attentive" visual processing. However, few studies have investigated the role of "implicit" visual processing in the recognition of emotions gener- ated by unrecognized visual stimuli. Methods: We recorded the skin con- ductance response (SCR) of a 67 year-old patient with posterior Cortical Atrophy (PCA), as well as 5 matched controls, during the presentation of neutral, positive and negative affective stimuli. We compared SCR of pictures depicting persons and objects selected from the International Af- fective Picture System (IAPS). We also recorded the SCR while the partic- ipants observed the same stimulus being degraded or camouflaged. After the presentation of each picture, the subjects were instructed to describe the picture and to indicate its valence (positive, negative or neutral). Results: Compared to controls, the patient with PCA exhibited the same pattern of SCR modulation between negative and positive stimulus, even when the stimulus presented was not recognized. In addition, while the controls failed to indicate the valence of degraded pictures, the patient with PCA was able to identify them. Conclusions: These results suggest that this patient could implicitly recognize emotions in the absence of vi- sual conscious knowledge. Furthermore, we found the patient made suc- cessful use of compensatory strategies to recognize the valence of visual stimuli despite the visual deficits. P4-094 REPRODUCIBILITY OF A FULLY AUTOMATED CHEMILUMINESCENT TAU ASSAY Anne Tweedie 1 , George Green 2 , Holly Soares 3 , Flora Berisha 4 , Robert Neely 5 , Irina Baburina 6 , Salvatore Salamone 7 , Daniel Kozo 8 , Shari Jackson 9 , Karen Ackles 9 , Deb Byrne 9 , Lisa DiMagno 9 , 1 Ortho Clinical Diagnostics, Webster, NewYork, United States; 2 Bristol-Myers Squibb, Princeton, New Jersey, United States; 3 Bristol-Myers Squibb, Wallingford, Connecticut, United States; 4 Bristol-Myers Squibb, Lawrenceville, New Jersey, United States; 5 Bristol-Myers Squibb, Princeton, New Jersey, United States; 6 Saladax Biomedical, Inc., Bethlehem, Pennsylvania, United States; 7 Saladax Biomedical, Inc., Bethlehem, New Jersey, United States; 8 Saladax Biomedical, Inc., Bethlehem, Pennsylvania, United States; 9 Ortho Clinical Diagnostics, Rochester, New York, United States. Contact e-mail: Atweedi2@its.jnj.com Background: Tau concentration in cerebrospinal fluid (CSF) has been shown to correlate with Alzheimer’s disease (AD) progression and is a key measure- ment in the identification of people at risk for progressing to Alzheimer’s dis- ease (AD) and the proper management of people with AD. The use of such markers for clinical purposes requires assays that can deliver consistent analyt- ical and clinical performance. We report on the VITROS Ò Immunodiagnostic Products Tau Assay 1 assay under development for the measurement of tau in CSF on the VITROS Immunodiagnostic Systems. The objective of this study was to demonstrate consistent performance of the assay across lots and across instrument platforms. Methods: Two lots of VITROS Tau assay reagent were tested on the VITROS Ò ECiQ and VITROS Ò 3600 Immunodiagnostic Sys- tems to evaluate lot-to-lot and system-to-system variability. Fifty individual CSF samples and controls were run in singleton with controls run at the be- ginning and end of each test. The results from the VITROS 3600 using Lot 1 reagent were used as the control condition and linear regression was per- formed, as well as the %CV calculated across all runs. Results: The sys- tem-to-system comparison (Lot 1 - ECiQ vs. Lot 1 - 3600) slope, intercept, and correlation (r) results were 0.9913, 2.87, and 1.000 respectively. The lot-to-lot comparison (Lot 2 - 3600 vs. Lot 1 - 3600) slope, intercept, and cor- relation (r) results were 0.979, -8.70, and 0.999 respectively. Across lots and systems (Lot 2 - ECiQ vs. Lot 1 - 3600) slope, intercept, and correlation (r) results were 0.981, 0.65, and 1.000 respectively. The mean doses of the con- trols (n ¼ 8 per control) ranged from 64.8 - 4145 pg/mL with an overall %CV from 0.9% - 4.1%. The mean doses of the individual CSF samples (n ¼ 4 per CSF) ranged from 56.2 - 1412 pg/mL with a %CV from 0.5% - 5.8%, overall 2.4%CV across lots and systems. Conclusions: The VITROS Immunodiag- nostic Products Tau Assay 1 has demonstrated consistent results across re- agent lots and across systems. P4-095 VASCULAR FACTORS AS PREDICTORS OF SEVERE DEMENTIA AND MORTALITY IN ALZHEIMER’S DISEASE: THE CACHE COUNTY DEMENTIA PROGRESSION STUDY JoAnn Tschanz 1 , Sarah Schwartz 1 , Mac Gilbert 1 , Joseph Wanzek 1 , Chelsea Sanders 1 , Michelle Mielke 2 , Chris Corcoran 1 , Maria Norton 1 , Constantine Lyketsos 3 , 1 Utah State University, Logan, Utah, United States; 2 Mayo Clinic, Rochester, Minnesota, United States; 3 Johns Hopkins University, Baltimore, Maryland, United States. Contact e-mail: joann. tschanz@usu.edu Background: Some vascular risk factors, conditions, and treatments have been found to predict the rate of dementia progression in Alzheimer’s disease (AD).In the Cache county study, atrial fibrillation and hypertension were asso- ciated with faster decline and select antihypertensive and "statin" medications were associated with slower decline. We extend our prior work by examining whether vascular factors also predict important clinical outcomes of severe de- mentia and mortality. Methods: AD cases (n¼333, 66% female) were diag- nosed by consensus panel. The mean time between dementia onset and diagnosis was 1.63 (SD¼1.25) years. Participants were re-examined every 6-18 months with the Mini-mental State Exam (MMSE), Clinical Dementia Rating (CDR) for up to 16 years. Severe dementia was defined as MMSE <¼ 10 or CDR¼3. Mortality was monitored through obituaries. History of hypertension, high cholesterol (HC), diabetes, stroke, myocardial infarction (MI), and coronary artery bypass graft or CABG was obtained by self-report prior to dementia onset or caregiver report thereafter. Anti-hypertensive and statin use were determined at each visit. These predictors were examined in Cox Regression models to estimate hazard of severe dementia and death, test- ing the following covariates: dementia onset age, duration, gender, APOE, and education (at least high school or less). Results: Mean (SD) age of onset of AD was 84.27 (SD¼6.39). Seventy-nine subjects developed severe demen- tia and 297 died over the period of observation. In individual Cox regression models, MI was associated with higher hazard of severe dementia (Hazard Ratios, 95% CI ¼ 1.61, 0.97-2.69) and HC was associated with lower hazard (0.56, 0.35-0.89), although only HC remained significant with the inclusion of covariates. Examination of HC and treatment (HC-no statins, HC-statins vs. neither) revealed reduction in hazard for severe dementia for HC-satins (0.48, 0.25-0.93). Additionally, only HC-statins (0.64, 0.45-0.89) was associ- ated with a lower hazard of death. Antihypertensives predicted neither time to severe dementia or death. Conclusions: These results extend our previous finding that statin use among persons with AD and history of high cholesterol is associated with slower progression. Further research is indicated to examine whether medication characteristics (i.e., penetration of blood brain barrier) have differential effects. P4-096 APOE- 3 4 AND CLINICAL PROGRESSION IN PARKINSON’S DISEASE Sarah Monsell 1 , Katherine Bristow 2 , Lilah Besser 3 , Harvey Checkoway 3 , Irene Litvan 4 , Walter Kukull 3 , 1 National Alzheimer’s Coordinating Center, Seattle, Washington, United States; 2 NACC/University of Washington, Table 1 Results of two Cox Regression models. The left part of the table shows hazard ratios for predictors of severe dementia, whereas the right part of the table shows hazard ratios for predictors of death. Severe dementia Mortality Hazard Ratio P value 95% CI Hazard Ratio P value 95% CI HC-no statins 0.65 0.11 0.38-1.11 0.84 0.23 0.63-1.17 HC-statins 0.48 0.029 0.25-0.93 0.64 0.008 0.45-0.89 Dementia duration 1.25 0.006 1.07-1.46 0.80 <0.001 0.72-0.89 Age 1.00 0.86 0.96-1.05 1.09 <0.001 1.07-1.11 < High school 1.99 0.023 1.10-3.59 NA NA NA Female sex NA NA NA 0.69 0.005 0.54-0.90 Poster Presentations: P4 P738