuropsychological tests included the WMS-R measures of verbal memory [immediate recall & learning], visual memory (immediate recall and non-verbal learning), general memory (visual and verbal combined), attention/concentration, delayed recall (visual and verbal memory); the WAIS-R measures of Block Design (visuospatial abilities) and Vocabulary; the Digit Symbol test (visuomotor coordination & speed); Trails A & B (psychomotor speed and working memory); the Boston Naming test (semantic memory and related language abilities); word fluency (language abilities and executive function); and fruit fluency (language abilities and verbal productivity). Each measure was assessed individually in ANCOVAs adjusted for age, education and gender to examine whether there were cross-sectional differences in cognitive performance on the tests between disease categories. Results suggested that there were significant differences between groups for word and fruit fluency, Trails A&B, Block Design and Digit Symbol, but not for general, verbal, semantic or visual memories, delayed recall, attention/concentration or vocabulary [142]. Specifically, the authors found that participants classified as having clinical disease performed significantly worse on Trails A, Block Design and Digit Symbol compared to participants classified as having no disease. Participants categorized as having subclinical disease performed significantly worse on fruit fluency and the Digit Symbol test also compared to participants with no disease. Finally, participants with clinical disease performed significantly worse than participants with subclinical disease on Trails A & B. 34 An important contribution of this study is the finding that CVD at subclinical levels is associated with poorer cognitive function after controlling for age, gender and education compared to no CVD. It is possible that other covariates not included in the analyses such as depressive symptoms could confound the results. In this case, the observed effect estimates would be negatively biased away from the null. A more serious problem, however, is that almost 40% of the population of adults eligible to participate in the CHS study chose not to participate, and those who did agree to participate were younger, more educated and had less CVD than the population that was initially contacted by CHS investigators. Therefore, the actual population prevalence of clinical and subclinical disease may be underestimated in the CHS population, which would then mean that the disease-cognition associations are also underestimated in the Saxton et al. study. Furthermore, by analyzing participants who were classified into groups based on disease status, this study does not examine associations between the range of cognitive function by levels of atherosclerosis. It would have been interesting for the authors to have additionally reported a quantification of change in cognitive performance per unit of IMT, perhaps stratified by disease status. This would have provided an estimate of the magnitude of the difference in cognition per unit IMT, which could have then been used to compare between studies. In addition, the authors did not conduct tests for trend in neuropsychological performance by disease category, which would have provided additional information about the relationship between atherosclerosis and cognition. 35 The second CHS study published by Rosano et al. (2005) focused on 409 of the CHS participants from the Pennsylvania site who had been cognitively tested and for whom electron beam computed tomography (EBCT) measures of coronary artery calcium (CAC) were obtained. The subjects were predominantly white (76.3%) and female (61%), and had a mean age of 78.2 years. Classifications of cognition were made based on a neuropsychological evaluation and or medical examination. In this study, the authors examined whether CAC (in quartiles) predicted one of three categories of cognitive outcomes: 1.) demented (n=31, 7.6%), 2.) MCI (n=69, 16.9%), or 3.) normal cognition (n=309, 75.6%). The demented and MCI categories were combined in analyses due to small numbers. Unadjusted analyses demonstrated that there was a borderline significant trend of increasing proportions of participants in the cognitive impairment/dementia group with increasing quartiles of CAC (p=0.049). However there was no difference in the age-adjusted proportion of cognitively impaired/demented participants with CAC in quartiles 2,3 & 4 compared to CAC in quartile 1 (p=0.23) [139]. In multivariate logistic regression models adjusted for age, sex, race and prevalent CVD, there was an increase in odds of cognitive impairment/dementia for participants with CAC in the 2 nd (OR=1.66, 95% CI = 0.8, 3.4), 3 rd (OR = 2.07, 95% CI = 1.1, 4.2) and 4 th (OR = 1.65, 95% CI = 0.8, 3.4) quartiles compared to participants with CAC in the 1 st quartile. Participants with CAC in quartiles 2, 3 or 4 had an increased odds of cognitive impairment/dementia (OR = 1.76, 95% CI = 0.9, 3.3) compared to those with CAC in quartile 1. The authors did not test for trend in risk of cognitive impairment/dementia with increasing quartiles of CAC. It would also have been 36 useful for the authors to have dichotomized CAC at the 2 nd quartile to assess whether there was no evidence for a “threshold” effect, given that the highest odds of impairment/dementia were associated with CAC in the 3 rd quartile. In addition, the study population was 56.3% of the source population. At least 15% of those who were lost to follow-up before cognitive assessment were too ill to participate, could not travel or had died. Therefore, it is possible that those who were lost to follow-up would have been categorized in the higher quartiles of CAC and or in the MCI/demented group. Thus, the observed effect estimates are probably underestimates of the true association. Furthermore, while we know that the investigators collected information on educational level and depressive symptoms, these covariates were not included in adjusted models, which suggests that the effect estimates presented may be confounded by these two factors and that the true risk estimates may be somewhat lower than those presented. Johnston et al. (2004) studied cognitive impairment and decline in 4,006 right- handed adults with a mean age of 74.7 years without a history of stroke, transient ischemic attack or carotid endarectomy participating in the CHS study. Cognitive impairment was defined as a score of less than 80 on the modified MMSE or less than 19 on the Digit Symbol test (1.5 SD’s below the mean score for the study population), and cognitive decline over 5 years of follow-up was defined as an average decrease of at least 1 point per year on the modified MMSE, which corresponded to the lowest 25 th percentile of the study population and was considered clinically significant. The investigators compared participants with 37 ≥75% carotid artery stenosis (“high grade”) and 1-74% carotid artery stenosis to those with no stenosis, and participants with IMT in the 2 nd , 3 rd and 4 th quartiles to those with IMT in the 1 st . Separate comparisons were made for left and right carotid arteries. Wald tests were used to test the equality of left- and right-sided coefficients. The investigators hypothesized that since the majority of right-handed individuals are left-hemisphere dominant, disease of the left carotid artery should have a greater effect on cognitive dysfunction than disease in the right carotid artery. On the other hand, if carotid artery disease (CAD) is a marker of other underlying risk factors for vascular disease, then both left and right-sided CAD should be equally associated with cognitive dysfunction. Secondly, they hypothesized that CIMT of the left and right common carotid arteries should both be associated with cognitive dysfunction since they assumed that IMT was a marker for underlying vascular disease rather than a direct cause of cerebral ischemia and cognitive impairment. 32 (0.8%) of participants had high-grade stenosis ( ≥75%) of the left internal carotid artery, 2,438 (61.5%) had 1-74% stenosis, and 1,497 (37.4%) had no stenosis. Scores on both the modified MMSE and Digit Symbol Test were significantly lower and cognitive impairment was significantly more common in participants with high-grade left-sided stenosis compared to those with no stenosis [66]. For participants with right-sided stenosis, the difference in modified MMSE was also significant but less pronounced, while scores on the Digit Symbol Test were lower but not significantly. Participants with high-grade right-sided stenosis were also more likely to be cognitively impaired than those with no right-sided stenosis but the difference was not statistically significant. Greater IMT of both the left and 38 right arteries was associated with significantly lower scores on both the modified MMSE and the Digit Symbol Test, and with significantly more common cognitive impairment. Compared to participants with no left-sided stenosis, participants with ≥75% stenosis were 5.7 times more likely to be cognitively impaired (OR= 6.7, 95% CI = 2.4, 1.8) after adjustment for demographic and vascular risk factors including age, sex, ethnicity, education, hypertension, LDL and HDL cholesterol, diabetes, smoking and history of congestive heart failure as well as contralateral stenosis, and compared to participants with no right-sided stenosis, participants with ≥75% stenosis were 20% more likely to be cognitively impaired (OR= 1.2, 95% CI = 0.33, 4.7). Given the different patterns observed between left-sided and right-sided disease, and since the adjustment for stenosis of the right carotid artery did not change the associations between stenosis in the left carotid artery and cognitive impairment, the authors concluded that carotid artery stenosis is not simply a marker for vascular disease and its risk factors. Rather, the observations suggest that carotid stenosis is directly related to brain dysfunction resulting in cognitive impairment. Compared to participants with left-sided IMT in the 1 st quartile, participants with CIMT in the 4 th quartile were 50% more likely to be cognitively impaired (OR = 1.5, 95% CI = 0.98, 2.3), and 40% more likely to decline cognitively over the follow-up period (OR = 1.4, 95% CI = 1.0, 1.8). Since associations between IMT and cognition were similar for left- and right-sided disease, the authors concluded that IMT is a marker for underlying risk factors and generalized atherosclerosis rather than a direct cause of cognitive impairment. The CHS cohort was one of the older cohorts included in this review. Furthermore, since the study design did not exclude 39 individuals with history of CVD as did many of other studies reviewed, the cohort was relatively less healthy, at least with respect to cardiovascular health, compared to cohorts in other studies included in this review. The methods used to quantify IMT in the CHS study are the same as those used in the ARIC study and thus also differ from the USC methods. Thus, as was the case for the ARIC IMT measure, the CHS IMT measure and the USC CIMT measure are not exactly comparable [145]. Also similar to the ARIC method, CHS method to calculate mean IMT uses smaller number of measurements, suggesting that there may greater variability in their data. In addition, CHS methods involve subjectivity from readers who identify locations of maximum IMT along the artery. Similar to the ARIC studies, both scenarios could also result in greater “noise” in CHS IMT data and effect estimates that could be biased towards the null. Furthermore, the CHS method of quantifying the mean IMT using the identification of areas of maximum thickness probably also results in greater absolute estimates of subclinical atherosclerosis relative to the USC method. As stated above for the ARIC study, all subjects in the CHS cohort have IMT measured using this method, and there is no reason to think that differential misclassification by the cognitive outcome is an issue. However, similar to the scenario predicted for the ARIC studies, since it is possible that the CHS study measure of IMT is greater overall, then a greater thickness of IMT will be associated with a given level of cognitive performance than what might be found in the analyses using USC data. 40 Muller et al. (2006) examined cross-sectional associations between sub-clinical atherosclerosis and cognitive function in 396 independently living Dutch men with mean age 60 years. The investigators had measurements of CIMT (measured in both the left and right distal common carotid arteries), PWV and ABI, which were used to categorize study subjects as “sub-clinical CVD” if one of the following was true: CIMT or PWV in the highest quartiles of the study population distribution or evidence of PAD, as defined as an ABI <0.90. Subjects were also categorized as “no CVD” or as “prevalent CVD” if participants reported a physician diagnosis of coronary artery disease, PAD or stroke. Using this categorization scheme, 13.6% of participants had prevalent CVD and 31.3% were classified as having sub-clinical CVD. A battery of 10 cognitive tests was used to assess cognitive function and included: MMSE (global cognition), Rey Auditory Learning test (verbal episodic memory), Doors test (visual memory), Digit Span forward and backward (short-term and working memory), verbal fluency, Digit Symbol (cognitive and perceptual speed), Trail Making A&B (attention and mental flexibility) and the DART (verbal intelligence and per-morbid intelligence). Cognitive domains of memory (Rey and Doors tests), processing capacity/speed (Digit Span and Trails A) and executive function (verbal fluency and Trails B) were created by summing z-transformed scores of individual tests that assessed abilities in each domain. Compound scores were used as dependent variables in linear regression analyses to examine associations between cognition and categories of CVD. Age and education-adjusted mean scores were significantly lower on the Doors test and borderline significantly lower on the DART for subjects with prevalent or sub-clinical CVD compared to 41 those with no CVD [103]. There were no differences between categories of CVD for the other cognitive tests. As IMT increased in thickness (mm), scores significantly decreased in the memory domain ( β = -1.32, 95% CI = -2.63, -0.01) and decreased in the executive function domain ( β = -0.89, 95% CI = -2.16, 0.38), but did not change in processing capacity/speed domain ( β = 0.75, 95% CI = -1.60, 3.10). Compared to those with no CVD, subjects with sub-clinical CVD had significantly lower scores in the memory domain ( β = -0.45, 95% CI = -0.83, -0.07) and lower scores on both processing speed/capacity and executive function domains. Subjects with prevalent CVD had lower scores on memory and executive function compared to those with no CVD but not on processing speed. When subjects were grouped into MMSE score >28 versus ≤ 28, there was a significant difference between categories of CVD, with those categorized as having prevalent CVD having proportionately more with MMSE ≤ 28 and those categorized as having no CVD having proportionately more with MMSE >28. The authors did not find evidence of an interaction between atherosclerosis and age, suggesting that atherosclerosis has the same effect on cognition in men younger than 60 years and older than 60 years. While this study partially drew from a population-based sample, the final study population was comprised of volunteers. The investigators did not use any additional health-related eligibility criteria other than participants must be independently living and able to attend the clinic for examinations, nor did they disclose how study participants compared to the 70% who were invited but declined to participate. If study participants were healthier than the larger group, it is possible 42 that effect estimates could be underestimating the true associations between atherosclerosis and cognitive function. On the other hand, if participants were motivated to participate because of being in ill health (and participation meant access to better medical care) then it is possible that observed associations are actually overestimates of the true associations in the larger population of independently living Dutch men. As part of a cross-sectional study of the association between alcohol consumption and cognitive performance among 2,215 participants in the Northern Manhattan Study, Wright et al. (2006) analyzed whether maximal carotid plaque thickness (MCPT) mediates this relationship. Participants were 63% Hispanic, 20% black and 15% white, 59% women and with a mean age of 67.9 years and had never been diagnosed with a stroke. MCPT was measured in atherosclerotic plaques of the internal and common carotid arteries and bifurcations identified by ultrasound. 58% of study participants had at least one atherosclerotic plaque and the mean MCPT was 1.1 mm. Cognitive status was assessed using the MMSE; the mean score for the study population was 27. The authors reported that MCPT was not univariately associated with performance on the MMSE, but did not provide any additional information. When MCPT was added to generalized linear models of the association between categories of reported alcohol intake and performance on the MMSE that were adjusted for demographic, numerous CVD risk factors and depression there was essentially no change in the estimated ORs, indicating that MCPT did not mediate the association [181]. The authors did not examine prevalence or size of 43 carotid plaques as an independent variable. Given the mean MMSE score in the study population, a similar critique applies as discussed earlier with respect to the MMSE not being the best choice to discriminate between cognitive abilities among cognitively intact individuals. Woo et al. examined cross-sectional correlates of low ABI in a sample of 4,000 community-dwelling Chinese adult volunteers with a mean age of 72.5 years. PAD was defined as ABI <0.9 and was present in 6.9% of the sample. Cognitive function was assessed with the MMSE, and cognitive impairment was defined as MMSE score <24. The authors found that the 25.3% of adults who were classified as cognitively impaired had significantly lower mean ABI than those who were cognitively normal (1.08 ±0.12 vs. 1.02 ±0.13, respectively p=0.0005) [180]. In multivariate logistic regression models adjusted for age and gender only, cognitive impairment was significantly associated with ABI <0.9 (OR for PAD = 1.75, 95% CI = 1.33, 2.30). In an additional multivariate model that was co-adjusted for self- reported diabetes, hypertension, myocardial infarction, smoking, alcohol consumption, vitamin C intake and mobility, cognitive impairment remained significantly independently associated with low ABI (OR for PAD = 1.58, 95% CI = 1.19, 2.09). Unlike other studies included in this review, the investigators in the Woo et al. study looked at cognitive impairment as an independent variable and atherosclerosis (as measured by ABI) as the dependent variable. Thus, given the way the investigators analyzed the data, cognitive impairment independently predicted low ABI. However, since this was a cross-sectional study, we don’t have 44 information about the timing of cognitive impairment or atherosclerosis in relation to each other. Nevertheless, taken with those of other studies, the results of this study suggest a bi-directionality of the relationship between cognition and atherosclerosis. Thus, it is possible that a third factor not included in the analysis influences both atherosclerosis and cognition. Finally, the scenario described for the Muller et al. study could also apply to this study since the study population was comprised of volunteers. Price et al. (2006) examined the association between atherosclerosis measured at baseline and cognitive function assessed after 10 years in 700 adults aged 55-74 (mean age 63.6) participating in the population-based Edinburgh Artery Study (EAS). Two cut-points (0.90 and 0.95) were used to divide ABI into low and normal categories; 16.6% of participants had ABI ≤ 0.95). A battery of six neuropsychological tests were administered to subjects which included the NART (premorbid cognition), Logical Memory Subsets of the Wechsler Memory Scale (immediate and delayed verbal declarative memory), Raven’s Progressive Matrices (non-verbal reasoning), Verbal Fluency Test (verbal fluency – executive function) and the Digit Symbol Test (information processing speed). Test scores were divided into tertiles and age, sex and NART-adjusted odds ratios comparing performance in the lowest tertile to the highest were estimated using logistic regression methods. Subjects in the middle tertile of cognitive performance were dropped from these analyses. 12.4% of subjects in the study had symptomatic CVD, which was defined as a positive history of MI, angina, stroke or intermittent claudication and the mean 45 ABI was 1.06 (±0.16). Baseline ABI was significantly positively correlated with performance on the Digit Symbol test (r=0.13, p =0.01, adjusted for age, sex and NART), but not correlated with Delayed Memory, Raven’s Matrices or Verbal Fluency [125]. Adjusted mean scores on the Digit Symbol and Verbal Fluency were significantly or borderline significantly lower, respectively in subjects with ABI ≤ 0.95 compared to those with ABI > 0.95. Subjects with ABI either ≤ 0.95 or ≤ 0.90 were more likely to score in the lowest tertiles of Verbal Fluency and Digit Symbol than the highest compared to subjects with ABI either >0.95 or >0.90. Subjects with ABI ≤ 0.90 were 1.6 times more likely to score in the lowest tertile of the Digit Symbol Test than the highest tertile compared to subjects with ABI >0.90 (adjusted OR = 2.6, 95% CI = 1.2-5.5). A strength of this study is the prospective design in which atherosclerosis was measured a decade prior to assessing cognition, thus establishing some temporality of exposure and outcome. A major limitation to the study is the potential differential loss to follow-up. The original EAS cohort included almost 1,600 men and women; at the ten-year follow-up, almost 400 or one- quarter of the original participants had died. It is fair to assume that those participants who died were more likely to be in either the ABI ≤ 0.95 or ≤ 0.90 group. However, this loss would not necessarily affect the observed associations. More importantly, participants were invited for cognitive testing and of the 1,000 participants receiving invitations, 717 (65%) agreed to undergo cognitive assessment. Participants tested were younger and healthier with respect to CVD, CVD risk factors and mean ABI than the total cohort. If the participants who refused cognitive testing also did so because of anticipated difficulties with cognitive testing, 46 i.e., were more likely to have cognitive dysfunctions, then differential loss to follow- up may have occurred, where those most likely to have both lower cognitive function and atherosclerosis were not included in the final study sample. Therefore, the observed associations may be biased towards the null and that the true association between atherosclerosis and cognition may be greater. Hanon et al. (2005) conducted a study of 308 consecutive elderly ambulatory adults over age 60 with memory loss (mean age 78 ± 8 years) attending a geriatric outpatient clinic in France. Exclusion criteria were major depression, psychiatric deficits, other metabolic disorders and dementias other than AD and VaD. Arterial stiffness was assessed in participants during their hospital visit using measurements of carotid-femoral pulse wave velocity (PWV). Cognitive function was also evaluated during their clinic visit using the MMSE as well as the Cognitive Efficiency Profile (CEP), a battery of neuropsychological tests which included measures of immediate and delayed memory, language, visuoperceptual and visuospatial capacities, praxia, gnosia, executive function, attention and judgment with a maximum score of 100 (higher scores indicating better cognitive function). Patients were classified into four groups: Alzheimer’s disease (AD) [according to DSM-IV criteria [5]; n=126 (41%)], Vascular dementia (VaD) [using NINDS- AIREN criteria [136]; n=19 (6%)], mild cognitive impairment (MCI) [by Petersen criteria [124] and using results from the cognitive testing: memory and executive functioning scores<mean-1.5 SD; n=83 (27%)] and normal cognition [using results from the cognitive testing; scores> mean-1.5 SD; n=80 (26%)]. Results from an 47 ANCOVA adjusted for age, gender, SBP, education level, anti-hypertensive therapy and presence of CVD indicated that mean PWV was significantly higher in subjects with VaD (15.2 ± 3.9 m/s, p<0.001), AD (13.3 ± 2.9 m/s, p<0.001) or MCI (12.6 ± 2.6 m/s, p<0.01) than in those without cognitive impairment (11.5 ± 2.0 m/s) [57]. Multivariate regression models adjusted for age, gender, SBP and education level showed that for each 2 m/s increase in PWV, the odds of AD were 1.73 (95% CI = 1.27, 2.47) and VaD were 3.52 (95% CI = 1.87, 8.05) compared to normal cognitive function. Furthermore, for every 2 m/s increase in PWV, there was a significant decrease in MMSE ( β = -0.091, SE =0.028, p<0.001) and CEP ( β=-0.029, SE=0.009, p<0.001) scores, indicating that subjects with higher PWV have worse cognitive function. Results of the study indicate that atherosclerosis as measured by PWV is an independent predictor of cognitive function. Scuteri et al. (2005) conducted a hospital-based study of 84 elderly adults with memory deficits (mean age 78 ± 5 years) in Italy. Exclusion criteria were history of stroke or atrial defibrillation. Arterial stiffness was assessed in participants during their hospital visit using measurements of carotid-femoral PWV normalized for blood pressure. Cognitive function was evaluated using the MMSE during the hospital visit. Patients referred to the hospital for memory deficits were categorized into three groups based on CT brain imaging: normal brain imaging (22.6%), subcortical microvascular lesions (48.8%) and cortical atrophy (27.4%). A regression analysis adjusted for age, sex, education, diabetes, LDL and HDL cholesterol, prevalent CVD, anti-hypertensive medications and nitrates demonstrated 48 that PWV (m/s) normalized for blood pressure was significantly inversely correlated with MMSE ( β = -0.28, p<0.01) [143]. Results from an ANOVA revealed a significant trend in PWV between groups; the cortical atrophy group (14.2 ± 1.7) had higher PWV than the microvascular lesion (12.8 ± 2.1) or normal imaging groups (13.1 ±2.2) (p-trend = 0.02). This is the only study of arterial stiffness included in this review to normalize PWV values to mean arterial blood pressure (MBP) [PWV(norm) = 100 x (PWV/MBP)]. Blood pressure levels are known to significantly affect PWV values. Thus, normalization to blood pressure should minimize the contribution of blood pressure on the detected effect of PWV on cognition, and thus make the results more of a “pure” look at the effect of atherosclerosis on cognitive function. Fujiwara et al. (2005) conducted a cross-sectional analysis of 352 elderly community-dwelling volunteers for a preventive medicine-screening program in Japan over the age of 70 years and without a history of CVD. Arterial stiffness was assessed in participants using measurements of brachial-femoral PWV (cm/s), and divided into tertiles with the lowest tertile of PWV serving as the reference group. Cognitive function was evaluated using the MMSE, and scores were used to divide participants into two groups, normal cognitive function (MMSE ≥ 24, n=90.9%) and cognitively impaired (MMSE <24, n=9.1%). Univariate analyses indicated that cognitively impaired individuals were significantly more likely to have PWV in higher tertiles (p=0.002). Compared to individuals with PWV in the lowest tertile (<1,750 cm/s), those with PWV in the middle tertile (1,750-2,070 cm/s) were 9.66 49 times as likely (95% CI = 1.15, 80.93, p=0.037) and those with PWV in the highest tertile (>2,070 cm/s) were 3.42 times as likely (95% CI = 0.37, 31.96, p=0.282) to be cognitively impaired after adjusting for age, sex, education, pulse pressure, ABI > 1.0, hemoglobin A1c, albumin and history of hyperlipidemia [48]. Thus, the finding that the highest risk of cognitive impairment was in middle tertile of PWV and PWV in the highest tertile was not associated with an increased risk of cognitive impairment suggest that PWV is not related to cognitive function in dose-dependent manner. However, this was a study of elderly volunteers; it is possible that individuals who would have been in the highest tertile of PWV and also cognitively impaired did not volunteer for the study because they were too ill to participate, or were deceased. Therefore, it is possible that the adults with PWV in the highest tertile who were included in the study represent a healthier group than the larger population of which they are a part. Selective survival bias would suggest that the observed OR could be underestimating the true OR. Furthermore, this study included only 32 individuals (9% of the study population) with cognitive impairment. Finally, the estimated confidence intervals are very wide, implying that the observed ORs should be interpreted with caution. Fukuhara et al. (2006) conducted a cross-sectional analysis of 203 elderly participants all aged 85 years in a population-based cohort study in Japan. Arterial stiffness was assessed by brachial-ankle PWV (m/s). Cognitive function was evaluated using the MMSE, and scores were used to divide participants into two groups, normal cognitive function (MMSE ≥ 24, n=63.1%)) and cognitively 50 impaired (MMSE <24, n=36.9%). Univariate analyses showed that cognitively impaired individuals were significantly more likely to have higher PWV (25.0 ± 0.8 m/s) than cognitively normal subjects (22.9 ± 0.5 m/s) (p<0.05) [49]. PWV (m/s) was significantly negatively associated with MMSE score after controlling for sex, education, hemoglobin A1c, BMI, smoking, systolic blood pressure, pulse pressure, habitual alcohol intake, total serum cholesterol