This article is mentioned in:

Introduction

Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide, which is a serious risk to human health (1,2). However, the pathogenesis of HCC has not been fully elucidated. Hepatitis B virus (HBV) chronic infection is considered one of the major risk factors for the development of HCC (3–5).

The HBV genome is an incomplete double-stranded circular structure, containing 4 open reading frames (ORF): S, C, P and X (6,7). ORF region mutations could alter viral replication and virulence force, which lead to a persistent virus infection and severe liver cell damage, and eventually result in the development of HCC (8,9). The precore region encodes the hepatitis B e-antigen (HBeAg), which has been associated with an increased risk of HCC statistically. Certain studies have shown that HBV X gene point mutations can affect the expression of HBeAg and increase the viral replication capacity (8,10,11). In recent years, the association between HBV mutations and the incidence of HCC has focused on the X gene region, the former C gene region and pre-S gene region, and the X gene region is the most important.

Previously, certain studies have identified that the X gene region T1653, V1753, T1762/A1764 and other point mutations may be associated with persistent HBV infection and the development of HCC (6,10–13). However, there remains certain controversy and further research is required. Therefore, the present meta-analysis was performed to investigate the association between HBV X gene point mutations and the development of HCC systematically and comprehensively.

Materials and methods

Search strategy

Two authors (Wang and Zeng) searched PubMed, EMBASE, the Cochrane Library and Chinese National Knowledge Infrastructure for the relevant studies. The key words included: ‘Hepatitis B virus X gene’, ‘mutation’, ‘liver cancer or hepatocellular carcinoma’. The searches were limited to human subjects. Language restriction was not imposed on the search process. In addition, the reference lists of the included studies were checked manually for other potentially eligible studies. This process was repeated until no additional associated studies could be identified.

Inclusion criteria

The inclusion criteria included: i) Study design for the prevalence of case-control; ii) the diagnoses of chronic hepatitis B, liver cirrhosis and HCC were according to the guidelines of the American Association for the Study of Liver Diseases (14); iii) all HBsAg-positive patients were infected with HBV >2 years and the HBV DNA level was in accordance with the test standard; and iv) the reported outcomes of patients was HCC.

Exclusion criteria

The exclusion criteria included: i) Patients with HCV, HDV or human immunodeficiency virus infection; ii) patients with alcoholic liver disease, autoimmune disease or drug-induced liver disease; iii) patients with antiviral treatment; iv) if similar research was reported by the same author, only the recent study or high-quality study was included in this analysis.

Data extraction and quality evaluation

Data extraction and quality evaluation of studies were conducted by 2 independent authors (Wang and Chen). The extracted data included first author, publication year, country, study design, quality score, cases of patients (number of events and total patients), age, gender, alanine aminotransferase and HBV DNA level, genotypes, mutation sites and detection method. The evaluation standard was in accordance with the methods of the study by Liu et al (15). Discussion or a third investigator aided in solving any disagreements.

Statistical analysis

The odds ratios (ORs) with 95% confidence intervals (CIs) of binary end points were analyzed. Heterogeneity was checked using χ2 test, P-values and I2. The random effects model was used when P<0.1; otherwise the fixed effects model was used when P≥0.1. Sensitivity analysis was conducted by eliminating one study in turn in the analysis. Potential publication bias was evaluated by visual inspection of the Begg funnel plots in which the log ORs were plotted against their standard errors. All the data were calculated by RevMan 5.0 software (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration).

Results

Study characteristics

According to the inclusion and exclusion criteria of the literature, 11 studies involving 2,502 patients were included in this meta-analysis, of whom 2,801 had HCC. These studies included 1 study in Chinese and 10 in English (6,11,13,16–23). A detailed flow chart explaining the inclusion of studies is shown in Fig. 1. The extracted information from the studies included: i) The publication year of the studies was between 1999 and 2013; ii) the study design were case-control studies; iii) the countries of the studies included mainland China, Hong Kong, South Korea and Thailand. A summary of the 11 included studies is shown in Table I. The details of the quality criteria based on factors are listed in Table II.

