The Cross-sectional Association Between Blood Pressure and Alzheimer s Disease in a Biracial Community Population of Older Persons

Transcription

1 Journal of Gerontology: MEDICAL SCIENCES 2000, Vol. 55A, No. 3, M130 M136 Copyright 2000 by The Gerontological Society of America The Cross-sectional Association Between Blood Pressure and Alzheimer s Disease in a Biracial Community Population of Older Persons Martha Clare Morris, 1 Paul A. Scherr, 2 Liesi E. Hebert, 1 David A. Bennett, 1 Robert S. Wilson, 1 Robert J. Glynn, 3 and Denis A. Evans 1 1 Rush Institute for Healthy Aging, Rush University and Rush-Presbyterian St. Luke s Medical Center, Chicago, Illinois. 2 Health Care and Aging Studies Branch, Centers for Disease Control and Prevention, Atlanta, Georgia. 3 Division of Preventive Medicine, Department of Medicine, Brigham and Women s Hospital and Harvard Medical School, Boston, Massachusetts. Background. The relation of blood pressure to Alzheimer s disease (AD) is complex because both an association of high blood pressure with increased risk of the disease and lower blood pressure as a consequence of the disease are possible. Methods. We examined the cross-sectional association of blood pressure and AD in the Chicago Health and Aging Project (CHAP), a study of a geographically defined, biracial community. After in-home interviews with 6,162 residents 65 years, a stratified random sample of 729 participants was clinically evaluated; 709 had blood pressures measured, and 243 were diagnosed with AD. Results. In logistic regression models adjusted for age, sex, education, and race there was no association between blood pressure measured as a continuous variable and Alzheimer s disease. In categorical analyses, however, prevalence of Alzheimer s disease was significantly higher among persons with low systolic pressure ( 130 mmhg) compared with the referent group of mmhg (odds ratio [OR] 2.2, 95% confidence interval [CI]: 1.2,4.1), and with low diastolic pressure ( 70 mmhg) compared to the referent of mmhg (OR 1.8, 95% CI: 1.1,3.1). High systolic and diastolic categories were not statistically different from the referent group, although there was some evidence that the associations differed by race. The odds ratios changed little with further adjustment for apolipoprotein E genotype, antihypertensive medications, body mass, stroke, diabetes, and heart disease. Conclusion. These findings are consistent with previous studies showing associations between low blood pressure and AD, but longitudinal studies are needed to characterize cause-and-effect associations. N OT only is blood pressure a possible risk factor for Alzheimer s disease, it may also be affected by this disease process (1). Most investigations of blood pressure and Alzheimer s disease (AD) have been case-control studies of clinical groups (2 7) and have shown either no association (2,3) or lower blood pressure among AD patients (4 7). Epidemiologic studies are limited to a few cross-sectional community studies reporting higher prevalence of AD (8) and dementia severity (9) among persons with low blood pressure, and one small longitudinal study which suggested that high blood pressure is a risk factor for AD, and that low blood pressure is an outcome of the disease process (10,11). The results of several large community studies of the association between blood pressure and cognitive function also varied widely (12 19). We examined the cross-sectional association between blood pressure and AD among participants of the Chicago Health and Aging Project, a large, biracial community study of older persons. METHODS The Chicago Health and Aging Project (CHAP) is a study of Alzheimer s disease among 6,162 older residents of three adjacent neighborhoods on the south side of Chicago. The initial component of the study, conducted from 1993 to 1997, included a door-to-door census of the community, inhome interviews with residents aged 65 years and older, and clinical evaluations for AD and other neurologic diseases in a stratified random sample of the participants. The study was approved by the Human Investigation Committee of Rush-Presbyterian St. Luke s Medical Center, Chicago. Population. The community census