and left ventricular hypertrophy ( β = -0.157 in MMSE score per m/s PWV, p=0.003). When subjects with ABI < 0.9 were excluded from the analysis, PWV was still significantly negatively associated with MMSE scores and the association was strengthened ( β = -0.196 in MMSE score per m/s PWV, p=0.001). The lack of clarity in the methods section of this paper regarding study procedures and analytic methods made the study difficult to evaluate and was a major limitation of this study. One study in the literature examined the cross-sectional association between cognition and blood vessel function as a surrogate measure of general atherosclerotic risk burden. Moser et al. (2004) obtained plethysmographic measures of forearm blood flow before and after infusion with vasoactive agents in 14 volunteers (7 men and 7 women), mean age 73 ± 6 years who all had unequivocal diagnoses of atherosclerotic vascular disease and a history of at least one of the following: angina, myocardial infarction, coronary angioplasty, coronary artery stent and peripheral vascular disease [102]. Exclusion criteria included coronary artery bypass grafting, valve replacement, carotid endarectomy, stroke, head injury with loss of consciousness, other neurological disorders, diagnosis of dementia or severe 51 psychiatric disorder. Acetycholine was administered to measure endothelium- dependent vasodilation, nitroprusside to measure endothelium-independent vasodilation and verapamil to measure smooth muscle function. The three agents were infused in one arm only. The ratio of blood flow in the infused arm versus the noninfused arm relative to flow in the two arms at baseline was calculated as a measure of the degree to which forearm resistance vessels can dilate when induced to do so, with higher values indicating healthier vessel function. Neuropsychological functioning was assessed using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), which includes measures of five cognitive domains: immediate and delayed memory, language, attention and visuospatial/visuoconstructive skills. The Beck Depression Inventory II was used to assess depressive symptoms. Partial correlation coefficients estimated the associations between RBANS total score and the percent change in blood vessel flow under each of the three agents. All measures of blood flow were positively associated with cognition. Controlling for age, correlations between RBANS total score and percent change in blood flow with verapamil was 0.73 (p=0.005), with acetylcholine was 0.39 (p=0.184) and with nitroprusside was 0.42 (p=0.158). Controlling for depression, correlations between RBANS total score and percent change in blood flow with verapamil was 0.78 (p=0.002), with acetylcholine was 0.51 (p=0.076) and with nitroprusside was 0.56 (p=0.048). The authors also tested correlations between neuropsychological performance and conventional vascular- related variables including total, HDL and LDL cholesterol, triglycerides, systolic and diastolic blood pressure, and glucose, and found the associations, in general, to 52 be much weaker and non-statistically significant than those with blood flow. The authors concluded that the findings provided evidence of a relationship between resistance vessel function and neuropsychological performance. The strong relationship between verapamil and cognition they interpreted to suggest that smooth muscle function may share a stronger relationship with neuropsychological status than does endothelial function, however, concluded that larger studies are necessary to further elucidate this relationship. This was a pilot study among a small number of volunteers. Simple analyses were conducted which were not simultaneously adjusted for a number of other factors known to influence cognitive performance. Therefore, while the results of this study suggest that the relationship between cognition and vessel function is stronger than that for conventional factors, the preliminary nature of the study requires that the results be interpreted with caution. One study used the medical histories of participants to count the number of CVD pathologies, which was then used as a surrogate measure for burden of atherosclerosis in analyses of cross-sectional associations with cognitive impairment. Van Exel et al. (2002) studied 599 participants in the population-based Leiden-85 study in the Netherlands who were all 85 years old. Total number of cardiovascular pathologies in the categories 1) arterial surgery, 2) stroke, 3) intermittent claudication, 4) MI, 5) angina or myocardial ischemia were counted, and a summary variable indicating burden of atherosclerosis was created with possible values 0, 1 and ≥ 2. Cognitive function was assessed using the MMSE, the Letter Digit Coding test (processing speed), the 40-item abbreviated Stroop test (attention) and the 12- 53 Word Learning test (immediate and delayed memory). Subjects were classified as cognitive impaired (MMSE scores <28) or cognitively normal (MMSE ≥ 28). Median values on the other tests were used to categorize subjects into good and poor cognitive ability groups in the cognitive domains assessed by the measures. Severe cognitive impairment (MMSE ≤ 18) was present in 16.6% of subjects; 38% of subjects had no CVD pathology, 34% had one, and 28% had two or more CVD pathologies. The mean MMSE score was significantly lower as categories of CVD burden increased; the mean score for zero CVD pathologies was 26 (26-27), for one CVD pathology it was 26 (26-27) and for two CVD pathologies it was 25 (24-27), p- trend = 0.003 [165]. The odds of cognitive impairment increased with increasing CVD pathologies (p-trend = 0.003); compared to subjects with zero CVD pathologies, subjects with one CVD pathology were 20% more likely to be cognitively impaired (OR=1.2, 95% CI = 0.8-1.8), and subjects with two CVD pathologies were twice as likely to be cognitively impaired (OR = 2.0, 95% CI = 1.3- 3.0). Poor cognitive speed also increased over strata of CVD pathologies, with an OR of 1.2 (0.7-1.8) for one pathology and 2.1 (1.2-3.5) for two (p-trend = 0.07). Memory was not affected by the number of CVD pathologies. Effect estimates were reportedly not altered by the addition of education, depressive symptoms and use of NSAIDS or CVD drugs to regression models. ORs increased and maintained their significance when subjects with a history of stroke were excluded from analyses. A strength of this study was that it was population-based. The inclusion of elderly adults 85 years or older who are survivors from a larger birth cohort does raise the question as to how informative the studies’ findings are for the general population. 54 One study in the literature examined the cross-sectional association between atherosclerosis and cognition using retinal microvascular abnormalities as a surrogate measure of atherosclerosis. Middle-aged African-American and Caucasian participants in the ARIC study who had negative histories of stroke, had not used central nervous system-relevant medications and for whom retinal photographs were obtained (n=7,526, mean age 53.8 years) were included in a study by Wong et al. (2002). Retinal photographs were evaluated for microvascular abnormalities which included retinopathies (lesions including microaneurysms, retinal hemorrhages, soft and hard exudates, macular edema, intraretinal microvascular abnormalities, venous beading, new vessels at the disc or elsewhere, vitreous hemorrhage, disc swelling or laser photocoagulation scars), arteriovenous nicking, focal and general arteriolar narrowing (as indicated by the arteriole-to-venule ratio). Cognitive function was assessed with Delayed Word Recall, Digit Symbol and Word Fluency tests. Cognitive impairment was defined as 2 SDs below the mean scores on each test for the study population; 149 (2.0%) subjects were identified using this criterion on the Word Fluency Test; 260 (3.5%) using the Digit Symbol Test and 299 (4.0%) using the Delayed Word Recall Test. Multivariable logistic regression methods were used to estimate odds ratios of cognitive impairment while adjusting for possible confounding factors; separate analyses were conducted using cut offs from each of the three tests to define the outcome of cognitive impairment. Mean scores on all three tests were significantly lower among participants with any retinopathy, retinal hemorrhage, microaneurysm or soft exudates compared to participants without these lesions after adjusting for age, sex, race, study center, education, occupation, 55 diabetes, fasting glucose, hypertension, mean arterial blood pressure, IMT, smoking, alcohol consumption, fasting and total HDL cholesterol levels and triglyceride levels [179]. There were no differences in mean cognitive scores of participants with arteriovenous nicking or generalized arteriolar narrowing compared to those without the abnormalities. Participants with focal narrowing actually scored higher on the Digit Symbol and Word Fluency tests and significantly higher on the Delayed Word Recall test compared to those without the abnormality, which could indicate that residual confounding was present despite efforts to adjust for confounders or could be chance findings. The odds of cognitive impairment were significantly higher among participants with any retinopathy compared to those without, using Delayed Word Recall (OR = 2.60, 95% CI = 1.70-3.99), Digit Symbol (OR = 1.91, 95% CI = 1.04-3.49) or Word Fluency (OR =2.03, 95% CI =1.07-3.86) tests to define cognitive impairment after controlling for the covariates mentioned above. Odds of cognitive impairment were also elevated or significantly elevated for participants with retinal hemorrhage, microaneurysm, soft exudates, arteriovenous nicking or generalized arteriolar narrowing compared to those without. The Delayed Word Recall test was the cognitive measure that was most strongly and consistently associated with cognitive impairment. There was no significant interaction by diabetes or hypertension and results were similar in analyses by subgroups of age, sex and race. The change in cognitive scores over the 6-year follow-up period of the study was small and thus analyses of change in cognitive function over time were not conducted. It was difficult to determine which participants were included in the various analyses from the tables provided because the number of participants 56 differed between tables and in at least one instance, did not add up to numbers presented in a previous table. The authors did not provide an explanation to account for the discrepancies in the numbers. Summary In conclusion, this literature review provides fairly consistent evidence supporting an association between subclinical atherosclerosis and cognitive function in non- cognitively impaired and non-demented populations. The strength of the evidence derives from the consistency of findings 1) across a number of demographically diverse populations, including a number of ethnic/racial groups (Caucasians in European and the US, African-Americans, and Asian populations in Japan and China) as well as in both middle-aged and elderly populations, 2) using various methods to assess atherosclerosis at a number of vascular sites and at various stages of severity along the continuum of atherosclerotic disease, and 3) using a number of different measures of cognitive abilities that assess different domains of cognition. The associations between subclinical atherosclerosis and cognitive function reported in the literature vary in magnitude. There are a number of possible reasons for this including differences in the populations studied, the units of measurements of 57 atherosclerosis and the definition of the cognitive outcome. If the proposed biological model, in which the cumulative effect of subclinical atherosclerosis was an important contributing factor in determining an individual’s level of cognition at any given time, were true then one might predict that subclinical atherosclerosis earlier in the disease continuum (eg. IMT) should be more strongly correlated with small differences in cognitive function than with cognitive impairment. One would also predict from this model that subclinical atherosclerosis at more advances stages of disease (eg. carotid plaques) should be more strongly correlated with cognitive dysfunction of greater severity (eg. cognitive impairment or dementia) than with milder differences in cognitive function. Only one study in the review had measures of both IMT and carotid plaques for the same population, allowing for an examination of this hypothesis [61]. Consistent with the prediction, plaques in the carotid arteries were more strongly associated with all types of dementia than was wall thickness of the carotid arteries. Unfortunately the authors did not look at MMSE scores continuously so the IMT-cognitive functioning portion of the above argument cannot be examined. While the different methods used to assess atherosclerosis add to strength of evidence of an association between subclinical atherosclerosis and cognitive function, for some methods of measuring atherosclerosis, there was only one or two studies that assessed the relationship with cognition. Thus, in some instances there were just not sufficient numbers of studies to allow comparisons be made between studies. In addition, the reviewed studies varied in the number and choice of 58 covariates included in analyses, also limiting their comparability. Education and depression/depressive symptoms are two important covariates, which are likely to confound observed associations between atherosclerosis and cognition. There is an expected inverse association between atherosclerosis and cognitive function, such that a greater extent of atherosclerosis is associated with lower cognitive function. Effect estimates from studies that failed to adjust analyses for education probably overestimate the inverse association (observed ORs would be negatively biased away from the null), since education is positively associated with cognitive function and negatively associated with atherosclerosis. Studies that failed to adjust for depression probably also overestimate the inverse association between atherosclerosis and cognitive function, assuming depression is frequently found to be negatively associated with aspects of cognitive function, and positively associated with atherosclerosis. Finally, as alluded to above, while neuropsychological tests are used to assess specific cognitive abilities, many tests including those used by studies in this review do not “cleanly” measure a single cognitive ability. Thus, there is some degree of overlap in what cognitive abilities the tests assess. Therefore, the use of multiple individual tests as outcomes or composite scores that combine several tests as cognitive outcomes can be expected to result in outcomes that are correlated. One methodological solution would be to use a correction in the alpha level for multiple comparisons. None of the studies included in this review took such an approach. 59 Questions that Remain Unanswered After reviewing the literature on atherosclerosis and cognition, a number of questions remain to be addressed. Are the observed associations between CVD risk factors and cognitive function mediated by atherosclerosis? Are certain areas of cognition more affected by subclinical atherosclerosis than others? At what stage in the atherosclerotic process and at what level of atherosclerosis severity are cognition and different cognitive domains affected? Establishing a timeline for when atherosclerosis may begin to exert an effect on cognitive functioning is critical for an understanding of the mechanisms by which atherosclerosis may act as well as for developing and targeting appropriate interventions. An ideal study to address a number of these issues would use more than one method to measure atherosclerosis and would assess several different cognitive domains over at least ten years of follow-up time in a large population-based sample that varied in the extent of subclinical atherosclerosis. 60 Chapter 2 - Subclinical Atherosclerosis Contributes to Differences in Cognitive Function in Healthy Adults without Cardiovascular Disease Abstract Atherosclerosis is the most common pathologic process underlying cardiovascular disease (CVD). It is not well known whether subclinical atherosclerosis is an independent risk factor for lower cognitive function among individuals without clinically evident CVD. We examined cross-sectional associations between subclinical atherosclerosis and cognitive function in a community-based sample of healthy adult participants enrolled in the BVAIT study (n=504, mean age 61 years). Carotid artery intima-media thickness (CIMT), coronary (CAC) and abdominal aortic calcium (AAC) were used to measure subclinical atherosclerosis. Cognitive function was assessed with a battery of neuropsychological tests. A principal components analysis was used to extract five uncorrelated cognitive factors from scores on individual tests, and a measure of global cognition was derived. Multivariable linear regression was used to examine the association between subclinical atherosclerosis and cognitive function, adjusting for other correlates of cognition. Increasing thickness of CIMT was associated with significantly lower scores on the verbal learning factor ( β = -0.07 per 0.1 mm increase CIMT [SE( β)=0.03], p=0.01). CAC and AAC were not individually associated with any of the cognitive factors. This study provides evidence that increasing CIMT is weakly 61 associated with lower verbal learning abilities in a population of healthy middle-to- older aged adults without clinically evident CVD. Introduction Interest in the association between cardiovascular disease (CVD), its risk factors and cognitive impairment, cognitive decline and dementia [41] has been building. Factors such as hypertension [159], diabetes [82] and elevated homocysteine [132] that are known to increase CVD risk, have also been linked to diminished cognitive function. Furthermore, there is a growing consensus that vascular disease is part of the pathology of Alzheimer’s disease (AD), which accounts for the majority of cases of dementia, and that AD and vascular dementia, often considered the second most common type of dementia, may not be completely separate entities as previously believed [87]. While atherosclerosis is the most common pathologic process underlying CVD [8], few community-based studies have examined the relationship between atherosclerosis and cognition [21,29,66,104]. Furthermore, most studies have focused attention on late stages of atherosclerosis when studying elderly demented populations with clinical CVD [38]. As such, it is not well known whether subclinical atherosclerosis is an independent risk factor for impaired cognitive function among non-demented individuals without clinically evident CVD. Establishing a timeline for when atherosclerosis may begin to exert an effect on cognitive function is critical for an understanding of the mechanisms by which 62 atherosclerosis may act as well as for developing and targeting appropriate interventions [75]. To address these questions, we examined cross-sectional associations between subclinical atherosclerosis, using measures of carotid artery intima-media thickness (CIMT), coronary artery calcium (CAC) and abdominal aortic calcium (AAC), and cognitive function assessed with a battery of neuropsychological tests in a community-based sample of healthy hyperhomocysteinemic middle-aged and older adults. Methods Study participants Healthy hyperhomocysteinemic adults > 40 years old who were randomized in the B-Vitamin Atherosclerosis Intervention Trial (BVAIT) were the focus of the present study. Data obtained for participants at their baseline visit were used in the current analysis. Briefly, men and postmenopausal women > 40 years old were eligible for BVAIT if they had Hcy >8.5 µmol/L. Of 5,309 individuals who were prescreened by telephone, 4,803 were ineligible or refused to be enrolled. Exclusions (n=1,039) were made for any clinical signs or symptoms of CVD (n=151), diabetes mellitus or fasting serum glucose >126 mg/dL (n=131), triglyceride (TG) levels >500 mg/dL (n=2), hypertension [systolic blood pressure (SBP) >160 mmHg and/or diastolic 63 blood pressure (DBP) >100 mmHg)] (n=11), untreated thyroid disease (n=2), creatinine clearance <70 ml/min (n=4), a life threatening disease with prognosis <5 years (n=113), alcohol intake of >5 drinks per day/substance abuse (n=1) or unwillingness to stop taking vitamin supplements (n=624). A total of 506 subjects were randomized in BVAIT; all signed a written informed consent approved by the Institutional Review Board at the University of Southern California. Measurement of Subclinical Atherosclerosis CIMT Using high resolution B-mode ultrasound, the right common carotid artery (CCA) was imaged using methods described previously [59]. An image analyst measured CIMT of the distal CCA far wall with automated computerized edge detection using an in-house software package (Prowin, patents 2005, 2006), as described elsewhere [146]. CIMT was the average of approximately 70-100 individual measurements between the intima–lumen and media–adventitia interfaces along a 1-cm length just distal to the carotid artery bulb. Coronary and abdominal aortic calcium Multidetector spiral computed tomography methodology using an Mx-8000 4-S-CT scanner (Philips, formerly Marconi, Cleveland, Ohio) was used to image the coronary arteries and thoracic abdominal aorta using a method similar to that of Carr et al. [24]. Contiguous, noninterlaced slices were acquired with a table increment of 64 Measurement of Cognitive Function A battery of cognitive and neuropsychological tests [90] was administered in a standardized order to all subjects by one trained psychometrist. The battery was designed to assess a broad array of cognitive functions and abilities, particularly episodic memory and executive tasks thought to be vulnerable to aging, and included the following tests: • Symbol Digit Modalities Test (SDMT): complex scanning and visual tracking, processing speed, attention • Trail Making Test Part B (Trails-B): visual conceptual and visuomotor tracking, divided attention, cognitive flexibility, psychomotor speed • Judgment of Line Orientation, Form H (JLO): visuospatial ability • Block Design [Wechsler Adult Intelligence Scale, 3 rd Edition (WAIS-III)]: visuoconstruction ability 65 • Letter-Number Sequencing [Wechsler Memory Scale, 3 rd Edition (WMS-III)] (LNS): attention, concentration, and working memory • Category fluency (animal naming, 60 seconds) (Animals): category fluency • Boston Naming Test, 30-item version (BNT): naming/semantic memory • Shipley Institute of Living Scale (Shipley), Abstraction Subset: general intellectual functioning • California Verbal Learning Test, 2 nd edition (CVLT-II), immediate recall (IR) and delayed recall (DR): verbal learning, conceptual ability and memory • Logical Memory I and II (paragraph recall, IR and DR) (WMS-III): logical memory • Faces I (IR) and II (DR) (WMS-III): nonverbal memory, visuospatial processing. The Center for Epidemiologic Studies Depression Scale (CES-D) scale was used to assess mood [127]. Of the 506 randomized subjects, two did not have cognitive testing; 504 (99.6%) subjects were included in the present study. Of the 504 subjects, 7 (1.4%) did not have measures of AAC, either due to technical problems that prevented reading or scoring of the scan (n=4) or missing phantom scores (n=3), and thus were not included in analyses that used AAC. 66 Biochemical and Behavioral Factors Blood pressure, body height and weight were measured and body mass index (BMI) was calculated (kg/m 2 ). Blood samples were drawn after a minimum 8-hour fasting period. Total fasting Hcy was determined in plasma using reverse phase HPLC with a C18 column on a Waters HPLC instrument equipped with a WISP automatic injector and attached to a fluorimeter [6]. A solution containing 40 nmol/ml Hcy was used as a standard and for column calibration. For quality control, pooled plasma spiked with different quantities of Hcy was used. Coefficient of variation for the assay was 7.8%. Total cholesterol, total TG, high-density lipoprotein cholesterol (HDL-C) and low- density lipoprotein cholesterol (LDL-C) (calculated) were measured by standardized enzymatic assay methodology [92]. Fasting serum glucose levels were measured using the glucose oxidase technique on a Beckman Glucose II analyzer (Beckman Instruments, Brea, California). Smoking questionnaires were used to determine smoking status. Statistical Analysis Characteristics of subjects and mean values for CVD risk factors were summarized. For subjects (n=24) who were unable or refused to complete one or more tests in the battery, age-gender- and education-specific mean values from the BVAIT study population were imputed. Small reductions (on the order of 1.2% – 1.6%) in variances of the tests resulted from imputations, which were made for < 0.7% of the 67 total number of tests. Mean CIMT was analyzed its continuous form, and a variable was created for CIMT categorized into quartiles. CAC and AAC were categorized as either present or absent (calcium score >0 versus 0). A composite variable indicating burden of subclinical atherosclerosis was created by summarizing a dichotomous CIMT variable ( ≥3 rd quartile vs. < 3 rd quartile), CAC and AAC, and ranged from 0 (no on all three) to 3 (yes for all three). For data reduction purposes, a principal components analysis with an orthogonal varimax rotation was performed on the 14 cognitive tests in the neuropsychological battery and consecutive uncorrelated factors were extracted. Following methods of Cattell [25], a scree plot of successive eigenvalues was used to identify at what number of principal components the plot leveled off; this led to a decision to retain five factors. The five factors accounted for 72.4% of the total variance, and for descriptive purposes, each were assigned a name that reflected high factor loadings of individual cognitive tests (i.e., loadings with an absolute value > 45). The resulting factors generally reflected cognitive abilities in areas of 1) executive function (high factor loadings on SDMT, Trails-B, LNS, JLO, Block Design and Shipley), 2) verbal learning (high factor loadings on CVLT- IR and DR), 3) logical memory (high factor loadings on paragraph recall - IR and DR), 4) visual episodic memory (high factor loadings on Faces I and II), and 5) semantic memory (high factor loadings on Animals and BNT) (Table 4). For each subject, a factor score for each of the five component factors was calculated. In addition, we created a measure of global cognition, which was calculated as the weighted sum of scores 68 on each of the individual tests in the neuropsychological battery; weighting of each test was based on the sum of the inverse covariances of scores with other tests. The resulting measure was then divided by its standard deviation (SD) so as to interpret results per SD of global cognition and ensure a consistent interpretation of results with those of the five cognitive factors. Table 4. Factor loadings of tests in the neuropsychological battery from principal components analysis Cognitive Factor Neuropsychological Test Executive Function Verbal Learning Logical Memory Visual Memory Semantic Memory Symbol Digit 0.66515 0.31613 0.16422 0.23729 0.09245 Paragraph Recall (Immediate) 0.17672 0.07565 0.89928 0.09454 0.10640 Faces (Immediate) 0.08397 0.20812 0.06988 0.87210 0.07339 Letter Number Sequencing 0.60418 0.10678 0.32011 0.09214 0.08590 Judgment of Line Orientation 0.62112 -0.14949 -0.06180 0.06530 0.34471 Boston Naming Test 0.41489 0.04893 0.17304 0.22991 0.63973 Paragraph Recall (Delayed) 0.16027 0.17168 0.88698 0.03936 0.13593 Faces (Delayed) 0.22149 0.10689 0.06357 0.85362 0.13383 CVLT (Immediate) 0.14911 0.87059 0.14813 0.15518 0.15820 Block Design 0.69538 0.08762 0.01813 -0.01488 0.34970 Animals 0.16695 0.26870 0.16468 0.07711 0.79169 Shipley 0.75950 0.16916 0.18321 0.10221 0.19000 CVLT (Delayed) 0.18793 0.87410 0.10605 0.17741 0.10282 Trail Making Test -0.76922 -0.19433 -0.13847 -0.17427 0.05413 Multivariable linear regression methods used the five factor scores and the global cognitive measure as dependent variables to examine the association with subclinical atherosclerosis (for CIMT: continuous, quartiles categories and dichotomous forms; dichotomous variables for CAC and AAC) and atherosclerosis burden (composite 69 variable). Separate models were run for each cognitive measure. Beta coefficients ( β) and their standard errors were estimated from regression models to assess the association between atherosclerosis and cognition. For the continuous CIMT variable, β represented the average unit change in the factor score per 0.1 mm of CIMT. For other categorical atherosclerosis variables, βs represented differences from the mean of a reference group (subjects with atherosclerosis at the lowest level or without atherosclerosis as defined by the variable). Models were adjusted for demographic characteristics that were significant univariate predictors of global cognitive function: age (40-46, 47-54, 55-60, 61-66, 67-74 and ≥75 years), gender, race/ethnicity (indicator variables for Caucasian, African-American, Latino and Asian-American/Pacific Islander/Native American), highest educational level achieved (high school or less, some college, Bachelor’s degree and graduate/professional degree), household income (<30,000, 30,000-49,999, 50,000- 69,999, 70,000-99,999 and ≥100,000 dollars/year) and mood (CES-D = 0, 1-3, 4-8, 9-20 and >20). Continuous variables including LDL-C, SBP and fasting Hcy for which a clearly linear relationship with cognitive function could not be established in our population or for which previous investigators have reported non-linear relationships [134,184] were categorized and modeled as a series of indicator variables. Potential confounding of the atherosclerosis-cognition association by CVD risk factors including LDL-C, Hcy and SBP, and smoking was controlled for by including these