Risk factors for meningioma in postmenopausal

0 downloads 0 Views 116KB Size Report
Jul 12, 2011 - Departments of Neurology (D.R.J, J.E.H.) and Health Sciences Research (J.E.O., R.A.V., A.H.W., J.R.C.), ...... JAMA. 1998;279(9):669–674. 34.
Neuro-Oncology 13(9):1011 –1019, 2011. doi:10.1093/neuonc/nor081 Advance Access publication July 12, 2011

N E U RO - O N CO LO GY

Risk factors for meningioma in postmenopausal women: results from the Iowa Women’s Health Study Derek R. Johnson, Janet E. Olson, Robert A. Vierkant, Julie E. Hammack, Alice H. Wang, Aaron R. Folsom, Beth A. Virnig, and James R. Cerhan Departments of Neurology (D.R.J, J.E.H.) and Health Sciences Research (J.E.O., R.A.V., A.H.W., J.R.C.), College of Medicine, Mayo Clinic, Rochester, Minnesota; and Epidemiology and Community Health (A.R.F) and Health Policy & Management (B.A.V.), School of Public Health, University of Minnesota, Minneapolis, Minnesota

Few risk factors for meningioma, aside from increasing age and female sex, have been identified. We investigated risk factors for meningioma in elderly women, a group with a high incidence. We evaluated associations of demographic, lifestyle, medical history, and anthropometric variables with risk of meningioma in the Iowa Women’s Health Study (IWHS), a population-based, prospective cohort study. Risk factors were collected via questionnaires mailed in 1986 and 1992. Incident meningiomas were identified via linkages to Medicare. Cox regression models were used to examine the association of risk factors with meningioma incidence. The mean age at baseline of the 27,791 women in the analysis cohort was 69.3 years (range, 65.0 –84.6 years). During 291,021 person-years of follow-up, 125 incident meningiomas were identified. After adjusting for age, lower levels of physical activity (relative risk [RR] , 0.68 for high versus low; P for trend 5 .039), greater body mass index (BMI; RR, 2.14 for ≥35 versus 19.5 – 24.9 kg/m2; P for trend 5 .0019), greater height (RR, 2.04 for >66 versus ≤62 inches; P for trend 5 .013), and a history of uterine fibroids (RR, 1.72; 95% confidence interval, 1.19, 2.50) were positively associated with meningioma risk in multivariate analysis. BMI at age 18 and 30 years were not associated with risk. There were no associations with menstrual or reproductive factors or other medical history and lifestyle factors. Physical activity, BMI, height, and history of uterine fibroids were associated with meningioma risk in older women. The positive association with height suggests a

Received January 5, 2011; accepted April 20, 2011. Corresponding Author: James R. Cerhan, MD, PhD, Department of Health Sciences Research, Mayo Clinic, 200 First Street SW, Rochester, MN 55905 ([email protected]).

role for early life influences on risk, whereas the associations with BMI and physical activity suggest a role for modifiable factors later in life. Keywords: cohort studies, epidemiology, primary brain tumor, risk factors.

M

eningioma is the most common benign intracranial neoplasm in adults, and postmenopausal women are the age-sex group with the highest incidence of these tumors.1 Despite the fact that meningiomas are relatively common in comparison to other primary intracranial tumors, relatively little is known about their etiology. The only proven exogenous risk factor for meningioma is exposure to radiation.2 Although high-dose radiation is associated with a significant increase in meningioma risk, this exposure is thought to account for only a small proportion of meningiomas at the population level.3 In addition, a number of endogenous risk factors have been suggested to be associated with risk of meningioma.1 For example, several studies suggest that meningioma risk increases with greater height and body mass index (BMI),4,5 and 1 recent study reported an inverse association with physical activity.4 Furthermore, evidence suggests a role for both endogenous and exogenous sex hormones in modifying meningioma risk.6 Meningiomas frequently contain sex hormone receptors,7,8 and the risk of meningioma in women is double the risk in men.1 Investigations into the role of hormone replacement therapy (HRT) in meningioma risk have yielded mixed results, but as a whole, the current literature suggests that HRT is associated with increased risk.9 – 11 Analysis of meningioma risk factors may provide insight into pathogenesis of this tumor. Because there

# The Author(s) 2011. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Johnson et al.: Risk factors for meningioma in post-menopausal women

have been relatively few large prospective cohort studies of meningioma risk factors,4,11 – 13 we analyzed data from the Iowa Women’s Health Study (IWHS), a large, population-based, prospective cohort study. Our analysis included previously studied factors (eg, smoking, alcohol use, reproductive factors, height, and weight), understudied factors (eg, physical activity and rural or farm residence), and novel factors (eg, weight over the life-course, body fat distribution, and certain medical conditions, such as uterine fibroids).

