New Model Performs Well for Detecting Alcohol-Associated Liver Disease

WEDNESDAY, March 4, 2026 (HealthDay News) — A novel predictive model demonstrates good performance for detecting metabolic dysfunction and alcohol-associated liver disease (MetALD) and alcohol-associated liver disease (ALD), according to a study published online Feb. 25 in Gastroenterology.

Federica Tavaglione, M.D., Ph.D., from the University of California at San Diego in La Jolla, and colleagues developed a novel, accurate, and scalable indirect alcohol biomarker panel to screen for MetALD-ALD in general population settings. The derivation cohort included 503 community-dwelling adults with overweight/obesity and steatotic liver disease with magnetic resonance imaging/magnetic resonance elastography assessment and phosphatidylethanol (PEth) testing. Bidirectional stepwise logistic regression was used to develop the optimal predictive model, which was validated internally using 2,000 bootstrap samples. The area under the receiver operating characteristic curve (AUROC) was used to assess model performance. External validation was conducted in an independent cohort of 1,777 Swedish individuals with PEth measurements.

The researchers included sex, mean corpuscular volume, gamma-glutamyltransferase, high-density lipoprotein cholesterol, and hemoglobin A1c in the final predictive model (MetALD-ALD Prediction Index [MAPI]). Good model performance was seen in the derivation and validation cohorts (AUROC, 0.76 and 0.75, respectively). Among all commonly used indirect alcohol biomarkers, MAPI was the top-performing model, based on AUROC.

“Our goal was to build something practical,” Tavaglione said in a statement. “These lab values are already part of standard care, so MAPI can be implemented immediately without adding cost or complexity for clinics.”

Several authors disclosed ties to the biopharmaceutical industry.

Abstract/Full Text (subscription or payment may be required)