Incident depression and mortality among people with different types of dementia: results from a longitudinal cohort study

Machine learning models identify predictive features of patient mortality across dementia types

BACKGROUND

Dementia care is challenging due to the divergent trajectories in disease progression and outcomes. Predictive models are needed to flag patients at risk of near-term mortality and identify factors contributing to mortality risk across different dementia types.

METHODS

Here, we developed machine-learning models predicting dementia patient mortality at four different survival thresholds using a dataset of 45,275 unique participants and 163,782 visit records from the U.S. National Alzheimer's Coordinating Center (NACC). We built multi-factorial XGBoost models using a small set of mortality predictors and conducted stratified analyses with dementiatype-specific models.

RESULTS

Our models achieved an area under the receiver operating characteristic curve (AUC-ROC) of over 0.82 utilizing nine parsimonious features for all 1-, 3-, 5-, and 10-year thresholds. The trained models mainly consisted of dementia-related predictors such as specific neuropsychological tests and were minimally affected by other age-related causes of death, e.g., stroke and cardiovascular conditions. Notably, stratified analyses revealed shared and distinct predictors of mortality across eight dementia types. Unsupervised clustering of mortality predictors grouped vascular dementia with depression and Lewy body dementia with frontotemporal lobar dementia.

CONCLUSIONS

This study demonstrates the feasibility of flagging dementia patients at risk of mortality for personalized clinical management. Parsimonious machine-learning models can be used to predict dementia patient mortality with a limited set of clinical features, and dementiatype-specific models can be applied to heterogeneous dementia patient populations.

Alzheimer's polygenic risk scores, APOE, Alzheimer's disease risk, and dementia-related blood biomarker levels in a population-based cohort study followed over 17 years

BACKGROUND

In order to utilize polygenic risk scores (PRSs) for Alzheimer's disease (AD) in a meaningful way, influential factors (i.e. training set) and prediction across groups such as APOE e4 (APOE4) genotype as well as associations to dementia-related biomarkers should be explored. Therefore, we examined the association of APOE4 and various PRSs, based on training sets that utilized differing AD definitions, with incident AD and all-cause dementia (ACD) within 17 years, and with levels of phosphorylated tau181 (P-tau181), neurofilament light (NfL), and glial fibrillary acidic protein (GFAP) in blood. Secondarily, effect modification by APOE4 status and sex was examined.

METHODS

In this prospective, population-based cohort study and nested case-control study, 9,940 participants in Germany were enrolled between 2000 and 2002 by their general practitioners and followed for up to 17 years. Participants were included in this study if dementia status and genetic data were available. A subsample of participants additionally had measurements of P-tau181, NfL, and GFAP obtained from blood samples. Cox and logistic regression analyses were used to assess the association of genetic risk (APOE genotype and PRSnoAPOE) with incident ACD/AD and log-transformed blood levels of P-tau181, NfL, and GFAP.

RESULTS

Five thousand seven hundred sixty-five participants (54% female, aged 50-75years at baseline) were included in this study, of whom 464 received an all-cause dementia diagnosis within 17 years. The PRSs were not more predictive of dementia than APOE4. An APOE4 specific relationship was apparent with PRSs only exhibiting associations to dementia among APOE4 carriers. In the nested case-control study including biomarkers (n = 712), APOE4 status and polygenic risk were significantly associated to levels of GFAP in blood.

CONCLUSIONS

The use of PRSs may be beneficial for increased precision in risk estimates among APOE4 carriers. While APOE4 may play a crucial etiological role in initial disease processes such as Aβ deposition, the PRS may be an indicator of further disease drivers as well as astrocyte activation. Further research is necessary to confirm these findings, especially the association to GFAP.

Association of depressive disorders and dementia with mortality among older people with hip fracture

BACKGROUND

Hip fracture (HF) is a significant cause of mortality among older people. Almost half of the patients with HF have dementia, which increases the mortality risk further. Cognitive impairment is associated with depressive disorders (DDs) and both dementia and DDs are independent risk factors for poor outcome after HF. However, most studies that evaluate mortality risk after HF separate these conditions.

