Lifestyle and vascular risk effects on MRI-based biomarkers of Alzheimer's disease: a cross-sectional study of middle-aged adults from the broader New York City area

Abstract

Objective: To investigate the effects of lifestyle and vascular-related risk factors for Alzheimer's disease (AD) on in vivo MRI-based brain atrophy in asymptomatic young to middle-aged adults.

Design: Cross-sectional, observational.

Setting: Broader New York City area. Two research centres affiliated with the Alzheimer's disease Core Center at New York University School of Medicine.

Participants: We studied 116 cognitively normal healthy research participants aged 30-60 years, who completed a three-dimensional T1-weighted volumetric MRI and had lifestyle (diet, physical activity and intellectual enrichment), vascular risk (overweight, hypertension, insulin resistance, elevated cholesterol and homocysteine) and cognition (memory, executive function, language) data. Estimates of cortical thickness for entorhinal (EC), posterior cingulate, orbitofrontal, inferior and middle temporal cortex were obtained by use of automated segmentation tools. We applied confirmatory factor analysis and structural equation modelling to evaluate the associations between lifestyle, vascular risk, brain and cognition.

Results: Adherence to a Mediterranean-style diet (MeDi) and insulin sensitivity were both positively associated with MRI-based cortical thickness (diet: β_s≥0.26, insulin sensitivity β_s≥0.58, P≤0.008). After accounting for vascular risk, EC in turn explained variance in memory (P≤0.001). None of the other lifestyle and vascular risk variables were associated with brain thickness. In addition, the path associations between intellectual enrichment and better cognition were significant (β_s≥0.25 P≤0.001), as were those between overweight and lower cognition (β_s≥-0.22, P≤0.01).

Conclusions: In cognitively normal middle-aged adults, MeDi and insulin sensitivity explained cortical thickness in key brain regions for AD, and EC thickness predicted memory performance in turn. Intellectual activity and overweight were associated with cognitive performance through different pathways. Our findings support further investigation of lifestyle and vascular risk factor modification against brain ageing and AD. More studies with larger samples are needed to replicate these research findings in more diverse, community-based settings.

Keywords: alzheimer’s disease; brain Imaging; brain aging; lifestyle; vascular risk.

Figure 1

Assessing the impact of lifestyle and vascular risk factors on brain ageing and cognition. Results of the following SEMs. (A) Relationships between lifestyle, vascular risk and brain, excluding cognition from the model. (B) Relationships between lifestyle, brain and cognition, excluding vascular risk from the model. (C) Relationships between vascular risk, brain and cognition, excluding lifestyle from the model. Red lines indicate the significant effects observed in our data. Purple lines indicate significant alternate paths. Blue lines indicate significant correlational effects. Grey lines indicate associations which were tested but found to be non-significant. Numbers beside the lines are standardised beta coefficients and corresponding P values: *P<0.01; **P<0.001. Not shown here, all brain ROI variables were correlated with one another (see online supplementary etable 1). Direct associations between lifestyle and vascular risk factors are found in online supplementary etable 2. Variables in squares are measured variables (eg, one variable represents this construct). Variables in circles are latent variables (multiple variables were averaged using a CFA to represent that construct). Straight lines indicate causal relationships, curved lines indicate correlational relationships. Age was entered as a covariate. Brain biomarkers were examined with and without adjusting for total intracranial volume. CFA, confirmatory factor analysis; ROI, region of interest; SEMs, structural equation models.

Figure 2

Assessing the impact of lifestyle and vascular risk factors on limbic brain structures and cognition. Results of the following SEMs. (A) Relationships between lifestyle, vascular risk and limbic brain, excluding cognition from the model. (B) Relationships between lifestyle, limbic brain and cognition, excluding vascular risk from the model. (C) Relationships between vascular risk, limbic brain and cognition, excluding lifestyle from the model. See legend to figure 1. Age was entered as a covariate. Brain measures were examined with and without adjusting for total intracranial volume. *P<0.01; **P<0.001. EC, entorhinal cortex; PCC, posterior cingulate cortex; SEMs, structural equation models.