The physical and social exposome affects human aging, and brain clocks may track its effects. However, most studies neglect multidomain exposures (physical, social and political) across diverse settings globally and their associations with brain aging. In this study, we characterized the associations between 73 country-level physical and social exposomal factors and multimodal brain age in 18,701 participants from 34 countries (healthy individuals and those with Alzheimer's disease, frontotemporal lobar degeneration or mild cognitive impairment). Exposome effects were assessed using generalized additive models and meta-analytic frameworks. Aggregated exposome models explained up to 15.5-fold more variance than individual exposures (delta Akaike information criterion (ΔAIC): 2,034-3,127). Physical exposome was primarily associated with accelerated structural brain aging (limbic, subcortical and cerebellar regions), whereas social exposome was more strongly associated with functional brain aging (frontotemporal and limbic networks). Exposome burden accounted for 3.3-9.1-fold higher risk of accelerated aging, exceeding effects of clinical diagnoses. Findings were out-of-sample validated in cross-sectional and longitudinal designs, remained consistent across clinical subgroups and persisted after adjustment for demographics, age correction bias, cognition, scanner type and data quality. The exposome accelerates brain aging in health and disease, underscoring the need to address physical, social and political inequities.

BACKGROUND:Most people with dementia live in LMICs, underscoring the need for LMIC-specific identification of high-risk individuals. This study aimed to develop and validate a simple dementia risk prediction model for these settings. METHODS:Data from seven 10/66 Study sites were analyzed. Over 100 candidate predictors were screened based on existing models and the 2024 Lancet Commission, including LMIC-specific variables (eg, food insecurity and household assets). Predictors were selected using LASSO and modelled with the Fine-Gray method to generate a risk score. Predictive accuracy was pooled via meta-analysis. RESULTS:11143 participants were included, among whom 1069 (9.6%) developed dementia during follow-up. A five-factor risk score comprising age, social engagement, physical activity, hypertension, and difficulty in handling money was developed. The pooled c-statistic was 0.75 (95% CI: 0.72-0.78), with good calibration across sites. Decision curve analysis showed a modest net benefit, with variation across countries. CONCLUSION:It is possible to predict incident dementia with reasonable accuracy using a simple model across different LMICs. Our findings support the use of context-specific risk assessment tools to identify individuals at elevated dementia risk in LMIC settings, which may inform resource allocation for dementia care services and public health planning.

INTRODUCTION:Cognitive trajectories may clarify how type 2 diabetes (T2D) and impaired fasting glucose (IFG) relate to dementia risk, but longitudinal associations remain unclear, particularly in the context of stroke. METHODS:Data from 5,631 dementia- and stroke-free older adults (mean age 75 years) from 7 international population-based cohorts were analyzed. Linear mixed-effects models estimated cognitive trajectories during stroke-free and post-stroke follow-up. Glucose status was defined by fasting glucose and prior T2D diagnosis. RESULTS:Over 6.6 years of follow-up (4.5% with incident stroke), T2D was associated with lower baseline cognitive performance compared with normal fasting glucose (-0.14 SD, 95% CI -0.21 to -0.07), but not with faster cognitive decline during stroke-free or post-stroke follow-up. IFG was not associated with lower cognitive performance or faster decline. DISCUSSION:In older adults, T2D was associated with persistently lower cognitive performance but not faster decline, suggesting adverse cognitive effects may be established before late life.