Methods Participants The IWHS cohort was previously described in detail.14 In brief, in 1986, a questionnaire was mailed to 99,826 women aged 55 – 69 years who were randomly selected from Iowa driver’s license files; 41,836 women returned the questionnaire (response rate, 42.7%). There were only minor demographic differences between respondents and nonrespondents, and respondents had somewhat lower mortality and cancer incidence rates, particularly for smoking-related cancers.15 Follow-up questionnaires were mailed in 1987, 1989, 1992, 1997, and 2004. The response rate for the 1992 questionnaire was 79%. Only data from the 1986 and 1992 questionnaires were used in this analysis. Exposure Measurement Risk factor data collected in 1986 included demographic characteristics (eg, date of birth, area of residence [city, rural nonfarm, and farm], and highest level of education), selected medical history (eg, diabetes, benign breast disease, endometriosis, uterine fibroids, and polycystic ovaries), anthropometric characteristics (eg, height, waist circumference, hip circumference, and weight at various ages, including at baseline and ages 18, 30, and 50 years), menstrual factors (eg, age at menarche, age at menopause, and history of oophorectomy), reproductive history (eg, number of live births and age at first live birth), exogenous estrogen use (ie, history of oral contraceptives and use of HRT), and lifestyle factors (eg, alcohol use, smoking history, and frequency of participation in moderate and vigorous physical activity). Additional data collected on the 1992 questionnaire included updated weight, smoking history, use of HRT, lifetime use of alcohol, and diabetes history. A number of calculated variables were derived from the reported baseline and historical anthropometric data. Waist-to-hip ratio was calculated from measurements taken with a paper tape measure enclosed with the questionnaire. Written instructions were provided for having a friend measure circumferences of the waist (1 inch above the umbilicus) and hips (maximal protrusion). This protocol has been shown to be valid (intraclass correlation coefficient with measures by trained technician, ≥0.84) and reliable (intraclass correlation of measures at 2 different time periods, ≥0.85).16

1012

NEURO-ONCOLOGY



SEPTEMBER 2011

Ovulatory years was calculated as age at menopause minus age at menarche, further removing time pregnant. A 3-level physical activity score (low, medium, and high) was calculated from the physical activity questions; this scale has been shown to strongly predict total mortality and cardiovascular disease mortality in this cohort.17 All medical history variables were coded as ever versus never. BMI in 1986 and 1992 was classified into the following categories defined by the World Health Organization (WHO):18 underweight (BMI, ,18.5), normal weight (BMI, 18.5 – 24.9), overweight (BMI, 25.0– 29.9), obese (BMI, 30.0 – 34.9), and severely obese (BMI, ≥35). BMI at earlier ages, as well as all other anthropometric variables and ovulatory years, were categorized into approximate quartiles among all women included in the analysis. Outcome Assessment Incident meningiomas in the cohort were identified by linkage to Medicare files.19 In brief, the cohort was linked to the Medicare enrollment data for 1986 –2004 using an established methodology.20 We obtained Medicare hospitalization data (MedPAR file) for 1986 – 2004 and outpatient and carrier files (physician and other suppliers) for 1991 –2004 (Part B claims were not available before 1991). Meningioma cases were identified from Medicare data by using the International Classification of Diseases, Ninth Revision, codes 192.1, 192.3, 225.2, 225.4, and 237.6 (neoplasms of the cerebral or spinal meninges). To be classified as a case, we required at least 1 of the diagnosis codes from the MedPAR (hospital) file or any 2 diagnoses .30 days apart in the carrier or outpatient files. We used the date of the first claim that ultimately fulfilled our case criteria as the date of diagnosis. Clinical information regarding cases, such as symptomatology and whether diagnoses were pathologic or radiographic, were not available. Statistical Analyses Before data analysis, we defined several exclusions (not mutually exclusive) from the original sample. Because participants who were not enrolled in both parts of Medicare or who were enrolled in managed care plans are unlikely to have complete claims histories,19 we excluded all women who never enrolled in both Part A (includes inpatient care in hospitals and nursing homes) and Part B (outpatient care, which has detailed claims data available starting in 1992) of Medicare for at least 1 month on or after 1 January 1993 (n ¼ 3014). We also excluded women who did not complete the 1992 questionnaire (n ¼ 8819), because it contained many items used in this analysis. Finally, we excluded women with a meningioma in Medicare claims before 1993 (n ¼ 46) or any history of cancer or cancer chemotherapy as reported on the 1986 or 1992 questionnaires or through linkage to the Iowa SEER Cancer Registry through 1992 (n ¼ 6723).

