Abstract Background Ten years after the original publication, the Centiloid framework is now broadly used to harmonize amyloid‐PET quantification, facilitate data sharing and comparison across cohorts, and even assist visual interpretation in clinical settings. We evaluated the global implementation of Centiloids by comparing their distribution and corresponding positivity thresholds across cohorts. Methods We gathered data from publicly available cohorts and reached out to investigators across the world to collect cross‐sectional Centiloids, demographic and clinical information, and visual read data. Gaussian mixture models (GMM, k=2) were fitted to Centiloid values for each cohort and cutoffs were calculated as mean + 2SD of the lower Gaussian. When visual reads were available, we determined Centiloid cutoffs that maximized correspondence with visual reads (Cohen's kappa). Data was combined across cohorts using random effects meta‐analyses. Results As of January 2025, we included 37 cohorts ( n = 41,678 participants) with heterogeneous pipelines, radiotracers, and clinical and demographic characteristics (Table‐1). The low Gaussian peaks ranged from ‐9 to 10CL; the meta‐analysis identified a common peak at 1CL. The second peaks were more heterogeneous (range=38‐102CL; meta‐analysis outcome=64CL). Across cohorts, the proportion of cognitively unimpaired versus impaired participants impacted the position of both peaks, with better separation in cohorts enriched in impaired individuals (Figure 1C). The meta‐analysis indicated a GMM‐based cutoff of 19CL (95%CI: 16‐21CL, Figure 1B); subgroup analyses showed no evidence of significant effect between single versus multicenter settings (17 versus 20CL, p = 0.30), MRI‐based or PET‐only processing (18 versus 19CL, p = 0.88), and no evidence of difference across radiotracers (Flutemetamol: 16CL; PIB: 17CL, Flutafuranol: 18CL, Florbetaben: 19CL, Florbetapir: 20CL, p = 0.78). In a subset of 29,496 participants with visual reads available, binary visual reads corresponded well to Centiloids (common kappa=0.86, Figure 2A). The visual read‐based cutoff of 24CL (95%CI: 21‐27CL, Figure 2B) maximized correspondence between visual read and quantification and was slightly higher than the GMM‐based cutoff. All meta‐analysis models showed high non‐random heterogeneity (I 2 >80%) across studies, suggesting non‐random differences in peaks and cutoffs. Conclusions Meta‐analysis‐based cutoffs align well with thresholds from the existing literature. High heterogeneity among studies underscores the need to investigate contributing factors, raising concerns about applying common cutoffs.
A Randomized, Double-Blind, Controlled, Parallel Group Study with Amustaline/Glutathione Pathogen Reduced Red Blood Cells in Regions at Potential Risk for Zika Virus Transfusion-Transmitted Infections (RedeS Study): -- Protocol for a Phase 3, Randomized, C
Abstract Background Red blood cell (RBC) transfusion may be lifesaving, however transfusion-transmitted infections (TTI) and transfusion-associated graft-versus-host disease (TA-GVHD) still pose a threat despite current donor selection and testing requirements. Amustaline (S-303)/glutathione (GSH) pathogen reduction (PR) is designed to inactivate infectious microbes and leukocytes in RBCs. The RedeS study evaluates the safety and efficacy of amustaline/GSH PR-RBCs compared to conventional RBCs in a broad spectrum of patients requiring acute and/or repeated RBC transfusion support. Methods RedeS is a Phase 3, prospective, multi-center, randomized, double-blind, active-controlled, parallel-design study, with a 6-month extension option to evaluate patients requiring repeated transfusions including red cell exchange (RCE).Eligible subjects will be stratified according to site, baseline bleeding status and participation in the extension period. The study arms include a transfusion period of 28-days for anemia with any bleeding or non-bleeding patients or a 28-day+6-month extension for non-bleeding patients with chronic anemia requiring repeated simple transfusions, or for patients with sickle cell disease (SCD) expected to receive three RCE procedures. The study will seek to enroll 600-800 subjects in the overall safety analysis, including ~140 repeatedly-transfused subjects in the 6-month extension including up to 30 subjects receiving three RCE procedures. The primary safety endpoints are the proportion of patients with any treatment-emergent adverse events (TEAEs) related to study RBC transfusion through 28 days after the last study transfusion, and the proportion of patients with treatment-emergent antibodies with confirmed specificity to PR-RBCs. The primary efficacy endpoint is the hemoglobin (Hb) increment in subjects who are non-bleeding at baseline, adjusted for the grams of Hb transfused and averaged over multiple transfusions. With 352 evaluable transfused non-bleeding subjects (176 per treatment group) from either arm, the study will have 80% power to demonstrate non-inferiority, defined as a Hb increment treatment difference of no more than 15% of the Control group mean, assuming a Control arm coefficient of variation of 50%. Discussion RedeS will characterize the safety of amustaline/GSH PR-RBCs and is designed to demonstrate the non-inferiority of PR-RBCs to conventional RBCs with respect to Hb increment, a common pragmatic clinical assessment of RBC transfusion effectiveness.
