Cigarette Smoking and Risk of Meningioma: The Effect of Gender

In the most recent report from the Central Brain Tumor Registry of the United States (CBTRUS), intracranial meningiomas are identified as the most frequently reported primary brain tumor in adults within the United States (1). The increasing awareness of the import of these tumors has led to a desire to investigate possible risk factors with ionizing radiation (IR) the most consistently confirmed risk exposure (2–9). Few other factors have been identified (2) although a number of investigators have examined the role of cigarette smoking (10–21). The examination of this exposure is an intriguing one given both smoking's well-known association with a wide range of cancers as well as its potential antiestrogenic effects. Notably, an inverse association of cigarette smoking has been reported for tumors such as endometrial cancer (22) which like meningioma may have a hormonal etiology. Reports of an association between smoking and meningioma have been inconsistent when examined across gender (10–13). However, when stratified by gender, several projects have suggested a variable effect (14–21) with women at decreased and men at increased risk. Several reports have also suggested confounding of risk by menopausal status (20) as well as exposure to diagnostic (15) or therapeutic (16) IR although no confirmation of these results exist. With the exception of the Million Women Study Cohort (20), previous studies have been hampered by small sample size or an inability to control for potential confounding variables in the statistical analysis. This report compares self-reported smoking history in 1,433 persons with intracranial meningioma to those of 1,349 controls. The large sample size of this population-based study will provide a more precise estimate of any association by gender. Moreover, the multiple covariates included in the data collection allow for the first time the joint statistical control of potential confounding factors such as education, body mass index, and menopausal status.

Eligible case subjects include all persons diagnosed from May 1, 2006 to April 28, 2011 with a histologically confirmed intracranial meningioma among residents of the states of Connecticut, Massachusetts, and North Carolina as well as the Alameda, San Francisco, Contra Costa, Marin, San Mateo, and Santa Clara counties of California and the Brazoria, Fort Bend, Harris, Montgomery, Chambers, Galveston, Liberty, and Waller counties of Texas. Cases were identified through the Rapid Case Ascertainment systems and state cancer registries of the respective sites and were between the ages of 20 and 79 years at time of diagnosis. Controls were selected by random-digit-dialing by an outside consulting firm (Kreider Research) and were matched to cases by 5-year age interval, sex, and state of residence. Study subjects with a previous history of meningioma and/or a brain lesion of unknown pathology were excluded. Subjects were English or Spanish speaking. The study, consent forms, and questionnaire were approved by the Institutional Review Boards at the Yale University School of Medicine, Brigham and Women's Hospital, the University of California at San Francisco, the MD Anderson Cancer Center, and the Duke University School of Medicine. The study was also approved by the State of Connecticut Department of Public Health Human Investigation Committee with some data directly obtained from the Connecticut Tumor Registry in the Connecticut Department of Public Health as well as the MA Tumor Registry.

The physicians of each eligible case were contacted to request permission to approach the case. Cases approved for contact by their physicians and controls identified by Kreider Research were sent an introductory letter. Approximately 1 or 2 weeks later, a trained interviewer contacted the potential study subject by telephone to administer the interview. Interviews took an average of 52 minutes. Proxies provided information for 9 cases and no controls. The questionnaire included detailed questions on demographics, family history of cancer, pregnancy and menstrual history, exogenous hormone history, and medical history. Subjects who had smoked a total of 100 cigarettes or more in their lifetime were defined as “ever smokers.” Smokers were asked the age at which they started (and for past smokers the age at which they stopped) smoking cigarettes, the number of cigarettes smoked per day, and the total number of years of smoking. Subjects who answered “0” to the question “in a typical week over the past year, on how many days did you consume an alcoholic beverage of any type (beer, wine, hard liquor)?” were defined as nondrinkers. In defining exposure to therapeutic IR, subjects were asked whether they had ever undergone radiation treatment to the head, neck, face, or chest. For exposure to diagnostic radiation, subjects were questioned whether they had ever received a dental X-ray (bitewing, full mouth, or panorex) a cerebral angiogram or a computed tomograph (CT) of the head. Risk factor and screening information were truncated at the date of diagnosis for cases and the date of interview for controls (hereafter referred to as the reference date).

