We read with great interest the article by Hur et al., entitled “Extrahepatic malignancies and antiviral drugs for chronic hepatitis B: A nationwide cohort study” published recently in Clinical and Molecular Hepatology [1]. The study provides important insights into the long-term risks of extrahepatic malignancies in chronic hepatitis B patients treated with entecavir (ETV) or tenofovir disoproxil fumarate (TDF). We would like to raise some additional methodological concerns that could help strengthen the study’s findings.

First, the current use of administrative codes to define the study population of viral hepatitis could lead to potential misclassification bias. While large administrative databases provide valuable population-wide data, reliance on ICD-10-CM coding systems for disease classification can introduce bias in diagnosis definitions [2], a limitation frequently mentioned in population-based real-world studies [3,4]. A recent validation study has shown that the positive predictive value (PPV) of ICD-10 codes for hepatitis B and C can be variable and often suboptimal [5]. Lacking precision in diagnostic codes may lead to false-positive diagnoses of hepatitis B and C, potentially influencing the observed relationship between antiviral therapy and cancer risk, and making it challenging to draw accurate conclusions about the comparative effects of ETV and TDF on subsequent malignancies.

The issue of comedication is another important factor that could introduce confounding into the study’s findings. Common medications such as aspirin or other nonsteroidal anti-inflammatory drugs (NSAIDs) have known effects on cancer risk, particularly for gastrointestinal cancers, due to their anti-inflammatory and antithrombotic properties [6]. Aspirin, for instance, has been reported to reduce the risk of esophageal and gastric cancers in real-world observeations [7]. Similarly, aside from aspirin and NSAIDs, other common medications, such as statins, which were also reported to be associated with gastric cancer risk reduction, could further complicate the interpretation of results [8]. Given that in the current study, the detailed profile of comedication use were not provided, readers were not able to determine whether study populations were more likely to be taking medications that could influence cancer risk. To improve the robustness of the findings, future studies were recommended include a detailed profile of cancer-associated comedication status to assess the potential impact of these medications on incident malignancy outcomes.

Lastly, adherence to antiviral therapy was not measured in the study. Poor adherence could lead to poor viral control and potentially increasing the risk of both intrahepatic and extrahepatic malignancies [9]. It would be valuable for future studies to assess adherence by utilizing pharmacy refill data or patient self-reports, which would allow researchers to further examine the potential role of adherence levels and malignancy outcomes.