Patients with metabolic dysfunction-associated steatotic liver disease could be classified into 5 reproducible subgroups with distinct genetic profiles and disease trajectories, including a polygenic subgroup with relatively mild metabolic abnormalities but substantial long-term liver risk.
Researchers conducted a retrospective analysis of electronic health record-linked genomic data from the Mayo Clinic Biobank and Tapestry Study. The Mayo Clinic cohorts included 2,042 patients with metabolic dysfunction-associated steatotic liver disease (MASLD). The Tapestry cohort included 2,560 patients after those with incomplete clinical data were excluded. The Mayo cohort was divided into development (n = 1,532) and validation (n = 510) groups. Tapestry served as an independent validation cohort. The researchers used latent class analysis of 13 clinical and demographic variables to derive the subgroups and evaluated their reproducibility across the development cohort.
The researchers characterized differences in genetic profiles, comorbidities, medication use, and longitudinal disease progression across the subgroups. Outcomes included metabolic dysfunction-associated steatohepatitis (MASH), fibrosis, cirrhosis, hepatocellular carcinoma, ischemic heart disease, acute renal failure, sleep apnea, major depression, and liver transplantation.
The analysis identified 5 groups: cardiometabolic MASLD without obesity (C1), male-predominant cardiorenal MASLD (C2), female-predominant MASLD with obesity and mood disorders (C3), polygenic MASLD (C4), and polygenic MASH (C5). In the development cohort, C1 accounted for 31% of patients, C2 for 27%, C3 for 24%, C4 for 9%, and C5 for 10%. C2 and C3 had heterogeneous metabolic profiles, whereas C4 and C5 showed more liver-focused phenotypes and distinct genetic architectures.
The subtypes were reproducible in the validation cohorts, with similar subgroup sizes and clinical characteristics. The researchers also observed similar longitudinal patterns in Tapestry despite the cohort's younger age. The classification also showed comparable performance among patients with type 2 diabetes or obesity.
Notably, C4 combined a comparatively mild metabolic profile with substantial liver-specific risk. This subgroup had the lowest incidence of several metabolic and extrahepatic comorbidities, including ischemic heart disease, but showed increased long-term risks of fibrosis and cirrhosis. Likelihood of liver transplantation occurred in 16% of patients in C4 in the Mayo cohort and 10% in Tapestry over 10 years. The researchers noted that C4 was enriched for liver-related genetic risk, including the TM6SF2 E167K variant.
C5 showed another liver-dominant disease trajectory. Compared with C1, patients in C5 had more than twice the likelihood of developing MASH, fibrosis, and acute renal failure during follow-up. C5 also had an elevated long-term cirrhosis risk and was characterized by elevated liver enzymes and enrichment for the PNPLA3 I148M risk variant.
In contrast, C1 through C3 showed more systemic or extrahepatic patterns. C2 was predominantly male and had a cardiorenal profile, whereas C3 was predominantly female and was characterized by obesity, mood disorders, greater antidepressant use, and a high prevalence of type 2 diabetes and sleep apnea. The researchers broadly characterized C1 through C3 as systemic or extrahepatic-dominant subtypes and C4 and C5 as intrahepatic-dominant subtypes.
The study had several limitations. The analysis included only clinical variables significantly associated with MASLD, potentially excluding other informative measures, such as gamma-glutamyl transferase, platelet count, and blood pressure. Requiring complete clinical data may also have limited characterization of disease progression. In addition, subgroup assignment could have been influenced by probabilities derived from the development cohort, and assigning each patient to a single subgroup did not capture uncertainty among patients near subgroup boundaries.
The findings suggest that integrating clinical and genetic information may help distinguish heterogeneous MASLD phenotypes and their differing patterns of progression, although further research is needed to address uncertainty in subgroup assignment and evaluate the framework in additional settings.
“These findings underscore the need for individualized clinical management based on genetic and metabolic risk profiles,” wrote lead study author Tahmina Sultana Priya, of Mayo Clinic, and colleagues.
The study was funded by Mayo Clinic. Andres J. Acosta and Alina M. Allen reported financial relationships with industry. The remaining researchers reported no conflicts related to the study.
Source: Nature Communications
