Abstract

Little is known about the risk of non-Hodgkin lymphoma (NHL) in nonsmokers who are exposed to environmental tobacco smoke (ETS). Previous research on NHL and ETS has not included men or examined doses of ETS exposure during childhood. The Cancer Prevention Study II Nutrition Cohort collected information on smoking habits and exposure to ETS during childhood and adulthood. Among 61,326 never-smoking men and women, 884 incident cases of NHL were identified between 1992 and 2009. Multivariable-adjusted relative risks and 95% confidence intervals were calculated using Cox proportional hazards regression to identify associations between ETS and NHL risk. Compared with no exposure to ETS as a child or an adult, childhood and/or adult ETS exposure was not associated with NHL overall. There was a positive association between the number of smokers in the house as a child (P for trend = 0.05) and exposure to 6 or more hours per week of ETS as an adult (relative risk = 2.37, 95% confidence interval: 1.12, 5.04) with follicular lymphoma risk. Adult ETS exposure was associated with a lower risk of diffuse large B-cell lymphoma (relative risk = 0.68, 95% confidence interval: 0.48, 0.97). This study suggests that adult and childhood ETS exposure may affect the risk of NHL, and that the associations differ by histological subtype.

Tobacco smoke is a known human carcinogen for at least 15 different cancer sites (1). Environmental tobacco smoke (ETS), or secondhand smoke, consists of both the smoke exhaled by smokers, as well as the sidestream smoke produced by the burning cigarette. This type of smoke contains the same carcinogenic compounds inhaled during active smoking and is an established risk factor for lung cancer (2).

Although active smoking has not been consistently associated with the risk of non-Hodgkin lymphoma (NHL) overall, there is growing evidence of an association with risk of the follicular lymphoma subtype, particularly in women (3–9). Recently, a large prospective study of never-smoking female California teachers showed a 50% higher risk of NHL and a nearly 3-fold higher risk of follicular lymphoma among women exposed to ETS for 40 or more years compared with those exposed for 5 years or less (10). However, no research on ETS and the risk of NHL has included men or examined dose-response associations of ETS exposure in childhood. Further research is needed to clarify possible associations between ETS and the risk of NHL subtypes in studies that include men.

The American Cancer Society's Cancer Prevention Study II (CPS-II) Nutrition Cohort is a large nationwide prospective study of both men and women that includes more than 60,000 never smokers. Information on intensity of adult ETS exposure, as well as intensity and duration of childhood ETS exposure was collected from participants in this cohort. This information was used to examine the associations of childhood and adult exposure to ETS with NHL and its subtypes.

METHODS

Study population

Subjects in this analysis were selected from the CPS-II Nutrition Cohort, a prospective study of cancer incidence and death in 184,187 men and women from the United States, described in detail elsewhere (11). Briefly, the Nutrition Cohort is a subcohort of the approximately 1.2 million subjects in the CPS-II, a prospective mortality study established by the American Cancer Society in 1982. Participants in the larger study were recruited nationally and completed a 4-page questionnaire at enrollment. Subjects, who were primarily between the ages of 50 and 74 years in 1992, were recruited from 21 states with population-based state cancer registries to participate in the Nutrition Cohort. Participants completed a 10-page mailed questionnaire that included information on demographic, medical, behavioral, environmental, occupational, and dietary factors. Follow-up questionnaires were sent to cohort members every 2 years beginning in 1997 to ascertain cancer diagnoses. Responses to all surveys were received from at least 87% of living participants after multiple mailings. The current study encompasses follow-up through June 2009. All aspects of the CPS-II Nutrition Cohort Study have been approved by the Emory University institutional review board (Atlanta, Georgia).

These analyses excluded subjects from the CPS-II Nutrition Cohort who were lost to follow-up (n = 6,251); reported a personal history of cancer other than nonmelanoma skin cancer at baseline in 1992 (n = 22,863); reported ever smoking 100 or more cigarettes or reported ever smoking pipes or cigars (n = 92,769); were missing information on childhood or adult exposure to ETS in 1992 (n = 953); or reported a diagnosis of a hematopoietic cancer in the first survey interval that could not be verified (n = 25). The final analytical cohort included 61,326 men and women who had never smoked.

Measures of childhood ETS exposure

In 1992, subjects were asked, “During the period from birth to age 18, did you ever live for more than 1 year with someone who smoked on a daily basis? (no, yes).” Subjects who answered “yes” to this question were asked “For how many years? (1–3, 4–6, 7–9, 10–12, 13–15, or 16–18 years),” and “Was this usually … (One, two, or more than two persons).” To be included in the analytical cohort, subjects were required to have answered at least 1 of the 3 questions. Subjects who answered “yes” to the first question or who provided a valid response to the number of years or number of persons were considered exposed to ETS in childhood.

Measures of adult ETS exposure

In 1982, participants were asked, “Whether or not you smoke, on the average, how many hours a day are you exposed to cigarette smoke of others?” followed by write-in responses for “at home,” “at work,” and “in other areas.” Subjects with missing values were assumed to be unexposed, because there were no instructions to enter “0” if not exposed. Values from the 3 location responses were summed to determine total hours per day of ETS exposure in 1982, and subjects who had values greater than 0 hours per day of exposure were considered to have past adult ETS exposure.

