Letter: Positivity of High-Sensitivity HBsAg Test Was Significantly Associated With Poorer Prognosis in Patients With Non-HBV-Related HCC—Authors' Reply

Abstract

We thank Dr. Li [1] for his interest in our study [2]. We demonstrated that 11.8% (34/288) of patients with non-hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) and previous HBV infection (pHBV) were positive for HBsAg by a high-sensitivity hepatitis B surface antigen (hHBsAg+) assay, and that they exhibited more advanced stages of HCC, higher AFP levels and increased rates of vascular invasion. In patients with non-HBV-related HCC, hHBsAg+, rather than pHBV, was significantly and independently associated with a poorer prognosis [2]. We concur with Dr. Li that the presence of competing risks in the Cox proportional hazard model may have led to an overestimation. Consequently, we performed another analysis using a competing risk model. Among the 401 patients analysed, 44 (10.1%) died from causes unrelated to HCC or decompensated cirrhosis. The cumulative incidence of non-HCC- or liver-related deaths was considered a competing risk factor. We conducted a Fine-Gray proportional hazard regression analysis to identify the factors associated with overall survival across several patient cohorts: the entire cohort of patients with non-HBV-related HCC (n = 401), non-B-non-C aetiology cohort (n = 226), BCLC stage B/C/D cohort (n = 153) and non-B-non-C aetiology and BCLC stage B/C/D cohorts (n = 105) (Table 1). In addition to factors such as BCLC stage, Child-Pugh-grade B/C, platelet count, AST, γ-GTP and AFP levels in the whole cohort, hHBsAg+ (HR: 2.555; 95% CI: 1.372–4.758; p = 0.003) was significantly associated with shorter overall survival in the competing risks model.

Similarly, in the non-B-non-C-aetiology cohort (HR: 4.502; p < 0.001), BCLC stage B/C/D cohort (HR: 4.549; p < 0.001) and combined non-B-non-C-aetiology and BCLC stage B/C/D cohorts (HR: 7.072; p < 0.001), hHBsAg+ was also significantly associated with shorter overall survival in the competing risk model (Table 1). Therefore, hHBsAg+ was identified as a significant prognostic factor, even when using the competing risks model.

Additionally, we concur with Dr. Li that the consideration of additional potential confounding factors such as treatment methods, nutritional status and family income may further strengthen the findings. We accounted for novel variables such as treatment modalities and nutritional status and conducted a Fine-Gray proportional hazard regression analysis. Nutritional status was evaluated using the Geriatric Nutritional Risk Index [3, 4]. Treatment modalities were classified into curative (surgery or radiofrequency ablation) and non-curative (TACE, systemic chemotherapy or nontreatment) categories. The Fine-Gray competing proportional hazard regression analysis revealed that, in addition to factors such as BCLC stage, Child-Pugh grade B/C, platelet count, AST levels, γ-GTP levels and AFP levels, non-curative treatments (HR: 1.833; 95% CI: 1.186–2.833; p = 0.0064) and hHBsAg+ (HR: 2.392; 95% CI: 1.293–4.615; p = 0.009) were significantly associated with shorter overall survival.

We agree that conducting subgroup analyses based on gender is important because of the significant gender difference of men having a higher probability of developing HCC than women. However, only five hHBsAg+ cases were identified in this cohort of women owing to biological factors that limited the inclusion of a larger number of women for a more thorough investigation. We hope that this issue will be addressed in future studies.

Naohiro Yasuura: conceptualization, investigation, writing – review and editing, data curation, formal analysis. Goki Suda: conceptualization, methodology, data curation, formal analysis, writing – original draft, funding acquisition. Masatsugu Ohara: investigation, writing – review and editing, formal analysis. Naoya Sakamoto: writing – review and editing, investigation, funding acquisition.

The Ethics Committee of Hokkaido University Hospital confirmed that the study protocol (IRB number: 019-0400) conformed to the ethical guidelines of the Declaration of Helsinki. Patients who provided written informed consent or did not decline to participate were included in this study. The ethics committee approved the inclusion of patients who did not decline to participate in the study instead of seeking for written informed consent from the included patients.

Professor Naoya Sakamoto received lecture fees from Bristol Myers Squibb and Pharmaceutical K. K.; grants and endowments from MSD K. K. and Chugai Pharmaceutical Co., Ltd.; and research grants from Gilead Sciences Inc. and AbbVie Inc. Dr. Goki Suda received research grants from Bristol Myers Squibb, MSD K. K., and Gilead Sciences. The authors declare no conflicts of interest.

This article is linked to Yasuura et al papers. To view these articles, visit https://doi.org/10.1111/apt.18229 and https://doi.org/10.1111/apt.18300.