Infection with Hepatitis B and C Viruses and Risk of Lymphoid Malignancies in the European Prospective Investigation into Cancer and Nutrition (EPIC)

Approximately 74,000, 32,000, and 12,000 new cases of non–Hodgkin lymphoma (NHL), multiple myeloma (MM), and Hodgkin lymphoma (HL), respectively, were estimated to have occurred in the European Union in 2008 (1). When combined, these 3 lymphoid tissue malignancies, which arise in the vast majority from B-cell lymphocytes (2), represent the fourth most frequent cancer in the European Union. Upward incidence trends for NHL (3) and MM (4) were observed in most European countries and the United States until the mid-1990s when incidence rates leveled off. The pattern of incidence trends for HL was less clear, with a general tendency to remain stable or to decline in the last decades in many European countries (5) and the United States (6).

Immunodeficiency and infection with certain viruses [Epstein-Barr, T-cell leukemia/lymphoma, Kaposi sarcoma herpes virus, and human immunodeficiency virus (HIV)] and bacteria (Helicobacter pylori) are the best documented causes of lymphoid tissue malignancies, although they account for a relatively small fraction of these diseases and are typically associated with specific histological subtypes or localizations (7).

A moderate, but consistent, association of hepatitis C virus (HCV) infection with NHL risk has gradually emerged over the last 2 decades (8–10) and led the World Health Organization (WHO)/International Agency for Research on Cancer (IARC) Monograph Working Group to consider that HCV does cause NHL, especially B-cell lymphoma (11). Data on the association of HCV with MM and HL risk are scantier, but findings were similar to those for NHL (9). There is also evidence of an association between hepatitis B virus (HBV) infection and NHL risk (12), although there are fewer studies on this topic (11) than on the association with HCV. Notably, an approximately 2-fold increased NHL risk has recently been reported among chronic HBV carriers in a large cohort study from South Korea (13).

With the exception of the South Korean study, most evidence on the relationship between HCV or HBV and NHL, MM, and HL derives from case–control studies (9, 12). Therefore, we took advantage of the European Prospective Investigation into Cancer and Nutrition (EPIC) to further elucidate the issue.

EPIC is a large prospective cohort study with more than 500,000 participants enrolled from 23 centers in Denmark, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom (14). Eligibility in the EPIC study was based essentially on geographic and administrative boundaries, although some special population groups were also included. The Utrecht subcohort in the Netherlands, for instance, was based on women who underwent breast cancer screening, and the French subcohort was based on women who were members of the health insurance system or state-school employees. In addition, all 5 Spanish subcohorts, and the Turin and Ragusa subcohorts in Italy, included high proportions (46%–100%) of blood donors. Participation rates are not available from all participating centers (14). Between 1991 and 2000, standardized lifestyle and personal history questionnaires were collected (median year at enrolment: 1995; 5th–95th percentile: 1993–1998). Blood samples were also collected at enrolment and fractionated into aliquots of serum, plasma, red blood cells, and buffy coats. One half of each aliquot was stored locally at the corresponding EPIC center, and the other half at the IARC (Lyon, France, storage at −196°C in liquid nitrogen), with the exception of Denmark (storage at −150°C in nitrogen vapor) and Sweden (storage in −80°C freezer). Ethical approval for this study has been obtained from all participating centers and the IARC Internal Review Board.

Vital status follow-up in EPIC (99% complete) is collected through record linkage with regional and/or national mortality registries in all countries except Germany and Greece, where data are collected actively. Cancer incidence is determined through record linkage with regional cancer registries (Denmark, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom; complete, for the present report, up to December 2004) or via a combination of methods, including the use of health insurance records, contacts with pathology registries, and active follow-up through participants and their next of kin (France, Germany, and Greece; complete up to December 2006).

The present study does not include data from France, due to incomplete coding for lymphoid tissue malignancies, or from Norway (only 9 NHLs and no MM and HL with blood samples available). In addition, 25% and 75% participants from Italy and Spain, respectively, were excluded because they were blood donors and therefore prescreened for exposure, or seropositivity, to markers of HCV and HBV infection.