Figure 1. Flow diagram of the relevant studies included in the meta-analysis. HCV, hepatitis C virus.

Table I. Characteristics and clinical data of the studies included in the meta-analysis.

Table I.

Characteristics and clinical data of the studies included in the meta-analysis.

Authors (year)	Country	Design	Quality score	E/C	Age, years	Gender (m/f)	ALT level (U/l)	HBV DNA (log copy/ml)	HBeAg (+) patients (%)	Genotype	Mutation sites	Detection method	Refs.
Takahashi et al (1999)	Japan	PCC	6–10	58/271	NA	NA	NA	NA	108 (37.6)	C	T1653, V1753, T1762, A1764, T1762/A1764	Sequencing	(16)
Tanaka et al (2006)	Japan/HK	PCC	≥10	148/180	50.2±10.7	261/69	46 (8–773)	5.4±1.4	109 (33.2)	C1, C2	T1653, V1753, T1762/A1764, V1765, A1896, A1899	Sequencing	(17)
Shinkai et al (2007)	Japan	PCC	≥10	80/80	55±8	135/25	85±133	5.7±1.4	60 (37.5)	C2	T1653, C1479, T1485, H1499, A1613, V1753, T1762/A1764, A1896	Sequencing	(13)
Wang et al (2007)	China	PCC	≤6	47/164	49.8±11.6	176/35	63.9±41.6	NA	112 (53.1)	Ba, C1, C2	T1653, V1753, T1762/A1764, T1856, T1858, A1896, A1898, A1899	Sequencing	(18)
Kim et al (2008)	Korea	PCC	6–10	60/124	45.9±17.3	134/50	90.8±132.1	NA	115 (62.5)	C	T1653, V1753, T1762/A1764	Sequencing	(19)
Kim et al (2009)	Korea	PCC	≥10	135/135	44.3±7.9	224/46	NA	3.45±3.8	76 (28.1)	C2	T1653, A1689, V1753, T1762/A1764, T1846, A1850, C1858, A1896	Sequencing	(6)
Choi et al (2009)	Korea	PCC	≤6	42/46	57.3±9.3	57/31	89.8±110.5	6.2±1.5	39 (44.3)	C2	M1385, A1485, B1499, B1574, A1613, T1631, T1653, V1753, T1762/A1764	Sequencing	(20)
Tangkijvanich et al (2010)	Thailand	PCC	≥10	60/60	55.7±9.8	104/16	161.1±116.9	5.9±1.4	32 (30)	B, C	A1613, T1653, V1753, T1762/A1764, T1766/A1768, C1858C	Sequencing	(21)
Shi et al (2013)	China	PCC	≤6	43/55	58.0±9.58	63/32	NA	4.67±0.91	46 (47.9)	NA	A1440, C1467, A1479, T1485, T1653, V1753, T1762/A1764	Sequencing	(11)
Li et al (2013)	China	PCC	6–10	102/105	NA	NA	NA	NA	NA	C2	T1653, V1753, T1762, A1764, T1766, A1768	Sequencing	(22)
Lyu et al (2013)	Korea	PCC	6–10	318/234	55 (30–74)	452/100	37 (9–774)	4.09±2.39	271 (49.1)	C2	T1653, V1753, T1762/A1764, A1896	Sequencing	(23)

[i] ALT, alamine aminotransferase; E/C, no. of experiments/no. of controls; HBeAg, hepatitis B e-antigen; HBV, hepatitis B virus; m/f, male/female; NA, not available; PCC, prevalence case-control.

Table II. Details of the quality criteria for studies included in the meta-analysis.

Table II.

Details of the quality criteria for studies included in the meta-analysis.

	Score
Quality parameter	0	1	2
Study design	Cohort or nested case-control	Incidence case-control	Prevalence case-control
Sample size of cases	<50	50–100	≥100
Source of samples	None	1 hospital	≥2 hospitals
Mutation detection method	None	Others	DNA sequencing
Matching of cases and controls
(1)	None	Age or gender	Age and gender
(2)	None	HBeAg or genotype	HBeAg and genotype

[i] HBeAg, hepatitis B e-antigen.