identified a total of 66,114 residents, 8,501 of whom were aged 65 years and older (59.1% black, 40.7% white, and 0.2% other racial groups). Before participation could be obtained for the population interview, 439 had died and 249 had moved. Of the remaining 7,813 older residents, 1,651 (21.1%) declined, and 6,162 participated in the initial population interviews (78.9% overall; 81.4% of blacks; 75.1% of whites). Participants were similar to nonparticipants in gender (60.7% and 59.0% female) and age (mean years of 73.7 and 73.0, respectively). The population interviews required, on average, one hour and 25 minutes to complete. Sample for clinical evaluation for AD. Clinical evaluations for AD were conducted for a stratified random sam- M130

2 BLOOD PRESSURE AND ALZHEIMER S DISEASE M131 ple of the CHAP participants. Sampling was conducted randomly within strata defined by age (65 69, 70 74, 75 79, 80 84, and 85 years), race (black, white, and other), gender, and performance on the East Boston Immediate and Delayed Memory Tests (good, intermediate, and poor). A total of 1,056 persons were selected in the sample; 86 died and 9 moved before participation. Of the remaining 961 persons, 729 participated in the clinical evaluations (75.8% overall; 68.8% of blacks, 86.2% of whites). Nonparticipants did not differ significantly from participants by level of blood pressure, age, gender, body mass index, or histories of heart attack, stroke, or diabetes. The overall participation among blacks in both the population interview and clinical evaluation was 9% lower than the overall participation among whites. The median time interval between the population interview and the clinical evaluation was 4.9 months. We excluded from the analyses 2 persons with no blood pressure measurements, 3 persons with no data on years of education, 1 person not of black or white race, and 14 persons for whom data were insufficient to complete an evaluation for AD. This left 709 persons for the analyses of blood pressure and AD. Alzheimer s disease and vascular dementia. The clinical evaluations were performed in the participants homes and included: (a) a battery of cognitive tests (20 29) administered by trained technicians and reviewed by a neuropsychologist and the examining neurologist; (b) a structured health history (including information on medication use) and neurologic examination performed by both a specially trained nurse clinician (30,31) and a neurologist; and (c) laboratory testing. In addition, informant interviews were conducted for a 10% random sample and for all cognitively impaired participants. Magnetic resonance imaging (MRI) was performed in cases where dementia was evident and there was uncertainty as to whether a stroke had occurred. The diagnosis of Alzheimer s disease was based on criteria of the Joint Work Group of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer s Disease and Related Disorders Association [NINCDS-ADRDA (32)]. The NINCDS-ADRDA criteria exclude from the probable Alzheimer s disease category those persons who meet criteria for clinical diagnosis of AD but also have another coexisting condition that could be responsible for dementia. Because this restriction makes dementing illness mutually exclusive, any risk factor for another dementing illness will protect against AD from the classification system alone. We did not use the restriction in the primary analysis presented here because elevated blood pressure is a risk factor for stroke and vascular dementia. Vascular dementia was diagnosed using the criteria proposed by the National Institute of Neurological Disorders and Stroke and the Association Internationale pour la Recherche et l Enseignement en Neurosciences International Workshop [NINDS-AIREN criteria (33)], but with MRI examination restricted to persons in whom there was evidence of dementia and uncertainty as to whether a stroke had occurred, or its relation to dementia. Blood pressure measurement. Blood pressure was measured both at the population interview and at the clinical evaluation according to the protocol of the Hypertension Detection and Follow-Up Program (34). On each occasion, two consecutive blood pressures were measured with 30 