Occupation and Risk of Non-Hodgkin Lymphoma and Its Subtypes: A Pooled Analysis from the InterLymph Consortium

Background: Various occupations have been associated with an elevated risk of non-Hodgkin lymphoma (NHL), but results have been inconsistent across studies. Objectives: We investigated occupational risk of NHL and of four common NHL subtypes with particular focus on occupations of a priori interest. Methods: We conducted a pooled analysis of 10,046 cases and 12,025 controls from 10 NHL studies participating in the InterLymph Consortium. We harmonized the occupational coding using the 1968 International Standard Classification of Occupations (ISCO-1968) and grouped occupations previously associated with NHL into 25 a priori groups. Odds ratios (ORs) adjusted for center, age, and sex were determined for NHL overall and for the following four subtypes: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), and peripheral T-cell lymphoma (PTCL). Results: We confirmed previously reported positive associations between NHL and farming occupations [field crop/vegetable farm workers OR = 1.26; 95% confidence interval (CI): 1.05, 1.51; general farm workers OR = 1.19; 95% CI: 1.03, 1.37]; we also confirmed associations of NHL with specific occupations such as women’s hairdressers (OR = 1.34; 95% CI: 1.02, 1.74), charworkers/cleaners (OR = 1.17; 95% CI: 1.01, 1.36), spray-painters (OR = 2.07; 95% CI: 1.30, 3.29), electrical wiremen (OR = 1.24; 95% CI: 1.00, 1.54), and carpenters (OR = 1.42; 95% CI: 1.04, 1.93). We observed subtype-specific associations for DLBCL and CLL/SLL in women’s hairdressers and for DLBCL and PTCL in textile workers. Conclusions: Our pooled analysis of 10 international studies adds to evidence suggesting that farming, hairdressing, and textile industry–related exposures may contribute to NHL risk. Associations with women’s hairdresser and textile occupations may be specific for certain NHL subtypes. Citation: ‘t Mannetje A, De Roos AJ, Boffetta P, Vermeulen R, Benke G, Fritschi L, Brennan P, Foretova L, Maynadié M, Becker N, Nieters A, Staines A, Campagna M, Chiu B, Clavel J, de Sanjose S, Hartge P, Holly EA, Bracci P, Linet MS, Monnereau A, Orsi L, Purdue MP, Rothman N, Lan Q, Kane E, Seniori Costantini A, Miligi L, Spinelli JJ, Zheng T, Cocco P, Kricker A. 2016. Occupation and risk of non-Hodgkin lymphoma and its subtypes: a pooled analysis from the InterLymph Consortium. Environ Health Perspect 124:396–405; http://dx.doi.org/10.1289/ehp.1409294


Introduction
Non-Hodgkin lymphoma (NHL) comprises a group of malignancies that are common in industrialized countries. Studies of occupational risk factors have proven valuable for generating hypotheses regarding the possible environmental causes of NHL, and over the past four decades, these studies have produced a number of strong leads (Schottenfeld and Fraumeni 2006). In particular, occupations involving exposure to pesticides and solvents have been repeatedly associated with NHL. Other occupational risk factors have been hypothesized; these include infectious agents, sunlight, organic dusts (including flour dust, textile dust, and wood dust), mineral dusts, metals, and ionizing radiation. Nevertheless, even repeatedly observed associations (e.g., employment as farmer) have not been entirely consistent across studies. A well-defined set of occupations and potential exposures relevant to NHL etiology has yet to be established.
Among the potential reasons for the lack of consistency in previous findings is the idea that individual case-control studies lack the power to provide stable estimates of relative risk for less-common occupations and are susceptible to chance findings because of the large number of occupations evaluated. Studies differ somewhat in how occupational details are recorded, coded, analyzed, *These authors contributed equally to this work. and reported, making comparison difficult, and they may not be comparable in terms of the NHL subtypes included and tumor classifications used. Finally, there may be true differences in risk associated with the same occupation across different study regions owing to local differences in population characteristics, exposure patterns, and NHL subtype distribution.
To determine the extent of agreement with previous findings in the large pooled dataset of InterLymph consortium studies, we conducted an analysis of occupations in relation to NHL using a uniform classification of occupations and NHL pathology. Our aims were a) to confirm the relationship of occupations of a priori interest to NHL and its subtypes, and b) to estimate the contribution of specific occupations of a priori interest to the incidence of NHL and its subtypes.

