A series of cohort studies revealed the real-world incidence rates of severe acute liver injury following the initiation of drugs suspected to be hepatotoxic.
In the study, published in JAMA Internal Medicine, investigators analyzed the data of nearly 8 million patients without preexisting liver or biliary disease who initiated one of 194 suspected hepatotoxic medications between 2000 and 2021. They noted that the patients had a mean age of 64.4 years, and 92.5% of them identified as male. Further, 55.1% of the patients had polypharmacy—defined as taking five or more drugs—and a significant proportion of them had comorbidities such as diabetes, hyperlipidemia, and obesity.
The investigators found that 17 of the drugs had the highest rates of severe acute liver injury (ALI), ranging from 5 to more than 10 events per 10,000 person-years. Among these 17 drugs, 64% of them were not included in the highest hepatotoxicity category based on published case reports. Antimicrobial drugs made up the majority (64%) of the drugs with the highest rates of severe ALI in the study. The incidence rates ranged from 0 events per 10,000 person-years for some drugs to as high as 86.4 events per 10,000 person-years for the antiretroviral drug stavudine.
Sensitivity analyses were then conducted to assess the robustness of the findings. Adjusting for comorbidities, stratifying by sex, limiting the analysis to younger patients, and examining drug-drug interactions and polypharmacy did not substantially alter the results. The study provided data on the sample sizes, number of events, and incidence rates for all drug groups (1 to 5), which were categorized based on the observed rates of severe ALI.
The investigators emphasized that relying solely on case report counts did not accurately reflect the observed rates of severe ALI after initiating these drugs. They proposed that their systematic approach of using data from real-world electronic health records to measure ALI rates—while accounting for the number of people exposed and censoring non–drug-related causes—provided a more objective way to screen for and investigate drug hepatotoxicity signals.
The findings suggested that patients initiating the drugs identified as high risk may need closer liver function monitoring. The investigators noted that electronic health record systems could be leveraged to alert clinicians about the potential for severe ALI when ordering these drugs and to recommend appropriate laboratory monitoring.
However, the study had some limitations. The predominantly male study population and the inability to capture all herbal supplements or drugs could limit the generalizability of the findings. Additionally, the reliance on laboratory tests for ALI ascertainment may have introduced surveillance bias, and the use of U.S. Department of Veterans Affairs (VA) data might have led to an underestimation of the rates caused by patients receiving care outside the VA system.
This study demonstrated that real-world evidence could identify the drugs most likely to induce hepatotoxicity and serve as a tool to investigate safety signals that are currently based primarily on published case reports. The investigators indicated that future studies should evaluate ALI rates in patients with chronic liver disease to better guide drug safety in that population. The data and methodology used in this study can inform further research and clinical decision-making regarding drug-induced liver injury.
Full disclosures can be found in the original study.
