Predicting dementia in primary care patients with a cardiovascular health metric: A prospective population-based study

Abstract

Background: Improving cardiovascular health possibly decreases the risk of dementia. Primary care practices offer a suitable setting for monitoring and controlling cardiovascular risk factors in the older population. The purpose of the study is to examine the association of a cardiovascular health metric including six behaviors and blood parameters with the risk of dementia in primary care patients. Methods: Participants (N = 3547) were insurants aged ≥55 of the largest German statutory health insurance company, who were enrolled in a six-year prospective population-based study. Smoking, physical activity, body mass index, blood pressure, total cholesterol, and fasting glucose were assessed by general practitioners at routine examinations. Using recommended cut-offs for each factor, the patients' cardiovascular health was classified as ideal, moderate, or poor. Behaviors and blood parameters sub-scores, as well as a total score, were calculated. Dementia diagnoses were retrieved from health insurance claims data. Results are presented as hazard ratios (HRs) and 95 % confidence intervals (95 % CIs). Results: Over the course of the study 296 new cases of dementia occurred. Adjusted for age, sex, and education, current smoking (HR = 1.77, 95 % CI 1.09-2.85), moderate (1.38, 1.05-1.81) or poor (1.81, 1.32-2.47) levels of physical activity, and poor fasting glucose levels (1.43, 1.02-2.02) were associated with an increased risk of dementia. Body mass index, blood pressure, and cholesterol were not associated with dementia. Separate summary scores for behaviors and blood values, as well as a total score showed no association with dementia. Sensitivity analyses with differently defined endpoints led to similar results. Conclusions: Due to complex relationships of body-mass index and blood pressure with dementia individual components cancelled each other out and rendered the sum-scores meaningless for the prediction of dementia.