Study population. Included in our analyses
were 10 NHL case-control studies that participate in the InterLymph consortium, had collected information on occupation from cases and controls, and were willing to contribute their data to the pooled analysis (see Table 1 for the acronyms used to refer to each study, details about study designs and locations, and citations to general references for each study). The InterLymph consortium of international investigators undertakes research projects to pool data across studies that explore the etiology of lymphoid malignancies. The set of harmonized core variables, including age, sex, study center (region), smoking status, and NHL subtype, was directly obtained from the InterLymph data coordinating center. Variables on occupational history were obtained from the principal investigators of each participating study. We applied the lymphoma classification scheme for epidemiologic research developed by InterLymph investigators (Morton et al. 2007) to all participating InterLymph studies. All cases classified as "lymphoid neoplasms" according to this classification, except multiple myeloma and Hodgkin lymphoma, were included in this analysis.
Occupational history. For the purpose of our pooled analyses, the data on occupation were classified into a standard internationally recognized occupational classification scheme, the International Standard Classification of Occupations 1968 (ISCO-68) [International Labour Office (ILO) 1981]. Depending on the original occupational classification used by the individual studies and on whether a full-text description of the occupation was available, the ISCO-68 code for each job recorded was determined by one of the following methods: a) a direct conversion of the original classification to the ISCO-68 classification (for the Yale and UCSF1 studies); b) a direct conversion from the original classification to the ISCO-68 classification followed by checking the correctness of each ISCO-68 code by comparing it with the free-text information on the occupation (for the NCI-SEER study); c) using the free-text information on the occupation to individually assign the ISCO-68 code (for the BC, Nebraska, UK, and NSW studies); or d) directly using the original occupational codes for those studies that used ISCO-68 as their original classification (for the Epilymph, Italy, and ENGELA studies). Eight of the 10 studies collected the full occupational history of cases and controls including all occupations held for at least 1 year and starting and ending years, and 2 studies (Nebraska, BC) recorded only the longest-held occupation.
We defined occupational groups of a priori interest for NHL based on the peer-reviewed  (Table 2). After discussions among three of the authors (A.'tM., A.J.D., R.V.), 25 occupational groups were constructed that included jobs associated with NHL in previous studies other than the 10 case-control studies included in our pooled analysis.
We also studied occupations within a group separately up to the detail of the 5-digit ISCO-68 code to explore whether an association was restricted to specific occupations within the group. For example, crop farmers were studied as a group, and specific occupations within this group such as orchard farmers and rice farmers were also studied separately.
Statistical analyses. Unconditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for Table 2. Occupational groups of a priori interest.
Criteria for presentation of results. The present analysis involved many specific occupations within the 25 a priori groups for which previous research demonstrated an association with an increased relative risk of NHL: 925 of > 2,000 relevant codes in the ISCO-68 classification were involved in this analysis. We set criteria to determine which associations to include in the results. We present results for ever employment and for > 10 years employment for all NHL and each of the four subtypes for each occupational group of a priori interest regardless of whether the estimates were statistically significant, with the exception of occupational groups with < 10 cases or < 10 controls. One occupational group in the analyses of all NHL (undertakers) and two groups in the analyses of the four subtypes (pulp & paper workers and petroleum workers) were excluded from the results because they had < 10 cases or < 10 controls. Additionally, we report associations with specific occupational titles included within the occupational groups of interest if we estimated a statistically significant OR (> 1.10 or < 0.90, for ever employment or > 10 years employment) based on men and women combined for all NHL or for any one of the four subtypes.
ORs were also calculated for the 1,286 occupations that were not included in the 25 groups of a priori interest. These results are not presented here but are available upon request.

Results
The 10 case-control studies included 10,046 cases and 12,025 controls (Table 1). Of the cases, 50% were from Europe, 43% were from North America, and 7% were from Australia. The year of diagnosis ranged from 1988 to 2004, and 52.4% of cases were male. The mean ± standard deviation (SD) age at interview was 57.6 ± 12.8 years for cases and 55.4 ± 14.2 years for controls. The mean year of first employment (in the 8 studies with full occupational history) was 1959 (± 16 years; range, 1915-2003) for cases and 1961 (± 16 years; range, 1912-2002) for controls. Of the four subtypes selected for separate analyses, DLBCL formed the largest group with 3,061 cases (52.4% male), followed by FL (2,140 cases; 45.6% male), CLL/SLL (1,014 cases; 59.3% male), and PTCL (632 cases; 56.5% male).
