HBV Infection Status and the Risk of Cholangiocarcinoma in Asia: A Meta-Analysis

Background. The inconsistent finding was between hepatitis B virus (HBV) infections and cholangiocarcinoma (CCA). This meta-analysis is to explore this relationship in Asia. Methods. A literature search was performed using PubMed, Web of Science, and Cochrane Library to October 30, 2015. Pooled incidence rate and OR with 95% CI were calculated using STATA 11.0. Results. Thirty-nine studies were included. The pooled incidence rate of CCA patients with HBV infection was 31% (95% CI 22%–39%). The pooled OR showed increased risk of CCA incidence with HBV infection (OR = 2.72, 95% CI 1.90–3.88), especially in ICC (OR = 3.184, 95% CI 2.356–4.302), while it showed no risk in ECC (OR = 1.407, 95% CI 0.925–2.141). Also, the pooled OR showed increased risk of ICC and ECC incidence (OR = 6.857, 95% CI 4.421–10.633 and OR = 1.740, 95% CI 1.260–2.404) in patients with HBsAg+/HBcAb+. The pooled OR showed increased risk of ICC incidence (OR = 1.410, 95% CI 1.095–1.816) in patients with HBsAg−/HBcAb+. Conclusion. It is suggested that HBV infection is associated with an increased risk of CCA in Asia. Two HBV infection models (HBsAg+/HBcAb+ and HBsAg−/HBcAb+) increase the risk of CCA, and patients with HBsAg−/HBcAb+ also had a risk of ICC. This trial is registered with PROSPERO CRD42015029264.


Introduction
Cholangiocarcinoma (CCA) is the second commonest primary liver tumor worldwide after hepatocellular carcinoma (HCC) [1]. In the World Health Organization (WHO), CCA is classified as intrahepatic cholangiocarcinoma (ICC) or extrahepatic cholangiocarcinoma (ECC). It accounts for 3% of all gastrointestinal tumors [2]. During the past 3 decades, the overall incidence rate of CCA has appeared to have significantly increased [1]. Hepatobiliary malignancies account for 13% of cancer-related deaths and 10%-20% of these are attributable to CCA worldwide [3]. Currently, CCA has not been extensively investigated, possibly as a result of its presumptive rarity and worrisome prognosis at the time of diagnosis. It has been shown that CCA is characterized by a low survival rate with a median survival of less than 24 months after diagnosis [4].
Defining the risk factors for CCA benefits the search for better ways to prevent the occurrence of this disease. To date, it has been shown that viral infection (mainly hepatitis B and hepatitis C infection), gallstone formation, choledochoenteric anastomosis, oil sands carcinogenicity, chemicals and radiations exposure, biliary system development abnormalities, and so on are the major risk factors for CCA [5]. However, data from low hepatitis B (HBV) epidemic area and HBV infection is not a risk factor for CCA [6]. Thus, we conducted this meta-analysis for evaluating the association between HBV infection and CCA (including ICC and ECC) in Asian countries which are high HBV epidemic areas.
Also, two main treatment options in hepatitis B are available: interferon and nucleos(t)ide analogue (NUCs). There are five NUCs currently approved for HBV treatment: lamivudine (LAM), adefovir (ADV), telbivudine (LdT), entecavir (ETV), and tenofovir disoproxil fumarate (TDF). Treatment with TDF or ETV is more effective due to a high antiviral potency and has a high barrier to resistance reducing the risk of drug resistance and treatment failure [7]. So, HBV DNA could be inhibited during treatment. Even, some patients could achieve serological conversion of HBsAg (HBsAg−/HBcAb+) whereas, for patients with curative resection of HBV-related HCC, positive HBcAb is associated with higher risk of early recurrence and poorer survival [8] and hepatitis B reactivation in HBsAg−/HBcAb+ patients receiving rituximab for lymphoma [9]. According to those, the risk of cancer was still existing in patients with HBsAg−/HBcAb+. Thus, this meta-analysis was also conducted for the association between HBV infection model and CCA risk.

Search Strategy.
We have registered this trial with PROS-PERO (http://www.crd.york.ac.uk/PROSPERO/) (trial number: CRD42015029264). Our study was performed according to the recommendations of the Moose [10] (see Supplementary Table 1 of the Supplementary Material available online at http://dx.doi.org/10.1155/2016/3417976). PubMed was searched combining the terms of ("hepatitis B" OR HBV OR CHB) AND ("bile duct neoplasms" OR cholangiocarcinoma).

