Tumor Biology

, Volume 35, Issue 11, pp 11579–11586

Reproductive factors in relation to risk of brain tumors in women: an updated meta-analysis of 27 independent studies

  • Hailiang Zong
  • Hongsheng Xu
  • Zhongqun Geng
  • Chong Ma
  • Xing Ming
  • Ming Shang
  • Kai Li
  • Xiaoguang He
  • Hai Du
  • Jianping Zhao
  • Lei Cao
Research Article

DOI: 10.1007/s13277-014-2448-1

Cite this article as:
Zong, H., Xu, H., Geng, Z. et al. Tumor Biol. (2014) 35: 11579. doi:10.1007/s13277-014-2448-1

Abstract

Previous studies on the association between reproductive factors and brain tumor risk in women have provided inconclusive findings. Thus, an updated meta-analysis was performed to obtain more precise risk estimates for brain tumor regarding several common reproductive factors. A comprehensive literature search for relevant publications in the PubMed and Embase databases was carried out from their inception up to June 20, 2014. Pooled relative risks (RRs) with corresponding 95 % confidence intervals (CIs) were calculated. There were 27 independent studies with a total of 12,129 cases and 1,433,915 controls included into the present meta-analysis. We found that an elevated risk of brain tumors (RR = 1.17, 95 % CI 1.06–1.29, P = 0.002), particularly glioma (RR = 1.33, 95 % CI 1.15–1.54, P < 0.001), was related to older age at menarche. Interestingly, stratified analysis by type of brain tumors showed that the longer duration of breast feeding was associated with the risk of meningioma negatively but glioma positively (for meningioma: RR = 0.76, 95 % CI 0.64–0.91, P = 0.002; for glioma: RR = 1.70, 95 % CI 1.14–2.55, P = 0.010). No significant association was observed when estimating the roles of other reproductive factors including parity, age at first birth, menopausal status, and age at menopause in brain tumorigenesis. Our study suggests that older age at menarche is a risk factor of brain tumors and glioma in particular. Additionally, more studies are warranted to further elucidate roles and mechanisms of common reproductive factors in the risk of brain tumors.

Keywords

Reproductive factors Meningioma Glioma Meta-analysis 

Introduction

Brain tumors are in general highly fatal malignancies, and glioma and meningioma are the most common types of primary brain tumors in adults [1]. Although there is progress in the diagnosis and treatment of brain tumors, little is known about the etiology. Ionizing radiation is the only well-established environmental risk factor for brain tumors [2]. Other factors, for instance rare hereditary syndromes, immune suppression giving rise to brain lymphomas, and single nucleotide polymorphisms, are implicated in brain tumorigenesis [3, 4, 5].

The incidence of meningioma is about 2-fold higher in women than in men, while the incidence of glioma is about 1.5- or 2-fold higher in men than in women, suggesting gender bias in the risk of meningioma and glioma [6, 7]. Besides, increasing evidence supports that sex hormone receptors are expressed in meningioma and glioma, which implicates potential roles of sex hormones in the etiology of brain tumors [8, 9]. Sex hormones exposure may play critical roles in brain tumors development by affecting the cell proliferation and apoptosis in tumor cells [10]. A number of epidemiological studies have been performed to assess the effect of reproductive and exogenous hormone factors on brain tumor risk among different populations; however, the effect remains obscure due to inconsistent and inconclusive reports [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. Two previous meta-analyses by the same research group have investigated the influence of exogenous and endogenous hormones in relation to glioma and meningioma risk, respectively [31, 32]. Unfortunately, they did not estimate the role of reproductive factors in whole brain tumors. Besides, studies newly published were not included. Moreover, the modifying effects of certain common reproductive factors, such as breast feeding, were not estimated in those meta-analyses. To precisely determine the influence of common female reproductive factors in brain tumorigenesis, we performed an updated meta-analysis by pooling all available publications to date.

