Risk of Non-Hodgkin Lymphoma among Patients with Hepatitis B Virus and Hepatitis C Virus in Taiwan: A Nationwide Cohort Study

Simple Summary Non-Hodgkin lymphoma (NHL) is difficult to diagnose and has a high mortality rate. Large-scale database research is necessary to examine and strengthen the correlation between viral hepatitis and NHL. This retrospective cohort study analyzed differences in the risk of developing NHL for patients with hepatitis to elucidate these relationships by using nationwide data from Taiwan’s National Health Insurance Research Database. In this study, the incidence rate of NHL in patients with hepatitis B was 0.22%, and in patients with hepatitis C, the incidence rate of NHL was 0.35%. These comparisons indicate that patients with HBV or HCV have a higher incidence of NHL (OR, 2.37; 95% CI, 1.93–2.91). Abstract Hepatitis B virus (HBV) and hepatitis C virus (HCV) are associated with an increased risk of developing non-Hodgkin lymphoma (NHL); however, adequate data corroborating these associations are lacking. Therefore, a study based on the national database was performed to investigate the correlation between HBV and HCV with NHL in Taiwan. This research was a retrospective cohort study using a nationally representative database established by the Health and Welfare Data Science Center of the Ministry of Health and Welfare, Taiwan. The participants were patients with HBV and HCV, analyzed using the propensity score matching method. The study results indicated that the incidence rate of NHL (0.13%) was significantly higher than that in patients from the general population. After controlling related variables, the hazard ratio (HR) of the incidence of NHL in patients with hepatitis was 2.37 (95% CI, 1.93–2.91). Furthermore, the incidence of NHL in patients with HBV was significantly higher than in patients from the general population (HR, 2.49; 95% CI, 1.94–3.19). The incidence of NHL in patients with HCV was significantly higher than in patients from the general population (HR, 2.36; 95% CI, 1.73–3.22). This study indicated that HBV and HCV significantly increase the risk of NHL.


Introduction
Hepatitis B virus (HBV) and hepatitis C virus (HCV) infections are key public health problems worldwide. It has been reported that HBV and HCV are relatively common in Taiwan [1,2]. Taiwan has been controlling viral hepatitis B infections through a large-scale In this study, secondary data were analyzed, examining information from 2000 to 2016 obtained from the Longitudinal Health Insurance Database (LHID), which was established by Taiwan's Ministry of Health and Welfare (MOHW). The LHID contains data on the medical records and causes of death for 2 million randomly sampled participants in Taiwan's National Health Insurance (NHI) program. Taiwan's NHI program has enrolled up to 99% of Taiwanese residents since 1995. Hence, the LHID represents the utilization of healthcare in Taiwan. The LHID contains anonymous data to protect the privacy of beneficiaries and is maintained by the Health and Welfare Data Science Center. Therefore, the requirement for informed consent was waived for this study, and the study protocol was approved after ethical review by the Institutional Review Board of China Medical University Hospital, Taiwan (No: CMUH107-REC2-004).

