Polygenic Risk Score Predicts Prognosis in Nonobese Metabolic Dysfunction-Associated Steatotic Liver Disease but not...

Prognosis in nonobese vs obese metabolic dysfunction-associated steatotic liver disease remains uncertain, and genetic effects may differ by obesity. We examined whether a 2-variant polygenic risk score (PNPLA3 rs738409 and TM6SF2 rs58542926) predicts liver-related events and overall survival across obesity strata.

This multicenter Japan Genetic Study of Nonalcoholic Fatty Liver Disease Research (JAGUAR) cohort included 1304 Japanese patients with biopsy-proven metabolic dysfunction-associated steatotic liver disease. We examined the relationship between the 2-variant polygenic risk score and liver-related events or overall survival according to obesity status. External validation used an independent Korean, nonbiopsied cohort; validation focused on overall survival, including a clinically enriched subset with fatty liver and Fibrosis-4 index >1.3.

Median follow-up was 8.1 years (range, 1.0-24.0). Overall, 103 liver-related events (7.9%) and 65 deaths (5.0%) occurred; 10-year liver-related event incidence was 10.1% and 10-year overall survival was 95.0%. Liver-related event risk did not differ by obesity (10.0% vs10.4%). Two-variant polygenic risk score independently predicted liver-related events in both strata (per 0.1-unit: obese hazard ratio, 1.207; nonobese hazard ratio, 1.454). For overall survival, the 2-variant polygenic risk score was significant only in nonobese patients (hazard ratio, 1.365), not in obese patients. Adding the 2-variant polygenic risk score improved discrimination in nonobese patients for both liver-related events and overall survival but not for overall survival in obese patients. In external validation, the 2-variant polygenic risk score was not significantly associated with overall survival, including in the enriched subset.

In biopsy-proven metabolic dysfunction-associated steatotic liver disease, the 2-variant polygenic risk score improves liver-related event prediction regardless of obesity and shows the strongest association with overall survival among nonobese patients; these overall survival associations were not reproduced in a predominantly mild-spectrum, nonbiopsied cohort.