In 1992, participants were asked the question, “Whether or not you smoke, about how many hours per week are you currently exposed to the smoke of others …,” followed by 2-digit write-in answers for “at home (hours/week)” and “in other places (hours/week)” and were asked to enter “0” if unexposed. To be included in the analytical cohort, subjects were required to have answered at least 1 of the questions with a value of 0 or greater. Values from the 2 responses were summed to determine total hours per week of ETS exposure in 1992, and subjects who had values greater than 0 hours per week of exposure were considered currently exposed to ETS as an adult.

Case ascertainment

This analysis includes 884 subjects with NHL diagnosed between the date of enrollment and June 30, 2009. The majority of cases (n = 668) were identified by self-report of NHL on follow-up surveys and subsequently verified by medical record abstraction or linkage with state cancer registries. There were 15 cases identified through self-report of another cancer, which were subsequently verified as NHL during medical record abstraction or linkage with state cancer registries. An additional 201 cases were identified as NHL through automated linkage of the entire cohort with the National Death Index; 72% of these were subsequently verified by linkage with the state cancer registries.

Histological subtypes were defined using the InterLymph Pathology Working Group guidelines (12), based on the 2008 revised World Health Organization classification of tumors of hematopoietic and lymphoid tissues (13). Using the International Classification of Disease for Oncology, Second Edition and Third Edition, the subtypes were classified as 22% diffuse large B-cell lymphoma (DLBCL), 22% chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), 12% follicular lymphoma, 17% multiple myeloma, 20% other B/T-cell NHL, and 7% unknown subtype.

Statistical analyses

Childhood, past adult, and current adult ETS measurements were combined to create a measure of any ETS exposure (no, yes), and a mutually exclusive childhood or adult ETS exposure variable (no exposure, childhood only, adult only, childhood and adult). Subjects reporting no childhood or adult ETS exposure were used as the reference group in all analyses. Childhood ETS exposure was further examined as any childhood exposure (no, yes), years living with a smoker (<10, 10–15, 16–18 years), and number of smokers in the household (1, ≥2). Adult ETS exposure was examined as any adult exposure (no, yes), status (past, current), location of exposure including household (no, yes) or other location (no, yes), and as tertiles of the total hours of exposure/week (1, 2–5, ≥6 hours/week). Descriptive statistics were calculated using distributions of categorical variables and means of continuous variables.

Person-years of follow-up for each participant were calculated as the amount of time from completion of the CPS-II Nutrition Cohort questionnaire in 1992/1993 to the date of 1) diagnosis of NHL; 2) a questionnaire reporting cancer other than NHL; 3) death occurring between the last returned survey and the next mailed survey; 4) return of the last questionnaire; 5) the last questionnaire on which the participant was known to be cancer free if they reported a hematopoietic cancer that could not be verified; or 6) the first questionnaire on which a participant reported a history of active smoking. Cox proportional hazards regression (14) was used to compute multivariable adjusted rate ratios and 95% confidence intervals for the association between ETS variables and NHL, as well as NHL subtype incidence. All models were stratified on age and additionally adjusted for sex (male, female), race (white, other), education (high school or less, some college, college graduate), family history of hematopoietic cancer (no, yes), body mass index (weight (kg)/height (m)2) (<25.00, 25.00–29.9, ≥30.00), height (sex-specific quintiles), physical activity (0, 0.1–17.4, ≥17.5 metabolic equivalents of energy in hours/week), alcohol use (nondrinker or <1 or ≥1 drinks/day, or missing data), hormone replacement therapy (no use, estrogen only, combined estrogen and progesterone only, mixed use/other use/missing data), and nonsteroidal antiinflammatory drug use (never, former, current, missing data). Models adjusted for age and sex were similar to multivariable models. Statistical tests of linear trend were calculated using a continuous ordinal variable indicating a 1-unit change per category with and without the referent group of no exposure. All P values are 2-sided.

Several sensitivity analyses were conducted. To reduce confounding by adult ETS exposure, we ran models of childhood ETS exposure excluding subjects with any current adult exposure to ETS. Similarly, models of adult ETS exposure were run excluding subjects reporting any childhood ETS exposure. To evaluate the effect of missing ETS data in 1982, we ran models incorporating only 1992 ETS data. Models were also run separately by sex. All analyses were performed using SAS, version 9.2, software (SAS Institute, Inc., Cary, North Carolina).

RESULTS

In this study of never smoking men and women, only 16% of subjects reported no exposure to ETS as children or adults, whereas 53% lived with at least 1 smoker as children, and 70% reported either past or current adult exposure to ETS (Table 1). Subjects reporting no exposure to ETS were slightly older, less obese, drank less alcohol, and were less likely to use nonsteroidal antiinflammatory drugs than subjects reporting any exposure to ETS. Those exposed to ETS both as children and as adults were younger, more obese, drank the most alcohol, and were most likely to use nonsteroidal antiinflammatory drugs than other subjects. Subjects with any ETS exposure in childhood had the least education. Subjects whose only ETS exposure was in childhood were more likely to be women, whereas subjects exposed only as adults included a higher proportion of men. Other risk factors were generally similarly distributed among ETS exposure groups. Among subjects who lived with smokers as children, most reported living with only 1 smoker and lived with a smoker during the duration of childhood (for 16–18 years). Approximately half of the subjects exposed to ETS in adulthood had current exposure at baseline, and the other half reported past adult exposure only. Subjects exposed to ETS as adults had more current hours per week of exposure and were more likely to be exposed at home if they reported exposure in childhood as well.

Table 1.