Background Left-right hippocampal volumetric asymmetry and atrophy are implicated in neurodegenerative and neuropsychiatric disorders, yet their molecular basis in healthy adults remains poorly understood. Methods We conducted a meta-analysis of epigenome-wide association studies across six population-based cohorts (n = 8156; 53% women; mean age = 60.7 years) to identify DNA methylation signatures associated with left and right hippocampal volumes (LHCV, RHCV) and hippocampal asymmetry (i.e, differences between left and right volumes divided by their sums). Findings We identified five CpGs and 262 differentially methylated regions associated with LHCV, nine CpGs and 246 regions with RHCV, one CpG and 16 regions with asymmetry. Cross-omics integration uncovered 15 LHCV-related and 13 RHCV-related methylation-gene expression pairs, with five overlapping genes primarily involved in immune regulation. LHCV-specific genes were involved in cellular signalling, and Mendelian randomisation (MR) analyses supported a potential causal association between brain expression of DIP2C and increased risk of major depressive disorder. RHCV-specific genes were involved in neuronal differentiation pathways, with MR analyses suggesting that brain-tissue expression of BAIAP2, MACF1, SLC16A5, and CORO1B was associated with neuropsychiatric disorders. We also identified sex-specific patterns with hippocampal asymmetry. Notably, baseline methylation at these sites predicted hippocampal atrophy rates, explaining >10% of the variation. Associations with multiple healthy dietary patterns suggest modifiable influences on hippocampal structure. Interpretation These findings highlight distinct methylation profiles as potential biomarkers or therapeutic targets for neuropsychiatric and neurodegenerative conditions. Funding Institutional funds, Federal Ministry of Education and Research of Germany, Alzheimer's Association.

OBJECTIVES:To explore the characteristics of cognitive super-agers and determine which definition best explains variance in outcomes. METHOD:'Super-agers', defined as (i) performance in at least one cognitive domain comparable to adults in their 30s-40s and at least average performance in other domains (SA1), (ii) superior performance for age (SA2) or (iii) no cognitive decline over time (SA3), were compared with 'typical agers' in a sample of 1611 Australian adults aged 65-106. Logistic regression was used to determine which variables best discriminated between super-agers and typical agers for each definition. Separate generalised linear models were used to determine variance in functional, physical and psychosocial outcomes. RESULTS:SA1 (n = 459) and SA2 (n = 236) had higher education and occupational complexity, less hypertension and lower depression scores; SA3 (n = 266) had higher physical activity, lower depression and hearing impairment. All cohorts had greater brain volumes and white matter integrity. Super-ageing was associated with less functional impairment (10.1-16.1%, p < 0.001). SA1 and SA2 were associated with faster walking speed (5.4-6.4%, p < 0.001) and better psychological wellbeing (5-6.5%, p = 0.001-0.003). CONCLUSION:This study demonstrates differences in factors associated with either superior cognition in older age or maintaining cognitive abilities. SA1 and SA2 were most strongly associated with positive outcomes.

Ethnic disparities in cerebral small vessel disease (CSVD) have been reported, but no systematic synthesis has compared markers and vascular risk factors across populations. We conducted a PROSPERO-registered systematic review and meta-analysis (CRD42024518105) including over two million community-dwelling adults. A two-tier ethnicity framework was applied. Tier 1 used broad US Office of Management and Budget-aligned categories to maximize comparability, showing that Asians with higher cerebral microbleeds, Whites with greater metabolic risks, Blacks with higher hypertension, and Hispanics with elevated diabetes. Tier 2 enabled finer intra-Asian subgrouping, revealing Chinese cohorts with greater white matter hyperintensity (WMH) and hemorrhagic burden, Japanese with a lacunar-predominant profile but lower microbleeds, and Koreans with blood pressure-linked WMH. Ethnicity also moderated key risk-lesion associations, including stronger effects of diabetes on WMH and lacunes in Chinese and of systolic blood pressure on WMHs in Koreans. These findings highlight that CSVD phenotypes differ across populations, reflecting complex interactions between vascular, genetic, and environmental factors. HIGHLIGHTS: First meta-analysis of CSVD markers stratified by global ethnic groups. Tier 1: Asians had higher CMBs and lacunes; Whites had higher metabolic risk. Tier 2: Chinese showed higher WMHs and CMBs; Japanese had more lacunes, fewer CMBs. Risk-CSVD associations varied by ethnicity, including diabetes and blood pressure. Results suggest need for ethnicity-informed prevention and harmonised reporting.

