Comparison of nucleoside and nucleotide analogs in the recurrence of hepatitis B virus‐related hepatocellular carcinoma after surgical resection: A multicenter study

Abstract Background Antiviral therapy should reduce the recurrence of hepatitis B virus‐related hepatocellular carcinoma (HBV‐related HCC) after surgical resection. However, there is little research on whether various antiviral drugs have different prognostic effects in patients with HBV‐related HCC after curative liver resection. The present study compared the effects of nucleotide analog (NtA) and nucleoside analog (NsA) antiviral therapies after surgical resection on the prognosis of HBV‐related HCC. Methods A total of 1303 patients with HBV‐related HCC who received curative hepatectomy at five institutes between April 2014 and April 2019 were retrospectively enrolled and analyzed. Propensity matching analysis was used to compare the outcomes of HCC patients given NsA versus NtA therapy. Subgroup analysis of patients treated with entecavir (ETV) and tenofovir disoproxil fumarate (TDF) was also performed. Results Among 1303 patients, 759 (58.2%) patients developed recurrence, and 460 (35.3%) patients died. Multivariable analyses revealed that NtA therapy significantly decreased the risk of HCC recurrence (hazard ratio [HR], 0.64; 95% confidence interval [CI], 0.51–0.80; p < 0.001) and HCC‐related death (HR, 0.52; 95% CI, 0.36–0.76; p = 0.001) compared to that with NsA therapy. Subgroup analysis showed that TDF treatment was associated with significantly lower rates of HCC recurrence (HR, 0.64; 95% CI, 0.49–0.83; p = 0.001) and death (HR, 0.32; 95% CI, 0.20–0.50; p < 0.001) than ETV treatment. Conclusions Nucleotide analog treatment, but not NsA treatment, significantly reduced the risk of HCC recurrence in patients with HBV‐related HCC and improved overall survival after curative hepatic resection.


| INTRODUCTION
Hepatocellular carcinoma (HCC) is the sixth most frequently diagnosed cancer and the third most frequent cause of death worldwide. HCC is the fourth most common malignant tumor in China and the second leading cause of death 1,2 largely because hepatitis B virus (HBV) infection, which is the major cause of HCC development, is epidemic in China. 3,4 Hepatectomy is the mainstay of HCC treatment, and it leads to expected outcomes (5-year survival of 60%-80%) in well-selected candidates. 5,6 However, a high recurrence rate of approximately 60% after curative hepatic resection at 5 years is the main factor affecting the prognosis of HCC. 7 High serum levels of HBV-DNA are an important predictor of HCC recurrence. 8,9 Previous welldesigned studies demonstrated that antiviral therapy reduced the risk of postoperative recurrence of HBV-related HCC. [10][11][12] Recently, Choi et al. found that patients treated with tenofovir disoproxil fumarate (TDF) had a lower risk of postoperative HCC recurrence than those treated with entecavir (ETV), 13 but Lee et al. found no significant difference in the risk of recurrence and death between the ETV and TDF groups. 14 Currently, direct-acting antivirals include nucleotide analogs (NtAs) and nucleoside analogs (NsAs). The commonly used nucleos(t)ide analogs (NUCs) that act as direct-acting antivirals include but are not limited to TDF and ETV. Additionally, the potential impact of direct-acting antivirals on HCC recurrence is still a controversial topic. It is still unclear whether NtAs and NsAs have different effects on HCC recurrence rates in patients with HBV-related HCC after curative resection.
Therefore, we aimed to compare the different effects of NtAs and NsAs on HCC recurrence and overall survival (OS) in patients with HBV-related HCC after curative resection.

| Study population
Using a retrospectively collected database, we identified patients with HBV-associated HCC who underwent therapeutic liver resection in five participating institutions between April 2014 and April 2019 ( Figure 1). The participating institutions included The Affiliated Cancer Hospital of Guizhou Medical University, the Affiliated Hospital of Chengdu University, the People's Hospital of Leshan, Sichuan Provincial People's Hospital, and West China Hospital of Sichuan University. HBV-related HCC was defined as progression to HCC after the detection of serum hepatitis B surface antigen (HBsAg) for at least 6 months. 15,16 The patients were divided into NsA and NtA groups.
Baseline patient and tumor characteristics were obtained from electronic medical records at each medical center and included patient demographics, HBsAg, hepatitis B virus e antigen, serum HBV-DNA load, alpha fetoprotein (AFP), coagulation function, liver function, renal function, and hematological parameters. Hepatic pathologists assessed tumor characteristics in the excised specimens. The antiviral choice for each patient was based on their socioeconomic status and the preferences of each doctor.
This study was performed according to the World Medical Association Declaration of Helsinki and approved by the Ethics Committee on Biomedical Research, West China Hospital of Sichuan University (IRB No. 2021-311).

