The Risk of Developing Hepatocellular Carcinoma Persists in Chronic Hepatitis B Patients Even After the Long-Term Administration of Nucleos(t)ide Analogs

1 Introduction

Hepatocellular carcinoma (HCC) is currently the third most common cause of cancer-related death globally [1], and chronic hepatitis B virus (HBV) infection is recognized as a significant risk factor. HBV infection poses a significant global public health challenge. Approximately 296 million individuals worldwide are estimated to be hepatitis B surface antigen (HBs antigen)-positive, with 820,000 deaths occurring each year due to HBV-related liver cirrhosis and hepatocellular carcinoma [2].

Currently, the anti-HBV therapies available for clinical use are interferon (IFN) and nucleos(t)ide analogs (NUCs) [3, 4]. Treatment with NUCs lowers serum HBV DNA titers and reduces the risk of HCC [5-8]. Even in patients who achieve viral suppression after several years of NUC therapy, the risk of HCC persists [5, 8]. The proportion of patients with HBV infection as the etiology of HCC has not decreased worldwide [9-11].

Currently, many chronic hepatitis B patients receive long-term NUC treatment. However, it is not clear whether the risk of HCC changes after prolonged treatment with NUCs, that is, treatment lasting for more than 5 or 10 years in chronic hepatitis B (CHB) patients. Evaluating the risk of HCC in patients receiving long-term NUC treatment and the effect of long-term NUC treatment on HCC is important for HCC surveillance in CHB patients receiving NUC treatment. In the present study, we collected data from CHB patients who received initial treatment with NUCs between July 2000 and March 2019 at multiple centers. Here, we identified the risk factors linked to the development of incident HCC in CHB patients receiving long-term NUC treatment and the suppressive effect of prolonged NUC therapy on the occurrence of incident HCC.

2 Methods

3 Results

3.2 Cumulative Incidence of HCC After the Initiation of NUC Treatment and Risk Factors for Incident HCC

Among the 737 patients, 147 developed HCC; 63 patients were positive and 81 were negative for HBe antigen, and 81 patients had cirrhosis, whereas 66 did not. The cumulative incidence of HCC after the initiation of NUC was 11.2% at 5 years, 18.4% at 10 years, and 23.1% at 15 years (Figure 1A). The Cox proportional hazards model revealed that older age, male sex, the presence of cirrhosis, a low platelet count, a low ALT level, and a high γ-GTP level at the time of NUC treatment initiation were independent risk factors for incident HCC (Table 2). The HBs antigen was identified as a risk factor for incident HCC in the univariate analysis; however, it was excluded from the multivariate analysis because 49.0% of the data were missing (Table 2). The type of NUC at the start of NUC treatment was not identified as a risk factor for incident HCC in either the univariate analysis (Table 2) or the multivariate analysis (Supporting Information S3: Table S2). When patients were analyzed separately by HBe antigen status, the Cox proportional hazards model identified older age and cirrhosis as independent risk factors for incident HCC in HBe antigen-positive patients, whereas older age, male sex, low platelet count, and high γ-GTP levels were independent risk factors in HBe antigen-negative patients (Supporting Information S3: Tables S3–S6).

TABLE 2. Risk factors at the start of NUC for incident HCC.

