Associations of CAIDE dementia risk score with glucose and lipid‐related metabolic blood biomarkers in the FINGER study participants

Abstract Background The Cardiovascular risk factors, aging, and dementia (CAIDE) risk score is a validated tool estimating dementia risk. We investigated the association of CAIDE score with 12 markers of glucose and lipid metabolism, in the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) participants. Methods The FINGER trial had 1260 participants, aged 60–77 years, with a CAIDE score ≥6, without substantial cognitive impairment. At baseline, a panel of 12 glucose and lipid metabolism blood markers (adiponectin, adipsin, c‐peptide, ghrelin, gastric inhibitor polypeptide (GIP), glucagon‐like‐peptide (GLP‐1), glucagon, insulin, leptin, plasminogen activator in plasma (PAI‐1), resistin, and visfatin) were measured using Bio‐Plex‐200. CAIDE dementia risk score (age, sex, education, blood pressure, body mass index (BMI), cholesterol, physical activity) was also available. Linear regression models were fit using CAIDE score as the outcome, metabolic markers as predictors and APOE4 as covariate. We also tested the interactions between the metabolic markers and APOE4. Results Participants with the metabolic markers (n=813) were younger (p=0.01), had lower levels of HDL (p=0.01) and higher levels of systolic blood pressure (p= 0.02), HbA1c (p=0.00), BMI (p=0.00), CRP (p=0.02) and higher CAIDE risk score (p=0.02) but slightly better memory (p=0.00) compared to those without the metabolic markers (n=447). Initially, CAIDE score was tested with each metabolic marker separately. Interactions between metabolic markers and APOE4 were also tested. Higher CAIDE score was associated with lower adiponectin (β coefficient= ‐0.01, p=0.01), ghrelin (β = ‐0.13, p<0.001), and higher insulin (β =0.16, p<0.001), leptin (β=0.15 p<0.001), PAI‐1 (β =0.14, p<0.001) and c‐peptide (β=0.15 p<0.001). No interactions between metabolic markers and APOE were significant except for glucagon x APOE4 (β= ‐2.10, p=0.021). Next, the metabolic markers with the significant associations were added to a mutually adjusted model. The associations of CAIDE score with ghrelin (β= ‐0.22, p<0.001) and insulin (β=0.124, p=0.03) remained significant. Conclusion Increased dementia risk in older people is linked to glucose and lipid imbalance. The identification of phenotypes through different metabolic markers and the role of genetic factors on these metabolic markers can lead to a better stratification of the population at risk of developing dementia.