Johnson et al.: Risk factors for meningioma in post-menopausal women

The number of person-years was accumulated from the date that participants met the enrollment criteria until the occurrence of meningioma, Medicare Part A or B disenrollment, death, or 31 December 2004. Women aged ,65 who were enrolled in Medicare due to disability or end-stage renal disease only contributed to analyses starting at age 65 years. We calculated age-specific incidence rates by apportioning each woman’s follow-up into 5-year age categories, allowing a woman to accumulate person-years in multiple age groups, and dividing the observed number of meningioma events by the aggregate personyears. Relative risks (RRs) and corresponding 95% confidence intervals (CIs) were calculated as a measure of association between the risk factors of interest and meningioma incidence and were estimated using Cox proportional hazards regression. Incidence was modeled as function of age.21 Follow-up for women aged ≥65 years at study start date began at their age as of 1 January 1993, whereas follow-up for all other women (n ¼ 5738) began at time of Medicare enrollment, typically at age 65 years. Potential risk factors were apportioned into 4 different groups: demographic and lifestyle, menstrual and reproductive, medical history, and anthropometric. We formally examined associations of risk factors with incidence of meningioma using tests for trend, ordering the categorized values from lowest to highest and including the resulting 1 degree-of-freedom term in the Cox regression model. Trend tests for BMI grouped by WHO categories were run after excluding underweight subjects, and trend tests for number of live births were run after excluding nulliparous women. Tests for all other risk factors were based on the full complement of categories. Because of Medicare rules, we were not able to report the exact number of individuals for exposure categories with counts of ,11 women. Analyses proceeded using a 3-step approach. We first examined associations of meningioma risk with each factor of interest, accounting only for age (modeled as the time variable). Next, we examined the within-group independent associations of those variables found to be significant in univariate analysis in step 1 by simultaneously including each in a multivariate Cox regression analysis. For these analyses, we used a P value inclusion criterion of .10 to reduce the possibility of type II error. Separate analyses were carried out for each of the 4 risk factor groups (demographic and lifestyle, menstrual and reproductive, medical history, and anthropometric). Because of multicollinearity issues, BMI and weight could not be simultaneously included in the anthropometric model; thus, multivariate models excluded weight. Similarly, BMI values across different time points could not be simultaneously included in the anthropometric model; to resolve this issue, we included the BMI measure from 1986, because it had the highest level of significance in the univariate analyses. Finally, all variables still retaining significance in each of the group-specific analyses were simultaneously included in one, final multivariate model. All statistical tests were 2-sided, and all analyses

were carried out using the SAS (SAS Institute) and S-Plus (Insightful) software systems.

Results There were a total of 27,791 women in the at-risk cohort following the application of our exclusion criteria for this analysis. The mean subject age at the beginning of our analysis period was 69.3 years (range, 65.0 – 84.6 years). During 291,021 person-years of follow-up (mean follow-up time, 10.5 years), 125 incident meningiomas were identified. Mean age at meningioma diagnosis was 75.1 years (range, 65.1 – 85.7 years). The overall incidence was 43.0 cases per 100,000 personyears; Table 1 shows age-specific incidence rates in this cohort. After adjusting for age, we did not observe any associations for the demographic or lifestyle factors with risk of meningioma (Table 2), with the exception of physical activity. Compared with women with a low level of physical activity, women at a moderate (RR, 0.57) or high level (RR, 0.61) were at lower risk of developing meningioma (P for trend ¼ .013). With respect to menstrual and reproductive factors (Table 3), we did not observe any statistically significant associations. However, there were suggestive elevations in risk for women with a history of unilateral (RR, 1.39) or bilateral (RR, 1.48) oophorectomy compared with women who had not undergone oophorectomy. We evaluated multiple medical history factors (Table 4), chosen because they potentially influence, or are influenced by, insulin resistance or circulating hormone levels. There was no association between meningioma and adult-onset diabetes, endometriosis, or noncancerous cysts or tumors of the ovary. We did observe elevated risks for benign breast disease (RR, 1.38), history of uterine fibroids (RR, 1.70), and polycystic ovaries (RR, 1.90), although only the association for uterine fibroids was statistically significant at P , .05. There were no associations with use of oral contraceptives or HRT. Among anthropometric variables, there were strong positive associations of meningioma risk with greater weight in 1986 and 1992, BMI in 1986 and 1992 (although weaker than BMI for 1986), and waist and hip circumferences in 1986, as well as a trend toward Table 1. Overall and age-specific meningioma incidence rates in the Iowa Women’s Health Study cohort, 1993–2004 No. of person-years

No. of events

Incidencea

Age group, years

No. of women

65–69

16,362

54,798

22

40.1

70–74

24,488

97,690

36

36.9

75–79 ≥80

23,914 12,976

89,408 49,126

44 23

49.2 46.8

Overall

27,791

291,021

125

43.0

a

Rate per 100,000 person-years.