OBJECTIVES:Higher levels of positive affect (PA) are consistently associated with lower depressive symptoms in older adults; however, the relative importance of specific PA components remain unclear. This study examined which discrete PA facets show the strongest associations with depressive symptoms and whether these associations reflect genetic or environmental influences using twin methodology and longitudinal network analysis. METHODS:Data were drawn from 540 older twins (mean age = 71 years, 65% female) participating in the Older Australian Twins Study across three waves spanning four years. PA was assessed using the Positive and Negative Affect Schedule (PANAS), and depressive symptoms were measured using the Geriatric Depression Scale (GDS-15). Additive genetic, common environmental, and unique environmental (ACE) models were applied to estimate variance components. Bayesian Gaussian Graphical Models were used to characterise affective networks, Directed Acyclic Graphs examined conditional directional associations, and hierarchical regression analyses tested longitudinal associations between baseline PA components and subsequent depressive symptoms. RESULTS:ACE modelling indicated that overall PA was predominantly influenced by environmental factors (53.1%), whereas depressive symptoms showed a stronger genetic contribution (59.5%). Across twin pairs and timepoints, network analyses consistently identified enthusiasm as the most central PA component. Enthusiasm and feeling active showed the strongest negative associations with depressive symptoms (partial correlations ranging from -0.12 to -0.18). Longitudinally, enthusiasm showed stronger concurrent associations with depressive symptoms, whereas feeling active demonstrated more sustained associations over the four-year follow-up. These patterns were consistent across monozygotic and dizygotic twin pairs. CONCLUSIONS:Specific PA states, particularly feeling enthusiastic and feeling active, show consistent associations with depressive symptoms in older adulthood and appear largely environmentally influenced. These findings highlight discrete emotional experiences that may be relevant targets for low-risk, non-pharmacological strategies aimed at supporting psychological well-being in aging populations, while complementing established approaches to depression prevention and treatment.

BACKGROUND:Cerebral small vessel disease (SVD) is a major cause of ischemic stroke, intracerebral hemorrhage, and dementia. Despite its importance, there are few studies of its prevalence and how cerebral SVD varies across the world, different age ranges, sexes, and magnetic resonance imaging (MRI) parameters. SVD can be estimated using MRI neuroimaging markers, including white matter hyperintensities (WMHs), lacunes, cerebral microbleeds (CMBs), and perivascular spaces (PVS). AIMS:This study aimed to document the global prevalence of SVD based on population-based or large community-based MRI studies and to determine how SVD prevalence varies by region, mean age, and sex. With SVD neuroimaging markers being the standard to assess SVD prevalence, we aimed to investigate how different MRI acquisition parameters may influence its prevalence. SUMMARY OF REVIEW:In this systematic review and meta-analysis, articles were searched from the Ovid MEDLINE and EMBASE databases between 1 January 2000 and 31 March 2024, without language restrictions. Title and abstract screening, full-text review, and data extraction were performed by at least two independent reviewers. The prevalence of SVD, subject demographic information, and MRI acquisition parameters were extracted. The Risk of Bias for Non-randomized Studies tool was used. The protocol was registered on PROSPERO (CRD42022311133). Of 14,582 studies identified, 246 studies spanning 40 countries were included in the systematic review. In the meta-analysis, 85 studies (88 cohorts) from 17 regions (n = 1,562,765) were included. The quality of studies was high (mean score 7.67 out of 8, ranging between 5 and 8). The pooled prevalence of moderate-to-severe WMH was 18.9%, and the pooled mean of WMH volume was 4.4 mL. Pooled prevalences of lacunes, cerebral microbleeds (CMBs), and moderate-to-severe perivascular spaces (PVS) were 11.2%, 10.3%, and 22.6%, respectively. A lower lacune prevalence (7.3% vs 13.3%; adjusted OR (aOR) [95% confidence interval (CI)]: 0.45 [0.30-0.68]) but higher PVS prevalence (30.9% vs 19.6%; aOR [95% CI]: 12.15 [2.12-69.46]) was found in Europe compared with Asia. A higher mean age of the studies was associated with a higher prevalence of most SVD markers, except for PVS. There was an overall trend of more lacunes and CMBs in males. MRI field strength, sequence used, and slice thickness could potentially influence the reported SVD prevalence, especially for WMH volume and CMB count. There was high heterogeneity in the studies (>95%) that was not resolved by performing analyses stratified by Global Burden of Disease (GBD) regions, age groups, study design, or MRI parameters. CONCLUSION:This systematic review and meta-analysis based on large MRI studies demonstrated that SVD is a common health problem affecting about one-fifth of the adult population. SVD prevalence differs in regions separated by geographical regions. SVD prevalence is higher with increasing age. There is an overall trend of more lacunes and CMBs in males. WMH volume and CMB are SVD markers prone to the variability of MRI acquisition parameters, and a harmonised SVD scanning protocol should be used. More studies from middle- and low-income regions would benefit the estimation of a truly global prevalence of SVD.