Materials and methods

Search strategy for eligible studies

Eligible studies on the association between reproductive factors and brain tumor risk were identified by literature search in databases of PubMed and Embase from their inception up to June 20, 2014. The following terms was used: brain tumors, brain neoplasms, brain cancer, central nervous system tumors, glioma, or meningioma; reproductive factors, menstrual factors, age at menarche, parity, age at first birth, menopausal status, age at menopause, breast feeding, pregnancy, gravidity, or reproductive history; and incidence, risk factor, or risk. No language restriction was imposed. The references of all retrieved studies were further screened for additional eligible articles. If studies were repeated, only the most recent or complete study was included into this meta-analysis.

Inclusion/exclusion criteria

Studies were included if they conformed to the following inclusion criteria: (1) studies on the association between reproductive factors and brain tumor risk, (2) studies in cohort or case-control design, (3) articles providing odds ratios (ORs), relative risks (RRs), or hazard ratios (HRs) with corresponding 95 % confidence intervals (CIs).

The exclusion criteria were as follows: (1) case-only study, (2) animal studies, (3) case reports, (4) studies not related to the association between reproductive factors and brain tumor risk, and (5) studies with duplicated data.

Data extraction

Two investigators carefully and independently extracted data from each available study. Disagreements on all terms were settled by discussion. The following terms were extracted: first author, publication year, study design, country of origins, characteristics of subjects, source of controls, sample size, study period, age of subjects, matching factor, adjusted factors, RRs or HRs or ORs with 95 % CIs, and type of brain tumors. We did not contact with any author for information for each study.

Statistical analysis

The relationship between reproductive factors and brain tumor risk was estimated by calculating pooled RRs with 95 % CIs. The between-study heterogeneity was evaluated using Cochran’s Q statistic test and I2 test [33, 34]. P < 0.05 and I2 > 50 % suggested significant heterogeneity between all studies, and the random effects model (the DerSimonian and Laird method) was adopted for combined analysis [35]; otherwise, the fixed effects model (the Mantel-Haenszel method) was applied for pooled analysis [36]. Stratified analysis by type of brain tumors was performed to evaluated roles of reproductive factors in brain tumorigenesis. Sensitivity analysis by omitting single study was also carried out for further assessment of association and source of heterogeneity. Potential publication bias was assessed by Begg’s funnel plot analysis and Egger’s test [37, 38]. All analyses were performed using the Stata 12.0 software (StataCorp, College Station, TX, USA). P < 0.05 implicated statistical significance.

Results

Characteristics of all available studies

As shown in Fig. 1, we initially identified 306 studies after a comprehensive literature in databases of PubMed and Embase. According to the inclusion criteria, 20 publications with 27 independent studies on the relationship between reproductive factors and brain tumor risk were finally included into this meta-analysis [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. A total of 12,129 cases and 1,433,915 controls were involved. Table 1 summarized the characteristics of all included studies. Of 27 studies, 13 were about the role of reproductive factors in meningioma risk and 14 were about the glioma risk. Based on the study design, 14 individual studies were in population-based case-control designs, 7 in prospective cohort designs, 3 in hospital-based case-control designs, 2 in nested case-control designs, and still 1 in a mixed design.
Fig. 1

Flow chart for eligible cohort studies

Table 1

Characteristics of all studies included into this study

Study

Year

Type of tumor

Origins

Age (years)

No. of cases/controls

Study design

Study period

Adjusted or matching factors

Anic GM (1)

2014

Meningioma

USA

19–90

247/695

PCC

NA

Frequency, race, residence, age, gender

Anic GM (2)

2014

Glioma

USA

18–90

507/695

PCC

NA

Frequency, race, residence, age, gender

Claus EB

2013

Meningioma

USA

29–79

1,127/1,109

PCC

2006–2011

Age, geography, frequency, race, education, smoking, alcohol use, BMI

Johnson DR

2011

Meningioma

USA

55–85.7

125/291,021

Prospective cohort study

1986–2004

Age

Kabat GC

2011

Glioma

USA

50–78

174/225,355

Prospective cohort study

1995–2001

Age, race, age at menarche, parous/nulliparous, age at menopause, history of hysterectomy, and smoking status

