Non-Hodgkin's lymphoma and hepatitis C virus: A case-control study from northern and southern Italy

Authors


Abstract

HCV has been associated with NHL, but the evidence from case series and case-control studies is not totally consistent. Between 1999 and 2002, we conducted a hospital case-control study on the association between HCV, HBV and NHL in 2 areas of Italy where HCV infection is relatively frequent. Cases (n = 225, median age 59 years) were consecutive patients with a new diagnosis of NHL admitted to local specialized and general hospitals. Controls (n = 504, median age 63 years) were patients with a wide spectrum of acute conditions admitted to the same hospitals as cases. HCV prevalence was 19.6% among NHL cases and 8.9% among controls (adjusted OR = 2.6, 95% CI 1.6–4.3). The ORs for HCV were similar for low-grade and intermediate-/high-grade B-cell NHL (3.2 and 2.4, respectively) as well as for nodal and extranodal NHL (2.7 and 2.6, respectively). Positivity for HBsAg was found in 3.8% of cases and 0.9% of controls (OR = 4.1, 95% CI 1.2–14.4). An elevated OR was also found for history of hepatitis C (OR = 4.7, 95% CI 2.3–9.5). History of blood transfusion before 1990 was associated with HCV positivity among controls but not with NHL risk. In conclusion, HCV infection was associated with an increase in NHL risk, and the fraction of NHL cases attributable to HCV was 12.4% (range 6.3–18.5%). © 2004 Wiley-Liss, Inc.

HCV is an RNA virus belonging to the family of flaviviruses. The prevalence of IgG antibodies against HCV varies widely between and within countries worldwide, ranging from approximately 1% in most Western countries1, 2, 3 to >10% in some parts of Egypt,4 Italy,5, 6 South Korea7 and Japan.8

HCV is hepatotropic and causes hepatitis, liver cirrhosis and hepatocellular carcinoma.9 Sometimes, it also induces autoimmune manifestations, most notably EMC. EMC is characterized by cutaneous vasculitis, nephritis, peripheral neuropathy and clonal B-cell lymphoproliferations that can evolve into NHL.10, 11, 12

However, elevated proportions of HCV were reported in NHL patients not affected by EMC. A systematic review of 66 studies of NHL, including over 6,000 NHL patients from Europe, Asia and the Americas, showed high prevalence of HCV positivity, especially in Italy (19.7% of 2,734 cases) and Japan (11.3% of 771 cases).13 Studies where some comparison between HCV prevalence in NHL patients and control subjects was possible have generally suggested relative risks between 2 and 4.14, 15, 16, 17, 18, 19 Negative findings20, 21 or weak associations22 in some areas of North America and northern Europe may be due to a very low prevalence of HCV infection in the study populations, particularly in the past. Furthermore, a few studies have also suggested an association between NHL and HBV infection.23

Elucidation of the role of hepatitis viruses in lymphomagenesis is of great importance since NHL is one of the few cancers still increasing in most populations worldwide.24 We conducted a case-control study of NHL in 2 Italian areas, one in the north and one in the south, which show a substantial difference in HCV prevalence (3.2%6 and 5.1%,25 respectively).

Abbreviations

CI, confidence interval; CLL, chronic lymphocytic leukemia; DLBCL, diffuse large B cell lymphoma; EMC, essential mixed cryglobulinemia; HBsAg, hepatitis B superficial antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; LIA, line immunoassay; LiPA, line probe assay; MALT, mucosa-associated lymphoid tissue; MEIA, microparticle enzyme immunoassay; NHL, non-Hodgkin's lymphoma; OR, odds ratio; SLL, small lymphocytic lymphoma.

MATERIAL AND METHODS

Between January 1999 and July 2002, we conducted a case-control study on the association between HCV, HBV and hemolymphopoietic tumors, including NHL, as well as hepatocellular carcinoma in the province of Pordenone, northeast Italy, and the town of Naples, in the south.