AIMS

To investigate whether dementia with depressive disorders (DDwD) affects the mortality risk at 12, 24, and 36 months after HF among older people.

METHODS

Patients with acute HF (n = 404) were included in this retrospective analysis of two randomized controlled trials performed in orthopedic and geriatric departments. Depressive symptoms were assessed using the Geriatric Depression Scale and cognitive function was assessed using the Mini-Mental State Examination. A consultant geriatrician made final depressive disorder and dementia diagnoses using the Diagnostic and Statistical Manual of Mental Disorders criteria, with support from assessments and medical records. The 12-, 24- and 36-month mortality after HF was analyzed using logistic regression models adjusted for covariates.

RESULTS

In analyses adjusted for age, sex, comorbidity, pre-fracture walking ability, and fracture type, patients with DDwD had increased mortality risks at 12 [odds ratio (OR) 4.67, 95% confidence interval (CI) 1.75-12.51], 24 (OR 3.61, 95% CI 1.71-7.60), and 36 (OR 4.53, 95% CI 2.24-9.14) months. Similar results were obtained for patients with dementia, but not depressive disorders, alone.

CONCLUSION

DDwD is an important risk factor for increased mortality at 12, 24, and 36 months after HF among older people. Routinely assessments after HF for cognitive- and depressive disorders could identify patients at risk for increased mortality, and enable early interventions.

TRIAL REGISTRATION

RCT2: International Standard Randomized Controlled Trial Number Register, trial registration number: ISRCTN15738119.

Amyloid-beta misfolding and GFAP predict risk of clinical Alzheimer's disease diagnosis within 17 years

INTRODUCTION

Blood-based biomarkers for Alzheimer's disease (AD) are urgently needed. Here, four plasma biomarkers were measured at baseline in a community-based cohort followed over 17 years, and the association with clinical AD risk was determined.

METHODS

Amyloid beta (Aβ) misfolding status as a structure-based biomarker as well as phosphorylated tau 181 (P-tau181), glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) concentration levels were determined at baseline in heparin plasma from 68 participants who were diagnosed with AD and 240 controls without dementia diagnosis throughout follow-up.

RESULTS

Aβ misfolding exhibited high disease prediction accuracy of AD diagnosis within 17 years. Among the concentration markers, GFAP showed the best performance, followed by NfL and P-tau181. The combination of Aβ misfolding and GFAP increased the accuracy.

DISCUSSION

Aβ misfolding and GFAP showed a strong ability to predict clinical AD risk and may be important early AD risk markers. Aβ misfolding illustrated its potential as a prescreening tool for AD risk stratification in older adults.

Predictive Models and Features of Patient Mortality across Dementia Types

Dementia care is challenging due to the divergent trajectories in disease progression and outcomes. Predictive models are needed to identify patients at risk of near-term mortality. Here, we developed machine learning models predicting survival using a dataset of 45,275 unique participants and 163,782 visit records from the U.S. National Alzheimer's Coordinating Center (NACC). Our models achieved an AUC-ROC of over 0.82 utilizing nine parsimonious features for all one-, three-, five-, and ten-year thresholds. The trained models mainly consisted of dementia-related predictors such as specific neuropsychological tests and were minimally affected by other age-related causes of death, e.g., stroke and cardiovascular conditions. Notably, stratified analyses revealed shared and distinct predictors of mortality across eight dementia types. Unsupervised clustering of mortality predictors grouped vascular dementia with depression and Lewy body dementia with frontotemporal lobar dementia. This study demonstrates the feasibility of flagging dementia patients at risk of mortality for personalized clinical management.

Association of Kidney Function With Development of Alzheimer Disease and Other Dementias and Dementia-Related Blood Biomarkers

IMPORTANCE

Previous research has suggested an association of kidney function with risk of Alzheimer disease (AD) or other dementias and dementia-related blood biomarkers, but a distinct association remains unclear.

