|Home | About | Journals | Submit | Contact Us | Français|
To assess the association between animal exposures and non-Hodgkin lymphoma (NHL).
Exposure data were collected from 1,591 cases and 2,515 controls during in-person interviews in a population-based case-control study of NHL in the San Francisco Bay Area. Odds ratios (ORs) and 95% confidence intervals (CIs) were adjusted for potential confounders.
Pet owners had a reduced risk of NHL (OR=0.71,CI=0.52 –0.97) and diffuse large-cell and immunoblastic large-cell (DLCL;OR=0.58,CI=0.39 –0.87) compared with those who never had owned a pet. Ever having owned dogs and/or cats was associated with reduced risk of all NHL (OR=0.71,CI=0.54–0.94) and of DLCL (OR=0.60,CI=0.42–0.86). Longer duration of cat ownership (p-trend=0.008), dog ownership (p-trend=0.04), and dog and/or cat ownership (p-trend =0.004) was inversely associated with risk of NHL. Ownership of pets other than cats and dogs was associated with a reduced risk of NHL (OR=0.64,CI=0.55–0.74) and DLCL (OR=0.58,CI=0.47 –0.71). Exposure to cattle for ≥5 years was associated with an increased risk of NHL (OR=1.6,CI=1.0–2.5) as was exposure to pigs for all NHL (OR=1.8,CI=1.2–2.6) and for DLCL (OR=2.0,CI=1.2–3.4).
The association between animal exposure and NHL warrants further investigation in pooled analyses.
In 2008 in the United States 66,120 newly diagnosed cases and 19,160 deaths from non-Hodgkin Lymphoma (NHL) are expected (1). Few risk factors for NHL have been identified and recent investigation has focused on environmental factors that could explain the past increase in NHL incidence including viral, chemical, lifestyle and occupational exposures[reviewed in (2)]. The literature on environmental exposures and risk of NHL is inconsistent and few etiologic exposures have been identified [reviewed in (2)]. NHL arises from cells of the immune system and infection, immune deficiency, and autoimmunity are among the strongest established risk factors (2).
Many studies have assessed relationships between occupations and NHL incidence with some implicating exposure to chemicals or other industry-specific hazards. Elevated risks for NHL have been associated with exposure to cattle (3) and other livestock (3–5). Some (6) but not other (7, 8) studies have shown that contact with unusual pets was associated with an increased risk of NHL. We assessed domestic and farm-animal exposure history in relation to NHL through in-person interviews in our population-based case-control study of NHL in the San Francisco Bay Area.
Detailed methods of patient recruitment have been reported (8, 9). Briefly, rapid case ascertainment was used to identify NHL patients within one month of diagnosis in six San Francisco Bay Area counties. Eligible patients were diagnosed between 1988–1993, were between 21–74 years of age and resided in the six Bay Area counties at the time of diagnosis. A total of 1591 (72%) eligible NHL patients completed interviews. Control participants (n=2515, 78% response rate) were identified using random-digit dial and were frequency matched to the cases by sex, county of residence and age in 5-year groups. Eligibility criteria were the same as for cases with the exception of NHL diagnosis. Diagnostic pathology materials were re-reviewed and classified using the Working Formulation by the expert study pathologist. For our subtype analyses, all diffuse large-cell and immunoblastic large-cell (DLCL) subtypes were combined and follicular lymphoma (FL) subtypes were combined to reflect the WHO classification (10). The subtype results are presented for DLCL, FL, small-lymphocytic lymphoma (SLL) and “other NHL” (10) due to sample size constraints for the smaller subgroups.
Study protocols and procedures were approved by the UCSF Committee on Human Research and all participants provided written informed consent prior to interview. Structured interviews were conducted in-person by trained interviewers in the study participants' homes or at a place convenient to the participant. Detailed questions were asked regarding history of occupation, use of therapeutic drugs, immunizations, allergies, viral infections (including HIV), and lifestyle factors. Questions pertaining to history of farming, agricultural work with animals, and pet ownership also were asked. Most questions pertained to incidence of exposures or activities up until one year before diagnosis (cases) or interview (controls). Bloods were drawn from consenting participants who had no history of chemotherapy within the past 3 months. Blood was tested for HIV using the autoantigen enzyme-linked immunosorbent assay with a Western blot assay confirmation.