Cases

Cases include patients aged between 18 and 84 years with incident histologically confirmed NHL. They had not received any cytotoxic treatment and were admitted as inpatients to the National Cancer Institute, Aviano; the “Santa Maria degli Angeli” General Hospital, Pordenone; the “Pascale” National Cancer Institute, Naples; and 4 general hospitals, Naples. These institutions attract the majority of NHL cases in the study areas. Histologic specimens were classified according to the International Classification of Diseases for Oncology (second edition),26 which was updated to include categories in the Revised European–American Lymphoma (REAL)/World Heath Organization (WHO) classification.27, 28 One pathologist in each of the study areas reviewed histologic diagnoses. NHL cases that were primarily extranodal at onset were distinguished from nodal NHL cases. A total of 240 NHL cases were identified. Two refused to participate in the study, and 7 had to be excluded for one of the following reasons: (i) NHL diagnosis 2 years before interview (1 patient), (ii) previous diagnosis of cancer other than skin cancer (4 patients) or (iii) concurrent diagnosis of hepatocellular carcinoma (2 patients). Six NHL cases were interviewed, but blood samples could not be obtained, thus leaving 225 cases (median age 59 years) for whom questionnaire and blood samples were available.

Controls

The comparison group included patients aged between 18 and 84 years admitted as inpatients for a wide spectrum of acute conditions to the same hospitals where NHL cases had been interviewed. Specifically excluded from the control group were patients whose hospital admission was caused by malignant diseases, conditions related to alcohol and tobacco consumption (e.g., respiratory diseases, peptic ulcer, etc.) or hepatitis viruses (e.g., hepatitis, cirrhosis, esophageal varices, etc.) as well as any chronic diseases that might have resulted in substantial lifestyle modifications (e.g., diabetes, cardiocerebrovascular diseases, etc.). Hematologic, allergic and autoimmune diseases were also excluded. Comorbidity for the diseases listed above was not, however, an exclusion criterion. A total of 554 controls were contacted, and 550 participated. Blood samples were available for 504 controls (median age 63 years). Of these, 27% were admitted to the hospital for trauma, 23% for nontraumatic orthopedic diseases, 22% for acute surgical conditions, 14% for eye diseases and 14% for a variety of other illnesses. Cases and controls were frequency-matched by center (Aviano/Pordenone and Naples), gender and age (in 5-year bands). Controls were slightly older than cases as age matching was conducted according to the age distribution of the whole group of cases in the entire study (which included lymphohematopoietic neoplasms other than NHL and hepatocellular carcinoma).

All NHL cases had an HIV test as part of their routine management and were HIV-negative. To the best of our knowledge, no control subjects had a history of HIV infection or AIDS, as expected in our relatively old population.

All participants signed an informed consent, according to the recommendations of the Ethical Committee of the Aviano National Cancer Institute, which approved the study.

Questionnaire

Five trained interviewers administered a structured questionnaire to both cases and controls during their hospital stay. It contained information on sociodemographic indicators, personal habits (including smoking and alcohol drinking) and dietary habits. Information was also sought with regard to behaviors and exposures that entailed risk of HCV transmission and diseases that affected the liver or the immune system.

HCV and HBV testing

Each case and control provided a 15 ml sample of blood (7.5 ml in vacutainer tubes with anticoagulant and 7.5 ml in vacutainer tubes without anticoagulant) the day the interview took place (generally the first day of hospital stay). Blood samples were kept refrigerated and processed within 12 hr from collection in either the General Hospital of Pordenone or the “Pascale” Cancer Institute of Naples. They were centrifuged at 1,500g for 10 min, extracted and distributed into different cryotubes of serum, buffy coat and red blood cells. All samples were stored at −80°C. Each cryotube was labeled and coded with the subject's full identification number. Periodically, samples were sent to the Laboratory of Microbiology and Immunology, General Hospital of Pordenone, where they were stored at −80°C and tested in a blinded fashion.

