Optimize nucleot(s)ide analogues’ to prevent hepatocellular carcinoma in patients with chronic hepatitis B: a lesson from real-world evidence

© The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, sharing, adaptation, distribution and reproduction in any medium or format, for any purpose, even commercially, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Optimize nucleot(s)ide analogues’ to prevent hepatocellular carcinoma in patients with chronic hepatitis B: a lesson from real-world evidence


INTRODUCTION
Hepatitis B virus (HBV) infection is a major global health problem that causes life-threatening liver diseases, including hepatocellular carcinoma (HCC), cirrhosis, and end-stage liver disease. Globally, 292 million people are infected with HBV, and about 1.0 million people die annually from HBV-related liver diseases, with an increasing trend [1,2] . Moreover, HBV infection causes more than 50% of HCC cases worldwide [3] . The incidence of HBV-associated HCC varies geographically, depending on regional exposure to risk factors and the availability of healthcare resources [2,4] . Most HCC cases occur in East Asia and sub-Saharan Africa, where viral hepatitis is an endemic [5] . The prevalence of HBV infection has been decreasing since the successful introduction of HBV vaccination programs, which are the primary prevention methods for HCC. However, in many countries, the overall clinical burden of HCC is still growing.
Three preventive measures may reduce HBV-related HCC development. Primary preventions may be achieved by vaccination, change in behavioral patterns, and prevention of mother-to-child-transmission. Secondary prevention includes effective antiviral treatment against the development of HBV-related HCC and surveillance measures to detect early stage of HCC in those infected by HBV. Tertiary prevention is an effort to reduce the recurrence of HBV-related HCC in patients who have been successfully treated for HCC.
Currently, in the field of chronic hepatitis B (CHB), there is much interest in real-world evidence (RWE) that complements the knowledge gained from traditional clinical trials and provides important information about uncommon events, long-term clinical outcomes such as HCC, comparative treatment effectiveness, disease burden, and many more topics. For the past decade, progress in research design and advancements in statistical methodology have allowed the widespread adoption of RWE in clinical CHB research.
In this context, the aims of the present review are to (1) outline the secondary prevention of HBV-related HCC by nucleos(t)ide analogues (NUCs), with a specific focus on RWE; and (2) suggest possible strategies to further reduce the risk of HCC in patients with CHB.

HEPATOCELLULAR CARCINOMA DEVELOPMENT IN PATIENTS WITH CHRONIC HEPATITIS B
HBV contributes to hepatocarcinogenesis via three mechanisms. The first involves chronic inflammation and hepatocyte regeneration due to persistent and sustained HBV replication. During this repetitive process, DNA mutation causes carcinogenesis [6] . Persistent and uncontrolled HBV replication is associated with an increased risk of HCC. One large-scale, untreated cohort study of more than 3,000 patients with CHB clearly showed a dose -response relationship between the serum HBV DNA level and the incremental risk of HCC [7] . In a long-term follow-up of the same cohort, persistently high viral load was associated with a higher rate of HCC than sustained low viral load [8] . Another study showed that serum hepatitis B surface antigen (HBsAg) level was also associated with HCC risk, especially in patients with low level viremia (HBV DNA < 2000 IU/mL) [9] . In this regard, antiviral treatment to decrease the development of HBV-associated HCC mainly focuses on reducing the level of HBV DNA, which is a primary function of NUCs [ Table 1]. The second mechanism involves chromosomal instability, whereby part of the HBV genome becomes integrated into the host hepatocyte genome, resulting in host DNA [10] . This mechanism of hepatocarcinogenesis can lead to HBV-associated HCC in patients with CHB who have no cirrhosis. Indeed, patients with occult hepatitis B infection (OBI) still carry the risk of HCC by this mechanism despite a negative serum HBsAg profile [11,12] . Therefore, the risk of HCC is not completely eliminated despite remarkable advances in antiviral treatment.
Liver cirrhosis itself is well known as a strong risk factor for HCC, and is characterized by tissue fibrosis and the conversion of normal liver architecture into structurally abnormal nodules [13] . If cirrhosis, regardless of etiology, is established, then the risk of HCC remains despite an effective antiviral treatment. Indeed, HBV-related HCC risk reduction by long-term antiviral treatment is more pronounced in patients without cirrhosis [14] . Several mechanisms have been postulated that induce hepatocarcinogenesis from cirrhotic liver, including telomere dysfunction, impaired hepatocyte proliferation by loss of replicative competition, and promoting tumor cell proliferation by an altered milieu [15] .