Baseline Characteristics of Nonsmoking Men and Women by Environmental Tobacco Smoke Exposure in the Cancer Prevention Study II Nutrition Cohort, 1992–2009

Characteristic ETS Exposure, %
 
No Exposure (n = 9,971) Childhood Only (n = 8,308) Adult Only (n = 19,056) Childhood and Adult (n = 23,991) 
Age at baseline, years 64.3 (6.78)a 63.3 (6.7)a 63 (6.4)a 61.8 (6.31)a 
Sex     
 Male 29.8 22.4 37.3 29.3 
 Female 70.2 77.6 62.7 70.7 
Race     
 White 96.9 97.8 96.6 97.6 
 Other 3.1 2.2 3.4 2.4 
Education     
 High school or less 34.0 36.9 33.5 37.3 
 Some college 26.0 26.9 26.8 27.6 
 College graduate 40.0 36.3 39.7 35.1 
Body mass indexb     
 <25.00 52.2 52.2 46.0 44.9 
 25.00–29.99 35.7 35.2 39.1 38.3 
 ≥30 12.1 12.6 14.9 16.7 
Physical activity, METS/week     
 0 9.8 9.7 8.6 8.6 
 0.1–17.4 65.8 64.9 64.9 64.2 
 ≥17.5 24.4 25.4 26.6 27.1 
Alcohol intake, drinks/day     
 Nondrinker 62.7 53.6 53.2 46.7 
 <1 26.7 33.2 34.4 39.1 
 ≥1 6.3 9.3 8.6 11.3 
Family history of hematopoietic cancer 3.7 3.5 3.7 3.4 
NSAID use     
 Never 22.2 19.7 18.9 17.0 
 Former 24.3 25.1 24.8 24.7 
 Current 48.0 50.4 51.1 53.2 
HRT use (in women)     
 Never 44.4 41.1 46.0 44.4 
 ERT only 33.8 35.7 32.9 33.9 
 CHRT only 10.3 11.1 10.0 10.6 
 Mixed/other 11.5 12.0 11.1 11.1 
Childhood smoke exposure, years     
 <10  10.7  9.9 
 10–15  16.0  15.5 
 16–18  71.2  72.8 
Smokers in house as child     
 1 Smoker in house  84.2  81.1 
 ≥2 Smokers in house  14.5  17.7 
Adult ETS status     
 Past only   51.6 44.5 
 Current   48.4 55.5 
Current ETS exposure in the home   7.6 12.5 
Current ETS exposure outside the home   44.6 49.5 
Current ETS exposure, hours/week 0 (0)a 0 (0)a 4.1 (12.09)a 5.7 (14.77)a 
Characteristic ETS Exposure, %
 
No Exposure (n = 9,971) Childhood Only (n = 8,308) Adult Only (n = 19,056) Childhood and Adult (n = 23,991) 
Age at baseline, years 64.3 (6.78)a 63.3 (6.7)a 63 (6.4)a 61.8 (6.31)a 
Sex     
 Male 29.8 22.4 37.3 29.3 
 Female 70.2 77.6 62.7 70.7 
Race     
 White 96.9 97.8 96.6 97.6 
 Other 3.1 2.2 3.4 2.4 
Education     
 High school or less 34.0 36.9 33.5 37.3 
 Some college 26.0 26.9 26.8 27.6 
 College graduate 40.0 36.3 39.7 35.1 
Body mass indexb     
 <25.00 52.2 52.2 46.0 44.9 
 25.00–29.99 35.7 35.2 39.1 38.3 
 ≥30 12.1 12.6 14.9 16.7 
Physical activity, METS/week     
 0 9.8 9.7 8.6 8.6 
 0.1–17.4 65.8 64.9 64.9 64.2 
 ≥17.5 24.4 25.4 26.6 27.1 
Alcohol intake, drinks/day     
 Nondrinker 62.7 53.6 53.2 46.7 
 <1 26.7 33.2 34.4 39.1 
 ≥1 6.3 9.3 8.6 11.3 
Family history of hematopoietic cancer 3.7 3.5 3.7 3.4 
NSAID use     
 Never 22.2 19.7 18.9 17.0 
 Former 24.3 25.1 24.8 24.7 
 Current 48.0 50.4 51.1 53.2 
HRT use (in women)     
 Never 44.4 41.1 46.0 44.4 
 ERT only 33.8 35.7 32.9 33.9 
 CHRT only 10.3 11.1 10.0 10.6 
 Mixed/other 11.5 12.0 11.1 11.1 
Childhood smoke exposure, years     
 <10  10.7  9.9 
 10–15  16.0  15.5 
 16–18  71.2  72.8 
Smokers in house as child     
 1 Smoker in house  84.2  81.1 
 ≥2 Smokers in house  14.5  17.7 
Adult ETS status     
 Past only   51.6 44.5 
 Current   48.4 55.5 
Current ETS exposure in the home   7.6 12.5 
Current ETS exposure outside the home   44.6 49.5 
Current ETS exposure, hours/week 0 (0)a 0 (0)a 4.1 (12.09)a 5.7 (14.77)a 

Abbreviations: CHRT, combined estrogen and progesterone hormone replacement therapy; ERT, estrogen replacement therapy; ETS, environmental tobacco smoke; HRT, hormone replacement therapy; METS, metabolic equivalents of energy; NSAID, nonsteroidal antiinflammatory drug.

a Value expressed as mean (standard deviation).

b Calculated as weight (kg)/height (m)2.

