Medical History, Lifestyle, Family History, and Occupational Risk Factors for Chronic Lymphocytic Leukemia/Small Lymphocytic Lymphoma: The InterLymph Non-Hodgkin Lymphoma Subtypes Project

Abstract

Background

Chronic lymphocytic leukemia (CLL) and small lymphocytic lymphoma (SLL) are two subtypes of non-Hodgkin lymphoma. A number of studies have evaluated associations between risk factors and CLL/SLL risk. However, these associations remain inconsistent or lacked confirmation. This may be due, in part, to the inadequate sample size of CLL/SLL cases.

Methods

We performed a pooled analysis of 2440 CLL/SLL cases and 15186 controls from 13 case-control studies from Europe, North America, and Australia. We evaluated associations of medical history, family history, lifestyle, and occupational risk factors with CLL/SLL risk. Multivariate logistic regression analyses were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs).

Results

We confirmed prior inverse associations with any atopic condition and recreational sun exposure. We also confirmed prior elevated associations with usual adult height, hepatitis C virus seropositivity, living or working on a farm, and family history of any hematological malignancy. Novel associations were identified with hairdresser occupation (OR = 1.77, 95% CI = 1.05 to 2.98) and blood transfusion history (OR = 0.79, 95% CI = 0.66 to 0.94). We also found smoking to have modest protective effect (OR = 0.9, 95% CI = 0.81 to 0.99). All exposures showed evidence of independent effects.

Conclusions

We have identified or confirmed several independent risk factors for CLL/SLL supporting a role for genetics (through family history), immune function (through allergy and sun), infection (through hepatitis C virus), and height, and other pathways of immune response. Given that CLL/SLL has more than 30 susceptibility loci identified to date, studies evaluating the interaction among genetic and nongenetic factors are warranted.

Chronic lymphocytic leukemia (CLL) and small lymphocytic lymphoma (SLL) are two subtypes of non-Hodgkin lymphoma (NHL). They are clinically defined by the presence of a clonal population of B-cell lymphocytes that have a characteristic immunophenotype (1). Because of their common immunophenotype and their similar clinical course, CLL and SLL are typically combined together despite the more prominent nodal involvement in SLL.

CLL/SLL is one of the most common B-cell malignancies in individuals of Caucasian descent and is very rare in individuals of Asian descent. The incidence of CLL/SLL has remained fairly constant over time with an annual incidence between 4 and 10 per 100000 (2–6). However, recent incidence studies of CLL in Asian populations have shown an increase in incidence suggesting an environmental role (7,8). CLL/SLL incidence is nearly twice as high in men as in women, and CLL/SLL incidence increases with age, with a median age at CLL/SLL diagnosis of 73 years (3–5,9).

A number of studies have been conducted to identify risk factors for CLL/SLL. The strongest and most consistent risk factor for CLL/SLL is family history of hematologic malignancy, regardless of how it is defined (i.e., defined by any family history of lymphoma, or focused on a family history of NHL, or specifically on a family history of CLL/SLL) (10–12). However, with other risk factors, these finding have yet to be confirmed (e.g., atopic associations), lacked consistency (e.g., smoking associations), or have not been assessed in a multivariate setting to explore relative independence. The lack of consistency may be due, in part, to inadequate sample size of CLL/SLL cases, especially for those factors that may be more modestly linked to CLL/SLL. To advance our understanding of the etiology of CLL/SLL, we conducted one of the largest epidemiological studies of CLL/SLL to date by investigating associations with lifestyle, medical history, family history, and selected occupational risk factors in a pooled analysis of 2440 cases and 15186 controls from 13 case-control studies from Europe, North America, and Australia as part of the International Lymphoma Epidemiology Consortium (InterLymph) NHL Subtypes Project. Moreover, we considered exposures not only individually but also jointly to assess independence among exposures.

Methods

Study Population

Detailed methodology for the InterLymph NHL Subtypes Project is provided elsewhere in this issue. Studies eligible for inclusion in this pooled analysis fulfilled the following criteria: 1) case-control design; 2) inclusion of incident CLL/SLL cases with the clonality of the circulating B-cell lymphocytes confirmed by flow cytometry; and 3) availability of individual-level data for at least several risk factors of interest by December 31, 2011. Most studies excluded individuals with a known history of solid organ transplantation or HIV/AIDS. Contributing studies were approved by local ethics review committees, and all participants provided written, informed consent before interview.

NHL Subtype Ascertainment and Harmonization

Cases were classified according to the World Health Organization classification (13,14) using guidelines from the InterLymph Pathology Working Group (15,16). Most studies had some form of centralized pathology review by at least one expert hematopathologist to confirm the diagnoses. Each participating study’s pathology review procedures, rules for NHL subtype classification, and NHL subtype distribution were then reviewed independently by an interdisciplinary team of pathologists and epidemiologists from InterLymph.

Risk Factor Ascertainment and Harmonization

Each study collected data on putative CLL/SLL risk factors in a standardized, structured format by in-person or telephone interviews (typically computer-assisted) or self-administered questionnaires. Risk factors selected for inclusion in this analysis were lifestyle, medical history, family history, and occupational risk factors with data from at least four studies. Centralized harmonization of de-identified, individual-level data from each study was a key element of the project. Each exposure variable was harmonized individually; data were then reviewed for consistency among related exposure variables. Details of the collected data and data harmonization rules are provided elsewhere in this issue.

Statistical Analysis

We first performed analyses to evaluate risk of CLL/SLL with each exposure variable using unconditional logistic regression models adjusted for age, race/ethnicity, sex, and study (i.e., the “basic model”). The statistical significance of each exposure was evaluated by a likelihood ratio test, comparing models with and without the exposure variable of interest, with P values less than 0.05 identifying putatively influential factors. Individuals with missing data for the exposure variable of interest were excluded. To evaluate effect heterogeneity among the 13 studies, we performed a separate logistic regression within each study and then quantified the variability of the coefficients by the H statistic, adapting the definition by Higgins and Thompson to categorical variables (17). No meaningful heterogeneity was observed (results not shown).

