Plasma metabolic profiles predict future dementia and dementia subtypes: a prospective analysis of 274,160 participants

Abstract

Background: Blood-based biomarkers for dementia are gaining attention due to their non-invasive nature and feasibility in regular healthcare settings. Here, we explored the associations between 249 metabolites with all-cause dementia (ACD), Alzheimer's disease (AD), and vascular dementia (VaD) and assessed their predictive potential.

Methods: This study included 274,160 participants from the UK Biobank. Cox proportional hazard models were employed to investigate longitudinal associations between metabolites and dementia. The importance of these metabolites was quantified using machine learning algorithms, and a metabolic risk score (MetRS) was subsequently developed for each dementia type. We further investigated how MetRS stratified the risk of dementia onset and assessed its predictive performance, both alone and in combination with demographic and cognitive predictors.

Results: During a median follow-up of 14.01 years, 5274 participants developed dementia. Of the 249 metabolites examined, 143 were significantly associated with incident ACD, 130 with AD, and 140 with VaD. Among metabolites significantly associated with dementia, lipoprotein lipid concentrations, linoleic acid, sphingomyelin, glucose, and branched-chain amino acids ranked top in importance. Individuals within the top tertile of MetRS faced a significantly greater risk of developing dementia than those in the lowest tertile. When MetRS was combined with demographic and cognitive predictors, the model yielded the area under the receiver operating characteristic curve (AUC) values of 0.857 for ACD, 0.861 for AD, and 0.873 for VaD.

Conclusions: We conducted the largest metabolome investigation of dementia to date, for the first time revealed the metabolite importance ranking, and highlighted the contribution of plasma metabolites for dementia prediction.

Keywords: Alzheimer’s disease; Dementia; Metabolomics; Plasma; Prediction; Vascular dementia.

Fig. 1

Dementia-associated metabolites in the association analysis. Significant associations (FDR–corrected Q value < 0.05) are shown, with red and blue colors respectively indicate the positive and negative effect directions and circle size proportional to the effect size. The most significant association for each metabolite group is labeled. Abbreviations: ACD, all-cause dementia; AD, Alzheimer’s disease; ApoB, apolipoprotein B; CE, cholesteryl esters; FC, free cholesterol; Glu, glucose; His, histidine; LA, linoleic acid; LDL, low-density lipoprotein; L-LDL, large LDL; LDL-L, total lipids in LDL; LDL-P, concentration of LDL particles; M-VLDL-L, total lipids in medium VLDL; PC, phosphatidylcholines; PL, phospholipids; PUFA, polyunsaturated fatty acids; SM, sphingomyelin; VaD, vascular dementia; Val, valine; VLDL-TG, triglycerides in VLD

Fig. 2

Metabolites importance ranking and SHAP visualization of modeling based on incident ACD populations. A Metabolites that survived FDR corrections in the association analysis further underwent sequential forward selection. The bar chart illustrates the importance of metabolites (left axis), ranked in ascending order. The line chart depicts cumulative area under the curve (AUC) values (right axis) as metabolites are included in successive iterations. The metabolites ultimately selected for MetRS calculation are highlighted in red. B Individual SHAP values of the selected metabolites are ranked according to their contributions. The x-axis represents the scale of the SHAP values for every metabolite, indicating their contribution to the prediction. The color range corresponds to each metabolic value, from blue (low value) to red (high value). Abbreviations: AUC, area under the curve; SHAP, SHapley Additive exPlanations

Fig. 3

MetRS stratifies the risk of dementia onset. A–C Observed event rate for incident ACD, AD and VaD, plotted against MetRS percentiles over the entire study population. Blue dots represent males and red dots represent females. The size of each dot is proportional to age. D–F Cumulative risk over the observation time for incident ACD, AD and VaD, stratified by MetRS tertiles (light blue, bottom tertile; blue, median tertile; dark blue, top tertile). The shaded area indicates the 95% CI of the survival curves. G Regression results of MetRS and dementia outcomes in all participants and subgroups. Model 1, MetRS; Model2, MetRS + demographic indicators; Model 3, MetRS + demographic indicators + cognitive indicators. Abbreviations: ACD, all-cause dementia; AD, Alzheimer’s disease; VaD, vascular dementia

Fig. 4

Prediction of incident ACD, AD, and VaD. Receiver operating characteristic (ROC) curves show the predictive performance of MetRS, either alone or in combination with demographic and cognitive indicators, for all incident cases (A–C), as well as for over 10-year incident cases (D–F) of ACD, AD, and VaD. The dotted line indicates an AUC of 0.50 for comparison. AUC estimates and 95% CIs are shown in Table S10. Abbreviations: AUC, area under the curve; ACD, all-cause dementia; AD, Alzheimer’s disease; VaD, vascular dementia