$386 Poster Session P3: Epidemiology and Risk Factors of Alzheimer's Disease were retained if significant. Results: In the final modal, MCI (OR = 28.9, p < 0.001), LLD (OR = 5.2, p = 0.010), APOE (OR = 3.1, p = 0.002), and age (OR = 1.1, p = 0.004 for each additional year) were significant main effects, with a trend for MCI by LLD interaction (p = 0.084). CVD (p = 0.357), education (p = 0.365), and gender (p = 0.407) were not significant. LLD by APOE interaction was significant (p = 0.045) with effect of LLD greater among NC (OR = 5.2) than among MCI (OR = 1.52). Analyses were repeated using only the sub-sample with CA at follow-up with comparable results. Conclusions: These results suggest that risk for development of dementia may be an interaction of presence of MCI, LLD, and the APOE gene, suggesting the incremental effect of LLD is greatest among those with no e4 allele. • RELIGIOSITY PREDICTS DEMENTIA IN ISRAELI MEN Michael Davidson 1 , Michal Schnaider Beeri* 2, Jeremy M. Silverman 2, James Schmeidler 2, Shlomo Noy 1, Ramit Ravona-Springer 1, Uri Goldbomt I . 1Sheba Medical Center, Tel-Hashomer, Israel; 2Mount Sinai School of Medicine, New York, NY, USA. Contact e-mail: michal.beeri@ mssm.edu Background: Although dementia has been associated with a variety of demographic and life style characteristics, its association with religiosity has not been examined. We assessed this association by following up participants in a 1963 cohort study of cardiovascular risk factors of Israeli men 40 to 65 years old, for the diagnosis of dementia in 1999. Methods: The original cohort was composed of 10.059 Jewish civil servants and municipal employees, of whom 2912 were presumed alive in 1999. The present analysis includes 1619 subjects (mean age in 1963= 44.4; sd = 4.0) of whom 17.9% were diagnosed with dementia. Their educational content was classified as only religious (29.0%), only secular (30.8%) or mixed (40.1%). They characterized their beliefs and practices on a five point scale of religiosity from 1= least religious to 5= most religious. Stepwise logistic regression was used to predict dementia from religious education and perceived religiosity, controlling for age, socioeconomic status (an index of extent of education and occupation), and geographical area of birth (six categories). Results: After controlling for the covariates, both educational content and religiosity were significantly associated with dementia; each remained significant after controlling for the other. Specifically, the OR for dementia was 2.0 (95% CI 1.5-2.8) for only religious relative to mixed education, 1.3 (95% CI 0.9-1.8) for only religious relative to only secular education, and 1.6 (95% CI 1.1-2.2) for only secular relative to mixed education. The OR for a unit increase in religiosity was 1.2 (95% CI 1.1-1.2). The prevalence of dementia ranged from 9.8% in the least religious category to 28.6% in the most religious category. Conclusions: The associations of two measures of religiosity with dementia persisted after controlling for sociodemographic characteristics known to be associated with dementia. The relationship between religiosity and dementia may be explained by differences in mode and content of the learned material, mode of cognitive activity (such as repetitive reading of familiar material), life style (for example, nutrition) or culturally biased assessment instruments. • PUBLIC-USE DATA FROM NACC Thomas D. Koepsell*, Walter A. Kukull, Duane Beekly, Gerald van Belle, Roger Higdon, Annette Fitzpatrick, Brenda Kurland. University of Washington, Seattle, WA, USA. Contact e-mail: koepsell@u.washington.edu Background: Since 1984, 32 Alzheimer's Disease Centers (ADCs) in the U.S. have conducted clinical and laboratory research on patients with AD and related disorders. The National Alzheimer's Coordinating Center (NACC) was established in 1999 to standardize data and facilitate research use of data across ADCs. In October, 2003, the NACC Steering Committee authorized public release of selected data on nearly 60,000 patients evaluated at an ADC. Objective(s): To describe this major new data resource. Methods: Patients in the public-use data set include 59,083 patients whose initial clinical evaluation at an ADC occurred before June, 2002. Data available on all patients include 45 items from the ADCs' Minimum Data