Bayesian Forecasting Tool to Predict the Need for Antidote in Acute Acetaminophen Overdose. Issue 8 (26th July 2017)

• Record Type: Journal Article
• Authors: Desrochers, Julie; Wojciechowski, Jessica; Klein‐Schwartz, Wendy; Gobburu, Jogarao V.S.; Gopalakrishnan, Mathangi
• Is Part Of: Pharmacotherapy. Volume 37:Issue 8(2017)
• Page Start: 916
• Page End: 926
• Publication Date: 2017-07-26
• Subjects: acetaminophen overdose -- Rumack–Matthew nomogram -- pharmacometrics -- Bayesian forecasting -- emergency physicians; Chemotherapy -- Periodicals; Pharmacology -- Periodicals; Drug Therapy -- Periodicals; Pharmacology -- Periodicals; 615.1
• DOI: 10.1002/phar.1972
• Languages: English
• ISSNs: 0277-0008

Abstract

Abstract : Study Objective: Acetaminophen (APAP) overdose is the leading cause of acute liver injury in the United States. Patients with elevated plasma acetaminophen concentrations (PACs) require hepatoprotective treatment with N ‐acetylcysteine (NAC). These patients have been primarily risk‐stratified using the Rumack–Matthew nomogram. Previous studies of acute APAP overdoses found that the nomogram failed to accurately predict the need for the antidote. The objectives of this study were to develop a population pharmacokinetic (PK) model for APAP following acute overdose and evaluate the utility of population PK model–based Bayesian forecasting in NAC administration decisions. Design, Patients and Measurements: Limited APAP concentrations from a retrospective cohort of acute overdosed subjects from the Maryland Poison Center were used to develop the population PK model and to investigate the effect of type of APAP products and other prognostic factors. The externally validated population PK model was used a prior for Bayesian forecasting to predict the individual PK profile when one or two observed PACs were available. The utility of Bayesian forecasted APAP concentration–time profiles inferred from one (first) or two (first and second) PAC observations were also tested in their ability to predict the observed NAC decisions. Main Results: A one‐compartment model with first‐order absorption and elimination adequately described the data with single activated charcoal and APAP products as significant covariates on absorption and bioavailability. The Bayesian forecasted individual concentration–time profiles had acceptable bias (6.2% and 9.8%) and accuracy (40.5% and 41.9%) when either one or two PACs were considered, respectively. The sensitivity and negative predictive value of the Bayesian forecasted NAC decisions using one PAC were 84% and 92.6%, respectively. Conclusion: The population PK analysis provided a platform for acceptably predicting an individual's concentration–time profile following acute APAP overdose with at least one PAC, and the individual's covariate profile, and can potentially be used for making early NAC administration decisions.