CHEMOPREVENTION OF HCC BY INTERFERON AND NUCS
Currently available antiviral treatments for CHB include interferon-a and NUCs. Interferon has been widely used as an antiviral treatment for many decades until orally available NUCs were introduced. Interferon-based antiviral treatment needs finite duration of treatment resulting in durable host immune control over HBV. Randomized trials and meta-analysis have demonstrated that interferon-based treatment reduces the risk of HCC by 40%-50% in treated compared to untreated patients [16][17][18][19] . Although interferons reduce the risk of HCC, their detailed usage and mechanisms are beyond the scope of this review. NUCs are an oral antiviral treatment; seven are currently available, namely, lamivudine, adefovir, telbivudine, entecavir (ETV), tenofovir disoproxil fumarate (TDF), tenofovir alafenamide, and besifovir. According to the international treatment guidelines for CHB, ETV, TDF, and tenofovir alafenamide are equally recommended as the preferred agents for treating CHB [20][21][22] .
Chemopreventive effects of various NUCs from previous studies are summarized in Table 1. Lamivudine, the first available NUC, showed a protective effect against HCC development in a randomized study with a hazard ratio (HR) of 0.49 [23] . This study is the only randomized, placebo-controlled trial comparing HCC prevention by an NUC and a placebo. Subsequent studies also show that lamivudine significantly reduces the risk of HCC in patients with CHB [24][25][26] . Two meta-analyses have indicated that treatment with NUCs reduced the risk of HBV-associated HCC by 74%-78% compared with treatment without NUCs [27,28] . However, the risk of HCC remained higher in patients with resistance-related virological breakthrough than in those with sustained virological response [29] , implying that sustained and uncontrolled HBV replication may not reduce the risk of HCC, even under NUC treatment.
Compared with the older generation of NUCs that includes lamivudine, adefovir, and telbivudine, ETV shows greater efficacy because it has a higher genetic barrier to viral resistance. Thus, ETV has begun to replace the old generation of NUCs among patients with resistant HBV and is recommended as a firstline antiviral treatment by the international guidelines since late 2000. A propensity score (PS)-matching study from Japan reported that patients with CHB who were treated with ETV showed a lower risk of 5-year cumulative HCC (3.7%) than an untreated historical cohort (13.7%; P < 0.001) [30] . In a cohort study involving 1980 patients with cirrhosis, treatment with ETV reduced the risk of HCC compared with controls (HR = 0.55) [31] . A Taiwanese multicenter study found that ETV treatment in patients with cirrhosis was associated with a 60% reduction in HCC risk compared with a cirrhotic control group [32] . TDF has been widely available since early 2010. In a simulation study, TDF was associated with a reduced incidence of HCC in patients without cirrhosis compared with untreated patients, showing a 60% reduction after 5 years of TDF treatment [14] . A multicenter international retrospective study also revealed that 5 years of TDF treatment reduced the risk of HCC in patients with CHB who had cirrhosis, compared with untreated patients [33] . A long-term follow-up study of Caucasian patients with CHB reported that treatment using either ETV or TDF significantly reduced the risk of HCC [34] . A recent meta-analysis involving 23 observational studies with 59,201 patients in the immune-active phase of infection showed that antiviral treatment decreased the risk of HCC compared with controls, with a HR of 0.5. The same study also analyzed 10 observational studies involving patients with cirrhosis and found that antiviral treatment reduced the risk of HCC, with a HR of 0.6 [35] .