OBJECTIVE

To evaluate the association of kidney function with risk of diagnosis of incident AD or dementia within 17 years and with the blood biomarkers neurofilament light (NfL), phosphorylated tau181 (p-tau181), and glial fibrillary acidic protein (GFAP).

DESIGN, SETTING, AND PARTICIPANTS

In this prospective, population-based cohort study and nested case-control study, 9940 participants in Germany were enrolled between 2000 and 2002 by their general practitioners and followed up for up to 17 years. Participants were included if information on dementia status and creatinine/cystatin C measurements were available. A subsample of participants additionally had measurements of NfL, p-tau181, and GFAP obtained from blood samples. Statistical analysis was performed from January 3 to November 25, 2022.

EXPOSURES

Impaired kidney function, based on estimated glomerular filtration rate less than 60 mL/min/1.73 m2 according to the 2021 Chronic Kidney Disease Epidemiology Collaboration creatinine-cystatin C equation.

MAIN OUTCOMES AND MEASURES

All-cause dementia, AD, and vascular dementia diagnosis, as well as log-transformed levels of NfL, p-tau181, and GFAP in blood.

RESULTS

Of 6256 participants (3402 women [54.4%]; mean [SD] age at baseline, 61.7 [6.6] years), 510 received an all-cause dementia diagnosis within 17 years of baseline. The dementia-related blood biomarker nested case-control sample included 766 participants. After adjusting for age and sex, impaired kidney function at baseline was not associated with a higher risk of all-cause dementia (hazard ratio [HR], 0.95; 95% CI, 0.69-1.29), AD (HR, 0.94; 95% CI, 0.55-1.63), or vascular dementia diagnosis (HR, 1.06; 95% CI, 0.65-1.70) within 17 years. In the cross-sectional analysis, after adjusting for age and sex, impaired kidney function was significantly associated with NfL and p-tau181 levels in blood (NfL: β = 0.47 and P < .001; p-tau181: β = 0.21 and P = .003). After adjusting for age and sex, significant associations with GFAP levels were evident only among men (men: β = 0.31 and P = .006; women: β = -0.12 and P = .11).

CONCLUSIONS AND RELEVANCE

In this population-based study of community-dwelling adults, reduced kidney function was associated with increased levels of dementia-related blood biomarkers but not increased dementia risk. Kidney function might influence the accuracy of dementia-related blood biomarkers and should be considered in clinical translation.

Association of plasma biomarkers, p-tau181, glial fibrillary acidic protein, and neurofilament light, with intermediate and long-term clinical Alzheimer's disease risk: Results from a prospective cohort followed over 17 years

INTRODUCTION

Blood biomarkers for Alzheimer's disease (AD) are the future of AD risk assessment. The aim of this study was to determine the association between plasma-measured phosphorylated tau (p-tau181), glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) levels and risk of clinical AD incidence with consideration to the impact of cardiovascular health.

METHODS

Within a community-based cohort, biomarker levels were measured at baseline using single molecule array technology in 768 participants (aged 50-75) followed over 17 years. Associations among biomarkers and AD, vascular dementia, and mixed dementia incidence were assessed.

RESULTS

GFAP was associated with clinical AD incidence even more than a decade before diagnosis (9-17 years), while p-tau181 and NfL were associated with more intermediate AD risk (within 9 years). Significant interaction was detected between cardiovascular health and p-tau181/NfL.

DISCUSSION

GFAP may be an early AD biomarker increasing before p-tau181 and NfL and the effect modifying role of cardiovascular health should be considered in biomarker risk stratification.

Long-term low-dose acetylsalicylic use shows protective potential for the development of both vascular dementia and Alzheimer's disease in patients with coronary heart disease but not in other individuals from the general population: results from two large cohort studies

BACKGROUND

No population-based cohort study investigated a potential inverse association between long-term low-dose acetylsalicylic acid (ASA) use and all-cause dementia and its two most common sub-types Alzheimer's disease (AD) and vascular dementia (VD) so far.