Wang SS

2011

Glioma

Duarte

NA

357/822

Mixed

1993–2001

NA

Korhonen K

2010

Meningioma

Finland

20–69

264/505

PCC

2000–2002

Age, residential area, family history with brain tumors

Michaud DS (1)

2010

Meningioma

Europe

20–83

194/276,212

Prospective cohort study

1992–2008

Smoking status, education, body mass index and menopausal status

Michaud DS (2)

2010

Glioma

Europe

20–83

193/276,212

Prospective cohort study

1992–2008

Smoking status, education, body mass index and menopausal status

Felini MJ

2009

Glioma

USA

≥20

619/650

PCC

1991–2004

Age, gender, ethnicity, frequency

Benson VS (1)

2008

Meningioma

UK

50–65

390/124,967

Prospective cohort study

1996–2005

Age, height, BMI, strenuous exercise, socioeconomic level, smoking, alcohol intake

Benson VS (2)

2008

Glioma

UK

50–65

646/124,967

Prospective cohort study

1996–2005

Age, height, BMI, strenuous exercise, socioeconomic level, smoking, alcohol intake

Wigertz A (1)

2008

Meningioma

Europe

18–69

906/1,774

PCC

2000–2004

Age and geographic region

Wigertz A (2)

2008

Glioma

Europe

18–69

626/1,774

PCC

2000–2004

Age and geographic region

Custer B

2006

Meningioma

USA

≥18

143/286

PCC

1995–1998

Age, education, smoking history, alcohol consumption, BMI, age at menarche, parity

Lee E

2006

Meningioma

USA

NA

219/260

HCC

1987–1992

Age, ethnicity, hospital, smoking, pregnancy, thyroid disorders, radiation treatment, menopause

Silvera SA

2006

Glioma

USA

40–75.4

120/89,835

Prospective cohort study

1980–2000

Age, study center, randomization group, age at menarche, parity, menopausal status, age at first live birth

Hatch EE (1)

2005

Meningioma

USA

18–90

151/436

HCC

1994–1998

Age, ethnicity, hospital, distance of residence from hospital, marital status, education

Hatch EE (2)

2005

Glioma

USA

18–90

212/436

HCC

1994–1998

Age, ethnicity, hospital, distance of residence from hospital, marital status, education

Huang K

2004

Glioma

USA

18–80

341/527

PCC

1995–1997

Age, menopausal status

Schlehofer B (1)

1999

Meningioma

Mixed

20–80

331/1,123

PCC

1980–1991

Age, ethnicity, residential area, menopausal status

Schlehofer B (2)

1999

Glioma

Mixed

20–80

1,178/1,987

PCC

1980–1991

Age, ethnicity, residential area, menopausal status

Cicuttini FM

1997

Glioma

Australia

20–70

166/170

PCC

1987–1991

Menopausal status

Lambe M (1)

1997

Meningioma

Sweden

15–65

1,088/5,440

NCC

1958–1990

Age

Lambe M (2)

1997

Glioma

Sweden

15–65

1,657/8,285

NCC

1958–1990

Age

Preston-Martin S

1995

Meningioma

USA

20–74

81/155

PCC

1978–1985

Age

Hochberg F

1990

Glioma

USA

15–81

67/59

PCC

1977–1981

Age, age at menarche

NA not available, PCC population-based case-control studies, NCC nested case-control studies, HCC hospital-based case-control studies

Association between reproductive factors and brain tumor risk

The pooled RRs revealed that older age at menarche was related to an increased risk of brain tumors (RR = 1.17, 95 % CI 1.06–1.29, P = 0.002; I2 = 28.4 %, P for heterogeneity = 0.116) (Table 2, Fig. 2). No significant association was observed when estimating the roles of other reproductive factors including parity, age at first birth, menopausal status, breast feeding, and age at menopause in brain tumorigenesis (Table 2). Sensitivity analysis by sequential omission of each eligible study confirmed the findings previously mentioned (data not shown).
Table 2

Summary results of meta-analysis

Contrasts

No. of studies

RR (95 % CI)