NEURO-ONCOLOGY



SEPTEMBER 2011

1013

Johnson et al.: Risk factors for meningioma in post-menopausal women

Table 2. Association of demographic and lifestyle factors with risk of meningioma, Iowa Women’s Health Study, 1993– 2004 Factor

No. of person-years

No. of cases

RRa (95% CI)

Pb

Education High school or less More than high school

171,469 118,964

72 53

1.00 (reference) 1.16 (0.57– 2.35)

.74

97,026 131,987

47 55

1.00 (reference) 0.86 (0.58– 1.27)

.33

60,093

23

0.79 (0.48– 1.31)

198,849

86

1.00 (reference)

88,400

34

0.90 (0.60– 1.33)

198,849

86

1.00 (reference)

36,982 49,203

14 20

0.88 (0.50– 1.56) 0.95 (0.58– 1.54)

244,777 43,891

111 13

1.00 (reference) 0.66 (0.37– 1.17)

.16

129,248 81,043

72 26

1.00 (reference) 0.57 (0.37– 0.90)

.013

75,341

26

0.61 (0.39– 0.96)

Residence City (population, .10,000 persons) Nonfarm rural to city (population, ,10,000 persons) Farm Smoking history Never Ever Pack-years of smoking 0 1–19 ≥20

.59

.75

Regular alcohol use No Yes Physical activity index Low Medium High a

Cox proportional hazards regression analysis, adjusted for age. P values are based on the test for trend. Statistically significant or suggestive findings (P , .10) are shown in boldface font.

b

increased risk with greater height (Table 5). The association for waist-to-hip ratio was weaker than BMI, waist circumference, or hip circumference individually. BMI earlier in life (age, 18 or 30 years) was a much weaker risk factor than BMI at age 50 or in 1986/1993. After adjustment for BMI in 1986, associations with waist circumference (P for trend ¼ .33) and hip circumference (P for trend ¼ .35) attenuated and were no longer statistically significant (data not shown). Multivariate Cox regression analysis yielded 4 factors that remained statistically significant in multivariate modeling: WHO BMI category in 1986 (RR, 2.14, 95% CI, 1.35–3.41, for ≥35 versus 19.5–24.9 kg/m2; P for trend ¼ .0019), quartile of height (RR, 2.04, 95% CI, 1.14 –3.54, for .66 versus ≤62 inches; P for trend ¼ .013), level of physical activity (RR, 0.68, 95% CI, 0.43 – 1.07, for high versus low index; P for trend ¼ .039), and history of uterine fibroids (RR, 1.72, 95% CI, 1.19 – 2.50; P ¼ .0042). To rule out the possibility that underlying subclinical disease was affecting these associations, we ran a sensitivity analysis excluding the first year of follow-up for each woman. The results were essentially unchanged from those presented above (data not shown).

Discussion This large prospective cohort study of postmenopausal women offers insight into risk factors for meningioma

1014

NEURO-ONCOLOGY



SEPTEMBER 2011

in a demographic group with the highest incidence of these tumors. We observed a positive association between history of uterine fibroids and risk of meningioma; to the best of our knowledge, this has not been previously reported. We reproduce associations between meningioma risk and greater height, greater BMI, and lower level of physical activity that had previously been reported in the Million Women Study, a large cohort study of women conducted in the United Kingdom.4 Finally, we report for the first time that BMI in the years directly preceding meningioma diagnosis is a stronger predictor of risk than BMI earlier in life. The correlation between BMI and cancer risk is not unique to meningioma: it has been described for multiple other types of cancer.22 For example, among postmenopausal women, high BMI is associated with an increased risk of breast cancer.23 The mechanism of the BMI and cancer association is not entirely clear; one proposed explanation is that BMI influences cancer risk by influencing sex hormone levels and/or insulin resistance,24 which could be relevant to meningioma. In our cohort, BMI in 1992 (just before baseline of the analysis) was a significant predictor of meningioma risk, as was BMI at age 50 years. However, BMI at ages 30 or 18 years was not associated with meningioma risk, suggesting that current or recent hormonal/metabolic effects of BMI around the time of diagnosis may be more important than historical measurements.

Johnson et al.: Risk factors for meningioma in post-menopausal women

Table 3. Association of menstrual and reproductive factors with risk of meningioma, Iowa Women’s Health Study, 1993– 2004 Characteristic

No. of person-years

No. of cases

RRa (95% CI)

Pb

121,899 131,771

54 54

1.00 (reference) 0.92 (0.63– 1.35)

.95

34,542

16

1.04 (0.60– 1.82)

Age at menarche, years ≤12 13 –14 ≥15 Age at menopause, years ,45

64,958

31

1.00 (reference)

45 –49

72,987

23

0.66 (0.39– 1.13)

50 –54 ≥55

110,631 30,489

49 16

0.93 (0.59– 1.45) 1.07 (0.59– 1.96)

211,747 24,115

82 13

1.00 (reference) 1.39 (0.78– 2.50)

.71

History of oophorectomy None Unilateral

50,768

29

1.48 (0.97– 2.26)

Ovulatory years ≤27.7

Bilateral

72,488

40

1.00 (reference)

27.8–33.2

72,164

28

0.70 (0.43– 1.14)

33.3–36.7 .36.7

72,974 73,395

28 29

0.69 (0.43– 1.12) 0.71 (0.44– 1.15)