Lymphoid tissue malignancies reported in the EPIC study were initially classified according to the second revision of the International Classification of Diseases for Oncology (ICD-O-2). They were subsequently recoded according to the WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues, Third Edition (2). ICD-O-2 codes that could not be unequivocally translated to a WHO diagnosis were categorized as lymphoma not otherwise specified (NOS). B-cell proliferations of uncertain malignant potential (n = 12) were not considered in the present report, leaving 1,275 potentially eligible lymphoid tissue malignancies. After a few additional exclusions (82 prevalent tumors; 36 participants with a history of cancer; and 127 with insufficient serum samples), 1,030 lymphoid tissue malignancies (cases) were available for matching.

Two controls were selected for each case by incidence density sampling from all cohort participants who were alive and cancer-free at the time cases were diagnosed and for whom blood samples were available. They were matched with cases by center, sex, age (±12 months at blood collection), date (±3 months) and time of day at blood collection, and fasting status. No matched controls were available for 7 cases and only 1 control was available for 18 cases, thus leaving 1,023 cases and 2,028 controls. Cases eventually consisted of 628 B-cell NHLs [(including 164 chronic lymphocytic leukemia (CLL)], 36 T-cell NHLs, 75 lymphoma NOS (collectively referred to as NHL, n = 739), 238 MMs, and 46 HLs.

Serologic assays were all conducted in the same laboratory (INSERM) using tests licensed in France after approval of analytic performance by a national regulatory authority. Cutoffs recommended by manufacturers were used to define positivity to each test.

Detection of antibodies against HCV (anti-HCV) was carried out on serum or plasma (for the Swedish centers), using a third-generation enzyme immunoassay (NANBASE C-96 3.0; General Biologicals Corporation; refs. 15, 16). Anti-HCV-seropositive samples were confirmed using a second enzyme immunoassay (HEPANOSTIKA HCV Ultra; Beijing United Biomedica Co. Ltd. ref. 17).

Antibodies against hepatitis B core antigen (anti-HBc, IgM, and IgG), a marker of past HBV exposure, were detected using an enzyme-linked fluorescent immunoassay based on an inhibition principle (VIDAS Anti-HBc Total II; BioMerieux; ref. 18). Testing for hepatitis B surface antigen (HBsAg), a marker of chronic carriage of HBV, was carried out using an enzyme-linked fluorescent immunoassay (VIDAS HBsAg Ultra; BioMerieux; ref. 19). HBsAg-seropositive samples were confirmed by a MicroElisa test (HEPANOSTIKA HBsAg Ultra; BioMerieux).

On account of the limited quantities of plasma available, testing for anti-HBc had to be omitted on plasma samples from Sweden. For the same reason, a different test for HBsAg had to be used on plasma samples from Sweden, that is, a solid-phase enzyme immunoassay [SURASE B-96 (TMB), General Biologicals Corporation].

Conditional logistic regression, stratified by case–control set, was used to estimate odds ratios (OR) and corresponding 95% confidence intervals (CI) for NHL, MM, and HL by anti-HCV, anti-HBc, and HBsAg seropositivity. ORs were also calculated for the combination of the 3 malignancies overall and by time interval (<3, ≥3 years) between blood sample collection and tumor diagnosis. The use of different time intervals (e.g., <2, ≥2; or <5, ≥5 years) yielded similar results (data not shown).

Table 1 shows the mean age of participants at enrolment and hence at blood collection, the proportion of men and the number of NHL, MM, and HL cases, as well as the corresponding number of controls and their hepatitis viral marker status by country. The mean age at enrolment was 57.9 (5th–95th percentile: 43.5–70.8), but it varied by center. On account of the matching procedure, age and sex distribution was the same in cases and controls. Anti-HCV seropositivity ranged from 0% in Germany and Greece to 5.3% in Spain. Anti-HBc seropositivity was least frequent in Denmark (1.9%) and most frequent in Greece (45.9%). Finally, HBsAg seropositivity went from 0% in the Netherlands and the United Kingdom to 2.7% in Greece (Table 1). Only one individual (a control subject from Italy) was seropositive for both anti-HCV and HBsAg. Mean age at cancer diagnosis or reference date (among controls) was 62.7 (5th–95th percentile: 48.8–75.5) but it varied by type of lymphoid malignancy (i.e., 62.9, 63.1, and 58.9 for NHL, MM, and HL, respectively; data not shown).