Does cohort heterogeneity affect relationships between AD biomarkers and cognition?
Abstract Background Extensive evidence links cognition, tau, and neurodegeneration in Alzheimer's disease (AD). However, in U.S. POINTER imaging (a community‐recruited clinical trial cohort selected for high cardiovascular risk, ethnoracial diversity, and increased risk of cognitive decline and dementia) tau PET is more closely linked to cognition than in ADNI. Here, we further explore how ADNI and POINTER differ with respect to relationships between cross‐sectional AD biomarkers, cortical thickness, and different cognitive domains. Method U.S. POINTER imaging (Baseline; N = 775; 68.9±5.2 years, 62.8% female) and ADNI ( N = 405; 69.9±5.4 years, 58% female) were matched by age, sex, clinical status, and APOE4 genotype (Table 1). We examined associations between baseline AD biomarkers (Amyloid‐beta (Aβ) PET in centiloids, harmonized tau PET measures of MK6240 in POINTER and FTP in ADNI), regional cortical thickness, and verbal memory (Delayed Story Recall), executive function (Trail‐Making B), and global cognition (Preclinical Alzheimer's Cognitive Composite scores) across each cohort, controlling for age, sex, and education. Result Aβ burden in ADNI was associated with worse executive function ( p < .001) and global cognition ( p < .001) but not memory, whereas in POINTER Aβ burden was not associated with any cognitive domain (Figure 1A). Among Aβ+ individuals, tau burden was associated with poorer performance across domains in ADNI, but only with poorer verbal memory in POINTER (Figures 1B&1C). Additionally, there were stronger positive associations between cortical thickness and executive function and global cognition compared to verbal memory, particularly in ADNI Aβ+ individuals even after adjusting for entorhinal tau ( p < .01; Figure 2). Conclusion Cognitive performance was linked to all biomarkers (Aβ, tau, cortical thickness) but these associations were consistently stronger in ADNI compared to POINTER, particularly in Aβ+ individuals. The stronger thickness‐cognition relationships in ADNI were not driven by tau, indicating that there is a tau‐independent role of thickness in explaining cognition. In ADNI, there were subtle differences between cognitive domains such that executive function was more closely linked than verbal memory to Aβ and cortical thickness, whereas verbal memory was linked to tau in both cohorts. Our study highlights that frequently‐reported AD biomarker associations may be attenuated in heterogeneous, community‐recruited cohorts.