.054

.16

Live birthsc Nulliparous Parous

1.00 (reference) 1.86 (0.82– 4.22)

.14

.88

No. of live births 1–2 3–4

92,888 116,207

39 55

1.00 (reference) 1.14 (0.75– 1.72)

≥5

55,020

23

1.01 (0.60– 1.69)

Age at first live birth ,20 20 –24 ≥25

53,306

27

1.00 (reference)

133,936

47

0.69 (0.43– 1.11)

75,493

40

1.04 (0.64– 1.71)

.67

a

Cox proportional hazards regression analysis, adjusted for age. P values are based on the test for trend. Statistically significant or suggestive findings (P , .10) are shown in boldface font. Per Medicare data use rules, the number person-years and number of cases are not reported for exposures in which ≥1 category has ,11 cases.

b c

Greater height has been suggested as a risk factor for meningioma in cohort and case-control studies, although this association is not seen in all analyses.25 – 27 In the IWHS cohort, greater height was an independent risk factor for meningioma in multivariate analysis. Increasing height has been previously shown to be a risk factor for a number of other tumor types, both benign and malignant.28 The mechanism behind this association is not clear; it has been hypothesized to be related to a metabolic environment associated with rapid growth in childhood and adolescence.29 Alternatively, it may be simply due to larger body size and a greater amount of tissue at risk of neoplastic transformation. To our knowledge, no analyses correlating height and meningeal surface area or volume have been published, but there is virtually no correlation between height and brain weight or volume, plausible surrogate markers for meningeal surface.30,31 This suggests that the increase in meningioma risk associated with height is not simply due to a greater amount of meningeal tissue.

The inverse association of physical activity with meningioma risk in our study is the first replication of a recent finding reported in the Million Women Study cohort.3 Although the association appears to be independent of BMI, physical activity is also thought to modify cancer risk by the effects of insulin resistance and hormonal milieu. For example, increasing total physical activity is negatively associated with the concentrations of estrone, estradiol, and androstenedione in postmenopausal women,24,32 and positively associated with increased insulin sensitivity.33 Given previously described associations of meningiomas with female sex and reproductive history, a number of reproductive candidate risk factors were examined. We found no association of meningioma risk with age at menarche, age at menopause, number of ovulatory years, age at first live birth, or number of live births. We did observe an elevated risk for parous compare to nulliparous women, although this was not statistically significant, perhaps due to low power (,11 nulliparous cases). Some previous large cohort studies

NEURO-ONCOLOGY



SEPTEMBER 2011

1015

Johnson et al.: Risk factors for meningioma in post-menopausal women

Table 4. Association of medical history factors with risk of meningioma, Iowa Women’s Health Study, 1993–2004 Characteristic

No. of person-years

No. of cases

RRa (95% CI)

Pb

268,902 22,119

114 11

1.00 (reference) 1.18 (0.64– 2.18)

.60

229,379 60,062

92 33

1.00 (reference) 1.38 (0.93– 2.05)

.11

1.00 (reference) 0.81 (0.30– 2.16)

.67

1.00 (reference) 1.70 (1.17– 2.46)

.005

1.00 (reference) 1.90 (0.61– 5.91)

.27

Adult-onset diabetes No Yes Benign breast disease No Yes Endometriosisc No Yes Uterine fibroids No Yes

216,488 69,799

79 43

Polycystic ovaries (Stein-Leventhal syndrome)c No Yes Other noncancerous cysts or tumors of the ovary No Yes

252,545 31,402

108 15

1.00 (reference) 1.13 (0.66– 1.93)

.67

234,242 55,805

105 20

1.00 (reference) 0.82 (0.50– 1.33)

.42

140,824 129,065

57 61

1.00 (reference) 1.17 (0.81– 1.68)

.40

Oral contraceptive use Never Ever Hormone replacement therapy Never Ever a

Cox proportional hazards regression analysis, adjusted for age. Statistically significant or suggestive findings (P , .10) are shown in boldface font. Per Medicare data use rules, the number of person-years and number of cases are not reported for exposures in which ≥1 category has ,11 cases.

b c

have shown an association between parity and meningioma risk,12,34 whereas others have not.11 Analysis of the Nurses’ Health Study (NHS) cohort showed a nonsignificant trend towards increased meningioma risk in parous women compared to nulliparous women.12 The mean age at diagnosis in the NHS cohort was 54 years old. A large population-based European cohort reported no overall association between parity and meningioma, but in a subgroup analysis of women aged ,50 years, number of live births was associated positively with meningioma risk.34 The absence of an association between number of births and meningioma risk among the parous women in the postmenopausal IWHS cohort is concordant with the European data,34 adding weight to the idea that multiparity is not a meningioma risk factor in older women. There has been a great deal of interest in the potential influence of HRT on meningioma risk.5 – 7,9 No association was seen in the IWHS, but the structure of the questionnaires did not include information on hormone dose or formulation (ie, estrogen alone or in combination with progesterone), so the null result does not preclude an association. A history of uterine fibroids was correlated positively with meningioma risk in the IWHS cohort. This association has not been previously been reported and requires replication. The association is not likely to be causal. It is