We then examined the relationship between CLL/SLL risk and each putative risk factor considering possible effect modification and accounting for other potential confounders. To consider possible effect modification, we repeated the above logistic regression analyses, but now stratified individuals by age, sex, race/ethnicity, region, study, study design (i.e., population-based versus hospital- or clinic-based), or other putative risk factors identified in the analysis. Forest plots illustrated the results from the stratified analyses to identify possible modifiers of the effect of an exposure variable of interest. No evidence of effect modification was observed (results not shown). To account for other potential confounders, we conducted two analyses. First, we evaluated the risk estimate for each putative risk factor in a series of models that adjusted for one other putative risk factor individually as well as age, race/ethnicity, sex, and study (pairwise adjustment modeling). Second, we conducted a single logistic regression model including all putative risk factors, this time including a separate missing category for each variable to ensure that the whole study population was included in the analysis (i.e., not dropped due to missing data). Finally, we conducted a forward step-wise logistic regression with all putative risk factors, adjusting for age, sex, race/ethnicity, and study, to identify our “full model.”

Because controls for most of the 13 studies were frequency matched by age and sex to all cases (e.g., all NHL subtypes) rather than just to CLL/SLL, we conducted sensitivity analyses using a subset of controls from each study that were frequency matched by age and sex to CLL/SLL cases. The results from these sensitivity analyses were very similar to the results obtained using the full set of controls (results not shown); thus, we retained the full set of controls for our main analyses to increase statistical power.

Results

Table 1 shows the distribution of the 2440 CLL/SLL cases and 15186 controls across the 13 participating studies, along with characteristics for age, race/ethnicity, sex, and social economic status. Relative to controls, cases tended to be older (over the age of 60 years) and were more likely to be men. Cases had a median age at diagnosis of 64 years (range 28–93) compared with 60 years (range 17–97) at interview in controls. The majority (>95%) of cases and controls were non-Hispanic whites. There was no evidence of a difference in distribution of social economic status across cases and controls (P = .19).

Table 1.

Characteristics of CLL/SLL cases and controls included in the InterLymph NHL Subtypes Project*

	Controls	Cases
	No. (%)	No. (%)
Total	15186 (86.2)	2440 (13.8)
Study
North America	5848 (38.5)	881 (36.1)
British Columbia	845 (5.6)	42 (1.7)
Iowa/Minnesota	1245 (8.2)	244 (10.0)
Mayo Clinic	1314 (8.7)	376 (15.4)
NCI-SEER	1055 (6.9)	133 (5.5)
Nebraska (newer)	533 (3.5)	29 (1.2)
University of Rochester	139 (0.9)	7 (0.3)
Yale	717 (4.7)	50 (2.0)
Europe	8644 (56.9)	1530 (62.7)
Engela	722 (4.8)	132 (5.4)
EpiLymph	2460 (16.2)	414 (17.0)
Italy multicenter	1771 (11.7)	214 (8.8)
Italy (Aviano-Naples)	504 (3.3)	18 (0.7)
SCALE	3187 (21.0)	752 (30.8)
Australia
New South Wales	694 (4.6)	29 (1.2)
Design
Population based	11093 (73.0)	1656 (67.9)
Hospital based	4093 (27.0)	784 (32.1)
Age
<30	831 (5.5)	1 (0.0)
30–39	1190 (7.8)	29 (1.2)
40–49	2044 (13.5)	205 (8.4)
50–59	3326 (21.9)	603 (24.7)
60–69	4386 (28.9)	925 (37.9)
70–79	2974 (19.6)	565 (23.2)
≥80	435 (2.9)	84 (3.4)
Missing	0 (0.0)	28 (1.1)
Sex
Men	8472 (55.8)	1614 (66.1)
Women	6714 (44.2)	826 (33.9)
Race/ethnicity
White, non-Hispanic	14303 (94.2)	2336 (95.7)
Black	199 (1.3)	16 (0.7)
Asian	189 (1.2)	5 (0.2)
Hispanic	95 (0.6)	9 (0.4)
Other/unknown/missing	400 (2.6)	74 (3.0)
Social economic status
Low	6141 (40.4)	1076 (44.1)
Medium	4655 (30.7)	684 (28.0)
High	4139 (27.3)	560 (23.0)
Other/missing	251 (1.7)	120 (4.9)
NHL classification
World Health Organization		1982 (81.2)
Working Formulation		458 (18.8)

* CLL/SLL = chronic lymphocytic leukemia/small lymphocytic lymphoma; NCI-SEER = National Cancer Institute--Surveillance, Epidemiology, and End Results; NHL = non-Hodgkin lymphoma; SCALE = Scandinavian Lymphoma Etiology Study.

Basic Model Results

Medical History.

