- Email: [email protected]
THE RELATIONSHIP BETWEEN CEREBRAL GLUCOSE METABOLISM AND COGNITION IN ADVANCE
P350
Poster Presentations: Sunday, July 16, 2017
to low space so that a patient’s AD progression can be easily visualized. Meanwhile, the probability measure is calculated by SVM with fuller features and minimum loss of prediction power. Along with other clinical tests, this new approach has the potential to integrate multiple features to assist in tracking AD progression. P1-262
ASSESSING RETINAL VASCULAR BIOMARKERS FOR ALZHEIMER’S DISEASE USING ULTRA-WIDEFIELD IMAGING (UWFI)
Lajos Csincsik1, Erin Flynn2, Enrico Pellegrini3, Giorgos Papanastasiou3, Tom MacGillivray3, Craig W. Ritchie3, Tunde Peto4, Imre Lengyel4, 1 Queen’s University Belfast, Belfast, United Kingdom; 2The George Washington University, Washington DC, USA; 3University of Edinburgh, Edinburgh, United Kingdom; 4Queen’s University Belfast, Belfast, United Kingdom. Contact e-mail: [email protected] Background: Pathological changes in the eye have been reported in a range of neurodegenerative diseases including changes in retinal vascular parameters (RVPs). The purpose of this study was to assess whether there are measurable RVP changes associated with Alzheimer’s disease (AD) using UWFI and whether any progression related changes can be detected during a 2 year period. Methods: UWFI were acquired of 20 healthy controls (HC; (MMSE>28) and 13 participants with AD (MMSE<26) using an OPTOS TX 200 scanning laser ophthalmoscope at baseline (BL) and after a two-year follow-up (FU). The study had full local Ethical Committee approval. Images were analysed computationally for several different vascular parameters including arterial and venular branching, vessel width gradient (WG), fractal dimensions (FD) of the vascular network patters and vessel tortuosity. Analysis was carried out either on the whole image or in a circular zone (Zone C) centred on the optic disc (OD) stretching from 0.5 ODDs to 2 ODDs from the OD boundary. The image and zone was further divided into four quadrants centred on the OD. The data was analysed using Student’s t-test; results with p<0.05 were reported. Results: In total, 78 images were included in the analysis (46 HC and 32 AD). There was no significant difference in age between HC and AD groups (77.767.2 vs. 71.3610.2; p>0.1). In the AD group at BL there was a decreased arterial FD (HC¼1.360.02; AD¼1.260.02; P<0.01) when all quadrant was analysed together, however this difference had become non-significant at FU. There was also an increased venular WG (HC¼6.05x10-361x10-3; AD¼8.23x10-362x10-3; p<0.01) when all quadrant was analysed together in AD group at BL and this difference had remained significant at FU too (HC¼6.26x10-361x103 ; AD¼8.99x10-363x10-3; p<0.05). From BL to FU our analysis could not detect progression in any of the RVPs. Conclusions: Our data from this pilot study suggests that retinal vascular biomarkers measured on UWFI can distinguish between HC and AD but might not be sensitive enough to detect progression during a 2-year period. Larger participant numbers and longer FU period might be needed to detect progression of RVPs. P1-263
WHAT HAPPENS WHEN TWO AD AMYLOID BIOMARKERS HAVE DISCORDANT RESULTS IN THE SAME PATIENT?