Table 2 shows the association between seropositivity to hepatitis viral markers and NHL, MM, and HL risk. The ORs associated with anti-HCV seropositivity were 1.30 (95% CI: 0.55–3.07), 0.55 (95% CI: 0.06–5.39), and 4.00 (95% CI: 0.36–44.1) for NHL, MM, and HL, respectively. The corresponding ORs by anti-HBc seropositivity were 1.25 (95% CI: 0.86–1.84), 0.56 (95% CI: 0.25–1.25), and 2.26 (95% CI: 0.49–10.4). Nonsignificant direct associations were found between seropositivity to HBsAg and the risk of NHL (OR = 1.78; 95% CI: 0.78–4.04) and HL (OR = 2.00; 95% CI: 0.13–32.0) whereas a statistically significant association was found with MM (OR = 4.00; 95% CI: 1.00–16.0 (Table 2)). When 164 CCL cases and their 326 matched controls were evaluated separately, no CLL and 2 controls were found to be seropositive to HBsAg. Subtraction of CLL from the NHL group led to an OR for seropositivity to HBsAg of 2.20 (95% CI: 0.93–5.18). Restriction to 690 B-cell lymphomas (i.e., NHL of B lineage and MM) and their corresponding 1,369 matched controls led to an OR of 2.11 (95% CI: 0.96–4.63) for seropositivity to HBsAg (data not shown).

The OR by seropositivity to HBsAg for the combination of the 3 malignancies was 2.21 (95% CI: 1.12–4.33) and did not differ by time interval between blood sample collection and tumor diagnosis. The OR was 2.22 (95% CI: 0.74–6.63) less than 3 years after and 2.20 (95% CI: 0.93–5.18) 3 years or more after blood collection (Table 3). There was no significant heterogeneity in the findings on HBsAg across countries (χ², 6 degrees of freedom = 3.7; P = 0.717). The association between the combination of lymphoid malignancies and seropositivity to HBsAg was also assessed separately in countries with high HBV prevalence (Greece, Italy, and Spain) and in low prevalence countries. Similar ORs (2.00; 95% CI: 0.75–5.33; and 2.41; 95% CI: 0.95–6.13, respectively) were found (data not shown).

The present study is the second largest population-based cohort study on the relationship between HCV or HBV infection and the risk of lymphoid malignancies (13). No clear association was found between anti-HCV seropositivity and the risk of NHL, MM, or HL. Conversely, an association was discovered with HBsAg seropositivity that was statistically significant for MM and for the combination of the 3 malignancies.

The majority of the evidence regarding the role of HCV or HBV infection in the etiology of lymphoid tissue malignancies derives from case–control studies (9, 12). Except for a cohort study of more than 600,000 South Korean workers and their dependents, among whom 9% were HBsAg-seropositive but information on anti-HCV was not available (13), all prospective information on the topic derives from relatively small follow-up studies (20) or cohorts of individuals seropositive to anti-HCV (21, 22), HBsAg (23, 24), or both (25, 26). The pooled relative risk for NHL among anti-HCV-seropositive individuals in prospective studies was 2.03 (95% CI: 1.37–3.02) (9, 25). The range of ORs for NHL according to HBsAg seropositivity in 5 cohort studies from England and Wales (23), the United States (25), Taiwan (24), Australia (26), and South Korea (13) was similar (1.7–3.5) to that reported for anti-HCV seropositivity. With respect to MM, only one cohort study on HCV (OR = 2.42, 95% CI: 0.97–4.99; ref. 22) and one on HBV [hazard ratio (HR) = 0.90; 95% CI: 0.56–1.43; ref. 13] were available. No cohort studies reported data on the relationship between HL and anti-HCV, but the large cohort study from South Korea suggested no association between HBsAg seropositivity and HL (HR = 0.99; 95% CI: 0.48–2.04; ref. 13).