Suffering from any atopic condition (including allergy, hay fever, asthma, or eczema) was inversely associated with CLL/SLL risk (odds ratio [OR] = 0.86, 95% confidence interval [CI] = 0.78 to 0.95; Table 2). Within specific atopic conditions, similar effect sizes were noted for allergy, food allergy, and hay fever conditions, although not statistically significant for the latter two (Table 2). Little to no evidence of a reduced risk was observed for asthma (OR = 0.99, 95% CI = 0.83 to 1.17) and eczema (OR = 0.96, 95% CI = 0.82 to 1.13). When we excluded individuals who had atopic conditions diagnosed within 2 years of age of CLL/SLL diagnosis or interview, the effects sizes remained consistent for any atopy, (OR = 0.82, 95% CI = 0.72 to 0.93), allergy (OR = 0.83, 95% CI = 0.67 to 1.03), and hay fever (OR = 0.83, 95% CI = 0.69 to 0.99). Having a history of transfusion also was inversely associated with CLL/SLL (OR = 0.79, 95% CI = 0.66 to 0.94; Table 2). When we stratified the study sample based on study design, we observed nonsignificant differences in effects between population-based case-control studies (OR = 0.89, 95% CI = 0.68 to 1.16) and hospital-based case-control studies (OR = 0.71, 95% CI = 0.56 to 0.91). Further, the association was more notable for those who received transfusion after 1990 (OR = 0.68, 95% CI = 0.49 to 0.94). There was also a weak upward trend with adult height (per 10cm change) when considered as a continuous variable (OR = 1.10, 95% CI = 1.02 to 1.19). We observed no significant associations with CLL/SLL risk with history of any of the specific autoimmune diseases, weight, body mass index, reproductive history, oral contraceptive use, and hormone replacement therapy (results not shown).

Table 2.

Basic model results for medical exposures*

	Controls	Cases	OR (95% CI)†	P
	No. (%)	No. (%)
Any atopic disorder‡
No	9733 (64.1)	1640 (67.2)	1.00 (referent)	.003
Yes	5192 (34.2)	705 (28.9)	0.86 (0.78 to 0.95)
Allergy§
No	9796 (70.9)	1720 (72.7)	1.00 (referent)	.020
Yes	3165 (22.9)	462 (19.5)	0.87 (0.77 to 0.98)
Food allergy
No	11708 (84.7)	1988 (84.1)	1.00 (referent)	.176
Yes	899 (6.5)	104 (4.4)	0.86 (0.69 to 1.07)
Asthma
No	12507 (83.1)	2012 (82.7)	1.00 (referent)	.859
Yes	1150 (7.6)	176 (7.2)	0.99 (0.83 to 1.17)
Hay fever
No	9984 (68.7)	1639 (67.9)	1.00 (referent)	.056
Yes	2513 (17.3)	311 (12.9)	0.88 (0.76 to 1.01)
Eczema
No	12486 (85.0)	2057 (84.9)	1.00 (referent)	.650
Yes	1426 (9.7)	203 (8.4)	0.96 (0.82 to 1.13)
Blood transfusion
No	7419 (73.5)	1000 (68.2)	1.00 (referent)	.008
Yes	1459 (14.5)	168 (11.5)	0.79 (0.66 to 0.94)
Age at first transfusion
No transfusion	7419 (73.5)	1000 (68.2)	1.00 (referent)	.133
First transfusion at age <25	338 (3.4)	32 (2.2)	0.80 (0.55 to 1.16)
First transfusion at age 25–39	419 (4.2)	45 (3.1)	0.81 (0.59 to 1.12)
First transfusion at age 40–54	338 (3.4)	40 (2.7)	0.79 (0.56 to 1.11)
First transfusion at age 55+	364 (3.6)	51 (3.5)	0.77 (0.56 to 1.05)
Total number of blood transfusions
No transfusion	7419 (73.5)	1000 (68.2)	1.00 (referent)	.036
1 transfusion	961 (9.5)	118 (8.0)	0.81 (0.66 to 0.99)
2 transfusions	273 (2.7)	35 (2.4)	0.90 (0.62 to 1.30)
3+ transfusions	170 (1.7)	13 (0.9)	0.61 (0.34 to 1.10)
Transfusion, but number unknown	55 (0.5)	2 (0.1)	0.32 (0.08 to 1.34)
Number of years from first transfusion to date of diagnosis
No transfusion	7419 (73.5)	1000 (68.2)	1.00 (referent)	.038
<20 y	646 (6.4)	72 (4.9)	0.71 (0.55 to 0.92)
20–39 y	563 (5.6)	67 (4.6)	0.89 (0.68 to 1.16)
40+ y	250 (2.5)	29 (2.0)	0.80 (0.53 to 1.19)
Transfusion before 1990
No transfusion	7419 (73.5)	1000 (68.2)	1.00 (referent)	.020
Transfusion before 1990	985 (9.8)	115 (7.8)	0.81 (0.66 to 1.01)
Transfusion 1990+	369 (3.7)	45 (3.1)	0.68 (0.49 to 0.94)
Transfusion year unknown	105 (1.0)	8 (0.5)	1.45 (0.62 to 3.38)
Missing, no transfusion data	1211 (12.0)	299 (20.4)
Adult height
Quartile 1 (low)	2763 (24.1)	398 (20.4)	1.00 (referent)	.079
Quartile 2	2752 (24.0)	471 (24.1)	1.12 (0.96 to 1.30)
Quartile 3	2649 (23.1)	450 (23.0)	1.14 (0.97 to 1.33)
Quartile 4 (high)	2740 (23.9)	475 (24.3)	1.23 (1.05 to 1.44)
Adult height (10cm)
Continuous	10904 (100.0)	1794 (100.0)	1.10 (1.02 to 1.19)	.015
Serology hepatitis C virus infection
No	5259 (68.4)	973 (77.5)	1.00 (referent)	.009
Yes	95 (1.2)	21 (1.7)	2.08 (1.23 to 3.49)

* CI = confidence interval; OR = odds ratio.

† OR (95% CI) adjusted for age, sex, race/ethnicity, and study.

^‡ Atopic disorders include asthma, eczema, hay fever, or other allergies, excluding drug allergies.

^§ Allergy excluding drug allergies and other atopic conditions, including hay fever, asthma, and eczema.

As previously published by InterLymph, having a hepatitis C virus seropositivity was strongly associated with CLL/SLL (OR = 2.08, 95% CI = 1.23 to 3.49; Table 2).

Family History.