CONTROVERSIES OVER THE EFFICACY OF HCC RISK REDUCTION AMONG NUC TYPES
Although ETV shows better efficacy in suppressing HBV DNA than old-generation NUCs, the HCC risk reduction is poorer than expected. The aforementioned Japanese study revealed that ETV treatment is associated with a significantly lower risk of HCC than treatment with lamivudine (P = 0.043) [30] . However, a large-scale retrospective study of 5,374 patients with CHB treated using either ETV or lamivudine demonstrated that ETV treatment conferred no lower risk of HCC than lamivudine (HR = 1.08, P = 0.48) [36] .
Another retrospective study used PS-matching analysis and showed similar risk for HCC between the two NUCs [37] .
A nationwide Korean cohort study by our group reported that patients with CHB who were treated using TDF had a significantly lower risk of HCC than those treated with ETV [38] . Many subsequent studies reported controversial findings [39][40][41][42][43] . A large-scale cohort study of 29,350 patients with CHB from Hong Kong demonstrated that TDF was associated with a lower risk of HCC than ETV, which was consistent with the nationwide Korean cohort study [40] . In contrast, a multicenter retrospective study from Korea showed no difference in the risk of HCC between these two NUCs [41] , and a multinational study also failed to show any significant difference in risk of HCC between ETV and TDF [42] . Recently, our group conducted a meta-analysis comparing the preventive effect against HCC between these two NUCs [44] . We included 15 studies comprising 61,787 patients with CHB who were treated using either ETV or TDF. The results suggested that TDF was associated with a significantly lower risk of HCC than ETV (HR = 0.80, P = 0.003) [44] . Even meta-analyses have presented controversial findings, depending on the characteristics of the included studies, ethnicity, and inclusion of decompensated cirrhosis [45][46][47][48] . Notably, our group recently used PS-matching to demonstrate that TDF was associated with significantly higher recurrence-free and overall survival rates than ETV in patients who had undergone curative-intent hepatectomy for HBVassociated HCC [49] . Collectively, most studies have either revealed that there is no difference in risk between the two treatments, or that TDF is associated with a lower risk of HCC [50] . No studies have favored ETV treatment over TDF in regards to HCC risk reduction. This controversy will likely remain unresolved until a randomized controlled trial is performed. However, it is unlikely that any such trial will be carried out, because HCC has a low occurrence, and any plausible trial would require a long-term observational period and a large number of patients [50] .

ADHERENCE AND DURATION OF TREATMENT INFLUENCE NUC'S CHEMOPREVENTIVE EFFECTS
Other factors affecting HCC prevention by NUCs include treatment adherence and duration. A study of 894 patients with CHB who were treated with ETV reported that poor adherence, defined as < 90% compliance, was associated with a higher mortality and greater risk of HCC, particularly among patients with cirrhosis [51] . Interestingly, the same group also reported that low levels of HBV DNA, despite the ETV treatment, were not predictive of HCC if patients showed good adherence to ETV treatment. This result emphasized the importance of drug compliance [52] . A nationwide Korean cohort study indicated that the incidence of HCC was incremental in a dose-dependent manner among poor-, intermediate-, and goodadherence groups, although the difference was not significant after multivariable adjustments [53] .
The treatment duration of NUCs for chemoprevention of HCC follows the treatment guidelines of CHB [20][21][22] . Currently, the realistic goal of CHB treatment is a so-called "functional cure", defined as the loss of HBsAg, regardless of HBsAb appearance. Therefore, once NUCs are started, treatment should be continued until this goal is achieved, even though controversies exist. A retrospective study involving 1,951 Caucasian patients with CHB demonstrated that the risk of HCC decreased after the first 5 years of ETV or TDF treatment [34] . However, a Korean retrospective study showed that the incidence of HCC did not significantly differ before and after 5 years of ETV treatment, suggesting that the lon-term risk of HCC in Asian patients with CHB may continue to persist [54] . This issue should be further validated and different NUCs must be compared to infer more conclusively.