Our present findings on HBsAg seropositivity and the risk of NHL are, therefore, in agreement with the studies available on the topic, although the positive association in our study was not statistically significant. Conversely, an association between HBsAg seropositivity and MM risk has not been reported before, and the one between anti-HCV seropositivity and NHL and MM in the EPIC study was weaker than in most previous work (9), although not incompatible, on account of our broad CI ranges. Very few HL cases were available in the EPIC study and the disease was included for the sake of completeness.

The similarity between seropositivity to HBsAg (0.8%) and anti-HCV (0.9%) in EPIC participants was unexpected. The European Centre for Disease Prevention and Control reported that HCV infection has been the most common type of hepatitis in the European Union over the last decades, but surveillance data are severely incomplete (27). The non-negligible exposure to HBV infection among EPIC participants was confirmed by the 9% seroprevalence of anti-HBc, a marker of HBV exposure that was not, however, associated with the risk of NHL, MM or HL. The low HBsAg:anti-HBc seroprevalence ratio (1:11) suggested that in EPIC participants who had been infected with HBV the infection seldom became an active chronic infection. This finding would be consistent with acquisition of the infection in adulthood, possibly through sexual intercourse, rather than at birth or during childhood (28).

The hint of an association of predominately B-cell lymphoid tissue malignancies with active chronic HBV infection, but not with past exposure, suggests that HBV might promote B-cell lymphomagenesis in a way similar to the mechanism proposed for HCV (i.e., chronic antigen stimulation of B-cells), possibly enhanced by engagement of B-cell surface receptors and costimulatory molecules (9). Both HCV and HBV can cause extrahepatic autoimmune diseases, such as cryoglobulinemia and immunocomplex deposition in polyarteritis nodosa, respectively (24, 29, 30). Both viruses are lymphotropic and can be detected in mononuclear cells (31) and the possibility of a direct oncogenic mechanism has been evaluated (32). Integration in the host cell genome, leading to a wide range of genetic alterations, has been shown for HBV but not HCV (32). Conversely, the reversibility of some virus-associated NHL after antiviral treatment has been reported for HCV but has never been evaluated for HBV-associated lymphoid tissue malignancies (29).

In agreement with the South Korean cohort (13), the association with HBsAg seropositivity in our study was not restricted to lymphoid malignancies that had arisen less than 3 years after blood collection, thus providing some reassurance against the possibility of reverse causation (i.e., reactivation of occult HBV infection due to immunologic alterations that precede tumor diagnosis). So far, reactivation of occult HBV infection has only been documented in anti–HBc-seropositive patients with hematologic malignancies undergoing chemotherapy (33).

The present study has strengths and weaknesses. It included nearly 1,000 lymphoid tissue malignancies, but anti-HCV and HBsAg seroprevalence was less than 1%. Different factors contributed to this low prevalence of hepatitis viruses. First, we had to exclude a substantial proportion of EPIC participants who were blood donors from Italy and Spain (i.e., the 2 countries in Europe where population-based prevalence of HCV infection is highest; refs. 34, 35). A “healthy volunteer” effect is also possible, as at least anti-HCV seropositivity was somewhat lower than that in the control groups of case–control studies on NHL carried out in some EPIC countries, such as Italy (34), Germany (36), Spain (35), and, to a lesser extent, Denmark and Sweden (37). Very high prevalence of HBV markers in Greece was expected, as the country had experience of a severe epidemic of HBV infection at least until the end of the 1970s (38).