There was no association between any of the ETS exposure variables and the risk of NHL overall (Table 2). However, any ETS exposure and exposure in both childhood and adulthood were associated with a higher risk of follicular lymphoma, although not significantly. Conversely, any ETS exposure and ETS exposure only in adulthood were significantly inversely associated with the risk of DLBCL. ETS exposure was not associated with the risk of CLL/SLL or multiple myeloma. Results were similar by sex (Web Table 1 available at http://aje.oxfordjournals.org/).

Table 2.

Relative Risk of NHL by Exposure to Environmental Tobacco Smoke Among Never Smokers in the Cancer Prevention Study II Nutrition Cohort, 1992–2009

ETS Exposure Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any exposure 682,055 732 0.98 0.82, 1.17 153 0.69 0.49, 0.98 151 0.84 0.59, 1.20 97 1.70 0.93, 3.11 131 1.14 0.72, 1.80 
 Childhood only 110,464 121 1.01 0.80, 1.29 26 0.76 0.46, 1.24 24 0.82 0.49, 1.36 16 1.71 0.80, 3.62 23 1.30 0.72, 2.33 
 Adult only 250,106 293 1.00 0.82, 1.22 56 0.66 0.44, 0.98 64 0.90 0.60, 1.34 34 1.56 0.81, 3.03 57 1.27 0.77, 2.09 
 Childhood and adult 321,485 318 0.94 0.77, 1.14 71 0.70 0.48, 1.04 63 0.79 0.52, 1.19 47 1.83 0.96, 3.47 51 0.96 0.58, 1.60 
ETS Exposure Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any exposure 682,055 732 0.98 0.82, 1.17 153 0.69 0.49, 0.98 151 0.84 0.59, 1.20 97 1.70 0.93, 3.11 131 1.14 0.72, 1.80 
 Childhood only 110,464 121 1.01 0.80, 1.29 26 0.76 0.46, 1.24 24 0.82 0.49, 1.36 16 1.71 0.80, 3.62 23 1.30 0.72, 2.33 
 Adult only 250,106 293 1.00 0.82, 1.22 56 0.66 0.44, 0.98 64 0.90 0.60, 1.34 34 1.56 0.81, 3.03 57 1.27 0.77, 2.09 
 Childhood and adult 321,485 318 0.94 0.77, 1.14 71 0.70 0.48, 1.04 63 0.79 0.52, 1.19 47 1.83 0.96, 3.47 51 0.96 0.58, 1.60 

Abbreviations: CI, confidence interval; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; DLBCL, diffuse large B-cell lymphoma; ETS, environmental tobacco smoke; NHL, non-Hodgkin lymphoma; RR, relative risk.

a Models are adjusted for age, sex, race, education, family history of hematopoietic cancer, height, body mass index (weight (kg)/height (m)2), physical activity, alcohol use, nonsteroidal antiinflammatory use, and hormone replacement therapy use.

Results of analyses for childhood ETS exposure in relation to NHL overall and NHL subtypes are shown in Table 3. Any childhood exposure was associated with a borderline higher risk of follicular lymphoma with evidence of a positive trend for the number of smokers in the childhood household (P for trend with reference = 0.05; P for trend without reference = 0.69). Conversely, for any childhood ETS exposure compared with no childhood or adult exposure, there was a nonsignificant lower risk of DLBCL (relative risk = 0.72, 95% confidence interval: 0.50, 1.04). There was a significantly lower risk of DLBCL among subjects living with a smoker for 16–18 years during childhood compared with those with no ETS exposure. In sensitivity analyses excluding subjects with any adult ETS exposure (not shown), the relative risks were similar to those shown in Table 3. There were no associations of childhood ETS exposure with the risk of NHL overall, CLL/SLL, or multiple myeloma.

Table 3.

Relative Risk of NHL by Childhood Exposure to Environmental Tobacco Smoke Among Never Smokers in the Cancer Prevention Study-II Nutrition Cohort, 1992–2009