seconds between measurements, using standard sphygmomanometers on seated subjects with the arm resting at heart level. Subjects were fitted with one of four standard cuff sizes after measurement of the mid-arm circumference. The first and fifth Korotkoff phases were recorded as systolic and diastolic measurements, respectively. The average of the four blood pressure measurements was used in all analyses except for a group of 66 persons for whom blood pressure was missing at one of the two visits; in this case the average from one visit was used. Covariates. Gender and race were recorded as part of the census and verified at the population interview. Race questions and categories were those used by the 1990 U.S. Census. Age was computed from self-reported birth date and date of the population interview. Education (reported highest grade or years of education) was obtained during the population interview and verified by informant for cognitively impaired participants at the clinical evaluation. Body mass index (BMI) was computed as weight (kg) divided by height (m 2 ). Weight was computed as the average of two measurements (one taken at the population interview and one at the clinical evaluation) using a digital freestanding scale placed on a hard, flat surface and with the participant s shoes removed. Height was self-reported in the population interview. Use of digitalis, antihypertensives, and diabetic medications was defined as use at either the population interview or the clinical evaluation as determined by direct inspection of all medications taken within the previous 2 weeks. History of diabetes was defined as use of antidiabetic medication at either visit or participant report of clinically diagnosed diabetes, sugar in the urine, or high blood sugar. Myocardial infarction was defined as clinically diagnosed heart attack, coronary thrombosis, coronary occlusion, or myocardial infarction, as reported by the participant during the population interview. History of stroke was defined as probable or possible stroke as diagnosed at the clinical evaluation by a neurologist on the basis of a uniform, structured examination and medical history (31,33,35). The presence of the apolipoprotein E 4 allele was based on genotyping of blood drawn on 646 of the participants using methods adapted from those of Hixson and Vernier (36). Statistical analysis. We used logistic regression analyses to estimate the odds ratios for AD by level of blood pressure, and SUDAAN software to estimate relationships in the total population and calculate variances controlled for the sample design and participation (37). Before terms for blood pressure were considered, we determined a best fitting model of the association between AD and demographic variables (age, sex, education, and race). We examined model residuals and significance levels (p.05) of linear and polynomial (or curvilinear) terms for age and education, and interactions between all variables. None of the higher order or interaction terms was significant. The basic model used to examine the association of blood pressure to AD included age (modeled in years and centered at the mean 80

3 M132 MORRIS ET AL. years), sex (female referent), race (white referent), and education (modeled in years and centered at the mean 12 years). We considered continuous terms for blood pressure including linear, polynomial (up to the second power), and squareroot transformations as well as categorical terms with a priori-defined categories of 10 mmhg. Multiple-adjusted models included terms from the basic-adjusted model as well as terms for apolipoprotein E 4 genotype, antihypertensive medication use, and BMI. We examined the blood pressure- AD association in 12 basic adjusted models stratified by each of race (black/white), age (65 79 years, 80 years), sex, education ( 12 years, 12 years), apolipoprotein E 4 genotype (at least one allele, no 4 allele), and antihypertensive medication use (yes/no). The likelihood ratio test was used to assess race interactions with blood pressure categories (38). Analyses including apolipoprotein E genotype were reweighted to adjust for additional missing values. RESULTS Blood pressure and covariates. The analyses of blood