Clinical bottom line
Drug-induced liver injury (DILI) remains a challenging diagnosis in clinical practice due to its heterogeneous presentation and the absence of specific biomarkers. Current management centers on early recognition, withdrawal of the offending agent, and supportive care. Pharmacotherapies such as N-acetylcysteine show benefit in select cases, but no universally effective treatment exists. Advances in diagnostic algorithms, including causality assessment tools and emerging machine learning models, enhance detection and prognostication. Novel oncological agents pose increasing hepatotoxicity risks, necessitating vigilant monitoring. Multidisciplinary approaches and adherence to evolving consensus guidelines improve patient outcomes (Andrade et al., 2020; Li et al., 2021; Riveiro-Barciela et al., 2025).
What the evidence shows
Andrade et al. (2020) provide a comprehensive overview of diagnostic and prognostic strategies in suspected DILI, emphasizing the importance of detailed drug history, exclusion of alternative causes, and use of structured causality assessment methods such as RUCAM. Their work highlights that early identification and cessation of the culprit drug remain the cornerstone of management.
Li et al. (2021) review current pharmacotherapies for DILI, noting that while N-acetylcysteine is effective in acetaminophen toxicity and some idiosyncratic cases, other agents like corticosteroids and ursodeoxycholic acid have limited and inconsistent evidence. They stress the need for individualized treatment based on injury pattern and severity.
Riveiro-Barciela et al. (2025) discuss challenges posed by novel oncological agents, including immune checkpoint inhibitors and targeted therapies, which can induce diverse liver injury patterns. They advocate for multidisciplinary management and tailored monitoring protocols to mitigate hepatotoxicity.
Hayashi et al. (2026) summarize practical updates from FDA DILI specialists, underscoring the role of standardized definitions, proactive monitoring during clinical trials, and the potential of emerging biomarkers to improve early detection and risk stratification.
Caveats and uncertainty
DILI diagnosis is primarily clinical and exclusionary; no definitive diagnostic test exists, complicating timely recognition.
Heterogeneity in DILI phenotypes and mechanisms limits generalizability of treatment recommendations.
Pharmacotherapies beyond N-acetylcysteine lack robust randomized controlled trial evidence.
Machine learning models for DILI prediction are promising but require further validation before routine clinical use.
Management guidelines continue to evolve alongside expanding drug formularies and novel therapeutics.
How this may change practice
Clinicians should maintain a high index of suspicion for DILI in patients with new liver enzyme elevations, especially when on high-risk medications.
Early drug discontinuation remains critical; prompt referral to hepatology may be warranted for severe or unclear cases.
Use of structured causality assessment tools can standardize evaluation and improve diagnostic confidence.
Emerging machine learning algorithms may soon assist in predicting DILI risk and guiding monitoring strategies.
Awareness of hepatotoxic potential of novel cancer therapies is essential for multidisciplinary care coordination.
Future integration of validated biomarkers and pharmacotherapies may enable more targeted and effective DILI management.