HCV

Sera were screened for antibodies against HCV using a third-generation MEIA (AxSYM HCV, version 3.0; Abbott, Wiesbaden, Germany). The assay detects antibodies to putative structural (core region) and nonstructural proteins of the HCV genome. Its sensitivity is estimated to be 99% in patients with chronic liver disease, with a specificity of 97% in panels of sera.29 Positive samples were tested for HCV antibodies using a third-generation LIA (Innogenetics, Ghent, Belgium) and for serum HCV RNA using Amplicor version 2.0 (Roche, Pleasanton, CA). The limit of detection for the Roche Amplicor 2.0 test for HCV RNA was 50 IU/ml. HCV RNA was also tested for in all NHL cases who were negative for HCV antibodies, to exclude false-negative findings due to NHL-related immunodeficiency. Only one NHL case who was HCV antibody-negative was positive for HCV RNA. Finally, HCV genotyping of HCV RNA-positive samples was performed using a second-generation LiPA (Innogenetics). Samples were considered HCV+ if HCV antibodies or HCV RNA were detected.

HBV

Testing for anti-HBsAg was performed using MEIA (AxSYM AUSAB, Abbott). Samples that gave borderline results were retested. HBsAg was tested using MEIA (AxSYM HBsAg, version 2; Abbott) and confirmed using a neutralization test (AxSYM, Abbott). Samples were deemed HBsAg+ if positive at the neutralization test. Fifteen NHL cases and 41 controls, reporting vaccination against HBV, were not included in the analyses of HBV markers.

Adjusted ORs and corresponding 95% CIs were calculated by means of unconditional multiple logistic regression, including age (in 5-year groups plus a term for age as a continuous variable), gender, center, years of education and place of birth.30 The method described by Bruzzi et al.31 was used to calculate the population-attributable risk.

RESULTS

Table I shows the distribution of NHL cases and controls by gender, age group, study center and selected variables. Place of birth in the south vs. north of Italy was associated with an OR of 2.1 (95% CI 1.2–3.6), whereas education level and occupation were unrelated to NHL risk. In addition, tobacco smoking and alcohol drinking were not significantly associated with NHL risk (data not shown).

Table I. OR and Corresponding 95% CI1 for NHL by Selected Sociodemographic Factors (225 Cases and 504 Controls): Italy, 1999–2002
 Cases number (%)Controls number (%)OR(95% CI)
  • 1

    Estimated from unconditional logistic regression adjusted for gender, age, center, education and place of birth when appropriate.

  • 2

    The sum does not add up to the total because of some missing values.

  • 3

    Reference category.

Gender    
 Male120 (53.3)341 (67.7)  
 Female105 (46.7)163 (32.3)  
Age (years)    
 <4547 (20.8)104 (20.6)  
 45–64107 (47.6)177 (35.1)  
 ≥6571 (31.6)223 (44.3)  
Study center    
 Aviano/Pordenone127 (56.4)280 (55.6)  
 Naples98 (43.6)224 (44.4)  
Education (years)    
 <797 (43.1)251 (49.8)13 
 7–1169 (30.7)127 (25.2)1.37(0.90–2.09)
 ≥1259 (26.2)126 (25.0)1.17(0.75–1.82)
 χmath image trend   0.61, p = 0.43
Place of birth    
 North98 (43.6)248 (49.2)13 
 South127 (56.4)256 (50.8)2.07(1.21–3.55)
Occupation2    
 White collar75 (34.6)177 (35.8)13 
 Blue collar87 (40.1)226 (45.7)1.02(0.66–1.59)
 Farmer14 (6.4)27 (5.5)1.38(0.63–3.02)
 Housewife41 (18.9)64 (13.0)1.09(0.58–2.05)

HCV prevalence was 19.6% among NHL cases and 8.9% among controls (OR = 2.6, 95% CI 1.6–4.3) (Table II). The OR for HCV was similar for low-grade B-cell NHL (OR = 3.2) and intermediate- and high-grade B-cell NHL (OR = 2.4). Prevalence >30% was found among cases of SLL/CLL, lymphoplasmacytic, marginal zone and MALT NHL, though there were small numbers for each of these specific histologic subtypes. None of the follicular and 12.5% of the T-cell NHL cases was HCV+. ORs for HCV positivity were also similar in nodal (OR = 2.7) and extranodal (OR = 2.6) NHL.

Table II. OR and Corresponding 95% CI1 for NHL by HCV Positivity (225 Cases and 504 Controls): Italy, 1999–2002
CategoryHCV+ number (%)HCV number (%)OR (95% CI)
  • 1

    Estimated from unconditional logistic regression adjusted for gender, age, center, education and place of birth.

