CORRESPONDENCE Re: Detection of Epstein-Barr Virus in Invasive Breast Cancers

Bonnet and co-workers (1) presented evidence on the presence of EpsteinBarr virus (EBV) sequences in breast cancer cells. We wanted to examine whether epidemiologic evidence would support the role of EBV infection in breast cancer. The tumorigenicity of EBV appears best established for Hodgkin’s disease in Europe and North America; some 40% of Hodgkin’s disease cases are attributed to EBV infection (2), and most show viral genomes in serum (3). We reason that, if breast cancer were related to EBV, we should find an epidemiologic association between breast cancer and Hodgkin’s disease. Saliva is the main route of EBV infection (4), and family members of women with breast cancer should show an excess of Hodgkin’s disease if EBV were the shared etiologic agent. To test these hypotheses, we used the nationwide Swedish Family-Cancer Database on 9.6 million people organized in families and containing cancer data from the Swedish Cancer Registry from years 1958 through 1996 (5). Standardized incidence ratios (SIRs) were calculated based on age-standardized national rates available from the database; 95% confidence intervals (95% CI) were calculated assuming a Poisson distribution. We first analyzed Hodgkin’s disease incidence as a second cancer in women diagnosed with breast cancer. There were 21 cases giving an SIR of 1.05 (95% CI 4 0.65–1.55). Analyzing breast cancer incidence after Hodgkin’s disease showed an SIR of 2.32 (n 4 45; 95% CI 4 1.69–3.04), but the excess was only observed greater than or equal to 5 years after Hodgkin’s disease, a well-known effect of radiotherapy (6); only two cases occurred 0–4 years after Hodgkin’s disease diagnosis, while 7.44 cases were expected. As a second hypothesis, we examined breast cancers and Hodgkin’s disease incidence in daughters of mothers who had either disease, but there was no increase in any SIR (Table 1). We found 88 couples in whom the wife presented with breast cancer and the husband with Hodgkin’s disease; the SIR for the wife with breast cancer was 0.94, and the SIR for the husband with Hodgkin’s disease was 1.01. In one couple, both spouses presented with Hodgkin’s disease compared with 1.6 couples expected. We thus failed to verify the hypotheses linking breast cancer and EBV, using Hodgkin’s disease as a surrogate marker for EBV infection. Elsewhere, we have found epidemiologic evidence incriminating EBV. With the use of the above database, we studied second cancers after squamous cell carcinoma of the skin and found statistically significant increases for salivary gland and nasal cancers in men and women (7). Nasopharyngeal cancer was also in excess, but the number of cases was small. Since salivary glands and nasopharynx are known target tissues for EBV-related carcinogenesis (4), we attributed the findings to the presence of this virus among patients who may be immunocompromised. The finding of elevated risk of salivary gland cancer has been remarkably consistent, although unexplained, in at least six previous follow-up studies on skin cancer (7). Recently, we were able to reproduce the findings for salivary gland and nasal cancers as second cancers in a follow-up study of patients with in situ squamous cell carcinoma of the skin. Although EBV-attributable cases of salivary gland and nasal cancers would be few in incidence, the results suggest that the virus may gain control once the immunologic surveillance falters, as it may locally in a tumor, such as breast cancer. This mechanism is consistent with the results of Bonnet et al. (4).