As expected, having a family history of any hematological malignancy among first-degree relatives was strongly associated with CLL/SLL risk (OR = 2.17, 95% CI = 1.77 to 2.65; Table 3). Risks were also elevated when looking at family history of NHL (OR = 1.92, 95% CI = 1.41 to 2.61) and family history of leukemia (OR = 2.41, 95% CI = 1.85 to 3.14). The risk increased slightly for first-degree relatives who were men compared with women for any family history hematological malignancy. Although not statistically significant (P > .05), the effects were elevated when looking at family history of myeloma (OR = 2.00, 95% CI = 0.92 to 4.34).

Table 3.

Basic model results for family history*

	Controls	Cases	OR (95% CI)*	P
	No. (%)	No. (%)
Any hematologic malignancy
No	8362 (74.0)	1178 (71.0)	1.00 (referent)	<.001
Yes	493 (4.4)	153 (9.2)	2.17 (1.77 to 2.65)
Family history of non-Hodgkin lymphoma
No	7924 (74.9)	1137 (74.5)	1.00 (referent)	<.001
Yes	210 (2.1)	62 (4.1)	1.92 (1.42 to 2.61)
Family history of leukemia
No	7777 (74.5)	1109 (73.0)	1.00 (referent)	<.001
Yes	237 (2.3)	84 (5.5)	2.41 (1.85 to 3.14)
Any hematologic malignancy, male relative
No	7785 (74.5)	1112 (73.2)	1.00 (referent)	<.001
Yes	229 (2.2)	81 (5.3)	2.32 (1.77 to 3.04)
Any hematologic malignancy, female relative
No	7782 (74.5)	1133 (74.5)	1.00 (referent)	<.001
Yes	232 (2.2)	60 (3.9)	1.79 (1.32 to 2.43)

* CI = confidence interval; OR = odds ratio.

† OR (95% CI) adjusted for age, sex, race/ethnicity, and study.

Occupation.

History of living or working on a farm was significantly associated with CLL/SLL risk (OR = 1.21, 95% CI = 1.07 to 1.36; Table 4). However, the association slightly attenuated when evaluating separately history of working on a farm (OR = 1.16, 95% CI = 1.00 to 1.35) or history of living on a farm (OR = 1.12, 95% CI = 0.97 to 1.30). These data were further corroborated with the occupational data. Here, we observed an association with farming occupation (OR = 1.23, 95% CI = 1.04 to 1.45). The effects vary, however, when we further classified farm work into animal farmers (OR = 0.64, 95% CI = 0.43 to 0.96) or crop farmers (OR = 1.19, 95% CI = 0.93 to 1.52); the mixed animal and crop farmer had elevated risks (OR = 1.32, 95% CI = 1.08 to 1.61). Of all other occupations evaluated, only hairdressers had an increased CLL/SLL risk (OR = 1.77, 95% CI = 1.05 to 2.98), although this and other occupations analyses were based on limited numbers of cases (Table 4).

Table 4.

Basic model results for occupational exposures*

	Controls	Cases	OR (95% CI)†	P
	No. (%)	No. (%)
Ever lived or worked on a farm
No	7088 (59.8)	760 (45.2)	1.00 (referent)	.002
Yes	4514 (38.1)	835 (49.7)	1.21 (1.07 to 1.36)
Ever lived on a farm
No	3230 (53.1)	511 (47.1)	1.00 (referent)	.130
Yes	2617 (43.0)	487 (44.9)	1.12 (0.97 to 1.30)
Ever worked on a farm
No	8081 (81.6)	1019 (73.5)	1.00 (referent)	.051
Yes	1626 (16.4)	290 (20.9)	1.16 (1.00 to 1.35)
Farmer
No	7544 (85.8)	806 (77.3)	1.00 (referent)	.019
Yes	2151 (14.2)	236 (22.6)	1.23 (1.04 to 1.45)
Animal farmer
No	8517 (96.8)	1013 (97.1)	1.00 (referent)	.024
Yes	278 (3.2)	29 (2.8)	0.64 (0.43 to 0.96)
Crop farmer
No	8304 (94.4)	952 (91.3)	1.00 (referent)	.185
Yes	491 (5.6)	90 (8.6)	1.19 (0.93 to 1.52)
Mixed animal and crop farmer
No	7597 (91.9)	867 (85.5)	1.00 (referent)	.008
Yes	665 (8.1)	146 (14.4)	1.32 (1.08 to 1.61)
Hairdresser
No	8690 (98.8)	1024 (98.2)	1.00 (referent)	.044
Yes	105 (1.2)	18 (1.7)	1.77 (1.05 to 2.98)

* CI = confidence interval; OR = odds ratio.

† OR (95% CI) adjusted for age, sex, race/ethnicity, and study.

Lifestyle Factors.

As in previous InterLymph reports, a moderate inverse association was noted between sun exposure and CLL/SLL risk, especially for recreational sun exposure (Table 5). We also observed a modest reduced risk of CLL/SLL in ever cigarette smoking (OR = 0.90, 95% CI = 0.81 to 0.99). This effect was further reduced when looking at current cigarette smoking (OR = 0.82, 95% CI = 0.71 to 0.94; Table 5), with a weak nonsignificant inverse trend with cigarettes smoked per day and with duration of smoking. Regular use of hair dyes (by men and women) did not modify the risk of CLL/SLL when evaluating ever use, type of hair dye, color, duration, or frequency (Table 5). However, use of hair dyes before 1980 had an increased risk of CLL/SLL (OR = 1.36, 95% CI = 1.0 to 1.86). No evidence of an association was noted for alcohol consumption and physical activity (results not shown).

Table 5.