Inclusion and Exclusion Criteria.
For estimating the incidence rate of CCA, ICC, or ECC, the studies were included if they meet the criteria as follows: (i) studies only containing the cancer group; (ii) exposure to HBV infection; (iii) providing enough information for calculating the incidence rate; and (iv) the number of cases being not less than 50.
For estimating the OR of CCA, ICC, or ECC, the studies were included if they meet the criteria as follows: (i) exposure to HBV infection; (ii) the outcome being CCA, ICC, or ECC incidence; (iii) providing risk estimates with 95% confidence interval (CI) or available information to calculate them; and (iv) published full-text report in English language.
Abstracts and reviews, letters, case reports, and studies that did not provide sufficient data to calculate the risk estimates were excluded. Two investigators (Hao Zhang and Biqing Zhu) independently selected studies, and any discrepancies were resolved by the third investigator (Wenting Zeng).

Quality
Assessment. The quality of the included studies was assessed independently by two authors (Hao Zhang and Jianxin Liang) using the Newcastle-Ottawa Scale (NOS) [46]. The NOS is for observed studies and consists of 3 parameters of quality: selection, comparability, and exposure/outcome assessment. The NOS assigns a maximum of 4 points for selection, 2 for comparability, and 3 for exposure or outcome. We assigned NOS scores of 1-3, 4-6, and 7-9 for low-, intermediate-, and high-quality studies. Discrepancies were settled by consensus after joint reevaluation of the original studies with the third author (Jianxin Liang).

Data Extraction.
The following information was extracted from each study: first author, publication time, the sample size, country, number of exposures in cases and controls; risk estimate; and 95% CI. The data were collected independently by two investigators (Hao Zhang and Biqing Zhu). When the literature citations were controversial, these investigators discussed them and reached a consensus on inclusion or exclusion.

Data
Analysis. The overincidence rate of CCA, ICC, and ECC with HBV infection was calculated by effect size (ES) and the corresponding 95% confidence interval (95% CI). The risk of HBV infection outcomes was estimated by odds ratio (OR) with the corresponding 95% CI. It was considered statistically significant when < 0.05. In the forest plots, OR > 1 represented a risk effect and OR < 1 represented a protective effect. Statistical heterogeneity of results was appraised using a Chi-Square based test and 2 statistic. Only when the analysis fullfilled both > 0.10 and 2 < 50%, the heterogeneity was considered not significant. The fixedeffects model was used when literature heterogeneity did not exist; otherwise, the random-effects model was employed. Sensitivity analysis was conducted by modification of the inclusion criteria of this meta-analysis. The pooled proportion of vertical transmission of toxoplasmosis was calculated by STATA 11.0 software (Stata Corporation, College Station, TX, USA) and the publication bias was considered significant when value was less than 0.05 in either Begg's test or Egger's test.

HBV Infection and the Risk of ECC.
For pooled analysis of incidence rates of ECC with HBV infection, 2 studies with 483 participants were included [11,20]. These studies were from Japan and China. The overall pooled incidence rates were 18.9% (95% CI −0.7%-38.5%, = 0.059), calculated with the random-effects model ( < 0.001, 2 = 100%, Figure 4   a random-effects model ( < 0.001, 2 = 88.6%) found no statistically significant increased risk of ECC incidence with HBV infection (OR = 1.407, 95% CI 0.925-2.141, = 0.110, Figure 4(b)). No publication bias was observed by Begg's test ( = 1.0) and Egger's test ( = 0.579). The sensitivity analysis showed that the result that HBV infection was not a high risk for patients progressing into ECC was not changed (Figure 4(c)). However, the adjusted OR of ECC with HBV infection found statistically significant increased risk of ECC incidence with HBV infection (OR = 2.6, 95% CI 2.0-3.4), which was only in one study [35].

