Poster Session P3: Epidemiology and Risk Factors of Alzheimer's Disease ~ 5 ] AGE AT ONSET AND FAMILIAL PATTERNS OF RISK IN AD Jeremy M. Silverman*, Michal Schnaider-Beeri, Hillel Grossman, Joy Wang, Elana Pulver. Mount Sinai School of Medicine, New York, NY,, USA. Contact e-mail: jeremy.silverman@mssm.edu Background: Earlier evidence suggests that relatives of very late onset AD probands (onset age: 85+ years) have a reduced risk of AD compared with relatives of earlier late onset AD probands (onset age: 60-84 years). The changing levels of familial risk across the late life span remains very little explored. Objective(s): To examine patterns of familial risk associated with age at onset across 773 late onset AD probands stratified into four onset groups determined by finding approximately equal sized quartiles (cut points: age 69, 76, 82) and 403 nondemented elderly probands (greater than age 82). Methods: Informant-based assessment was conducted on 3122 60+ year old siblings and parents (relatives) of the AD probands and 1542 60+ year old relatives of the nondemented controls. Survival curves were constructed and compared for the four AD proband groups of relatives as well as tbose of the nondemented elderly. Results: Increasingly lower levels of risk: were observed with each later onset group with the lowest levels in the nondemented proband group. Significant differences (P < 0.05, most P < 0.0001) were observed for 9 of the 10 pairwise comparisons, the tenth (comparing relatives of AD probands with onset 60=-68 v. 69=-75) showed a trend level difference (p = 0.08). Conclusions: With respect to proband onset age, relatives of late onset AD probands do not face a uniform level of risk for AD. There are incremental reductions in the level of risk to relatives of late AD probands as age at onset increases. ~ T H E EFFECT OF APOE ON THE RELATIONSHIP BETWEEN DIABETES AND DEMENTIA: RESULTS FROM THE CANADIAN STUDY OF HEALTH AND AGING Chris MacKnight*, Kenneth Rockwood. Dalhousie University, Halifax, NS, Canada. Contact e-mail: chris.macknight@dal.ca Background: We have previously found that diabetes is related to incident vascular cognitive impairment, though not to Alzheimer's disease. Objec- tive: To investigate the effect of the E4 allele of apolipoprotein E on the relationship between diabetes and dementia. Methods: Data are from phases I and II of the Canadian Study of Health and Aging (1991/92 and 1996/97). This involved 10,263 individuals at baseline. Apolipoprotein E results were available on 1183 subjects, of whom 134 were diabetic. Diabetes was determined at baseline by a combination of history, medication use and laboratory variables. Incident dementia was diagnosed by conference after a comprehensive evaluation, by traditional criteria. The cause of dementia was divided into Alzheimer's disease, vascular cognitive impairment and other. ApoE was determined from blood samples taken at phase II. Logistic regression was performed, and crude and adjusted odds ratios are reported. ApoE status was defined both by any vs. no E4 and number of E4 alleles. Results: There were 162 subjects with incident Alzheimer's disease and 137 with incident vascular cognitive impairment. Diabetes was not related to incident dementia, with no effect modification from ApoE (OR 1.07, 95%CI 0.68-1.69). Results were similar for incident Alzheimer's disease (OR 0.99, 95%CI 0.57-1.73). Crude and adjusted (for age, sex, education and ApoE status) odds ratios for incident vascular cognitive impairment were significant, and unchanged by the addition of ApoE to the model (OR 1.88, 95%CI 1.16--3.06 and OR 1.86, 95%CI 1.14-3.04). Conclusions: Diabetes is a risk factor for vascular cognitive impairment, though not for Alzheimer's disease. Apolipoprotein E did not affect this relationship. $39t ~ D O E S SPOUSAL DEPRESSION AFFECT COGNITIVE DECLINE IN THEIR PARTNER? Ladson Hinton* 1, Mary Haan 2, Nancy West 2. i UC Davis, Sacramento, CA, USA; 2University of Michigan, Ann Arbog MI, USA. Contact e-mail: Iadson. hinton @ ucdmc, ucdavis, edu Background: While there has been considerable attention to the role of individual factors in cognitive decline, less research has focused on social contextual factors. The marital relationship and characteristics of a spouse form an important aspect of the social contexts of many older adults. Objective: To investigate the effect of spousal depressive symptoms on cognitive decline in their partner over a three year period in a cohort of older Latinos. Methods: Participants for this study were 310 spousal pairs who were participants in a large cohort study of Latino elderly (age 60 and above) in the Sacramento area. Three annual waves of data were used to model the relationship betweeu spousal depressive symptoms (CES-D score) and cognitive decline (3 MS score) in their partner. Repeated measures analysis (mixed models) was used to evaluate the effect of spousal baseline depressive symptoms on 3-year cognitive decline in their partner. Partner's age, education, cultural orientation, baseline stroke and type 2 diabetes were included in the models. Results: Wives were younger and more often Spanish speaking compared with their husbands. Wives' baseline CES-D scores and 3MS scores were higher than their husbands'. In the adjusted models, spousal CES-D scores predicted (P < 0.05) decline in their partner's 3MS score over three years of follow-up. For both husbands and wives, additional variables associated (p < 0.05) with decline in the 3MS included being older and having less formal education. Conclusion: Husbands and wives experience cognitive decline in response to their partner's depressive symptoms. ~ - ~ EFFECT OF FOLIC ACID FORTIFICATION ON SERUM FOLATE LEVELS IN AFRICAN AMERICAN ELDERLY Floyd B. Willis*, Neill Graff-Radfurd, John Lucas, Francine Parfitt. Mayo Clinic Jacksonville, Jacksonville, FL, USA. Contact e-mail: willis.floyd@mayo.edu Background: There is a large amount of epidemiological evidence showing association between common moderately elevated plasma homocysteine levels and increased risk of vascular disease, stroke, and Alzheimer's Disease (AD). Folic acid functions as a substrate in the metabolism of homocysteine and numerous studies have shown that dietary supplementation of folic acid can lower serum homocysteine levels. Compared to Caucasians, elderiy African Americans (AA) have been shown to be at higher risk of morbidity and mortaIity from vascular disease and AD. Therefore, it is important to study the role that folate may play in risk modification for AAs. In 1996, the Food and Drug Administration (FDA) sought to reduce risk of neural tube defects in pregnancy by mandating folic acid supplementation in all enriched US flour, rice, and other grain products with a minimum of 140 micrograms (~g) of folate per I00 grams (g) of products. The mandatory compliance date was January 1, 1998. Jacques et a1 showed that this action resulted in increased serum folate leveis and decreased homocysteine levels in a subset of persons from the Framingham Heart Study. To our knowledge, there is no such published data on elderly AAs and little insight oi1 whether folate supplementation is beneficial to this group. Objective: 1) To examine the effect of folic acid fortification on serum folate levels of a group of elderly AAs. Methods: Longitudinal study involving 112 elderly AAs Table 1. Summary of measurements before and after January i, 1998 Variable Before After Change P-value January 1, January 1, (After-Before) 1988 1988 Folate I~g/L (N = 108) 7.8 (5.0) 10.3 (4.9) 2.5 (6.6) <0.001 Homocysteine Ixmol/L (N = 112) 11.5 (4.9) 12.7(6.5) 1.3 (6.1) 0.03 Creafinine mg/dl (N = 109) 1.1 (0.4) 1.1 (0.7) 0.1 (0.4) 0.13 Vitamin B-I2 ng/L (N = 93) 534 (277) 556 (298) 23 (272) 0.4 - the sample mean is given with the sample standard deviation in parenthesis