OBJECTIVES:Subjective wellbeing is a self-defined appraisal of happiness that includes judgments of life satisfaction and overall affect. Health-related quality of life is an appraisal of wellbeing focused on one's health status. Factors influencing these different aspects of wellbeing in old age remain unclear, with particular uncertainty around cognition. This study examined these relationships through assessments over 12 years. METHOD:One-thousand-and-thirty-seven people without dementia were recruited as part of a population-based cohort study. Participants completed a neuropsychological battery, medical history, and measures of life satisfaction, positive affect, health-related quality of life, depression, and anxiety biennially over 12 years. Participants also completed personality measures and informants completed measures of function. Linear mixed models were used to examine whether cognition and other variables were associated with wellbeing measures. RESULTS:Life satisfaction, positive affect, and health-related quality of life were moderately correlated at baseline. Cognition had a small, but statistically significant longitudinal association with each of these outcomes after adjusting for potential confounders. Physical health, function, personality, and anxiety also predicted wellbeing measures. CONCLUSION:Cognition is associated with subjective wellbeing and health-related quality of life in older people. Physical health, function, personality, and anxiety are also important predictors of wellbeing in old age.

Background Late-life depression (LLD) is associated with increased risk of cognitive decline and dementia, yet clinically relevant neuroimaging markers of short-term cognitive decline remain uncertain. Methods We conducted a two-stage study. In the Shanghai Action to Prevent the Elderly from Dementia (SHAPE) cohort, baseline MRI was compared among 40 LLD participants with stable cognition (LLDcs), 38 LLD participants with cognitive decline over two years (LLDcd), and 47 healthy controls. Measures included peak width of skeletonized mean diffusivity (PSMD), global and tract-specific difference in distribution function based on mean diffusivity (DDFMD), grey-matter measures, and regional diffusion tensor image analysis along the perivascular space (ALPS) indices. Within-group partial correlations between MRI measures and cognitive performance were also examined. Candidate markers were then evaluated in the Sydney Memory and Ageing Study (MAS) LLD sample (21 LLDcs, 16 LLDcd) using SHAPE-trained single-marker Firth logistic models applied without refitting. Results In SHAPE, LLDcd showed greater white-matter microstructural burden than LLDcs and controls, including higher PSMD, lower global and tract-specific DDFMD, alongside smaller left nucleus accumbens (L_NAcc) volume. Within LLDcd, poorer white-matter integrity was associated with worse orientation and other cognitive domains. In MAS, global DDFMD and L_NAcc volume were the two markers whose discrimination of LLDcd survived correction for multiple comparisons. Conclusions White-matter microstructural burden and smaller L_NAcc volume may help characterize LLD patients at higher near-term risk of cognitive decline. These findings warrant evaluation in larger longitudinal LLD cohorts.

