In the present study, using MR, we provide evidence for a putative causal effect of BMI on cortical thickness across the human brain, with increased BMI associated with lower cortical thickness. The most prominent effects were observed in the frontal and temporal regions, notably the precentral and fusiform gyri. We furthermore found a concurrent association of CRP and VAT with lower cortical thickness, both globally and regionally across brain regions, largely overlapping with those associated with increased BMI.
Our finding of an association between increased BMI and lower cortical thickness particularly in the temporal cortex is in line with findings from latest large-scale multicenter analyses 10,47 that congruently report a relationship between increased BMI and lower temporal cortical thickness. In fact, the most up-to-date and largest mega-analysis from the ENIGMA consortium based on 6,420 participants indicated the strongest regional effect sizes for the association between increased BMI and lower cortical thickness in the fusiform gyrus 6, which was supported by the findings of our MR analyses. More recent large-scale studies by the ENIGMA consortium and others further corroborate the association between increased BMI and lower temporal thickness 9,48
Two separate MR studies have also recently presented evidence for an impact of obesity on cortical thickness 49,50. Limitations in both studies, however, resulted in smaller numbers of associations seen across the brain, and in the case of Chen et al. 2023b 50 there is evidence for both decreased and increased cortical thickness with increasing BMI. Both studies focused only on adiposity and included regional thickness data that had been corrected for global thickness; in the current study we used summary statistics that had not undergone such correction, in line with analyses in observational studies, to interrogate region-specific impacts 6. Incorporating these considerations, it is apparent that such differences in methodology would have led to very different interpretations. Additionally, Chen et al. used GWAS summary statistics for both the adiposity and cortical measures that were generated including individuals from the UK Biobank, which is likely to have biased their results due to the large number of overlapping samples in the exposure and outcome datasets 50.
The main findings of our hypothesis-driven study hold several relevant implications for future research. First, our genetic results support the hypothesis of a causal effect of BMI on brain structural decline. Importantly, this is consistent with previous findings from longitudinal human neuroimaging studies reporting BMI related brain structural atrophy over time 15 as well as with recent reports on accelerated brain ageing in obesity 5,51.
Our finding of a putative causal effect of increased BMI on lower cortical thickness raises the question of the underlying biological mechanisms that mediate the effect of BMI on cortical structural decline. In this regard, our finding of a parallel association of VAT and CRP with changes in brain structure, that largely correspond to BMI-associated changes appears noteworthy. These findings align with prior research linking obesity, inflammation, and altered brain physiology globally and in cortical regions in particular 52–54. Notably, past studies have emphasized the role of VAT in generating pro-inflammatory cytokines, potentially exerting adverse effects on the CNS. More specifically, pro-inflammatory molecules released by lymphocytes and M1 macrophages residing in the VAT were shown to induce apoptosis trough microglia stimulation inside of the CNS, a possible mechanism for cortical thinning 55–57. Our results underscore the need for future mechanistic investigations in this domain. Chronic low-grade inflammation, which has been shown to be present in multiple neuropsychiatric disorders and obesity, may represent a substantial shared underlying biological mechanism. With the discoveries of meningeal lymphatic vessels surrounding the brain and signalling cascades activated by peripheral cytokines that result in an increased permeability of the blood brain barrier, it is only recently that researchers are starting to uncover pathways by which systemic inflammation affects brain function 58,59. Moreover, there is a bidirectional relationship between chronic low-grade inflammation and dysregulated mitochondrial function 60,61. Since mitochondrial integrity is essential to ensure extraordinary energy demands during synaptic transmission and plasticity 62, this might be another potential mechanism by which obesity related peripheral maladjustments affect brain structure. In this regard, future studies should furthermore take into account the microbiome, more specifically the role of the gut-brain axis in the context of obesity. Gut dysbiosis can lead to an unbalanced immune response and influence systemic energy regulation 63,64. Overall, the biological correlations in combination with the data we have presented are promising for possibly establishing a link between obesity, energy balance, neuroinflammation and cortical changes in the brain. Thus, chronic low-grade inflammation and dysregulated energy balance, inherent to BMI, ermerge as plausible mechanisms underlying the observed reduction in cortical thickness attributed to BMI 65.