Evidence of heterogeneity in relative risks (p < 0.05, Q-test for heterogeneity) across the four NHL subtypes was present for "women's hairdressers," metal workers, "printing pressmen," textile workers, and "cabinetmakers" (Table 4). "Printing pressmen," however, had very small numbers of cases and controls (< 10) for all analyses except for CLL/SLL. Attributable fraction. We estimated the proportion of NHL and of each subtype that Abbreviations: -, < 10 cases or < 10 controls; NA, 0 cases or controls; a Results are not presented for the undertakers occupational group because they included < 10 cases or < 10 controls. b Results are presented for a specific occupational title within an occupational group if there was a statistically significantly increased or decreased risk of NHL associated with ever or > 10 years employment for men and women combined; results are excluded when there were < 10 cases or < 10 controls. c Adjusted for age, sex, and study center.
volume 124 | number 4 | April 2016 • Environmental Health Perspectives was attributable to the main occupational groups (farmers, textile workers, hairdressers, wood workers, painters) or to specific occupations (e.g., "women's hairdressers," "spraypainters") for which an elevated relative risk had been observed (p < 0.05). AFs for NHL were low, between 0.3% for "women's hairdressers" and 0.63% for "general farm workers," and were somewhat higher for the rarer individual subtypes: 1.49% for "women's hairdressers" and CLL/SLL and ≥ 3.69% for the textile worker group and PTCL. AFs differed by sex in a number of occupations, reflecting the scarcity of men or women in a particular occupation.

Discussion
We found evidence that NHL was associated with employment as textile workers, hairdressers, and farm workers, as well as with employment as painters, printers, wood workers, metal workers, medical workers, electrical workers, and cleaners. The statistically Abbreviations: -, < 10 cases or < 10 controls; NA, 0 cases or controls. a Results are not presented for the following occupational groups because they included < 10 cases or < 10 controls: Pulp & paper workers, petroleum workers, and undertakers. b Results are presented for a specific occupational title within an occupational group if there was a statistically significantly increased or decreased OR for at least one subtype associated with ever employment or > 10 years employment for men and women combined; results are excluded when there were < 10 cases or < 10 controls. c Adjusted for age, sex, and study center. d Q-test for heterogeneity across the four subtypes, based on ORs for ever employment in the occupation for men and women combined.
significant heterogeneity in relative risk estimates among subtypes suggested that employment as "women's hairdressers" was particularly associated with DLBCL and CLL/SLL and employment as textile workers with DLBCL and PTCL. Our pooled analysis used a uniform classification of NHL diagnosis and was substantially larger than any individual study. A limitation of our study is that grouping workers according to job title disregards the wide qualitative and quantitative variation in exposure that may occur for workers with the same job title (McGuire et al. 1998). Even if an association between job title and disease is found, the potentially causative agents are unknown, although they are likely to be common rather than rare exposures within the occupational group. The international nature of this study also implies that only associations for occupations with internationally comparable exposure profiles can be detected and that some misclassification will be introduced owing to the recoding of different occupational classifications into a single one. An advantage of using job titles rather than specific exposures is that recall by participants is less likely to be influenced by their disease status, making differential misclassification also less likely. The multiple comparisons of a job title-based approach, however, suggest a vulnerability to false positive findings. Results are therefore focused on the a priori-selected occupational groups (24 were eligible) extracted from earlier NHL research. We discuss below the findings from our study that are consistent with previously reported associations, and we also discuss occupational exposures that might be implicated as etiologic agents.
We confirmed the previously reported association of NHL with crop farming occupations (Blair et al. 1992;Keller-Byrne et al. 1997), but not with animal farming (Amadori et al. 1995;Boffetta and de Vocht 2007;Lee et al. 2002), which was negatively associated with CLL/SLL. This finding suggests that risk estimates for all farming and all NHL combined may be uninformative and that future studies will need to consider both NHL subtype and farming type to identify the possible specific farming exposures that may be involved in these associations.