Patricio Alexis Chrem Mendez1, Ismael Luis Calandri1, Lucia Pertierra1, Allegri Ricardo2, Silvia Vazquez1, 1FLENI, Buenos Aires, Argentina; 2 Neurological Research Institute Raul Carrea, Buenos Aires City, Argentina. Contact e-mail: [email protected]
Background: Due to the development of new diagnostic tools for Alzheimer disease (AD), new clinical scenarios could be found. AD biomarkers increase diagnostic certainty and the risk of having AD. Therefore, they are used for the selection of participants in prodromal AD clinical protocols, but also in the clinical setting to differentiate overlapping dementia syndromes. However, what happens when two AD biomarkers have discordant results in the same patient? The objective of this study is to assess discordant rates between AD (Alzheimer Disease) biomarkers and to analyze longitudinal follow-up of these patients. Methods: In our Memory Center, we studied 144 patients with both PETPiB and Ab42 values in CSF. We established the percentage of agreement between biomarkers comparing the cut-off of Ab42 values in CSF and the PET-PIB results. We reviewed the clinical records of patients with discordant results taking into account changes in diagnosis and dementia conversion rate. Results: The percentage of discordance in patients seen in our clinic was 13.8 % (20/144). When we considered only the group of probable AD patients, the rate was only 5.5%. The discordance was due mainly to Ab42 positive with PET-PIB negative (80%). We excluded 9 patients due to lack of prospective clinical follow up in our center. The clinical diagnosis of these 11 patients were: MCI or probable AD in only 2 cases. The clinical diagnosis for the rest of the cases was heterogeneous and included FTD spectrum, encephalitis and psychiatric diseases. The clinical diagnosis after the medical team received the results was changed in only 4 cases. 4 patients converted to dementia in longitudinal follow up. Conclusions: Discordance between CSF and nuclear methods is more frequent when we evaluated patients with low clinical suspicious for AD. This discordance could be due in part to different sensitivity of the 2 methods but also could reflect a more heterogeneous non AD population. P1-264
THE RELATIONSHIP BETWEEN CEREBRAL GLUCOSE METABOLISM AND COGNITION IN ADVANCE
Sarah N. Forrester1, Jeannie-Marie S. Leoutsakos1, Andres M. Lozano2, Lisa Fosdick2, Steven Targum3, Constantine Lyketsos1, Gwenn S. Smith1, The ADvance Study Group, 1Johns Hopkins University School of Medicine, Baltimore, MD, USA; 2Functional Neuromodulation, Ltd, Charlottesville, VA, USA; 3Clintara, Boston, MA, USA. Contact e-mail: [email protected] Background: Consistent reductions in cerebral glucose metabolism measured with PET and [18F]-2-deoxy-2-fluoro-Dglucose ([18F]FDG) have been found in heteromodal association cortices in patients with AD and MCI compared to controls (Smith et al, 1992, Drzezga et al, 2003 ). Further, cerebral glucose metabolism is a sensitive marker for disease progression in AD that can predict decline in cognition (Landau et al 2010). In a multisite, double-blind study (ADvance) of the effects of deep brain stimulation of the fornix (DBS-f) in mild AD, cerebral glucose metabolism, as well as cognitive outcomes (ADAS-Cog, CDR, and CVLT), were included as outcome measures. The objective of the current analysis is to determine the sensitivity of glucose metabolism to predicting changes in cognition associated with DBS-f. Methods: Forty-two mild AD patients were randomized one month after electrode implantation, to continuous stimulation or sham groups, and were monitored for 24 months. The effect of DBS-f on glucose metabolism was measured using logistic regression and cognitive test scores at 6 months and 12 months were predicted by change in glucose
Poster Presentations: Sunday, July 16, 2017
metabolism at these same timepoints using lagged analysis. The regions of interest included regions increased with DBS-f that comprise regions spared (pre- and post-central gyrus, cuneus and cerebellum,) and regions affected (temporal and parietal association cortex and hippocampus) in AD. Results: Regression analysis showed that treatment predicted increased hippocampal glucose metabolism after 6-months of DBS in AD patients 65 years and older (b ¼ 0.566, p ¼ 0.040). The lagged analysis revealed that increases in glucose metabolism in the hippocampus, post-central gyrus (b ¼ 0.156, p ¼ 0.025), and the cerebellum (b ¼ 0.162, p ¼ 0.019) at 6-months predicted an increase in CVLT at 12 months, not the other measures. Conclusions: Mild AD patients treated with DBS show increased hippocampal glucose metabolism. Our findings confirm and extend the association between DBS and glucose metabolism in the hippocampus found in the main ADvance outcome paper (Lozano et al, 2016). Further, glucose metabolism at 6-months predicted increases in verbal memory at 12-months. Future analyses will evaluate whether the predictive effect at 6 months continues beyond 12 months. Grant# 4R01AG042165-05.