HBV Infection Status and the Risk of Cancer.
Two HBV infection models, which were HBsAg+/HBcAb+ and HBsAg−/HBcAb+, were included in this meta-analysis. Patients with HBsAg+/HBcAb+ showed a high risk of ICC (OR = 6.857, 95% CI 4.421-10.633) with calculating in a fixeffects model ( = 0.447, 2 = 0%, Figure 5(a)). As only two studies were analyzed, the publication bias and sensitivity analysis cannot be performed. The risk of ICC was increased in patients with HBsAg−/HBcAb+ (OR = 1.410, 95% CI 1.095-1.816) with calculating in a fix-effects model ( = 0.202, 2 = 37.4%, Figure 5(b)). No publication bias was observed by Begg's test ( = 0.96) and Egger's test ( = 0.270). As only three studies were analyzed, the sensitivity analysis cannot be performed.
Patients with HBsAg+/HBcAb+ showed a high risk of ECC (OR = 1.740, 95% CI 1.260-2.404) with calculating in a fix-effects model ( = 0.914, 2 = 0%, Figure 5(c)). No publication bias was observed by Begg's test ( = 0.734) and Egger's test ( = 0.627). The sensitivity analysis showed that the result that HBV infection was a high risk for patients progressing into ECC was not changed ( Figure 5(d)). The risk of ECC was not increased of patients with HBsAg−/HBcAb+ (OR = 1.049, 95% CI 0.881-1.249) with calculating in a fixeffects model ( = 0.382, 2 = 2.1%, Figure 5(e)). No publication bias was observed by Begg's test ( = 0.734) and Egger's test ( = 0.993). As only three studies were analyzed, the sensitivity analysis cannot be performed. The sensitivity analysis showed that the result that HBV infection was not a high risk for patients progressing into ECC was not changed ( Figure 5(f)).

Discussion
HBV is widely epidemic in Asian countries [48]. Previous study has concluded that HBV infection is associated with an increased risk of ICC [49]. This meta-analysis was employed to estimate the CCA incidence rate in HBV infection patients and the CCA risk of HBV infection models in Asia, including ICC and ECC.
In this study, HBV infection leading to CCA was about 31%, and this rate (39%) was also high in China. HBV infection, where pooled OR is 2.72 (95% CI 1.90-3.88) and      adjusted OR is 5.903 (95% CI 3.110-11.207), is a risk factor of CCA. This phenomenon also occurred in the patients with ICC. As all we know, HBV is a strong risk factor for HCC. And hepatocytes and cholangiocytes have the same progenitor cell; therefore, it can be postulated that HBV can induce carcinogenesis in both hepatocytes and cholangiocytes by the same mechanism [50]. Furthermore, studies show that HBV has been suggested to be involved in the pathogenesis of ICC through the inflammatory process [51,52], which further supports the potential role of HBV infection in the pathogenesis of cholangiocarcinoma. Also, recent clinical surveys in China have detected HBV DNA in the tissue species from bile duct cancer.
Interestingly, the risk of ICC was also excluded in patients with the model of HBsAg−/HBcAb+ that termed non-B non-C. The main mechanism may be that these patients were occult HBV infection. It is reported that even a history of HBV infection (positive HBcAb+ with HBsAg−) can be a risk factor for HCC [53]. HBcAb positive status may reflect occult HBV infection and could be associated with an increase in the risk of carcinogenesis [54]. Also, a recent study reported that positive HBcAb is associated with a higher risk of early intrahepatic recurrence and poorer relapse-free survival of HBV-related HCC patients after curative resection [8], which means that HBcAb may play a role in the HCC patients.
Although the incidence rate of ECC was not significantly different in HBV infection patients and HBV infection was not a risk of patients progressing into ECC, the HBV infection status was much different. Patients with HBsAg+/HBcAb+ showed a high risk of ECC, while patients with HBsAg−/HBcAb+ showed no risk of ECC. The mechanism may be associated with HBsAg status. Though few studies have explained the link between HBsAg and ECC, a recent report showed that HBsAg could stimulate proliferation and functional modification of hepatocytes via LEF-1 through the Wnt pathway at the premalignant stage of HCC [55]. Also, high HBsAg levels were associated with late recurrence (after 2 years) after curative resection in HBV-related HCC [56]. So, HBsAg might use a similar action to affect ECC.
In general, our meta-analysis results were similar to those recently published by others about this subject [57]. However, this meta-analysis has several weaknesses that should be considered. First, the included studies were all observational studies that might introduce selection and recall biases. Second, lots of results were based on the random-effect model analysis, which might not reduce the strength of evidence. Third, a small number of studies were used to analyze the HBV infection status and the risk of cancer. Finally, the included studies were mostly from China. Thus, in the future, more studies from other counties need to be performed to explain the link between HBV infection and risk of CCA.

Conclusion
This study suggests that HBV infection is associated with an increased risk of CCA in Asia. The two HBV infection models (HBsAg+/HBcAb+ and HBsAg−/HBcAb+) increase the risk of CCA, while the model of HBsAg−/HBcAb+ increases the risk of ICC. More studies, especially from low epidemic areas, were needed to explore the connection between HBV infection model and CCA risk.           the study and revised the article. Wenting Zeng is guarantor of the article. All authors approved the final version of the manuscript. Hao Zhang and Biqing Zhu contributed equally to this work.