Abstract
BACKGROUND: Steatotic liver disease is associated with cardiovascular risk, yet the relevance of hepatic fibro-inflammatory activity to cerebral small vessel disease remains unclear. We investigated whether liver MRI phenotypes, iron-corrected T1 (cT1; fibro-inflammatory injury) and proton density fat fraction (PDFF; hepatic fat), and brain MRI white matter hyperintensity (WMH; cerebral small-vessel disease burden) predict incident composite neuropsychiatric events and mortality.</p>
METHODS: This prospective study included 29,344 UK Biobank participants with liver and brain MRI. Participants were classified into four phenotypes based on liver imaging impairment (cT1 ≥ 800 ms) and high WMH burden (>80th percentile, age-adjusted). Cause-specific Cox models estimated risks for incident composite neuropsychiatric events (stroke, dementia, or late-life depression), liver events, and all-cause mortality, adjusting for prespecified covariates.</p>
RESULTS: Over a median follow-up of 5.7 years, 530 participants developed neuropsychiatric events. Higher cT1 was associated with neuropsychiatric events (per 100 ms, hazard ratio [HR] 1.31, 95% confidence interval [CI] 1.04-1.65) and mortality (1.29, 1.04-1.60). Liver fat (PDFF) was not associated with composite neuropsychiatric events (1.00, 0.98-1.03) or mortality (1.00, 0.97-1.03). High WMH burden associated with neuropsychiatric events (1.26, 1.08-1.48) and mortality (1.17, 1.01-1.35). Participants with both elevated cT1 and high WMH had a higher risk of neuropsychiatric events (2.30, 1.00-5.29) and liver events (7.13, 3.32-15.33).</p>
CONCLUSIONS: Fibro-inflammatory liver injury, rather than steatosis, was the primary liver MRI correlate of incident composite neuropsychiatric events and mortality. Combined liver-brain MRI phenotyping identified a high-risk subgroup, consistent with shared microvascular vulnerability.</p>