factors in the regression models. We explored whether gender, age and education level were effect modifiers by including interaction terms for each of these variables with the independent subclinical atherosclerosis variable in the 70 models. If the interaction was significant (p-value of <0.05), additional models were run by the stratifying factor. All analyses used SAS version 9.0 (SAS Institute Inc., Cary, NC, USA.). Results Characteristics of study subjects and mean values of CIMT, CAC and AAC are summarized in Table 5. The majority of the study subjects were men (61.1%), non- Hispanic White (64.7%) and highly educated (59.3% with a Bachelor’s or graduate degree). Most subjects were overweight (mean BMI = 28.1 kg/m 2 ); 38% reported having smoked currently or in the past. Table 5. Baseline Characteristics for BVAIT Subjects with Cognitive Testing (n=504) Variable Mean ± SD or Number (%) Age (years) 60.8 ± 9.9 40-49 71 (14.1%) 50-59 147 (29.2%) 60-69 182 (36.1%) 70-79 95 (18.9%) 80+ 9 (1.8%) Gender Male 308 (61.1%) Female 196 (38.9%) Race/Ethnicity Non-Hispanic White 326 (64.7%) Non-Hispanic Black 75 (14.9%) Hispanic 55 (10.9%) Asian/Pacific Island/Native American 48 (9.5%) 71 Table 5, continued Variable Mean ± SD or Number (%) Educational Level High school or less 60 (11.9%) Some college 145 (28.8%) Bachelor’s degree 131 (26.0%) Graduate/professional degree 168 (33.3%) Current/Former Smoker 191 (38.0%) Body-Mass Index (kg/m 2 ) 28.1 ± 5.0 Blood Pressure (mmHg) Systolic 129.6 ± 16.9 Diastolic 80.8 ± 10.4 Total Cholesterol (mg/dL) 221.2 ± 39.4 LDL Cholesterol (mg/dL) 138.5 ± 36.4 HDL Cholesterol (mg/dL) 56.8 ± 15.2 Triglycerides (mg/dL) 134.4 ± 132.0 Glucose (mg/dL) 100.0 ± 11.9 Homocysteine (nmol/mL) 10.3 ± 3.0 CIMT (mm) 0.75 ± 0.15 Aortic Calcium, present* 347 (70.0%) Coronary Calcium, present 222 (44.1%) CES-D score 6.3 ±6.7 Depression (CES-D >16) 37 (7.3%) * n=497 † n=500 Increasing thickness of CIMT was associated with significantly lower scores on the verbal learning factor ( β = -0.07 per 0.1 mm increase in CIMT [SE( β)=0.03], p=0.01), but not in other areas of cognitive function (Table 6). Compared to subjects with CIMT in the lowest quartile (<0.65 mm), subjects with CIMT in the highest quartile ( ≥0.83 mm) performed significantly lower on verbal learning ( β = -0.26 72 [SE( β)=0.12], p=0.04). The presence of CAC or AAC was not significantly individually associated with lower scores on any of the cognitive factors, or on the measure of global cognition (results not shown). Global cognitive abilities were lower, but not significantly, among individuals with evidence of subclinical atherosclerosis at three vascular sites (i.e., the greatest burden) compared to individuals with evidence at two or less sites (Table 6). Table 6. Associations between Cognitive Factor Scores and Subclinical Atherosclerosis Measures from Regression Models* for 504 BVAIT Subjects 73 CognitiveFactor β(SE), p-value† Subclinical Atherosclerosis Measure Executive Function Verbal Learning Logical Memory Visual Memory Semantic Memory Global Cognition Mean CIMT (per 0.1 mm) 0.02 (0.03) 0.44 -0.07 (0.03) 0.01 - 0.01 (0.03) 0.81 0.03 (0.03) 0.38 - 0.04(0.03) 0.15 -0.02(0.03) 0.44 Composite score‡ 3 vs. 0-2 -0.07(0.12) 0.57 0.01 (0.13) 0.94 -0.17 (0.14) 0.22 -0.05 (0.14) 0.71 -0.16 (0.13) 0.22 -0.19 (0.11) 0.09 *Adjusted for age, gender, race/ethnicity, education, income, CES-D score, Hcy, SBP, LDL-C, smoking status † p-value for comparison of one category to reference category ‡ n=498, composite score dichotomized at 3 versus 0, 1 and 2 74 Given the significant inverse association between CIMT and verbal learning and the significant main effects of age, gender and education on cognition, effect modification by age, gender and level of education was explored. Associations between CIMT and verbal learning differed significantly by gender (p=0.02) but not by age or education (p-values interaction >0.05). An association between increasing CIMT and poorer verbal learning was detected in men ( β =-0.10 per 0.1 mm increase CIMT [SE( β)=0.03], p=0.004) but not in women. Compared to men with CIMT <3 rd quartile, men with the thickest CIMT (CIMT ≥3 rd quartile) scored almost 1/3 SD lower on verbal learning ( β =-0.31 [SE( β)=0.13], p=0.016) (Table 7). Table 7. Associations between Subclinical Atherosclerosis and Cognitive Performance from Gender Stratified Adjusted Regression Models Subclinical Atherosclerosis Measure Verbal Learning factor β [SE( β)], p-value Male (n=308) Female (n=196) p(interaction) Mean CIMT (per 0.1 mm) -0.10 (0.03), 0.004 0.04 (0.07), 0.56 0.02 CIMT ≥0.83 mm vs. <0.83 mm -0.31 (0.13), 0.016 0.19 (0.17), 0.26 0.006 Discussion This study provides evidence that increasing thickness of CIMT is weakly associated with lower verbal learning abilities among healthy middle-to-older aged individuals without clinically evident CVD but with elevated Hcy. The observed association remained robust after adjustment for CVD risk factors, including LDL-C, Hcy, SBP and smoking, suggesting that CIMT is an independent correlate of verbal learning. 75 Arterial calcium measures of subclinical atherosclerosis were not associated with lower cognitive function in any area assessed. Results from this study also suggest that individuals with the greatest burden of subclinical atherosclerosis may have lower global cognitive abilities compared to those with less burden. Taken together, our findings suggest that in healthy, non-demented adults, early subclinical atherosclerosis (i.e., increasing thickness of CIMT) is associated with lower verbal learning, whereas later stage subclinical atherosclerosis (calcifications of the arteries) is not. We did not find evidence that executive function, logical, visual and semantic memory were associated with subclinical atherosclerosis. The associations observed in this population were small. Lower verbal learning performance per 0.1 mm CIMT corresponds to approximately 7% of an SD in the factor score, which is still well within the range of what is considered clinically “normal” [113]. The difference in verbal learning performance (nearly 30% of a SD lower) was more pronounced comparing individuals with CIMT in the highest quartile to the lowest. Stratified analyses suggest that the association between subclinical atherosclerosis and cognition may differ by gender, with an association between increasing CIMT and lower verbal learning apparent for men but not for women. This observation could be explained by the possibility that a threshold in the level of CIMT exists which must be reached for cognition associations to be observed. It is possible that this threshold was reached in men but not women in the study population, a possibility supported by the fact that mean CIMT was greater for male subjects than female subjects. 76 Given the statistical tests performed, we cannot rule out the possibility that the association detected between increasing CIMT and lower verbal learning abilities could be due to chance. However, by employing a principal components analysis to derive uncorrelated cognitive factors, we minimized multiple statistically significant findings due to correlated outcomes. Furthermore, our objectives were to address questions about different stages of subclinical atherosclerosis as well as different areas of cognitive function, thus necessitating multiple analyses. Atherosclerosis could be acting through one or more possible mechanisms to exert an effect on cognitive function. Endothelial dysfunction represents the earliest stage of atherosclerosis [91]. Endothelial cells make up one component of the blood-brain barrier (BBB), which functions to maintain a constant intracerebral milieu critical for proper brain function [60]. By disrupting the function of endothelial cells, atherosclerosis could theoretically increase permeability of the BBB [100,149]. Breaches to this barrier could allow neurotoxins and other substances from which the brain is normally protected into the cerebral environment. Over time, the cumulative effect of these substances within the brain could result in detectable losses of cognitive abilities. The diminished delivery of oxygen resulting from reductions in blood flow is another possible mechanism by which atherosclerosis may affect cognition. Ruitenberg et al. [140] showed that greater cerebral blood flow velocity was related to a lower prevalence of cognitive decline and dementia. Results of that study suggest that cerebral hypoperfusion may precede and contribute to the onset of 77 cognitive dysfunction and dementia. It is possible that deficits in certain cognitive domains, such as verbal learning, may be explained by their association with areas of the brain that are more vulnerable to the effects of cerebral hypoxemia resulting from vascular disease. Areas of the brain irrigated by long penetrating arteries such as the hippocampus are less able to tolerate the effects of hypoperfusion [137]. An fMRI study of CVLT performance (the test with high factor loadings on the verbal learning factor) showed that an area of the brain most active during tasks associated with this test was the hippocampus [65]. Thus, in populations that are cognitively normal, small decreases in oxygen supply to the brain associated with subclinical atherosclerosis may explain some of the differences in cognitive performance in middle age and the cumulative effect of small decrements over time may contribute to observable declines in cognitive function. A third possibility is that CIMT may be a surrogate measure for intracranial atherosclerosis and its resultant localized pathology of infarction or ischemia due to blockage or deterioration of cerebral arterioles [30]. This is consistent with the fact that CIMT predicts stroke risk [18], and would suggest that individuals with greater carotid atherosclerosis may also have more cerebral atherosclerosis, which individually leads to declines in cognition. Other population-based studies that used CIMT as a measure of atherosclerosis reported associations between atherosclerosis and cognitive function. In the ARIC study (adults aged 45-64 years), CIMT was cross-sectionally correlated with psychomotor performance in men and women, and with verbal learning in women 78 after adjustment for basic demographic factors [29], but not longitudinally after 6 years [75]. In contrast, we found evidence of an association between CIMT and decreased verbal learning in men but not in women. In a study of 400 Dutch middle-aged and elderly men aged 40-80 years [104], increased CIMT was significantly associated with lower memory (combined scores for verbal and visual). Investigators in that study used the Rey verbal learning test to assess verbal memory, and the Doors test to assess visual memory. Our observed association between lower verbal learning performance and increasing CIMT among men is partly in line with these findings, although visual memory was not associated with CIMT in our sample. A limitation of this cross-sectional study is the inability to address directionality of associations. We controlled for a number of factors including education and income that could be associated with both cognitive function and subclinical atherosclerosis in order to minimize their contribution to the observed associations. Strengths of this study include the battery of neuropsychological tests that allowed an examination of a broad range of cognitive abilities, as well as highly reliable and well-validated measures of subclinical atherosclerosis. This study is limited by small numbers of elderly adults >80 years old and women aged 40-49 years given the selection criteria for postmenopausal women. Therefore, the findings of the study may not be generalizable to elderly adults and to premenopausal women. In addition, given that BVAIT selected for subjects with Hcy >8.5 µmol/L, results from this study may not be generalizable to populations with lower Hcy. Previous research has shown that 79 elevated Hcy is associated with atherosclerosis [19,131] as well as with deficits in cognitive performance, cognitive decline and dementia [133]. It is therefore possible that given study selection criteria, we may have been more likely to detect an association, if one existed, between subclinical atherosclerosis and lower cognitive function. It is important to note, however, that mean Hcy levels in the study population (10.3 µmol/L) are still within what is considered the normal range [96]. In summary, our findings suggest that in a population of otherwise healthy middle- to-older aged hyperhomocysteinemic adults without clinically evident CVD, 1) subclinical atherosclerosis, measured by CIMT but not CAC or AAC, contributes to lower verbal learning abilities but not other areas of cognitive function through mechanisms independent of standard cardiovascular disease risk factors assessed in this study; 2) this association may be limited to men perhaps because they met a threshold level of subclinical atherosclerosis necessary for an association with cognition to be observed. Additional studies are needed to further elucidate the relationship between subclinical atherosclerosis and cognitive function among healthy adults without clinically evident CVD. 80 Chapter 3 - Metabolic Syndrome and Cognitive Function in Healthy Middle-Aged and Older Adults without Diabetes Abstract Few studies have addressed whether the metabolic syndrome (MetS) and its individual components are associated with cognitive function in middle-aged and older populations, as well as whether specific areas of cognition are more affected than others. We examined the cross-sectional association between MetS and six areas of cognitive function in healthy cognitively intact adults without diabetes (n=853, mean age 61 years) participating in two intervention trials. The National Cholesterol Education Program (NCEP) criteria were used to identify subjects with MetS. Cognitive function was assessed prior to randomization with a neuropsychological battery. A principal components analysis was used to extract five uncorrelated factors interpreted to represent five areas of cognition, and a measure of global cognition was calculated. MetS was weakly associated with lower verbal learning ( β = -0.16 [SE( β) = 0.09], p = 0.08). As the number of MetS criteria increased, global cognition and verbal learning decreased (p-trend both = 0.01). Hypertension was the only MetS risk factor that was independently correlated with lower verbal learning ( β = -0.26 [SE( β) = 0.07], p = 0.0004), semantic memory ( β = - 0.15 [SE( β) = 0.07], p = 0.03) and global cognition ( β = -0.12 [SE( β) = 0.06], p = 0.06). This study adds to the evidence of an association between MetS and lower cognitive function among healthy middle aged and older adults without CVD and 81 diabetes, as well as confirms the correlation between hypertension and lower cognition. Introduction Risk factors for cardiovascular disease (CVD) including hypertension and metabolic disturbances, such as diabetes, have been linked to the development of Alzheimer’s disease (AD) and vascular dementia (VaD) [13,81,99]. Several population- and community-based studies of non-demented individuals have reported associations between these and other CVD risk factors and cognitive dysfunction or decline [39,68,134], while others have not [4,122,133]. The metabolic syndrome (MetS), a clustering of risk factors that includes 1) abdominal obesity, 2) hypertriglyceridemia, 3) low levels of high density lipoprotein cholesterol (HDL-C), 4) hypertension and 5) hyperglycemia, was defined to summarize a constellation of CVD risk factors [108]. MetS is associated with increased risk of CVD [85] and diabetes [129], and has also been linked to the incidence [67] and prevalence of dementia [58,67] and prevalence of AD [83,166]. Whether MetS is correlated with cognitive dysfunction apart from each risk factor individually has been addressed by a limited number of studies [36,52,78,182184]. These studies found some evidence that MetS was associated with cognitive decline, but tended to focus on elderly populations [36,52,78,182184], and used measures that assessed limited areas of cognition [52,182184]. Questions therefore remain as to whether MetS is associated with cognitive function in middle-aged and older 82 populations, and with dysfunction in specific areas. Thus, we examined whether MetS was cross-sectionally associated with lower cognitive function in six areas among healthy cognitively intact postmenopausal women enrolled in the Women's Isoflavone Soy Health (WISH) Trial, and otherwise healthy cognitively intact men and women with elevated plasma homocysteine (Hcy) enrolled in the B-Vitamin Atherosclerosis Intervention Trial (BVAIT) (http://www.clinicaltrial.gov/). We also investigated which MetS component factors are independently correlated with cognitive function. Study protocols for both trials excluded diabetic persons from participation; thus this report is unique in that it focuses on MetS in a non-diabetic population. Methods Study participants Healthy postmenopausal women who were enrolled in WISH and otherwise healthy hyperhomocysteinemic adults who were enrolled in BVAIT were the focus of the present study. Data obtained for participants at their baseline visit prior to randomization were used in the current analysis. Briefly, postmenopausal women > 30 years old were eligible for WISH; men and postmenopausal women > 40 years old were eligible for BVAIT if they had Hcy >8.5 µmol/L. Of 6,372 individuals who were prescreened by telephone, 5,516 were ineligible or refused to be enrolled. Exclusions (n=1,100) were made for any clinical 83 signs or symptoms of CVD (n=165), diabetes mellitus or fasting serum glucose >126 mg/dL (n=137), triglyceride (TG) levels >500 mg/dL (n=3), hypertension [systolic blood pressure (SBP) >160 mmHg and/or diastolic blood pressure (DBP) >100 mmHg)] (n=13), untreated thyroid disease (n=4), creatinine clearance <70 ml/min (BVAIT) or serum creatinine >2.0 mg/dL (WISH) (n=6), a life threatening disease with prognosis <5 years (n=124), alcohol intake of >5 drinks per day/substance abuse (n=4), unwillingness to stop taking food or vitamin supplements (n=628), current use of hormone therapy (WISH) (n=15) or food allergy (WISH) (n=3). A total of 856 subjects were randomized (350 in WISH and 506 in BVAIT); all signed a written informed consent approved by the Institutional Review Board at the University of Southern California. Measurements All subjects were administered a battery of cognitive tests [90] in a standardized order by one trained psychometrist. The battery was designed to assess a broad array of cognitive abilities, with an emphasis on specific tasks used to detect age- associated change in middle-aged and elderly populations, particularly episodic memory and executive function, and included the following tests: • Symbol Digit Modalities Test (SDMT) • Trail Making Test Part B (Trails-B) • Judgment of Line Orientation, Form H (JLO) • Block Design [Wechsler Adult Intelligence Scale, 3 rd Edition (WAIS-III)] 84 • Letter-Number Sequencing [Wechsler Memory Scale, 3 rd Edition (WMS-III)] (LNS) • Category fluency (animal naming, 60 seconds) (Animals) • Boston Naming Test, 30-item version (BNT) • Shipley Institute of Living Scale Abstraction Subtest (Shipley) • California Verbal Learning Test, 2 nd edition (CVLT-II), immediate recall (IR) and delayed recall (DR) • Logical Memory I and II (paragraph recall, IR and DR) (WMS-III) • Faces I (IR) and II (DR) (WMS-III) The Center for Epidemiologic Studies Depression Scale (CES-D) scale [127] was used to assess mood. Of the randomized subjects, three did not have cognitive testing; 853 (99.6%) subjects were included in the present study. Fasting total cholesterol (TC), HDL-C and TG were measured by enzymatic assay methodology, and LDL cholesterol (LDL-C) was computed with the Friedewald equation [9,47]. Lipid measurements were standardized to the CDC Lipid Standardization Program [92]. Fasting serum glucose levels were measured with spectrophotometry on an Olympus 5400 (Quest Diagnostics, West Hills, California). Blood pressure, body height, weight and waist circumference (WC) were measured, and body mass index (BMI) was calculated (kg/m 2 ). A smoking questionnaire was used to determine smoking status. 85 Study participants were identified as having MetS using a modified NCEP definition [55,108] if ≥ 3 of the following risk determinants were present: 1) abdominal obesity: WC >102 cm (>40 in) for men or >88 cm (>35 in) for women; 2) TG ≥ 150 mg/dL; 3) HDL-C < 40 mg/dL for men or < 50 mg/dL for women; 4) blood pressure ≥130/ ≥85 mmHg or current use of anti-hypertensive medications; 5) fasting glucose level ≥ 110 mg/dL. Participants who did not meet ≥ 3 of the NCEP criteria were classified as not having MetS. Since both trials excluded diabetic persons, individuals identified with MetS in the present analysis represent MetS among non- diabetic populations. Statistical Analysis We conducted χ 2 or t-tests to assess whether baseline demographic and CVD risk factors significantly differed between study subjects with and without MetS. For subjects (n=44) who were unable to complete one or more tests in the neuropsychological battery, age-gender- and education-specific mean values from the study population were imputed. Small reductions (averaging 3.7%) in variances of the tests resulted from imputations, which were made for < 0.7% of the total number of tests administered. For data reduction purposes, a principal components analysis with an orthogonal varimax rotation was performed on the 14 cognitive tests in the neuropsychological battery and consecutive uncorrelated factors were extracted. Following methods of Cattell [25], a scree plot of successive eigenvalues was used to identify the number 86 of principal components at which the plot leveled off; this led to a decision to retain five factors. The five factors accounted for 72.4% of the total variance, and were interpreted by assigning a name to each that reflected high factor loadings of individual cognitive tests (i.e., loadings with an absolute value > 45). The resulting factors generally reflected cognitive abilities in areas of 1) executive function (high factor loadings on SDMT, Trails-B, LNS, JLO, Block Design and Shipley), 2) verbal learning (high factor loadings on CVLT- IR and DR), 3) logical memory (high factor loadings on paragraph recall - IR and DR), 4) visual episodic memory (high factor loadings on Faces I and II), and 5) semantic memory (high factor loadings on Animals and BNT). Incidentally, factor loadings were consistent with those derived for the BVAIT population, thus leading to comparable factors with those in previous analyses. For each subject, a factor score for each of the five component factors was calculated. In addition, we created a measure of global cognition, which was calculated as the weighted sum of scores on each of the individual tests in the neuropsychological battery; weighting of each test was based on the sum of the inverse covariances of scores with other tests. The resulting measure was then divided by its standard deviation (SD) so as to interpret results per SD of global cognition and ensure a consistent interpretation of results with those of the five factors. General linear models were used to examine the association between presence of MetS (independent variable) and the six measures of cognition (dependent variables) adjusting for age, gender, race/ethnicity, highest educational level achieved, 87 household income, study and CES-D score. Beta coefficients ( β) and their standard errors were estimated from regression models to assess the association between MetS and cognition. Separate models were run for each of the six measures of cognition. For the variable expressing the presence/absence of MetS, βs represented average differences, in SD units, of the cognitive measure comparing subjects with MetS to those without MetS. To assess whether there was a trend in cognitive functioning with increasing number of MetS criteria, a term for the number of MetS criteria present was tested in the regression models. The five MetS criteria were modeled individually to determine whether each criterion was a significant individual correlate of cognitive function, and together to determine whether each criterion was a significant independent correlate of cognitive function. Criteria were modeled both as continuous (divided by their SD) and categorical variables, using NCEP cut points to define categories (i.e., hypertensive versus non-hypertensive). βs were interpreted either as 1) the mean difference in the cognitive measure relative to subjects with a one SD lower value of the particular MetS criterion (criteria modeled continuously), or 2) the mean difference in the cognitive measure relative to subjects not meeting the particular MetS criterion (criteria modeled categorically). All analyses used SAS version 9.1 (SAS Institute Inc., Cary, NC, USA.). Results The mean (SD) age of the study population was 60.9 (8.8) years. The majority were women (63.9%), Caucasian (64.5%) and highly educated (59.6% with a Bachelor’s 88 or graduate degree) (Table 8). Nearly one-third (31.9%) was currently using medications for hypertension. 112 (13.1%) of the study population fit the expanded NCEP definition for MetS. Of those, 75 (67%) met three criteria, 30 (26.8%) met four, and 7 (6.3%) met all five criteria. Among those with MetS, the most common criterion met was hypertriglyceridemia (n=90, 80.4%), followed by hypertension (n=86, 76.8%), abdominal obesity (n=82, 73.2%), low HDL-C (n=66, 58.9%) and hyperglycemia (n=56, 50%). Subjects with MetS were more likely to be Hispanic and, as expected, to have higher blood pressure, fasting glucose, TG and BMI and lower HDL-C compared to those without MetS (Table 8). Table 8. Baseline Characteristics for Study Subjects with Cognitive Testing (n=853) by the Presence of the Metabolic Syndrome Variable With the Metabolic Syndrome* (n=112, 13.1%) Without the Metabolic Syndrome* (n=741, 86.9%) p-value† Age (years) 61.8 ± 8.2 60.8 ± 8.9 0.28 Gender 0.17 Male 47 (42.0%) 261 (35.2%) Female 65 (58.0%) 480(64.8%) Race/Ethnicity 0.009 Non-Hispanic White 64 (57.1%) 486 (65.6%) Non-Hispanic Black 15 (13.4%) 83 (11.2%) Hispanic 25 (22.3%) 87 (11.7%) Asian/Pacific Island/Native American 8 (7.1%) 85 (11.5%) Educational Level 0.25 High school or less 16 (14.3%) 80 (10.8%) Some college 39 (34.8%) 210 (28.3%) Bachelor’s degree 25 (22.3%) 192 (25.9%) Graduate/professional degree 32 (28.6%) 259 (34.9%) 89 Table 8, continued Household Income 0.23 Low 32 (28.6%) 165 (22.4%) Medium 15 (13.4%) 115 (15.6%) Moderate 24 (21.4%) 140 (19.0%) High 41 (36.6%) 317 (43.0% Current/Former Smoker 44 (39.3%) 290 (39.2%) 0.98 Body-Mass Index (kg/m 2 ) 32.6 ± 5.1 26.7 ± 4.7 <0.0001 Blood Pressure (mmHg) Systolic 132.4 ± 15.5 123.7 ± 16.6 <0.0001 Diastolic 81.9 ± 10.2 77.9 ± 10.0 <0.0001 Total Cholesterol (mg/dL)‡ 221.0 ± 32.9 220.7 ± 37.1 0.92 LDL Cholesterol (mg/dL)‡ 136.1 ± 29.4 137.0 ± 34.5 0.77 HDL Cholesterol (mg/dL)§ 46.3 ± 10.6 61.5 ± 16.0 <0.0001 Triglycerides (mg/dL)§ 194.5 ± 72.1 110.8 ± 57.9 <0.0001 Glucose (mg/dL) ║ 109.3 ± 12.8 96.8 ± 9.2 <0.0001 CES-D score‡ 11.3 ± 9.2 11.8 ± 8.6 0.51 *Mean ± SD or Number (%) †p-value for comparison of subjects with the metabolic syndrome to subjects without the metabolic syndrome ‡ n=851 § n=852 ║ n=848 Participants with MetS had lower scores on the verbal learning factor ( β = -0.16 [SE( β) = 0.09], p = 0.08) compared to participants without MetS (Table 9). For each additional MetS criterion present, scores decreased by 6% of a SD on the measure of global cognition ( β = -0.06 [SE( β) = 0.03], p = 0.01), and by 7% of a SD on the verbal learning factor ( β = -0.07 [SE( β) = 0.03], p = 0.01). The MetS was not associated with lower executive function, logical memory or visual memory (Table 9). 