Unconditional logistic models were computed using SAS (v.9.1; SAS Institute, Cary, NC) to obtain odds ratios (ORs) as estimates of relative risks (hereafter called risk) and 95% confidence intervals (CIs). All statistical tests were two-sided and all models were adjusted for sex and age in 5-year groups. Potential confounders were included in full multivariable models if their inclusion changed OR estimates ≥10%. Potential confounders included race/ethnicity, education level, self-report of animal and plant allergies, number of siblings and age of first pet or farm exposure. Analyses were restricted to HIV-negative participants (N=1262 cases, N=2094 controls). Cut points for pet ownership and farm animal exposure were based on 5- or 10-year increments.
Distribution by education, race/ethnicity and sex for cases and controls are presented for total NHL and the four major NHL histologic subtypes (Table 1). The mean age for all HIV-negative NHL cases and controls was 57 and 54 years, respectively. Cases were somewhat less educated and a greater proportion of cases were men. This was consistent for total NHL, DLCL, FL and SLL but not the “other NHL” subtypes.
Ever having owned a pet was associated with a reduced risk of all NHL (OR=0.71,CI=0.52–0.97) and of “other NHL” (OR=0.60,CI=0.36–1.0) compared with those who never had owned a pet (Table 2). Ever having owned a dog (OR=0.79,CI=0.64–0.97) or a dog and/or cat (OR=0.71,CI=0.54–0.94) was associated with reduced risk of all NHL. Longer duration of cat ownership (p-trend=0.008), dog (p-trend=0.04), and dog and/or cat ownership (p-trend=0.004) was inversely associated with risk of all NHL. All analyses of cat and dog ownership included those who had owned other pets. Ever having owned birds was not associated with risk of NHL. Ever having owned pets other than cats and dogs was associated with a reduced risk of NHL (OR=0.64,CI=0.55–0.74) although there was no linear trend effect of duration (p-trend=0.39). Age when first owned a pet was not associated with risk of NHL.
The association between pet ownership and NHL risk differed by histologic subtype (Table 2). Any prior pet ownership was associated with a reduced risk of DLCL (OR=0.58,CI=0.39–0.87) as was cat (OR=0.84,CI=0.69–1.00), dog (OR=0.75,CI=0.57–1.00, p-trend=0.005) and dog and/or cat ownership (OR=0.60,CI=0.42–0.86, p-trend=0.008). Ever having owned birds (OR=0.68,CI=0.53–0.86) or pets other than cats and dogs also was associated with a reduced risk of DLCL (OR=0.58,CI=0.47–0.71). Ever having owned pets other than cats and dogs was associated with a reduced risk of FL (OR=0.78,CI=0.62–0.99), SLL (OR=0.63,CI=0.45–0.89) and “other NHL” (OR=0.59,CI=0.46–0.77). There were no other consistent associations between pet ownership and the FL, SLL and the “other NHL” subtypes.
Because those who reported never having owned pets comprised a very small proportion of participants, we compared characteristics of pet owners and non-owners. Compared with participants who had not owned a pet, those who ever had owned a pet were less likely to be Hispanic (7%vs.13%;χ2, P<0.0001), African-American (5%vs.12%;χ2,P=0.0002) or Asian (5%vs.24%;χ2,P<0.0001). Demographic characteristics, sex (χ2,P=0.92) and education level (χ2,P=0.10), were not associated with pet ownership. History of animal (8%vs.6%;χ2,P=0.36) and plant allergies (15%and13%;χ2,P=0.50) also were similar for pet owners and non-owners, respectively whereas pet owners were less likely to have three or more siblings(22%vs.37%;χ2,P<0.0001). Adjustment for race, ethnicity, education level, age at first pet exposure, allergies and number of siblings did not affect the overall or subtype associations between pet ownership or farm animal exposure and NHL risk (data not shown).
Neither living nor working on a farm was associated with NHL risk regardless of duration or age when first lived on a farm (Table 3). Exposure to cattle for ≥5 years was associated with an increased risk of all NHL (OR=1.6,CI=1.0–2.5). Exposure to pigs was associated with an increased risk of all NHL (OR=1.5,CI=1.1–2.1) and DLCL (OR=1.6,CI=0.99–2.5) with increased risk for those with ≥5 years exposure (all NHL, OR=1.8,CI=1.2–2.6; DLCL, OR=2.0,CI=1.2–3.4).