To date, 2,228 eligible cases and 2,604 eligible controls have been identified. Ninety-eight percent of eligible cases had a consenting physician. Among those cases, 65% participated in the interview portion of the study while 52% of eligible controls participated in the interview. Six hundred sixty-six cases were ineligible due to out-of-state residency (45), language (70), recurrent meningioma (83), incarcerated (3), age (50), spinal meningioma (144), pathology unavailable for review (56), mental or medical (i.e., deaf) illness (96), deceased (cause of death other than meningioma; 76), another pathology (i.e., lung metastasis; ref. 16), or other (27). Eighty-five controls were ineligible due to out-of-state residency (6), language (8), a history of previous brain tumor unknown pathology (8), age group (1), mental or medical illness (53), deceased (3), or other (9). Interviewed and noninterviewed cases were similar with respect to age, sex, and residence. Interviewed and noninterviewed controls did not differ by sex or residence but did differ by age with interviewed controls older than noninterviewed controls. The sample used in this analysis includes 1,433 case and 1,349 control subjects.

The initial portion of the statistical analysis included descriptive statistics. T-tests, χ², and Fisher exact tests were used to examine the association between meningioma risk and independent covariates. To assess the odds of meningioma associated with risk factors, conditional logistic regression was used to provide maximum likelihood estimates of the ORs [adjusted for age, alcohol use (yes/no), race (white vs. nonwhite), education (≤16 vs. >16 years), and body mass index] with 95% CI using the statistical package PC-SAS version 9.2 (25). (As the variables income and education were colinear, only education was included as the data were more complete). Linear trend was assessed across ordered categories. As prior studies examined the association between cigarette smoking and meningioma risk by menopausal status (pre vs. post; ref. 19), receipt of a full-mouth dental X-ray (ever/never; ref. 15), and radiotherapy to the head (16), we also examined the effect of these variables in the final model.

An electronic search of the MEDLINE, ISI Web of Science, and EMBASE databases from 1970 to August 2011 identified 6 case–control (15–19,21) and 1 cohort (20) studies quantifying associations between cigarette smoking and meningioma by gender (Table 1). To be eligible for inclusion, publications had to include original data and to present gender-specific OR or relative risk quantifying the association between cigarette smoking (ever vs. never) and meningioma risk. Using the inverse variance mixed effects model of DerSimonian and Laird (23), separate meta-analyses were conducted for males and for females with the RevMan v5.1.2 software (Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen, Sweden). Study heterogeneity was assessed with the I² statistic (24). To assess the presence of reporting bias, funnel plots graphing estimates of study precision against the ORs were created and visually inspected.

Descriptive statistics for the study sample are provided in Table 2. The mean age was 57.5 years for cases versus 57.4 years for controls (P = 0.74). The majority of study subjects were female and White. Cases and controls did not differ by age, race, sex, and geographic location by design. Controls were more likely to have 16 or more years of schooling and to have a salary greater than $75,000.

Table 3 compares reported smoking histories for cases and controls. There was a significant interaction between ever having smoked and sex (P = 0.01) supporting the stratification of risk estimates by sex. Regardless of sex, cases and controls did not differ significantly by mean age at first use, last use, or mean duration. However, women smoked less than did men with an older age at first use, younger age at last use, and shorter duration than did men. Among cases, smokers were significantly older at age of diagnosis than were nonsmokers (60.4 vs. 56.0 years for males, P < 0.01 and 59.4 vs. 56.4 years for females, P < 0.01, respectively).

Women who reported ever having smoked cigarettes were at significantly decreased risk of meningioma (aOR, 0.8; 95% CI, 0.7–0.9) relative to women who had never smoked. Conversely, among men, ever smokers had an increased risk of meningioma (aOR, 1.3; 95% CI, 1.0–1.7) relative to never smokers. Risk for females did not vary by duration or amount of use while for men an elevated risk was seen with increased duration and increased number of pack-years; (OR, 1.6; 95% CI, 1.1–2.2) for men with a 13 or more pack-year history.

We attempted to examine previously reported effect modification by menopausal status as well as by history of diagnostic or therapeutic radiation. Among women, we tested for an interaction with smoking exposure by menopausal status however no significant differences were seen with ever (P = 0.26), current (P = 0.28), or past (P = 0.41) smoking status. When controlled for a history (ever/never) of bitewing, full mouth, or panorex dental films, a history of head CT, or a history of prior radiotherapy to the head, neck, face, or chest there is no evidence of effect modification.

