Throughout much of the 20th century, the rates of non-Hodgkin lymphoma (NHL) have increased in the U.S. and other developed countries for reasons that are still not fully understood. A recent article indicated that residential exposure to U.S. Environmental Protection Agency (U.S. EPA)-designated Superfund sites was a significant risk factor for an individual's risk to develop NHL (Webber & Stone, 2017). In this article, we are reporting on the disparities in the age adjusted prevalence rates of NHL for males and females as associated with individual residential proximity to Kentucky's Superfund sites. In Kentucky, like national and global trends, age-adjusted NHL rates in males consistently exceed rates in females (Al-Hamadani et al., 2015; Devesa & Fears, 1992; Roman & Smith, 2011), even though NHL is typically a cancer that impacts the older population where females outnumber males.
Although there are known links between autoimmune diseases and NHL (Ekstrom Smedby et al., 2008), and autoimmune diseases are more frequent in females (Ansell et al., 2011), males continue to experience higher rates of all types of NHL. Differences among males and females are less pronounced in regions of the world with the highest human development indices; however, those areas experience higher cases of Burkitt's lymphoma (Roman & Smith, 2011). Higher NHL rates in males are seen across all subtypes, from more common intranodal subtypes such as diffuse large B-cell lymphoma (Hedstrom et al., 2015), follicular lymphoma (Nabhan et al., 2016), and mantle cell lymphoma (Aschebrook-Kilfoy, Caces, Ollberding, Smith, & Chiu, 2013), to the rarest forms of extranodal NHL such as primary central nervous system lymphoma (Villano, Koshy, Shaikh, Dolecek, & McCarthy, 2011) and primary gastric lymphoma (Padhi et al., 2012).
Male-female differences were observed not only in NHL rates but also in NHL comorbid diseases. A European observational study, including 40 countries, reported that males were more likely to be diagnosed with cutaneous melanoma and NHL, whereas in females the association between NHL and melanoma was negative (Allam et al., 2015). In addition, significant associations were found between NHL and renal cell carcinoma in males, but not in females (Lossos, Ferrell, Duncan, & Lossos, 2011). Another study found that a high body mass index at age 18 is associated with a significantly higher NHL risk for females, but there was no such association observed in males (Kelly et al., 2012).
Males with certain subtypes of NHL do not appear to respond as well to the immunochemotherapeutic compound rituximab as females do (Pfreundschuh et al., 2010; Riihijarvi, Taskinen, Jerkeman, & Leppa, 2011). Yet, other studies showed that only female patients responded well to the chemotherapeutic compound lenalidomide (Eve et al., 2012) and that, based on a murine T-cell lymphoma model, aspirin demonstrated greater antitumor properties in females than in males (Kumar, Vishvakarma, Bharti, & Singh, 2012).
Endocrine regulation might partially explain the differences noted in the prevalence of NHL, survival rates, and differences in response to NHL treatments in males and females (Yakimchuk et al., 2011). In vivo murine models revealed that the estrogen receptor beta-agonists inhibited the proliferation, vascularization, and dissemination of lymphoid tumors (Yakimchuk et al., 2014). Reduction in serum interleukin-6 triggered by 17-beta-estradiol is another mechanism by which estrogen might lower NHL risk (Horesh & Horowitz, 2014; Rachon, Mysliwska, Suchecka-Rachon, Wieckiewicz, & Mysliwski, 2002). Pregnancy has also been shown to be a protective factor for NHL prevalence (Horesh & Horowitz, 2014; Prescott et al., 2009), as has the use of oral contraceptives (Lee, Bracci, & Holly, 2008).
Differences among males and females in NHL prevalence and treatment response appear to indicate that efficacy of chemotherapeutics is possibly connected to unidentified and sex-specific polymorphisms in genes that code for glutathione S-transferases (Cho et al., 2010; Riihijarvi et al., 2011).
This correlational study used 1998-2012 cancer registry records, obtained in 2014 from the Kentucky Cancer Registry, for the first diagnosis of intranodal or extranodal NHL; cancer records from adjacent states were not available for analyses. The following variables were included in the analyses at the individual level: sex, race, and ethnicity; age at diagnosis; family history of NHL; county of residence; Appalachia residence; and Beale Code for the level of urbanization. Using the geographic coordinates for the patient's residential address, the census tract was identified for 82.3% of the NHL cases, while for 17.7% the residential ZIP centroid was used instead. We obtained the census tract Topologically Integrated Geographic Encoding and Referencing (TIGER) file from the
2010 U.S. Census website. With the TIGER file, we could identify the residential census tract only for 82.3% of cases; the rest (17.7%) could not be placed in a census tract, most likely because their listed address was a post office box or rural route. For those we could not place in a census tract, we had to estimate where they lived by placing their residence in the exact middle (centroid) of their ZIP code.
According to the 2010 U.S. Census, Kentucky included 1,115 census tracts. Of the 1,115, 734 tracts reported cases of NHL between 1995 and 2012. Only 145 census tracts (13%) in Kentucky had Superfund sites located within their borders and there was a maximum of 5 sites per tract. At the time of this study in 2014, the U.S. EPA website listed 133 Superfund sites located in Kentucky. The exposure risk is defined as the patient's residential proximity to Superfund sites and it was operationalized as an ordinal variable with three categories: 0 = exposure risk beyond 10 km; 1 = exposure risk within 10 km, but beyond...