Mediation analyses link cardiometabolic factors and liver fat with white matter hyperintensities and cognitive performance: A UK Biobank study

Daniel Edvard Askeland-Gjerde,Lars T. Westlye,Patrik J. Andersson,Max Korbmacher,Ann-Marie G. de Lange,Dennis van der Meer,Olav B. Smeland,Sigrun Halvorsen,Ole A. Andreassen,Tiril P. Gurholt

DOI: https://doi.org/10.1101/2024.10.17.24315664

2024-10-18

Abstract:Liver fat is associated with cardiometabolic disease, cerebrovascular disease, and dementia. Cerebrovascular disease, most frequently cerebral small vessel disease, identified by MRI as white matter hyperintensities (WMH), often contributes to dementia. However, liver fat′s role in the relationship between cardiometabolic risk, WMH, and cognitive performance is unclear. In the UK Biobank cohort (n = 32,628; 52.6% female; mean age 64.2 ± 7.7 years; n = 23,467 cognitive performance subsample), we used linear regression to investigate associations between cardiometabolic factors measured at baseline and liver fat, WMH, and cognitive performance measured at follow-up, on average, 9.3 ± 2.0 years later. We used structural equation modeling to investigate whether liver fat mediates associations between cardiometabolic factors and WMH and whether WMH mediates associations between liver fat and cognitive performance. Nearly all cardiometabolic factors were significantly associated with liver fat (|r| in [0.03, 0.41], p in [1.4x10 , 0]) and WMH (|r| in [0.05, 0.14], p in [1.5x10 , 2.7x10 ]) in regression models. Liver fat was associated with WMH (r=0.09, p=3x10 ) and cognitive performance (r=-0.03, p=1.5x10 ). Liver fat mediated the associations between cardiometabolic factors and WMH (β-mediation| in [0.01, 0.03], p-mediation in [5.7x10 , 0]) and WMH mediated the associations between liver fat and cognitive performance (β-mediation = -0.01, p-mediation ≈ 0). Our findings indicate that liver fat mediates associations between cardiometabolic factors and WMH and that WMH mediates the association between liver fat and cognitive performance. This suggests that liver fat might be important for understanding the effects of cardiometabolic factors on cerebrovascular disease and cognitive function. Experimental studies are warranted to determine relevant targets for preventing vascular-driven cognitive impairment.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to explore the mediating role of liver fat in the relationship among cardiometabolic factors, white matter hyperintensities (WMH), and cognitive performance. Specifically, the study aims to verify the following hypotheses:

1. **Relationship between cardiometabolic risk factors and white matter hyperintensities**: The study hypothesizes that there is an association between cardiometabolic risk factors and white matter hyperintensities, and this association may be achieved through the mediating role of liver fat.
2. **Relationship between liver fat and cognitive performance**: The study hypothesizes that there is an association between liver fat and cognitive performance, and this association may be achieved through the mediating role of white matter hyperintensities.
3. **Gender differences**: The study also hypothesizes that the above - mentioned associations may have gender differences.

### Research background
Liver fat is related to cardiometabolic diseases, cerebrovascular diseases, and dementia. Cerebrovascular diseases, especially small - vessel diseases, are usually identified by white matter hyperintensities (WMH) detected by magnetic resonance imaging (MRI) and are often an important contributing factor to dementia. However, the role of liver fat in the relationship among cardiometabolic risk, white matter hyperintensities, and cognitive performance is not clear.

### Research methods
The study used data from the UK Biobank, with a sample size of 32,628 people, of which 52.6% were women and the average age was 64.2 years. Linear regression analysis and structural equation modeling (SEM) were used to explore the relationships among cardiometabolic factors, liver fat, white matter hyperintensities, and cognitive performance.

### Main findings
1. **Cardiometabolic factors and liver fat**: Almost all cardiometabolic factors were significantly associated with liver fat (|r| ranged from [0.03, 0.41], and p - values ranged from [1.4x10^-8, 0]).
2. **Cardiometabolic factors and white matter hyperintensities**: Almost all cardiometabolic factors were significantly associated with white matter hyperintensities (|r| ranged from [0.05, 0.14], and p - values ranged from [1.5x10^-13, 2.7x10^-148]).
3. **Liver fat and white matter hyperintensities**: Liver fat was significantly associated with white matter hyperintensities (r = 0.09, p = 3x10^-64).
4. **Liver fat and cognitive performance**: Liver fat was significantly associated with cognitive performance (r=- 0.03, p = 1.5x10^-7).
5. **Mediating effects**:
- Liver fat played a mediating role in the association between cardiometabolic factors and white matter hyperintensities (|βmediating| ranged from [0.01, 0.03], and p - mediating ranged from [5.7x10^-9, 0]).
- White matter hyperintensities played a mediating role in the association between liver fat and cognitive performance (βmediating=- 0.01, p - mediating≈0).

### Conclusion
The research results show that liver fat plays a mediating role in the association between cardiometabolic factors and white matter hyperintensities, and white matter hyperintensities play a mediating role in the association between liver fat and cognitive performance. This suggests that liver fat may be an important factor in understanding the impact of cardiometabolic factors on cerebrovascular diseases and cognitive function. Future experimental studies are needed to determine relevant targets for preventing vascular - driven cognitive impairment.