pressure and AD included 380 black and 329 white participants. The black participants were more likely than white participants to have high systolic and diastolic blood pressure (Figure 1). Blood pressure was associated with a number of factors that could modify or confound associations between blood pressure and AD. Low diastolic blood pressure ( 70 mmhg) was associated with higher prevalence of chronic disease, including diabetes, clinical stroke, myocardial infarction, and digitalis use (Table 1). Use of antihypertensive medication was most prevalent among persons with high diastolic pressure. Digitalis use was most prevalent among persons with low systolic pressure and among those with low diastolic blood pressure. The prevalence of the apolipoprotein E 4 allele and of clinical stroke was highest in the lower and upper extremes of systolic blood pressure. Mean body mass increased with increasing systolic and diastolic pressure. among persons with low systolic pressure ( 130 mmhg) compared with the referent group of mmhg, and among persons with low diastolic pressure ( 70 mmhg) compared with the referent group of mmhg (Table 2). Further adjustment for the presence of at least one apolipoprotein E 4 allele, BMI, and antihypertensive medication use (the multiple-adjusted model) did not materially alter the associations with the lowest categories of systolic and diastolic pressure, but the 95% CI for low diastolic pressure was slightly wider and nonsignificant. There was a suggestion in the multiple-adjusted model of an increased association for high systolic pressure ( 160 mmhg). The odds ratio increased from 1.3 in the basic-adjusted model to 2.5 in the multiple-adjusted model, although the CI was wide and nonsignificant, indicating uncertainty in the estimated odds ratio (95% CI: 0.9, 6.8). Analyses by race. We examined the association between levels of blood pressure and AD within black and white racial groups (Table 2). Among blacks, the odds ratio for AD in the basic-adjusted model was increased in the Prevalence of AD and vascular disease. Among the 709 persons in the analyses, there were 243 cases of AD (166 among blacks and 77 among whites). In 181 of these cases, AD was the only dementing illness; in 44 cases the participants had AD and vascular dementia; and in 18 cases the participants had AD and a nonvascular dementing illness. There were 51 cases of vascular dementia (the 44 cases who also had AD and 7 cases with vascular dementia only). The prevalence of AD increased with age and was higher among blacks than whites in all age groups. Cross-sectional association of blood pressure and AD. When examined as continuous variables, neither systolic nor diastolic blood pressure was associated with AD in the total sample or within black or white racial groups in logistic regression models adjusted for age, sex, and education. Per 10 mmhg increase in blood pressure, the odds ratios for AD in the total population were OR 0.91 (95% confidence interval [CI]: 0.79, 1.04) for systolic pressure, and OR 0.92 (95% CI: 0.76, 1.25) for diastolic pressure. In categorical analyses, however, the prevalence of Alzheimer s disease was significantly higher in the total sample Figure 1. Population distribution of (A) systolic and (B) diastolic blood pressure for 380 black and 329 white participants of the clinical evaluation sample, the Chicago Health and Aging Project, Percentages are sample data weighted to account for the stratified sampling design.

4 BLOOD PRESSURE AND ALZHEIMER S DISEASE M133 lowest category ( 70 mmhg) of diastolic pressure (OR 1.9, 95% CI: 1.0, 3.7) with marginal statistical significance. Adjustment for all potential confounders lowered the odds ratio to 1.5, and it was no longer statistically significant (95% CI: 0.6, 3.6). The prevalence of AD among blacks for low systolic pressure or for higher levels of systolic and diastolic pressure was not significantly different from the referent levels. There were fewer cases of AD among whites and thus less precise estimates of the association by categories of blood pressure. We observed higher prevalence of AD among whites in the lowest systolic category ( 130 mmhg) and in the two highest diastolic categories