In addition, serum samples have to be used very sparingly in long-term multihypothesis cohort studies such as EPIC. We were, therefore, unable to test for HCV RNA, HBV DNA, and HBeAg or the presence of other potentially relevant infections. Lack of information on HIV seropositivity is of special concern as HIV infection is associated with HCV and HBV infection and with increased risk of lymphoid malignancies (7). There are, however, good reasons to believe that the number of HIV-infected individuals among participants in the EPIC study was negligible. The earliest WHO estimates about HIV prevalence in Europe suggested that in 1998 there were fewer than 10 HIV-infected persons per 100,000 population in all countries included in the present study and less than 3 per 100,000 in northern European countries (39). Most important, in the mid-1990s, the majority of HIV infections in Europe involved individuals (e.g., intravenous drug users) who were younger than the predominantly middle-aged participants in the EPIC study.

Finally, the low number of NHL associated with hepatitis viral markers prevented us from conducting separate analyses by histologic subtype. Morphologically, however, lymphomagenesis linked to chronic antigenic stimulation from HCV, and possibly HBV, can involve a variety of lymphoid tissue malignancies (9). A broad spectrum of histologic subtypes was indeed found among the 30 cases seropositive for anti-HCV or HBsAg (e.g., 5 diffuse large cell lymphomas, 4 follicular lymphomas, 2 CLLs, 1 extranodal marginal zone, and lymphoplasmocytic lymphoma in addition to 7 MMs, 7 lymphoma NOS, and 3 HLs).

In conclusion, the present nested case–control study from EPIC neither weakened nor strengthened the evidence of an association between HCV and NHL or other lymphoid tissue malignancies. However, it lent some additional support to the possibility that chronic HBV infection may have a role in lymphomagenesis. The proportion of lymphoid tissue malignancies attributable to HBV should be very low in European countries but can be high among HBsAg-seropositive individuals. In the large South Korean cohort study, for instance, the risk of NHL attributable to HBV among HBsAg-seropositive individuals was 43% (13). Treatment directed at control of HBV infection should, therefore, be evaluated in HBsAg-seropositive patients with lymphoid tissue malignancies.

In addition, the non-negligible prevalence of markers of HBV infection in healthy EPIC participants from countries where HBV vaccination of children or adolescents is not recommended (Denmark, the Netherlands, and the United Kingdom; www.who.int) indirectly supports the WHO call for universal HBV vaccination even in low-endemicity countries (40).

No potential conflicts of interest were disclosed.

The authors thank Bertrand Hemon for technical assistance.

This project received financial support from the International Agency for Research on Cancer. The EPIC study was funded by the European Commission: Public Health and Consumer Protection Directorate 1993–2004; Research Directorate-General 2005; Ligue contre le Cancer (France); Société 3M (France); Mutuelle Générale de l'Education Nationale; Institut National de la Santé et de la Recherche Médicale (INSERM); German Cancer Aid; German Cancer Research Center; German Federal Ministry of Education and Research (BMBF 01 E0 0803, A. Nieters); Danish Cancer Society; ISCIII (RETIC-RD06/0020) of the Spanish Ministry of Health; the participating regional governments and institutions of Spain; the CIBER Epidemiología y Salud Pública (CIBERESP), Spain; Cancer Research UK; Greek Ministry of Health and Social Solidarity; Hellenic Health Foundation and Stavros Niarchos Foundation; Medical Research Council, United Kingdom; the Stroke Association, United Kingdom; British Heart Foundation; Department of Health, United Kingdom; Food Standards Agency, United Kingdom; Hellenic Ministry of Health; The Stavros Niarchos Foundation; Hellenic Health Foundation; Italian Association for Research on Cancer; Italian National Research Council; Dutch Ministry of Public Health, Welfare and Sports (VWS); Netherlands Cancer Registry (NKR); LK Research Funds; Dutch Prevention Funds; Dutch ZON (Zorg Onderzoek Nederland); World Cancer Research Fund (WCRF); Statistics Netherlands (the Netherlands); Swedish Cancer Society; Swedish Scientific Council; Regional Government of Skane, Sweden; and Norwegian Cancer Society.

During the work on the manuscript, M. Lise was the recipient of a fellowship from the Associazione Italiana per la Ricerca sul Cancro, which he carried out at the International Agency for Research on Cancer.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.