Basic model results for lifestyle exposures*

	Controls	Cases	OR (95% CI)†	P
	No. (%)	No. (%)
Total sun exposure (h/wk)
Quartile 1 (low)	1241 (18.0)	162 (15.6)	1.00 (referent)	.005
Quartile 2	1326 (19.2)	148 (14.3)	0.82 (0.64 to 1.05)
Quartile 3	1339 (19.4)	202 (19.5)	1.09 (0.86 to 1.37)
Quartile 4 (high)	1437 (20.8)	173 (16.7)	0.75 (0.59 to 0.96)
Recreational sun exposure (h/wk)
Quartile 1 (low)	1987 (20.5)	380 (24.5)	1.00 (referent)	.029
Quartile 2	2141 (22.1)	295 (19.0)	0.81 (0.68 to 0.96)
Quartile 3	1789 (18.5)	270 (17.4)	0.88 (0.74 to 1.05)
Quartile 4 (high)	2653 (27.4)	428 (27.6)	0.80 (0.69 to 0.94)
History of cigarette smoking‡
No	5721 (39.9)	945 (39.4)	1.00 (referent)	.038
Yes	7406 (51.6)	1246 (52.0)	0.90 (0.81 to 0.99)
Smoking status as of ~1 y before diagnosis/interview
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	.026
Former smoker	3976 (27.7)	769 (32.1)	0.93 (0.83 to 1.04)
Current smoker	2945 (20.5)	378 (15.8)	0.82 (0.71 to 0.94)
Smoker, status unknown	485 (3.4)	99 (4.1)	1.05 (0.79 to 1.38)
Age started smoking cigarettes
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	.105
<14 y	667 (4.7)	100 (4.2)	0.84 (0.66 to 1.06.)
14 to <18 y	2804 (19.6)	449 (18.7)	0.87 (0.76 to 0.99)
18 to <20 y	1466 (10.2)	240 (10)	0.83 (0.71 to 0.98)
20+ y	1950 (13.6)	367 (15.3)	0.96 (0.84 to 1.11)
Smoker, age start unknown	519 (3.6)	90 (3.8)	1.07 (0.82 to 1.39)
Frequency of cigarette smoking
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	.068
1–10 cigarettes/day	2547 (17.8)	426 (17.8)	0.93 (0.82 to 1.06)
11–20 cigarettes/day	3105 (21.7)	545 (22.7)	0.9 (0.80 to 1.02)
21–30 cigarettes/day	771 (5.4)	121 (5)	0.9 (0.72 to 1.11)
30+ cigarettes/day	742 (5.2)	107 (4.5)	0.72 (0.57 to 0.90)
Smoker, cigarettes/day unknown	241 (1.7)	47 (2)	1.06 (0.76 to 1.48)
Duration of cigarette smoking
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	.165
1–20 y	2414 (16.8)	336 (14)	0.97 (0.84 to 1.12)
21–30 y	1535 (10.7)	246 (10.3)	0.92 (0.79 to 1.08)
30–39 y	1560 (10.9)	274 (11.4)	0.82 (0.71 to 0.96)
40+ y	1757 (12.3)	359 (15)	0.88 (0.76 to 1.01)
Smoking duration unknown	140 (1)	31 (1.3)	1.00 (0.67 to 1.51)
Years since quitting cigarette smoking
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	<.001
Former smoker, quit >25 y ago	1236 (8.6)	261 (10.9)	0.90 (0.77 to 1.05)
Former smoker, quit >15 to 25 y ago	1023 (7.1)	212 (8.8)	1.00 (0.85 to 1.19)
Former smoker, quit >5 to 15 y ago	1097 (7.6)	190 (7.9)	0.91 (0.76 to 1.09)
Former smoker, quit ≤5 y ago	561 (3.9)	80 (3.3)	0.80 (0.62 to 1.03)
Former smoker, unknown when quit	59 (0.4)	26 (1.1)	2.54 (1.53 to 4.21)
Lifetime cigarette exposure
Nonsmoker	5721 (39.9)	945 (39.4)	1.00 (referent)	.023
1–10 pack-years	2175 (15.2)	341 (14.2)	1.03 (0.90 to 1.19)
>10–20 pack-years	1487 (10.4)	238 (9.9)	0.88 (0.75 to 1.04)
>20–35 pack-years	1613 (11.2)	275 (11.5)	0.84 (0.72 to 0.98)
>35 pack-years	1827 (12.7)	335 (14)	0.82 (0.70 to 0.95)
Smoker, pack-years unknown	304 (2.1)	57 (2.4)	0.96 (0.71 to 1.31)
Ever used hair dyes
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.502
Ever hair dye	2881 (35.6)	284 (28.0)	1.08 (0.86 to 1.37)
Men	3950 (48.7)	602 (59.4)
Type of hair dye used
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.072
Temporary only	206 (2.5)	14 (1.4)	0.74 (0.41 to 1.34)
Permanent	2380 (29.4)	259 (25.5)	1.16 (0.91 to 1.48)
Ever hair dye, type unknown	295 (3.6)	11 (1.1)	0.52 (0.23 to 1.15)
Men	3950 (48.7)	602 (59.4)
Color of hair dye used
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.375
Light	894 (11.0)	98 (9.7)	1.13 (0.84 to 1.51)
Dark	1677 (20.7)	172 (17.0)	1.11 (0.86 to 1.44)
Ever hair dye, color unknown	310 (3.8)	14 (1.4)	0.65 (0.32 to 1.30)
Men	3950 (48.7)	602 (59.4)
Duration of hair dye use
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.459
1–8 y	921 (11.4)	72 (7.1)	1.00 (0.71 to 1.42)
9–19 y	677 (8.4)	71 (7.0)	1.22 (0.86 to 1.72)
20+ y	770 (9.5)	95 (9.4)	1.26 (0.92 to 1.73)
Ever hair dye, duration unknown	513 (6.3)	46 (4.5)	0.89 (0.57 to 1.36)
Men	3950 (48.7)	602 (59.4)
Frequency of hair dye use
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.096
1–5 times/y	835 (10.3)	64 (6.3)	1.08 (0.76 to 1.53)
6–11 times/y	849 (10.5)	82 (8.1)	1.03 (0.74 to 1.43)
12+ times/y	507 (6.3)	78 (7.7)	1.51 (1.09 to 2.10)
Ever hair dye, frequency unknown	690 (8.5)	60 (5.9)	0.89 (0.62 to 1.28)
Men	3950 (48.7)	602 (59.4)
Used hair dyes before 1980
Never hair dye	1241 (15.3)	120 (11.8)	1.00 (referent)	.152
Ever hair dye use <1980	895 (11.0)	127 (12.5)	1.36 (1.00 to 1.86)
Hair dye use only 1980+	990 (12.2)	96 (9.5)	1.06 (0.76 to 1.46)
Hair dye use, time period unknown	996 (12.3)	61 (6.0)	0.85 (0.57 to 1.27)
Men	3950 (48.7)	602 (59.4)