METHODS

Cox regression models with inverse probability of treatment weighting to model the underlying cardiovascular risk were used to assess the associations of low-dose ASA use with all-cause dementia, AD, and VD incidence in community-dwelling older adults from the German ESTHER study (N = 5258) and the UK Biobank (N = 305,394). Inclusion criteria were age of 55 years or older and completed drug assessment. Meta-analyses of the individual participant data from the two prospective cohort studies were performed.

RESULTS

Four hundred seventy-six cases of all-cause dementia, 157 cases of AD, and 183 cases of VD were diagnosed over a median of 14.3 years of follow-up in ESTHER. In the UK Biobank, 5584 participants were diagnosed with all-cause dementia, 2029 with AD, and 1437 with VD over a median of 11.6 years. The meta-analysis of both cohorts revealed a weak reduction in hazards for all-cause dementia (hazard ratio (HR) [95% confidence interval (CI)]: 0.96 [0.93 to 0.99]). The strongest protective effect of low-dose ASA was observed in participants with coronary heart disease (CHD) in both cohorts, and a significant interaction was detected. In particular, in meta-analysis, a 31% reduction in hazard for AD, 69% for VD and 34% for all-cause dementia were observed (HR [95% CI]: 0.69 [0.59 to 0.80], 0.31 [0.27 to 0.35], 0.46 [0.42 to 0.50], respectively). Furthermore, compared to non-users, users of low-dose ASA for 10 years or longer (who likely use it because they have CHD or a related diagnosis putting them at an increased risk for cardiovascular events) demonstrated a strong protective effect on all dementia outcomes, especially for VD (HR [95% CI]: 0.48 [0.42 to 0.56]) whereas no protective associations were observed with shorter low-dose ASA use.

CONCLUSIONS

The protective potential of low-dose ASA for all-cause dementia, AD, and VD seems to strongly depend on pre-existing CHD and the willingness of patients to take it for a minimum of ten years.

Weight loss in Alzheimer's disease, vascular dementia and dementia with Lewy bodies: Impact on mortality and hospitalization by dementia subtype

OBJECTIVES

Loss of weight is associated with cognitive decline as well as several adverse outcomes in dementia. The aim of this study was to assess whether weight loss is associated with mortality and hospitalization in dementia subtypes.

METHODS

A cohort of 11,607 patients with dementia in Alzheimer's disease (AD), vascular dementia (VD), or dementia with Lewy bodies (DLB) was assembled from a large dementia care health records database in Southeast London. A natural language processing algorithm was developed to established whether loss of weight was recorded around the time of dementia diagnosis. Cox proportional hazard models were applied to examine the associations of reported weight loss with mortality and emergency hospitalization.

RESULTS

Weight loss around the time of dementia was recorded in 25.5% of the whole sample and was most common in patients with DLB. A weight loss-related increased risk for mortality was detected after adjustment for confounders (Hazard ratio (HR):1.07; 95% confidence interval (CI):1.02-1.15) and in patients with AD (HR: 1.11; 95% CI: 1.04-1.20), but not in DLB and VD. Weight loss was associated with a significantly increased emergency hospitalization risk (HR: 1.14; 95% CI: 1.08-1.20) and in all three subtypes.

CONCLUSIONS

While there were associations with increased hospitalization risk for all three subtype diagnoses, weight loss was only associated with increased mortality in AD. Weight loss should be considered as an accompanying symptom in dementia and interventions should be considered to ameliorate risk of adverse outcomes.