ETS Exposure No. of Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any childhood exposure 431,948 439 0.96 0.79, 1.16 97 0.72 0.50, 1.04 87 0.80 0.54, 1.17 63 1.79 0.96, 3.35 74 1.05 0.65, 1.70 
Years living with a smoker as a child                 
 <10 43,589 49 1.05 0.76, 1.45 0.67 0.32, 1.38 10 0.88 0.44, 1.76 2.48 1.04, 5.91 1.26 0.58, 2.75 
 10–15 68,219 71 0.97 0.73, 1.29 24 1.13 0.68, 1.87 12 0.69 0.36, 1.32 1.60 0.67, 3.81 11 0.96 0.46, 1.98 
 16–18 312,670 311 0.94 0.77, 1.15 63 0.65 0.43, 0.96 64 0.82 0.54, 1.23 42 1.65 0.86, 3.17 54 1.07 0.64, 1.76 
  P for trendb   0.50  0.06  0.24  0.21  0.81 
  P for trendc   0.58  0.34  0.86  0.41  0.82 
No. of smokers in the house as a child                 
 1 352,784 368 0.97 0.80, 1.17 80 0.72 0.49, 1.05 73 0.80 0.54, 1.19 52 1.78 0.95, 3.36 62 1.06 0.65, 1.73 
 ≥2 73,925 66 0.91 0.68, 1.23 15 0.70 0.39, 1.28 14 0.82 0.44, 1.53 11 2.00 0.87, 4.61 11 0.99 0.47, 2.06 
  P for trendb   0.56  0.16  0.22  0.05  0.82 
  P for trendc   0.53  0.82  0.82  0.69  0.85 
ETS Exposure No. of Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any childhood exposure 431,948 439 0.96 0.79, 1.16 97 0.72 0.50, 1.04 87 0.80 0.54, 1.17 63 1.79 0.96, 3.35 74 1.05 0.65, 1.70 
Years living with a smoker as a child                 
 <10 43,589 49 1.05 0.76, 1.45 0.67 0.32, 1.38 10 0.88 0.44, 1.76 2.48 1.04, 5.91 1.26 0.58, 2.75 
 10–15 68,219 71 0.97 0.73, 1.29 24 1.13 0.68, 1.87 12 0.69 0.36, 1.32 1.60 0.67, 3.81 11 0.96 0.46, 1.98 
 16–18 312,670 311 0.94 0.77, 1.15 63 0.65 0.43, 0.96 64 0.82 0.54, 1.23 42 1.65 0.86, 3.17 54 1.07 0.64, 1.76 
  P for trendb   0.50  0.06  0.24  0.21  0.81 
  P for trendc   0.58  0.34  0.86  0.41  0.82 
No. of smokers in the house as a child                 
 1 352,784 368 0.97 0.80, 1.17 80 0.72 0.49, 1.05 73 0.80 0.54, 1.19 52 1.78 0.95, 3.36 62 1.06 0.65, 1.73 
 ≥2 73,925 66 0.91 0.68, 1.23 15 0.70 0.39, 1.28 14 0.82 0.44, 1.53 11 2.00 0.87, 4.61 11 0.99 0.47, 2.06 
  P for trendb   0.56  0.16  0.22  0.05  0.82 
  P for trendc   0.53  0.82  0.82  0.69  0.85 

Abbreviations: CI, confidence interval; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; DLBCL, diffuse large B-cell lymphoma; ETS, environmental tobacco smoke; NHL, non-Hodgkin lymphoma; RR, relative risk.

a Models are adjusted for age, sex, race, education, family history of hematopoietic cancer, height, body mass index (weight (kg)/height (m)2), physical activity, alcohol use, nonsteroidal antiinflammatory use, and hormone replacement therapy use.

bP for trend that includes the reference category.

cP for trend that does not include the reference category.

Results of analyses for adult ETS exposure are shown in Table 4. Current ETS exposure and household exposure were positively associated with the risk of follicular lymphoma. Additionally, there was evidence of a dose-response relationship between hours per week of ETS exposure and risk (P for trend with reference = 0.03; P for trend without reference = 0.26) with nearly 2.4-fold higher risk for 6 or more hours per week. Compared with no ETS exposure, any adult ETS exposure, current exposure, and exposure outside the home were associated with a statistically significantly lower risk of DLBCL. In sensitivity analyses excluding subjects with any childhood ETS exposure (not shown), relative risks were similar to those in Table 4. Exposure to ETS as an adult was not associated with risk of NHL overall, CLL/SLL, or multiple myeloma.

Table 4.

Relative Risk of NHL by Adult Exposure to Environmental Tobacco Smoke Among Never Smokers in the Cancer Prevention Study-II Nutrition Cohort, 1992–2009

ETS Exposure No. of Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any adult exposure 571,591 611 0.97 0.81,1.16 127 0.68 0.48, 0.97 127 0.84 0.58, 1.22 81 1.70 0.92, 3.14 108 1.11 0.70, 1.76 
Adult exposure status                 
 Past only 271,112 296 0.96 0.79, 1.17 70 0.79 0.53, 1.16 64 0.85 0.57, 1.27 35 1.51 0.78, 2.92 56 1.17 0.71, 1.92 
 Current 300,479 315 0.98 0.80, 1.19 57 0.58 0.38, 0.87 63 0.84 0.56, 1.26 46 1.90 1.00, 3.62 52 1.04 0.63, 1.74 
Current adult household exposure 58,749 53 0.91 0.66, 1.25 11 0.60 0.30, 1.17 12 0.89 0.46, 1.72 11 2.37 1.03, 5.43 0.81 0.34, 1.91 
Current adult exposure outside the home 272,014 289 0.99 0.81, 1.21 53 0.59 0.39, 0.90 54 0.79 0.52, 1.20 38 1.73 0.90, 3.35 52 1.14 0.69, 1.90 
Total current adult exposure, hours/ week                 
 1 93,470 88 0.86 0.66, 1.13 19 0.63 0.36, 1.09 16 0.66 0.37, 1.19 13 1.70 0.77, 3.74 14 0.87 0.44, 1.71 
 2–5 116,051 129 1.03 0.81, 1.31 20 0.52 0.30, 0.89 23 0.79 0.47, 1.33 16 1.71 0.80, 3.65 24 1.23 0.69, 2.22 
 ≥6 90,958 98 1.04 0.80, 1.35 18 0.60 0.34, 1.06 24 1.11 0.66, 1.86 17 2.37 1.12, 5.04 14 0.99 0.50, 1.96 
  P for trendb   0.53  0.06  0.95  0.03  0.53 
  P for trendc   0.20  0.78  0.09  0.26  0.61 
ETS Exposure No. of Person-Years Histological Subtype
 