			Univariate analysis			Multivariate analysis
			HR	95% CI	p-value	HR	95% CI	p-value
Age	years old	per 1-year increase	1.06	1.04–1.07	< 0.0001	1.05	1.03–1.07	< 0.0001
Sex		Male/female	1.54	1.08–2.18	0.0134	1.69	1.11–2.57	0.0116
Liver fibrosis		Cirrhosis/noncirrhosis	6.13	4.43–8.50	< 0.0001	2.63	1.61–4.28	< 0.0001
BMI	kg/m2	≥ 23/< 23	1.07	0.73–1.57	0.7197
Average daily alcohol intake	g	≥ 20/< 20	1.10	0.70–1.71	0.6817
Diabetes mellitus		Presence/absence	2.19	1.47–3.25	0.0003	1.01	0.61–1.66	0.9838
NUC at the start of NUC treatment		LAM, ETV/TDF, TAF	6.65	0.93–47.7	0.0593
HBs antigen	IU/mL	≥ 250/< 250	9.67	1.34–69.7	0.0006
HBe antigen		Positive/negative	0.85	0.61–1.19	0.3455
HBV DNA	Log IU/mL	≥ 3.0/< 3.0	1.68	0.62–4.53	0.2672
HBcrAg	Log U/mL	≥ 3.0/< 3.0	0.48	0.09–2.49	0.3848
Platelet	× 104/μL	per 1 × 104/μL increase	0.86	0.83–0.89	< 0.0001	0.94	0.89–0.99	0.0135
AST	U/L	per 1 U/L increase	0.999	0.998–1.001	0.4719
ALT	U/L	per 1 U/L increase	0.99902	0.99782–0.99999	0.0467	0.99852	0.99679–0.99992	0.0361
γ-GTP	U/L	per 1 U/L increase	1.002	1.001–1.003	0.0008	1.003	1.002–1.005	0.0005
Total bilirubin	mg/dL	per 1 mg/dL increase	1.11	1.04–1.16	0.0047	1.09	0.94–1.26	0.2433
Albumin	g/dL	per 1 g/dL increase	0.43	0.33–0.55	< 0.0001	0.84	0.57–1.27	0.4115
PT	%	per 1% increase	0.97	0.96–0.98	< 0.0001	1.00	0.99–1.02	0.8059
AFP	ng/mL	per 1 ng/mL increase	1.001	1.000–1.002	0.0470	1.00	0.99–1.02	0.3875
Creatinine	mg/dL	per 1 mg/dL increase	1.11	0.61–1.51	0.6549

3.4 Cumulative Incidence of HCC After 5 or 10 Years of NUC Treatment

To examine the risk of incident HCC more than 5 or 10 years after NUC treatment initiation, we conducted a landmark analysis using the 5-year or 10-year mark as the starting point. After 5 years of NUC treatment, among the 737 patients, 593 patients, including 92 cirrhotic patients, continued NUC treatment without developing HCC (Supporting Information S3: Table S9). After 10 years of NUC treatment, among the 737 patients, 421 patients, including 59 cirrhotic patients, continued NUC treatment without developing HCC (Supporting Information S3: Table S10). Landmark analysis revealed that the 5-year and 10-year cumulative incidence rates of HCC starting 5 years after NUC treatment initiation were 8.1% and 13.4%, respectively (Figure 2A). The 5-year cumulative incidence of HCC starting 10 years after NUC treatment initiation was 5.8% (Figure 2B).

4 Discussion

We clearly demonstrated that the risk of HCC occurrence persists for a long time after NUC treatment initiation in the present study. Additionally, we revealed that older age, the presence of cirrhosis, and a high γ-GTP level are risk factors for subsequent HCC occurrence 5 years after the start of NUC treatment, whereas the presence of cirrhosis and diabetes at baseline are risk factors for subsequent incident HCC 10 years after the start of NUC treatment. The mean observation period after the initiation of NUC treatment in the present study exceeded 12 years, making this study valuable.

Our study revealed that in noncirrhotic patients, the risk persists at a comparable level even more than 10 years after starting NUC treatment. The factors related to HBV-induced incident HCC include the accumulation of genetic abnormalities due to chronic inflammation and regeneration, genome instability due to the integration of the HBV genome into the host genome [14, 15], DNA methylation [16], and direct carcinogenic effects of the HBx protein [17] or HBV preS/S proteins [18]. NUC treatment suppresses chronic hepatitis and inhibits incident HCC by suppressing serum HBV DNA but does not eliminate HBV genome integration, DNA methylation, the HBx protein, or the HBV preS/S proteins. The persistence of the risk of incident HCC more than 15 years after the initiation of NUC treatment in noncirrhotic patients may reflect that the risk of incident HCC due to HBV genome integration, DNA methylation, the HBx protein, and the HBV preS/S proteins persists over time. In cirrhotic patients, the cumulative HCC incidence after 5 and 10 years of NUC treatment was lower than that immediately after NUC treatment initiation. The observation of a suppressive effect on incident HCC after long-term NUC administration in cirrhotic patients suggests that chronic inflammation plays a significant role in the development of incident HCC in this population. However, the data in Table 5 suggest that patients who developed HCC within 5 years were more likely to have advanced cirrhosis than patients who developed HCC more than 5 years after treatment initiation. The reduction in HCC incidence after long-term NUC treatment in cirrhotic patients may simply be due to a decline in the number of high-risk patients over time.