1016

NEURO-ONCOLOGY



SEPTEMBER 2011

possible that these conditions share a genetic predisposition, but it is more likely that these 2 conditions simply have significantly overlapping risk factors. As in meningioma, insulin resistance and elevated circulating sex hormone levels have been proposed as risk factors for uterine fibroma.35 This analysis of the IWHS reports meningioma risk factors in a large cohort of postmenopausal women. The prospective design minimizes the likelihood of selection and recall bias, and the population-based design increases the external validity. Linkage of the IWHS to Medicare databases allowed for the identification of meningiomas without the need to rely on subject selfreport. We were unable to use or link to the Iowa SEER Cancer Registry because meningioma was not a reportable cancer for most of the follow-up period. Although we were only able to identify cases that occurred in women aged ≥65 years (Medicare eligibility), most of the cohort had already reached this age, and those who had not all aged into Medicare age within 3 years. Our statistical analysis took into account the left truncation, and age-specific (to age 84 years) and overall rates of meningioma agree well with recent SEER data for white women, suggesting relatively complete case ascertainment. This study design also has some potential limitations. The IWHS population is exclusively female and

Johnson et al.: Risk factors for meningioma in post-menopausal women

Table 5. Association of medical history factors with risk of meningioma, Iowa Women’s Health Study, 1993–2004 Characteristic

No. of person-years

No. of cases

RRa (95% CI)

Pb

27 39

1.00 (reference) 1.21 (0.74–1.97)

.082

Height in 1986, inches ≤62 63 –64 65 –66 .66 Weight in 1986, pounds

77,215 92,298 74,247

32

1.24 (0.74–2.07)

47,261

27

1.65 (0.97–2.82)

≤132

75,165

21

1.00 (reference)

133– 147 148– 168

70,380 75,429

21 31

1.07 (0.58–1.95) 1.47 (0.85–2.56)

.168

70,047

52

2.66 (1.61–4.42)

Body mass index in 1986, kg/m2 18.5–24.9 114,906

41

1.00 (reference)

25.0–29.9

109,637

36

0.92 (0.59–1.44)

30.0–34.9 ≥35.0+

45,652 18,422

35 13

2.14 (1.36–3.36) 1.99 (1.06–3.71)

22 20

1.00 (reference) 0.92 (0.50–1.69)

.00002

.00072 d

Waist circumference in 1986, inches ≤30.25 30.26–33.50

71,735 70,290

33.51–37.75

72,378

35

1.56 (0.92–2.67)

.37.75 66,892 Hip circumference in 1986, inches

44

2.13 (1.28–3.56)

≤38.25

72,597

21

1.00 (reference)

38.26–40.50 40.51–43.13

71,227 67,268

22 33

1.07 (0.59–1.94) 1.70 (0.98–2.93)

.43.13

69,168

45

2.26 (1.34–3.79)

Waist to hip ratio ≤0.77

73,501

27

1.00 (reference)

0.78–0.83

73,192

29

1.07 (0.63–1.81)

0.84–0.89 .0.89

72,704 70,766

28 41

1.04 (0.61–1.76) 1.56 (0.96–2.54)

≤132 133– 150

72,645 80,627

19 30

1.00 (reference) 1.43 (0.80–2.54)

151– 170

68,350

33

1.86 (1.06–3.26)

.170 66,368 Body mass index in 1992, kg/m2c

41

2.39 (1.39–4.12)

.00055

.00036

.082

Weight in 1992, pounds

18.5–24.9

1.00 (reference)

25.0–29.9 30.0–34.9

0.97 (0.63–1.52) 1.91 (1.21–3.01)

35.0+

1.37 (0.68–2.74)

Body mass index at age 18 years, kg/m2 ≤19.61 72,358

25

1.00 (reference)

19.62–21.19

73,631

37

1.46 (0.88–2.42)

21.20–23.04 .23.04

72,170 71,629

26 36

1.05 (0.60–1.81) 1.46 (0.88–2.43)

27 35

1.00 (reference) 1.28 (0.77–2.11)

.00074

.023 d

.33

Body mass index at age 30 years, kg/m2 ≤21.07 21.08–22.70

72,177 73,035

22.71–24.74

72,883

28

1.02 (0.60–1.74)

.24.74

68,527

32

1.25 (0.75–2.08)

.60

Continued

NEURO-ONCOLOGY



SEPTEMBER 2011

1017

Johnson et al.: Risk factors for meningioma in post-menopausal women

Table 5. Continued Characteristic

No. of person-years

No. of cases

RRa (95% CI)

