Drs. Marinella and Moseley raise an interesting point about the possible association between non-Hodgkin B-cell lymphoma (NHL) and chronic HCV infection. As they mention, Ferri and colleagues [1] detected HCV viremia in 32% (16 of 50) of unselected patients with NHL compared with 3% of patients with Hodgkin lymphoma and 1.3% of healthy controls. At first glance, this marked discrepancy in the prevalence of HCV infection between patients with NHL and healthy controls cannot be ignored, but careful review of this study raises two concerns. First, blood exposure was reported in 10 of 50 patients, but the authors state that such exposure did not correlate with HCV infection without providing the supporting data. Second, patients with NHL were reported to be "Italian-born heterosexuals with no history of drug use." Could these patients have had other risk factors for HCV infection, such as tattooing; hemodialysis; high-risk sexual behavior; or undisclosed parenteral exposures, including intranasal cocaine use [2]? The absence of such information and the relatively small number of patients with NHL studied make it difficult to interpret these findings.

The development of NHL is probably a multifactorial process, and a causal relation with one or more oncogenic viruses remains possible. To satisfy the Bradford-Hill guidelines for causality in the domain of carcinogenesis [3], the cause should precede the effect and the association should be consistent, should be biologically plausible, and should show a strong association as well as a dose-response relation. Long-term follow-up studies in patients with post-transfusion HCV infection have failed to report an increased incidence of lymphoreticular malignant conditions, but these studies were not designed to investigate this association. For that purpose, large case–control studies, akin to those showing an association between Helicobacter pylori and gastric carcinoma [4, 5], are needed. If, as a result, an epidemiologic association is confirmed, questions about biological plausibility and mechanisms of oncogenesis of HCV in lymphoreticular malignant conditions would need to be answered. Until then, the association between HCV infection and NHL, if any, remains unclear.