Previous studies reported that advanced-stage liver fibrosis at the start of NUC treatment is an important predictor of incident HCC despite antiviral therapy [19-21]. Additionally, it has been reported that an elevated FIB-4 index, which estimates fibrosis, after starting NUC treatment is a risk factor for the development of incident HCC during NUC therapy [22, 23]. Although previous research has showed that NUC treatment can improve fibrosis, even in patients with cirrhosis [24, 25], our study revealed that individuals with cirrhosis at the initiation of NUC therapy remain at a significantly greater risk of developing HCC than those without cirrhosis, even more than 10 years after starting NUC treatment. From the perspective of suppressing incident HCC, initiating NUC treatment in HBV patients before they progress to cirrhosis may be critical.

Huang CF et al. reported that a γ-GTP level > 25 U/L at 6 months after NUC treatment initiation is a risk factor associated with the development of HCC in the following 6 months [26]. Sometani E et al. reported that a high γ-GTP level was a risk factor for HCC in CHB patients treated with NUCs for more than 1 year who had suppressed HBV DNA titers [27]. γ-GTP is a well-known biological marker of excessive ethanol consumption and is associated with metabolic dysfunction-associated steatotic liver disease (MASLD) [28]. In patients with chronic hepatitis C without alcohol consumption or diabetes mellitus, γ-GTP at the time of sustained virological response (SVR) has been reported to be correlated with growth differentiation factor 15 (GDF15), a marker of mitochondrial dysfunction. Additionally, single nucleotide polymorphisms (SNPs) in ZNF827 and GDF15 have been associated with high GGT levels [29]. In the present study, GDF15 and SNPs were not examined, which is a limitation of the present study. Patients with higher alcohol intake had higher γ-GTP levels than did those with lower alcohol intake, and patients with diabetes mellitus had higher γ-GTP levels than did those without diabetes mellitus. The finding that a high γ-GTP level is a risk factor for HCC suggests that the presence of comorbid alcohol-associated liver disease (ALD) or MASLD is a risk factor for HCC.

Diabetes was identified as contributing to the development of HCC more than 10 years after NUC treatment initiation. Because the comorbidity rate of diabetes is high in MASLD patients [30], the coexistence of MASLD may have contributed to the development of HCC more than 10 years after NUC treatment initiation. It has been reported that in CHB patients with diabetes, the hyperglycemic burden is associated with the development of HCC, regardless of the use of anti-HBV therapy or baseline HBV DNA levels [31]. Furthermore, sodium-glucose cotransporter 2 (SGLT2) inhibitors reduce the risk of incident HCC in CHB patients with diabetes [32]. Future studies are needed to determine whether glycemic control reduces the risk of incident HCC in patients with diabetes receiving NUC treatment.

Interestingly, in the present study, the HBs antigen, HBe antigen, hepatitis B virus core-related antigen (HBcrAg), and HBV DNA were not risk factors for incident HCC under NUC treatment initiation. Recently, new HBV markers, including HBV RNA [33], iTACT-HBcrAg [12], and the hepatitis B surface antigen glycan isomer (HBsAgGi)/HBs antigen ratio [34], have been reported to be valuable predictors of HCC incidence during NUC therapy. These novel HBV markers were not examined in the present study. In the future, the predictive ability of these novel HBV markers for incident HCC during NUC treatment is expected to be examined in cohorts with longer observation periods, such as in the present study.

In conclusion, the HCC incidence rate persists for at least 15 years after NUC treatment initiation. Cirrhosis at the start of NUC therapy was strongly associated with a higher risk of HCC, even after 10 years. However, in noncirrhotic patients, the risk remains at a similar level even after a decade of NUC treatment.

Conflicts of Interest

Hayato Hikita and Tomohide Tatsumi are Editorial Board members of Hepatology Research and coauthors of this article. To minimize bias, they were excluded from all editorial decision-making related to the acceptance of this article for publication. Tetsuo Takehara is a former member of the Hepatology Research Editorial Board and a coauthor of this article. Tetsuo Takehara received research grants from Gilead Sciences Inc., GSK plc, and Bristol Myers Squibb. Tetsuo Takehara and Hayato Hikita received lecture fees from Gilead Sciences.