In addition to the findings for CRP and VAT, it should be noted that none of the other examined factors, including glucose metabolism, serum triglycerides, lipoproteins and blood pressure, exhibited a consistent pattern of associations with brain structure. This finding is striking considering previous biological research suggesting a potential connection between the vascular system and brain integrity including findings of associations between higher BMI and white matter hyperintensities, which are known to be related to cerebrovascular diseases 65. While clarification of the mechanistic underpinnings between the putative causal link of BMI and brain structure will rely on preclinical research (e.g., animal models), future translational research should take advantage of methodological progress in neuroimaging to clarify the relationship between BMI and brain structural decline: Investigation of perivascular spaces and glymphatic clearance through ultra-fast magnetic encephalography represent promising approaches for the investigation of intermediate phenotypes related to neuroinflammation and neurovascular changes in obesity in-vivo 66,67. In summary, our study supports the notion that VAT and low-grade inflammation, rather than shifts in metabolic serum markers as previous beliefs linking obesity primarily with vascular factors have underestimated its direct impact on cognition. Recent findings challenge this view, suggesting that obesity contributes not only to vascular dementia but also, notably, to Alzheimer's dementia 68
While the overwhelming relevance of obesity and related metabolic dysregulation as a cardiovascular risk factor has long been recognized by research and clinical practice, our results advocate for the relevance of increased BMI as a risk factor for the development of neurodegenerative disorders. Increased BMI has been shown to act as an important risk factor for several neurodegenerative and psychiatric disorders and has recently been stated to be the most important modifiable risk factor for dementia in the US 69,70. The putative causal effect of increased BMI on cortical thickness decline observed in the present study appears to be in line with the aforementioned reports. It could thus be speculated that lower cortical thickness might act as a relevant mediator in the association between adverse metabolic conditions and neuropsychiatric disorders. Thus, atrophy of brain regions located in the temporal lobe has long been known to be associated with the subsequent development of cognitive impairment and dementia 71–73. Future studies should thus aim to further clarify the role of BMI related thickness decline in the fusiform temporal and precentral frontal gyrus as the most prominent findings in the present study, in the development of neuropsychiatric disorders.
This study has a number of strengths and limitations. Strengths include the application of MR analyses to test hypotheses based on previous observational studies, using genetic information from large-scale meta-analyses of BMI as well as a comprehensive investigation of genetic determinants influencing cortical thickness 27. These data allowed us to investigate both global as well as regional specific associations between BMI and cortical thickness, with the finding of causal associations between BMI and brain structure complementing previous cross-sectional and longitudinal neuroimaging studies. The regional specific analyses allowed us to confirm a putative causal effect of BMI particularly in the temporal cortex, thus providing candidate brain regions for subsequent mechanistic studies on underlying biological mediators. The inclusion of a variety of relevant related phenotypes covering a broad spectrum of metabolic, cardiovascular and inflammatory traits represents another strength as it allowed to disentangle and identify potential candidate systems for future mechanistic research on the impact of BMI on brain health.
Limitations to our study should also be acknowledged. First, while we were able to investigate associations between BMI and cortical thickness, we could not investigate whether the thickness of the cortex globally or in particular regions influence BMI due to the current paucity of cortical thickness-associated genetic variants. Further, current data do not allow us to infer the exact timing of potential causal effects of BMI on brain structure. It thus remains unclear during which developmental stages and at which speed increased BMI might cause brain structural decline. Future GWAS meta-analyses and longitudinal cohort studies are warranted to address these open research questions.
Collectively, the present study corroborates the notion of a putative causal effect of BMI on brain structural volume decline. Our observation of a causal effect of increased BMI on lower temporal and frontal cortical thickness calls for increased attention towards the relevance of obesity and related metabolic conditions as modifiable risk factors for brain health. Moreover, our findings suggest that visceral adipose tissue and low-grade inflammation may be critically linked phenotypes in understanding the impact of BMI on brain structure. These findings highlight the need for future experimental investigations aimed at unravelling the potential cascade of mechanisms and identifying intervention opportunities within the intricate connection between weight gain, adipose tissue, inflammation, and structural brain changes. Future research and preventive efforts should aim to further explore the biological mechanisms through which BMI might influence brain structural decline and clarify the relationship between BMI related brain structural impairment and specific domains of neurocognitive functioning.