Prediction of clinical diagnosis of Alzheimer's disease, vascular, mixed, and all-cause dementia by a polygenic risk score and APOE status in a community-based cohort prospectively followed over 17 years

The strongest genetic risk factor for Alzheimer's disease (AD) is the ε4 allele of Apolipoprotein E (APOE) and recent genome-wide association meta-analyses have confirmed additional associated genetic loci with smaller effects. The aim of this study was to investigate the ability of an AD polygenic risk score (PRS) and APOE status to predict clinical diagnosis of AD, vascular (VD), mixed (MD), and all-cause dementia in a community-based cohort prospectively followed over 17 years and secondarily across age, sex, and education strata. A PRS encompassing genetic variants reaching genome-wide significant associations to AD (excluding APOE) from the most recent genome-wide association meta-analysis data was calculated and APOE status was determined in 5203 participants. During follow-up, 103, 111, 58, and 359 participants were diagnosed with AD, VD, MD, and all-cause dementia, respectively. Prediction ability of AD, VD, MD, and all-cause dementia by the PRS and APOE was assessed by multiple logistic regression and receiver operating characteristic curve analyses. The PRS per standard deviation increase in score and APOE4 positivity (≥1 ε4 allele) were significantly associated with greater odds of AD (OR, 95% CI: PRS: 1.70, 1.45-1.99; APOE4: 3.34, 2.24-4.99) and AD prediction accuracy was significantly improved when adding the PRS to a base model of age, sex, and education (ASE) (c-statistics: ASE, 0.772; ASE + PRS, 0.810). The PRS enriched the ability of APOE to discern AD with stronger associations than to VD, MD, or all-cause dementia in a prospective community-based cohort.

Depression as a Risk Factor for Dementia and Alzheimer's Disease

Preventing the onset of dementia and Alzheimer’s disease (AD), improving the diagnosis, and slowing the progression of these diseases remain a challenge. The aim of this study was to elucidate the association between depression and dementia/AD and to identify possible relationships between these diseases and different sociodemographic and clinical features. In this regard, a case-control study was conducted in Spain in 2018–2019. The definition of a case was: A person ≥ 65 years old with dementia and/or AD and a score of 5–7 on the Global Deterioration Scale (GDS). The sample consisted of 125 controls; among the cases, 96 had dementia and 74 had AD. The predictor variables were depression, dyslipidemia, type 2 diabetes mellitus, and hypertension. The results showed that depression, diabetes mellitus, and older age were associated with an increased likelihood of developing AD, with an Odds Ratio (OR) of 12.9 (95% confidence interval (CI): 4.3–39.9), 2.8 (95% CI: 1.1–7.1) and 1.15 (95% CI: 1.1–1.2), respectively. Those subjects with treated dyslipidemia were less likely to develop AD (OR 0.47, 95% CI: 0.22–1.1). Therefore, depression and diabetes mellitus increase the risk of dementia, whereas treated dyslipidemia has been shown to reduce this risk.

Depression in People With Dementia and Caregiver Outcomes: Results From the European Right Time Place Care Study

OBJECTIVE

To investigate the cross-sectional associations between depression in people with dementia and both caregiver burden and quality of life in 8 European countries, and to test these associations compared with the presence of other neuropsychiatric symptoms.

DESIGN

Cross-sectional study.

SETTING AND PARTICIPANTS

In total, 1223 dyads comprised of informal caregivers and people with dementia living in a community-dwelling setting, recruited from the Right Time Place Care study, a cohort survey from 8 European countries.

MEASURES

To test the associations between depression (according to the Cornell Scale for Depression in Dementia) and informal caregiver burden (defined by the Zarit scale and hours of supervision in terms of Resource Utilization in Dementia), distress (defined by the Neuropsychiatric Inventory Questionnaire distress score), and quality of life (according to the visual analogue scale and 12-item General Health Questionnaire).

RESULTS

Linear regressions showed an association between depression and main outcomes (Zarit scale: β 3.7; P = .001; hours of supervision: β 1.7; P = .004; Neuropsychiatric Inventory Questionnaire distress score: β 1.2; P = .002). A similar association was found concerning psychological and overall well-being (12-item General Health Questionnaire: β 1.8; P < .001; Euroqol Visual Analogue Scale: β -4.1; P = .003). Both associations remained significant despite the presence of other NPS and after adjusting for confounders.

CONCLUSIONS AND IMPLICATIONS

Further studies are needed to assess whether providing tailored strategies for optimizing diagnosis and managing of depression in people with dementia might improve caregiver quality of life and reduce their burden in the community-dwelling setting.