WHO SHOULD BE TREATED WITH NUCS TO REDUCE THE RISK OF HCC?
The natural course of CHB is generally classified into four different phases based on serum HBV replicative status, with serum alanine aminotransferase (ALT) level used as a surrogate marker for significant liver inflammation. These phases include immune-tolerant, immune-active, inactive carrier, and hepatitis B e-antigen (HBeAg)-negative hepatitis. Current international treatment guidelines for CHB universally recommend antiviral treatment in patients with immune-active phase or HbeAg-negative hepatitis [20][21][22] . Moreover, patients with cirrhosis are advised to initiate antiviral treatment, regardless of serum HBV DNA level [20][21][22] . Most patients with CHB at a given time can be classified into one phase of the natural CHB course. However, CHB is a dynamic disease that interacts with HBV and host immunity. Hence, the phases of CHB change based on patient age, immunity, and viral replication status. In practice, a few patients with CHB, especially those without cirrhosis, reside in the so called "gray zone", which is not addressed in treatment guidelines. Current international treatment guidelines state that NUC should be initiated based on two main factors: serum HBV DNA and ALT level. Without evidence of cirrhosis, NUC should be started if the serum HBV DNA level is high (≥ 20,000 IU/mL for HBeAg-positive patients and ≥ 2,000 IU/mL for HBeAg-negative patients) or if ALT levels are abnormal [2× the upper limit of normal (ULN)]. However, current treatment criteria leave several unresolved questions. Specifically, are patients with CHB who do not meet the treatment initiation criteria definitely safe from developing HCC without the antiviral treatment? Current treatment initiation criteria stipulate that elevated serum ALT can be used as a surrogate marker of active liver inflammation because liver biopsy is not generally conducted for this purpose. However, mounting evidence indicates that serum ALT levels may not be a good criterion for deciding the initiation of antiviral treatment because it is altered under other non-HBV conditions, such as combined fatty liver disease and heavy alcohol assumption. In addition, previous studies have shown a poor association between serum ALT levels and the degree of necroinflammation in the hepatocytes of biopsied specimens, especially when serum ALT level is only slightly elevated (ALT < 2× ULN) [55,56] . In other words, the current treatment initiation criteria may delay antiviral treatment until significant active liver disease becomes apparent. This assumes that HBV infection is basically harmless in HBV carriers until many decades have passed and only when active liver disease is present [57] . Traditionally, the immune-tolerant phase refers to the immunologically and histologically dormant stage not requiring antiviral treatment, according to the treatment guidelines. In this regard, recent immunological studies have revealed that patients in the immune-tolerant and immune-active phases show a similar HBVspecific T-cell response, indicating that the immune-tolerant phase should not be considered completely benign, as previously assumed [58] . Another study suggested that hepatocarcinogenic processes, such as HBV DNA integration and clonal hepatocyte expansion, were already present during the early phase of CHB, whereas traditional concepts understand the immune-tolerant phase as benign [59] . Indeed, these in vitro studies were supported by RWE. A large-scale retrospective study from Korea demonstrated that untreated patients in the immune-tolerant phase had a higher risk of HCC than treated immune-active patients [60] . Another multicenter retrospective study from Korea showed that NUC treatment of patients currently in the immune-tolerant phase reduces the risk of HCC, even if the serum ALT level is below the upper limit of normal [61] . Another gray zone includes patients who are HBeAg-negative and have a high viral load, but have normal or slightly above normal ALT levels. As mentioned above, current treatment guidelines allow treatment initiation if the serum ALT level is more than 2× ULN. However, a retrospective study suggested that untreated HBeAg-negative, high viral-load patients without significant ALT elevation carried a higher risk of HCC than treated active phase patients with elevated serum ALT [62] . The degree of liver fibrosis resulting from hepatocyte injury by necroinflammation is also associated with the development of HCC.
To measure the degree of liver fibrosis, liver biopsy has traditionally been used. However, the clinical utility of liver biopsy is limited by its invasiveness, sampling variability, and inter-observer variability. Therefore, several non-invasive tests of fibrosis have been developed and validated [63][64][65] . These non-invasive assessments of liver fibrosis can aid the treatment decision-making process if a patient does not meet the traditional treatment criteria. Indeed, international guidelines recommend treatment initiation in patients with significant fibrosis by non-invasive assessments [20][21][22] . Liver biopsy should be preserved in selected cases where serologic and non-invasive tests are inconclusive in determining the disease activity. In addition, patients with first-degree family history of HCC, extrahepatic manifestations, or age > 40 years may be considered for treatments with antiviral agents considering their increased risk of HCC [66] .
Taken together, to prevent HBV-related HCC by using appropriate and timely antiviral treatment, the current treatment initiation criteria might be inadequate and could be expanded based on serum HBV DNA and ALT levels, degree of liver fibrosis, and personal characteristics [57] . Lastly, antiviral treatment is not recommended for patients with OBI by current international guidelines. Unlike overt HBV infection, patients with OBI are HBsAg negate but may have intermittent very low level of serum HBV DNA [11,12] . Indeed, patients with OBI still carry the risk of HCC and the risk of HCC may increase with combined known liver diseases such as chronic HCV infection. However, this area should be further investigated by future studies.