P351
gest that the effect of gender on the rate of transitioning between groups is negligible. P1-266
WITHDRAWN
P1-267
BDNF VAL66MET INCREASES RATE OF MEMORY DECLINE, HIPPOCAMPAL VOLUME LOSS AND TAU ACCUMULATION IN AUTOSOMAL DOMINANT ALZHEIMER’S DISEASE
Yen Ying Lim1, Jason Hassenstab2, Carlos Cruchaga3, Alison M. Goate4, Anne M. Fagan5, Tammie L. S. Benzinger5, Paul Maruff1,6, Colin L. Masters1, Eric McDade7, John C. Morris5, Randall J. Bateman5, Dominantly Inherited Alzheimer Network, 1The Florey Institute of Neuroscience and Mental Health, Parkville, Australia; 2Washington University in St. Louis, St Louis, MO, USA; 3Washington University in St. Louis, Saint Louis, MO, USA; 4Icahn School of Medicine at Mount Sinai, New York, NY, USA; 5Washington University School of Medicine, St. Louis, MO, USA; 6Cogstate Ltd., Melbourne, Australia; 7Washington University at St. Louis, St. Louis, MO, USA. Contact e-mail: [email protected] Background: The brain-derived neurotrophic factor (BDNF) Val66-
P1-265
RATES OF PROGRESSION BETWEEN DIAGNOSTIC GROUPS TOWARDS ALZHEIMER’S DISEASE: COMPARISON BETWEEN COHORT STUDIES
Stephanie Evans1, Kevin McRae-McKee1, Alison Ower1, Christoforos Hadjichrysanthou1, Mei Mei Wong1, Frank de Wolf1,2, Roy Anderson1, 1Imperial College London, London, United Kingdom; 2 Janssen Prevention Center, Leiden, Netherlands. Contact e-mail: s.evans@ imperial.ac.uk Background: There exist a large number of cohort studies that have been used to identify genetic and biological risk factors for developing Alzheimer’s disease (AD). However, there is disagreement between studies as to how strongly these risk factors affect the rate of progression through diagnostic groups towards AD. We have calculated the probability of transitioning through diagnostic groups in a number of AD specific cohort studies, and considered how uncertainty around the strength of the effect of these risk factors affects estimates of the distribution of individuals in each diagnostic group in an AD clinical trial simulator. Methods: Using a generalised linear mixed model we have calculated the odds ratios of an individual transitioning through diagnostic groups, given the combination of risk factors that they possess, in a number of cohort studies. The odds ratios have been transformed into transition probability estimates and have been used within an AD clinical trial simulator to study the effect that uncertainty around the magnitude of these odds ratios, and therefore transition probabilities, has on predictions of the number of patients that would develop AD in a clinical trial. Results: We have compared the number of people that develop AD in the clinical trial simulation under a range of transition probabilities with data from real-world non-AD specific cohort studies. By using the number of people in each diagnostic group to calibrate the parameters in the clinical trial simulator, we have provided a better understanding of the effect of known AD risk factors on the probability of transitioning through diagnostic groups. Conclusions: Current data suggests that the most significant risk factor for developing AD is being in a state of mild cognitive impairment and being APOE positive. Results also sug-
Met polymorphism (rs6265) is implicated in synaptic excitation and neuronal integrity. In autosomal dominant Alzheimer’s disease (ADAD), mutation carriers (MC) who also carry the Met66 allele show worse memory and higher levels of cerebrospinal fluid (CSF) tau, but equivalent amyloid levels compared to MC Val66 homozygotes at baseline. The aim of this study was to determine the extent to which the BDNF Val66Met polymorphism affects changes in memory, brain volume, tau and Ab in ADAD prospectively. Methods: Prospective neuropsychological, biomarker and neuroimaging data collected from the Dominantly Inherited Alzheimer