  • 2

    Reference category.

  • 3

    Includes 3 lymphomas of unknown primary site.

  • 4

    Includes one splenic lymphoma with villous lymphocytes who was HCV.

Controls45 (8.9)459 (91.1)12
NHL Cases44 (19.6)181 (80.4)2.64 (1.62–4.30)
Cases, by grade and histologic subtype   
 Low-grade B-cell16 (20.5)62 (79.5)3.21 (1.63–6.33)
  SLL/CLL6 (33.3)12 (66.7) 
  Lymphoplasmacytic3 (30.0)7 (70.0) 
  Follicular0 (0.0)36 (100.0) 
  Marginal zone2 (50.0)2 (50.0) 
  MALT5 (50.0)5 (50.0) 
 Intermediate- and high-grade B-cell24 (19.3)100 (80.7)2.39 (1.32–4.31)
  Mantle cell0 (0.0)2 (100.0) 
  DLBCL19 (20.4)74 (79.6) 
  DLBCL, immunoblastic3 (15.8)16 (84.2) 
  Burkitt2 (20.0)8 (80.0) 
 T-cell2 (12.5)14 (87.5)1.66 (0.34–8.24)
 Unknown2 (28.6)5 (71.4)3.63 (0.63–21.03)
Cases, by primary site   
 Nodal330 (19.4)125 (80.6)2.65 (1.53–4.59)
 Extranodal414 (20.0)56 (80.0)2.61 (1.29–5.28)

Thirty-six NHL cases (81.8% of HCV+ cases) and 30 controls (66.7% of HCV+ controls) were positive for HCV RNA (Table III). The OR of NHL associated with the presence of HCV RNA was, therefore, 3.1 (95% CI 1.8–5.4). The association with the 2a/2c HCV genotype (OR = 5.2) was not significantly stronger than that with the 1b HCV genotype (OR = 2.3) (Table III).

Table III. OR and Corresponding 95% CI1 of NHL by HCV RNA Positivity and Viral Genotype (225 Cases and 504 Controls): Italy, 1999–2002
HCV RNACases number (%)Controls number (%)OR(95% CI)
  • 1

    Estimated from unconditional logistic regression adjusted for gender, age, center, education and place of birth.

  • 2

    Reference category.

Negative189 (84.0)474 (94.1)12 
Positive36 (16.0)30 (5.9)3.14(1.83–5.38)
Genotype    
 1a/1b1 (0.2) 
 1b18 (8.0)20 (4.0)2.33(1.16–4.68)
 2a/2c16 (7.2)8 (1.5)5.19(2.12–12.69)
 3a1 (0.4)
 Unknown1 (0.4)1 (0.2) 

Markers of HBV infection and self-reported history of different types of hepatitis are shown in Table IV. Positivity for anti-HBsAg was not associated with NHL risk (OR = 0.8), though the few carriers of HBsAg showed an OR of 4.1 (95% CI 1.2–14.4). Only one HBsAg+ NHL case was coinfected with HCV. A significantly elevated OR was found for history of hepatitis C (OR = 4.7, 95% CI 2.3–9.5) but not for history of hepatitis A or B. No risk trend emerged according to age at diagnosis of hepatitis C (data not shown in Tables). Three additional cases and one control reported a history of not otherwise specified hepatitis. Liver disease was found concomitantly with NHL diagnosis in 13 cases, including 7 with chronic hepatitis, 3 with liver cirrhosis and 3 with not otherwise specified liver diseases. When these 13 NHL cases (10 of whom were HCV+) were excluded, the association between NHL and HCV positivity was slightly weakened but still statistically significant (OR = 2.1, 95% CI 1.2–3.4).

Table IV. OR and Corresponding 95% CI1 for NHL by Markers of HBV and History of Hepatitis (225 Cases and 504 Controls): Italy, 1999–2002
 Cases number (%)Controls number (%)OR(95% CI)
  • 1

    Estimated from unconditional logistic regression adjusted for gender, age, center, education and place of birth.

  • 2

    The sum does not add up to the total because of some missing values.