Risk factors and clinical relevance of MRI vascular markers in the U.S. POINTER study: a cross‐cohort comparison with the Framingham Heart Study
Abstract Background Our objective is to examine the generalizability of the relationships between four MRI markers of CVD: DTI‐derived free water (FW), fractional anisotropy (FA), perivascular space (PVS) and white matter hyperintensities (WMH) volume, with vascular risk factors (VRF) and cognition in the baseline POINTER Imaging and Framingham Heart Study (FHS) cohorts. Method We studied 907 U.S. POINTER and 797 FHS Offspring participants with available baseline demographic, clinical and imaging data (Table 1). Continuous FW, FA, WMH (log‐transformed) and PVS measurements were regressed against total cranial volume (TCV). Cognitive outcomes included a Global Cognitive Composite and executive function composite score for POINTER and FHS. For each MRI marker, we used linear and logistic regressions to assess the effect of age, sex, hypertension, diabetes and smoking status, adjusting for ethnicity. We then assessed, for each cognitive outcome, the individual effect of MRI markers using separate linear regressions adjusted for demographic and clinical factors, as well as TCV‐adjusted gray matter volume to control for atrophy. Result POINTER and FHS differed across most demographic and biomarker characteristics (Table 1). Despite these differences, there were similar associations between demographic and VRF and the 4 MRI vascular markers in POINTER and FHS (Figure 1): increased age was associated with worse outcomes for the 4 markers, hypertension was associated with increased FW, PVS and lower FA, diabetes was associated with increased FW only, smoking status did not impact any of the four markers. Only exception was that female had lower FW and increased WMH in POINTER and increased PVS in FHS. Associations between CVD measures and cognition were also similar across cohorts: reduced cognitive performance was significantly associated with both decreased FA and increased FW (Figure 2), but not with WMH nor PVS. Conclusion Our study explores four imaging markers representing diverse mechanisms by which CVD could impact brain structure and function in the POINTER and FHS studies. Despite differences in cohort characteristics, we found mostly consistent CVD relationships with demographic, VRF, and cognitive measures between cohorts which supports the generalizability of the relation of these markers with VRF and cognition identified in the POINTER cohort.
Evaluation of Alzheimer's disease biomarkers and cognition in non‐Hispanic Black and non‐Hispanic White participants in ADNI and US POINTER Imaging
Abstract Background Given low representation of diverse populations in Alzheimer's disease (AD) research, it remains unclear if AD biomarker relationships observed primarily in non‐Hispanic White (NHW) cohorts generalize to other groups. Here we evaluate associations between key AD biomarkers and cognitive performance among cognitively unimpaired non‐Hispanic Blck (NHB) and NHW participants across 2 cohorts with unique sample characteristics. Method Participants included cognitively unimpaired NHB and NHW participants with Aβ and tau PET in ADNI (69 NHB, 402 NHW) and POINTER Imaging (135 NHB, 637 NHW). We examined harmonized measures of global Aβ (in centiloids), entorhinal cortex (EC) tau (quantified with MK6240 in POINTER and FTP in ADNI) and Preclinical Alzheimer's Cognitive Composite (PACC) scores. Regression models within each ethnoracial group evaluated relationships between 1) Aβ and EC tau, and 2) EC tau and PACC performance. Interaction models within each cohort were also conducted to assess differences in these associations by ethnoracial group. All models were adjusted for age and gender. Models with PACC additionally adjusted for education. Result In US POINTER, NHBs had lower Aβ and EC tau burden compared to NHWs (Table 1). In ADNI, there was no difference in EC tau between groups, and no differenence in Aβ after adjusting for age and gender. In US POINTER, the positive Aβ‐EC tau relationship was observed in both ethnoracial groups whereas in ADNI, Aβ was associated with higher EC tau in the NHW but not the NHB group. In both cohorts, EC tau was associated with worse PACC performance in both NHB and NHW and there were no differences in this association by ethnoracial group in interaction models (Figure 1). Conclusion NHBs had lower Aβ and EC tau than NHWs in POINTER but not ADNI, and there were inconsistent associations between Aβ and EC tau across cohorts. Despite these differences, AD biomarkers appear similarly related to cognitive outcomes across ethnoracial groups in both cohorts. Future analyses will attempt to further determine the generalizability of AD biomarker relationships in diverse samples by considering matching on relevant health and demographic characteristics, and inclusion of additional cohorts with harmonized biomarker and cognitive measurements.