NHL
 
DLBCL
 
CLL/SLL
 
Follicular
 
Multiple Myeloma
 
No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI No. of Cases RRa 95% CI 
No exposure 128,509 152 1.00 Referent 42 1.00 Referent 39 1.00 Referent 12 1.00 Referent 22 1.00 Referent 
Any adult exposure 571,591 611 0.97 0.81,1.16 127 0.68 0.48, 0.97 127 0.84 0.58, 1.22 81 1.70 0.92, 3.14 108 1.11 0.70, 1.76 
Adult exposure status                 
 Past only 271,112 296 0.96 0.79, 1.17 70 0.79 0.53, 1.16 64 0.85 0.57, 1.27 35 1.51 0.78, 2.92 56 1.17 0.71, 1.92 
 Current 300,479 315 0.98 0.80, 1.19 57 0.58 0.38, 0.87 63 0.84 0.56, 1.26 46 1.90 1.00, 3.62 52 1.04 0.63, 1.74 
Current adult household exposure 58,749 53 0.91 0.66, 1.25 11 0.60 0.30, 1.17 12 0.89 0.46, 1.72 11 2.37 1.03, 5.43 0.81 0.34, 1.91 
Current adult exposure outside the home 272,014 289 0.99 0.81, 1.21 53 0.59 0.39, 0.90 54 0.79 0.52, 1.20 38 1.73 0.90, 3.35 52 1.14 0.69, 1.90 
Total current adult exposure, hours/ week                 
 1 93,470 88 0.86 0.66, 1.13 19 0.63 0.36, 1.09 16 0.66 0.37, 1.19 13 1.70 0.77, 3.74 14 0.87 0.44, 1.71 
 2–5 116,051 129 1.03 0.81, 1.31 20 0.52 0.30, 0.89 23 0.79 0.47, 1.33 16 1.71 0.80, 3.65 24 1.23 0.69, 2.22 
 ≥6 90,958 98 1.04 0.80, 1.35 18 0.60 0.34, 1.06 24 1.11 0.66, 1.86 17 2.37 1.12, 5.04 14 0.99 0.50, 1.96 
  P for trendb   0.53  0.06  0.95  0.03  0.53 
  P for trendc   0.20  0.78  0.09  0.26  0.61 

Abbreviations: CI, confidence interval; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; DLBCL, diffuse large B-cell lymphoma; ETS, environmental tobacco smoke; NHL, non-Hodgkin lymphoma; RR, relative risk.

a Models are adjusted for age, sex, race, education, family history of hematopoietic cancer, height, body mass index (weight (kg)/height (m)2), physical activity, alcohol use, nonsteroidal antiinflammatory use, and hormone replacement therapy use.

bP for trend that includes the reference category.

cP for trend that does not include the reference category.

DISCUSSION

This large prospective study provided the opportunity to examine exposure to both childhood and adult ETS in relation to risk of NHL overall and NHL subtypes among never smokers. The evidence for a positive association between ETS exposure and risk of follicular lymphoma was supported by a statistically significant linear trend for the number of smokers in the house during childhood, as well as for the number of hours per week of current ETS exposure during adulthood. For DLBCL, current ETS exposure was associated with lower risk, although there were no dose-response relationships with the intensity of childhood exposure or the intensity of current adult exposure. NHL overall, CLL/SLL, and multiple myeloma were not associated with ETS, nor were there differences by sex.

In the California Teachers Study (CTS) (10), the only other study to examine ETS and NHL subtypes, positive associations of childhood and adult ETS exposure with risk of follicular lymphoma were similar to those found in our study. The linear dose-response relationships for both greater intensity and duration of ETS exposure were also similar in both studies. It is notable that the intensity of exposure was measured qualitatively in the CTS (a little, fairly, very smoky) compared with the quantitative measure used in the CPS-II (hours/week), yet the results remained consistent. The similar results using 2 different measures of exposure provide support for the validity of the observed dose-response associations with follicular lymphoma. The greater proportion of follicular cases and stronger ETS and follicular lymphoma association in the CTS are likely driving the overall NHL association in the CTS, which was not observed in the CPS-II. The association between ETS and follicular lymphoma observed in the CTS and in our study is supported by accumulating evidence of an association with active cigarette smoking in women (5, 7, 10, 15), although results from studies of active smoking and the risk of follicular lymphoma in men and women combined are inconsistent (4, 16–18). In this study, data were limited to detect meaningful differences by sex. It has been suggested that inconsistency in the active smoking and follicular lymphoma association across studies might be due to confounding by ETS among never smokers (10). However, in a previous analysis of the CPS-II, ETS did not confound the association between active smoking and follicular lymphoma (5). The prospective studies of active smoking and follicular lymphoma in women have shown a 2-fold higher risk for current smokers compared with never smokers (5, 7, 10, 15), which is similar to that found for ETS exposure in childhood and adulthood. This finding is initially surprising given the lower carcinogenic dose inhaled from ETS compared with that from active smoking (2). However, the reference groups in ETS studies comprise never smokers unexposed to ETS, whereas never smokers in most active smoking studies include those with exposure to ETS (84% of CPS-II never smokers). If ETS exposure actually increases the rate of follicular lymphoma, then the reference rate is lower in studies of ETS than in studies of active smoking, making a doubling of relative risk in both studies possible. A second possibility is that subjects exposed to ETS as both children and adults have been exposed to ETS for a longer duration and at potentially more vulnerable ages compared with those exposed to active smoking, perhaps explaining relative risks similar to those of active smoking.