Pb .0036

Body mass index at age 50 years, kg/m2 ≤22.73

72,469

20

1.00 (reference)

22.74–24.96 24.97–27.55

75,132 72,969

31 30

1.49 (0.85–2.62) 1.49 (0.84–2.62)

.27.55

69,786

43

2.24 (1.32–3.80)

a

Cox proportional hazards regression analysis, adjusted for age. b P values are based on the test for trend. Statistically significant or suggestive findings (P , .10) are shown in boldface font. c Per Medicare data use rules, the number of person-years and number of cases are not reported for exposures in which ≥1 category has ,11 cases. d Test excludes women with a body mass index ,18.5.

predominantly white, so the risk factors identified may not apply to males or nonwhite ethnic groups. The initial response rate was only 42%, although this should not impact internal validity. Furthermore, we previously found only minor demographic differences between responders and nonresponders and that the association of BMI with cancer occurrence was not appreciably affected by nonresponse bias.15 Some items on the standard questionnaire required subjects to recall elements of their medical history, as well as information from the past, such as weight at age 18 years. However, because these data were prospectively collected there is no reason to suspect a systematic recall bias, although power to identify associations would have been reduced by random reporting error. In addition, this study analyzed incident cases of meningioma, but these tumors often have a long subclinical course, and many of the meningiomas diagnosed during the follow-up period likely developed prior to study entry. Moreover, no screening was performed to detect clinically silent, undiagnosed meningiomas. In aggregate, these data suggest that meningioma risk in postmenopausal women is influenced by hormonal/

metabolic factors that potentially exert their influence both early and late in life. In childhood and adolescence, the metabolic environment associated with rapid growth and eventual tall stature may predispose to meningioma decades in the future. Risk factors in adult life, such as obesity and low level of physical activity, modulate insulin resistance and circulating hormone levels in postmenopausal women and may also lead to an increased risk of meningioma.

Acknowledgment We thank Sondra Buehler for her editorial assistance.

Funding This work was supported by the National Institutes of Health (R01 CA39742). Conflict of interest: None declared.

References 1.

Wiemels J, Wrensch M, Claus EB. Epidemiology and etiology of menin-

8.

Tilzer LL, Plapp FV, Evans JP, Stone D, Alward K. Steroid receptor pro-

9.

Benson VS, Pirie K, Green J, et al. Hormone replacement therapy and

teins in human meningiomas. Cancer. 1982;49(4):633 –636.

gioma. J Neurooncol. 2010;99(3):307 –314. 2.

Umansky F, Shoshan Y, Rosenthal G, Fraifeld S, Spektor S. Radiation-induced meningioma. Neurosurg Focus. 2008;24(5):E7.

3.

McKinney PA. Brain tumours: incidence, survival, and aetiology. Journal of neurology, neurosurgery, and psychiatry. 2004;75(Suppl 2:ii):12– 17.

4.

meningioma

primary glioma and meningioma tumours in the Million Women Study

2007;110(3):471 –476.

what

do

we

tell

our

patients?

Cancer.

11. Michaud DS, Gallo V, Schlehofer B, et al. Reproductive factors and exogenous hormone use in relation to risk of glioma and meningioma

Increased prevalence of obesity and obesity-related postoperative com-

in a large European cohort study. Cancer Epidemiol Biomarkers Prev.

in

male

patients

with

meningiomas.

Neurosurgery.

2010;19(10):2562 –2569. 12. Jhawar BS, Fuchs CS, Colditz GA, Stampfer MJ. Sex steroid hormone

2007;61(4):754 –760; discussion 760 –751. Korhonen K, Salminen T, Raitanen J, Auvinen A, Isola J, Haapasalo H.

exposures and risk for meningioma. J Neurosurg. 2003;99(5):848 – 853.

Female predominance in meningiomas can not be explained by differ-

13. Mills PK, Preston-Martin S, Annegers JF, Beeson WL, Phillips RL, Fraser

ences in progesterone, estrogen, or androgen receptor expression.

GE. Risk factors for tumors of the brain and cranial meninges in Seventh-Day Adventists. Neuroepidemiology. 1989;8(5):266 –275.

J Neurooncol. 2006;80(1):1 –7. 7.

risk:

Aghi MK, Eskandar EN, Carter BS, Curry WT, Jr, Barker FG, 2nd. plications

6.

Study. Int J Cancer. 2010;127(7):1692–1698. 10. Claus EB, Black PM, Bondy ML, et al. Exogenous hormone use and

Benson VS, Pirie K, Green J, Casabonne D, Beral V. Lifestyle factors and cohort. Br J Cancer. 2008;99(1):185 –190.

5.

incidence of central nervous system tumours in the Million Women

Cahill DW, Bashirelahi N, Solomon LW, Dalton T, Salcman M, Ducker TB.

Estrogen

and

progesterone

receptors

in

meningiomas.