of mmhg and 90 mmhg (Table 2). The odds ratios were statistically significant and remained so, although with very wide confidence intervals when we controlled for all potential confounders (OR 4.0 for systolic 130 mmhg, 95% CI: 1.4, 11.3 and OR 7.1, 95% CI: 1.1, 45.0) for diastolic pressure 90 mmhg. We conducted separate analyses to test whether the associations between blood pressure and AD differed by race; the differences were statistically significant for both systolic pressure ( , p.001) and diastolic pressure ( , p.001). Analyses considering stroke and chronic disease. To examine whether the observed associations in the extreme categories of blood pressure may be due to the confounding effects of stroke and other chronic disease, we added terms to the adjusted logistic models for clinical stroke, myocardial infarction, digitalis use, or diabetes. The inclusion of these terms did not alter the odds ratios for AD by level of blood pressure as estimated in the multipleadjusted models for the total sample or within racial groups. Table 1. Population Distribution* of Characteristics From the Stratified Sample Evaluated for Alzheimer s Disease by Categories of Systolic and Diastolic Blood Pressure Systolic Blood Pressure (mmhg) Diastolic Blood Pressure (mmhg) Characteristic n Mean age (years) Male (%) Mean education (years) Apolipoprotein E4 1 or 2 alleles (%) Mean body mass index Any antihypertensive medications (%) Digitalis (%) Myocardial infarction (%) Stroke (%) Diabetes (%) Note: Subjects (N 709) evaluated in the Chicago Health and Aging Project, *Population distributions are computed from weighted sample distributions. Table 2. Basic- and Multiple-Adjusted Odds Ratios (95% CI) of Alzheimer s Disease in Relation to Levels of Blood Pressure in the Total Sample and by Race Systolic Pressure (mmhg) Diastolic Pressure (mmhg) Model n AD Total Sample Basic Adjusted (1.2, 4.1) (0.6, 2.2) (0.5, 2.6) (0.6, 3.2) (1.1, 3.1) (0.8, 2.7) (0.4, 2.1) Multiple Adjusted (1.0, 4.4) (0.6, 2.8) (0.6, 3.9) (0.9, 6.8) (0.8, 3.2) (0.8, 3.1) (0.4, 2.4) Blacks Basic Adjusted (0.9, 4.0) (0.3, 1.8) (0.4, 3.3) (0.4, 3.4) (1.0, 3.7) (0.5, 2.6) (0.2, 1.5) Multiple Adjusted (0.6, 4.0) (0.4, 2.5) (0.4, 5.2) (0.6, 9.8) (0.6, 3.6) (0.6, 3.3) (0.1, 1.3) Whites Basic Adjusted (1.4, 10.0) (1.1, 7.0) (0.3, 3.6) (0.7, 8.0) (0.8, 5.1) (1.0, 7.3) (1.4, 17.8) Multiple Adjusted (1.4, 11.3) (0.9, 7.1) (0.4, 4.8) (0.6, 15.9) (0.9, 5.6) (1.0, 7.7) (1.1, 45.0) Notes: Basic-adjusted models are adjusted for age, sex, race, education, and the stratified sampling design. Multiple-adjusted models are adjusted for age, sex, race, education, apolipoprotein E4, body mass index, and antihypertensive medications and the stratified sampling design. Subjects were clinically evaluated for Alzheimer s disease in the Chicago Health and Aging Project,

5 M134 MORRIS ET AL. We also repeated all analyses using the NINCDS/ ADRDA criteria exactly as written, that exclude cases of AD with coexisting disease. In each model, the odds ratios for the more exclusive definition of AD were closer to the null and not statistically significant. Analyses within subgroups. Because the association between blood pressure and AD may differ for various subgroups, we examined the association within categories of sex, education ( 12 years, 12 years), age (65 79 years, 80 years), apolipoprotein E genotype (at least one 4 allele, no 4 allele), antihypertensive medication use, and chronic disease (clinical stroke, myocardial infarction, digitalis use, or diabetes); see Table 3. The associations were not materially different within categories of these variables, as indicated by the overlapping confidence intervals. The small increased prevalence observed in the total sample for the lowest level of systolic pressure was most evident among women, persons with no 4 allele, and those aged years. In the subgroup of persons with either clinical stroke, heart disease, or diabetes, persons with high systolic pressure ( 160 mmhg) had a significantly increased prevalence of AD compared with persons with systolic pressure of mmhg (OR 2.4, 95% CI: 1.1, 4.1). DISCUSSION We found limited evidence for a cross-sectional association between blood pressure and