Study Participants
In this study, patients newly diagnosed as having HBV or HCV between 2002 and 2013 were selected as the research group, to ensure that each participant had undergone a follow-up period of at least 3 years [14]. Hepatitis was defined as the main diagnosis in 3 or more outpatient clinic visits within one year for HBV ( We selected patients from the general population as a control group for comparison. To reduce outcome bias and avoid selection bias, we used the propensity score matching (PSM) method, which entailed patient and control matching by age and sex at a ratio of 1 (patients with chronic hepatitis) to 5 (patients from the general population) to obtain comparisons for increasing the comparability between the hepatitis cohort and the control population [15,16]. To increase the accuracy of the results, we excluded those patients who had been diagnosed with NHL before chronic hepatitis. We assigned the comparison the same index date as the hepatitis cohort, according to a matching identification. Furthermore, patients who received NHL diagnoses before the index date were also excluded to reduce research bias.
A total of 324,942 participants were included from 2002 to 2013, of whom 54,157 were patients with HBV or HCV, and 270,785 were patients from the general population. The screening process for the selection of research participants is shown in Figure 1.
(PSM) method, which entailed patient and control matching by age and sex at a ratio of 1 (patients with chronic hepatitis) to 5 (patients from the general population) to obtain comparisons for increasing the comparability between the hepatitis cohort and the control population [15,16]. To increase the accuracy of the results, we excluded those patients who had been diagnosed with NHL before chronic hepatitis. We assigned the comparison the same index date as the hepatitis cohort, according to a matching identification. Furthermore, patients who received NHL diagnoses before the index date were also excluded to reduce research bias.
A total of 324,942 participants were included from 2002 to 2013, of whom 54,157 were patients with HBV or HCV, and 270,785 were patients from the general population. The screening process for the selection of research participants is shown in Figure 1.

Study Design
The study design was a retrospective cohort study to examine the risk of NHL in patients with HBV or HCV. The date of diagnosis of chronic hepatitis was defined as the observation start date for participants in the study group, and after matching, the same date was assigned as the observation start date for corresponding members of the control group. All participants were tracked from the observation start date until death, or the

Study Design
The study design was a retrospective cohort study to examine the risk of NHL in patients with HBV or HCV. The date of diagnosis of chronic hepatitis was defined as the observation start date for participants in the study group, and after matching, the same date was assigned as the observation start date for corresponding members of the control group. All participants were tracked from the observation start date until death, or the participant being diagnosed with NHL, or the end date of the study. The definition of NHL in the study was based on ICD-9-CM diagnostic code 202.8 and ICD-10-CM diagnostic codes C85.8 and C85.9. The control variables in this study were the patient's sex, age, and related

Statistical Analysis
All analyses in the study were performed using SAS version 9.4, and statistical significance was defined as p-values < 0.05. Descriptive statistics were used to analyze the numbers and percentages of patient characteristics (sex, age, and health status) and other variables; chi-squared analysis was performed to compare the differences in various variables. The Cox proportional hazards model was used to determine the relationship between the occurrence of NHL for patients with HBV or HCV by controlling the relevant variables.

The Baseline Characteristic Distribution of Study Subjects after Matching
This study examined the occurrence of NHL in patients newly diagnosed with hepatitis between 2002 and 2013. After excluding patients who had previously been diagnosed with NHL, a total of 324,942 participants were included, 54,157 of whom were patients with hepatitis and 270,785 were patients from the general population. Table 1 lists the distribution of the variables of the study participants. Patients were matched by sex and age, and chi-squared tests were performed to determine any differences between the patients with hepatitis and paired patients from the general population. No statistically significant difference (p > 0.05) was observed in the matching variables, including gender and age.  Table 2 presents the bivariate analysis for each variable and the associated occurrence of NHL. Among all patients with hepatitis, the HBV group comprised 37,656 patients (69.53%), the HCV group comprised 14,365 patients (26.52%), and 2136 patients were diagnosed as having both HBV and HCV. A total of 419 patients from this study developed NHL, and the overall incidence rate of NHL was 0.13%. The incidence rate of NHL in patients with hepatitis was 0.25%, which was significantly higher than the incidence rate of NHL in patients from the general population, which was 0.10%. For the variable "age", older patients exhibited a higher incidence of NHL. The incidence rate of NHL in patients ≥ 65 years was 0.28%, which was significantly higher than that in patients aged 20 to 44 (0.07%) and 45 to 54 (0.07%). For patients with comorbidities, the highest incidence of NHL was 1.35%, which was for patients who had received allografts. For patients with diabetes and hypertension, the respective incidence rates of NHL were 0.17% and 0.16%; for patients with rheumatoid arthritis, the incidence rate of NHL was 0.32%; the incidence rate of NHL in patients with lupus erythematosus was 0.33%; the incidence rate of NHL in patients with psoriasis was 0.33%; and the incidence rate of NHL in patients with human immunodeficiency virus was 0.66%. All of the above differences were statistically significant (p < 0.001).   Table 3). After controlling other related variables, compared with general patients, the relative risk of in patients with HBV developing NHL was 2.49 times higher (95% CI, 1.94-3. 19), and for patients with HCV, the relative risk was 2.36 times higher (95% CI, 1.73-3.22). For patients with HBV and HCV, a higher relative risk of NHL was observed, but this was not statistically significant (HR, 1.20; 95% CI, 0.45-3.23).

