Back

Predicting Real-life Drinking Scenarios through a Physiological Digital Twin Incorporating Secondary Alcohol Markers

Podeus, H.; Simonsson, C.; Jakobsson, G.; Kronstrand, R.; Nyman, E.; Lövfors, W.; Cedersund, G.

2025-11-15 systems biology
10.1101/2025.11.14.686953 bioRxiv
Show abstract

Alcohol consumption poses significant societal challenges, necessitating accurate tools for detecting at-risk drinking. Various biomarkers reflect alcohol intake over different timeframes. Blood alcohol concentration (BAC) and breath alcohol concentration (BrAC) are commonly used for short-term detection, particularly in forensic contexts such as driving under the influence of alcohol (DUIA) cases. Moreover, the rapid kinetics of BAC and BrAC limit their utility in determining the precise timing of intake--an essential factor in legal cases involving defenses like the hipflask argument. To address this, secondary metabolites including ethyl glucuronide (EtG), ethyl sulphate (EtS), and urine alcohol concentration (UAC) offer slower, more time-sensitive profiles. Combining these markers could enable a more accurate reconstruction of past alcohol consumption events. Traditionally, mathematical models have been used to help extract information from the dynamics of alcohol-related markers. Existing mathematical models typically focus on primary markers or single secondary markers in isolation. In this study, we present an extended mathematical model that integrates all markers; BAC, EtG, EtS, and UAC into a unified framework, expanding our previous physiological twin model. Our updated model enables personalized simulations of alcohol metabolism and intake timing, which in turn creates the fundament for enhancing forensic assessments and supporting applications where accurate temporal analysis of alcohol consumption is critical. To facilitate accessibility and practical use of this analysis, we have implemented and provided an interactive web application.

Published in Scientific Reports (predicted rank #3) · training set

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.