AD in this large, biracial community study. There was no association between blood pressure as a continuous variable and AD. In categorical analyses, however, persons with low systolic and diastolic pressure had moderately increased prevalence. There was an increased prevalence of AD among white participants with high diastolic pressure that was not observed in the total sample, among black participants, or within other subgroups. However, because there were few whites in the high blood pressure categories, and no association with high blood pressure in the total sample, the racial differences must be viewed with caution. In another analysis of the entire CHAP population (Morris and colleagues, unpublished manuscript), we found no association between blood pressure and cognitive function among white participants, and small decreases in cognitive test scores of from 1 to 5 percentiles for black participants in the extreme ranges of both systolic pressure ( 100 mmhg and 180 mmhg) and diastolic pressure ( 60 mmhg and 100 mmhg). In the present study, stroke and the apo E 4 genotype were more prevalent among persons in the extremes of the systolic blood pressure range, and both were strongly associated with Alzheimer s disease. Adjustment for these variables in the analysis did not change the associations between blood pressure and AD. However, there was some evidence of interaction such that persons with both high systolic pressure and either stroke, myocardial infarction, or diabetes also had increased likelihood of AD. One possible explanation for these findings could be that cerebrovascular Table 3. Basic Adjusted Odds Ratios (95% CI) for Prevalent Disease in Relation to Levels of Blood Pressure Within Categories of Other Risk Factors in 709 Sample Participants Systolic Pressure (mmhg) Diastolic Pressure (mmhg) Risk Factors n AD Age (years)* (1.1, 7.5) (0.3, 2.8) (0.3, 4.5) (0.4, 5.6) (1.3, 6.5) (0.8, 4.8) (0.3, 2.9) (0.7, 3.0) (0.6, 3.2) (0.4, 2.5) (0.4, 2.6) (0.6, 2.1) (0.5, 2.2) (0.4, 4.8) Apolipoprotein E* Any 4 allele (0.5, 6.1) (0.4, 5.7) (0.6, 7.6) (0.7, 15.8) (0.9, 6.4) (0.7, 6.7) (0.4, 4.4) No 4 allele (2.0, 8.9) (0.9, 4.0) (0.2, 2.3) (0.6, 4.4) (0.7, 2.7) (0.5, 2.9) (0.1, 1.2) Stroke, heart disease, diabetes* Yes (1.0, 7.0) (0.5, 4.0) (0.4, 4.3) (1.1, 4.1) (1.1, 5.0) (0.8, 5.0) (0.5, 5.6) No (1.0, 5.0) (0.4, 2.5) (0.4, 2.9) (0.4, 5.5) (0.6, 2.9) (0.4, 2.7) (0.2, 1.7) Antihypertensive medications* Yes (1.1, 6.2) (0.5, 2.6) (0.4, 3.5) (0.4, 3.5) (1.1, 4.6) (0.5, 2.8) (0.4, 3.1) No (0.6, 3.7) (0.4, 3.3) (0.3, 4.0) (0.4, 6.2) (0.6, 2.8) (0.6, 4.0) (0.2, 3.0) Sex** Females (1.5, 6.9) (0.6, 3.0) (0.3, 2.1) (0.4, 3.0) (0.9, 3.8) (0.7, 3.2) (0.4, 3.1) Males (0.4, 3.0) (0.2, 2.3) (0.4, 6.1) (0.3, 6.1) (0.8, 4.2) (0.5, 3.9) (0.2, 2.4) Notes: Subjects were clinically evaluated for Alzheimer s disease in the Chicago Health and Aging Project, *Models are adjusted for age, sex, race, education, and the stratified sampling design. **Models adjusted for age, education, race, and the stratified sampling design.

6 BLOOD PRESSURE AND ALZHEIMER S DISEASE M135 disease modifies the clinical expression of Alzheimer s disease (39). Population studies of Alzheimer s disease and blood pressure are few in number. To our knowledge, ours is the only biracial study to examine this issue. This type of study is less affected by bias due to patient selection factors than are clinically based case-control studies, because all persons in the population have a chance of selection for clinical evaluation for AD. In population studies, low participation rates may be a source of bias if associated with both the exposure and disease status. Participation in the CHAP study was moderately high, although somewhat lower among black residents. Other characteristics, such as level of blood pressure, age, gender, weight, and the presence of chronic conditions, did not differ significantly by participation status. This reduces concerns that the findings may be biased or nonrepresentative of the total population. Further, the CHAP study used structured and uniform procedures for the diagnosis of Alzheimer s disease and vascular dementia, thereby lessening the chance for misclassification or bias in disease