NEURO-ONCOLOGY

abdominal obesity with multiple health outcomes in older women: the Iowa Women’s Health Study. Arch Intern Med. 2000;160(14):2117–2128.

J Neurosurg. 1984;60(5):985 –993.

1018

14. Folsom AR, Kushi LH, Anderson KE, et al. Associations of general and



SEPTEMBER 2011

Johnson et al.: Risk factors for meningioma in post-menopausal women

15. Bisgard KM, Folsom AR, Hong CP, Sellers TA. Mortality and cancer rates

25. Helseth A, Tretli S. Pre-morbid height and weight as risk factors for devel-

in nonrespondents to a prospective study of older women: 5-year

opment of central nervous system neoplasms. Neuroepidemiology. 1989;8(6):277 –282.

follow-up. Am J Epidemiol. 1994;139(10):990– 1000. 16. Kushi LH, Kaye SA, Folsom AR, Soler JT, Prineas RJ. Accuracy and

26. Tulinius H, Sigfusson N, Sigvaldason H, Bjarnadottir K, Tryggvadottir L. Risk

reliability of self-measurement of body girths. Am J Epidemiol.

factors for malignant diseases: a cohort study on a population of 22,946 Icelanders. Cancer Epidemiol Biomarkers Prev. 1997;6(11):863–873.

1988;128(4):740 – 748. 17. Kushi LH, Fee RM, Folsom AR, Mink PJ, Anderson KE, Sellers TA.

27. Lee M, Wrensch M, Miike R. Dietary and tobacco risk factors for adult

Physical activity and mortality in postmenopausal women. JAMA.

onset glioma in the San Francisco Bay Area (California, USA). Cancer Causes Control. 1997;8(1):13– 24.

1997;277(16):1287 –1292. 18. World Health Organization. Obesity: Preventing and Managing the

28. Gunnell D, Okasha M, Smith GD, Oliver SE, Sandhu J, Holly JM. Height,

Global Epidemic. WHO Obesity Technical Report Series 894. Geneva,

leg length, and cancer risk: a systematic review. Epidemiol Rev. 2001;23(2):313 –342.

Switzerland: World Health Organization; 2000. 19. Virnig B, Durham SB, Folsom AR, Cerhan J. Linking the Iowa Women’s Health Study cohort to Medicare data: linkage results and application to hip fracture. Am J Epidemiol. 2010;172(3):327 –333. cancer

related

health

services

research

using

cancer. Best Pract Res Clin Endocrinol Metab. 2002;16(2):225–241. 30. Heymsfield SB, Gallagher D, Mayer L, Beetsch J, Pietrobelli A. Scaling of

20. Potosky AL, Riley GF, Lubitz JD, Mentnech RM, Kessler LG. Potential for

29. Okasha M, Gunnell D, Holly J, Davey Smith G. Childhood growth and adult

a

linked

Medicare-tumor registry database. Med Care. 1993;31(8):732 – 748. 21. Korn EL, Graubard BI, Midthune D. Time-to-event analysis of longitudinal follow-up of a survey: choice of the time-scale. Am J Epidemiol. 1997;145(1):72 –80.

human body composition to stature: new insights into body mass index. Am J Clin Nutr. 2007;86(1):82 –91. 31. Witelson SF, Beresh H, Kigar DL. Intelligence and brain size in 100 postmortem brains: sex, lateralization and age factors. Brain. 2006;129(Pt 2):386 –398. 32. Neilson HK, Friedenreich CM, Brockton NT, Millikan RC. Physical

22. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham CL, Anis AH.

activity and postmenopausal breast cancer: proposed biologic mechan-

The incidence of co-morbidities related to obesity and overweight:

isms and areas for future research. Cancer Epidemiol Biomarkers Prev.

a systematic review and meta-analysis. BMC Public Health. 2009;9:88. 23. Lahmann PH, Hoffmann K, Allen N, et al. Body size and breast cancer risk: findings from the European Prospective Investigation into Cancer And Nutrition (EPIC). Int J Cancer. 2004;111(5):762 – 771.

2009;18(1):11 –27. 33. Mayer-Davis EJ, D’Agostino R, Jr, Karter AJ, et al. Intensity and amount of physical activity in relation to insulin sensitivity: the Insulin Resistance Atherosclerosis Study. JAMA. 1998;279(9):669 –674. 34. Wigertz A, Lonn S, Hall P, et al. Reproductive factors and risk of meningioma

24. McTiernan A, Wu L, Chen C, et al. Relation of BMI and physical activity

and glioma. Cancer Epidemiol Biomarkers Prev. 2008;17(10):2663–2670.

to sex hormones in postmenopausal women. Obesity (Silver Spring).

35. Okolo S. Incidence, aetiology and epidemiology of uterine fibroids. Best

2006;14(9):1662 – 1677.

Pract Res Clin Obstet Gynaecol. 2008;22(4):571 –588.

NEURO-ONCOLOGY



SEPTEMBER 2011

1019