Artificial intelligence analysis of more than 14,000 patients with metabolic dysfunction‑associated steatohepatitis (MASH) identified cardiovascular and kidney conditions as the most consistent predictors of rapid liver scarring, serious long‑term complications, and the highest health‑care costs.
Study design and data sources
Researchers accessed the OM1 Real‑World Data Cloud, an electronic health‑record and claims repository covering over 340 million U.S. individuals. They extracted 14,707 patients diagnosed with MASH who also had a Fibrosis‑4 (FIB‑4) score recorded within 90 days of diagnosis, spanning January 2013 to September 2022.
Using the PhenOM AI platform, the team compiled diagnoses, procedures, medications, laboratory values, and demographic information from the year preceding each patient’s index date. The AI grouped these variables into “phenotypic signals” and tested them against three outcomes: rapid fibrosis progression, long‑term liver‑ and cardiovascular‑related clinical events, and health‑care costs in the top decile of the cohort.
Key findings on disease progression
Among 1,795 patients evaluated for fibrosis trajectory, a history of anemia and thrombocytopenia—particularly iron‑deficiency anemia—showed the largest absolute difference between rapid progressors and non‑progressors. Cardiovascular diagnoses such as atrial fibrillation, coronary artery disease, and congestive heart failure displayed the strongest relative association with swift fibrosis advancement. Heart‑failure‑related procedures, prior hospitalizations, and kidney disorders including acute kidney injury and chronic kidney disease (CKD) also distinguished fast‑progressing cases.
Related: Users Blocked From Accessing Website
When looking at longer‑term outcomes in a separate group of 13,880 patients, chronic conditions like diabetes and hypertension, together with kidney and cardiac diagnoses, were linked to the development of cirrhosis or other liver complications. In patients without prior cardiovascular disease, the same set of cardiac and renal conditions, along with diabetes, predicted subsequent heart failure, myocardial infarction, stroke, or unstable angina.
Cost analysis of 10,133 patients with adequate claims data revealed that the highest‑cost subgroup—constituting the top 10 %—was characterized by a combination of heart‑failure‑related procedure codes, infection and sepsis codes, cardiac diagnoses, and diagnostic testing.
Hospitalization codes paired with kidney diagnoses, as well as gastrointestinal diagnoses coupled with abdominal imaging, also marked the most expensive cases.
Kidney health matters.
Related: The Importance Of Regular Dental Checkups For Families
Across all three analytic streams, cardiovascular and kidney‑related diagnoses repeatedly surfaced as the strongest phenotypic signals. The authors suggest this pattern may reflect shared risk factors between heart disease and MASH, as well as the bidirectional relationship between type 2 diabetes and MASH that heightens CKD risk.
Given the reliance on real‑world data, the study acknowledges limitations such as incomplete medical records and the use of diagnosis codes in place of biopsy or imaging confirmation. The investigators call for further research to refine patient‑risk stratification and to explore more targeted monitoring strategies.
The consistent link between cardiovascular‑renal phenotypes and severe MASH outcomes could inform clinical pathways that prioritize early, non‑invasive screening for patients with these comorbidities, potentially reducing the need for invasive biopsies.