In addition to our own data, the review identified 6 studies which merited inclusion in the meta-analyses [The 1980 study by Preston-Martin (18) was dropped for lack of numeric detail]. The meta-analysis of females included 2,015 cases and 1,178,932 controls. Females who reported ever smoking were at significantly decreased risk of meningioma relative to never smokers in the meta-analysis (OR, 0.82; 95% CI, 0.68–0.98; Fig. 1). Results of a sensitivity analysis, conducted by carrying out the cumulative meta-analysis with each study systematically omitted, one at a time with replacement, did not indicate that any one study was exerting undue influence on the summary measure. Moderate study heterogeneity was detected in the meta-analysis of females (I² = 53%), but this heterogeneity is entirely due to a single study (I² = 0% when the study of Hu and colleagues is dropped from the analysis).

The meta-analysis of males included 599 cases and 754 controls. Ever smokers had a significantly increased risk of meningioma relative to never smokers (OR, 1.39; 95% CI, 1.08–1.79; Fig. 2). Sensitivity analyses did not indicate that any one study was exerting undue influence on the summary measure. Only minimal study heterogeneity was detected (I² = 17%). Funnel plot results lessened concern for the presence of substantial publication bias for either sex (data not shown).

This is the largest and the most recent case–control study to examine the relationship between cigarette smoking and meningioma risk. Unlike previous studies, we were able to both stratify by gender and control for a number of confounding factors such as education, alcohol use, and body mass index. In these data, active cigarette smoking was associated with an increased risk in men but a decreased risk in females. A number of previous authors have examined the relationship between cigarette smoking and meningioma risk with inconsistent results when males and females are grouped together (10–13) but as formally examined by our meta-analysis, remarkably consistent results [with the exception of the early study by Preston-Martin (18) which included 185 females cases from the Los Angeles area] when stratified by gender.

The finding of a protective effect of smoking among women in our study is intriguing in light of the suggestive but poorly defined role for hormonal factors for meningioma (26). An association between hormones and meningioma risk has been suggested by the increased incidence of the disease in women versus men, the presence of hormone (particularly progesterone) receptors on some meningiomas, an association between breast cancer, uterine fibroids, endometriosis and meningioma risk (27), indications that meningiomas change in size during the luteal phase of the menstrual cycle and pregnancy, and in vitro proliferation of meningioma cell lines in culture after exposure to estrogens. In the one previous case–control study to examine risk by menopausal status a stronger effect was noted in premenopausal women (19) although we were not able to detect such an effect, potentially due to the smaller number of premenopausal women in our data. Cigarette smoking is hypothesized to be antiestrogenic by enhancing the metabolism of estradiol to inactive cathechol estrogens, increasing the binding of estrogen by serum sex hormone–binding globulin, as well as decreasing adipose-derived estrogen (28). The effect of smoking has been examined in a number of hormone-associated cancers including breast for which results have been inconsistent and endometrial (9) for which smoking has been consistently associated with decreased risk. In addition to a hormonal difference, the observed variation in risk associated with cigarette smoking for women versus men may be due to other factors including differences in patterns of cigarette use by gender (28,29). Smoking may also serve as a marker for other variables associated with risk in men but not women including alcohol use, weight (and hence amount of adipose tissue), and socioeconomic variables, although these variables were controlled for in our analyses.

Strengths to the study include the population-based study design, large sample size, and relatively consistent magnitude and direction of risk estimates. Histologic confirmation was obtained for all case subjects suggesting that these results may only be applicable to lesions that are deemed in need of surgery rather than conservative management.

Limitations for this study include the possibility of misreporting of cigarette smoking by study participants. Self-reporting of cigarette smoking may also vary by gender although data that correlate thiocyanate and cotinine levels in male and female study subjects with self-reported cigarette use suggest that self-report is a reliable and cost efficient means to measure smoking behavior in both men and women (30, 31). Differential recall by case–control status is possible although a widespread knowledge of any association between meningioma and smoking among the general public is unlikely given the limited research on this topic. We noted lower than expected (although in line with other recent studies of brain tumors) response rates among control subjects. Cases and controls did not differ by race, age, sex, or geographic site but did differ with respect to education and income with controls reporting higher income and education than controls, suggesting a greater willingness among persons of higher socioeconomic status to participate in epidemiology research. Although these variables were adjusted for in all analyses, such differences in socioeconomic status, a factor likely related to cigarette smoking use, may lead to bias in risk estimation, although the opposite direction of risks identified here seems to argue against such a bias.