Discussion
In this study, the incidence rate of NHL in patients with HBV was 0.22%, and in patients with HCV, the incidence rate of NHL was 0.35%. These comparisons indicate that patients with HBV or HCV have a higher incidence of NHL (OR, 2.37; 95% CI, 1.93-2.91). This study was a retrospective cohort follow-up study that used secondary database analysis to screen patients with HBV and HCV through disease diagnosis codes and prescribed medications, and used the PSM method to obtain a control group. In the investigation of the two hepatitis conditions and the occurrence of NHL, we determined that patients with HBV and patients with HCV both had a higher risk of developing NHL and that the higher relative risk between the groups occurred in patients with HBV.
The mechanism of HBV infection in NHL is still unclear. One hypothesis is that HBV can directly infect lymphocytes and be incorporated into the host genome, leading to the overexpression of oncogenes or downregulation of tumor suppressor genes. Additionally, viral hepatitis B replication and viral antigens may also induce the expression and release of hematopoietic tumor growth factors, leading to the proliferation of cloned lymphocytes [17]. HCV may indirectly affect B-cells and regulate oncogenic transformation through intracellular viral proteins. Such indirect effects include chronic antigen stimulation by HCV and the existence of potential antigen-selection-driven processes in the development of NHL in patients with HCV [18].
Further evidence of the indirect effect of HCV on NHL is the upregulation of the tumor necrosis factor family during chronic HCV infection [19]. The HCV genome produces structural (nucleocapsid, E1, E2) and nonstructural proteins that may result in the E2 protein of HCV causing chronic antigen-driven polyclonal B-cell proliferation [20]. Although the existence of HCV lymphocytes is disputed, CD81 on B-cells is a branded HCV internalization receptor, and the costimulatory receptor B7.2 (CD86) mediates HCV to memory B-cells [21]. HCV produces nitric oxide synthase and reactive-oxygen-induced DNA damage [22] and upregulates host B-cell receptor signaling in patients with HCV [23]. However, the underlying carcinogenic mechanisms are still unclear.
Taborelli et al. (2016) conducted a case-control study in Italy and reported that 3.7% of patients with NHL also developed HBV, which was a rate significantly higher than the 1.7% observed in the control group [24]. However, studies have produced inconsistent results, which may be due to differences in research designs, ethnicities, geography, and environments [12,13,25]. To increase the accuracy of the results in the present study, the selected participants with hepatitis had all been prescribed antiviral medications for chronic hepatitis; this confirmed the positive correlation between HBV infection and the development of NHL in the Taiwanese population.
A meta-analysis comprehensively assessed the association between HBV and NHL. The pooled estimates of 58 studies included the risk of NHL in patients with HBV, which significantly increased the overall odds ratio (sOR, 2.50; 95% CI, 2.20-2.83) regardless of study design (case-control study: sOR, 2.47; 95% CI, 2.16-2.82; cohort study: sOR, 2.64; 95% CI, 1.78-3.91) [26]. Previous studies examining the link between HBV infection and NHL have yielded conflicting results. Studies from Europe, Japan, China, and South Korea have indicated HR ranges from 1.74 to 4.87 [11,27,28]. However, other studies have found no significant association between HBV infection and NHL [13,29,30]. Studies in Taiwan have also corroborated the association between hepatitis infection and NHL [31,32]. Kleinstern et al. published a case-control study in 2016 to examine the relationship between HBV and the occurrence of NHL. The study included healthy participants as the control group and reported that the relative risk of developing NHL in patients with HBV was 2.39 times higher (95% CI, 1.13-5.06). Although patients with HBsAg antibodies were also determined to have a low relative risk of developing NHL (OR, 0.68; 95% CI, 0.51-0.92), the risk was not statistically significant for natural antibody responses (OR, 0.76; 95% CI, 0.56-1.04) [33]. This showed a link between persistent HBV infection and B-NHL. Another study reported that acquired immunity from natural infections also increases the risk of B-NHL. Therefore, regardless of whether HBsAg is cleared, the risk of B-NHL is likely to be higher in patients with hepatitis [34].