90 Table 9. Linear Regression Models* Associating the Metabolic Syndrome with Six Measures of Cognition (n=853) Cognitive Measure MetS Present β coefficient [SE( β)] p-value Number of MetS Criteria Present† β coefficient [SE( β)] p-value Global Cognition -0.08 (0.08) 0.32 -0.06 (0.03) 0.01 Executive Function 0.02 (0.09) 0.87 -0.02 (0.03) 0.39 Verbal Learning -0.16 (0.09) 0.08 -0.07 (0.03) 0.01 Logical Memory 0.10 (0.10) 0.34 -0.01 (0.03) 0.79 Visual Memory -0.03 (0.10) 0.79 0.01 (0.03) 0.74 Semantic Memory - 0.10 (0.09) 0.29 -0.05 (0.03) 0.07 *Adjusted for age, gender, race/ethnicity, education, income, study, CES-D score † Continuous variable ranging from 0-5 ‡ Subjects with MetS: n=112,13.1%; subjects without MetS: n=741, 86.9 HDL-C levels were positively associated with scores on verbal learning ( β = 0.16 [SE( β) = 0.03] per 16mg/dL of HDL, p <0.0001). Increases in SBP and TG were individually associated with lower semantic memory (Table 10). Lower global cognitive abilities were correlated with individual decreases in HDL-C and increases in TG (Table 10). When mutually adjusted as continuous variables, no MetS component was a statistically significant independent correlate of lower cognitive function in any area (data not shown). Table 10. Linear Regression Models* Associating Components of the Metabolic Syndrome Individually Modeled as Continuous Variables with Cognitive Function (n = 853) Change in Cognitive Factor per SD of Component Metabolic Syndrome Component Verbal Learning β(SE) p-value† Semantic Memory β(SE) p-value† Global Cognition β(SE) p-value† SBP (per 17 mmHg) 0.003 (0.04) 0.94 -0.07 (0.04) 0.06 -0.02 (0.03) 0.61 Glucose (per 11 mg/dL) -0.01 (0.03) 0.97 -0.05 (0.03) 0.14 0.02 (0.03) 0.51 HDL (per 16 mg/dL) 0.16 (0.03) <0.0001 -0.04 (0.03) 0.26 0.06 (0.02) 0.03 Waist Circumference (per 5 in) -0.04 (0.04) 0.28 0.06 (0.04) 0.10 -0.03 (0.03) 0.37 Triglycerides (per 66 mg/dL) -0.03 (0.03) 0.29 -0.07 (0.03) 0.04 -0.06 (0.03) 0.03 *Adjusted for age, gender, race/ethnicity, education, income, study, CES-D score † p-value for comparison per standard deviation 91 92 Using the NCEP-defined MetS criteria, hypertension was individually associated with lower verbal learning ( β = -0.22 [SE( β) = 0.07], p = 0.001), semantic memory ( β = -0.17 [SE( β) = 0.07], p = 0.01) and global cognition ( β = -0.11 [SE( β) = 0.06], p = 0.06). Hypertriglyceridemia and hyperglycemia were individually associated with lower semantic memory, and low HDL-C and hypertriglyceridemia were significantly individually associated with lower global cognition (Table 11). Abdominal obesity was significantly positively associated with semantic memory scores. When all MetS criteria were mutually adjusted in regression models, hypertension was the only MetS risk factor that was consistently independently inversely associated with verbal learning, semantic memory and global cognition (Table 11). MetS criteria (modeled either as continuous or categorical variables) were not associated (either individually or independently) with executive function, logical memory or visual memory (results not shown). Table 11. Linear Regression Models* Associating Individual NCEP Metabolic Syndrome Characteristics (Categorical Variables) with Cognitive Function (n=853) Metabolic Syndrome Criteria Verbal Learning Semantic Memory Global Cognition Individually modeled β(SE) p-value† Mutually adjusted β(SE) p-value† Individually modeled β(SE) p-value† Mutually adjusted β(SE) p-value† Individually modeled β(SE) p-value† Mutually adjusted β(SE) p-value† Hypertension -0.22 (0.07) 0.001 -0.22 (0.07) 0.001 -0.17 (0.07) 0.01 -0.16 (0.07) 0.02 -0.11 (0.06) 0.06 -0.11 (0.06) 0.07 Hyperglycemia -0.05 (0.10) 0.61 0.004 (0.10) 0.97 -0.18 (0.09) 0.06 -0.18 (0.10) 0.07 0.06 (0.09) 0.48 0.12 (0.09) 0.19 Low HDL-C Level -0.11 (0.08) 0.18 -0.12 (0.09) 0.18 -0.07 (0.08) 0.39 -0.07 (0.09) 0.44 -0.17 (0.08) 0.03 -0.14 (0.08) 0.08 Abdominal Obesity 0.06 (0.08) 0.41 - 0.03 (0.08) 0.74 0.16 (0.08) 0.04 0.23 (0.08) 0.003 0.05 (0.07) 0.47 -0.005 (0.07) 0.95 Hypertriglyceridemia -0.04 (0.07) 0.56 0.02 (0.08) 0.77 -0.14 (0.07) 0.06 -0.13 (0.08) 0.11 -0.16 (0.07) 0.015 -0.12 (0.07) 0.09 *Adjusted for age, gender, race/ethnicity, education, income, study, CES-D score † p-value for comparison of one category to reference category 93 94 Because the NCEP specifies gender-specific cutpoints for two of the MetS criteria (waist circumference and HDL-C levels), and because abdominal obesity was positively associated with semantic memory in the study population, we explored whether there were gender differences in the semantic memory associations. We found that mutually adjusted associations were several magnitudes higher in women for abdominal obesity ( β = 0.34 [SE( β) = 0.09], p = 0.0004) and hyperglycemia ( β = -0.32 [SE( β) = 0.14], p = 0.03) than men ( β = 0.08 [SE( β) = 0.14], p = 0.56, and ( β = -0.09 [SE( β) = 0.14], p = 0.52, respectively), but were not different for hypertension. Discussion This study provides evidence of an association between MetS and lower cognitive function among healthy, cognitively intact middle aged and older men and women without CVD or diabetes. We found that non-diabetic adults with MetS tended to perform lower on verbal learning tasks but not in other areas of cognition compared to those without MetS. A greater number of MetS criteria present were also associated with lower verbal learning, supporting the validity of the association between MetS and this factor. However, the magnitude of the mean difference, about a seventh of an SD, is modest when applied to individuals. Similar trends with increasing number of MetS criteria present and lower semantic memory and global cognition were also observed. Hypertension was the only MetS risk factor that was an independent correlate of lower function in these three areas of cognition. The 95 average age of our study population was at least 10 years less than populations evaluated in previous studies of MetS and cognition [36,52,78,182184]. In contrast to a study reporting that low HDL-C was the single MetS criteria independently associated with decreased memory function in elderly women [78], we found that hyperglycemia was an independent correlate of semantic memory among women in our study population. That study, however, examined baseline levels of HDL-C and memory assessed at a 12-year follow-up. Our study used baseline measures of both MetS criteria and cognition. While higher blood pressure on a continuous scale was not independently associated with lower verbal learning, hypertension was. One possible explanation of this finding is that the NCEP hypertension definition specifies a relevant threshold at which associations between elevations in blood pressure and cognition may be observed. To further investigate the role of hypertension in cognition, we assessed the contribution of hypertension to the effect of MetS on verbal learning abilities. We found that hypertension attenuated the MetS- verbal learning association by approximately 50%, suggesting that a large proportion of this association is due to the effect of hypertension. Results of this study suggest that the association between MetS and its component factors and cognition may be specific to certain areas of cognition, in addition to having an overall effect on global cognition. Specifically, verbal learning and semantic memory may be more vulnerable to the metabolic disturbances clustered in 96 MetS. Our finding that MetS is associated with verbal learning is consistent with the literature, in that MetS is a risk factor for diabetes [129] and previous studies have reported associations between diabetes and decreased verbal memory [151]. The extension of these findings to a study population of non-diabetic persons makes our findings unique. Mechanisms by which hyperglycemia may lead to decreased cognitive function have been proposed. In animal studies, chronic hyperglycemia results in decreased acetylcholine synthesis and loss of cortical neurons [81]. Cholinergic transmission is known to be important for learning and memory, and the cerebral cortex is thought to be where memory is “stored” in the brain [69]. Hypertension, which has been consistently linked [40,41] to decline in cognitive functioning, may act by accelerating arteriosclerosis in the cerebral microvasculature, leading to brain lacunar infarctions and or white matter lesions [81]. Previous studies examining the association between MetS or its components and cognition have used one [52,182,184] or at most a few [78,183] neuropsychological instruments to assess cognitive function, usually the Mini-Mental State Examination (MMSE) or a derivative. The MMSE was developed to evaluate mental status in elderly individuals and to screen for cognitive impairment or dementia [46,160]. While a popular measure in epidemiologic studies for its brevity and ease of administration, the MMSE does not provide a wide range of scores with which to assess variability in cognitively intact populations, nor does it allow for a comprehensive examination of different areas of cognition. A strength of the current 97 study is the neuropsychological battery that allowed for an assessment of a broad range of cognitive abilities. Furthermore, our approach to calculating cognitive factor scores using principal components ensures that the cognitive outcomes (except for the global cognition score) were uncorrelated. In addition, we controlled for a number of factors known to be associated with cognitive function. Nevertheless, the effects observed in this population were small. For verbal learning, the decrease in cognitive performance associated with the presence of MetS corresponds to approximately 16% of SD in the factor score, which is still within the range of what is considered a clinically “normal” level of cognitive functioning [43,113,118]. This study is limited by small numbers of adults >80 years old. Thus we may have been underpowered to assess cognition at lower bounds of normal given the inverse association between age and cognition. In addition, the study’s findings may not generalize to elderly adults. Furthermore, given the BVAIT selection criteria for subjects with Hcy >8.5 µmol/L, it is possible that results from this study may not be generalizable to populations with lower Hcy. Given the cross-sectional design, we did not have information on duration of specific conditions, and moreover, were unable to address directionality of associations. Finally, the study population was comprised of healthy, well-educated volunteers for two clinical trials, and probably does not represent the entire range of cognitive abilities prevalent in the general population. Therefore, it is possible that we may have underestimated the true association between MetS and cognition. 98 In conclusion, this study adds to the evidence that MetS and its component factors are associated with lower cognitive functioning, specifically with lower verbal learning, semantic memory and global cognition in healthy, cognitively intact middle-aged and older adults without CVD or diabetes. Results from this study also suggest that individual components of MetS, particularly hypertension, may be at least as important a contributor to cognition as the MetS symptom cluster. 99 Chapter 4 – Grant Proposal: Abnormalities in the Retinal Microvasculature and Cognitive Function Research Plan Study Objectives In 2005, older U.S. adults aged 65 and older numbered 36.8 million and constituted 12.4% of the US population. This number is expected to double by the year 2030 when older adults are anticipated to reach 20% of the population. A substantial proportion of the population does not experience normal cognitive aging and suffer from adverse cognitive outcomes such as cognitive impairment (estimated population prevalence 17%) or dementia (estimated population prevalence 10%) [86]. It is critical to understand what factors contribute to the development of adverse cognitive outcomes early in life in order to prevent or delay their onset [128]. Cardiovascular disease (CVD) is pervasive in the Western world, and given that the human brain is a highly vascularized organ, potential CVD effects on cognition are a concern. While atherosclerosis of the cerebral microvasculature is most relevant to the brain and cognitive function, the microvasculature of the brain is difficult to assess in vivo, and thus other approaches are required to address this problem. Given similarities in anatomy and physiology, the retinal vessels provide a surrogate to non-invasively study vessels in the brain. Observable abnormalities in the retinal vasculature may be markers of cardiovascular disease and indicators of its affects on the cerebral microvasculature. 100 The Latino population comprises 12.5% of the US population as of the 2000 census [164] and is the fastest growing segment of the population. It is estimated that by the year 2050, Latinos will make up almost one-quarter of the population. Data in Latinos are lacking on the prevalence of certain retinal microvascular abnormalities including general and focal arteriolar narrowing that are indicative of mild to moderate and severe arteriosclerosis. Furthermore, few population-based epidemiologic studies of cognitive function among Latinos exist. The Los Angeles Latino Eye Study (LALES) study provides an unparalleled opportunity to further elucidate the relationship between arteriosclerosis and cognitive function using the wealth of existing data collected in LALES, a five-year population-based cross-sectional study of over 6,300 Latinos aged 40 years and older in Los Angeles County (NEI: EY11753 and EY03040). Extensive data on demographic characteristics, ocular and medical histories were collected on all LALES participants, and detailed medical and ocular examinations were performed. LALES included cognitive screening with the Cognitive Abilities Screening Instrument-Short Form (CASI-S), and a supplement to LALES is currently validating this instrument against the Spanish English Neuropsychological Assessment Scales (SENAS) in a subset of study participants. We propose utilizing this data to address hypotheses that disease of the cerebral microvasculature is associated with 1) decreased cognitive function, specifically in areas of verbal memory and conceptual abilities, and 2) brain abnormalities visualized by magnetic resonance imaging (MRI). To test our hypotheses, we will utilize the retinal vessels 101 as anatomic/physiologic surrogates of microvasculature in the brain, and will use the measurements of narrowed diameters of retinal vessels to reflect cerebral arteriosclerosis. We hypothesize that focal and generalized retinal arteriolar narrowing will be associated with decreased cognitive function. Our underlying assumption is that changes in the retinal vasculature that are signs of underlying pathology will be reflective of similar changes in the cerebral vasculature. Specific Aims 1. To describe the distribution of retinal vessel diameters and the prevalence of generalized arteriolar narrowing and retinal microvascular abnormalities in a Latino population. 2. To examine whether retinal microvascular abnormalities are associated with decreased cognitive function and cognitive function in specific domains of cognition. 3. To investigate the association between retinal microvascular abnormalities and cognitive impairment and dementia. 4. To assess whether retinal microvascular abnormalities are associated with cerebral atrophy, total lacune volume, white matter hyperintensity volume and hippocampal volume from magnetic resonance imaging (MRI) data. 102 Background and Significance The US population is aging and age-related health problems including cognitive dysfunction are increasing The US population is aging, due in part to advances in public health and medicine, which have reduced death rates for children, young and older adults. In 2005, older adults aged 65 and older numbered 36.8 million and constituted 12.4% of the US population. This number is expected to double by the year 2030 when older adults are anticipated to number 71.5 million or 20% of the population. The “baby boom” generation will reach retirement age during the years 2010 - 2030 [163], which will further contribute to the burgeoning of the older population. It can thus be expected that age-related health problems will become an increasing concern. Cognitive function represents the complex repertoire of abilities reflecting a dynamic interaction between the individual and the social environment throughout life [22,98]. Determinants of cognition are multifactorial and include a number of demographic and social factors such as socioeconomic status, race and ethnicity, social connectedness, occupational environment, and religion/spirituality. Factors related to emotional health including depression, psychological stress, anxiety and quality of life also contribute to cognitive function. Cardiovascular and metabolic risk factors including high blood pressure, overweight and obesity, diabetes and stroke; inflammatory factors, diet, alcohol intake, physical activity and genetics [138] have all been shown to be associated with cognitive function [112]. 103 Cognitive function follows an approximate trajectory over the human lifespan with cognitive growth apparent during childhood, a plateau in cognition with relative stability and little variation during middle and older ages, and cognitive decline in the latter years of life [22]. Thus with normal aging, some decline in cognition is expected, however, cognitive impairment and dementia are not necessary consequences of aging. Nevertheless, a substantial proportion of the population does not experience normal cognitive aging and suffer from cognitive impairment (estimated population prevalence 17%) or dementia (estimated population prevalence 10%) [86]. Cognitive dysfunction is a large personal and societal burden [70]. It is critical to understand what early life factors contribute to the development of such adverse cognitive outcomes in order to prevent or delay their onset. CVD is pervasive in the Western world Cardiovascular disease (CVD) is the leading cause of death and a major cause of disability in both men and women in the United States (US) [2,26,138]. CVD, including coronary heart disease, coronary artery disease, hypertension and stroke, is highly prevalent in the US population. In the year 2004, an estimated 1 in 3, or over 79 million Americans had cardiovascular disease (CVD) [95,138]. Each year, over one-third of all deaths are attributed to CVD [27,138]. As the US population ages, the incidence of CVD is likely to increase and the burdens associated with it will continue to grow [16]. 104 The development of CVD is preceded by atherosclerosis, the pathological process most commonly underlying CVD [8,15,23,167,170]. The atherosclerotic process is progressive, beginning with damage to arterial endothelium and progressing to vascular remodeling, plaque formation and stenosis [15]. Human autopsy studies suggest that atherosclerosis may appear in the first and second decades of life, is prevalent to some degree across all age groups thereafter, with more severe forms becoming apparent in middle-aged and older individuals [12,123,135,152,156,190]. Atherosclerosis may exist in subclinical stages for decades prior to becoming symptomatic. The prevalence of atherosclerosis varies in extent and severity in the population. Major risk factors for atherosclerosis include age, sex, postmenopausal status, elevated LDL cholesterol, elevated blood pressure, obesity, smoking, physical inactivity, family history of CVD, non-Caucasian race/ethnicity and diabetes [11,32,80,94,117,153,168,169]. CVD effects on cognitive function Vascular disease is known to contribute to cognitive dysfunction in the latter years of life [114]. Given the pervasiveness of CVD in the Western world, there is a growing concern about possible effects of CVD and atherosclerosis on cognitive function particularly among middle-aged and older adults. A number of studies have examined the association between atherosclerosis and cognition in adults. Some have found evidence supporting an association between atherosclerosis and reduced cognitive function, while other studies have reported weak or null associations [21,29,66,103]. These studies assessed medium and large vessels using various 105 measures of atherosclerosis including coronary artery plaques and stenosis and carotid artery intima-media thickness (IMT). The inconsistent findings from these studies could be explained by the fact that atherosclerosis in large vessels may not be reflective of disease in small vessels. Pruissen et al. showed that carotid artery IMT was significantly thicker in patients with large vessel disease than in patients with small vessel disease [126]. Therefore, using the assessment of disease in large vessels to estimate the effect of small vessel disease on cognition may lead to erroneous conclusions. The human brain is a highly vascularized organ, and the vast majority of brain blood vessels are in the size range of “microvasculature”, i.e., arterioles less than 200µm in diameter. It is known that the state of the cerebral vasculature relates to cognitive disease [64,115]. Atherosclerotic changes to the vasculature that are most relevant to the brain and thus to cognitive function are those affecting the microvasculature. Thus, it is more important to assess the cerebral microvasculature in order to examine the influence of vascular disease on cognition. Since the microvasculature of the brain is difficult to assess in vivo, other approaches are required to address this problem. 106 Rationale for using the retina as a marker for analogous structures in the brain The eye: a portal into the brain While imaging methods such as CAT, PET, MRI and fMRI are excellent approaches for visualizing the structure and function of the brain, they are costly, non-portable, time-consuming and not practical for large population-based studies. In vivo assessments of the microvasculature of the brain are technically difficult, and thus are usually restricted to post-mortem analyses. Another approach to this problem is to study simpler parts of the human brain, namely the retina. As such, the human visual system is regarded as the “portal into the brain” [37]. The retinal vessels share commonalities in anatomy, physiology and embryology with the cerebral small vessels. Embryologically, the retina is an extension of the diencephalon; both the retina and the brain share a similar pattern of vascularization during development [119,176]. Branches of blood vessels both of the eye and the brain originate from the carotid artery. Small arteries in both the brain and retina (approximately 100-400 µm) maintain the structural organization characteristic of larger arteries with three vacular layers (tunicas intima, media and adventitia) and an epithelial lining. Similar to cerebral small vessels, the retinal arteries are also functional end arteries that do not normally anastomose, although they have a dense capillary network [84]. The majority of visible retinal arteries are in the size range of 50 – 250 µm, which correspond to the size of cerebral small vessels (small arteries and arterioles) [84]. Both the cerebral and retinal arterioles (<100 µm) are composed 107 of 2 layers with a continuous elastic lamina and tunica media. Furthermore, as there exists a blood–brain barrier of the cerebral arteries, there is also a barrier between the blood and the retina [84]. Given the homology between retinal and cerebral microcirculations, the retinal vessels provide a unique opportunity to indirectly study vessels in the brain. Furthermore, the retina may be directly and non-invasively visualized in vivo [120,176]. In conclusion, with respect to the goal of evaluating associations between vascular disease and cognition, the anatomic and physiologic similarities between the retinal and cerebral vessels therefore make an assessment of disease in retinal vessels an alternative to assessing disease in large arteries. Retinal microvascular abnormalities are prevalent in the general population Several retinal microvascular characteristics have been described [119,161,176] and can be classified into categories: a) retinal arteriolar changes related to the retinal arterioles only, including focal arteriolar narrowing, generalized arteriolar narrowing and arteriovenous (AV) nicking; b) retinopathies (microvascular characteristics not explicitly arteriolar in nature), including microaneurysms, hemorrhages (blot, flame- shaped and vitreous), soft and hard exudates, cotton wool spots, macular edema, intraretinal microvascular abnormalities, venous beading, new vessels at the disc or elsewhere; c) deviations from optimal retinal vascular network geometry indicated by abnormal bifurcation angles and arteriolar branching coefficients, as well as tortuosity. Abnormal changes to the retinal microvasculature have been divided into 108 four overlapping phases (i.e., not necessarily sequential) with corresponding pathological changes and clinical signs: 1) a vasoconstrictive phase (elevated blood pressure leading to generalized and focal arteriolar narrowing); 2) a sclerotic phase (persistently elevated blood pressure leading also to generalized and focal arteriolar narrowing, tortuosity, increase in the angle of arteriolar branching); 3) an exudative phase (sustained hypertension, disruption of the blood-retinal barrier leading to narrowing of the arteriolar wall, impairment in blood flow and ischemic complications indicated by microaneurysms and retinal hemorrhages, soft and hard exudates); 4) a complications phase (longstanding hypertensive and atherosclerotic changes in vessels leading to retinal complications including central or branch retinal artery and vein occlusions, macroaneurysms, macular edema). It is important to note that most of what is known about stages of retinal abnormalities has been derived from animal data, and the exact pathophysiological bases of these clinical signs in humans remain to be determined. Several population-based studies have used standardized photographic grading methods to estimate the prevalence of retinal microvascular abnormalities among middle-aged, older and elderly black and white non-diabetic and mixed diabetic and non-diabetic populations [72,74,89,186]. Retinal microvascular abnormalities are known to be prevalent in non-diabetic populations, although more common among individuals with hypertension and diabetes [73]. In LALES, initial data suggest that the prevalence of any retinopathy was higher among diabetic than non-diabetic individuals, as well as among 109 participants with a history of hypertension than those without a history of hypertension. Prior to LALES, no study that described the population prevalence of retinal microvascular abnormalities included substantial numbers of Latino participants. Thus, published racial/ethnic contrasts are limited to whites and blacks, with blacks having a higher prevalence of some retinal microvascular abnormalities than whites [176]. Table 10 summarizes prevalences of retinal microvascular abnormalities among non-diabetic adults reported in population-based studies. Table 12. Prevalence of retinal microvascular abnormalities in non-diabetic adults from population-based studies Study Population Age (years) Prevalence ARIC [74] white & black adults 48-72 3-8% Beaver Dam [71] white adults 43-86 2-14% CHS [175,178] white & black adults 69-97 7-10% LALES* Latino adults 40-97 6% *preliminary data Data from previously published studies suggest that there may be an age-dependent variation in the prevalence of focal arteriolar narrowing, with prevalence increasing with increasing age [73,74], whereas the relationship between retinopathies and age is less clear [176]. In LALES, preliminary data suggest that the prevalence of retinopathies tends to increase with increasing age. The proposed project will generate data with which to investigate how generalized arteriolar narrowing varies with age among a Latino population. Findings from available studies are also inconsistent with respect to the relationship between retinal microvascular 110 abnormalities and gender. In LALES, initial data suggest that men tend to have a higher prevalence of retinopathies than women. Evidence suggests that retinal microvascular abnormalities reflect cerebral microvascular changes While the causes, systemic associations and clinical significance of retinal microvascular changes and abnormalities are not fully understood in humans, evidence indicates that they parallel changes to the cerebral vasculature [84], and may be reflective of cerebral microvascular disease [54]. Furthermore, retinal microvascular abnormalities are known to reflect arteriolar damage from hypertension and other processes [176]. As such, retinal microvascular abnormalities are hypothesized as markers of generalized vascular disease. Available evidence suggests that changes to the retinal microvasculature that occur with age and under disease conditions are reflective of changes occurring in the cerebral microvasculature. Both the retina and brain experience reductions in blood flow [119], decreased glucose and oxygen metabolism, and impairment of the structural integrity of the anatomy of the microvasculature with aging. Aging of the cerebral vasculature is associated with morphological changes including basement membrane thickening, a decrease in endothelial and pericyte cell populations and reductions in density [119]. While less data is available to evaluate age-related changes to the retina, age-related basement membrane thickening has also been observed. 111 Both the cerebral and retinal microvasculatures are known to undergo a continuum of morphological changes with increases in blood pressure [119]. Hypertension causes generalized retinal arteriolar narrowing in vessels without significant atherosclerosis [119,176]. In vessels with moderate atherosclerosis, focal arteriolar narrowing occurs in segments without sclerosis, and dilation in segments with sclerosis. Persistent elevations in blood pressure characterize a sclerotic phase of retinal microvascular changes, which is associated with hyperplasia of the tunica media and hyaline degeneration of the arteriolar wall. With sustained hypertension, breakdowns of the blood-retina barrier occur, with degeneration of vascular smooth muscle and endothelial cell necrosis leading to exudative changes. With longstanding hypertension and atherosclerosis in retinal vessels, complications develop, including arteriolar thrombosis, artery and vein occlusions and macular edema. Similar changes to the cerebral microvasculature are known to occur with hypertension, including hyaline atherosclerosis and luminal narrowing. Degeneration of the tunica media and internal elastic lamina, and replacement with fibrous tissue also lead to increased cerebral vessel tortuosity, increased vessel permeability, and breakdown of the blood-brain barrier. Fibrinoid necrosis is the pathological hallmark of acute hypertensive brain damage[119]. 