| Follow-up
All patients were followed up at 1 month postoperatively, every 3 months for the first 3 years and every 6 months for the next few years. The tumor evaluation and follow-up protocol included multiphasic contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI), physical examination, blood cell and differential counts, liver function tests, AFP levels, HBV markers, and HBV-DNA levels.
The primary endpoint was recurrence-free survival (RFS), and the secondary endpoint was OS. The index date was defined as the date of hepatectomy for HCC. RFS was defined as the interval between surgery and the first incidence of positive recurrence. OS was defined as the time interval between surgery and death of any cause or last follow-up. The last date of follow-up was 31 March 2021.
Tumor recurrence was defined as multiphasic contrastenhanced CT or MRI showing intrahepatic lesions with typical HCC enhancement characteristics, that is, contrast enhancement in the arterial phase and washout in the venous phase.

| Statistical analysis
To balance the baseline characteristics and minimize the effect of potential confounders, nearest-neighbor 1:2 propensity score matching (PSM) with a caliper size of 0.02 was used to reduce group differences in covariances between NsA and NtA patients (Figure 1). Propensity scores were calculated using the following 20 variables: age, sex, Barcelona Clinic Liver Cancer (BCLC) stage, tumor size, tumor number, microvascular invasion (MVI), diabetes, hypertension, serum HBV-DNA level, HBsAg, AFP, prothrombin time (PT), red blood cell count, hemoglobin, white blood cell count, platelets, aspartate aminotransferase (AST), alanine aminotransferase, albumin (ALB), and total bilirubin. Baseline characteristics were grouped into continuous and categorical variables. Continuous variables, which are reported as the means ± standard deviation, were compared between groups using the t-test or the Mann-Whitney U test. Categorical variables were compared using the χ 2 test or Fisher's exact test, and the results are expressed as numbers (n) or proportions (%). Cumulative RFS and OS rates were analyzed using the Kaplan-Meier method, and differences were compared using the logarithmic rank test. Univariable and multivariable Cox proportional risk regression analyses were performed to identify predictors associated with RFS and OS and assess risk factors that lead to recurrence and death.
All statistical analyses were performed using R statistical software version 4.04 (R Foundation for Statistical Computing) and SPSS software version 25.0 (SPSS). A twotailed p value < 0.05 was considered statistically significant.

| Patient characteristics
The study included 1303 patients: 1105 (84.8%) received NsA postoperatively; and 198 (15.2%) received NtA postoperatively. Table 1 shows the patient characteristics for  Propensity score matching adjustment resulted in 594 patients, and none of the parameters of the two groups continued to be significantly different ( Table 1). The characteristics between patients treated with ETV and TDF were not significantly different (Table 1).
Of the 567 PSM patients (396 in the NsA group and 198 in the NtA group), 228 (57.5%) patients in the NsA therapy group developed recurrence, and 90 (45.4%) patients in the NtA therapy group developed recurrence. There was a significant difference in RFS between patients who received NsA and NtA therapy, and the NtA group had a better RFS rate (HR, 0.67; 95% CI, 0.52-0.85; p = 0.001; Figure 3A). The 1-, 3-, and 5-year recurrence rates in the NsA treatment group were 34.8%, 55.9%, and 65.1%, respectively, and 19.9%, 42.4%, and 52.4%, respectively, in the NtA treatment group (p = 0.001, Figure 3A).
Multivariate Cox regression analyses of 1303 patients were performed to determine RFS predictors of recurrence in patients with HBV-related HCC after hepatectomy, and parameters significantly associated with recurrence risk in univariate analysis were incorporated into multivariate analysis. The multivariate Cox regression model included the entire cohort of 1303 patients. NtA treatment was associated with a significantly lower risk of HCC recurrence than NsA treatment (HR, 0.64; 95% CI, 0.51-0.80; p < 0.001;
Multivariable Cox regression analyses including 1303 patients revealed that NtA treatment was independent of other predictive factors, and the risk of death was significantly lower than NsA treatment (HR, 0.41; 95% CI, 0.29-0.59; p < 0.001;

