Exogenous hormone use, reproductive factors, and risk of intracranial meningioma in females

Eligible case patients included all persons diagnosed from May 1, 2006, to October 6, 2011, with a histologically confirmed intracranial meningioma among residents of the states of Connecticut, Massachusetts, and North Carolina as well as the California counties of Alameda, San Francisco, Contra Costa, Marin, San Mateo, and Santa Clara and the Texas counties of Brazoria, Fort Bend, Harris, Montgomery, Chambers, Galveston, Liberty, and Waller counties of Texas. Patients 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. Control individuals 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 patients with a history of meningioma and/or a brain lesion of unknown pathology were excluded. Individuals were English- or Spanish-language speaking. The study, consent forms, and questionnaire were approved by the institutional review boards at Yale University School of Medicine, Brigham and Women’s Hospital, University of California at San Francisco, M D Anderson Cancer Center, and 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 Massachusetts Tumor Registry. The physicians of each eligible patient were contacted to request permission to approach the individual. Patients approved for contact by their physicians, and control individuals identified by Kreider Research were sent an introductory letter. Approximately 1–2 weeks later, a trained interviewer contacted the potential study participant by telephone to administer the interview. Interviews took an average of 52 minutes. Proxies provided information for 9 patients and no controls. The questionnaire included detailed questions on demographics, family history of cancer, pregnancy and menstrual history, exogenous hormone history, and medical history. Women were considered perimenopausal if they reported no regular menses for up to 12 months and postmenopausal if they reported no regular menses for at least 12 months or had undergone a bilateral oophorectomy. Women who had missing or incomplete information on menstruation, who reported a hysterectomy, or who indicated use of exogenous hormones/intrauterine device while still menstruating were considered postmenopausal if they were 55 years of age or older, perimenopausal if they were between 46 and 55 years, or premenopausal if they were less than 46 years at interview. Individuals who had smoked a total of 100 cigarettes or more in their lifetime were defined as “ever” smokers. Those 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. Risk factor and screening information were truncated at the date of diagnosis for patients and the date of interview for controls (hereafter referred to as the reference date). To date, 2404 eligible patients and 2958 eligible controls have been identified. Ninety-eight percent of eligible patients had a consenting physician. Among those patients, 65% participated in the interview portion of the study, while 52% of eligible controls participated in the interview. Six hundred ninety-six patients were ineligible due to out-of-state residency (n = 48), language (n = 74), recurrent meningioma (n = 84), incarceration (n = 3), age (n = 50), spinal meningioma (n = 148), pathology unavailable for review (n = 75), mental or medical (for example, deafness) illness (n = 110), death (cause of death other than meningioma) (n = 79), another pathology (for example, lung metastasis) (n = 16), or other reason (n = 9). One hundred ten control individuals were ineligible due to out-of-state residency (n = 6), language (n = 8), a history of brain tumor of unknown pathology (n = 8), age group (n = 1), mental or medical illness (n = 70), death (n = 3), or other reaon (n = 14). Interviewed and noninterviewed patients 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 being older than noninterviewed controls. The sample used in this analysis includes the 1127 female case and 1092 female control individuals. The initial portion of the statistical analysis included descriptive statistics. We used t-tests, chi-square tests, and Fisher exact tests 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 odds ratios (adjusted for age, alcohol use [yes/no], race [white versus nonwhite], education [(≤16 years versus > 16 years of age) in addition to the endogenous and exogenous hormone variables with 95% confidence intervals using the statistical package PC-SAS version 9.2 (SAS Institute). (Because the variables of income and education were co-linear, only education was included, as the data were more complete.) Linear trends were assessed across ordered categories.