  • 3

    Excluded were 15 NHL cases and 41 controls who reported vaccination against hepatitis B.

  • 4

    Reference category.

HBV markers23    
 Anti-HbsAg157 (74.8)341 (73.6)14 
 Anti-HbsAg+45 (21.4)118 (25.5)0.84(0.55–1.27)
 HBsAg+8 (3.8)4 (0.9)4.07(1.15–14.42)
History of hepatitis A    
 No214 (95.1)490 (97.2)14 
 Yes11 (4.9)14 (2.8)2.07(0.89–4.82)
History of hepatitis B    
 No213 (94.7)487 (96.6)14 
 Yes12 (5.3)17 (3.4)1.46(0.67–3.19)
History of hepatitis C    
 No200 (88.9)491 (97.4)14 
 Yes25 (11.1)13 (2.6)4.65(2.27–9.51)

Selected factors possibly associated with HCV transmission are shown in Table V and evaluated as risk factors for NHL among cases and controls and as risk factors for HCV positivity among controls only. A somewhat increased NHL risk was seen for permanent tattoos (OR = 2.5, 95% CI 0.9–7.0). History of surgical operations showed a significant inverse association, as expected, since diseases requiring surgical operations affected a subgroup of controls. History of blood transfusion, tooth extraction, piercing (mainly for earrings) and employment in health professions were unrelated to NHL risk. Two cases and no controls reported ever using i.v. drugs. With respect to risk factors for HCV positivity in the control group, associations were found with old age (OR for 65 or older vs. younger than 45 years = 16.8, 95% CI 2.2–131.1), history of blood transfusion before 1990 (OR = 4.6, 95% CI 2.0–10.3) and presence of permanent tattoos (OR = 8.5, 95% CI 1.1–67.4). Conversely, surgical operations, tooth extraction, piercing and employment in health professions were not related to HCV positivity (Table V).

Table V. OR and Corresponding 95% CI1 for NHL Among Cases and Controls and of HCV Positivity Among Controls by Selected Variables: Italy, 1999–2002
 All subjectsControls only
Cases (n = 225) number (%)Controls (n = 504) number (%)OR (95% CI)HCV+ (n = 45) number (%)HCV (n = 459) number (%)OR (95% CI)
  • 1

    Estimated from unconditional logistic regression adjusted for gender, age, center, education and place of birth when appropriate.

  • 2

    The sum does not add up to the total because of some missing values.

  • 3

    Reference category.

Age (years)      
 <45 1 (2.2)103 (22.4)13
 45–54 4 (8.9)54 (11.8)7.92 (0.85–73.40)
 55–64 11 (24.4)108 (23.5)12.11 (1.49–98.16)
 ≥65 29 (64.4)194 (42.3)16.81 (2.16–131.06)
Study center      
 Aviano/Pordenone 21 (46.7)259 (56.4)13
 Naples 24 (53.3)200 (43.6)1.94 (0.47–8.01)
Blood transfusions      
 Never196 (87.1)438 (86.9)1331 (68.9)407 (88.7)13
 Ever29 (12.9)66 (13.1)0.98 (0.60–1.59)14 (31.1)52 (11.3)3.57 (1.67–7.62)
Year of first blood transfusion2      
 Never196 (87.5)438 (87.1)1331 (68.9)407 (88.9)13
 <199016 (7.1)47 (9.3)0.73 (0.40–1.35)12 (26.7)35 (7.6)4.55 (2.02–10.25)
 ≥199012 (5.4)18 (3.6)1.60 (0.73–3.50)2 (4.4)16 (3.5)1.46 (0.30–7.27)
Surgical operations      
 No61 (27.1)88 (17.5)137 (15.6)81 (17.6)13
 Yes164 (72.9)416 (82.5)0.56 (0.38–0.83)38 (84.4)378 (82.4)1.29 (0.53–3.17)
Tooth extraction2      
 Never27 (14.1)44 (9.1)132 (4.5)42 (9.6)13
 1–14107 (55.7)252 (52.3)0.65 (0.37–1.14)19 (43.2)233 (53.2)1.08 (0.23–5.18)
 ≥1558 (30.2)186 (38.6)0.63 (0.33–1.23)23 (52.3)163 (37.2)1.09 (0.21–5.61)
Piercing2      
 No112 (59.6)347 (71.8)1330 (68.2)317 (72.2)13
 Yes76 (40.4)136 (28.2)0.92 (0.47–1.81)14 (31.8)122 (27.8)0.94 (0.26–3.36)
Permanent tattoo2      
 No180 (96.3)473 (97.9)1342 (95.5)431 (98.2)13
 Yes7 (3.7)10 (2.1)2.46 (0.87–6.99)2 (4.5)8 (1.8)8.49 (1.07–67.37)
Employment in health professions      
 Never214 (95.1)486 (96.4)1344 (97.8)442 (96.3)13
 Ever11 (4.9)18 (3.6)1.06 (0.47–2.40)1 (2.2)17 (3.7)0.55 (0.07–4.47)