OTHER CONSIDERATIONS WITH NUC TREATMENT TO REDUCE HCC RISK
Many preventive interventions have been carried out in patients with CHB to lower the risk of HCC development, such as lifestyle modification and correct dietary habits [67] .

Alcohol
Alcohol consumption is an independent risk factor for HCC, and when patients with CHB drink alcohol, poorer outcomes are expected. The mechanisms of how alcohol interacts with HBV are not fully demonstrated, but it may accelerate viral replication, causing oxidative stress, or hinder the immune system [68] . Heavy alcohol consumption increases HCC risk significantly. A retrospective study in Taiwan compared alcoholic patients with CHB, non-alcoholic patients with CHB, and alcoholic patients. The results of a 10-year cumulative HCC incidence in these groups were 52.8%, 39.8%, and 25.6%, respectively (P < 0.001) [69] . In addition, alcohol consumption negatively affects the protective effect of NUCs on HCC. Cumulative alcohol consumption of more than 200 kg for a lifetime was posed as an independent risk factor for HCC in NUC-treated patients with CHB (HR = 2.21, P = 0.013) [30] .

Smoking
Smoking also confers an increased risk of HCC. According to a study in a European cohort, smoking, even former smoking, was highly related to HCC occurrence in patients with chronic viral hepatitis [former smoking, odds ratio (OR): 1.98, P = 0.092; current smoking, OR: 4.55, P = 0.001] [70] . Smoking also increases the risk of HCC-related mortality. Subjects who were free of malignancy were followed up until the end of the study period or death from HCC. The relative risk of HCC-related mortality was 1.4 in men and 1.1 in women [71] .

Metabolic syndrome
Fatty liver disease a crucial factor in HCC. Non-alcoholic fatty liver disease (NAFLD) is closely associated with diabetes, hyperlipidemia, and obesity. Diabetes is also associated with a two-fold increased risk of HCC [72] . Obese patients with CHB are more likely to have HCC. A prospective study was carried out in Taiwan in which 2,903 HBsAg-positive male patients were followed up for 14.7 years [73] . As body mass index (BMI) increased, so did the risk of HCC. Compared with normal-weighted patients (BMI = 18.5-24.9 kg/m 2 ), overweight patients (BMI = 25.0-29.9 kg/m 2 ) and obese patients (BMI > 30.0 kg/m 2 ) were more likely to be suffering from HCC or liver-related death, with HRs of 1.48 and 1.96, respectively [73] . To date, a specific diet affecting the clinical course of CHB is not identified. However, lifestyle modification with a healthy diet and proper exercise is important.

CONCLUSION
Recently, dramatic advancements have been made in HBV treatment; the development of an HBV vaccine and antiviral agents against HBV have tremendously improved clinical outcomes, including viral hepatitis and its fatal complications. Nevertheless, given the enormous clinical and socio-economical burdens of HCC caused by HBV, CHB should be properly managed with appropriate and timely use of antiviral treatments for secondary prevention. Unfortunately, no current drug can eliminate HBV nor can it completely eliminate the risk of HCC. However, many recent RWEs have provided important insights into secondary prevention of HCC, making clinicians reconsider the indications for antiviral treatment beyond the current treatment guidelines set forth for patients with CHB. Clinicians should be aware of the limitations of RWE before applying it in research, but the method may be used across a wide spectrum of CHB research through judicious selection of data sources, refinement of study designs, and appropriate analytic approaches. This will bring researchers a step closer to optimizing the secondary prevention of HCC in patients with CHB.

Authors' contributions
Responsible for the conception, review, drafting, and critical revision of the manuscript, and approved the final version of the manuscript: Lim J, Choi J

Availability of data and materials
Not available.

Financial support and sponsorship
There was no funding or support from industry for this study. The interpretation and reporting of the data were the sole responsibility of the authors.