The extent to which risk for meningiomas associated with exposure to cigarette smoke is modified by genotype is unknown and this is an important area for future study. Genetic variants in genes involved in the control of aromatic hydrocarbons have been implicated in meningioma risk, but not confirmed (32–34).

Given the important role of IR in meningioma risk, several previous groups have attempted to control for IR exposure when assessing risk associated with smoking. In their population-based case–control study including 200 cases of meningioma, Phillips and colleagues (15) assessed risk with cigarette smoking that occurred 10 or more years before the meningioma surgery and reported gender-specific findings quite similar to ours. Although the actual estimates were not presented, when the authors controlled for subjects who reported ever having a full mouth dental X-ray series, findings for active smoking were strengthened. Flint-Richter and colleagues (16) assessed the role of smoking in presumed radiation- and nonradiation–related meningiomas using data from the Tineas Capitas Cohort (3). They reported an increased risk associated with smoking for men. For women, they observed a significant inverse association of meningioma with smoking (OR, 0.32; 95% CI, 0.14–0.77) with a dose–response association (P < 0.01) in nonirradiated (mean dose, 1.5 Gy) women and a nonsignificant increase risk of meningioma in irradiated women. These findings lead the authors to speculate on the existence of an interaction between IR and smoking in meningioma risk for women. No effect modification by exposure to IR (either diagnostic or therapeutic) was appreciated in our analyses. Further study of the possible role of IR in the examination of smoking and meningioma risk is of interest. Studies such as this one allow for the collection of large numbers of persons with varying gene–environment combinations and hence comparison of the effect of exposures such as IR across genetic variant; our group plans to examine these interactions in future work.

Our results suggest a gender-specific relationship between smoking and intracranial meningioma risk. The large size of our data set (which includes information on important confounding variables) allows us to confirm a reduced risk for women who are active smokers and offers additional insight into what is likely a complex relationship between hormonal factors and meningioma risk.

The authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. No potential conflicts of interest were disclosed.

Conception and design: E.B. Claus, L. Calvocoressi, M.L. Bondy, J.M. Schildkraut, M.R. Wrensch, J.L. Wiemels

Development of methodology: E.B. Claus, K.M. Walsh, M.R. Wrensch, J.L. Wiemels

Acquisition of data: E.B. Claus, L. Calvocoressi, M.L. Bondy, J.M. Schildkraut, M.R. Wrensch, J.L. Wiemels

Analysis and interpretation of data: E.B. Claus, K.M. Walsh, L. Calvocoressi, M.L. Bondy, M.R. Wrensch, J.L. Wiemels

Writing, review, and/or revision of the manuscript: E.B. Claus, K.M. Walsh, L. Calvocoressi, M.L. Bondy, J.M. Schildkraut, M.R. Wrensch, J.L. Wiemels

Administrative, technical, or material support: E.B. Claus

Study supervision: E.B. Claus, L. Calvocoressi, M.L. Bondy, J.M. Schildkraut, M.R. Wrensch, J.L. Wiemels

The authors thank the cooperation of the following Connecticut Hospitals: Bridgeport Hospital, Bristol Hospital, Charlotte Hungerford Hospital, Danbury Hospital, Day-Kimball Hospital, Eastern Connecticut Health Network, Greenwich Hospital, Griffin Hospital, Hartford Hospital, John Dempsey Hospital, Johnson Memorial Hospital, Lawrence Memorial Hospital, Middlesex Hospital, MidState Medical Center, Hospital of Central Connecticut, New Milford Hospital, Norwalk Hospital, St. Francis Hospital and Medical Center, St. Mary's Hospital, Hospital of St. Raphael, St. Vincent's Medical Center, Stamford Hospital, Waterbury Hospital, William Backus Hospital, Windham Hospital, and Yale-New Haven Hospital.

This work was supported by NIH R01 grants CA109468, CA109461, CA109745, CA108473, and CA109475; NIH R25 grant CA112355 as well as by the Brain Science Foundation and the Meningioma Mommas.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.