diagnosis. As with any cross-sectional study, however, the study results are difficult to interpret because of the inability to distinguish between cause and effects of disease. In addition, because prevalent disease, by definition, includes a mixture of cases in all phases of disease, effects of AD on blood pressure may be obscured. This could explain, for example, the higher prevalence of stroke and the 4 allele in both extremes of the blood pressure range. The relation between blood pressure and Alzheimer s disease was examined among 1,642 participants of the Kungsholmen project in Sweden (8). Prevalent Alzheimer s disease was substantially higher among persons with systolic pressure less than 120 mmhg when compared with persons with levels of mmhg (OR 4.6). In this study, low diastolic pressure was also significantly inversely associated with prevalent disease. We were not able to examine the association for systolic levels lower than 120 mmhg because of an insufficient number of persons with blood pressure that low. Another Swedish study (10) examined the relation of blood pressure to incident Alzheimer s disease, and observed higher levels of blood pressure among a small number of persons diagnosed with AD 9 to 15 years later. Blood pressure levels declined in the entire sample, but those with dementia had a greater decline beginning several years before disease diagnosis. Results of the present study, showing moderately increased prevalence of Alzheimer s disease among persons with low levels of blood pressure, are consistent with an effect of Alzheimer s disease on blood pressure. There is also a suggestion of increased prevalence among persons with high diastolic pressure; this would support the theory that high blood pressure increases the risk of Alzheimer s disease. These results could also be explained by chance, or by confounding due to associations between blood pressure and advanced disease or to other unknown risk factors. Longitudinal population-based studies are needed to determine whether blood pressure is a risk factor for incident Alzheimer s disease, and also whether disease affects blood pressure. Acknowledgments This study was supported by grants AG and AG from the National Institute on Aging. The authors thank the communities of Washington Heights, Morgan Park, and Beverly, Illinois for their support, and gratefully acknowledge the work of study coordinators Michelle Bos and Holly Hadden, data manager Linda Jankowski, and their staffs. Address correspondence to Dr. Martha Clare Morris, Rush Institute for Healthy Aging, 1645 West Jackson, Suite 675, Chicago, IL mmorris@rush.edu References 1. Burke WJ, Coronado PG, Schmitt CA, Gillespie KM, Chung HD. Blood pressure regulation in Alzheimer s disease. J Autonom Nerv Sys. 1994;48: Waldemar G, Christiansen P, Larsson HB, et al. White matter magnetic resonance hyperintensities in dementia of the Alzheimer type: morphological and regional cerebral blood flow correlates. J Neurol Neurosurg Psych. 1994;57: Kozachuk WE, DeCarli C, Schapiro MB, Wagner EE, Rapport SI, Horwitz B. White matter hyperintensities in dementia of Alzheimer s type and in healthy subjects without cerebrovascular risk factors. A magnetic resonance imaging study. Arch Neurol. 1990;47: Kilander L, Boberg M, Lithell H. Peripheral glucose metabolism and insulin sensitivity in Alzheimer s disease. Acta Neurol Scand. 1993; 87: Landin K, Blennow K, Wallin A, Gottfries CG. Low blood pressure and blood glucose levels in Alzheimer s disease. Evidence for a hypometabolic disorder. J Intern Med. 1993;233: Wang SJ, Liao KK, Fuh JL, et al. Cardiovascular autonomic functions in Alzheimer s disease. Age Ageing. 1994;23: Tresch DD, Folstein MF, Rabins PV, Hazzard WR. Prevalence and significance of cardiovascular disease and hypertension in elderly patients with dementia and depression. J Am Geriatr Soc. 1985;33: Guo Z, Viitanen M, Fratiglioni L, Winblad B. Low blood pressure and dementia in elderly people: the Kungsholmen project. Br Med J. 1996; 312: Hogan DB, Ebly EM, Rockwood K. Weight, blood pressure, osmolarity, and glucose levels across various stages of Alzheimer s disease and dementia. Dementia Geriatr Cog Disord. 