* CI = confidence interval; OR = odds ratio.

† OR (95% CI) adjusted for age, sex, race/ethnicity, and study.

^‡ Smoked longer than 6 months or more than 100 cigarettes in lifetime.

Full Model Results.

To develop our full model, we selected risk factors that had statistical significance of P values less than .05 in the basic model analyses. However, for those risk factors that had additional variables that captured duration and intensity, we only selected those that had consistent evidence of association across the additional variables. The full model results are shown in Table 6. There is clear evidence that the selected risk factors are independent of one another with minimal evidence of confounding and interaction. This is evidenced by the minimal change in effect size obtained from the basic models (adjusted only for design variables) compared with that obtained from the full model (adjusted for the design variables and the other selected variables). This is also evidenced by our pairwise adjustment modeling.

Table 6.

Basic and full model results*

	Controls	Cases	Basic model	P	Full model	P
	No. (%)	No. (%)	OR (95% CI)†		OR (95% CI)‡
Any atopic disorder§
No	9733 (64.1)	1640 (67.2)	1.00 (referent)	.003	1.00 (referent)	.002
Yes	5192 (34.2)	705 (28.9)	0.86 (0.78 to 0.95)		0.85 (0.77 to 0.94)
Blood transfusion
No	7419 (73.5)	1000 (68.2)	1.00 (referent)	.008	1.00 (referent)	.011
Yes	1459 (14.5)	168 (11.5)	0.79 (0.66 to 0.94)		0.79 (0.66 to 0.95)
Adult height
Continuous (10cm)	10904 (100)	1794 (100)	1.10 (1.02 to 1.19)	.015	1.09 (1.01 to 1.17)	.020
Serology hepatitis C virus infection
No	5259 (68.4)	973 (77.5)	1.00 (referent)	.009	1.00 (referent)	.011
Yes	95 (1.2)	21 (1.7)	2.08 (1.23 to 3.49)		1.99 (1.16 to 3.41)
History of cigarette smoking
No	5721 (39.9)	945 (39.4)	1.00 (referent)	.038	1.00 (referent)	.082
Yes	7406 (51.6)	1246 (52)	0.90 (0.81 to 0.99)		0.91 (0.83 to 1.01)
Total sun exposure (hours/week)
Quartile 1 (low)	1241 (18.0)	162 (15.6)	1.00 (referent)	.005	1.00 (referent)	.003
Quartile 2	1326 (19.2)	148 (14.3)	0.82 (0.64 to 1.05)		0.81 (0.63 to 1.04)
Quartile 3	1339 (19.4)	202 (19.5)	1.09 (0.86 to 1.37)		1.06 (0.84 to 1.34)
Quartile 4 (high)	1437 (20.8)	173 (16.7)	0.75 (0.59 to 0.96)		0.71 (0.55 to 0.92)
First-degree family history, any hematologic malignancy
No	8362 (74.0)	1178 (71.0)	1.00 (referent)	<.001	1.00 (referent)	<.001
Yes	493 (4.4)	153 (9.2)	2.17 (1.77 to 2.65)		2.16 (1.76 to 2.65)
Ever lived or worked on a farm
No	7088 (59.8)	760 (45.2)	1.00 (referent)	.002	1.00 (referent)	.004
Yes	4514 (38.1)	835 (49.7)	1.21 (1.07 to 1.36)		1.20 (1.06 to 1.35)
Hairdresser
No	8690 (98.8)	1024 (98.2)	1.00 (referent)	.044	1.00 (referent)	.044
Yes	105 (1.2)	18 (1.7)	1.77 (1.05 to 2.98)		1.77 (1.05 to 3.01)

* CI = confidence interval; OR = odds ratio.

† OR (95% CI) adjusted for age, sex, race/ethnicity, and study.

^‡ OR (95% CI) adjusted for age, sex, race/ethnicity, study, and all other variables in full model.

^§ Atopic disorders include asthma, eczema, hay fever, or other allergies, excluding drug allergies.

Discussion

Individual level data from 2440 CLL cases and 15186 controls from 13 case-control studies were reanalyzed centrally through the InterLymph consortium to evaluate associations of medical history, family history, lifestyle, and occupational risk factors with CLL/SLL risk. We confirmed prior findings by InterLymph and others with additional cases and controls for a number of exposures. Specifically, we confirmed the previously reported strong increased risk with family history of hematological cancer (10–12), the increased risk with height (18,19), the increased risk with farming exposures (20–24), and the protective effect of UV radiation on CLL/SLL risk (25,26). Although we report a significant increased risk with hepatitis C virus herein, we had no additional new data beyond that reported in an earlier InterLymph pooled analysis (27).