Similar to those of our study, results of the CTS also showed an inverse association, albeit not statistically significant, between adult ETS exposure and the risk of DLBCL (10). In previous analyses of the CPS-II and in several other prospective cohort studies, current smokers had a lower risk of DLBCL compared with never smokers (4, 5, 7, 18). However, among all of those studies, there was no dose-response relationship, suggesting that there may be another factor associated with ETS exposure that reduces the risk of DLBCL. Taken together, neither the ETS nor the active smoking literature provides particularly strong evidence for an association of tobacco smoke with lower risk of DLBCL.

Although the carcinogenic dose from ETS is lower than that from active smoking (2), ETS is considered a human carcinogen (1). ETS is made up primarily of unfiltered sidestream smoke produced from the end of the smoldering cigarette. The concentrations of some compounds in sidestream smoke can be up to 10 times greater than those in mainstream smoke (i.e., exhaled smoke) (2). Studies in mice have shown condensates from sidestream smoke to be significantly more carcinogenic than condensates from mainstream smoke (2). There are several biologically plausible mechanisms that may explain an association between ETS and NHL. In a recent commentary (19), many of the proposed mechanisms related to smoking were discussed, including the presence of multiple known chemical carcinogens in tobacco smoke, the increased occurrence of t(14;18) chromosomal translocations in the peripheral blood of smokers, and possible detrimental effects of tobacco smoke on the immune system. Children are particularly vulnerable to toxins like ETS because of greater exposure relative to body weight and immature metabolism pathways (20).

It is possible that the opposing associations of ETS with follicular lymphoma and DLBCL are related. ETS could preferentially increase the risk of follicular lymphoma by initiating t(14;18) translocations, which are present in 90% of follicular lymphomas and in only 20%–30% of DLBCLs (13). If follicular lymphoma and DLBCL arise through a common molecular pathway, an increase in t(14;18) translocations could result in both more cases of follicular lymphoma and fewer cases of DLBCL. Notably, 25%–35% of follicular lymphomas transform to DLBCL (13), indicating that there is at least 1 shared pathway between the 2 subtypes. Alternatively, subjects with higher ETS exposure may access the health care system more often (because of other conditions, such as asthma) making them more likely to be diagnosed with follicular lymphoma before transformation to DLBCL, representing a diagnostic rather than a biological effect. Otherwise, ETS associations with follicular lymphoma and DLBCL may be independent or the result of chance.

The strengths of this study include its large prospective design, the ability to fully characterize never smokers, and dose measurements of both childhood and adult ETS exposures. The large size of the CPS-II Nutrition Cohort and 16 years of follow-up for incident cancer provided an adequate number of NHL cases to evaluate the risks of NHL subtypes in a never-smoking population. The prospective design reduces the possibility of recall bias based on disease status. The detailed questionnaires on active smoking history over follow-up ensured that the ETS and NHL relationship was not confounded by active smoking later in life. This is currently the only study that has published results on the duration and intensity of childhood ETS exposure. In addition, the quantitative measurement of adult exposure to hours per week of ETS allowed us to examine very small exposures to ETS and provided an alternative measure of intensity to that used in the CTS.

An important limitation of this study is the incomplete assessment of lifetime ETS exposure data. Although this study captures childhood exposure well, the adult ETS exposure is limited to the timing of baseline questionnaires in 1982 and 1992. The average ages of CPS-II Nutrition Cohort participants were 52 and 63 years at the times adult ETS exposure was measured; thus, early adult exposure was not captured. Childhood and adult ETS exposures are correlated, and it is possible that associations limited to childhood ETS may actually represent unmeasured early adult ETS exposure. It is likely that ETS exposure has changed for many subjects over time as a result of changes in living conditions, working conditions, and the advent of smoke-free air laws in public places. However, in both 1982 and 1992, complete smoking bans in public places were still uncommon (21). Therefore, both adult measurements of ETS exposure are representative of a time prior to smoke-free air laws and are more representative of ETS in public places in early adulthood than later measurements would be. Missing data assumptions in 1982 may underestimate adult ETS exposure, although the results of sensitivity analyses using only 1992 data were similar. Importantly, the misclassification of adult ETS exposure should bias our results toward the null. Other limitations include residual confounding and the possibility of bias created by censoring.

In conclusion, our results are consistent with those in both the ETS and active smoking research that indicate a positive association between exposure to tobacco smoke and the risk of follicular lymphoma. Furthermore, childhood ETS exposure may contribute to this risk, although additional research is needed to confirm this association. If confirmed, these data have important implications for understanding possible long-term carcinogenic effects of tobacco smoke exposure in children, 58% of whom have measurable levels of tobacco smoke exposure in the United States (2). This study supports public smoking bans and other policies that reduce exposure to ETS.

ACKNOWLEDGMENTS

Author affiliations: Epidemiology Research Program, American Cancer Society, Atlanta, Georgia (W. Ryan Diver, Lauren R. Teras, Mia M. Gaudet, Susan M. Gapstur).

This work was supported by the American Cancer Society, which funds the creation, maintenance, and updating of the Cancer Prevention Study-II cohort.

We thank each member of the study management group and acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention's National Program of Cancer Registries and cancer registries supported by the National Cancer Institute's Surveillance, Epidemiology, and End Results Program.

Conflict of interest: none declared.