Meta-analysis

All 11 studies involving 2,502 patients reported that the relevant outcome of the T1653, V1753 and T1762/A1764 point mutations were associated with the risk of HCC. The results of the meta-analysis showed that T1653 (OR, 3.11; 95% CI, 2.22–4.36) (Fig. 2), V1753 (OR, 2.55; 95% CI, 1.66–3.92) (Fig. 3) and T1762/A1764 (OR, 4.49; 95% CI, 2.86–7.07) (Fig. 4) increase the risk of HCC.

Figure 2. Forest plot for the odds ratios of T1653 for the risk of HCC. HCC, hepatocellular carcinoma; CI, confidence interval; M-H, Mantel-Haensze.

Figure 3. Forest plot for the odds ratios of V1753 for the risk of HCC. HCC, hepatocellular carcinoma; CI, confidence interval; M-H, Mantel-Haensze.

Figure 4. Forest plot for the odds ratios of T1762/A1764 for the risk of HCC. HCC, hepatocellular carcinoma; CI, confidence interval; M-H, Mantel-Haensze.

Subgroup analysis

The summary ORs for HBV X gene mutations were explored by HBeAg status, country and quality score. The summary OR of the T1762/A1764 double mutations in HBeAg (+) group were lower compared to the HBeAg (−) group, whereas the T1653 and V1753 mutations were higher. There is no significant difference in the results between subgroup and overall analyses except for the V1753 point mutation for the risk of HCC in patients from Korea. The result suggested that there was no statistically significant associations between the V1753 mutation and the risk of HCC in Korean patients (OR, 1.25; 95% CI, 0.69–2.26) (Table III). The summary ORs for T1653, V1753 and T1762/A1764 increased with decreasing study quality score (Table III).

Table III. Subgroup analyses based on the main characteristics of the included studies.

Table III.

Subgroup analyses based on the main characteristics of the included studies.

	T1653			V1753			T1762/A1764
Characteristics	T/P	OR (95% CI)	Model	T/P	OR (95% CI)	Model	T/P	OR (95% CI)	Model
HBeAg
+	5/392	3.25 (1.87–5.66)	Fixed	5/392	2.50 (1.52–4.12)	Fixed	6/536	3.15 (2.03–4.89)	Fixed
−	5/741	2.56 (1.78–3.68)	Fixed	5/741	1.84 (1.32–2.58)	Fixed	6/911	5.50 (2.64–11.46)	Fixed
Country
China	3/513	3.23 (1.80–5.79)	Fixed	3/513	5.17 (3.28–8.13)	Fixed	3/513	5.99 (2.17–16.53)	Random
Japan	3/775	3.78 (1.52–9.41)	Random	3/775	2.93 (2.11–4.06)	Fixed	3/775	4.27 (2.68–6.80)	Fixed
Korea	4/1,094	2.50 (1.77–3.54)	Fixed	4/1,094	1.25 (0.69–2.26)	Random	4/1,094	3.58 (1.46–8.79)	Random
Thailand	1/120	2.75 (1.04–7.29)	NA	1/120	2.51 (1.14–5.52)	NA	1/120	6.19 (2.42–15.83)	NA
Quality score
>6	8/2,108	3.05 (2.10–4.42)	Random	8/2,108	2.14 (1.39–3.28)	Random	8/2,108	3.66 (2.30–5.81)	Random
≤6	3/394	3.11 (1.74–5.59)	Fixed	3/394	5.18 (2.98–9.02)	Fixed	3/394	9.60 (5.14–17.93)	Fixed

[i] HBeAg, hepatitis B e-antigen; NA, not available; OR, odds ratio; T/P, no. of trials/no. of patients.

Sensitivity analysis and publication bias

The effect of a single study on the overall pooled analyses was investigated by sensitivity analysis to evaluate the heterogeneity of each study. There were no significant influences observed when one study was removed each turn. The funnel plot showed an asymmetrical distribution of these studies, indicating that publication bias existed in these results regarding an association between T1653 (Fig. 5A), V1753 (Fig. 5B) and T1762/A1764 (Fig. 5C) point mutations and the risk of HCC.