Association of neighborhood deprivation with Alzheimer's Disease pathologies, gray matter volume, white matter hyperintensities and cognition by sex, race or ethnicity and APOE4 status in Non‐demented Older Adults
Abstract Background Neighborhood socio‐economic conditions may increase Alzheimer's disease (AD) risk through amyloid, tau, atrophy or vascular pathways, potentially varying by genetic risk, race or ethnicity and sex. We examined associations between Area Deprivation Index (ADI) and cognition, AD pathology, gray matter (GM) and white matter hyperintensities (WMH) volume in cognitively unimpaired older adults by race or ethnicity, sex and APOE4 status. Method We included 1896 non‐demented older adults from US POINTER study with available cognitive evaluations, ADI level – a census–based neighborhood disadvantage metric (categorized as low, moderate and high) and a subsample with neuroimaging data ( N = 806). Linear/logistic regressions were used to assess interactions of ADI with sex, race or ethnicity, and APOE4 on (1) the Preclinical Alzheimer's Cognitive Composite, (PACC‐5), (2) MRI‐derived GM volume meta‐ROI (entorhinal, fusiform, parahippocampal, mid‐temporal, inferior temporal), (3) WMH and (4) AD pathology (amyloid and tau PET), adjusted by age, sex and intracranial volume when appropriate. Result African American participants ( p <0.001) and females ( p <0.001) were more likely to live in deprived areas. African American participants had significantly lower PACC‐5 ( p <0.001) and GM volume ( p <0.001), “Other or multiple races” participants showed significantly higher amyloid burden ( p = 0.001), and White participants higher WMH ( p = 0.013). Females showed higher PACC‐5 scores ( p <0.001), while APOE4 carriers showed lower PACC‐5 ( p = 0.042) and higher amyloid burden ( p <0.001) (Table 1). Race and ADI interacted on PACC‐5 ( p = 0.043) and showed a borderline interaction on Tau ( p = 0.079), with higher ADI linked to lower PACC‐5 in White, Hispanic, and “Other or Multiple Races” and higher Tau in Hispanic. Further, there was a borderline ADI*APOE4 status interaction on WMH ( p = 0.059) and a significant ADI*APOE4 interaction on amyloid ( p = 0.001), where APOE4 carriers in deprived areas showed higher WMH and lower amyloid burden (Figure 1). No significant ADI*sex interactions were observed. Conclusion Neighborhood disadvantage was more commonly observed in African American participants, but interaction with cognition was primarily seen in other racial/ethnic groups. APOE4 carriers showed increased amyloid burden, but those from more disadvantaged neighborhoods exhibited a higher vascular component alongside lower amyloid‐driven pathology. These results may indicate varying effects of social disadvantages across groups or sample bias.
Association of neighborhood social and built environment with physical activity in U.S. POINTER
Abstract Background Neighborhood social and built environment characteristics may promote physical activity, a modifiable factor for dementia risk, which may in turn promote cognitive and brain health. We explored the associations of the U.S. National Walkability Index ‐ a nationwide geographic data resource that ranks block groups according to their relative walkability – with cognitive function, gray matter (GM) and white matter hyperintensities (WMH), neighborhood disadvantage, sedentary behavior and frequency of physical activity in cognitively unimpaired older adults at increased Alzheimer's disease (AD) risk. Method We included baseline data from 1398 older adults (68.3 ± 5.17 years, 68.3% females, 30.5% APOE4 carriers) from the U.S. POINTER trial. Walkability was classified as: least, below average, above average and most walkable. Neighborhood disadvantage (low, moderate and high) was assessed using Area Deprivation Index (ADI) ( N = 1384). Physical inactivity was assessed using the Sedentary Behavior Questionnaire (SBQ, hours sitting/week, N = 1106), and physical activity was measured with a composite score from the Community Health Activities Model Program for Seniors (CHAMPS) questionnaire. Linear and logistic regressions examined associations between walkability and (1) physical activity and sedentary behavior; and (2) cognition assessed with the Preclinical Alzheimer's Cognitive Composite (PACC‐5), GM volume and WMHs, Additionally, associations between neighborhood disadvantage and physical activity and sedentary behavior were explored. Models were adjusted for age, sex, race, APOE4 and total intracranial volume. Result More deprived areas were generally less walkable (β = ‐0.008, p < 0.001) (Figure 1). Higher walkability (Above average: β = 0.192, p = 0.011 and most walkable: β = 0.183, p = 0.032) was associated with higher PACC‐5 (Figure 2A) scores but not with more preserved GM or lower WMH. Walkability scores were not associated with physical activity and sedentary measurements, while greater neighborhood disadvantage was linked to sedentary behavior (β = 0.126, p < 0.001) (Figure 2B). Conclusion These results suggest that physical inactivity in more deprived areas is associated with neighborhood socioeconomic disadvantages rather than built environment. Additionally, walkability scores were not associated with any measure of physical activity or brain health, suggesting the need to further investigate other health effects of neighborhood disadvantage and walkability.