Network (DIAN) over w2 years were analyzed in 81 preclinical mutation carriers (MC), all with a clinical dementia rating (CDR) score of 0 and estimated to be 11 years prior to clinical symptom onset, and 78 matched mutation non-carriers (NC). BDNF genotype was obtained for MCs (58 Val66 homozygotes, 23 Met66 carriers). Results: Compared to MC Val66 homozygotes, MC Met66 carriers showed greater decline in episodic memory (p<.001), loss of hippocampal volume (p¼.005), and increase of CSF tau (p<.001) (Figure 1). Cortical Ab accumulation was equivalent between MC Val66 homozygotes and MC Met66 carriers (p¼.427) (Figure 1). Compared to NCs, MC Val66 homozygotes showed greater increase in cortical Ab accumulation (p<.001) but equivalent rates of change in episodic memory decline (p¼.700), loss of hippocampal volume (p¼.215), and accumulation of CSF tau (p¼.266) (Figure 1). Conclusions: ADAD is associated with pathologically increased rates of Ab and tau accumulation, loss of hippocampal volume and decline in episodic memory. The results of the current study show that for MCs who also carry the BDNF Met66 allele, decline in episodic memory, loss of hippocampal volume and increase in CSF tau is substantially greater than for MCs who are Val66 homozygotes, despite equivalent rates of Ab accumulation. This is consistent with findings in preclinical sporadic AD, where amyloid positive Met66 carriers also show faster deterioration in episodic memory and hippocampal volume, but not Ab accumulation, when compared to Ab+ Val66 homozygotes. Together, these data suggest that the BDNF Val66Met polymorphism modifies the contributions to the neurodegenerative process in ADAD.
Poster Presentations: Sunday, July 16, 2017
to low space so that a patient’s AD progression can be easily visualized. Meanwhile, the probability measure is calculated by SVM with fuller features and minimum loss of prediction power. Along with other clinical tests, this new approach has the potential to integrate multiple features to assist in tracking AD progression. P1-262
ASSESSING RETINAL VASCULAR BIOMARKERS FOR ALZHEIMER’S DISEASE USING ULTRA-WIDEFIELD IMAGING (UWFI)
Lajos Csincsik1, Erin Flynn2, Enrico Pellegrini3, Giorgos Papanastasiou3, Tom MacGillivray3, Craig W. Ritchie3, Tunde Peto4, Imre Lengyel4, 1 Queen’s University Belfast, Belfast, United Kingdom; 2The George Washington University, Washington DC, USA; 3University of Edinburgh, Edinburgh, United Kingdom; 4Queen’s University Belfast, Belfast, United Kingdom. Contact e-mail: [email protected] Background: Pathological changes in the eye have been reported in a range of neurodegenerative diseases including changes in retinal vascular parameters (RVPs). The purpose of this study was to assess whether there are measurable RVP changes associated with Alzheimer’s disease (AD) using UWFI and whether any progression related changes can be detected during a 2 year period. Methods: UWFI were acquired of 20 healthy controls (HC; (MMSE>28) and 13 participants with AD (MMSE<26) using an OPTOS TX 200 scanning laser ophthalmoscope at baseline (BL) and after a two-year follow-up (FU). The study had full local Ethical Committee approval. Images were analysed computationally for several different vascular parameters including arterial and venular branching, vessel width gradient (WG), fractal dimensions (FD) of the vascular network patters and vessel tortuosity. Analysis was carried out either on the whole image or in a circular zone (Zone C) centred on the optic disc (OD) stretching from 0.5 ODDs to 2 ODDs from the OD boundary. The image and zone was further divided into four quadrants centred on the OD. The data was analysed using Student’s t-test; results with p<0.05 were reported. Results: In total, 78 images were included in the analysis (46 HC and 32 AD). There was no significant difference in