112 Evidence of association between retinal microvascular abnormalities and cognition Three epidemiologic studies contribute to the evidence that certain retinal microvascular abnormalities are cross-sectionally associated with poorer cognitive function among middle-aged, older and elderly adults [10,121,179]. Wong et al. [179] reported that the presence of any retinopathy was associated with poorer cognitive function among middle-aged persons without stroke. Included in the study were middle-aged African-American and Caucasian participants in the population-based ARIC study who had negative histories of stroke and had not used central nervous system-relevant medications (n=7,526, mean age 53.8 years). Retinal photographs were evaluated for microvascular abnormalities, which included any retinopathy, AV nicking, focal arteriolar narrowing and general arteriolar narrowing (as indicated by the arteriole-to-venule ratio). Cognitive function was assessed with Delayed Word Recall, Digit Symbol and Word Fluency tests; participants scoring 2 SDs below the mean scores for the study population on each test were defined as cognitively impaired. Mean scores on all three tests were significantly lower among participants with any retinopathy, retinal hemorrhages, microaneurysms or soft exudates compared to participants without these lesions after adjusting for a number of demographic and CVD risk factors. In addition, scores in the Delayed Word Recall test were significantly lower in persons with focal arteriolar narrowing compared to those without. There were no differences in mean cognitive scores of participants with AV nicking or generalized arteriolar narrowing compared to those without the abnormalities. The authors also reported that the odds of cognitive 113 impairment were significantly higher among participants with any retinopathy compared to those without (ORs of approximately 2). Odds of cognitive impairment were also increased for participants with retinal hemorrhage, microaneurysm, soft exudates, arteriovenous nicking or generalized arteriolar narrowing compared to those without. There was no significant interaction by diabetes or hypertension and results were similar in analyses by subgroups of age, sex and race. Baker et al. [10] also reported that the presence of any retinopathy was associated with lower psychomotor performance among 2,211 CHS participants (mean age 78 ± 4 years). Subjects included in the study were African-American (15.5%) and Caucasian without a prior history of stroke. Retinal photographs were evaluated as described above for the ARIC study, and similar classes of retinal signs were included in analyses, as were retinal vessel calibers. Cognitive function was assessed with the Digit Symbol Test (psychomotor performance) and the Modified Mini- Mental State Examination (3MSE); dementia was also studied as an outcome. All analyses were adjusted for age, gender, race, education, IMT, BMI, hypertension, diabetes status and cigarette smoking. Adjusted mean scores on the Digit Symbol Test were significantly lower among participants with any retinopathy present compared to those without a retinopathy, but were not different by the presence/absence of AV nicking, focal arteriolar narrowing, or by arteriolar or venular caliber. 3MSE scores were borderline significantly lower among participants with focal arteriolar narrowing present compared to those without focal narrowing, but were not different for any other retinal vessel characteristic assessed. 114 The odds of dementia were significantly greater for persons with focal arteriolar narrowing (OR = 1.99, 95% CI = 1.11-3.56); among persons with hypertension, the presence of any retinopathy was also associated with dementia (OR=2.10, 95% CI=1.04-4.24). It is known that vascular topographical geometry is not randomly organized, but rather tends to conform to principals that optimize the organization of the network, such as minimizing physical properties that would lead to losses in volume across the vascular network [106,107,147,187189]. Murray’s law expresses properties of branching coefficients that result in the most efficient circulation across a vascular network [120]. Patton et al. reported that suboptimal retinal vascular network geometry was associated with decreased cognitive ability among elderly adults [121]. Included in the study were community-dwelling elderly Scottish adults (n=321) who were surviving members of the 1921 Lothian Birth cohort in Scotland (all aged approximately 83 years at follow-up). All participants had been surveyed in 1932 at age 11 for their mental ability with the Moral House Test, a measure of IQ. Retinal photographs were evaluated for measurements of retinal vessel caliber and optimality of arteriolar bifurication geometry (indicated by a branching coefficient [BC] and bifurication angle [BA]). Rather than assessing the presence or absence of retinopathies, the authors quantified deviations from optimal arteriolar bifurication geometry by calculating differences between expected values of BD and BA under optimal geometry conditions and observed values of the BC and BA. Cognitive function was assessed with the Logical Memory component of the WMS- 115 R (immediate and delayed recall), the Controlled Oral Word Association Test (verbal fluency and Raven’s Standard Progressive Matrices (nonverbal reasoning). A measure of general cognitive ability was generated by extracting a component that reflected shared variance between the three measures. The MMSE was also used to assess cognitive status. Analyses were adjusted for gender, APOE e4 status, presence of diabetes, smoking status, history of cerebrovascular disease, social class, education, blood pressure, alcohol consumption, visual acuity and premorbid IQ (at age 11). Results of general linear models showed no association between focal and general arteriolar narrowing and any of the cognitive tests or the measure of general cognition. However, deviation from optimality of the median BC was significantly inversely associated with verbal fluency and general cognitive ability, and of BA was significantly inversely associated with logical memory. These studies suggest that retinopathies and less than optimal arteriolar bifurication geometry, which may result from changes in blood flow from sustained hypertension are associated with reduced cognitive function. Other retinal microvascular abnormalities such as general arteriolar narrowing, which is associated with earlier stages of arteriosclerosis and more recent elevations in blood pressure are not associated with reduced cognitive function. However, as two of the three studies focused on elderly populations and none of the studies included Latino participants, there is a significant need for additional population-based studies to further elucidate the use of retinal microvascular abnormalities in examining the role of cerebral microvascular disease in cognitive dysfunction among middle-aged and older 116 populations other than Caucasians and African-Americans. The proposed research will extend our knowledge to the Latino population and includes young, middle-aged and elderly participants. In addition, subsets of the study population will be included in analyses that assess multiple areas of cognition, and that use brain imaging data to investigate biological correlates of cognitive function. Evidence of association between retinal microvascular disease and brain abnormalities identified with MRI MRI measures of white matter lesions (WML) and lacunar infarcts (LI) are brain abnormalities that are indicators of cerebral small vessel disease. While the exact pathophysiological mechanisms of small vessel disease are unknown [116], several studies suggest that CVD risk factors including age, atherosclerosis and hypertension are also risk factors for cerebral small vessel disease. WML and LI are also believed to contribute to the development of dementia [63,77]. Several studies have investigated cross-sectional associations between retinal microvascular signs identified with fundus photography and brain abnormalities identified with MRI. Kwa et al. (2002) were among the first to demonstrate cross-sectional associations between pathologic changes in the retinal arteries and MRI signs of cerebral small vessel disease. In a study of 185 hospital patients (mean age 62 years) with symptomatic atherosclerosis, with and without hypertension, both brain MRI and retinal photographs were obtained. Cerebral small vessel disease was defined as the presence of either WML or LI on MR images. Retinal photos were assessed by two 117 ophthalmologists for retinal artery narrowing, crossings, sclerosis and tortuosity. 92% of patients with small vessel disease had at least one retinal abnormality compared to 77% of patients without small vessel disease (p<0.01). Retinal abnormalities were correlated with MRI signs of small vessel disease (r=0.20, p<0.01); this correlation was somewhat attenuated when adjusted for age (r=0.16, p = 0.03). Except for tortuosity, all retinal abnormalities were more prevalent in patients with cerebral small vessel disease, but correlations were statistically significant only for retinal arterial narrowing (r=0.17, p=0.03) and sclerosis (r=0.18, p=0.02). LI were associated with exudates (r=0.15, p=0.04) and WML correlated with narrowing (r=0.16, p=0.03) and sclerosis (r=0.28, p=0.001). The correlations remained when patients were divided into groups based on hypertensive status [84]. In contrast to Kwa et al. (2002), Ikram et al. (2006) reported that retinal vessel diameters were not cross-sectionally associated with WML or prevalent LI. Included in the study were 490 members randomly selected from the total cohort of the population-based Rotterdam Study who had a cerebral MRI scan and an ophthalmic exam at baseline (Dutch adults with mean age 68 years). 279 of the 490 also had follow-up MRIs approximately 4.5 years later. Diameters of the retinal arterioles and venules were measured (CRAE, CRVE) and the AVR was calculated using a semi-automated system. WML and LI were identified on MRI scans. Retinal arteriolar narrowing, venular dilation and AVR were not cross-sectionally associated with periventricular or subcortical WML or with the presence of LI. However, larger retinal venular diameters were associated with marked progression of both 118 periventricular and subcortical WML as well as with incident LI. WML associations persisted despite adjustment for a number of CVD risk factors including cholesterol, BMI, diabetes, presence of carotid artery plaques and blood pressure. Smaller retinal arteriolar diameters and the AVR were not associated with WML progression or with LI incidence [63]. A series of three analyses using the ARIC cohort have shown that retinopathy is associated with WML, cerebral atrophy and subclinical cerebral infarction. Included in the studies were 1,684 randomly selected members of the cohort (black and white healthy adults with mean age 62 years) who had both cerebral MRI and retinal photography. Brain scans were evaluated for WML, overall volume of periventricular and subcortical white matter signal abnormality, cerebral infarcts (lesions ≥3mm in maximum diameter in a vascular distribution), lacunar infarcts (infarct with maximum diameter <2 cm located in brain stem, thalamus, basal ganglia, internal capsule or deep cerebral white matter), and cerebral atrophy indicated by either sulcal widening or ventricular enlargement. Retinal photos were graded for retinopathy, AV nicking and focal arteriolar narrowing. Generalized arteriolar narrowing was estimated using a semiquantitative method and AVR was calculated. Wong et al. (2002) reported that persons with retinal microvascular abnormalities were significantly more likely to have higher grades of WMLs than persons without abnormalities. After adjustment for age, gender, race and a number of CVD risk factors, all retinopathies assessed except generalized arteriolar narrowing were significantly associated with prevalent WML [177]. In a second 119 report, Wong et al. (2003) reported that persons with any retinopathy, AV nicking and focal arteriolar narrowing were more likely to have higher grades of sulcal widening and ventricular enlargement (greater severity of atrophy) than those without these retinal abnormalities. People with generalized arteriolar narrowing were also more likely to have higher ventricular enlargement but not sulcal widening. After multivariable adjustment, persons with any retinopathy were 1.9 times as likely to have sulcal widening and 1.5 times as likely to have ventricular enlargement. Microaneurysms and retinal hemorrhages were the only two retinopathies to be individually associated with sulcal widening. Focal and generalized arteriolar narrowing were not significantly associated with either sulcal widening or ventricular enlargement [174]. Cooper et al. (2006) reported that cerebral infarction was significantly more prevalent in the presence of all retinal microvascular abnormalities assessed. After multivariable adjustment, AV nicking, focal arteriolar narrowing, retinal hemorrhages and microaneurysms were significantly associated with cerebral infarction; soft exudates and AVR were not. Further analysis revealed that the associations were limited to hypertensive persons only [33]. Longstreth et al. (2006) reported that certain retinal vessel abnormalities were associated cross-sectionally with prevalent and incident infarcts, and with white matter grade and worsening white matter grade. Included in the study were 1,717 members of the population-based Cardiovascular Health Study who underwent two cerebral MRI scans approximately 5 years apart and had retinal photography 120 performed at the time of the second MRI (adults without a history of transient ischemic attack or stroke with mean age 78 years). Neuroradiologists identified infarcts (areas of abnormal signal intensity ≥3mm in maximum diameter in a vascular distribution and without mass effect) and estimated white matter grade (on a 10 point system with higher being more abnormal). Change in the MRI findings from the first to second scan was also assessed, specifically for incident infarcts and worsening white matter grade. Retinal photographs were assessed by graders for the presence of four retinal microvascular abnormalities. Retinopathy was defined as the definite or probable presence of any of the following lesions in any of the four quadrants of the retina: microaneurysms, retinal hemorrhages, soft exudates or cotton wool spots, hard exudates, macular edema, intraretinal microvascular abnormalities, venous beading, new vessels at the disc or elsewhere, vitreous hemorrhage and laser photocoagulation scars. AV nicking and focal arteriolar narrowing were also defined by their definite or probable presence in any of the four quadrants. Retinal photographs were digitized and analyzed with a semi-automated system in which diameters of the retinal arterioles and venules were measured. Summary measures of CRAE, CRVE and AVR were calculated. After adjusting for age and gender, only AV nicking and AVR were significantly associated with prevalent and incident infarcts, and CRAE with incident infarcts. AVR was significantly associated with white matter grade; CRAE was significantly associated with worsening white matter grade. Adjusting for hypertension in the models attenuated the effects, while adjustment for diabetes had little effect [93]. 121 Summary Evidence from the three population-based studies examining associations between retinal microvascular characteristics and cognitive outcomes suggests that generalized narrowing and AV nicking, which are thought to be reflective of early stages of pathophysiological processes in the retinal circulation possibly associated with atherosclerosis and or acute or persistent elevations in blood pressure were not associated with either reduced cognitive function, cognitive impairment or dementia. Retinopathies, which are thought to reflect more advanced stages of pathophysiological processes in the retinal vasculature were associated with all cognitive outcomes reductions in cognitive function, cognitive impairment and dementia in both of the studies that assessed them. These studies also suggest that other retinal microvascular signs such as focal narrowing, and deviations from optimal retinal vascular geometry were associated with the outcomes. It is unclear from a mechanistic standpoint why focal narrowing and the suboptimal geometry, also reflective of earlier pathophysiological stages were associated with reduced cognitive function when AV nicking and generalized narrowing (also in the same class of severity) were not. Also important to note is that, the AVR was modeled continuously or as an ordinal categorical variable in two out of the three studies that used this measure, while other retinal abnormalities were examined as dichotomous (present/absent) variables. Some consistent patterns emerged from studies that had MRI data on white matter lesions and lacunar infarcts, however, there were also several inconsistencies. Generalized narrowing was associated with white matter lesions in one study and not associated in three others. Retinopathies were 122 associated with both white matter lesions and infarcts in one study and not associated with either in another study. As we plan to obtain objective measurements of both retinal arteriolar and venular diameters, calculate the AVR as a measure of generalized arteriolar narrowing, and have data on numerous retinal microvascular abnormalities including focal arteriolar narrowing, AV nicking and several retinopathies, the proposed research will contribute to the existing literature on retinal microvascular characteristics, cognitive function and volumetric brain measurements, and serve to clarify contradictory observations from previous studies. The Latino population is growing, yet is neglected in population-based studies The Latino population comprises 12.5% of the US population as of the 2000 census [164] and is the fastest growing segment of the population. It is estimated that by the year 2050, Latinos will make up almost one-quarter of the US population. While they comprise a heterogeneous group both in ethnicity and nativity, Latinos in general have a lower median household income, a higher percentage of families living below poverty level, lower educational levels and greater percentages in “blue collar” occupations compared to non-Latino whites [162]. Yet despite their lower relative SES, Latinos experience lower all-cause mortality than do non-Latino whites, a phenomenon described by researchers as the “Latino paradox” [97]. The distribution of certain CVD risk factors additionally distinguishes the Latino population from non-Latino populations. Smoking rates, for example, are lower among Latinos compared to non-Latino whites, yet average BMIs tend to be higher. 123 Diabetes is 1.7 times as prevalent among Mexican-Americans than among non- Hispanic whites of similar age [110]. Thus the unique constellation of health protective and disease risk factors among Latinos contribute to the distinctness of this population. In spite of their growing presence, most studies have not included substantial numbers of Latino participants, and only recently have studies such as the Sacramento Area Latino Study on Aging (SALSA) study been designed to focus specifically on CVD and cognitive outcomes in the Latino population. Therefore much of what is known about aging and cognitive outcomes is derived from non- Latino populations. Health promotion and disease prevention programs based on studies of white and black populations may therefore not be appropriately extended to Latinos. Data in Latinos are additionally lacking on the prevalence of certain retinal microvascular abnormalities including general and focal arteriolar narrowing that are indicative of mild to moderate and severe arteriosclerosis. Neither of the three studies examining the association between retinal microvascular abnormalities and cognition included sufficient numbers of Latino participants. Furthermore, few population-based epidemiologic studies of cognitive function among Latinos exist. The Los Angeles Latino Eye Study (LALES) study provides an unparalleled opportunity to further elucidate the relationship between the retinal microvasculature and cognitive function using the wealth of existing data collected in LALES. The Los Angeles Latino Eye Study (LALES) was originally designed as a five-year population-based cross-sectional study of over 6,300 Latinos primarily of Mexican 124 descent aged 40 years and older residing in Los Angeles County (NEI: EY11753 and EY03040). LALES was designed to survey health care and ocular disease in the Latino population, as well as to determine both modifiable and nonmodifiable risk factors associated with ocular disease. Extensive data on demographic characteristics, risk factors, ocular and medical histories were collected on all LALES participants, and detailed ocular examinations were performed. LALES included cognitive screening with the Cognitive Abilities Screening Instrument- Short Form. A supplement to LALES is currently validating this instrument against the Spanish English Neurological Assessment Scales, a more in-depth measure of cognitive function, in a subset of study participants. The proposed research project will utilize the wealth of existing data including retinal photographs collected in LALES, taking advantage of a unique opportunity to further elucidate the relationship between vascular disease and cognitive function. We propose utilizing this data to address hypotheses that disease of the cerebral microvasculature is associated with 1) decreased cognitive function, specifically in areas of verbal memory and conceptual abilities, and 2) brain abnormalities visualized by magnetic resonance imaging (MRI). To test our hypotheses, we will utilize the retinal vessels as anatomic/physiologic surrogates of microvasculature in the brain, and will use the measurements of diameters of retinal vessels to reflect cerebral arteriosclerosis. We hypothesize that focal and generalized retinal arteriolar narrowing will be associated with decreased cognitive function. Our underlying assumption is that changes in the retinal vasculature that are signs of underlying 125 pathology will be reflective of similar changes in and are a valid biomarker of the cerebral vasculature. Summary In summary, with respect to a goal of examining the relationship between vascular disease and cognition, ideally one would directly assess the cerebral vessels. Given that invasive methods are not possible to perform in vivo, and non-invasive imaging methods are impractical for large population-based studies, an alternative approach is required in which surrogates for the cerebral vessels are used. The retina presents such an opportunity and is more amenable to study in vivo given its accessibility. The retinal vessels are anatomically and physiologically similar to cerebral vessels and can be assessed non-invasively through direct visualization. Abnormalities in the retinal microvasculature reflect underlying pathology, may be markers of generalized vascular disease, and are fairly prevalent in the population. Limited studies have addressed correlations between retinal microvascular abnormalities and cognition and small vessel disease. Furthermore, no study to date has included substantial numbers of Latino participants and available data are significantly limited with respect to ethnic/racial variations in the observed associations. Therefore, much remains to be learned in this area. The proposed project will not only significantly contribute to an emerging area of investigation but will also fill gaps in the data on the growing Latino population. 126 Preliminary Studies/Previous Work Associations between CVD risk factors and cognitive function We have previously reported that carotid artery intima-media thickness (IMT), a measure of subclinical atherosclerosis early in the atherosclerotic continuum, is associated with decreased cognitive function in healthy men and women enrolled in the B-Vitamin Atherosclerosis Intervention Trial (BVAIT) [50]. BVAIT is a randomized, double-blind, placebo-controlled, single-center arterial imaging trial designed to test whether B-vitamin supplementation reduces the progression of subclinical atherosclerosis in subjects with elevated fasting plasma homocysteine levels (Hcy), but without clinically evident CVD. A secondary aim of this trial is to evaluate the impact of the B-vitamin intervention on cognitive performance. A total of 506 healthy men and women aged 40 years or older with Hcy >8.5 µmol/L were randomized into two treatment arms: daily oral folic acid 5 mg plus vitamin B 12 0.4 mg plus vitamin B 6 50 mg (combined into 1 pill) or daily B-vitamin placebo. We collected extensive data on demographic characteristics, biochemical variables and behavioral risk factors on all subjects prior to randomization. We had three measures of subclinical atherosclerosis: carotid artery IMT, coronary artery and abdominal aortic calcium (two measures of subclinical atherosclerosis later in the atherosclerotic continuum). We created a variable summarizing the presence or absence of atherosclerosis at the vascular sites assessed by the three measures whose value ranged from 0-3, with 0 indicating the lowest burden of atherosclerosis (CIMT <75 th percentile, absence of aortic or coronary calcium), and 3 indicating the highest 127 burden of atherosclerosis (CIMT >75 th percentile, presence of both aortic and coronary calcium). In BVAIT, cognitive function was assessed with a battery of fifteen neuropsychological tests. We used a principal components analysis to calculate five uncorrelated cognitive factors from scores on individual neuropsychological tests for each participant. Table 13 shows results from adjusted multivariable linear regression models estimating the association between subclinical atherosclerosis and cognitive function. β coefficients indicate the average difference in standard deviation units in performance on each of the five cognitive factors comparing subjects a) with thicker mean CIMT to thinner CIMT (per 0.1 mm) or, b) with the greatest burden of atherosclerosis to those with less burden of atherosclerosis. Table 13. Associations between Cognitive Factor Scores and Subclinical Atherosclerosis Measures from Regression Models* for 504 BVAIT Subjects 128 CognitiveFactor β(SE), p-value† Subclinical Atherosclerosis Measure Executive Function Verbal Learning Logical Memory Visual Memory Semantic Memory Global Cognition Mean CIMT (per 0.1 mm) 0.02 (0.03) 0.44 -0.07 (0.03) 0.01 - 0.01 (0.03) 0.81 0.03 (0.03) 0.38 - 0.04(0.03) 0.15 -0.02(0.03) 0.44 Composite score‡ 3 vs. 0-2 -0.07(0.12) 0.57 0.01 (0.13) 0.94 -0.17 (0.14) 0.22 -0.05 (0.14) 0.71 -0.16 (0.13) 0.22 -0.19 (0.11) 0.09 *Adjusted for age, gender, race/ethnicity, education, income, CES-D score, Hcy, SBP, LDL-C, smoking status † p-value for comparison of one category to reference category ‡ n=498, composite score dichotomized at 3 versus 0, 1 and 2 129 Increasing thickness of CIMT was associated with significantly lower performance in the verbal memory/conceptual ability area of cognition ( β = -0.07 per 0.1 mm increase CIMT [SE( β)=0.03], p=0.01), but not in other areas of cognitive function. Compared to subjects with CIMT in the lowest quartile (<0.65 mm), subjects with CIMT in the highest quartile ( ≥0.83 mm) had significantly lower performance in the verbal memory/conceptual ability domain ( β = -0.27 [SE( β)=0.12], p=0.03). The presence of calcium in the abdominal aortic or coronary arteries was not individually associated with any of the five areas of cognitive function. Our results suggest that in healthy, non-demented adults, CIMT, which occurs earlier in the atherosclerotic process, has more influence on cognitive function, specifically verbal memory and conceptual abilities, than calcifications of the arteries, which occur later in the atherosclerotic process. One explanation for this observation is that CIMT more closely reflects intracranial atherosclerosis and its resultant localized pathology of infarction or ischemia due to blockage or deterioration of cerebral arterioles than does calcification in other arteries. Another possibility is that individuals with greater carotid atherosclerosis also have more cerebral atherosclerosis, which individually leads to declines in cognition. We have also reported that metabolic syndrome, a summary measure of major cardiovascular and metabolic risk factors associated with CVD and diabetes, is associated with decreased cognitive function among healthy cognitively intact men and women without diabetes (n=853) from the combination of subjects randomized in BVAIT and the Women's Isoflavone Soy Health (WISH) Trial. [51]. WISH is a 130 randomized, double-blind, placebo-controlled, noninvasive ultrasonographic trial designed to test whether dietary supplementation of soy isoflavones reduces the progression of early carotid artery atherosclerosis in healthy postmenopausal women without clinically evident CVD. A secondary aim of this trial is to evaluate whether cognitive function would be improved with the administration of the soy isoflavones. Postmenopausal women aged 30 years or older were eligible for WISH. Exclusion criteria were similar to those for BVAIT, and additionally included current use of hormone therapy. A total of 350 eligible women were randomized into two treatment arms in WISH: daily 25g soy protein (150 mg total isoflavones) or daily total milk protein matched placebo (0 mg isoflavones). We used an expanded definition of the National Cholesterol Education Program (NCEP) criteria to identify subjects with the metabolic syndrome [85,108]. Cognitive outcomes included the five cognitive factors, as well as a measure of global cognition. Presence of the metabolic syndrome (i.e., meeting at least 3 of 5 NCEP criteria) was weakly associated with decreases in verbal memory and conceptual abilities ( β = -0.16 [SE( β) = 0.09], p = 0.08). As the number of MetS criteria increased, global cognition and verbal memory and conceptual abilities decreased (p-trend both = 0.01), and tended to decrease in semantic memory and visuospatial abilities (p-trend = 0.07). Hypertension was the only MetS risk factor that was independently correlated with decreased verbal memory and conceptual abilities ( β = -0.26 [SE( β) = 0.07], p = 0.0004), semantic memory and visuospatial abilities ( β = -0.15 [SE( β) = 0.07], p = 0.03) and global cognition ( β = -0.12 [SE( β) = 0.06], p = 0.06). Our results suggest that in healthy, non-demented adults without CVD and diabetes, the 131 metabolic syndrome, and particularly hypertension contribute to cognitive dysfunction. Associations between retinal microvascular abnormalities and cognitive function As noted above, our collaborators at the University of Melbourne have reported associations between certain retinal microvascular abnormalities and reduced cognitive function among middle-aged persons without stroke [179]. Mean scores on Delayed Word Recall, Digit Symbol and Word Fluency tests were significantly lower among participants with any retinopathy, retinal hemorrhage, microaneurysm or soft exudates compared to participants