Sociodemographic associations with AD biomarkers in POINTER, UBRAIN, and ADNI
Abstract Background In aging and dementia research, the relationships between modifiable risk factors and biomarkers often vary across cohorts. We aim to evaluate the consistency and generalizability of the association between the area deprivation index (ADI) ‐ an indicator of socioeconomic/resource disadvantage within a specific area ‐ and cognition and Alzheimer's disease (AD) biomarkers across three distinct research cohorts, two from the U.S. (U.S. POINTER Imaging, ADNI) and one from Spain (UBRAIN). Method We included a total of 3091 participants without dementia, 1879 from U.S. POINTER Imaging, 282 from ADNI and 930 from UBRAIN (Table 1). The ADI was calculated using validated indices from the US and Spain, with neighborhood deprivation ranked as Low, Moderate, and High in the US, and Low and High in Spain. Outcomes of interest included (1) the preclinical Alzheimer's cognitive composite (PACC‐5), (2) MRI‐derived gray matter volumes in AD‐vulnerable regions and (3) PET or CSF‐derived amyloid status (+/‐). We ran logistic regression models with ADI level as predictor adjusted by sex, age, and APOE4 and considered education and race or ethnicity (only US cohorts) in our analyses. Result Participants in more deprived areas had lower education across cohorts and were more likely to be female (U.S. POINTER) and African American (U.S. POINTER and ADNI, Table 1). Independent of age, sex and APOE4 status, higher ADI was associated with lower PACC in U.S. POINTER (Moderate: β=‐0.165, p <0.001; High β=‐0.412, p <0.001) and UBRAIN (ADI: High β=‐0.096, p = 0.022) but not in ADNI (Figure 1). After controlling for education, this association remained significant in U.S. POINTER (Moderate: β=‐0.166, p = 0.002; High: β=‐0.248, p = 0.007) but not in UBRAIN (β=‐0.021, p = 0.66). In U.S. POINTER, there was a trend towards an association of higher ADI with lower GM volume in AD‐vulnerable regions (β=‐0.126, p = 0.089) that was not observed in ADNI or UBRAIN (Figure 2). No associations between ADI and amyloid status were found across cohorts. Conclusion We observed differing associations in neighborhood deprivation and cognition across US cohorts, as well as between the US and Spain cohorts. Country‐specific factors and sample representativeness must be considered when investigating the pathophysiologic mechanisms underlying social determinants and disparities in AD.