age between HC and AD groups (77.767.2 vs. 71.3610.2; p>0.1). In the AD group at BL there was a decreased arterial FD (HC¼1.360.02; AD¼1.260.02; P<0.01) when all quadrant was analysed together, however this difference had become non-significant at FU. There was also an increased venular WG (HC¼6.05x10-361x10-3; AD¼8.23x10-362x10-3; p<0.01) when all quadrant was analysed together in AD group at BL and this difference had remained significant at FU too (HC¼6.26x10-361x103 ; AD¼8.99x10-363x10-3; p<0.05). From BL to FU our analysis could not detect progression in any of the RVPs. Conclusions: Our data from this pilot study suggests that retinal vascular biomarkers measured on UWFI can distinguish between HC and AD but might not be sensitive enough to detect progression during a 2-year period. Larger participant numbers and longer FU period might be needed to detect progression of RVPs. P1-263
WHAT HAPPENS WHEN TWO AD AMYLOID BIOMARKERS HAVE DISCORDANT RESULTS IN THE SAME PATIENT?
Patricio Alexis Chrem Mendez1, Ismael Luis Calandri1, Lucia Pertierra1, Allegri Ricardo2, Silvia Vazquez1, 1FLENI, Buenos Aires, Argentina; 2 Neurological Research Institute Raul Carrea, Buenos Aires City, Argentina. Contact e-mail: [email protected]
Background: Due to the development of new diagnostic tools for Alzheimer disease (AD), new clinical scenarios could be found. AD biomarkers increase diagnostic certainty and the risk of having AD. Therefore, they are used for the selection of participants in prodromal AD clinical protocols, but also in the clinical setting to differentiate overlapping dementia syndromes. However, what happens when two AD biomarkers have discordant results in the same patient? The objective of this study is to assess discordant rates between AD (Alzheimer Disease) biomarkers and to analyze longitudinal follow-up of these patients. Methods: In our Memory Center, we studied 144 patients with both PETPiB and Ab42 values in CSF. We established the percentage of agreement between biomarkers comparing the cut-off of Ab42 values in CSF and the PET-PIB results. We reviewed the clinical records of patients with discordant results taking into account changes in diagnosis and dementia conversion rate. Results: The percentage of discordance in patients seen in our clinic was 13.8 % (20/144). When we considered only the group of probable AD patients, the rate was only 5.5%. The discordance was due mainly to Ab42 positive with PET-PIB negative (80%). We excluded 9 patients due to lack of prospective clinical follow up in our center. The clinical diagnosis of these 11 patients were: MCI or probable AD in only 2 cases. The clinical diagnosis for the rest of the cases was heterogeneous and included FTD spectrum, encephalitis and psychiatric diseases. The clinical diagnosis after the medical team received the results was changed in only 4 cases. 4 patients converted to dementia in longitudinal follow up. Conclusions: Discordance between CSF and nuclear methods is more frequent when we evaluated patients with low clinical suspicious for AD. This discordance could be due in part to different sensitivity of the 2 methods but also could reflect a more heterogeneous non AD population. P1-264
THE RELATIONSHIP BETWEEN CEREBRAL GLUCOSE METABOLISM AND COGNITION IN ADVANCE