Descriptive statistics for the study sample are provided in Table 1 . The mean age was 57.2 years for patients versus 57.4 years for controls (p = 0.58). The majority of study subjects identified themselves as white. Patients and controls did not differ by age or geographic location by design. Controls were more likely to be white, to have 16 or more years of schooling, and to have an annual salary greater than $75,000. Results from the multivariate analysis are presented in Table 2 . Patients and controls did not differ by age at menarche, age at menopause, or number of full-term pregnancies, although the mean age at first live birth was higher in controls than patients (27.7 vs 25.9 years, respectively; p < 0.01). Breastfeeding was associated with a reduced risk of meningioma, significantly so if a woman had breastfed for 6 months or longer (OR = 0.75, 95% CI 0.60–0.94). Patients were more likely to report having undergone a bilateral oophorectomy, even after adjustment for a history of surgery for ovarian or endometrial cancer or fibroid tumors. Patients had a significantly higher BMI than did controls (28.0 vs 27.1, respectively, p < 0.01). Women in the second, third and fourth quartile had 1.06 (95% CI 0.83–1.35), 1.13 (95% CI 0.89–1.45), and 1.29 (95% CI 1.01–1.65) times the risk of women in the first quartile (p = 0.04, test for trend). As we have reported previously, current smoking status was associated with a significantly reduced risk of meningioma in females. 9 Interestingly, the use of alcohol was also associated with a reduced risk of meningioma. With respect to the use of exogenous hormones, overall there was no increase in risk associated with use (ever/ never) of OCs. When all women, regardless of menopausal status, were included in the analysis, an elevated risk was not seen for women currently using OCs (OR 1.45, 95% CI 0.94–2.24). When the analysis was restricted to include only premenopausal women, a significant increase in risk was appreciable for current OC users (OR 1.8, 95% CI 1.1–2.9), although the number of exposed women was low (only 87 patients and 71 controls were defined as current OC users). Although a similar pattern of risks was appreciated between HRT and meningioma risk, no statistically significant increase in risk was seen for “ever,” past, or current users relative to “never” users, with less than 10% of the sample reporting current HRT use. No association was present between the reported use of fertility medications and meningioma risk.

This is the largest and most recent case-control study to examine the relationship between hormonal or reproductive factors and meningioma risk. Unlike many previous studies, in the present one we were able to control for a number of confounding factors such as race, education, and cigarette smoking. We found that high BMI, current use of OCs in premenopausal women, and not smoking were significantly associated with meningioma risk. There was limited evidence of an association between reproductive variables and meningioma risk with younger age at first live birth and a history of breastfeeding associated with decreased risk. No significant association was seen between use of either fertility medication or HRT and risk. A number of studies have examined the association between meningioma risk and OC use 3 , 5 , 12 , 15 , 17 , 18 , 22 , 23 , 31 ( Table 3 ) as well as HRT 2 , 4 , 5 , 12 , 15 , 17 – 20 , 22 , 23 , 31 ( Table 4 ). Little evidence is seen for an increased risk associated with “ever” use of OC, although several studies (primarily cohort) and ours suggest an increase in risk with current OC use. The EPIC investigators 23 noted a higher risk for current versus “never” OC users (OR 3.61, 95% CI 1.75–7.46), whereas the Nurses Health Study 17 reported an elevated but not statistically increased risk (OR 1.34,95% CI 0.18–9.96) for current users. (The Million Women Study 3 did not report current use risk.) Few investigators have been able to examine the relationship between sex hormone receptor expression and OC use. The two case-control studies that did so reported conflicting results, the interpretation of which is complicated by differing definitions of receptor positivity across studies. 12 , 19 There is little evidence of an association between HRT use and risk of meningioma in any of the case-control studies (including ours), but three 2 , 17 , 23 of the four 18 prospective cohort studies (as well as two 4 , 19 of the three 5 retrospective cohorts) reported significantly elevated risks either for current 2 , 17 , 23 or “ever” 19 use of HRT ( Table 4 ). In a case-control analysis within the Nurses Health Study, 17 the relative risk of meningioma associated with current HRT for pre- and postmenopausal women was 2.48 (95% CI 1.29–4.77) and 1.86 (95% CI 1.07–3.24), respectively when compared with postmenopausal women who had never used hormones. 17 Results from the Million Women Study 2 (RR 1.34, 95% CI 1.03–1.75) and EPIC 23 cohorts (HR1.79, 95% CI 1.18–2.71) concur. The reason for such conflicting results by study design (for example, casecontrol vs cohort) is unclear but may point to common difficulties associated with case-control studies including recall bias or the inclusion of controls with significantly higher socioeconomic status than cases. Three studies examined risk by HRT preparation: one case-control 12 study found no association, but two cohort studies 2 , 19 reported increased risk with estrogen-only preparations of note given the generally low rates of estrogen and high rates of progesterone receptor expression in meningioma tumors. 