CORRESPONDENCE Re: Cancer Incidence in Denmark Following Exposure to Poliovirus Vaccine Contaminated With Simian Virus

We read with interest the recent article by Engels et al. (1) in which the authors performed a retrospective birthcohort analysis in Denmark following exposure to poliovirus vaccine contaminated with simian virus 40 (SV40) to clarify whether SV40 infection increases risk of mesothelioma, choroid plexus tumors, and non-Hodgkin’s lymphoma, or of cancers arising in children. The authors concluded that “exposure to SV40-contaminated poliovirus vaccine in Denmark was not associated with increased cancer incidence.” It is important to point out that several limitations have been recognized for this and similar epidemiologic studies addressing exposure to SV40contaminated poliovirus vaccines and the incidence of human cancers (Table 1) (2–4). Indeed, an evaluation by the Institute of Medicine Immunization and Safety Review Committee found that the epidemiologic data used in birthcohort studies to examine cancer rates in individuals potentially exposed to SV40-contaminated vaccines are inadequate to evaluate a causal relationship (2). The validity of observational studies, such as the retrospective analyses by Engels et al. (1), depends on the accuracy of the existing knowledge of the biologic properties of SV40 and the identification of the human population infected with the virus (2,4,5). Therefore, supportive evidence from experimental studies is required to draw causal inferences in human disease (4,6). Indeed, a recent case–control study (5) of 1793 cancer patients indicated that there is a statistically significant excess risk of SV40 associated with primary brain cancers (odds ratio [OR] 3.8, 95% confidence interval [CI] 2.6 to 5.7), primary bone cancers (OR 24.5, 95% CI 6.8 to 87.9), malignant mesothelioma (OR 15.1, 95% CI 9.2 to 25.0), and non-Hodgkin’s lymphoma (OR 5.4, 95% CI 3.1 to 9.3). A known confounding factor in the observational study by Engels et al. (1) is that the actual number of people infected with live SV40 through the use of contaminated poliovirus vaccines in Denmark or other countries is not known (2–4). It is also recognized that not all vaccine lots were contaminated with SV40, that formalin inactivation may have reduced the titer of live SV40 in the vaccine lots that were contaminated, and that successful infection rates by live SV40 are unknown (2–4). For example, only 19% of newborn children and 15% of infants aged 3–6 months at the time of receiving a known contaminated oral poliovirus vaccine excreted infectious SV40 in their stools for up to 5 weeks after vaccination (4), indicating an established infection. Therefore, an inability to identify the population actually infected with SV40 in Denmark through the use of contaminated poliovirus vaccines precludes a meaningful calculation of cancer incidence in relation to exposure to those vaccines. Furthermore, there is ample evidence that some individuals acquire SV40 infection from sources other than poliovirus vaccines (2,5), indicating that the individuals in the unexposed group identified by Engels et al. (1) may also have been infected with SV40. These shortcomings led the Institute of Medicine committee to recommend that no additional epidemiologic studies of individuals potentially exposed to contaminated poliovirus vaccine be initiated (2). Hence, future studies need to focus on how SV40 is transmitted in humans today, how it is distributed throughout the infected host, how the virus interacts with different tissues, and how the host responds immunologically to this infection.