AIM:Distressing dreams were previously reported to predict future all-cause dementia among predominantly white US participants aged 79-89 years, particularly in men. We investigated whether disturbing dreams (nightmares and bad dreams) were associated with all-cause and Alzheimer dementia (AD) among individuals aged 60-89 years from diverse international regions. METHODS:Data were from six longitudinal cohort studies across Brazil, China, France, Italy, South Korea, and Taiwan (n = 10,238, 42.5% men). Cox regressions with a random effect for study investigated associations between disturbing dreams and incident dementia, with all participants and stratified separately by sex and baseline age. Analyses examined (i) any disturbing dreams and (ii) disturbing dreams at least once a week. Fully adjusted analyses included three studies with covariates for sleep problems, medications, mental and physical health, cognition, and APOE ε4 status. RESULTS:Disturbing dreams were reported by 24.2% overall and all-cause dementia, and AD incidence was 10.8 and 5.3 per 1000 person-years, respectively. In fully adjusted analyses, having any disturbing dreams was associated with increased incidence of all-cause dementia among 60-69-year-olds (hazard ratio [HR] 3.93, 95% confidence interval [CI] 1.32-11.67). There were no significant effects for older individuals. In fully adjusted sex-stratified analyses, having disturbing dreams at least once a week was associated with AD only among men (HR 3.59, 95% CI 1.44-8.96). CONCLUSIONS:We found some evidence for disturbing dreams being associated with incident all-cause dementia among individuals aged 60-69 years and with AD among men. The mechanisms potentially underlying these associations remain to be clarified.

Background: Fermented dairy foods, such as yogurt and cheese, contain bioactive components that differ from those in non-dairy foods, but their associations with depression and dementia risk in later life remain unclear. Methods: We analyzed data from the Sydney Memory and Ageing Study, a community-dwelling cohort of adults aged 70-90 years, to examine associations between dairy intake and depressive symptoms (Geriatric Depression Scale-15), psychological distress (Kessler-10), and incident depression (physician diagnosis or antidepressant use) and dementia (DSM-IV criteria). Intake of yogurt, cheese, and non-fermented milk was assessed at baseline using a validated food-frequency questionnaire. Longitudinal associations were examined using Fine-Gray competing-risks models that accounted for death; cross-sectional associations were also assessed. Results: Among 966 participants (mean age: 78.3; 55.5% women), compared with no consumption, higher yogurt intake (one standard serving) was significantly associated with lower depressive symptom scores (adjusted β: -0.37 and -0.39 for quartiles 3-4 (mean: 88.5-164 g/day), and so was low-fat cheese intake (mean: 13.2 g/day) (adjusted β: -0.35). Over a mean follow-up of 3.3 years, 120 incident cases of depression and 68 deaths occurred: higher yogurt intake and low-fat cheese consumption (versus non-consumption) were associated with lower risk of depression (adjusted subdistribution hazard ratios 0.41 [95% CI 0.19-0.88] and 0.40 [0.21-0.78], respectively). No significant associations were observed for psychological distress, cognition, or incident dementia (a mean follow-up of 5.2 years, 100 incident cases, and 153 deaths); no associations were observed for regular cheese or milk intake. Conclusions: These findings suggest a potential role for fermented dairy foods, particularly yogurt and low-fat cheese intake, but not non-fermented milk, in mental well-being in later life.

Genetic susceptibility to Alzheimer's disease (AD) may influence the extent to which environmental factors shape cognition, with individuals at higher genetic risk potentially exhibiting greater sensitivity to environmental exposures. Sleep, an important factor for both cognitive function and AD risk, may further moderate genetic influences (A), including both measured (AP) and latent (AL) components, as well as shared (C) and non-shared environmental (E) contributions to cognition. This study leveraged data from the Interplay of Genes and Environment across Multiple Studies consortium (N = 3894; 1947 complete twin pairs, 842 monozygotic (MZ) pairs and 1105 dizygotic (DZ); Average age = 62.36 years, 38.75% female). Across six cognitive abilities, we examined whether an AD polygenic score (AD-PGS) moderated environmental influences on cognitive performance. We also examined whether sleep moderated genetic and environmental contributions on cognitive performance. Although the AD-PGS accounted for a negligible proportion of genetic variance as a main effect (B's = -0.004 to 0.02), we observed environment-by-PGS interactions. Increasing genetic risk for AD was associated with lower contributions from environmental experiences unique to each individual, on episodic memory, working memory, and verbal ability (B's = -0.03 to -0.05). These interaction effects, albeit small, were primarily observed with the AD-PGS including the APOE region. Hence, the role of person-specific environments on cognitive functioning was boosted for those at lower genetic risk for AD but reduced at greater genetic risk for AD. Although sleep moderation was minimal, results suggest that poorer sleep influences genetic influences on cognitive functioning. Statement of Significance This study provides novel insights into how genetic risk for Alzheimer's disease (AD-PGS), and sleep (duration and disturbances) moderate genetic and environmental contributions to cognitive performance. Although the AD-PGS accounts for minimal genetic variance, significant but small environment-by-PGS interactions emerged. Individuals with higher AD-PGS exhibited heightened responsiveness to nonshared environmental influences on episodic memory, working memory, and verbal ability, primarily driven by the APOE ε4 region. Poorer sleep may be associated with subtle variations in genetic contributions to cognition, though moderation effects were minimal. This work offers insight into the genetic mechanisms underlying sleep, AD risk, and cognition, underscoring the need to consider both genetic susceptibility and environmental influences when examining factors that contribute to individual differences in cognition.