without these lesions after adjusting for a number of demographic and CVD risk factors. Participants with AV nicking or generalized arteriolar narrowing did not significantly differ in performance on any of the cognitive tests compared to those without the abnormalities. The odds of cognitive impairment were significantly higher among participants with any retinopathy compared to those without, using Delayed Word Recall (OR = 2.60, 95% CI = 1.70-3.99), Digit Symbol (OR = 1.91, 95% CI = 1.04-3.49) or Word Fluency (OR =2.03, 95% CI =1.07-3.86) tests to define cognitive impairment after controlling for the covariates mentioned above. Odds of cognitive impairment were also increased for participants with retinal hemorrhage, microaneurysm, soft exudates, AV nicking or generalized arteriolar narrowing compared to those without. There was no significant interaction by diabetes or hypertension, and results were similar in analyses by subgroups of age, sex and race. 132 Research Design and Methods LALES Study Design LALES was originally designed as a five-year population-based cross-sectional study initiated in February 2000 of over 6,300 Latinos primarily of Mexican descent aged 40 years and older residing in six census tracts in the city of La Puente in Los Angeles County (NEI: EY11753 and EY03040). LALES was designed to survey health care and ocular disease in the Latino population, as well as to determine both modifiable and nonmodifiable risk indicators associated with ocular disease. Extensive data on demographic characteristics, risk factors for ocular disease, ocular and medical histories, and access to medical and ocular care were collected on all LALES participants, and detailed ocular examinations were performed. LALES included several operational strategies in its study design to accommodate the relatively young, working Latino population, including offering evening and weekend clinic hours, bilingual field and clinic staff, free child care at and transportation to the clinic. Approval for the study was obtained from the Los Angeles County/University of Southern California Medical Center Institutional Review Board. LALES had five specific aims: 1) to determine age-specific prevalence of blindness, visual impairment, and ocular disease among Latinos aged 40 years or older; 2) to determine the proportion of the prevalence of blindness and visual impairment attributed to refractive error, lens opacities, glaucoma, diabetic retinopathy (DR), and 133 age-related maculopathy; 3) to evaluate the importance of risk factors and the extent to which such factors may be associated with visual impairment and the prevalence of each ocular disease; 4) to determine the impact of blindness, visual impairment, and presence of ocular disease and comorbid medical conditions on self-reported visual impairment and health-related quality of life; and 5) to evaluate utilization of eye care and general health care services. Primary outcome measures in LALES included blindness, visual impairment, lens opacities, diabetic retinopathy, open- angle glaucoma as well as quality of life and health care utilization. The proposed study will examine cross-sectional associations between vascular disease and cognitive function using the retinal vessels as markers for analogous vessels in the brain. We will specifically use general and focal retinal arteriolar narrowing and retinal microvascular abnormalities as indicators of the prevalence and severity of retinal microvascular disease. We will use the sample of LALES participants for whom cognitive screening data was collected (n=2,931) as our study population. Additional cross-sectional analyses will focus on two sub-samples of the study population who 1) had more in-depth cognitive testing with the Spanish English Neuropsychological Assessment Scales and 2) had brain MRIs performed. Subject Ascertainment The LALES base population included all eligible individuals from all dwelling units within the 6 census tracts in or around La Puente, California. These individuals were identified by the LALES Survey Research Center, approached at their home, and 134 invited for a detailed clinical and eye examination, which was performed in a standardized manner at the LALES local eye examination center. The La Puente area was chosen because it is primarily residential, included a large proportion (83%) of Latinos, had enough individuals to obtain precise prevalence estimates of ocular disease, benefits from high levels of support and encouragement from community and church leaders and local health care practitioners, is in close proximity to Los Angeles County/University of Southern California Medical Center, and has a similar age distribution of the Latino population to that of Latinos in the United States. To determine household residence, LALES used the US Census definition of resident, which counts anyone who considers the home his or her permanent residence, lives and sleeps at the residence the majority of the time, or lives in the household at least six months of the year. Individuals were eligible to participate in LALES if they 1) resided in one of the selected La Puente census tracts, 2) self- identified as being Latino or of Latino heritage, and 3) were 40 years or older on the day of the household screening. Of 10,663 individuals screened for participation in LALES, 7,789 (73.0%) were eligible for LALES. Of those eligible, 6,357 were recruited and completed a clinical examination (included both medical and ocular components), for a participation rate of 82%. In 2001, the scope of LALES was expanded to include cognitive screening with the Cognitive Abilities Screening Instrument-Short version (CASI-S) through funding by a pilot project with the USC Alzheimer’s Disease Research Center (ADRC) (Administrative Supplement; PI Wendy Mack). 2,931 (46%) LALES 135 participants were administered the CASI-S as part of the questionnaire portion of their in-home visit. Demographic characteristics and select CVD risk factors of LALES participants who completed an ocular exam and were screened with CASI-S (n=2,584) are summarized in Table 14. Hypertension and diabetes, two major CVD risk factors are prevalent in the LALES population (30% and 17%, respectively), and the majority (76%) has less than a high school education. Table 14. Characteristics of LALES Subjects with Cognitive Screening with the CASI-S (n=2,584) Characteristic Mean ± SD or Number (%) Age (years) 54.6 ± 11.0 40-49 yrs 37% 50-59 yrs 30% 60-69 yrs 20% 70-79 yrs 10% ≥ 80yrs 3% Gender (female) 58% Country of birth United States 22% Mexico 66% Other 12% Marital status (married) 68% Educational level ≥12 years 34% Working status (employed) 41% Annual income level >$40,000 13% Covered by health insurance 64% History of hypertension 30% History of angina 4% History of diabetes 17% Current smoker 38% Since 2003-2004, an additional ADRC research project has been providing funding to supplement LALES with more in-depth cognitive testing with the Spanish English Neuropsychological Assessment Scales (SENAS). An elderly ( ≥60 years old) subsample (n=400, 15.5%) of LALES participants who had cognitive screening with the CASI-S are being selected to validate the CASI-S as a screening instrument for dementia against the SENAS, and to improve the detection of mild cognitive impairment (ADRC research project; PI: Wendy Mack). This research project is also funding an imaging component, in which MRIs are being obtained on an estimated 128 subjects anticipated to be identified as demented or cognitively impaired (based on SENAS scores), as well a random sample of cognitively unimpaired subjects for a total of 180 MRIs. Figure 2. LALES Cognitive Sub-Samples Quantitative measurement of retinal vessel calibers & AVR using retinal photos (U of Melbourne) n=2,584 Brain MRI study (volumetric measures) sample with CI/dementia (n=60) & cognitively intact (n=120) n=180 Validation study with SENAS (6 cognitive domains) sample >=60 years old n=400 Cognitive Screening with CASI-S n=2,931 LALES n=6,357 Medical Exam Questionnaire/Interview Ocular Exam - Retinal photographs graded for retinopathies (U of Wisconsin) 136 137 Sample sizes We plan to include in analyses LALES subjects and subsets of LALES subjects for whom retinal photographs were taken during their ocular exam, were screened with the CASI-S and tested with the SENAS, as well as those for whom brain imaging from MRI studies was obtained. Subgroups of participants will be included in each sub-analysis as specified below. Specific Aim 1 - To describe the distribution of retinal vessel diameters and the prevalence of generalized retinal arteriolar narrowing and microvascular abnormalities in a Latino population. This aim will be addressed with two subsamples of LALES participants: 1) to describe for the prevalence of any retinopathy, we will use the sample who completed an ocular exam with retinal photography and whose retinal photographs were graded for retinopathies by collaborators at the University of Wisconsin, Madison (n=6,357); 2) to describe the distribution of retinal vessel diameters (CRVE, CRAE) and generalized arteriolar narrowing (AVR), we will use the sample for whom retinal fundus photographs were graded for retinal vessel calibers by collaborators at University of Melbourne (n= 2,584). Specific Aim 2 - To examine whether retinal microvascular abnormalities are associated with decreased cognitive function (CASI-S) and cognitive function in specific domains of cognition (SENAS). This aim will be addressed with two subsamples of LALES participants: 1) to examine associations with overall cognitive function, we will use the sample of participants for whom both 138 CASI-S was obtained and retinal fundus photographs were available and graded for retinal vessel caliber (n= 2,584); 2) to examine associations with specific areas of cognitive function, we will use the sample of participants for whom more in-depth cognitive testing was done with the SENAS and for whom retinal fundus photographs were available and graded (n=400). Specific Aim 3 - To investigate the association between retinal microvascular measures and cognitive impairment and dementia (SENAS). This aim will be addressed using the subsample of LALES participants for whom more in-depth cognitive testing was done with the SENAS and for whom retinal fundus photographs were available and graded for retinal vessel calibers (n=400). Specific Aim 4 - To assess whether retinal microvascular abnormalities are associated with cerebral volume, total lacune volume and white matter hyperintensity volume from magnetic resonance imaging (MRI) data. This aim will be addressed using the subsample of LALES participants for whom MRIs are being obtained, and retinal fundus photographs were available and graded (n= 180). Data Collection A major advantage of this research proposal is that it does not require additional data collection from study participants, making it extremely cost-effective. Retinal photographs have already been obtained from LALES participants and CASI-S has been administered to participants. Validation with the SENAS is currently in progress as are the MRI studies. The proposed project will use existing retinal vessel 139 data, CASI-S and SENAS cognitive data. Grading of retinal vessel calibers is currently being done by collaborators at the University of Melbourne. We will obtain and process brain imaging data from MRI studies. These sources of data will be merged in a single database for analyses. In-home interview In LALES, a questionnaire assessing sociodemographic factors, ocular and medical histories and access to medical and ocular care was administered to participants during the in-home component of the study by a member of the LALES field staff. All data were collected via computer-assisted in-person interviews. Interviewers used the preferred language (English or Spanish) of each individual. After the in- home interview was complete, clinic appointments were arranged for participants to receive a free medical and ocular examination at the Local Eye Examination Center. Participants who were not able to complete an exam at the clinic were asked to undergo an exam in their homes by a trained ophthalmologist and a trained technician. This examination included presenting and best-corrected distance VA measurement (using the ETDRS examination protocol), IOP measurement with the Tono-Pen XL, 24-2 threshold VF testing (Oculus Easyfield, Oculus, Woodinville, WA), an anterior segment examination using a portable slit lamp, and a dilated fundus examination. 140 In-clinic medical and ocular exams The clinical examination included an in-clinic medical and ocular examination and an in-clinic interview. As the examination was performed, data were entered directly into the LALES Microsoft ACCESS database. The procedures for data collection were standardized and included a series of measurements taken by ophthalmic technicians, interviewers, and an ophthalmologist. The full in-clinic examination consisted of the following components: 1) verification of name, date of birth, age, gender, street address, and informed consent; 2) in-clinic interview using a questionnaire that included questions relating to quality of life and eye care utilization; 3) measurements of height, weight, and waist–hip ratio; 4) urine collection; 5) measurements of glycosylated hemoglobin and random blood glucose and storage of blood for future DNA analysis; 6) pulse rate and blood pressure; 7) lensometry; 8) visual acuity (VA) measured with the revised ETDRS charts 1, 2, and 3 and Lea symbol charts for illiterate participants (VA measurements were attempted at 1 m for those participants who read fewer than 20 letters at 4 m); 9) automated refraction with a Humphrey Automatic Refractor (Carl Zeiss Meditec, Dublin, CA) if the presenting VA was not 20/20 in either eye; 10) near vision with the participant's present reading prescription using the Modified ETDRS Near Vision Acuity Chart; 11) iris color grading using the Iris Color Classification System [144]; 12) pupil assessment with the Rosenbaum pupil screener (pocket version); 13) intraocular pressure (IOP) measurements with both a Goldman applanation tonometer and a Tono-Pen XL (Medtronic Ophthalmics, Jacksonville, FL); 14) anterior and posterior segment examinations at the slit lamp with a 90-diopter lens and indirect 141 ophthalmoscopy; 15) gonioscopy only for those participants with narrow angles observed in the slit-lamp examination; 16) lens grading using the LOCS II system [31]; 17) VF test using the Humphrey Automated Field Analyzer (Swedish interactive thresholding algorithm Standard 24-2); 18) fundus photography using the Topcon TRC 50EX Retinal Camera (Topcon Corporation of America, Paramus, NJ) (3 stereoscopic fields were done on all nondiabetic participants, and 7 stereoscopic fields on all diabetic participants; stereoscopic disc photography was also performed); and 19) central corneal thickness and axial length measurements using an ultrasonic A-scan/pachymeter DGH 4000B SBH IOL Computation module (DGH Tech Inc., Exton, PA). Summary: Data from interview or examination Data obtained from questionnaires administered during the in-home interview or clinical exam will be used to identify other co-morbidities and vascular risk factors. Participants were asked whether they were ever diagnosed or received treatment for a list of health conditions. Self-reported diagnoses of hypertension (or treatment for), diabetes, heart disease, peripheral vascular disease, glaucoma, macular degeneration, or diabetic retinopathy were queried. Participants were asked to list all current medications. Measured blood pressure and obesity [body mass index (BMI)] were obtained. To be consistent with the original definitions in LALES, we will identify subjects as diabetic if they either self-report that they have been diagnosed as diabetic, or have a non-fasting glucose > 200 mg/dL. Hypertensive subjects will 142 be identified as those who self-report a diagnosis of hypertension, or have a clinic- measured blood pressure > 140/90. Using CDC criteria, measured BMI will be used to define subjects as normal weight (BMI<25 kg/m 2 ), overweight (BMI>25 kg/m 2 ) or obese (BMI>30 kg/m 2 ). Serum lipids will be obtained on all subjects participating in the LALES supplement. Active and passive smoking and alcohol consumption asked on questionnaires will be used to determine smoking status. Cognitive function assessment: CASI-S and SENAS In 2001, the scope of screening in LALES was expanded to include cognitive screening with the Cognitive Abilities Screening Instrument-Short form (CASI-S) through funding by a pilot project of the USC Alzheimer’s Disease Research Center (ADRC) (Administrative Supplement; PI Wendy Mack). 2,931 (46%) LALES participants were administered the CASI-S as part of the in-home interview portion of the LALES study. Given the predominantly Latino population of Mexican ancestry in LALES, an important study design issue was the choice of measures to assess cognitive function that would be linguistically and culturally appropriate. The CASI-S was selected to screen cognitive function in LALES for its brevity and ease of instruction, as well as because it may be administered in either English or Spanish. The CASI-S is a four-item abbreviated version of the Cognitive Abilities Screening Instrument (CASI) [157], which was developed [158] as a brief screening instrument to identify demented individuals of various levels of education. The CASI-S was adapted for LALES to include computer-assisted administration and scoring; score range on the CASI-S is 0-33. Sensitivity and specificity of the CASI-S to identify 143 individuals with dementia for subjects whose education level is 10-22 years using a score ≤24 are 90% and 94%, respectively, and 96% and 80%, respectively for subjects with 4-9 years of education. For subjects with 0-3 years of education, using a CASI-S score ≤21 has a sensitivity of 94% and specificity of 68%. Since 2003-2004, an ADRC research project has been providing funding to augment LALES with an additional cognitive component. An elderly ( ≥60 years old) subsample (n=400, 13.6%) of LALES participants who had cognitive screening with the CASI-S are being selected to validate the CASI-S as a screening instrument for dementia against the Spanish English Neuropsychological Assessment Scale (SENAS) and to improve the detection of mild cognitive impairment. (ADRC research project; PI: Wendy Mack). This subset of participants are being administered the SENAS in either English or Spanish during a follow-up LALES visit. The SENAS was developed [105] to test a range of cognitive abilities relevant to the neuropsychological evaluation of older persons from different ethnic groups. Furthermore, considerations were given in its development to include a broad set of cognitive measures that would be psychometrically matched both to each other and between Spanish and English versions. Thus, the SENAS was chosen for the LALES supplement because its psychometric measures are applicable for mixed Hispanic and Caucasian samples of older adults. The SENAS was additionally chosen because it may be utilized to characterize cognitive profiles related to different brain pathologies such as Alzheimer’s disease, cerebrovascular disease (“vascular cognitive impairment”), Parkinson’s disease and frontotemporal lobe 144 dementia, and thus could be used to validate the CASI-S in screening for dementia. The SENAS consists of 13 scales that assess both verbal and non-verbal aspects of 6 cognitive domains: conceptual thinking, semantic memory, attention span, episodic memory, non-verbal/spatial abilities and verbal abilities (Table 15). Average internal consistency reliability coefficients for SENAS non-memory scales are 0.85 for non-Hispanic English speakers, 0.86 for Hispanic English speakers and 0.88 for Hispanic Spanish speakers, indicating uniformly high and consistent reliability across groups. Values on the entire battery of the SENAS are comparable to or superior to those reported for the Wechsler Adult Intelligence and Memory Scales (WAIS-III) [171]. The SENAS has also been shown to have high construct validity based on structural equation modeling relating scales to latent variables in both English and Spanish in four of the six domains (semantic memory, episodic memory, non-verbal/spatial abilities and verbal abilities) (i.e., inferences can legitimately be made from the domains assessed to actual cognitive abilities). 145 Table 15. Scales of SENAS neuropsychological test battery and abilities measured Cognitive Domain Verbal Measure Non-Verbal Measure Conceptual Thinking 1. Verbal Conceptual Thinking 2. Non-Verbal Conceptual Thinking Semantic Memory 3. Object Naming 4. Picture Association Attention Span 5. Verbal Attention Span 6. Visual Attention Span Episodic Memory 7. Word List Learning – I & 8. Word List Learning - II 9. Spatial Configuration Learning Non-Verbal/Spatial Abilities 10. Pattern Recognition 11. Spatial Localization Verbal Abilities 12. Verbal Comprehension 13. Verbal Expression reproduced from Mungas et al.[105] Retinal Fundus Photographs Fundus photography is commonly used in research and clinical settings to document the retina [141], and resulting retinal photographs can be used to identify characteristics of the retinal vasculature. In LALES, retinal fundus photographs were shot on 35mm Ektachrome 100 slide film (Kodak, Rochester, NY) with a Topcon TRC 50EX Retinal Camera and filed with study participants’ records. The stereoscopic fundus photographs were graded for retinopathies by collaborators at the Ocular Epidemiology Grading Center at the University of Wisconsin, Madison who were blinded to subject identity. Retinopathies assessed included microaneurysms, retinal hemorrhages, drusen (soft or hard exudates), macular edema, intraretinal microvascular abnormalities, venous beading, neovascularization on the 146 disc or elsewhere, vitreous hemorrhage and disc swelling and laser photocoagulation scars. Quantitative measurements of retinal vessel caliber are the basis for assessing retinal arteriolar and venular diameters, and calculating a measure of generalized retinal arteriolar narrowing. We have established a collaboration with the lab of Tien Wong, PhD at the Center for Eye Research Australia at the University of Melbourne in Melbourne, Australia. For each participant who completed an ocular exam, in which retinal fundus photographs were taken, two stereoscopic slides representing field 1 (F1) (centered on the temporal edge of the optic disc) of each eye will be sent to the Center for Eye Research Australia for grading. The slides will be digitized using a Nikon Scanner at 180dpi and saved as JPEG files of 1-2MB. The digital images will then be uploaded into a computerized semi-automated vessel measurement program, which will be used to quantify retinal arteriolar and venular calibers. This software measures the diameter of both the arterioles and venules that course through a specific region of the eye located 0.5 to 1.0 optic disc diameters from the optic disc margin. The average diameters for all arterioles and venules measured in this region will be summarized as the central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE). These methods are based on obtaining the biggest 6 arterioles and venules according to the formulas by Knudtson et al. [76]. A third measure known as the arterio-to-venule ratio (AVR) will be calculated from the ratio of CRAE:CRVE. The AVR is a relative measure of the caliber of arterioles to venules and thus adjusts partially for variable magnification 147 from differences in refractive errors. An AVR of 0.50 may be interpreted as the arteriolar caliber is on average, 50% as large as the venular caliber, while an AVR of 1.50 may be interpreted as the arteriolar caliber is on average 50% greater than the venular caliber. AVR will be used as an indicator of generalized arteriolar narrowing with lower numbers indicative of greater narrowing. These grading methods have been shown to be reliable [62]. Replicate grading studies for deriving the AVR have demonstrated intragrader correlation coefficients of 0.84, and intergrader correlation coefficients of 0.79, indicating that the measures are highly reproducible [62]. The combined data from grading retinal photographs (based on the work of collaborators at the University of Wisconsin and the University of Melbourne) will allow us to examine four categories of retinal microvascular abnormalities: 1) retinopathies (any microvascular characteristics not explicitly arteriolar in nature) including microaneurysms, hemorrhages (blot, flame-shaped and vitreous), soft and hard exudates, cotton wool spots, macular edema, intraretinal microvascular abnormalities, venous beading, new vessels at the disc or elsewhere; and three categories representing abnormalities specific to the retinal arterioles: 2) focal arteriolar narrowing; 3) generalized arteriolar narrowing; 4) arteriovenous nicking. MRI studies to identify brain abnormalities The ADRC supplement is also funding an imaging component of LALES, in which MRIs are being obtained on the estimated 128 subjects found to be demented or 148 cognitively impaired (based on SENAS scores), as well a random sample of cognitively unimpaired subjects for a total of 180 MRIs. The MRI studies are being performed at the Queen of the Valley MRI Center, an MRI center local to La Puente, on a GE Signa High Definition 1.5 Tesla LX 9.1 MR system. T1-, T2-, and proton density-weighted MR images are being acquired following imaging protocol developed at the UC Davis Alzheimer’s Disease Center. Raw data from the studies are being archived as the exams are completed. For the proposed research, we will need raw MRI data to be processed in order to obtain volumetric brain measurements. To achieve this aim, we are collaborating with Charles DeCarli at UC Davis who will coordinate the processing of MRI data to obtain volumetric measures of cortical white and grey matter, and identify cerebral lacunar infarcts and white matter hyperintensities. Brain MRI pixels will be classified into subcortical gray matter, white matter, hippocampal grey matter, ventricular CSF, and sulcal CSF using a computerized segmentation algorithm. All pixels within the cranium will be summed to compute total intracranial volume. The processed MRI data will include measures of cortical gray matter volume, white matter volume, ventricular and sulcal CSF (as measures of cerebral atrophy), hippocampal volume, total lacune volume and white matter hyperintensity volume. All volumetric measures will be computed as percentages of total intracranial volume to adjust for differences in brain size. 149 Data Analysis Data sources, entry and management This proposal utilizes existing data from LALES and the ADRC LALES supplement, which has been cleaned and entered in study databases as part of LALES, or is in the process of being collected, cleaned and entered for the ADRC supplement. At the start of our analysis, we will have databases pertaining to LALES demographic, clinical exam, retinal photography, CASI-S and SENAS data. We will have obtained an electronic database from our collaborators at the Center for Eye Research Australia, which will contain variables for CRAE, CRVE and AVR by study ID. We will have MRI data for cortical gray matter volume, white matter volume, cerebral volume, total lacune volume and white matter hyperintensity volume from our UC Davis collaborators. We will thus have data on demographic and lifestyle factors including age, gender, educational level, SES, smoking history and alcohol consumption; quality of life; mood/depression; biologic and physiologic measures including blood pressure/hypertension; risk factors including prevalent CVD; other comorbidities such as diabetes and ocular disorders/diseases for all subgroups of LALES participants included in the proposed analyses. Serum glucose levels will be available for participants with self-reported diagnoses of diabetes, n=439. Lipid levels will be obtained for the subsample of LALES participants for whom more in- depth cognitive testing is being done with the SENAS and for whom retinal fundus photographs were available and graded for retinal vessel calibers (n=400). 150 Sample size and study power. Study power is detailed below for the specific aims. Specific Aim 1 - To describe the distribution of retinal vessel diameters and the prevalence of generalized arteriolar narrowing and microvascular abnormalities. We will calculate simple descriptive statistics (mean, standard deviation, 95% confidence interval) for two subsets of LALES participants. 1) To estimate the prevalence of retinal microvascular abnormalities, we will use the 6,357 for whom retinal fundus photographs were graded for retinopathies. Preliminary data suggest that 6% of the LALES population with gradeable retinal fundus photographs had at least one retinopathy. For a two-sided 95% confidence interval and assuming a standard deviation of 5 (using data from the ARIC study [179]), we can expect a confidence interval width of 0.122. 2) To describe the distribution of retinal vessel diameters (continuous statistics), we will use the 2,584 LALES participants for whom retinal fundus photographs were graded for retinal vessel calibers. For a two- sided 95% confidence interval and assuming a standard deviation of 6 (using data from the Patton et. al study [121]), we can expect a confidence interval width of 0.23. To the estimate the prevalence of generalized arteriolar narrowing we will also use the 2,584 for whom retinal fundus photographs were graded for retinal vessel calibers. As generalized arteriolar narrowing will be defined as the lowest 20% of the sample distribution, per methods previously specified for the ARIC study [179], we thus expect 20% of the population (n=517) to have generalized narrowing. For a 151 two-sided 95% confidence interval, we can expect a confidence interval width of 0.015. Specific Aim 2 - To examine whether retinal microvascular abnormalities are associated with decreased cognitive function (CASI-S) and cognitive function in specific domains of cognition (SENAS). 