Figure 5. Funnel plot regarding the publication bias of (A) T1653, (B) V1753 and (C) T1762/A1764 for the risk of hepatocellular carcinoma. OR, odds ratio; SE, standard error.

Discussion

Meta-analysis is regarded as a qualitative and quantitative tool to solve those problems that remain controversial in clinical settings. The results of the meta-analysis were the highest level of evidence. The debate on the association between HBV gene mutations and development of HCC is ongoing. Recently, a series of studies on this subject have been published. Therefore, the present meta-analysis was performed and the results showed that T1653, V1753 and T1762/A1764 mutations could significantly increase the risk of HCC, particularly the T1762/A1764 double mutations.

HBV chronic infection is the most important risk factor for the development of HCC (24). Lin et al (25) considered that the transactivation function of the carboxyl terminus of the HBV X protein would regulate HBV DNA replication and transcription of liver cell proliferation and differentiation. As well as T1653, V1753 and T1762/A1764 mutations are all located in the carboxyl terminus region of the HBV X protein. Kim et al (19) suggested that T1653, V1753 and T1762/A1764 mutations would change the sequence of HBV X protein amino acids, further leading to the activation of proto-oncogenes and inactivation of the tumor suppressor gene, eventually causing the occurrence of HBV-related HCC. However, the mechanism of how the point mutations of HBV X gene region affect the biological function of HBV X protein remains to be elucidated in further studies.

Liu et al (15) considered that the HBeAg status would affect the point mutation type of the HBV X gene and development of HCC. Previous studies suggested that the expression of HBeAg is often significantly correlated with immune evasion and acted as an indicator of active viral replication (26,27). In the subgroup analysis, data on the HBeAg status was extracted, which was a potential confounder, and identified that T1762/A1764 double mutations could decrease HBeAg expression. However, a large number of studies showed that T1762/A1764 double mutations could enhance the virus replication. The contradiction indicated that the decrease of HBeAg expression did not equate to an improved development of HCC. Persistent chronic HBV infection may be due to HBV immune escape, which further aggravates the condition of patients and eventually results in HCC (28,29). Li et al (22) reported that T1762/A1764 double mutations could not predict the development of HCC. However, the present study identified that double mutations were more closely associated with the risk of HCC compared to T1653 or V1753 alone. Therefore, the combined mutations could lead to a higher incidence of liver cancer and improve the predictability of HCC.

The present findings showed that the summary ORs for T1653, V1753 and T1762/A1764 were higher in the low-quality compared to the high-quality studies. Potential confounders may have an important role in evaluating HBV mutations and the risk of HCC in low-quality studies. Yin et al (30) suggested that the average age of the patients with chronic hepatitis B was 10 years younger than that of the patients with HCC. Additionally, Yang et al (31) suggested that HBV mutations accumulated with increasing age. Therefore, the association between the HBV X gene point mutations and risk of HCC was more likely to be overestimated in the confounder-unmatched, low-quality studies.

The main characteristics embodied in the present study were: i) The association between HBV X gene mutations and development of HCC among various studies examined systematically and comprehensively, in order to have an improved understanding for the effect of HBV X protein on development of HCC; ii) a series of subgroup analyses were conducted to explore the effect of potential confounding factors on the development of HCC; and iii) the results suggested that these point mutations could be used as molecular markers of the risk of HCC. The limitations of the study were: i) The age, gender, genotype and other confounding factors could not be matched fully and the existence of various offsets requires further information and data to be confirmed; ii) only 3 HBV X gene mutations were analyzed in the meta-analysis, and there may be other gene mutations that affect the HBV X protein biological function as well as V1674, T1766 and A1768 mutations; and iii) the included studies were all observational case-control studies, as experimental studies could not be conducted in humans.

In the future, the mechanism of the HBV X gene region point mutations should focus on the biological function of the HBV X protein and the association with the development of HCC. In order to improve the prediction for HCC risk and reduce or even avoid the development of HCC, quicker and easier methods should be developed for the detection of HBV gene mutations.

Acknowledgements

The authors would like to thank Dr Longkun Li for providing methods of data analysis.

References