Sarah N. Forrester1, Jeannie-Marie S. Leoutsakos1, Andres M. Lozano2, Lisa Fosdick2, Steven Targum3, Constantine Lyketsos1, Gwenn S. Smith1, The ADvance Study Group, 1Johns Hopkins University School of Medicine, Baltimore, MD, USA; 2Functional Neuromodulation, Ltd, Charlottesville, VA, USA; 3Clintara, Boston, MA, USA. Contact e-mail: [email protected] Background: Consistent reductions in cerebral glucose metabolism measured with PET and [18F]-2-deoxy-2-fluoro-Dglucose ([18F]FDG) have been found in heteromodal association cortices in patients with AD and MCI compared to controls (Smith et al, 1992, Drzezga et al, 2003 ). Further, cerebral glucose metabolism is a sensitive marker for disease progression in AD that can predict decline in cognition (Landau et al 2010). In a multisite, double-blind study (ADvance) of the effects of deep brain stimulation of the fornix (DBS-f) in mild AD, cerebral glucose metabolism, as well as cognitive outcomes (ADAS-Cog, CDR, and CVLT), were included as outcome measures. The objective of the current analysis is to determine the sensitivity of glucose metabolism to predicting changes in cognition associated with DBS-f. Methods: Forty-two mild AD patients were randomized one month after electrode implantation, to continuous stimulation or sham groups, and were monitored for 24 months. The effect of DBS-f on glucose metabolism was measured using logistic regression and cognitive test scores at 6 months and 12 months were predicted by change in glucose
Poster Presentations: Sunday, July 16, 2017
metabolism at these same timepoints using lagged analysis. The regions of interest included regions increased with DBS-f that comprise regions spared (pre- and post-central gyrus, cuneus and cerebellum,) and regions affected (temporal and parietal association cortex and hippocampus) in AD. Results: Regression analysis showed that treatment predicted increased hippocampal glucose metabolism after 6-months of DBS in AD patients 65 years and older (b ¼ 0.566, p ¼ 0.040). The lagged analysis revealed that increases in glucose metabolism in the hippocampus, post-central gyrus (b ¼ 0.156, p ¼ 0.025), and the cerebellum (b ¼ 0.162, p ¼ 0.019) at 6-months predicted an increase in CVLT at 12 months, not the other measures. Conclusions: Mild AD patients treated with DBS show increased hippocampal glucose metabolism. Our findings confirm and extend the association between DBS and glucose metabolism in the hippocampus found in the main ADvance outcome paper (Lozano et al, 2016). Further, glucose metabolism at 6-months predicted increases in verbal memory at 12-months. Future analyses will evaluate whether the predictive effect at 6 months continues beyond 12 months. Grant# 4R01AG042165-05.
P351
gest that the effect of gender on the rate of transitioning between groups is negligible. P1-266
WITHDRAWN
P1-267
BDNF VAL66MET INCREASES RATE OF MEMORY DECLINE, HIPPOCAMPAL VOLUME LOSS AND TAU ACCUMULATION IN AUTOSOMAL DOMINANT ALZHEIMER’S DISEASE
Yen Ying Lim1, Jason Hassenstab2, Carlos Cruchaga3, Alison M. Goate4, Anne M. Fagan5, Tammie L. S. Benzinger5, Paul Maruff1,6, Colin L. Masters1, Eric McDade7, John C. Morris5, Randall J. Bateman5, Dominantly Inherited Alzheimer Network, 1The Florey Institute of Neuroscience and Mental Health, Parkville, Australia; 2Washington University in St. Louis, St Louis, MO, USA; 3Washington University in St. Louis, Saint Louis, MO, USA; 4Icahn School of Medicine at Mount Sinai, New York, NY, USA; 5Washington University School of Medicine, St. Louis, MO, USA; 6Cogstate Ltd., Melbourne, Australia; 7Washington University at St. Louis, St. Louis, MO, USA. Contact e-mail: [email protected] Background: The brain-derived neurotrophic factor (BDNF) Val66-
P1-265
RATES OF PROGRESSION BETWEEN DIAGNOSTIC GROUPS TOWARDS ALZHEIMER’S DISEASE: COMPARISON BETWEEN COHORT STUDIES