Abstract Blood-based biomarkers could transform Alzheimer disease (AD) detection by enabling scalable, less invasive assessment of underlying pathology, yet their applicability across globally diverse populations remains uncertain. We systematically reviewed 168 publications comprising 139 independent cohorts from Asia, Europe, North America, South America, Africa and Oceania. Random-effects meta-analyses pooled log-transformed ratios of mean biomarker concentrations for AD versus cognitively unimpaired individuals, mild cognitive impairment versus cognitively unimpaired individuals, and amyloid-β PET-positive versus amyloid-β PET-negative individuals. p-tau217, p-tau181 and glial fibrillary acidic protein showed the largest and most consistent group differences. In clinically defined comparisons, p-tau181 separation in mild cognitive impairment was lower in predominantly Asian than White cohorts, whereas glial fibrillary acidic protein separation in AD was higher in predominantly Asian cohorts. No significant between-population differences were observed in amyloid-defined comparisons. These findings support leading blood biomarkers as globally relevant indicators of AD pathology, but rigorous harmonized validation is needed before thresholds can be translated into equitable clinical practice.

Post-stroke depression (PSD) affects approximately one-third of stroke survivors and is associated with poorer rehabilitation outcomes, increased disability, and higher mortality. Despite its clinical importance, PSD is often studied as a binary outcome, obscuring heterogeneity in symptom trajectories over time. This study aimed to identify replicable longitudinal trajectories of PSD across independent cohorts and examine baseline predictors of trajectory membership. Data were analysed from three cohorts of the Stroke and Cognition Consortium (STROKOG; n=750). Depression was measured using the Geriatric Depression Scale (GDS-15) and the Hamilton Depression Rating Scale (HAMD-17). Latent class growth analysis was conducted within cohorts and in pooled GDS-15 and HAMD-17 datasets. Across six analyses, a consistent three-class solution emerged. The majority of participants (58–83%) showed no clinically significant depressive symptoms (No Depression). A second class (14–29%) displayed mild symptoms that generally remitted over time (Mild Remitting). A smaller subgroup (4–13%) exhibited symptoms in the moderate-severity range throughout follow-up, with trajectory direction varying across analyses. Baseline global cognitive impairment predicted Mild Remitting class membership across both pooled analyses. In pooled HAMD-17 analyses, female sex predicted membership in the Moderate Improving class, characterised by moderately severe symptoms that gradually resolved over time. In pooled GDS-15 analyses, older age, hypertension, and diabetes predicted membership in the Moderate Stable class. PSD follows a robust three-trajectory structure across independent cohorts, countries, and instruments. Early clinical factors, particularly cognitive impairment, sex, and vascular risk factors, may identify stroke survivors at risk for persistent depression, enabling earlier targeted intervention.