The observed associations for hairdressers were stronger for women's hairdressers than for other job titles within this occupational category, supporting a hypothesis of hair dye or other hair treatments more commonly used by women than by men or children as possible causes. Associations were present for DLBCL and CLL/SLL but were absent for FL. A previous pooled analysis of InterLymph studies reported associations with personal hair dye use for NHL subtypes FL and CLL/SLL (Zhang et al. 2008). Exposure from personal hair dye use is, however, not strictly comparable to the exposure experienced by hairdressers because hairdressers are exposed on a daily basis to a range of other compounds such as solvents and propellant gases, including dichloromethane and chlorofluorocarbons.
The observed associations between textilerelated occupations and NHL (DLBCL and PTCL) suggest a range of possible exposures that can occur in this environment, but the implication of multiple specific occupations within this group, which is involved in both fabric making and garment making, indicates that associations were not restricted to specific tasks in the textile industry (e.g., textile dyeing) but rather may be associated with more ubiquitous exposures (Siemiatycki et al. 1986).
We found associations with NHL for a number of other occupations potentially exposed to solvents. Among these occupations were cleaners, painters (especially spray-painters) with potential for exposure to solvents in paints and paint strippers, and machine tool operators, who may be exposed to a range of solvents including aliphatic hydrocarbon solvents, aromatic hydrocarbon solvents, chlorinated solvents, mineral oils, and diesel fuel and exhaust. Metal workers would also be exposed to metal dust and metal-working fluids. Although our findings of positive associations for these occupations may support a role for solvent exposure as a risk factor for NHL, other exposures may also be responsible.
Some solvent exposure would likely be implicated in two other occupational groups for which we observed an association with NHL: several specific occupations within the electrical and electronics-related group may also have exposure to electromagnetic fields (Mester et al. 2006), and carpenters may be exposed to wood dust, wood preservatives, formaldehyde, and molds in addition to solvents. Forestry workers could also be exposed to wood dust and potentially to pesticides and engine exhausts.
All teaching occupations combined were inversely associated with NHL, a finding that is the opposite of the results of a death certificatebased case-control study (Figgs et al. 1995) and a meta-analysis (Baker et al. 1999). We did observe a marked positive association for preprimary education teachers with CLL/SLL, which could point towards common childhood infections as a possible causal factor (Vineis et al. 2000). Long-term employment as a medical doctor, in which infectious agents may also play a role, was associated with FL.
Among the four NHL subtypes, the statistically significant heterogeneity in relative risk estimates suggested that "women's hairdressers" were at an increased risk for DLBCL and CLL/SLL, but not for FL, which was previously suggested to be associated with personal hair dye use (Zhang et al. 2008;Sangrajrang et al. 2011). Textile workers were another occupational group to show heterogeneity across NHL subtypes and were particularly at risk for DLBCL and PTCL. There was no significant heterogeneity in ORs for crop and mixed/unspecified farming among the four subtypes, although DLBCL and CLL/SLL appeared to be most strongly associated with farming occupations. We note the strong association between spray-painters and FL as well as the lack of adequate numbers for analysis in the other subtypes examined. The authors of a recent major analysis of NHL subtypes and a broad range of risk factors in the InterLymph consortium reported that certain occupations were associated with one or more subtypes, including spray-painters (FL), crop farmers (DLBCL, CLL/SLL), hairdressers (DLBCL, CLL/SLL), and medical doctors (FL). These analyses were adjusted for all other significant risk factors (Morton et al. 2014) and are consistent with our findings. However, our analysis based on occupational titles is not the proper setting in which to explore whether socioeconomic confounders for which we were unable to control might have generated some of our positive findings; such hypotheses need to be specifically addressed in dedicated analyses.

Conclusions
This pooled analysis supports a role for textile-, hairdressing-, and farming-related exposures in the development of NHL. Additional occupations associated with NHL or NHL subtypes include cleaners, painters, printers, and wood workers. The results by sex indicate that occupational exposures may play a role in NHL for both women and men, but the specific occupations involved differ between the sexes. The large numbers of participants and the application of standard NHL and occupational classification systems allowed us to make estimates of relative risk by NHL subtype, forming an important step towards improving our understanding of NHL etiology. The findings of the present study can be further refined at the next stage, after specific exposures are identified in detailed exposure studies.