Stephanie Evans1, Kevin McRae-McKee1, Alison Ower1, Christoforos Hadjichrysanthou1, Mei Mei Wong1, Frank de Wolf1,2, Roy Anderson1, 1Imperial College London, London, United Kingdom; 2 Janssen Prevention Center, Leiden, Netherlands. Contact e-mail: s.evans@ imperial.ac.uk Background: There exist a large number of cohort studies that have been used to identify genetic and biological risk factors for developing Alzheimer’s disease (AD). However, there is disagreement between studies as to how strongly these risk factors affect the rate of progression through diagnostic groups towards AD. We have calculated the probability of transitioning through diagnostic groups in a number of AD specific cohort studies, and considered how uncertainty around the strength of the effect of these risk factors affects estimates of the distribution of individuals in each diagnostic group in an AD clinical trial simulator. Methods: Using a generalised linear mixed model we have calculated the odds ratios of an individual transitioning through diagnostic groups, given the combination of risk factors that they possess, in a number of cohort studies. The odds ratios have been transformed into transition probability estimates and have been used within an AD clinical trial simulator to study the effect that uncertainty around the magnitude of these odds ratios, and therefore transition probabilities, has on predictions of the number of patients that would develop AD in a clinical trial. Results: We have compared the number of people that develop AD in the clinical trial simulation under a range of transition probabilities with data from real-world non-AD specific cohort studies. By using the number of people in each diagnostic group to calibrate the parameters in the clinical trial simulator, we have provided a better understanding of the effect of known AD risk factors on the probability of transitioning through diagnostic groups. Conclusions: Current data suggests that the most significant risk factor for developing AD is being in a state of mild cognitive impairment and being APOE positive. Results also sug-
Met polymorphism (rs6265) is implicated in synaptic excitation and neuronal integrity. In autosomal dominant Alzheimer’s disease (ADAD), mutation carriers (MC) who also carry the Met66 allele show worse memory and higher levels of cerebrospinal fluid (CSF) tau, but equivalent amyloid levels compared to MC Val66 homozygotes at baseline. The aim of this study was to determine the extent to which the BDNF Val66Met polymorphism affects changes in memory, brain volume, tau and Ab in ADAD prospectively. Methods: Prospective neuropsychological, biomarker and neuroimaging data collected from the Dominantly Inherited Alzheimer Network (DIAN) over w2 years were analyzed in 81 preclinical mutation carriers (MC), all with a clinical dementia rating (CDR) score of 0 and estimated to be 11 years prior to clinical symptom onset, and 78 matched mutation non-carriers (NC). BDNF genotype was obtained for MCs (58 Val66 homozygotes, 23 Met66 carriers). Results: Compared to MC Val66 homozygotes, MC Met66 carriers showed greater decline in episodic memory (p<.001), loss of hippocampal volume (p¼.005), and increase of CSF tau (p<.001) (Figure 1). Cortical Ab accumulation was equivalent between MC Val66 homozygotes and MC Met66 carriers (p¼.427) (Figure 1). Compared to NCs, MC Val66 homozygotes showed greater increase in cortical Ab accumulation (p<.001) but equivalent rates of change in episodic memory decline (p¼.700), loss of hippocampal volume (p¼.215), and accumulation of CSF tau (p¼.266) (Figure 1). Conclusions: ADAD is associated with pathologically increased rates of Ab and tau accumulation, loss of hippocampal volume and decline in episodic memory. The results of the current study show that for MCs who also carry the BDNF Met66 allele, decline in episodic memory, loss of hippocampal volume and increase in CSF tau is substantially greater than for MCs who are Val66 homozygotes, despite equivalent rates of Ab accumulation. This is consistent with findings in preclinical sporadic AD, where amyloid positive Met66 carriers also show faster deterioration in episodic memory and hippocampal volume, but not Ab accumulation, when compared to Ab+ Val66 homozygotes. Together, these data suggest that the BDNF Val66Met polymorphism modifies the contributions to the neurodegenerative process in ADAD.