A new agent-based modeling study published in JAMA Network Open reveals alarming projections about the growing burden of liver disease in the United States, with metabolic dysfunction-associated steatotic liver disease (MASLD) expected to affect over 121 million Americans by 2050.
Rising Disease Burden and Population Impact
The sophisticated analytical model, which simulated over 2.8 million individuals, projects MASLD prevalence will climb steadily from 33.7% (86.3 million people) in 2020 to 36.8% (101.2 million) by 2030, ultimately reaching 41.4% (121.9 million) by 2050. The most dramatic increase is expected among older adults, with a striking 300% rise in cases among those aged 80 or older.
Disease Progression and Severity Trends
The study reveals concerning patterns in disease progression. While in 2020, nearly half (48.4%) of MASLD patients had no fibrosis (F0), the disease profile is expected to become more severe by 2050. The proportion of patients with advanced fibrosis stages (F2-F4) is projected to increase significantly.
Metabolic dysfunction-associated steatohepatitis (MASH), a more severe form of the disease, is predicted to affect 7.9% of US adults (23.2 million people) by 2050, up from 5.8% (14.9 million) in 2020. Cases of MASH with clinically significant fibrosis are expected to surge from 6.7 million to 11.7 million individuals.
Critical Healthcare System Implications
The model predicts a substantial increase in severe liver-related complications:
- Hepatocellular carcinoma (HCC) cases are projected to nearly double, rising from 11,483 new cases annually in 2020-2025 to 22,440 cases per year by 2046-2050
- Liver transplant needs are expected to quadruple, increasing from 1,717 to 6,720 annual cases over the same period
- Annual liver-related deaths are predicted to more than triple, from 30,500 in 2020 to 95,300 by 2050
Methodology and Model Validation
The research employed an agent-based state transition model with a yearly cycle and lifetime horizon. This microsimulation approach captured population heterogeneity and individual-level variations, ensuring more accurate predictions. The model's validity was demonstrated by its accurate replication of US population growth from 2000 to 2020, with yearly estimates within 0.5% of observed data, and close alignment with National Health and Nutrition Examination Survey estimates.