With an additional CLL/SLL cases and controls, we supported an earlier InterLymph finding of a reduction in risk of CLL/SLL with history of any atopic condition (28). A concern with any association with IgE-mediated exposures, like atopic conditions, is reverse causality such that the inefficient immunological repose to allergens may be due to CLL/SLL disease and therefore explain the observed associations. With the larger sample size, we excluded CLL/SLL cases whose atopic diagnosis was within 10 years of CLL/SLL diagnosis to evaluate the role of reverse causality. The results (not shown) from these additional sensitivity analyses were consistent with the full data and show a risk reduction of 20% among those CLL/SLL cases reporting any atopic disease. Thus, the inverse association is unlikely to be due to reverse causality and suggests that induction of an increase IgE response by environmental exposures could be a factor in CLL/SLL pathogenesis (29). Future studies will be needed to test this hypothesis.

We observed an inverse association between blood transfusion and CLL/SLL. These results were consistent across other transfusion variables including number of transfusions, latency, and timeframe of transfusion (before 1990 or after). However, these results are inconsistent with the hypothesis that pathogens are transmitted through blood transfusion and therefore would presumably increase the risk of CLL/SLL. An earlier meta-analysis of case-control and cohort studies reported an increased risk with CLL/SLL (OR = 1.66, 95% CI = 1.08 to 2.56) (30). There was very little overlap of studies included in the earlier meta-analysis and the one reported herein. Although our pooled study of individual-level data is a strength, as is our large sample size, we were unable to explain the biologic mechanism behind our findings. Our results may be due to unknown confounding as we lacked details on the type of transfusion or the indication for the transfusion. Our results may also be due to selection bias because the effect was more inversely associated in hospital-based (OR = 0.71) compared with population-based (OR = 0.89) case-control studies.

We report a statistically significant inverse association between current cigarette smoking and CLL/SLL. The effect of this association is modest, but yet consistent across the other smoking variables with dose response relationships observed with pack-years, intensity, and duration of exposure, although not statistically significant for the latter two. In a recent InterLymph pooled study of smoking that included 1156 CLL/SLL cases and 4630 controls from seven case-control studies, which were also included herein, current cigarette smoking was found to have a nonsignificant reduced risk of CLL/SLL (OR = 0.82, 95% CI = 0.67 to 1.01) (31). With an additional six case-control studies included in this study, we found this effect to remain. Effect estimates from prospective cohort studies have been mixed with both elevated effects (32,33) and protective effects (34) reported with current cigarette smoking; however, none of these results were significant due to small numbers (n < 500 CLL/SLL cases). The biologic mechanism for this inverse association is unclear, but smoking may affect immune function (35). Replication of our findings is needed, but given that we observed a significant but weak effect, large sample sizes will be required or alternative approaches, such as evaluating biomarkers of exposure to cigarette smoke on CLL/SLL risk, will be needed.

We have identified an increased risk of CLL/SLL among hairdressers, which has not been previously reported. Hairdressers can be exposed to a wide variety of chemicals including organic solvents, dyes, and ammonia. The International Agency for Research on Cancer (IARC) categorized the occupational exposures of a hairdresser or barber as probably carcinogenic to humans (Group 2A) and the personal use of hair colorants as not classifiable as to its carcinogenicity to humans (Group 3) (36). Previous studies have explored the association with hematological malignancies or NHL in general with generally negative associations. Data on detailed exposure associated to this occupation were not available. In contrast to these findings among hairdressers, we have inconclusive results with exposure to hair dye use. A previous InterLymph analysis reported an increased CLL/SLL risk for hair dyes use before 1980 among women (OR = 1.50, 95% CI = 1.10 to 2.00) based on a subset of four case-control studies reported herein (37). However, with data from nine additional case-control studies, we found a slight attenuation with this finding (OR = 1.36, 95% CI = 1.00 to 1.86). Future studies will need detailed data of hair dye exposure to hone in on the effect of this exposure on CLL/SLL risk.

Our study has several strengths, including the ability to harmonize individual-level data, the extensive review of harmonization by workgroups to ensure accuracy, the large number of CLL/SLL cases and controls, and the large number of available exposures to simultaneously evaluate joint effects and perform sensitivity analyses. The findings for the exposures were fairly consistent across the studies with modest evidence of heterogeneity. Our study has several limitations, as well. Although case-control studies are subject to recall bias, it is unlikely to have a major effect herein because CLL/SLL has very few established risk factors. All studies used the older CLL/SLL diagnostic criteria that required absolute lymphocyte count more than 5×10⁹ cells/L compared with the new 2008 criteria of B-cell lymphocyte count more than 5×10⁹ cells/L. Under the new diagnostic change, at least a third of the Rai stage 0 CLL/SLL cases are reclassified to monoclonal B-cell lymphocytosis, a precursor condition to CLL/SLL (2). Although our sample size is large (even after accounting for the misclassification of cases), it is possible that some of our novel findings are due to chance.

Overall, the results of this pooled analysis provide additional evidence that a number of exposures are associated with CLL/SLL risk and that these exposures are independent of each other. A number of the exposures are not modifiable (e.g., race/ethnicity, sex, family history, height, and atopy), whereas some potentially modifiable exposures may decrease risk (e.g., UV radiation), whereas others may increase risk (e.g., farm exposures and hair products). Further studies are needed to confirm our smoking and transfusion findings, as well as detailed studies evaluating occupational exposures of farming and hairdressers. The biologic basis for these associations remains to be elucidated; however, our findings support that genetic factors, immune function, and infection have a role in CLL/SLL leukemogenesis. Given that CLL/SLL has more than 30 susceptibility loci identified to date (38–42), studies evaluating the interaction among these genetic and nongenetic factors is warranted.