Studies and meta-analyses have identified an association between HCV infection and the development of B-NHL. The estimated risk of lymphoma development is moderate, with an average odds ratio between two and three. However, this estimate varies widely and depends not only on the type of histology considered, but also on the geographic locations and ethnicities of the populations included in the trials [35].
A nested large-population case-control study in Korea investigated the suspected relationship between hepatitis and NHL. The results reported that the incidence of HBV and HCV in the NHL group (3.3% and 1.3%, respectively) was higher than that in the control group (0.9% and 0.3%, respectively; p < 0.001) [11]. In a previous study in Taiwan, chronic HCV infection was temporally associated with a twofold increase in the risk of lymphoid tumors from 2001 to 2005, especially the risk of NHL. Although 8.2% of participants received interferon-based treatment, the statistical results of the study were not significant due to the small number of participants [14].
The relative risk of developing NHL increases with increasing age. For example, compared with patients aged 20 to 44, the relative risk of developing NHL for patients aged 55 to 64 years old was 3.30 times higher (95% CI, 2.29-4.75), and the relative risk for developing NHL in patients ≥ 65 years old was 6.28 times higher (95% CI, 4.26-9.26). Among older adults, a registry-based case-control study for HCV infection and the risk of cancer indicated a positive association with diffuse large B-cell lymphoma (aOR, 1.57; 95% CI, 1.34-1.84) [36]. One study reported the adjusted ORs of HBV in patients with NHL who were <55 years compared with those who were ≥55 years-2.28 (p = 0.038) and 3.48 (p < 0.001), respectively-and the adjusted ORs for HCV in patients with NHL who were <55 years compared with those who were ≥55 years were 2.58 (p = 0.114) and 3.24 (p = 0.044), respectively [11]. The results of the present study indicate that patients with lupus erythematosus, human immunodeficiency virus, and organ transplant were all at a higher risk of developing NHL. This confirms the correlation between immunodeficiency and NHL. The cause of NHL may be related to viral infection, bacterial infection, or weakened immune function [6]. Early research also demonstrated that several risk factors were associated with lymphoma, including immunosuppression, several autoimmune disorders (rheumatoid arthritis, celiac disease, systemic lupus erythematosus, and Sjögren's syndrome), and certain infectious agents for NHL [37].
It has been indicated that serum HBV DNA and HCV RNA amounts reveal complex virological profiles which may be present in HBV and HCV co-infected patients [38]. In our study, patients co-infected with both HBV and HCV had a lower HR compared with HBV-and HCV-infected patients that may be due to having too few participants with NHL (n = 4). However, 95% CI is not statistically significant. The lower risk of NHL in the co-infected patients may be different depending on the geographical region and the study population. Large-scale prospective studies are warranted to assess the association between patients co-infected with both HBV and HCV and the subsequent development of NHL.
The International Agency for Research on Cancer assessed and confirmed the carcinogenicity of seven viral agents to humans, including HCV. HBV and HCV are indirect carcinogens, producing their effects through chronic inflammation [39]. A previous casecontrol study and meta-analysis demonstrated that chronic HBV infection was positively associated with NHL. However, acquired immunity (HBsAg−, anti-HBs+, anti-HBc+) by natural infection also increased the risk of developing B-NHL (adjusted OR, 2.25; 95% CI, 1.96-2.57) [34]. Vaccination against HBV and HCV may reduce the risk of hepatitis-related liver disease and NHL [11]. In addition, recent studies have indicated that anti-HBV and anti-HCV drugs for the treatment may contribute to the treatment of NHL [40][41][42].