Clinical Practice Guideline on Cognitive Assessments for the Early Detection of Cognitive Impairment in Primary Care: A report from the Alzheimer's Association
Abstract Background Cognitive impairment often goes undetected by clinicians due to time constraints, limited expertise, and lack of standardized tools. Structured cognitive assessments improve detection accuracy compared to unstructured clinical judgment. In 2013, the Alzheimer’s Association published recommendations to help primary care clinicians incorporate cognitive assessments into the Medicare Annual Wellness Visit. Since then, advances in cognitive assessment tools, availability of disease‐modifying therapies, and shifting expectations around early diagnosis have highlighted the need for updated guidance. In response, the Alzheimer’s Association is developing an evidence‐based clinical practice guideline for cognitive testing in primary care, using the GRADE approach, informed by a systematic review of the best available evidence. Method A multidisciplinary panel formulated a clinical question to inform the development of this guideline. A systematic review was conducted (search range: 1999‐2024) to evaluate the diagnostic accuracy of ten cognitive tests (5‐Cog, AD8, GPCOG, IQCODE, Mini‐Cog, MIS, MoCA, QDRS, RUDAS, and SLUMS) for the early detection of cognitive impairment (including MCI and dementia), in adults aged 55+ in ambulatory settings. The panel also narratively summarized data on emerging digital tools. Result Forty‐one studies meeting eligibility criteria were included in the systematic review. Results were analyzed separately by test (including short‐IQCODE and short‐MoCA) and for primary care and for ambulatory clinic settings (e.g., neurology, memory disorder clinics), as well as for English and Spanish language, resulting in 21 separate analyses. Only three analyses, all related to MoCA, allowed for meta‐analysis. The panel will use the evidence to inform conditional recommendations for or against each specific test. Conclusion By evaluating the diagnostic performance of selected tests, this guideline aims to support healthcare professionals in primary care settings in effective triaging of individuals who may benefit from further cognitive evaluation. To enhance real‐world applicability, considerations such as administration time, ease of scoring, and diagnostic accuracy ‐ along with appropriate cut‐offs by race/ethnicity, education, and language ‐ should be prioritized. This guideline will be regularly updated over time to reflect emerging evidence and the needs of diverse populations.
Abstract Background Ten years after the original publication, the Centiloid framework is now broadly used to harmonize amyloid‐PET quantification, facilitate data sharing and comparison across cohorts, and even assist visual interpretation in clinical settings. We evaluated the global implementation of Centiloids by comparing their distribution and corresponding positivity thresholds across cohorts. Methods We gathered data from publicly available cohorts and reached out to investigators across the world to collect cross‐sectional Centiloids, demographic and clinical information, and visual read data. Gaussian mixture models (GMM, k=2) were fitted to Centiloid values for each cohort and cutoffs were calculated as mean + 2SD of the lower Gaussian. When visual reads were available, we determined Centiloid cutoffs that maximized correspondence with visual reads (Cohen’s kappa). Data was combined across cohorts using random effects meta‐analyses. Results As of January 2025, we included 37 cohorts ( n = 41,678 participants) with heterogeneous pipelines, radiotracers, and clinical and demographic characteristics (Table‐1). The low Gaussian peaks ranged from ‐9 to 10CL; the meta‐analysis identified a common peak at 1CL. The second peaks were more heterogeneous (range=38‐102CL; meta‐analysis outcome=64CL). Across cohorts, the proportion of cognitively unimpaired versus impaired participants impacted the position of both peaks, with better separation in cohorts enriched in impaired individuals (Figure‐1C). The meta‐analysis indicated a GMM‐based cutoff of 19CL (95%CI: 16‐21CL, Figure‐1B); subgroup analyses showed no evidence of significant effect between single versus multicenter settings (17 versus 20CL, p = 0.30), MRI‐based or PET‐only processing (18 versus 19CL, p = 0.88), and no evidence of difference across radiotracers (Flutemetamol: 16CL; PIB: 17CL, Flutafuranol: 18CL, Florbetaben: 19CL, Florbetapir: 20CL, p = 0.78). In a subset of 29,496 participants with visual reads available, binary visual reads corresponded well to Centiloids (common kappa=0.86, Figure‐2A). The visual read‐based cutoff of 24CL (95%CI: 21‐27CL, Figure‐2B) maximized correspondence between visual read and quantification and was slightly higher than the GMM‐based cutoff. All meta‐analysis models showed high non‐random heterogeneity (I 2 >80%) across studies, suggesting non‐random differences in peaks and cutoffs. Conclusions Meta‐analysis‐based cutoffs align well with thresholds from the existing literature. High heterogeneity among studies underscores the need to investigate contributing factors, raising concerns about applying common cutoffs.