REFERENCES

1
IARC Working Group on the Evaluation of Carcinogenic Risks to Humans
Personal habits and indoor combustions. Volume 100 E. A review of human carcinogens
IARC Monogr Eval Carcinog Risks Hum
 , 
2012
, vol. 
100
 
Pt E
(pg. 
1
-
538
)
2
US Department of Health and Human Services
The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General
2006
Atlanta, GA
US Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health
3
Schöllkopf
C
Smedby
KE
Hjalgrim
H
, et al.  . 
Cigarette smoking and risk of non-Hodgkin's lymphoma—a population-based case-control study
Cancer Epidemiol Biomarkers Prev
 , 
2005
, vol. 
14
 
7
(pg. 
1791
-
1796
)
4
Nieters
A
Rohrmann
S
Becker
N
, et al.  . 
Smoking and lymphoma risk in the European Prospective Investigation Into Cancer and Nutrition
Am J Epidemiol
 , 
2008
, vol. 
167
 
9
(pg. 
1081
-
1089
)
5
Diver
WR
Patel
AV
Thun
MJ
, et al.  . 
The association between cigarette smoking and non-Hodgkin lymphoid neoplasms in a large US cohort study
Cancer Causes Control
 , 
2012
, vol. 
23
 
8
(pg. 
1231
-
1240
)
6
Morton
LM
Holford
TR
Leaderer
B
, et al.  . 
Cigarette smoking and risk of non-Hodgkin lymphoma subtypes among women
Br J Cancer
 , 
2003
, vol. 
89
 
11
(pg. 
2087
-
2092
)
7
Parker
AS
Cerhan
JR
Dick
F
, et al.  . 
Smoking and risk of non-Hodgkin lymphoma subtypes in a cohort of older women
Leuk Lymphoma
 , 
2000
, vol. 
37
 
3-4
(pg. 
341
-
349
)
8
Stagnaro
E
Ramazzotti
V
Crosignani
P
, et al.  . 
Smoking and hematolymphopoietic malignancies
Cancer Causes Control
 , 
2001
, vol. 
12
 
4
(pg. 
325
-
334
)
9
Zahm
SH
Weisenburger
DD
Holmes
FF
, et al.  . 
Tobacco and non-Hodgkin's lymphoma: combined analysis of three case-control studies (United States).
Cancer Causes Control
 , 
1997
, vol. 
8
 
2
(pg. 
159
-
166
)
10
Lu
Y
Wang
SS
Reynolds
P
, et al.  . 
Cigarette smoking, passive smoking, and non-Hodgkin lymphoma risk: evidence from the California Teachers Study
Am J Epidemiol
 , 
2011
, vol. 
174
 
5
(pg. 
563
-
573
)
11
Calle
EE
Rodriguez
C
Jacobs
EJ
, et al.  . 
The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and baseline characteristics
Cancer
 , 
2002
, vol. 
94
 
9
(pg. 
2490
-
2501
)
12
Turner
JJ
Morton
LM
Linet
MS
, et al.  . 
InterLymph hierarchical classification of lymphoid neoplasms for epidemiologic research based on the WHO classification (2008): update and future directions
Blood
 , 
2010
, vol. 
116
 
20
(pg. 
e90
-
e98
)
13
Swerdlow
SH
Campo
E
Harris
NL
, et al.  . 
WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues
 , 
2008
Lyon, France
IARC Press
14
Cox
DR
Regression models and life-tables
J R Stat Soc B
 , 
1972
, vol. 
34
 
2
(pg. 
187
-
220
)
15
Kroll
ME
Murphy
F
Pirie
K
, et al.  . 
Alcohol drinking, tobacco smoking and subtypes of haematological malignancy in the UK Million Women Study
Br J Cancer
 , 
2012
, vol. 
107
 
5
(pg. 
879
-
887
)
16
Herrinton
LJ
Friedman
GD
Cigarette smoking and risk of non-Hodgkin's lymphoma subtypes
Cancer Epidemiol Biomarkers Prev
 , 
1998
, vol. 
7
 
1
(pg. 
25
-
28
)
17
Lim
U
Morton
LM
Subar
AF
, et al.  . 
Alcohol, smoking, and body size in relation to incident Hodgkin's and non-Hodgkin's lymphoma risk
Am J Epidemiol
 , 
2007
, vol. 
166
 
6
(pg. 
697
-
708
)
18
Troy
JD
Hartge
P
Weissfeld
JL
, et al.  . 
Associations between anthropometry, cigarette smoking, alcohol consumption, and non-Hodgkin lymphoma in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial
Am J Epidemiol
 , 
2010
, vol. 
171
 
12
(pg. 
1270
-
1281
)
19
Morton
LM
Smoking and lymphoma: a small part of a complex story
Leuk Lymphoma
 , 
2012
, vol. 
53
 
10
(pg. 
1853
-
1854
)
20
Landrigan
PJ
Goldman
LR
Children's vulnerability to toxic chemicals: a challenge and opportunity to strengthen health and environmental policy
Health Aff
 , 
2011
, vol. 
30
 
5
(pg. 
842
-
850
)
21
American Nonsmokers Rights Foundation
Chronological table of U.S. population protected by 100% smokefree state or local laws
 
http://www.no-smoke.org/pdf/EffectivePopulationList.pdf Updated January 2, 2014 Accessed January 13, 2014

Author notes

Abbreviations: CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; CPS-II, Cancer Prevention Study-II; CTS, California Teachers Study; DLBCL, diffuse large B-cell lymphoma; ETS, environmental tobacco smoke, NHL, non-Hodgkin lymphoma.