A growing number of older adults now live in cities, raising important questions about how different features of urban built environments influence brain and cognitive health. Here we examine whether living in well-interconnected neighborhoods impacts the structure of the hippocampus in older adults. We analyze structural brain imaging and residential environment data from over 500 community-dwelling older adults (>70 years old) living in Sydney, with each participant having up to three repeated brain imaging assessments over a 6-year period. For each participant, we quantify neighborhood connectivity using the street intersection density within walkable network buffers surrounding their home and measure the volumes of the hippocampal head, body and tail. We find that older adults living in highly connected neighborhoods have a larger hippocampus tail, a brain area involved in spatial navigation. However, they also show a steeper decline in hippocampal tail volume over time, with a slight rebound beyond the age of 85. These results can help inform the design of age-friendly street networks that better support brain health.

BACKGROUND:Frailty and social isolation are independently associated with dementia risk. This study examined whether frailty predicts dementia differently for older adults living alone versus with others. METHODS:Participants were 967 community-dwelling, dementia-free older adults from the Sydney Memory and Ageing Study (2005-2020). Frailty was assessed using an adapted Fried Frailty Phenotype (robust, pre-frail, frail). Dementia was diagnosed via biennial neuropsychological testing and clinical consensus. Cox regressions analysed associations between baseline frailty, living arrangements (alone vs. with others), and incident dementia over 12 years. RESULT:Frail individuals had 1.74 times higher dementia risk than robust individuals (HR = 1.74, 95% CI: 1.16-2.60, p = .007). Living alone was not independently associated with dementia (HR = 1.03, p = .825). The interaction between frailty and living arrangements was non-significant (HR = 1.23, p = 0.600). Exploratory stratified analyses showed frailty predicted dementia among those living alone (HR = 1.92, 95% CI: 1.06-3.51, p = .033) but not those living with others (HR = 1.42, p = 0.224). CONCLUSION:Frailty is a significant predictor of dementia risk in older adults. While the formal interaction test was non-significant, exploratory findings suggest frail individuals living alone may be particularly vulnerable, warranting further investigation and potential targeting for combined physical and social interventions, pending replication.

Background and ObjectivesDementia risk prediction models developed for the general population perform poorly in stroke cohorts. Existing stroke-specific models are few and limited by short prediction horizons or reliance on neuroimaging. The aim of this study was to develop a clinically practical model for predicting 5-year dementia risk after stroke using commonly available variables and individual participant data from the Stroke and Cognition Consortium (STROKOG).MethodsData were pooled from 12 studies across 10 countries. Dementia was diagnosed mainly by expert panel consensus and algorithmic classification. Fine-Gray subdistribution hazard models estimated dementia probability, accounting for death as a competing event. Candidate predictors included routinely collected baseline clinical and stroke-related variables, selected through backward stepwise elimination. Model performance was evaluated using discrimination (C-index) and calibration for prediction up to 5 years after stroke. Internal-external cross-validation (IECV) assessed generalizability across studies, regions, and study periods.ResultsA total of 2,663 participants (mean age 67.0 years [SD 11.1]; 40% female) were followed for a median of 2.0 years (IQR 1.0-5.0), during which 655 developed dementia (8.7 per 100 person-years). The final model included age, sex, education, history of previous stroke, diabetes, stroke severity, 2 interactions (age x sex; age x stroke severity), and study-level variables including national current health expenditure. An Excel-based risk calculator is available in the Supplement (eAppendix 1). The model demonstrated strong discrimination (C-index: 0.81; 95% CI 0.75-0.87) and excellent calibration in the full data set used for development. In IECV, discrimination was acceptable across individual studies (pooled C-index: 0.70 [0.67-0.73]) and higher in recent (post-2010; 0.79 [0.76-0.82]) and European (0.74 [0.71-0.78]) cohorts. Risks were slightly overestimated in Asian cohorts. Case numbers were too small for reliable assessment in other regions.DiscussionWe developed and internally-externally validated a 5-year dementia risk model for stroke survivors using routinely available clinical variables. The model showed strong performance in the full development data set and generalized well to recent and European cohorts, although external validation in diverse populations is needed. This tool can help identify high-risk individuals for targeted cognitive monitoring and follow-up. By informing clinical decision making and resource planning, it offers a practical means to improve long-term outcomes.