Predicting Outcomes in Acetaminophen Overdose Using Real Patient Data: The Role of Protein Adducts in Determining Survival and Mortality

In the United States, acetaminophen (APAP) overdose is the leading cause of acute liver injury. Current assessment tools like the King’s College Criteria (KCC) cannot predict APAP dosage or overdose timing, crucial for guiding treatment decisions. The Model for APAP-Induced Liver Damage (MALD) employs a system of differential equations using biomarkers namely AST, ALT, and INR, to estimate overdose parameters and determine whether NAC therapy suffices or if transplantation is needed. However, these biomarkers are not specific to APAP toxicity. Our work enhances MALD by incorporating a fourth biomarker, APAP-protein adducts, validated with real patient data from the Acute Liver Failure Study Group. The inclusion of Adducts significantly improves the model’s predictive accuracy, reflected in higher ROC curve performance, further enhanced when patients with alcohol use are excluded. We also analyze the asymptotic behavior of the extended model and conduct sensitivity analyses on key parameters, aiming to refine predictive capabilities. Ultimately, this extension aims to improve early diagnosis and treatment decisions for APAP-induced liver injury, advancing beyond existing models like KCC by providing more precise overdose estimation and prognosis.