| DISCUSSION
In this multicenter study, we compared the clinical outcomes of 1303 patients with HBV-related HCC treated with NtA or NsA therapy after curative resection. We found that NtA therapy was associated with decreased recurrence and increased OS compared to NsA therapy for patients undergoing R0 liver resection of HBV-related HCC. The subgroup analysis revealed that the TDF group had higher RFS and OS rates than the ETV group. This was consistently observed in propensity score-matched and multivariable-adjusted analyses.
Although there has been progress in the management of HCC, the high recurrence rate of tumors remains a major problem for HCC patients undergoing curative resection. There are no internationally recognized effective adjuvant therapies to prevent the postoperative recurrence of HCC. Well-known risk factors for HCC recurrence include advanced BCLC stage, multiple tumors, satellite lesions, large tumors, microvascular invasion, and high HBV load. 8,17 For patients with high-risk factors for recurrence, clinical intervention should be actively pursued because these tumor characteristics are not easily improved. In addition to tumor characteristics, inflammation plays a key role in tumorigenesis, and the inflammatory microenvironment is an important part of tumorigenesis. 18 Many carcinogenic microbial infections, such as HBV, are be associated with some forms of chronic inflammation. 19 Therefore, host hepatitis virus load is a correctable risk factor for HCC recurrence after therapeutic resection. 8 Previous studies showed that high levels of serum HBV-DNA and HBsAg seropositivity were associated with the development and recurrence of HCC. 8,17 Active antiviral therapy with NUCs is an effective treatment for the prevention of HCC recurrence after curative resection. 11,12,20 It has been consistently reported that oral treatment with NUCs reduces the risk of postoperative recurrence of HBV-related HCC. 10,11,20 The primary mechanism of NUC therapy is inhibition of the activity of HBV polymerase, which halts HBV replication 21 and inhibits the direct and indirect carcinogenic mechanisms of HBV. 22 Antiviral therapy prevents HBV reactivation, inhibits hepatitis activity, and reduces the inflammation of liver tissues, which leads to the regression of liver fibrosis and cirrhosis. 23 Many recent studies have explored the differential effects of TDF and ETV on risk of HCC in patients with chronic HBV infection. Two meta-analyses and two cohort studies reported that TDF treatment was associated with a lower risk of HCC than ETV treatment. [24][25][26][27] These studies found that the virological response rate in the early period was higher in patients treated with TDF than in those treated with ETV. This suggests that TDF may have advantages over ETV in terms of the prevention of HCC. Recently, one study showed no significant difference in the rates of HCC recurrence and death between the ETV and TDF treatment groups, 14 and two other studies showed that the rates of recurrence and death in the TDF group were significantly lower than those in the ETV group. 13,28 None of the studies showed that ETV was better than TDF. The potential mechanism is not clear. However, a recent study by Murata et al. showed that NtA induced the expression of interferon-λ3 (IFN-λ3) and inhibited the production of HBsAg. 29 IFN-λ exhibited effective antitumor activity in a mouse model of cancer. 30-32 IFN-λ3 directly inhibits tumor growth via the induction of apoptosis and/or cell cycle arrest and enhances host immunity by modulating innate and adaptive immune responses. 33 In addition, another study by Murata et al. revealed that only NtA therapy (adefovir dipivoxil [ADV] and TDF) has an additional pharmacological effect in modulating lipopolysaccharide-mediated cytokine production, which was expected to have a favorable immune response to the elimination of HBV. 34 Our result is in line with the research by Choi et al. 13 However, included NtA and NsA therapy, and they included only TDF and ETV therapy. At present, some low-income countries/regions may have problems concerning access to TDF or the high price of TDF. Our study suggests that other NtAs, such as ADV, seem to have the same ability to reduce the risk of HCC recurrence after curative resection. For these cases, we think they can use ADV instead.
Some previous studies suggested that the serum HBV-DNA level or HBsAg seropositivity was an independent risk factor for HCC recurrence and death. 8,17 Conversely, some studies found that serum HBV-DNA or HBsAg was not associated with prognosis in HCC. 12,35 Our study found that a higher preoperative serum HBV-DNA level was not an independent risk factor for HCC recurrence, but it was an independent risk factor for death. One possible reason is that all patients were treated with antiviral agents, and the relative risk of viral status was low and was overshadowed by stronger risk factors.
Our study has several limitations. First, this study was a retrospective study. Although we attempted to compensate for potential bias by propensity matching variables associated with treatment outcomes, selection bias remained a possibility. Second, in the subgroup analysis, the TDF group was comprised of much fewer patients than the ETV group because TDF was not approved for use in China until after 2016. Due to limited experience and possible renal F I G U R E 4 Kaplan-Meier analysis of (A) RFS and (B) OS among 432 patients receiving ETV treatment or TDF treatment. ETV, entecavir; TDF, tenofovir disoproxil fumarate injury, physicians might recommend TDF more strictly and carefully, leading to patient selection bias. Third, our study explored the impact of two classes of antiviral drugs on the prognosis of HBV-related HCC, and the efficacy of antiviral agents in the same class of drugs may differ.
In conclusion, NtA, especially TDF, is a highly potent NUC that effectively reduces HCC recurrence and prolongs postoperative survival. Our study suggests that if patient circumstances permit, NtA therapy, especially TDF, should be preferred in patients with HBV-related HCC after curative resection. Our findings may have considerable clinical significance for the prevention of HCC recurrence in patients. *Variables found significant at p < 0.1 in univariable analyses were entered into multivariable Cox regression analyses.