1) Using CASI-S to assess cognitive function: We will use multiple linear regression analyses of the 2,584 participants for whom both CASI-S data and retinal fundus photographs were graded for both retinal microvascular abnormalities and vessel calibers. This will allow us to detect correlations of 0.06 or greater at 80% power and 2-sided alpha of 0.05. Adjusting the significance level for a 2-sided alpha of 0.01 to reflect tests of multiple retinal vessel measures (i.e., 4 categories of retinal microvascular abnormalities) we will be able to detect correlations of 0.07 or greater at 80% power. 2) Using SENAS to assess specific domains of cognitive function: We will use multiple linear regression analyses of the 400 participants ≥ 60 years for whom SENAS was obtained and retinal fundus photographs were graded. This will allow us to detect correlations of at least 0.17 at 80% power using a 2-sided alpha of 0.01. Specific Aim 3 - To investigate the association between retinal microvascular measures and cognitive impairment and dementia (SENAS). We will use multiple logistic regression analyses of the 400 LALES participants for whom more in-depth cognitive testing was done with the SENAS and for whom retinal photographs were graded. From the SALSA study estimating dementia prevalence in a Latino population [56], we estimate that approximately 5% or 20 152 participants will be demented. From a review of estimates for mild cognitive impairment [14], we estimate that approximately 10% or 40 participants will be cognitively impaired. Using the entire sample of 400 subjects with the following anticipated breakdown (60 demented or cognitively impaired, 340 unimpaired), we will be able to detect an odds ratio for the comparison of demented/cognitively impaired vs. unimpaired of 1.4 or greater for dichotomous exposure variables (e.g., presence/absence of the retinal microvascular abnormality) with 80% power and testing at a two-sided alpha of 0.01. For continuous or ordinal categorical variables (e.g., AVR), we will be able to detect odds ratios (expressed per SD of the independent variable) of 1.23 and greater, with 80% power and testing at a two-sided alpha level of 0.01. Specific Aim 4 - To assess whether retinal microvascular abnormalities are associated with volumetric brain measurements from magnetic resonance imaging (MRI) data. We will use multiple linear regression analyses of the 180 LALES participants (60 identified as demented/cognitively impaired, 120 unimpaired) for whom MRIs were obtained and retinal photographs were graded. For simple linear regression analyses of retinopathies and volumetric brain measurements, we will be able to detect correlation coefficients of 0.25 and greater with 80% power and a two-sided alpha of 0.01. 153 Statistical analysis plan Preparation of data for analyses Data files from multiple sources will be merged in a single database for analyses. Using the merged dataset, we will create variables needed for analyses pertaining to our specific aims. We will create a dichotomous (yes/no) variable representing any retinopathy (i.e., yes for the presence of any retinopathy, no for the absence of any retinopathy), as well as for total number of retinal microvascular abnormalities (i.e., sum of retinal abnormalities identified). We will also create an ordinal categorical variable (values 1-4, with higher values indicating greater severity) representing severity of retinal vessel abnormality. Determination of severity will be based on previous reports of human and animal data which suggest that retinal microvascular changes occur in four phases [176]. Categories will be assigned as follows: 1: generalized and focal arteriolar narrowing, 2: generalized and focal arteriolar narrowing with tortuosity, AV nicking or increases in the angle of arteriolar branching, 3: microaneurysms, retinal hemorrhages, soft or hard exudates and 4: central or branch retinal artery and vein occlusions, macroaneurysms, macular edema. We will create a variable for generalized arteriolar narrowing, which will be defined “present” for an AVR ≤ the lowest 20% of the sample distribution, and “absent” for an AVR > 20% of the sample distribution. We will use the total CASI-S score to examine cognitive function as a continuous variable. We will also create a dichotomous variable indicating cognitive dysfunction versus normal cognition using CASI-S data. Individuals will be 154 identified as having cognitive dysfunction if their age-, gender- and education- adjusted CASI-S scores are at or below the 10 th percentile for the LALES population. Individuals with CASI-S scores above the 10 th percentile for the LALES population will be categorized as cognitively normal. We will also use the total SENAS score to examine cognitive function as a continuous variable. We will calculate individual scores for each of the six cognitive domains assessed on SENAS. Five SENAS scales are used in the process of establishing clinical diagnoses of dementia or cognitive impairment. These include Verbal Memory, Verbal Attention Span, Object Naming, Verbal Abstraction and Pattern Recognition [45]. LALES participants will be classified as cognitively normal, cognitively impaired, or demented using SENAS scores as follows. First, individuals will be identified as having dementia if their demographically adjusted score or raw score on two of the five diagnostic SENAS tests was at or below the 10 th percentile of a non-demented normative sample. Individuals will be classified as having cognitive impairment if their demographically adjusted score on one of the five diagnostic SENAS tests was at or below 10 th percentile of the normative sample. Individuals who were not demented or not cognitively impaired will be classified as cognitively normal. We will examine distributions of our data and identify possible outliers of variables that specifically pertain to our analyses. We will investigate variables that may be non-normally distributed, and will explore various transformations of non-normal 155 data. We will also assess the extent of missing data, again for variables that specifically pertain to our analyses. We will begin with descriptive analyses that will include frequency distributions, means and standard deviations of all variables relevant to our analyses. We will conduct simple correlations of CASI-S and SENAS scores and characteristics of retinal vessels by categories of demographic factors including age, gender, education and SES, and CVD risk factors including BMI, smoking status, hypertension and diabetes. Analyses pertaining to Specific Aims Specific Aim 1 - To describe the distribution of retinal vessel diameters and the prevalence of generalized arteriolar narrowing (n=2,584), and prevalence of microvascular abnormalities (n=6,357). We will report on distributions of diameters of retinal vessels (CRAE, CRVE) (means, standard deviations, 95% confidence intervals) for the sample with retinal vessel caliber grading, and on the prevalence of retinal microvascular abnormalities (numbers, proportions, standard deviations, 95% confidence intervals) for the sample with retinal fundus photography grading. We will also report this data by strata based on gender, age decades (40-49, 50-59, 60-69, 70-79, ≥80 years), hypertension and diabetes status. We will calculate quartiles and quintiles of retinal vessel diameters, both for the entire population as well as for strata of the population stated above. We will report the prevalence of generalized arteriolar narrowing (AVR ≤ 20 th 156 percentile) (proportion, variance estimate). This descriptive aim will thus produce novel data on characteristics of retinal vessels in the Latino population. Specific Aim 2 - To examine whether retinal microvascular abnormalities are associated with decreased cognitive function (CASI-S) (n=2,584) and cognitive function in specific domains of cognition (SENAS) (n=400). We will use multiple linear and logistic regression methods to evaluate whether retinal microvascular abnormalities are associated with cognitive function. Variables for microvascular abnormalities modeled will include the presence vs. absence of generalized arteriolar narrowing and any retinopathy, as well as a continuous variable for total number of retinal microvascular abnormalities (i.e., sum of retinal abnormalities identified), and an ordinal categorical variable (ranging from 1-4) representing severity of retinal vessel abnormalities. Dependent variables will be CASI-S (total score) and SENAS (domain-specific scores). We will model CASI-S score as a continuous variable to assess the range of cognitive function. We will also model CASI-S score as a dichotomous variable (<10 th percentile vs. ≥10 th percentile of age-, gender-, education-specific values) to compare individuals with cognitive dysfunction to cognitively normal individuals. We will model scores on the six SENAS domains as continuous variables. Other independent variables of interest will include binary risk factors (diabetes and hypertension/use of antihypertensive agents), ordinal categorical risk factors (BMI categorized as normal, overweight, obese), and continuous risk factors (non-fasting 157 glucose, systolic and diastolic blood pressure, BMI). Covariates that will be included in models as confounders will include age, sex, educational level, SES and a measure of acculturation. We will also examine whether language with which tests were administered (English or Spanish) was associated with cognition. A baseline regression model will be run with the retinal vessel variable of interest plus age, sex, education and SES. Separate models will be run for each retinal measure stated above. The significance of the other independent variables of interest will be determined in relation to a baseline model, i.e., by evaluating a model with the baseline variables plus the specific risk factor of interest and comparing it to a model with baseline variables only with likelihood ratio tests. Formal testing of regression model assumptions will also be conducted. β coefficients (linear regression) or odds ratios (OR) (logistic regression) will be estimated from models and will be used to evaluate the direction and magnitude of associations between retinal vessel characteristics and cognitive function. Given the tests of multiple retinal vessel measures, all significance testing will occur at a two-sided alpha level of 0.01. Specific Aim 3 - To investigate the association between retinal microvascular measures and cognitive impairment and dementia (n=400). We will use unconditional binary logistic regression to evaluate whether the retinal vessel measures are associated with the presence of dementia/cognitive impairment defined by SENAS scores. Using binary logistic regression models, we will evaluate the relationship between retinal vessel measures and other risk factors and the 158 presence of dementia/cognitive impairment (compared to unimpaired subjects as a binary outcome variable). Following procedures described in 2, we will evaluate the associations of the same independent variables of interest listed above, consider the same covariates as confounders in the models, and similarly evaluate the independent significance of factors. ORs estimated from models and will be used to evaluate the direction and magnitude of associations. All significance testing will occur at a two-sided alpha level of 0.01. Specific Aim 4 - To assess whether retinal microvascular abnormalities are associated with volumetric brain measurements from magnetic resonance imaging (MRI) data (n=180). We will use multiple linear regression methods to assess the association between retinal microvascular abnormalities and volumetric brain measurements (volumes of ventricular and sulcal CSF, white matter, lacunes and hippocampus adjusted for total intracranial volume) adjusted for age, sex, education and SES. Timeline and Anticipated Difficulties Timeline • Grading of retinal photographs for CRVE and CRAE, and calculating AVR: Given an estimated two weeks to grade retinal vessel caliber from photographs for 100 subjects, we anticipate that vessel caliber data will be available in 6 months. 159 • Data cleaning and preparation for analyses: 6 months. • MRI scans will be done at Queen of the Valley MRI Center. Testing is currently in process and following a schedule in which equal numbers of dementia and control subjects are being tested. Anticipated completion is 3 years. • Data analyses pertaining to specific aims and manuscripts: 1-3 years. Manuscripts are planned as follows: one descriptive paper reporting on distribution of retinal vessel diameters and prevalence of microvascular abnormalities in the Latino population; two papers reporting on findings from analyses pertaining to associations between retinal microvascular abnormalities and cognitive function and dementia/cognitive impairment (separately); a fourth paper reporting findings from analyses of retinal microvascular abnormalities and volumetric brain measures. Anticipated difficulties One limitation of the proposed project is that we will not be able to assess the relationship between retinal vessel measures and atherosclerosis since the LALES study did not obtain a measurement of atherosclerosis (eg. carotid artery IMT). Thus we will not be able to explore the role of atherosclerosis in any associations we observe between abnormalities in the retinal vasculature and our cognitive outcomes or volumetric brain measures. An approach we will take to work with this absence of data will be to assess associations between retinal microvascular abnormalities 160 and established CVD risk factors including hypertension and diabetes, for which we will have data. Analyses described in this proposal are cross-sectional examinations of the data. Therefore, issues pertaining to timing (i.e., reverse causality) are relevant concerns for this project. However, we don’t believe there is evidence in the literature or biological plausibility to support cognitive dysfunction causing abnormalities in the retinal microvasculature. We will have an additional opportunity to address directionality of the retinal abnormalities-cognition associations in a subset of LALES subjects who are participating in the LALES follow-up (LALES-II). Approximately 2000 LALES participants for whom cognitive screening with CASI- S was obtained were recruited for LALES-II, in which a follow-up CASI-S was administered. We will therefore be able to examine longitudinal associations (i.e., change in CASI-S scores over 4 years) between retinal measures and cognitive function. LALES-II is currently in progress and expected to be completed in two years. Given the observational nature of the study, a concern would be that a third factor could be causing both cognitive dysfunction and retinal microvascular abnormalities. We are minimizing this possibility by controlling for known confounding factors such as age and education. Additionally, our thorough examination of the data and careful consideration of other possible risk factors that could be associated both with the measures of retinal vessel abnormalities and the cognitive outcomes of interest will further serve to minimize confounding. 161 Significance of proposal This proposal is very economical in that we will use existing data to address our specific aims. Furthermore, the proposed work will generate novel descriptive data on the prevalence of retinal microvascular abnormalities and the distribution of CASI-S scores in a Latino population for whom this data currently does not exist. In addressing the specific aims that involve hypothesis testing, this research will also contribute to what is currently known about associations between retinal microvascular abnormalities and cognitive function. The Latino population is the fastest growing segment of the US population, yet tends to be neglected in epidemiologic studies. None of the existing population-based epidemiologic studies addressing the relationship between retinal microvascular disease and cognition have included substantial numbers of Latino participants. We thus have the great potential to fill in gaps in our knowledge in this area of research. 162 Bibliography 1. Agatston, A.S., Janowitz, W.R., Hildner, F.J., Zusmer, N.R., Viamonte, M. and Detrano, R. (1990). "Quantification of coronary artery calcium using ultrafast computed tomography." J Am Coll Cardiol 15: 827-832. 2. AHA (2006). Heart Disease and Stroke Statistics—2006 Update. . Dallas, Texas, American Heart Association. 3. AHA. "American Heart Association Heart Attack and Angina Statistics." 2006, from http://www.americanheart.org/presenter.jhtml?identifier=4591. 4. Aleman, A., Muller, M., de Haan, E.H. and van der Schouw, Y.T. (2005). "Vascular risk factors and cognitive function in a sample of independently living men." Neurobiol Aging 26(4): 485-90. 5. APA (2000). American Psychiatric Association Diagnostic and statistical manual of mental disorders (DSM-IV-TR). Washington, DC, American Psychiatric Press, Inc. 6. Araki, A. and Sako, Y. (1987). "Determination of free and total homocysteine in human plasma by high-performance liquid chromatography with fluorescence detection." J Chromatogr 422: 43-52. 7. Auperin, A., Berr, C., Bonithon-Kopp, C., Touboul, P.J., Ruelland, I., Ducimetiere, P. and Alperovitch, A. (1996). "Ultrasonographic assessment of carotid wall characteristics and cognitive functions in a community sample of 59- to 71-year-olds. The EVA Study Group." Stroke 27(8): 1290-5. 8. Azen, S.P., Mack, W.J., Cashin-Hemphill, L., LaBree, L., Shircore, A.M., Selzer, R.H., Blankenhorn, D.H. and Hodis, H.N. (1996). "Progression of coronary artery disease predicts clinical coronary events. Long-term follow-up from the Cholesterol Lowering Atherosclerosis Study." Circulation 93(1): 34-41. 9. Bachorik, P.S. and Ross, J.W. (1995). "National Cholesterol Education Program recommendations for measurement of low-density lipoprotein cholesterol: executive summary. The National Cholesterol Education Program Working Group on Lipoprotein Measurement." Clin Chem 41(10): 1414-20. 10. Baker, M.L., Marino Larsen, E.K., Kuller, L.H., Klein, R., Klein, B.E., Siscovick, D.S., Bernick, C., Manolio, T.A. and Wong, T.Y. (2007). "Retinal microvascular signs, cognitive function, and dementia in older persons: the Cardiovascular Health Study." Stroke 38(7): 2041-7. 163 11. Bamba, V. and Rader, D.J. (2007). "Obesity and atherogenic dyslipidemia." Gastroenterology 132(6): 2181-90. 12. Berenson, G.S. (2001). "Bogalusa Heart Study: a long-term community study of a rural biracial (Black/White) population." Am J Med Sci 322(5): 293-300. 13. Biessels, G.J. and Kappelle, L.J. (2005). "Increased risk of Alzheimer's disease in Type II diabetes: insulin resistance of the brain or insulin-induced amyloid pathology?" Biochem Soc Trans 33(Pt 5): 1041-4. 14. Bischkopf, J., Busse, A. and Angermeyer, M.C. (2002). "Mild cognitive impairment--a review of prevalence, incidence and outcome according to current approaches." Acta Psychiatr Scand 106(6): 403-14. 15. Blankenhorn, D.H. and Hodis, H.N. (1994). "George Lyman Duff Memorial Lecture. Arterial imaging and atherosclerosis reversal." Arterioscler Thromb 14(2): 177-92. 16. Bonow, R.O., Smaha, L.A., Smith, S.C., Jr., Mensah, G.A. and Lenfant, C. (2002). "World Heart Day 2002: the international burden of cardiovascular disease: responding to the emerging global epidemic." Circulation 106(13): 1602-5. 17. Booze, C.F., Jr. and Staggs, C.M. (1987). "A comparison of postmortem coronary atherosclerosis findings in general aviation pilot fatalities." Aviat Space Environ Med 58(4): 297-300. 18. Bots, M.L., Baldassarre, D., Simon, A., de Groot, E., O'Leary D, H., Riley, W., Kastelein, J.J. and Grobbee, D.E. (2007). "Carotid intima-media thickness and coronary atherosclerosis: weak or strong relations?" Eur Heart J 28(4): 398-406. 19. Boushey, C.J., Beresford, S.A., Omenn, G.S. and Motulsky, A.G. (1995). "A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. Probable benefits of increasing folic acid intakes." JAMA 274(13): 1049-57. 20. Boustani, M., Peterson, B., Hanson, L., Harris, R. and Lohr, K.N. (2003). "Screening for dementia in primary care: a summary of the evidence for the U.S. Preventive Services Task Force." Ann Intern Med 138(11): 927-37. 164 21. Breteler, M.M., Claus, J.J., Grobbee, D.E. and Hofman, A. (1994). "Cardiovascular disease and distribution of cognitive function in elderly people: the Rotterdam Study." BMJ 308(6944): 1604-8. 22. Brunner, E.J. (2005). "Social and biological determinants of cognitive aging." Neurobiol Aging 26 Suppl 1: 17-20. 23. Buchwald, H., Matts, J.P., Fitch, L.L., Campos, C.T., Sanmarco, M.E., Amplatz, K., Castaneda-Zuniga, W.R., Hunter, D.W., Pearce, M.B., Bissett, J.K. and et al. (1992). "Changes in sequential coronary arteriograms and subsequent coronary events. Surgical Control of the Hyperlipidemias (POSCH) Group." JAMA 268(11): 1429-33. 24. Carr, J.J., Nelson, J.C., Wong, N.D., McNitt-Gray, M., Arad, Y., Jacobs, D.R., Jr., Sidney, S., Bild, D.E., Williams, O.D. and Detrano, R.C. (2005). "Calcified coronary artery plaque measurement with cardiac CT in population-based studies: standardized protocol of Multi-Ethnic Study of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults (CARDIA) study." Radiology 234(1): 35-43. 25. Cattell, R.B. (1966). "The scree test for the number of factors." Multivariate Behavioral Research 1: 245-276. 26. CDC. "Heart Disease Facts." 9/7/2006, from http://www.cdc.gov/DHDSP/library/fs_heart_disease.htm. Retrieved 7/24/07, 27. CDC. "Center for Disease Control and Prevention website; WISQARS Leading Causes of Death Reports, 1999 - 2003 ", from http://webapp.cdc.gov/sasweb/ncipc/leadcaus10.html. 28. CDC. "Center for Disease Control and Prevention website; WISQARS Leading Causes of Death Reports, 1999 - 2003 ", from http://webapp.cdc.gov/sasweb/ncipc/leadcaus10.html. 29. Cerhan, J.R., Folsom, A.R., Mortimer, J.A., Shahar, E., Knopman, D.S., McGovern, P.G., Hays, M.A., Crum, L.D. and Heiss, G. (1998). "Correlates of cognitive function in middle-aged adults. Atherosclerosis Risk in Communities (ARIC) Study Investigators." Gerontology 44(2): 95-105. 30. Chui, H.C., Mack, W.J., Jackson, J.E., Mungas, D., Reed, B.R., Tinklenberg, J., Chang, F.L., Skinner, K., Tasaki, C. and Jagust, W.J. (2000). "Clinical criteria for the diagnosis of vascular dementia: a multicenter study of comparability and interrater reliability." Arch Neurol 57(2): 191-6. 165 31. Chylack, L.T., Jr., Leske, M.C., McCarthy, D., Khu, P., Kashiwagi, T. and Sperduto, R. (1989). "Lens opacities classification system II (LOCS II)." Arch Ophthalmol 107(7): 991-7. 32. Conti, A.A., Macchi, C., Molino Lova, R., Conti, A. and Gensini, G.F. (2007). "Relationship between physical activity and cardiovascular disease. Selected historical highlights." J Sports Med Phys Fitness 47(1): 84-90. 33. Cooper, L.S., Wong, T.Y., Klein, R., Sharrett, A.R., Bryan, R.N., Hubbard, L.D., Couper, D.J., Heiss, G. and Sorlie, P.D. (2006). "Retinal microvascular abnormalities and MRI-defined subclinical cerebral infarction: the Atherosclerosis Risk in Communities Study." Stroke 37(1): 82-6. 34. Crowley, L.V. (2004). An Introduction to Human Disease:Pathology and Pathophysiological Correlations. Sudbury, MA, Jones and Bartlett Publishers. 35. Detrano, R., Kang, X., Mahaisavariya, P., Tang, W., Colombo, A., Molloi, S., Garner, D. and Nickerson, S. (1994). "Accuracy of quantifying coronary hydroxyapatite with electron beam tomography." Invest Radiol 29(8): 733-8. 36. Dik, M.G., Jonker, C., Comijs, H.C., Deeg, D.J., Kok, A., Yaffe, K. and Penninx, B.W. (2007). "Contribution of Metabolic Syndrome Components to Cognition in Older Persons." Diabetes Care. 37. Dowling, J.E. (1987). The Retina: An Approachable Part of the Brain. Cambridge, MA, The Belknap Press of Harvard University. 38. Elias, M.F., Elias, P.K., Robbins, M.A. and et al. (2001). Cardiovascular risk factors and cognitive functioning: an epidemiologic perspective. Neuropsychology of cardiovascular disease. S. R. Waldstein and M. F. Elias. Mahwah, NJ, Lawrence Erlbaum Associates, Inc.: 83-104. 39. Elias, M.F., Elias, P.K., Sullivan, L.M., Wolf, P.A. and D'Agostino, R.B. (2005). "Obesity, diabetes and cognitive deficit: The Framingham Heart Study." Neurobiol Aging 26 Suppl 1: 11-6. 40. Elias, M.F., Elias, P.K. and Wolf, P.A. (2000). Comparative effects of age and blood pressure on neuropsychological test performance: The Framingham Study. Behavior, health and aging. S. B. Manuck, R. Jennings, B. S. Rabin and A. Baum. Mahwah, NJ, Lawrence Erlbaum Associates: 199-223. 41. Elias, M.F. and Robbins, M.A. (1991). Cardiovascular Disease, Hypertension and Cognitive Function. Behavioral Aspects of Cardiovascular Disease. A. P. Shapiro and A. Baum. Hillsdale, NJ, Lawrence Erlbaum. 166 42. Elias, M.F., Wolf, P.A., D'Agostino, R.B., Cobb, J. and White, L.R. (1993). "Untreated blood pressure level is inversely related to cognitive functioning: the Framingham Study." Am J Epidemiol 138(6): 353-64. 43. Elwood, R.W. (1995). "The California Verbal Learning Test: psychometric characteristics and clinical application." Neuropsychol Rev 5(3): 173-201. 44. Enos, W.F., Holmes, R.H. and Beyer, J. (1953). "Coronary disease among United States soldiers killed in action in Korea; preliminary report." J Am Med Assoc 152(12): 1090-3. 45. Farias, S.T., Mungas, D. and Jagust, W. (2005). "Degree of discrepancy between self and other-reported everyday functioning by cognitive status: dementia, mild cognitive impairment, and healthy elders." Int J Geriatr Psychiatry 20(9): 827-34. 46. Feher, E.P., Mahurin, R.K., Doody, R.S., Cooke, N., Sims, J. and Pirozzolo, F.J. (1992). "Establishing the limits of the Mini-Mental State. Examination of 'subtests'." Arch Neurol 49(1): 87-92. 47. Friedewald, W.T., Levy, R.I. and Fredrickson, D.S. (1972). "Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge." Clin Chem 18(6): 499-502. 48. Fujiwara, Y., Chaves, P.H., Takahashi, R., Amano, H., Yoshida, H., Kumagai, S., Fujita, K., Wang, D.G. and Shinkai, S. (2005). "Arterial pulse wave velocity as a marker of poor cognitive function in an elderly community- dwelling population." J Gerontol A Biol Sci Med Sci 60(5): 607-12. 49. Fukuhara, M., Matsumura, K., Ansai, T., Takata, Y., Sonoki, K., Akifusa, S., Wakisaka, M., Hamasaki, T., Fujisawa, K., Yoshida, A., Fujii, K., Iida, M. and Takehara, T. (2006). "Prediction of cognitive function by arterial stiffness in the very elderly." Circ J 70(6): 756-61. 50. Gatto, N.M., Henderson, V.W., St. John, J.A., McCleary, C., Detrano, R., Hodis, H.N. and Mack, W.J. (2007). "Subclinical Atherosclerosis Contributes to Differences in Cognitive Function in Healthy Adults without Cardiovascular Disease." Stroke (submitted). 51. Gatto, N.M., Henderson, V.W., St. John, J.A., McCleary, C., Hodis, H.N. and Mack, W.J. (2007). "Metabolic Syndrome and Cognitive Function in Healthy Middle-Aged and Older Adults without Diabetes." Annals of Neurology (submitted). 167 52. Geroldi, C., Frisoni, G.B., Paolisso, G., Bandinelli, S., Lamponi, M., Abbatecola, A.M., Zanetti, O., Guralnik, J.M. and Ferrucci, L. (2005). "Insulin resistance in cognitive impairment: the InCHIANTI study." Arch Neurol 62(7): 1067- 72. 53. Golomb, J., de Leon, M.J., Kluger, A., George, A.E., Tarshish, C. and Ferris, S.H. (1993). "Hippocampal atrophy in normal aging. An association with recent memory impairment." Arch Neurol 50(9): 967-73. 54. Goto, I., Katsuki, S., Ikui, H., Kimoto, K. and Mimatsu, T. (1975). "Pathological studies on the intracerebral and retinal arteries in cerebrovascular and noncerebrovascular diseases." Stroke 6(3): 263-9. 55. Grundy, S.M., Cleeman, J.I., Daniels, S.R., Donato, K.A., Eckel, R.H., Franklin, B.A., Gordon, D.J., Krauss, R.M., Savage, P.J., Smith, S.C., Jr., Spertus, J.A. and Costa, F. (2005). "Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement." Circulation 112(17): 2735-52. 56. Haan, M.N., Miller, J.W., Aiello, A.E., Whitmer, R.A., Jagust, W.J., Mungas, D.M., Allen, L.H. and Green, R. (2007). "Homocysteine, B vitamins, and the incidence of dementia and cognitive impairment: results from the Sacramento Area Latino Study on Aging." Am J Clin Nutr 85(2): 511-7. 57. Hanon, O., Haulon, S., Lenoir, H., Seux, M.L., Rigaud, A.S., Safar, M., Girerd, X. and Forette, F. (2005). "Relationship between arterial stiffness and cognitive function in elderly subjects with complaints of memory loss." Stroke 36(10): 2193-7. 58. Hashizume, K., Suzuki, S., Hara, M., Komatsu, A. and Yamashita, K. (2006). "Metabolic syndrome and age-related dementia: endocrinological aspects of adaptation to aging." Mech Ageing Dev 127(6): 507-10. 59. Hodis, H.N., Mack, W.J., Lobo, R.A., Shoupe, D., Sevanian, A., Mahrer, P.R., Selzer, R.H., Liu Cr, C.R., Liu Ch, C.H. and Azen, S.P. (2001). "Estrogen in the prevention of atherosclerosis. A randomized, double-blind, placebo- controlled trial." Ann Intern Med 135(11): 939-53. 60. Hof, P.R., Trapp, B.D., de Vellis, J., Claudio, L. and Colman, D.R. (1999). The Cellular Components of Nervous Tissue. Fundamental Neuroscience. M. J. Zigmond, F. E. Bloom, S. C. Landis, J. L. Roberts and L. R. Squire. San Diego, Academic Press: 41-70. 168 61. Hofman, A., Ott, A., Breteler, M.M., Bots, M.L., Slooter, A.J., van Harskamp, F., van Duijn, C.N., Van Broeckhoven, C. and Grobbee, D.E. (1997). "Atherosclerosis, apolipoprotein E, and prevalence of dementia and Alzheimer's disease in the Rotterdam Study." Lancet 349(9046): 151-4. 62. Hubbard, L.D., Brothers, R.J., King, W.N., Clegg, L.X., Klein, R., Cooper, L.S., Sharrett, A.R., Davis, M.D. and Cai, J. (1999). "Methods for evaluation of retinal microvascular abnormalities associated with hypertension/sclerosis in the Atherosclerosis Risk in Communities Study." Ophthalmology 106(12): 2269-80. 63. Ikram, M.K., De Jong, F.J., Van Dijk, E.J., Prins, N.D., Hofman, A., Breteler, M.M. and De Jong, P.T. (2006). "Retinal vessel diameters and cerebral small vessel disease: the Rotterdam Scan Study." Brain 129(Pt 1): 182-8. 