8 Two case-control studies 12 , 20 examined risk by hormone receptor status with no evidence of an association. If current but not past use of oral hormones is associated with meningioma risk, this may suggest a promoter rather than initiator role of these medications; however, further analyses by preparation type and tumor hormone receptor are needed. To date, no study has examined risk concurrently by both preparation and hormone receptor status. With the exception of the Nurses Health Study, 17 no association has been documented 12 , 18 , 22 , 23 , 26 , 30 between age at menarche and meningioma risk, which is consistent with our findings. The Nurses Health Study 17 investigators observed relative risks of 1.29 (95% CI 0.86 – 1.92) and 1.97 (95% CI 1.06, 3.66) for menarche at ages 12–14 and 14 or older, respectively, compared with menarche before age 12. In general, age at menopause, 18 , 23 , 30 age at first birth, 3 , 12 , 18 , 21 , 23 , 30 and parity 3 , 10 , 18 , 21 , 23 , 30 have not been associated with meningioma risk. One case-control study 22 with 219 cases found a protective effect for pregnancy, which increased with number of pregnancies and age at first pregnancy. Data on breastfeeding and risk are limited 12 , 15 , 23 , 30 but mostly 12 , 15 , 30 similar to our findings of an inverse association with risk. An unexpected but previously reported finding is the positive association between bilateral oophorectomy and risk. 18 , 23 The explanation for this finding is unclear but may relate to the fact that all but 5 of these patients also underwent a hysterectomy and were more likely to receive estrogen-only HRT than were women not undergoing such a procedure. As previously detailed, 9 a history of ever smoking significantly decreased the risk of meningioma in women. Similar findings were noted for current and past smokers. The effects were similar in both pre- and postmenopausal women. The finding of a protective effect of smoking among women is intriguing (and consistent across studies) 3 , 23 in light of the suggestive but poorly defined role for hormonal factors for meningioma. Cigarette smoking is hypothesized to be anti-estrogenic by enhancing the metabolism of estradiol to inactive cathechol estrogens, increasing the binding of estrogen by serum sex hormone–binding globulin, and decreasing adipose-derived estrogen 1 Our findings of a positive association between increased BMI and meningioma risk are consistent with all 3 , 12 , 17 , 18 but one 22 previous study, most of which were not able to reach statistical significance given small sample size. Mechanisms by which obesity may be related to risk include decreased insulin sensitivity or increased inflammatory response (of interest given our previous findings in these data for immune-related factors and risk). 27 , 29 Limitations include the fact that exogenous hormone use was not validated but was determined through a telephone interview. We created a picture booklet of OCs and HRT to use as a memory aid. Despite this, approximately 25% of women could not recall the name of at least one preparation that they had used over their lifetime, a percentage that appears to correlate with data from other studies of OC use. 24 In addition, despite being the largest study to date of meningioma, the data are limited by small numbers of current OC users (as meningioma is generally diagnosed in postmenopausal women) and of current postmenopausal HRT users (likely due to the sharp reduction in use of HRT in the US over the past decade. 16 Differential recall by case-control status is possible, although widespread knowledge of any association between meningioma and hormonal therapies 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 individuals. Patients and controls did not differ by race, age, or geographical site but did differ with respect to education and income, with controls reporting higher income and education levels than patients, 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 exogenous hormone use, may lead to bias in risk estimation. In addition, we obtained information on a range of covariables that could impact the association between HRT and meningioma. Our ability to control for these variables enabled us to minimize potential overestimation of risk. Histological confirmation was obtained for all patients, suggesting that these results may only be applicable to lesions that are deemed in need of surgery rather than conservative management.

The examination of variants drawn from genes in hormones 13 , 14 , 25 and meningioma risk remains unexplored, with few genes examined in only small samples. The extent to which risk for meningioma associated with exposure to exogenous hormones is modified by genotype is unknown. Overall, our data suggest but do not confirm a role for hormone-related factors but are limited by the small numbers of patients currently on oral hormone medications and a lack of hormone receptor data, and this is an important area for future study.