The association of NHL with HCV positivity was reassessed separately in different strata of the study population, achieving consistent results (data not shown in tables). The OR for NHL in HCV+ individuals was similar in men (OR = 2.6, 95% CI 1.3–5.1) and women (OR = 2.8, 95% CI 1.3–6.1) and in individuals aged 60 years or older (OR = 3.0, 95% CI 1.6–5.4) and in younger ones (OR = 2.5, 95% CI 1.0–5.9). The association with HCV positivity was stronger, though not significantly, in Naples (OR = 3.6, 95% CI 1.9–7.0) than in Aviano/Pordenone (OR = 1.9, 95% CI 0.9–4.2). Part of the difference between the 2 study centers was accounted for by a greater proportion of follicular NHL cases, all of which were HCV, in Aviano/Pordenone than in Naples.

DISCUSSION

The present case-control study provides further evidence that HCV infection is associated with risk of NHL, notably B-cell NHL, among HIV individuals. Our present OR of 2.6 is consistent with the findings of several previous reports that included adequate control groups,13, 14, 15, 16, 17, 18, 19 notwithstanding wide variations in HCV prevalence in the studied populations.

The association with NHL was found in our study regardless of whether HCV positivity was defined on the basis of anti-HCV antibodies, which chiefly reflect past HCV infection, or carriage of HCV RNA, which indicates persistent infection. However, the majority of NHL cases (81.8%) and controls (66.7%) who were positive for anti-HCV antibodies were also positive for HCV RNA, thus obscuring whether persistent HCV infection is required for an increased NHL risk to be found. The OR for 2a/2c HCV genotype was higher than that for 1b genotype, but the difference was not statistically significant. Silvestri et al.32 showed an excess of 2a/2c genotype among HCV RNA+ NHL patients compared to blood donors and patients with chronic liver disease, suggesting that this genotype may be particularly involved in the pathogenesis of lymphoproliferative and autoimmune disorders.

The association with HCV positivity was found for a broad range of NHL subtypes, though it appeared stronger for SLL/CLL, lymphoplasmacytic, marginal zone and MALT NHL than for follicular NHL. Only one patient had splenic lymphoma with villous lymphocytes, the histologic subtype shown to regress following treatment of HCV infection.33

Although all NHL cases were reclassified according to the REAL/WHO classification,28 the distinction of NHL subtypes remains difficult and some histologic subtypes were rare. Caution must, therefore, be used in comparing the strength of the association with HCV positivity across different histologic subtypes. Especially high HCV prevalence for lymphoplasmacytic NHL16, 34, 35 and some types of MALT NHL14 is in agreement with a few previous reports. Moreover, we suggest a lack of HCV positivity among follicular NHL. The 2 major groups of indolent (low-grade) NHL and aggressive (intermediate- and high-grade) B-cell NHLs showed similar ORs.