64. Inzitari, D., Erkinjuntti, T., Wallin, A., Del Ser, T., Romanelli, M. and Pantoni, L. (2000). "Subcortical vascular dementia as a specific target for clinical trials." Ann N Y Acad Sci 903: 510-21. 65. Johnson, S.C., Saykin, A.J., Flashman, L.A., McAllister, T.W. and Sparling, M.B. (2001). "Brain activation on fMRI and verbal memory ability: functional neuroanatomic correlates of CVLT performance." J Int Neuropsychol Soc 7(1): 55-62. 66. Johnston, S.C., O'Meara, E.S., Manolio, T.A., Lefkowitz, D., O'Leary, D.H., Goldstein, S., Carlson, M.C., Fried, L.P. and Longstreth, W.T., Jr. (2004). "Cognitive impairment and decline are associated with carotid artery disease in patients without clinically evident cerebrovascular disease." Ann Intern Med 140(4): 237-47. 67. Kalmijn, S., Foley, D., White, L., Burchfiel, C.M., Curb, J.D., Petrovitch, H., Ross, G.W., Havlik, R.J. and Launer, L.J. (2000). "Metabolic cardiovascular syndrome and risk of dementia in Japanese-American elderly men. The Honolulu-Asia aging study." Arterioscler Thromb Vasc Biol 20(10): 2255- 60. 68. Kalmijn, S., van Boxtel, M.P., Verschuren, M.W., Jolles, J. and Launer, L.J. (2002). "Cigarette smoking and alcohol consumption in relation to cognitive performance in middle age." Am J Epidemiol 156(10): 936-44. 69. Kandel, E.R., Schwartz, J.H. and Jessell, T.M. (2000). Principles of Neural Science. New York, McGraw-Hill. 169 70. Kelman, H.R., Thomas, C., Kennedy, G.J. and Cheng, J. (1994). "Cognitive impairment and mortality in older community residents." Am J Public Health 84(8): 1255-60. 71. Klein, R., Klein, B.E., Knudtson, M.D., Wong, T.Y. and Tsai, M.Y. (2006). "Are inflammatory factors related to retinal vessel caliber? The Beaver Dam Eye Study." Arch Ophthalmol 124(1): 87-94. 72. Klein, R., Klein, B.E. and Moss, S.E. (1997). "The relation of systemic hypertension to changes in the retinal vasculature: the Beaver Dam Eye Study." Trans Am Ophthalmol Soc 95: 329-48; discussion 348-50. 73. Klein, R., Klein, B.E., Moss, S.E. and Wang, Q. (1993). "Blood pressure, hypertension and retinopathy in a population." Trans Am Ophthalmol Soc 91: 207-22; discussion 222-6. 74. Klein, R., Sharrett, A.R., Klein, B.E., Chambless, L.E., Cooper, L.S., Hubbard, L.D. and Evans, G. (2000). "Are retinal arteriolar abnormalities related to atherosclerosis?: The Atherosclerosis Risk in Communities Study." Arterioscler Thromb Vasc Biol 20(6): 1644-50. 75. Knopman, D., Boland, L.L., Mosley, T., Howard, G., Liao, D., Szklo, M., McGovern, P. and Folsom, A.R. (2001). "Cardiovascular risk factors and cognitive decline in middle-aged adults." Neurology 56(1): 42-8. 76. Knudtson, M.D., Lee, K.E., Hubbard, L.D., Wong, T.Y., Klein, R. and Klein, B.E. (2003). "Revised formulas for summarizing retinal vessel diameters." Curr Eye Res 27(3): 143-9. 77. Kobayashi, K., Akishita, M., Yu, W., Hashimoto, M., Ohni, M. and Toba, K. (2004). "Interrelationship between non-invasive measurements of atherosclerosis: flow-mediated dilation of brachial artery, carotid intima- media thickness and pulse wave velocity." Atherosclerosis 173(1): 13-8. 78. Komulainen, P., Lakka, T.A., Kivipelto, M., Hassinen, M., Helkala, E.L., Haapala, I., Nissinen, A. and Rauramaa, R. (2007). "Metabolic syndrome and cognitive function: a population-based follow-up study in elderly women." Dement Geriatr Cogn Disord 23(1): 29-34. 79. Kukull, W.A., Larson, E.B., Teri, L., Bowen, J., McCormick, W. and Pfanschmidt, M.L. (1994). "The Mini-Mental State Examination score and the clinical diagnosis of dementia." J Clin Epidemiol 47(9): 1061-7. 80. Kuller, L.H. (2004). "Ethnic differences in atherosclerosis, cardiovascular disease and lipid metabolism." Curr Opin Lipidol 15(2): 109-13. 170 81. Kumari, M., Brunner, E. and Fuhrer, R. (2000). "Minireview: mechanisms by which the metabolic syndrome and diabetes impair memory." J Gerontol A Biol Sci Med Sci 55(5): B228-32. 82. Kumari, M. and Marmot, M. (2005). "Diabetes and cognitive function in a middle-aged cohort: findings from the Whitehall II study." Neurology 65(10): 1597-603. 83. Kuusisto, J., Koivisto, K., Mykkanen, L., Helkala, E.L., Vanhanen, M., Hanninen, T., Kervinen, K., Kesaniemi, Y.A., Riekkinen, P.J. and Laakso, M. (1997). "Association between features of the insulin resistance syndrome and Alzheimer's disease independently of apolipoprotein E4 phenotype: cross sectional population based study." BMJ 315(7115): 1045-9. 84. Kwa, V.I., van der Sande, J.J., Stam, J., Tijmes, N. and Vrooland, J.L. (2002). "Retinal arterial changes correlate with cerebral small-vessel disease." Neurology 59(10): 1536-40. 85. Lakka, H.M., Laaksonen, D.E., Lakka, T.A., Niskanen, L.K., Kumpusalo, E., Tuomilehto, J. and Salonen, J.T. (2002). "The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men." Jama 288(21): 2709-16. 86. Larrea, F.A., Fisk, J.D., Graham, J.E. and Stadnyk, K. (2000). "Prevalence of cognitive impairment and dementia as defined by neuropsychological test performance." Neuroepidemiology 19(3): 121-9. 87. Launer, L.J. (2002). "Demonstrating the case that AD is a vascular disease: epidemiologic evidence." Ageing Res Rev 1(1): 61-77. 88. Launer, L.J., Masaki, K., Petrovitch, H., Foley, D. and Havlik, R.J. (1995). "The association between midlife blood pressure levels and late-life cognitive function. The Honolulu-Asia Aging Study." JAMA 274(23): 1846-51. 89. Leibowitz, H.M., Krueger, D.E., Maunder, L.R., Milton, R.C., Kini, M.M., Kahn, H.A., Nickerson, R.J., Pool, J., Colton, T.L., Ganley, J.P., Loewenstein, J.I. and Dawber, T.R. (1980). "The Framingham Eye Study monograph: An ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a general population of 2631 adults, 1973-1975." Surv Ophthalmol 24(Suppl): 335-610. 90. Lezak, M.D., Howieson, D.B. and Loring, D.W. (2004). Neuropsychological Assessment. New York, Oxford University Press. 171 91. Libby, P. (2001). "Current concepts of the pathogenesis of the acute coronary syndromes." Circulation 104(3): 365-72. 92. Lipid Research Clinics Program. The Manual of Laboratory Operations: Lipid and Lipoprotein Analysis. (May 1974). Bethesda, MD National Heart and Lung Institute. 1. 93. Longstreth, W., Jr., Larsen, E.K., Klein, R., Wong, T.Y., Sharrett, A.R., Lefkowitz, D. and Manolio, T.A. (2007). "Associations between findings on cranial magnetic resonance imaging and retinal photography in the elderly: the Cardiovascular Health Study." Am J Epidemiol 165(1): 78-84. 94. Luma, G.B. and Spiotta, R.T. (2006). "Hypertension in children and adolescents." Am Fam Physician 73(9): 1558-68. 95. Mahaisavariya, P., Detrano, R., Kang, X., Garner, D., Vo, A., Georgiou, D., Molloi, S. and Brundage, B.H. (1994). "Quantitation of in vitro coronary artery calcium using ultrafast computed tomography." Cathet Cardiovasc Diagn 32(4): 387-93. 96. Malinow, M.R., Bostom, A.G. and Krauss, R.M. (1999). "Homocyst(e)ine, diet, and cardiovascular diseases: a statement for healthcare professionals from the Nutrition Committee, American Heart Association." Circulation 99(1): 178- 82. 97. Markides, K.S. and Coreil, J. (1986). "The health of Hispanics in the southwestern United States: an epidemiologic paradox." Public Health Rep 101(3): 253-65. 98. Marmot, M. (2000). Multilevel Approaches to Understanding Social Determinants. Social Epidemiology. L. Berkman and I. Kawachi. Oxford, Oxford University Press: 349-367. 99. Martins, I.J., Hone, E., Foster, J.K., Sunram-Lea, S.I., Gnjec, A., Fuller, S.J., Nolan, D., Gandy, S.E. and Martins, R.N. (2006). "Apolipoprotein E, cholesterol metabolism, diabetes, and the convergence of risk factors for Alzheimer's disease and cardiovascular disease." Mol Psychiatry 11(8): 721- 36. 100. McCarron, R.M., Chen, Y., Tomori, T., Strasser, A., Mechoulam, R., Shohami, E. and Spatz, M. (2006). "Endothelial-mediated regulation of cerebral microcirculation." J Physiol Pharmacol 57 Suppl 11: 133-44. 172 101. Molloy, D.W., Alemayehu, E. and Roberts, R. (1991). "Reliability of a Standardized Mini-Mental State Examination compared with the traditional Mini-Mental State Examination." Am J Psychiatry 148(1): 102-5. 102. Moser, D.J., Hoth, K.F., Robinson, R.G., Paulsen, J.S., Sinkey, C.A., Benjamin, M.L., Schultz, S.K. and Haynes, W.G. (2004). "Blood vessel function and cognition in elderly patients with atherosclerosis." Stroke 35(11): e369-72. 103. Muller, M., Grobbee, D.E., Aleman, A., Bots, M. and van der Schouw, Y.T. (2006). "Cardiovascular disease and cognitive performance in middle-aged and elderly men." Atherosclerosis. 104. Muller, M., Grobbee, D.E., Aleman, A., Bots, M. and van der Schouw, Y.T. (2007). "Cardiovascular disease and cognitive performance in middle-aged and elderly men." Atherosclerosis 190(1): 143-9. 105. Mungas, D., Reed, B.R., Marshall, S.C. and Gonzalez, H.M. (2000). "Development of psychometrically matched English and Spanish language neuropsychological tests for older persons." Neuropsychology 14(2): 209-23. 106. Murray, C.D. (1926). "The Physiological Principle of Minimum Work: I. The Vascular System and the Cost of Blood Volume." Proc Natl Acad Sci U S A 12(3): 207-14. 107. Murray, C.D. (1926). "The Physiological Principle of Minimum Work: II. Oxygen Exchange in Capillaries." Proc Natl Acad Sci U S A 12(5): 299-304. 108. NCEP (2001). "Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III)." JAMA 285(19): 2486-97. 109. NCHS. "National Center for Health Statistics at the Centers for Disease Control and Prevention National Health Interview Survey." from http://www.cdc.gov/nchs/nhis.htm. 110. NDEP. "The Diabetes Epidemic Among Hispanic and Latino Americans." November 2005, from http://ndep.nih.gov/diabetes/pubs/FS_HispLatino_Eng.pdf. Retrieved July 2007, 111. NHLBI. "National Heart, Lung and Blood Institute Direct and Indirect Costs of Major Illnesses, U.S., 2006." from http://www.nhlbi.nih.gov/about/factbook/chapter4.htm. 173 112. NIH. "Cognitive and Emotional Health Project: The Healthy Brain." from http://trans.nih.gov/cehp/ReviewDocs.htm. Retrieved 7/24/07, 113. Norman, M.A., Evans, J.D., Miller, W.S. and Heaton, R.K. (2000). "Demographically corrected norms for the California Verbal Learning Test." J Clin Exp Neuropsychol 22(1): 80-94. 114. O'Brien, J.T. (2006). "Vascular cognitive impairment." Am J Geriatr Psychiatry 14(9): 724-33. 115. Pantoni, L. and Garcia, J.H. (1997). "Cognitive impairment and cellular/vascular changes in the cerebral white matter." Ann N Y Acad Sci 826: 92-102. 116. Pantoni, L. and Garcia, J.H. (1997). "Pathogenesis of leukoaraiosis: a review." Stroke 28(3): 652-9. 117. Paoletti, R., Bolego, C., Poli, A. and Cignarella, A. (2006). "Metabolic syndrome, inflammation and atherosclerosis." Vasc Health Risk Manag 2(2): 145-52. 118. Paolo, A.M., Troster, A.I. and Ryan, J.J. (1997). "California Verbal Learning Test: normative data for the elderly." J Clin Exp Neuropsychol 19(2): 220-34. 119. Patton, N., Aslam, T., Macgillivray, T., Pattie, A., Deary, I.J. and Dhillon, B. (2005). "Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: a rationale based on homology between cerebral and retinal microvasculatures." J Anat 206(4): 319-48. 120. Patton, N., Aslam, T.M., MacGillivray, T., Deary, I.J., Dhillon, B., Eikelboom, R.H., Yogesan, K. and Constable, I.J. (2006). "Retinal image analysis: concepts, applications and potential." Prog Retin Eye Res 25(1): 99-127. 121. Patton, N., Pattie, A., MacGillivray, T., Aslam, T., Dhillon, B., Gow, A., Starr, J.M., Whalley, L.J. and Deary, I.J. (2007). "The association between retinal vascular network geometry and cognitive ability in an elderly population." Invest Ophthalmol Vis Sci 48(5): 1995-2000. 122. Pavlik, V.N., Hyman, D.J. and Doody, R. (2005). "Cardiovascular risk factors and cognitive function in adults 30-59 years of age (NHANES III)." Neuroepidemiology 24(1-2): 42-50. 123. Pedersen, H.S., Mulvad, G., Newman, W.P., 3rd and Boudreau, D.A. (2003). "Atherosclerosis in coronary arteries and aorta among Greenlanders: an autopsy study." Atherosclerosis 170(1): 93-103. 174 124. Petersen, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tangalos, E.G. and Kokmen, E. (1999). "Mild cognitive impairment: clinical characterization and outcome." Arch Neurol 56(3): 303-8. 125. Price, J.F., McDowell, S., Whiteman, M.C., Deary, I.J., Stewart, M.C. and Fowkes, F.G. (2006). "Ankle brachial index as a predictor of cognitive impairment in the general population: ten-year follow-up of the Edinburgh Artery Study." J Am Geriatr Soc 54(5): 763-9. 126. Pruissen, D.M., Gerritsen, S.A., Prinsen, T.J., Dijk, J.M., Kappelle, L.J. and Algra, A. (2007). "Carotid intima-media thickness is different in large- and small-vessel ischemic stroke: the SMART study." Stroke 38(4): 1371-3. 127. Radloff, L.S. (1977). "The CES-D scale: A self-report depression scale for research in the general population. ." Applied Psychological Measurement 1: 385-401. 128. Raz, N. and Rodrigue, K.M. (2006). "Differential aging of the brain: patterns, cognitive correlates and modifiers." Neurosci Biobehav Rev 30(6): 730-48. 129. Reaven, G.M. (1988). "Banting lecture 1988. Role of insulin resistance in human disease." Diabetes 37(12): 1595-607. 130. Reed, J.E., Rumberger, J.A. and Davitt, P.J. (1994). "System for quantitative analysis of coronary calcification via electron beam computed tomography." Proc SPIE 2168: 36-43. 131. Refsum, H., Ueland, P.M., Nygard, O. and Vollset, S.E. (1998). "Homocysteine and cardiovascular disease." Annu Rev Med 49: 31-62. 132. Riggs, K.M., Spiro, A., 3rd, Tucker, K. and Rush, D. (1996). "Relations of vitamin B-12, vitamin B-6, folate, and homocysteine to cognitive performance in the Normative Aging Study." Am J Clin Nutr 63(3): 306-14. 133. Robbins, M.A., Elias, M.F., Budge, M.M., Brennan, S.L. and Elias, P.K. (2005). "Homocysteine, type 2 diabetes mellitus, and cognitive performance: The Maine-Syracuse Study." Clin Chem Lab Med 43(10): 1101-6. 134. Robbins, M.A., Elias, M.F., Elias, P.K. and Budge, M.M. (2005). "Blood pressure and cognitive function in an African-American and a Caucasian- American sample: the Maine-Syracuse Study." Psychosom Med 67(5): 707- 14. 175 135. Roger, V.L., Weston, S.A., Killian, J.M., Pfeifer, E.A., Belau, P.G., Kottke, T.E., Frye, R.L., Bailey, K.R. and Jacobsen, S.J. (2001). "Time trends in the prevalence of atherosclerosis: a population-based autopsy study." Am J Med 110(4): 267-73. 136. Roman, G., Tatemichi, T., Erkinjuntti, T., Cummings, J., Masdeu, J., Garcia, J., Amaducci, L., Orgogozo, J., Brun, A., Hofman, A.e.a. and (1993 ). "Vascular dementia: diagnostic criteria for research studies. Report of the NINDS- AIREN International Workshop." Neurology 43(2): 250-60. 137. Roman, G.C. (2004). "Brain hypoperfusion: a critical factor in vascular dementia." Neurol Res 26(5): 454-8. 138. Rosamond, W., Flegal, K., Friday, G., Furie, K., Go, A., Greenlund, K., Haase, N., Ho, M., Howard, V., Kissela, B., Kittner, S., Lloyd-Jones, D., McDermott, M., Meigs, J., Moy, C., Nichol, G., O'Donnell, C.J., Roger, V., Rumsfeld, J., Sorlie, P., Steinberger, J., Thom, T., Wasserthiel-Smoller, S. and Hong, Y. (2007). "Heart disease and stroke statistics--2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee." Circulation 115(5): e69-171. 139. Rosano, C., Naydeck, B., Kuller, L.H., Longstreth, W.T., Jr. and Newman, A.B. (2005). "Coronary artery calcium: associations with brain magnetic resonance imaging abnormalities and cognitive status." J Am Geriatr Soc 53(4): 609-15. 140. Ruitenberg, A., den Heijer, T., Bakker, S.L., van Swieten, J.C., Koudstaal, P.J., Hofman, A. and Breteler, M.M. (2005). "Cerebral hypoperfusion and clinical onset of dementia: the Rotterdam Study." Ann Neurol 57(6): 789-94. 141. Saine, P.J. and Tyler, M.E. (2002 ). Ophthalmic Photography: Retinal Photography, Angiography, and Electronic Imaging. Boston, Butterworth- Heinemann. 142. Saxton, J., Ratcliff, G., Newman, A., Belle, S., Fried, L., Yee, J. and Kuller, L. (2000). "Cognitive test performance and presence of subclinical cardiovascular disease in the cardiovascular health study." Neuroepidemiology 19(6): 312-9. 143. Scuteri, A., Brancati, A.M., Gianni, W., Assisi, A. and Volpe, M. (2005). "Arterial stiffness is an independent risk factor for cognitive impairment in the elderly: a pilot study." J Hypertens 23(6): 1211-6. 176 144. Seddon, J.M., Sahagian, C.R., Glynn, R.J., Sperduto, R.D. and Gragoudas, E.S. (1990). "Evaluation of an iris color classification system. The Eye Disorders Case-Control Study Group." Invest Ophthalmol Vis Sci 31(8): 1592-8. 145. Selzer, R.H., Hodis, H.N., Kwong-Fu, H., Mack, W.J., Lee, P.L., Liu, C.R. and Liu, C.H. (1994). "Evaluation of computerized edge tracking for quantifying intima-media thickness of the common carotid artery from B-mode ultrasound images." Atherosclerosis 111(1): 1-11. 146. Selzer, R.H., Mack, W.J., Lee, P.L., Kwong-Fu, H. and Hodis, H.N. (2001). "Improved common carotid elasticity and intima-media thickness measurements from computer analysis of sequential ultrasound frames." Atherosclerosis 154(1): 185-93. 147. Sherman, T.F. (1981). "On connecting large vessels to small. The meaning of Murray's law." J Gen Physiol 78(4): 431-53. 148. Singh-Manoux, A. and Marmot, M. (2005). "High blood pressure was associated with cognitive function in middle-age in the Whitehall II study." J Clin Epidemiol 58(12): 1308-15. 149. Skoog, I., Wallin, A., Fredman, P., Hesse, C., Aevarsson, O., Karlsson, I., Gottfries, C.G. and Blennow, K. (1998). "A population study on blood-brain barrier function in 85-year-olds: relation to Alzheimer's disease and vascular dementia." Neurology 50(4): 966-71. 150. Smith, A. (1982). Symbol Digit Modalities Test (SDMT). Manual (revised). Los Angeles, Western Psychological Services. 151. Strachan, M.W., Deary, I.J., Ewing, F.M. and Frier, B.M. (1997). "Is type II diabetes associated with an increased risk of cognitive dysfunction? A critical review of published studies." Diabetes Care 20(3): 438-45. 152. Strong, J.P., Malcom, G.T., McMahan, C.A., Tracy, R.E., Newman, W.P., 3rd, Herderick, E.E. and Cornhill, J.F. (1999). "Prevalence and extent of atherosclerosis in adolescents and young adults: implications for prevention from the Pathobiological Determinants of Atherosclerosis in Youth Study." JAMA 281(8): 727-35. 153. Sviridov, D. and Nestel, P.J. (2007). "Genetic factors affecting HDL levels, structure, metabolism and function." Curr Opin Lipidol 18(2): 157-63. 154. Swan, G.E., Carmelli, D. and Larue, A. (1998). "Systolic blood pressure tracking over 25 to 30 years and cognitive performance in older adults." Stroke 29(11): 2334-40. 177 155. Swan, G.E., LaRue, A., Carmelli, D., Reed, T.E. and Fabsitz, R.R. (1992). "Decline in cognitive performance in aging twins. Heritability and biobehavioral predictors from the National Heart, Lung, and Blood Institute Twin Study." Arch Neurol 49(5): 476-81. 156. Takei, H., Strong, J.P., Yutani, C. and Malcom, G.T. (2005). "Comparison of coronary and aortic atherosclerosis in youth from Japan and the USA." Atherosclerosis 180(1): 171-9. 157. Teng, E.L., Hasegawa, K., Homma, A., Imai, Y., Larson, E., Graves, A., Sugimoto, K., Yamaguchi, T., Sasaki, H., Chiu, D. and et al. (1994). "The Cognitive Abilities Screening Instrument (CASI): a practical test for cross- cultural epidemiological studies of dementia." Int Psychogeriatr 6(1): 45-58; discussion 62. 158. Teng, E.L., Larson, E.B., Lin, K.N., Graves, A.B. and Liu, H.C. (1998). Screening for Dementia: The Cognitive Abilities Screening Instrument-Short Version (CASI-Short). American Psychological Association. . San Francisco, CA. 159. Tervo, S., Kivipelto, M., Hanninen, T., Vanhanen, M., Hallikainen, M., Mannermaa, A. and Soininen, H. (2004). "Incidence and risk factors for mild cognitive impairment: a population-based three-year follow-up study of cognitively healthy elderly subjects." Dement Geriatr Cogn Disord 17(3): 196-203. 160. Tombaugh, T.N. and McIntyre, N.J. (1992). "The mini-mental state examination: a comprehensive review." J Am Geriatr Soc 40(9): 922-35. 161. Tso, M.O. and Jampol, L.M. (1982). "Pathophysiology of hypertensive retinopathy." Ophthalmology 89(10): 1132-45. 162. U.S. Bureau of the Census. "Current Population Survey." 12/28/2006, from http://www.census.gov/cps/. Retrieved 7/24/07, 163. U.S. DHHS Administration on Aging. "Statistics on the Aging Population." 7/12/07, from http://www.aoa.gov/prof/statistics/statistics.asp. Retrieved 7/20/07, 164. US Bureau of the Census (1993). We the American... Hispanics, U.S. Government Printing Office, Washington DC. 165. van Exel, E., Gussekloo, J., Houx, P., de Craen, A.J., Macfarlane, P.W., Bootsma-van der Wiel, A., Blauw, G.J. and Westendorp, R.G. (2002). "Atherosclerosis and cognitive impairment are linked in the elderly. The Leiden 85-plus Study." Atherosclerosis 165(2): 353-9. 178 166. Vanhanen, M., Koivisto, K., Moilanen, L., Helkala, E.L., Hanninen, T., Soininen, H., Kervinen, K., Kesaniemi, Y.A., Laakso, M. and Kuusisto, J. (2006). "Association of metabolic syndrome with Alzheimer disease: a population-based study." Neurology 67(5): 843-7. 167. Vigen, C., Hodis, H.N., Selzer, R.H., Mahrer, P.R. and Mack, W.J. (2005). "Relation of progression of coronary artery atherosclerosis to risk of cardiovascular events (from the Monitored Atherosclerosis Regression Study)." Am J Cardiol 95(11): 1277-82. 168. Vinereanu, D. (2006). "Risk factors for atherosclerotic disease: present and future." Herz 31 Suppl 3: 5-24. 169. von Eckardstein, A. (2005). "Risk factors for atherosclerotic vascular disease." Handb Exp Pharmacol(170): 71-105. 170. Waters, D., Craven, T.E. and Lesperance, J. (1993). "Prognostic significance of progression of coronary atherosclerosis." Circulation 87(4): 1067-75. 171. Wechsler, D. (1997). Wechsler Adult Intelligence Scale-III. San Antonio, The Psychological Corporation. 172. Wendelhag, I., Gustavsson, T., Suurkula, M., Berglund, G. and Wikstrand, J. (1991). "Ultrasound measurement of wall thickness in the carotid artery: fundamental principles and description of a computerized analysing system." Clin Physiol 11(6): 565-77. 173. Wikstrand, J. and Wiklund, O. (1992). "Frontiers in cardiovascular science. Quantitative measurements of atherosclerotic manifestations in humans." Arterioscler Thromb 12(1): 114-9. 174. Wong, T.Y. (2003). "Cerebral white matter lesions, retinopathy, and stroke." Stroke 34(11): e212. 175. Wong, T.Y., Hubbard, L.D., Klein, R., Marino, E.K., Kronmal, R., Sharrett, A.R., Siscovick, D.S., Burke, G. and Tielsch, J.M. (2002). "Retinal microvascular abnormalities and blood pressure in older people: the Cardiovascular Health Study." Br J Ophthalmol 86(9): 1007-13. 176. Wong, T.Y., Klein, R., Klein, B.E., Tielsch, J.M., Hubbard, L. and Nieto, F.J. (2001). "Retinal microvascular abnormalities and their relationship with hypertension, cardiovascular disease, and mortality." Surv Ophthalmol 46(1): 59-80. 179 177. Wong, T.Y., Klein, R., Sharrett, A.R., Couper, D.J., Klein, B.E., Liao, D.P., Hubbard, L.D. and Mosley, T.H. (2002). "Cerebral white matter lesions, retinopathy, and incident clinical stroke." JAMA 288(1): 67-74. 178. Wong, T.Y., Klein, R., Sharrett, A.R., Manolio, T.A., Hubbard, L.D., Marino, E.K., Kuller, L., Burke, G., Tracy, R.P., Polak, J.F., Gottdiener, J.S. and Siscovick, D.S. (2003). "The prevalence and risk factors of retinal microvascular abnormalities in older persons: The Cardiovascular Health Study." Ophthalmology 110(4): 658-66. 179. Wong, T.Y., Klein, R., Sharrett, A.R., Nieto, F.J., Boland, L.L., Couper, D.J., Mosley, T.H., Klein, B.E., Hubbard, L.D. and Szklo, M. (2002). "Retinal microvascular abnormalities and cognitive impairment in middle-aged persons: the Atherosclerosis Risk in Communities Study." Stroke 33(6): 1487-92. 180. Woo, J., Lynn, H., Wong, S.Y., Hong, A., Tang, Y.N., Lau, W.Y., Lau, E., Orwoll, E. and Kwok, T.C. (2006). "Correlates for a low ankle-brachial index in elderly Chinese." Atherosclerosis 186(2): 360-6. 181. Wright, C.B., Elkind, M.S., Rundek, T., Boden-Albala, B., Paik, M.C. and Sacco, R.L. (2006). "Alcohol intake, carotid plaque, and cognition: the Northern Manhattan Study." Stroke 37(5): 1160-4. 182. Xiong, G.L., Plassman, B.L., Helms, M.J. and Steffens, D.C. (2006). "Vascular risk factors and cognitive decline among elderly male twins." Neurology 67(9): 1586-91. 183. Yaffe, K., Haan, M., Blackwell, T., Cherkasova, E., Whitmer, R.A. and West, N. (2007). "Metabolic syndrome and cognitive decline in elderly Latinos: findings from the Sacramento Area Latino Study of Aging study." J Am Geriatr Soc 55(5): 758-62. 184. Yaffe, K., Kanaya, A., Lindquist, K., Simonsick, E.M., Harris, T., Shorr, R.I., Tylavsky, F.A. and Newman, A.B. (2004). "The metabolic syndrome, inflammation, and risk of cognitive decline." Jama 292(18): 2237-42. 185. Yaghoubi, S., Tang, W., Wang, S., Reed, J., Hsiai, J., Detrano, R. and Brundage, B. (1995). "Offline assessment of atherosclerotic coronary calcium from electron beam tomograms." Am J Card Imaging 9(4): 231-6. 186. Yu, T., Mitchell, P., Berry, G., Li, W. and Wang, J.J. (1998). "Retinopathy in older persons without diabetes and its relationship to hypertension." Arch Ophthalmol 116(1): 83-9. 180 187. Zamir, M. (1976). "Optimality principles in arterial branching." J Theor Biol 62(1): 227-51. 188. Zamir, M. and Medeiros, J.A. (1982). "Arterial branching in man and monkey." J Gen Physiol 79(3): 353-60. 189. Zamir, M., Medeiros, J.A. and Cunningham, T.K. (1979). "Arterial bifurcations in the human retina." J Gen Physiol 74(4): 537-48. 190. Zhdanov, V.S., Sternby, N.H., Vikhert, A.M. and Galakhov, I.E. (1999). "Development of atherosclerosis over a 25 year period: an epidemiological autopsy study in males of 11 towns." Int J Cardiol 68(1): 95-106.
Abstract (if available)
Abstract
Cardiovascular disease (CVD) is the leading cause of death and a major cause of disability in both men and women in the United States. Atherosclerosis, which is the most common pathological process underlying CVD, may begin in the first and second decades of life and is prevalent to some degree across all age groups thereafter. More severe forms of atherosclerosis tend to be found in middle-aged and older individuals. Major risk factors for atherosclerosis include increasing age, elevated LDL cholesterol, elevated blood pressure, obesity, smoking and diabetes. Cognitive function represents the complex repertoire of abilities reflecting a dynamic interaction between the individual and the social environment throughout life. Determinants of cognition are multifactorial and include a number of demographic, social, biological, physiologic and lifestyle factors. Evidence suggests that several risk factors for CVD may also be risk factors for cognitive dysfunction. A review of the literature summarizes the epidemiologic studies that have examined the association between subclinical atherosclerosis and cognitive function. Some studies reported associations between atherosclerosis and reduced cognitive function, while others reported weak or null associations. The studies tended to measure atherosclerosis in medium to large-sized arteries, and focused on older populations with more advanced atherosclerosis, with CVD, or who may have not been cognitively intact.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Dietary fats, atherosclerosis, and cardiovascular disease
PDF
Subclinical carotid atherosclerosis, psychosocial measures, and cognitive function in middle- to older-aged adults
PDF
Genetic studies of inflammation and cardiovascular disease
PDF
Association of subclinical atherosclerosis with plasma B-vitamin, cysteine, homocysteine, and cysteinyl glycine in a cardiovascular disease-free population
PDF
Genetic risk factors in multiple myeloma
PDF
Hormone therapy timing hypothesis and atherosclerosis
PDF
Association of carotid artery stiffness with measures of cognition in older adults
PDF
Subclinical atherosclerosis in overweight Latino youth: influence of cardiometabolic risk factors
PDF
Cross-sectional association of blood pressure, antihypertensive medications, MRI volumetric measures and cognitive function scores in an aging population
PDF
Genes and environment in prostate cancer risk and prognosis
PDF
Long term effects of air pollution on lung function in the European community respiratory health survey
PDF
Environmental risk factors of Multiple Sclerosis: a twin study
PDF
Fluid intake, micturition habits, associated medications and conditions as potential risk factors for bladder cancer
PDF
Relationship of blood pressure and antihypertensive medications to cognitive change in the BVAIT, WISH, and ELITE clinical trials
PDF
Genetic and environmental risk factors for childhood cancer
PDF
Early life risk factors for childhood asthma
PDF
Age related macular degeneration in Latinos: risk factors and impact on quality of life
PDF
Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer
PDF
Cognitive functioning following ovarian removal before or after natural menopause
PDF
Common immune-related factors and risk of non-Hodgkin lymphomy
Asset Metadata
Creator
Gatto, Nicole M.
(author)
Core Title
Cardiovascular disease risk factors and cognitive function
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Epidemiology
Publication Date
11/20/2007
Defense Date
08/20/2007
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
atherosclerosis,cardiovascular disease,cognitive function,intima-media thickness,memory,metabolic syndrome,OAI-PMH Harvest,subclinical
Language
English
Advisor
Mack, Wendy J. (
committee chair
), Gatz, Margaret (
committee member
), Gilliland, Frank D. (
committee member
), Henderson, Victor (
committee member
), Hodis, Howard Neil (
committee member
)
Creator Email
ngatto@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m936
Unique identifier
UC1326173
Identifier
etd-Gatto-20071120 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-485951 (legacy record id),usctheses-m936 (legacy record id)
Legacy Identifier
etd-Gatto-20071120.pdf
Dmrecord
485951
Document Type
Dissertation
Rights
Gatto, Nicole M.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
atherosclerosis
cardiovascular disease
cognitive function
intima-media thickness
memory
metabolic syndrome
subclinical