To date, it has been confirmed that the HBV DNA level is positively correlated with the up-regulated expression of PD-1 on T cells, which is closely linked to the formation of HCC immunosuppressive microenvironment [19, 20]. The PD-1/PD-L1 axis also plays an important role in HBV replication [21]. Recently, the combination regimens of PD-1/PD-L1 inhibitor and an antiangiogenic therapy have been proven to be an optimal treatment for advanced HCC [7, 8, 22, 23]. However, the interaction between HBV load and anti-PD-1/PD-L1 therapy remains controversial, particularly in patients who do not receive a continuous antiviral therapy. Several studies have reported high HBV load as a risk factor for HBV reactivation and hepatic impairment during anti-PD-1/PD-L1 therapy [24, 25], while other reports demonstrated that PD-1/PD-L1 inhibitors can be safe and effective in cancer patients with either chronic HBV or HCV infection [26, 27]. Different antiviral agents, such as lamivudine (LAM) or adefovir dipivoxil (ADV), may account for the controversial results in studies reporting HBV load to be a risk factor for HBV reactivation. However, there is no data demonstrating this issue, since the most commonly used agent was entecavir (ETV) or tenofovir (TDF) in the currently reported studies, where LAM and ADV were no longer recommended as first-line anti-viral regimens. Most of the relevant studies did not explore whether antiviral therapy can improve the efficacy and safety of anti-PD-1 treatment in combination with an antiangiogenic therapy. In the current study, we found that baseline HBV load did not affect the prognosis of HCC patients receiving anti-PD-1 combined with an antiangiogenic therapy, while PD-1 inhibitors did not aggravate HBV reactivation and hepatic impairment in patients given TAF prophylaxis.
Since evidence regarding whether HBV infection affects the prognosis of HCC patients receiving anti-PD-1 based therapy is scarce, patients with a high baseline HBV DNA level were always excluded from clinical trials regardless of the antiviral strategies utilized, limiting their efficiency and generalizability. In the KETNOTE-224 study, tumour response was comparable between patients with and without HBV/HCV infection [27]. Similarly, the CheckMate 040 study reported similar tumour responses among patients with advanced HCC, irrespective of HCC aetiology [26]. However, patients with a higher baseline HBV DNA level (usually > 500 IU/mL or > 2000 IU/mL) were excluded, and whether baseline HBV DNA level affected the clinical prognosis of HCC patients receiving anti-PD-1 based therapy was not assessed in the above clinical trials. In a retrospective study in China, the baseline HBV load was found to have no significant impact on the prognostic outcomes or rates of hepatic impairment during anti-PD-1 blockade [28]. According to our results, similar ORR and DCR were observed in patients with low and high baseline HBV DNA levels. In addition, there was no significant difference in PFS between patients with a higher or lower baseline HBV loads. Importantly, our data highlight that HBV load may not affect the prognosis of HCC patients receiving anti-PD-1 therapy combined with an antiangiogenic therapy.
Whether PD-1 inhibitors aggravate HBV reactivation and hepatic impairment is another concern of anti-PD-1-based therapies. In a phase Ib study comparing nivolumab with and without an HBV therapeutic vaccine, in virally suppressed patients with HBeAg (−) chronic HBV, PD-1 inhibitor was demonstrated to be well tolerated and led to HBsAg decline in most patients [29]. In a study comparing HBV reactivation between patients with low and high HBV DNA loads, who were undergoing anti-PD-1 blockade treatment, similar incidences of HBV reactivation and HBV-associated hepatitis were observed [30]. In the current study, only 2 of the 70 patients (2.9%) experienced HBV reactivation, which was a lower rate compared to patients with other cancer types in another study [14]. The reason for this discrepancy may be that all patients in our study simultaneously received TAF prophylaxis. Continuous and effective antiviral treatment was shown to improve the prognosis of HCC patients receiving anti-PD-1 blockade with high viral loads in our previous study (recently accepted article, https://doi.org/10.21037/atm-21-3020). Nevertheless, the specific role of TAF in the protection against HBV reactivation or hepatic impairment has not been elucidated, since TAF has been proven to have a greater plasma stability and higher renal safety than TDF. In addition, we did not observe any cases of HBV-related hepatic impairment during the follow-up period. Taken together, we suggest that HBV-HCC patients accept first-line antiviral prophylaxis such as TAF before and during the period of anti-PD-1-based therapy.
The current study is not free from certain limitations. First, this single-arm study was designed retrospectively, which may have caused bias in the selection of patients. The implications of this study need to be verified by future clinical studies with larger sample sizes. Second, the overall survival (OS) data were not included in the analysis, as the follow-up period was not long enough, and only two patients died until the observation deadline. Finally, patients with HCV infection were excluded from the final analysis, and the influence of HCV loads on these patients remains unclear.
In conclusion, our study provides evidence that baseline HBV loads do not affect the prognosis of HCC patients receiving anti-PD-1 in combination with antiangiogenic therapy, while PD-1 inhibitors do not aggravate HBV reactivation and hepatic impairment in patients given TAF prophylaxis. However, as this was a non-randomized retrospective study, our data should not be taken as non-biased or used to guide clinical decisions without a further proof derived from prospective clinical trials. Besides, future prospective studies should also pay attention to the interaction between HBV reactivation and the combination therapy of anti-PD-1 and anti-CTLA-4, since trials are ongoing to explore the possibility of CTLA-4 in combination with anti-PD-1/PD-L1.