HCV may thus differ from other infectious agents that tend to be associated with specific subsets of NHL, such as acute T-cell leukemia/lymphoma for human T-cell leukemia/lymphoma virus, endemic Burkitt's lymphoma and immunoblastic lymphoma of immunosuppressed patients for Epstein-Barr virus and MALT lymphoma of the stomach for Helicobacter pylori.36 The prevailing pathogenic hypothesis for HCV may explain the involvement of a broad spectrum of B-cell NHL. Specific B-cell clones may proliferate as a consequence of the chronic antigenic stimulation sustained by HCV. The Ig variable region genes expressed by B-NHL cells from HCV+ patients show somatic mutations indicative of an antigen-selection process.12 If this is the case, HCV infection may be associated with all those B-cell lymphomas (the vast majority28) whose cell of origin has been chronically stimulated by HCV antigens in germinal center-like structures, where vigorous clonal expansion of B cells takes place and the risk of mutations and malignant transformation is highest.37 Not expected to be associated with HCV, then, would be T-cell and mantle-cell NHL as well as a subset of CLL and Burkitt's lymphoma, in addition to Hodgkin's lymphoma.28 Our present findings and the aforementioned pathogenetic hypothesis would apply chiefly to immunocompetent individuals. HCV has not been demonstrated to increase NHL risk among, e.g., HIV+ subjects.38

The few chronic carriers of HBsAg appeared to have an increased NHL risk, too. Like HCV, HBV can replicate in extrahepatic tissues, including lymph nodes and bone marrow; but the virus has been so far evaluated in few NHL case series.23, 39 Kim et al.23 reported for HBsAg+ individuals in South Korea an OR for B-cell NHL of 4.57 (95% CI 2.03–10.29) after adjustment for age, gender, smoking, alcohol drinking, transfusion history and HCV positivity. Findings on HBV should, however, be interpreted with great caution as they are based on a small number of HBV carriers. Reactivation of HBV infection in NHL cases is possible but unlikely as all blood samples in the present study were obtained before any cytotoxic treatment had been administered.40

Our study also offered the opportunity to quantify further, by means of a comparison of HCV+ and HCV controls, major risk factors for HCV infection in Italy, chiefly blood transfusions. The proportion of HCV infection attributable to blood transfusion among controls was, however, only 22.0% (95% CI 6.3–37.8%), thus suggesting that the majority of HCV infections in Italy were acquired through parenteral routes of transmission other than transfusion, possibly because of the use of nondisposable needles up to the 1970s.5 The age distribution of HCV positivity in our control group showed a very strong cohort effect.

Hospital-based case-control studies are prone to selection bias, but it is unlikely that our OR for HCV was inflated by a lack of HCV+ control subjects. Indeed, the prevalence of HCV positivity in our control group and the HCV distribution by area of living, place of birth, age and known risk factors for HCV was consistent with the findings of previous HCV surveys in Italy.5, 6, 25 Selection bias with respect to NHL cases is of greater concern. Indeed, it is difficult to rule out that NHL cases who are HCV+ or present HCV-related complications are selectively referred to the specialized oncology departments participating in the present investigations. The exclusion of 13 NHL cases where it was unclear whether symptoms of hepatic disease or of NHL had triggered the diagnostic process did not materially modify our finding. Acquisition of HCV as a consequence of NHL treatment can also be confidently excluded since we restricted our study to newly diagnosed cases. A confounding effect of blood transfusions, which have been associated with NHL in some (but not all) previous studies on the topic,41 can also be ruled out since history of blood transfusions was not significantly related to NHL risk in our study. Accurate testing for HCV in a centralized laboratory and the revision of NHL diagnoses represent important strengths of our present study.

In conclusion, our study shows that HCV infection is associated with an increase in NHL risk. The fraction of NHL cases attributable to HCV, according to our study, is 12.4% (range 6.3–18.5%). Avoidance or treatment of HCV infection may thus help to reduce a sizeable number of NHL cases, notably in areas where HCV infection is frequent6 or currently increasing.42 The relationship between HBsAg carriage and NHL deserves further evaluation.

Acknowledgements

We thank Dr. C. La Vecchia for useful comments; Ms. O. Volpato for study coordination;, Drs. G. Laconca, M. Grimaldi and O. Manganelli for help in data collection; and Ms. I. Calderan and Ms. L Mei for editorial assistance. We also acknowledge the laboratory technicians of the Division of Microbiology and Immunology, General Hospital of Pordenone, for the HCV antibody and virologic testing.

Ancillary