Future experimental studies to further clarify the causal relationships between NHL and HBV and HCV infection may support the clinical necessity of such therapies. Although anti-HBV and anti-HCV agents have an impact on cancer risk, this study used HBV and HCV disease diagnostic codes in conjunction with medication usage to determine to a strict correlation between viral hepatitis medications and the risk of cancer in the participants. However, the main purpose of this study was to examine the risk of cancer in patients with HBV and HCV. Therefore, no further analysis of medications was necessary. We recommend that future studies further analyze and study the effects of various viral hepatitis medications on NHL.
The results of this study confirmed that patients with HBV or HCV have a higher risk of developing NHL, which is consistent with the results of previous studies. Compared with related studies, the present study has several strengths. First, we used a national sampling database for analysis. The MOHW database comprises a stratified random sample of two million individuals with information on sex, age, and location. The database is nationally representative and also avoids bias in the selection of research participants. Previous investigations concerning the hepatitis virus and NHL have mostly been casecontrol studies. This study, however, adopted the design of a generational follow-up study, which is a comparably more rigorous design. Finally, in addition to controlling risk factors such as sex, age, and the related comorbidities of the participants, subjects who had previously been diagnosed as having NHL were excluded before study participants were selected. PSM for the study and control groups further reduced sample selectivity bias and confirmed the representativeness and reliability of the results of this study.
This study also has limitations. The study was limited to the information provided in the LHID; information regarding patient living environments, tobacco and alcohol use, dietary habits, stress levels, and family disease history, laboratory parameters (e.g., HBV DNA, HCV RNA, and genotypes), among others, is unavailable in the database but may affect NHL factors. The LHID was also missing data regarding the active chronic hepatitis, cirrhosis, and NHL histology (such as B-or T-origin, indolent or aggressive, and the World Health Organization [WHO] classification subtypes). This study reduced confounding via adjusting comorbidity disease, but clinical situations (e.g., RA, SLE, and HIV infection) could be associated with lymphoma development. Future studies are warranted to evaluate the effects of hepatitis virus infections on the occurrence of NHL, or NHL histology after excluding related autoimmune diseases. Additionally, future research should be supplemented by clinical empirical medicine and by linking other databases or questionnaires to more rigorously and thoroughly analyze correlations or causality. Hepatitis prevention and treatment is a public health issue that Taiwan has prioritized, and cancer prevention has always been a priority medical issue in Taiwan. The results of this study may provide relevant units and clinical medical personnel guidance on hepatitis prevention and treatment strategies and serve as a reference for the future development of cancer prevention and treatment strategies for patients with hepatitis.

Conclusions
HBV and HCV infections play a significant role in the development of NHL. In patients with HBV and patients with HCV, both receiving antiviral agents, chronic coinfection with HBV and HCV was associated with an increased risk of NHL in a Taiwanese population. In addition, the relative risk of developing NHL increases with increasing age and some comorbidities (such as HPL, CKD, SLE, HIV, and organ transplant).  Informed Consent Statement: The database was anonymous; therefore, the requirement for informed consent was waived.

Data Availability Statement:
The National Health Insurance Database used to support the findings of this study was provided by the Health and Welfare Data Science Center, Ministry of Health and Welfare (HWDC, MOHW) under license and so cannot be made freely available. Requests for access to these data should be made to HWDC (https://dep.mohw.gov.tw/dos/cp-5119-59201-113.html, accessed on 11 November 2021).