Funding

Intramural Research Program of the National Cancer Institute/National Institutes of Health and National Cancer Institute/National Institutes of Health (R01 CA14690, U01 CA118444, and R01 CA92153-S1).

InterLymph annual meetings during 2010–2013 were supported by the Epidemiology and Genomics Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute/National Institutes of Health (2010–2013); Lymphoma Coalition (2010–2013); National Institutes of Health Office of Rare Diseases Research (2010); National Cancer Institute/National Institutes of Health (R13 CA159842 01; 2011); University of Cagliari, Provincial Administration of Cagliari, Banca di Credito Sardo, and Consorzio Industriale Sardo, Italy (2011); Intramural Research Program of the National Cancer Institute/National Institutes of Health (2012); and Faculté de Médecine de Dijon, Institut de Veille Sanitaire, Registre des hémopathies malignes de Côte d’Or, INSERM, Institut National du Cancer, Université de Bourgogne, Groupe Ouest Est d’Etude des Leucémies et Autres Maladies du Sang (GOELAMS), l’Institut Bergonié, The Lymphoma Study Association (LYSA), Registre Régional des Hémopathies de Basse Normandie, and the City of Dijon, France (2013). Meeting space at the 2013 Annual Meeting of the American Association for Cancer Research (AACR) was provided by the Molecular Epidemiology Group (MEG) of the AACR. Pooling of the occupation data was supported by the National Cancer Institute/National Institutes of Health (R03CA125831).

Individual studies were supported by the Canadian Institutes for Health Research (CIHR), Canadian Cancer Society, and Michael Smith Foundation for Health Research (British Columbia); Intramural Research Program of the National Cancer Institute/National Institutes of Health (Iowa/Minnesota); National Cancer Institute/National Institutes of Health (N01-CP-ES-11027; Kansas); National Cancer Institute/National Institutes of Health (R01 CA50850; Los Angeles); National Cancer Institute/National Institutes of Health (R01 CA92153 and P50 CA97274), Lymphoma Research Foundation (164738), and the Henry J. Predolin Foundation (Mayo Clinic); Intramural Research Program of the National Cancer Institute/National Institutes of Health and Public Health Service (contracts N01-PC-65064, N01-PC-67008, N01-PC-67009, N01-PC-67010, and N02-PC-71105; NCI-SEER); National Cancer Institute/National Institutes of Health (R01CA100555 and R03CA132153) and American Institute for Cancer Research (99B083; Nebraska [newer]); National Cancer Institute/National Institutes of Health (N01-CP-95618) and State of Nebraska Department of Health (LB-506; Nebraska [older]); National Cancer Institute/National Institutes of Health (R01CA45614, RO1CA154643-01A1, and R01CA104682; UCSF1); National Cancer Institute/National Institutes of Health (CA143947, CA150037, R01CA087014, R01CA104682, RO1CA122663, and RO1CA154643-01A1) [UCSF2]; National Heart Lung and Blood Institute/National Institutes of Health (hematology training grant award T32 HL007152), National Center for Research Resources/National Institutes of Health (UL 1 RR024160), and National Cancer Institute/National Institutes of Health (K23 CA102216 and P50 CA130805; University of Rochester]; National Cancer Institute/National Institutes of Health (CA62006 and CA165923; Yale); Association pour la Recherche contre le Cancer, Fondation de France, AFSSET, and a donation from Faberge employees (Engela); European Commission (QLK4-CT-2000-00422 and FOOD-CT-2006–023103), Spanish Ministry of Health (CIBERESP, PI11/01810, RCESP C03/09, RTICESP C03/10, and RTIC RD06/0020/0095), Rio Hortega (CM13/00232), Agència de Gestió d’Ajuts Universitaris i de Recerca–Generalitat de Catalunya (Catalonian Government, 2009SGR1465), National Institutes of Health (contract NO1-CO-12400 ), Italian Ministry of Education, University and Research (PRIN 2007 prot.2007WEJLZB, PRIN 2009 prot. 20092ZELR2), Italian Association for Cancer Research (IG grant 11855/2011), Federal Office for Radiation Protection (StSch4261 and StSch4420), José Carreras Leukemia Foundation (DJCLS-R04/08), German Federal Ministry for Education and Research (BMBF-01-EO-1303), Health Research Board, Ireland and Cancer Research Ireland, and Czech Republic MH CZ - DRO (MMCI, 00209805) [EpiLymph]; National Cancer Institute/National Institutes of Health (CA51086), European Community (Europe Against Cancer Programme), and Italian Alliance Against Cancer (Lega Italiana per la Lotta contro i Tumori; Italy, multicenter); Italian Association for Cancer Research (IG 10068; Italy, Aviano-Milan); Italian Association for Cancer Research (Italy, Aviano-Naples); Swedish Cancer Society (2009/659), Stockholm County Council (20110209), Strategic Research Program in Epidemiology at Karolinska Institut, Swedish Cancer Society (02 6661), Danish Cancer Research Foundation, Lundbeck Foundation (R19-A2364), Danish Cancer Society (DP 08-155), National Cancer Institute/National Institutes of Health (5R01 CA69669-02), and Plan Denmark [SCALE]; Leukaemia & Lymphoma Research (United Kingdom); and Australian National Health and Medical Research Council (ID990920), Cancer Council NSW, and University of Sydney Faculty of Medicine (New South Wales).

We thank the following individuals for their substantial contributions to this project: Dennis P. Robinson and Priya Ramar (Mayo Clinic College of Medicine) for their work at the InterLymph Data Coordinating Center in organizing, collating, harmonizing, and documenting of the data from the participating studies in the InterLymph Consortium; Michael Spriggs, Peter Hui, and Bill Wheeler (Information Management Services, Inc) for their programming support; and Noelle Richa Siegfried and Emily Smith (RTI International) for project coordination.