Obesity is a global health concern linked to cardiovascular disease and dementia. While an association between obesity and brain function has been suggested, the underlying mechanisms remain unclear. Studies have reported inconsistent associations between obesity and brain structure, with some showing global cerebral gray matter (GM) atrophy and others showing inconsistent effects on white matter (WM) volume. The distinction between central (abdominal) and peripheral obesity is crucial, as central obesity is strongly associated with metabolic syndrome, type-2 diabetes, myocardial infarction, and Alzheimer's disease. This study aimed to investigate the association between obesity, particularly central obesity, and brain structure using various measures of body composition (BMI, WHR, abdominal MRI, DXA, bioelectrical impedance) and brain MRI (GM and WM volumes, brain network efficiency). The researchers also investigated whether the associations were mediated by known cardiovascular risk factors. Understanding the relationship between obesity and brain structure is vital for developing effective prevention and treatment strategies for obesity-related cognitive decline and dementia.

Existing literature demonstrates a link between obesity and both cardiovascular disease and dementia, yet the underlying mechanisms remain poorly understood. Studies have reported associations between obesity and global cerebral gray matter (GM) atrophy, but the findings regarding white matter (WM) volume have been inconsistent, with some showing decreases and others showing increases. More recent research has employed diffusion-tensor MRI tractography to assess brain network integrity, revealing correlations between network metrics and cognition in both normal aging and disease states. The impact of obesity on the brain is further complicated by the differing effects of central (abdominal) versus peripheral obesity. Central obesity, characterized by a high android-to-gynoid ratio (increased visceral fat), is particularly linked to metabolic syndrome and increased risk of several diseases, including Alzheimer's disease. However, the effects of central obesity on the central nervous system (CNS) are not fully understood. The challenges of accurately measuring body fat and distribution at scale have led to the use of anthropometric measures (BMI, WHR) in clinical practice. However, abdominal MRI and DXA offer more precise assessments of body fat distribution and composition, providing a more nuanced understanding of the relationship between obesity and brain structure. Previous studies on the link between obesity and brain structure have often been limited by small sample sizes and reliance on BMI as the sole measure of obesity.

This study used data from 15,634 participants in the UK Biobank who underwent brain MRI, abdominal MRI, DXA, and bioelectric whole-body impedance. Obesity was assessed using BMI, WHR, abdominal MRI (visceral adipose tissue, subcutaneous adipose tissue, total adipose tissue volume, lean tissue volume), DXA (android-to-gynoid fat mass ratio, visceral adipose tissue mass, trunk-to-leg fat mass ratio, etc.), and bioelectric whole-body impedance (fat mass, fat-free mass). Brain MRI data included GM and WM volumes, specific brain regions (thalamus, caudate, putamen, pallidum, hippocampus, etc.), and brain network metrics derived from diffusion-tensor MRI tractography. Cognitive performance was assessed using several tests (visual memory, reaction time, prospective memory, trail making test). Blood biochemistry data included markers such as blood glucose, cholesterol, and HbA1c. Statistical analysis included analysis of covariance to test for group differences in neuroimaging outcomes across stratified BMI/WHR groups, partial correlations to examine associations between continuous indicators of obesity and neuroimaging outcomes, and mediation analysis using structural equation modeling to determine whether blood biochemistry or blood pressure mediated the effect of obesity on GM volume. The researchers adjusted for covariates such as sex, age, Townsend Deprivation Index, alcohol intake, smoking status, diabetes, and blood pressure.

Central obesity, as measured by WHR, abdominal MRI, and DXA, showed a significant negative association with global cerebral GM volume (p = 6.7 x 10⁻¹⁶ for anthropometric data, p = 8.3 x 10⁻⁸¹ for DXA, p = 0.0006 for abdominal MRI). Regional analysis revealed significant associations between central obesity and specific GM subcortical nuclei (thalamus, caudate, pallidum, nucleus accumbens). No significant associations were found between obesity and WM volume, brain network efficiency, or cognitive scores. Mediation analysis did not reveal any significant mediating effects of C-reactive protein, blood pressure, glucose, lipids, or other metabolic syndrome components. Higher body fat mass and fat-free mass from bioelectrical impedance were associated with lower GM volume. Similarly, higher total adipose tissue volume from abdominal MRI was associated with lower GM volume. DXA findings showed a significant negative effect of fat mass index (FMI) and lean body mass index (LBMI) on GM volume. Interaction analysis showed significant interaction effects between sex and measures of body fat mass, fat-free mass, abdominal adipose tissue volume.

This study provides robust evidence that central obesity is associated with GM atrophy, but not WM volume or network efficiency, in a large population-based cohort. The consistent findings across multiple measures of central obesity strengthen the study's conclusions. The lack of association with WM and network measures suggests that the CNS damage associated with central obesity primarily affects GM, potentially indicating neuronal loss rather than tract degeneration. The specific regional associations observed in subcortical GM structures known to be involved in energy balance and reward processing suggest potential mechanisms linking central obesity to brain structure. The lack of mediation by traditional metabolic syndrome markers indicates a more complex relationship between central obesity and GM atrophy, suggesting that other factors warrant further investigation. The findings highlight the importance of distinguishing between central and peripheral obesity when studying the effects of obesity on the brain. The relatively small effect sizes despite highly significant p-values underscore the importance of very large sample sizes for detecting subtle effects in complex systems.

This large-scale study demonstrates a significant association between central obesity and reduced gray matter volume in the brain, particularly in subcortical regions involved in energy balance and reward processing. This association was not mediated by known cardiovascular risk factors or markers of metabolic syndrome. The results highlight the importance of central obesity as a key factor influencing brain structure and warrant further research to elucidate the mechanisms linking central adiposity to GM atrophy. Longitudinal studies are needed to ascertain the causal relationship and the progression of these effects over time. Further research should focus on identifying the specific mediating factors responsible for this relationship and explore potential therapeutic interventions.

The cross-sectional design of the study limits causal inference. The UK Biobank sample, while large and population-based, may not perfectly represent the entire population, potentially introducing selection bias. While highly significant associations were found, the effect sizes were small in some cases, highlighting the complexity of the relationship between obesity and brain structure. The abdominal MRI data only covered the abdomen, precluding a complete assessment of whole-body fat distribution. Despite the large sample size, some analyses had reduced statistical power due to missing data in various measures.