Set, concerning sociodemographic characteristics, clinical diagnoses, comorbid conditions, Mini-Mental Status Examination scores, family history, last evaluation date, vital status, and availability of imaging test results and biological specimens. Additional neuropathologic data are included for 5,483 decedents with an autopsy. Results: Overall, 60% of database patients are female; 82% are white, 11% black, 1.4% Asian/Pacific Islander, and 4.2% of other races. The median year of birth was 1922, and the median year of first ADC evaluation was 1996. Some 38,703 patients had a clinical diagnosis of dementia, including 31,746 with probable or definite AD, with or without a coexisting neurological disorder. Among patients with dementia, 8,692 (23%) developed symptoms before age 65 years, and 13,940 had at least one first-degree relative with dementia. As of 2003, 16,744 (28%) had died, and 7,143 underwent autopsy. Conclusions: This new public-use database is among the largest data sets on Alzheimer's disease ever released to researchers. The data are probably best suited to clinical questions that can be addressed with common data elements on a large case series, including internal comparisons involving demographic subgroups, clinical features, diagnosis type, and survivorship. The neuropathological data can also be used to study associations between specific neuropathological features and their relationships with clinical and other patient characteristics. Data on image and specimen availability can be used to judge whether adequate sample size is available for special studies. Guidelines for access to, and appropriate use of, the new data will be discussed. • EVALUATION OF HFE (HEMOCHROMATOSIS) MUTATIONS AS GENETIC MODIFIERS IN SPORADIC AD AND MCI Daniel Berlin* 1,2, George Chong 3, Howard Chertkow 1,2, Howard Bergman 1,4, Natalie A. Phillips 1,5, Hyman M. Schipper 1,2. 1Bloomfield Centre for Research in Aging, Lady Davis Institute for Medical Research, S.M.B.D. Jewish General Hospital, Montreal, PQ, Canada; 2Department of Neurology and Neurosurgery, McGill University, Montreal, PQ, Canada; 3Department of Diagnostic Medicine, Lady Davis Institute for Medical Research, S.M.B.D. Jewish General Hospital, Montreal, PQ, Canada; 4Division of Geriatric Medicine, Department of Medicine, McGill University, Montreal, PQ, Canada; 5Department of Pyschology, Concordia University, Montreal, PQ, Canada. Contact e-mail: daniel.berlin @ mail.mcgilL ca Background: Extensive support has amassed for the notion that oxidative stress may be involved in the pathogenesis of Alzheimer disease (AD). Pathological brain iron deposition has been implicated as a source of neurotoxic reactive oxygen species in AD. Recent reports have examined whether DNA mutations involved in the iron overload disorder hereditary hemochromatosis (HH) can act as risk factors for AD, possibly by acceler- ating brain iron accumulation. One previous study suggested that presence of the mutant hfe-H63D allele may significantly lower the age of AD onset. Objective(s): Our objective was to determine the relationship of the hfe mutations to the demographic, clinical, and neuropsychological features of AD and MCI, and to evaluate whether an interaction existed between hfe and apololipoprotein E (apoE) status in our patient population. Methods: We genotyped 213 sporadic AD, 106 MCI, and 63 normal elderly control (NEC) individuals for the H63D and C282Y hfe mutations by polymerase chain reaction (PCR)/restriction fragment length polymorphism (RFLP) analysis. Results: We observed no statistically significant impact of the H63D or C282Y mutations on age at AD symptoms onset or diagnosis, age at onset of cognitive symptoms (AD and MCI combined), rates of MCI-to-AD conversion, or specific neuropsychological deficits. No interactions between hfe zygosity and apoE status were discerned. Patients homozygous for H63D exhibited a trend towards accelerated MCI-to-AD conversion rates and a subset of younger individuals (aged 55-75) exhibited an earlier onset of cognitive symptoms relative to wild-type hfe and H63D heterozygotes. Conclusions: Contrary to earlier reports, the results of the present study do not implicate the common life mutations as genetic modifiers of sporadic AD and MCI. Trends towards accelerated cognitive dysfunction in H63D homozygotes warrant further study.