1997;8: Skoog I, Lemfelt B, Landahl S, et al. 15-year longitudinal study of blood pressure and dementia. Lancet. 1996;347: Skoog I, Andreasson LA, Landahl S, Lernfelt B. A population-based study on blood pressure and brain atrophy in 85 year olds. Hypertension. 1998;32: Scherr PA, Hebert LE, Smith LA, Evans DA. Relation of blood pressure to cognitive performance in the elderly. Am J Epidemiol. 1991; 134: Kuusisto J, Koivisto K, Mykkanen L, et al. Essential hypertension and cognitive function. The role of hyperinsulinemia. Hypertension. 1993; 22: Swan GE, LaRue A, Carmelli D, Reed TE, Fabsitz RR. Decline in cognitive performance in aging twins. Arch Neurol. 1992;49: Wallace RB, Lemke JH, Morris MC, Goodenberger M, Kohout F, Hinrichs JV. Relationship of free-recall memory to hypertension in the elderly. The Iowa 65 Rural Health Study. J Chronic Dis. 1985;38: Launer LJ, Masaki K, Petrovitch H, Foley D, Havlik RJ. The association between mid-life blood pressure levels and late-life cognitive function. The Honolulu-Asia Aging Study. JAMA. 1995;274: Guo Z, Fratiglioni L, Winblad B, Viitanen M. Blood pressure and performance on the Mini-Mental State Examination in the very old. Am J Epidemiol. 1997;145: Farmer ME, Kittner SJ, Abbott RD, Wolz MM, Wolf PA, White LR. Longitudinally measured blood pressure, anti-hypertensive medication use and cognitive performance: the Framingham Study. J Clin Epidemiol. 1990;43: Elias MF, Wolf PA, D Agostino RB, Cobb J, White LR. Untreated blood pressure level is inversely related to cognitive functioning: the Framingham study. Am J Epidemiol. 1993;138:

7 M136 MORRIS ET AL. 20. Morris JC, Heyman A, Mohs RC, et al. CERAD Part I. Clinical and neuropsychological assessment of Alzheimer s disease. Neurology. 1989;39: Welsh KA, Butters N, Mohs RC, et al. CERAD Part V. A normative study of the neuropsychological batter. Neurology. 1994;44: Wechsler D. Wechsler Memory Scale Revised Manual. New York: Psychological Corporation; Smith A. Symbol Digit Modalities Test Manual Revised. Los Angeles: Western Psychological; Benton AL, Varney NR, Hamsher KD. Visuospatial judgement: a clinical test. Arch Neurol. 1978;35: Goodglass H, Kaplan E. The Assessment of Aphasia and Related Disorders. Philadelphia: Lea & Febiger; Ekstrom RB, French JW, Harman HH, Dermen D. Manual kit for factor-referenced cognitive tests. Princeton, NJ: Educational Testing Service; Cooper JA, Sugar HJ. Incidental and intentional recall in Parkinsons disease: an account based on diminished attentional resources. J Clin Exp Neuropsychol. 1993;15: Raven JC, Court JH, Raven J. Standard Progressive Matrices: Raven Manual, Section 3. Oxford, UK: Oxford Psychologists Press; Nelson HE. National Adult Reading Test (NART) Test Manual. Windsor, UK: NFER-NELSON Publishing; Bennett DA, Shannon KM, Beckett LA, Goetz CG, Wilson RS. Metric properties of nurses ratings of parkinsonian signs with a modified Unified Parkinson s Disease Rating Scale. Neurology. 1997;49: Goldstein LB, Bertels C, Davis JN. Interrater reliability of the NIH Stroke Scale. Arch Neurol. 1989;46: McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer s disease: report of the NINCDS- ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer s Disease. Neurology. 1984;34: Roman GC, Tatemichi TK, Erkinjunitti T, et al. Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN international workshop. Neurology. 1993;43: Hypertension Detection and Follow-Up Program Cooperative Group. Variability of blood pressure and the results of screening in the HDFP. J Chronic Dis. 1978;31: Foulkes MA, Wolf PH, Price TR, et al. The Stroke Data Bank: design, methods and baseline characteristics. Stroke. 1988;19: Hixson J, Vernier DO. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with Hha 1. J Lipid Res. 1990; 31: Shah BV, Barnwell BG, Bieler GS. SUDAAN s User s Manual, Release 7.0. Research Triangle Park, NC: Research Triangle Institute; Cox, DR. Analysis of Binary Data. London: Methuen; Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the clinical expression of Alzheimer s disease. JAMA. 1997;227: Received August 17, 1998 Accepted July 26, 1999 Decision Editor: William B. Ershler, MD Please mention The Journals of Gerontology when replying to advertisements.