Because a small proportion of all study participants reported ever having worked with cattle and pigs, we compared characteristics of those who worked with these farm animals with those who did not work with farm animals. Participants who worked with cattle or pigs, respectively, were more likely to have been men (75%vs.55%;χ2,P<0.0001; 77%vs.55%;χ2,P<0.0001), African-American (8%vs.5%;χ2,P=0.02; 12%vs.5%;χ2,P<0.0001), to have ≤12 years of education (44%vs.32%;χ2,P<0.0001; 46%vs.32%;χ2,P<0.0001) and were less likely to have been Asian (1%vs.8%;χ2,P<0.0001; 2%vs.8%;χ2,P<0.0001). There was no difference by Hispanic ethnicity (8%vs.7%;χ2, P=0.63; 9%vs.7%;χ2,P=0.21) or by history of animal (7%vs.8%;χ2,P=0.22; 8%vs.8%;χ2,P=0.68) or plant allergies (13%and15%;χ2,P=0.24; 13%and15%;χ2,P=0.33).
This large population-based case-control study provides support for a decreased risk between pet ownership and NHL and an increased risk between farm-animal exposure and NHL. An increasing total number of years of pet ownership was associated with a reduced risk of NHL. The association between pet ownership and NHL in earlier reports has been inconsistent (6–8). A study of household pets including dogs, cats, caged birds and exotic pets (e.g. lizards, fish, monkeys) in Sweden showed no association with NHL risk (7) whereas a small pooled study of Swedish NHL cases and controls found that contact with unusual pets (not dogs, cats and birds) was associated with an increased risk of NHL (6).
Given the evidence that animal exposure during infancy may reduce the prevalence of allergic sensitization and allergic disease later in childhood (11–13), it is possible that the association between pet ownership and NHL may be related to altered immune function and desensitization to allergens. Most recent studies have shown that exposure to cats and dogs early in infancy reduces the risk of allergic sensitization and the prevalence of some atopic diseases such as asthma and eczema (11, 14). Several studies have shown that the presence of animals and especially the presence of dogs in households is associated with increased levels of airborne and settled endotoxins (14, 15) and that higher levels of house-dust endotoxin have been associated with reduced risk of atopic diseases in early childhood (14, 16).
It is well established that immunocompetence and immune-related activity are associated with NHL development (9, 17), although specific mechanisms are lacking and our observations were not explained by early life exposure. Inverse associations between allergies and risk of NHL have been reported (18–21) as have childhood and immune-related factors (9, 17, 22). While we cannot rule out the role of immune function in explaining the reduced risk of NHL observed among pet owners, the proportion of animal and plant allergies was similar for pet owners and non-owners. The role of allergic reactions in NHL is likely to be complex and may be related to B-cell differentiation as allergic individuals have more efficient immune systems that are required for B-cell activation and differentiation (19, 23).
Similar to our results for cattle and pigs, elevated risks for NHL have been associated with exposure to livestock (3–5), mainly cattle (3). Exposure to zoonotic viruses among farm-animal workers also may be an underlying exposure associated with increased NHL risk, as has been proposed for slaughterhouse workers, meat inspectors, meat packagers and processors (24). Several methodologic and etiologic factors may be contributing to the inconsistent associations reported for NHL and livestock exposure. Assessing exposure to animals on a farm is complicated because these exposures are likely to interact with or be confounded by other farming-related exposures, such as other animals or pesticides (3). Several studies have reported associations between pesticide exposure and NHL (4, 25–27). However, the types of pesticides (herbicides vs. insecticides) used and the methods of application differ between crop and livestock farmers. One study reported elevated risks for NHL among a greater number of chemical classes of insecticides used on livestock than among groups of insecticides used on crops and NHL risk largely was restricted to the livestock farmers and not the crop farmers (25).
Factors related to potential for bias should be considered when interpreting these results. Methods used to diminish potential selection bias included random-digit-dial to identify age-, sex- and county-matched controls from the same population where the cases arose, and rapid case ascertainment was used to quickly identify all incident NHL cases to diminish survivor bias. Experienced interviewers, blinded to the study hypotheses, conducted in-person interviews with participants to diminish interviewer and response bias and no proxy interviews were conducted. The similarity of response rates for the cases and controls is a strength of this study, and refusal response rates were low. In addition, most interviews were conducted in participants' homes, allowing privacy and easy access to items to confirm some exposures. Nondifferential exposure misclassification can lead to case-control odds ratios that are biased toward the null. However, because animal exposure is not a known risk factor for NHL there was less likelihood of recall differential between cases and controls or of response bias associated with these data.
Whereas many studies have assessed relationships between specific exposures and NHL incidence, few have been associated consistently with risk of NHL (2). Further investigation using pooled analyses of multiple case-control studies is warranted to clarify the effect of pet and agricultural animal exposures on immune function and risk of NHL.
Support provided by grant number CA45614 and in part by grants CA89745, CA66529 and CA87014 from the National Cancer Institute, National Institutes of Health