P valuea

Tests for heterogeneity

I2 (%)

Pb

Brain tumors

 Age at menarche

20

1.17 (1.06–1.29)

0.002

28.4

0.116

 Parity

19

0.97 (0.82–1.14)

0.702

56.4

0.001

 Age at first birth

21

0.97 (0.88–1.08)

0.636

0.0

0.635

 Menopausal status

17

1.08 (0.92–1.27)

0.352

51.3

0.008

 Age at menopause

13

1.05 (0.90–1.22)

0.526

0.0

0.747

 Breast feeding

6

0.94 (0.68–1.31)

0.724

67.1

0.010

Meningioma

 Age at menarche

10

1.04 (0.91–1.19)

0.545

0.0

0.854

 Parity

10

1.09 (0.84–1.42)

0.522

64.8

0.002

 Age at first birth

11

0.89 (0.76–1.04)

0.131

0.0

0.584

 Menopausal status

9

1.13 (0.89–1.44)

0.323

58.2

0.014

 Age at menopause

6

1.03 (0.82–1.29)

0.816

5.5

0.382

 Breast feeding

4

0.76 (0.64–0.91)

0.002

0.0

0.919

Glioma

 Age at menarche

10

1.33 (1.15–1.54)

<0.001

43.3

0.070

 Parity

9

0.89 (0.77–1.03)

0.113

31.6

0.165

 Age at first birth

10

1.05 (0.92–1.21)

0.465

0.0

0.717

 Menopausal status

8

1.01 (0.86–1.19)

0.866

47.9

0.062

 Age at menopause

7

1.06 (0.88–1.29)

0.527

0.0

0.793

 Breast feeding

2

1.70 (1.14–2.55)

0.010

45.0

0.178

RR relative risk, CI confidence interval

aP value for the association between reproductive factors and brain tumor risk

bP value for heterogeneity analysis

Fig. 2

Forest plots of hip brain tumor risk associated with age at menarche

Association between reproductive factors and meningioma risk

Totally, 13 individual studies involving 5,266 cases and 703,274 controls were included to investigate the effect of reproductive factors on meningioma risk. Reproductive factors including age at menarche, parity, age at first birth, menopausal status, and age at menopause did not alter the risk of meningioma (Table 2). However, a negative association between breast feeding and risk of meningioma was found (RR = 0.76, 95 % CI 0.64–0.91, P = 0.002; I2 = 0.0 %, P for heterogeneity = 0.919) (Table 2, Fig. 3). Sensitivity analysis did not materially modify the findings (data not shown).
Fig. 3

Forest plots of hip brain tumor risk associated with breastfeeding

Association between reproductive factors and glioma risk

There were 14 independent studies with a total of 6,863 cases and 730,641 controls, estimating the relationship of reproductive factors with glioma risk. An increased risk of glioma was found to be associated with older age at menarche (RR = 1.33, 95 % CI 1.15–1.54, P < 0.001; I2 = 43.3 %, P for heterogeneity = 0.070) (Table 2, Fig. 2). Interestingly, a positive association between breast feeding and risk of glioma was found (RR = 1.70, 95 % CI 1.14–2.55, P = 0.010; I2 = 45.0 %, P for heterogeneity = 0.178) (Table 2, Fig. 3). No significant correlation was demonstrated concerning other reproductive factors (Table 2). Sensitivity analysis confirmed the pooled results (data not shown).

Publication bias

There was no publication bias in this study, as suggested by both Begg’s funnel plots analysis and Egger’s test (data not shown).

Discussion

The updated meta-analysis of 27 independent studies with 12,129 cases and 1,433,915 controls suggests that older age at menarche is a risk factor of brain tumors and glioma in particular. Interestingly, longer duration of breast feeding was negatively associated with meningioma risk but positively related to glioma risk, as suggested by the pooled RRs (for meningioma: RR = 0.76, 95 % CI 0.64–0.91, P = 0.002; for glioma: RR = 1.70, 95 % CI 1.14–2.55, P = 0.010). It could be concluded that the longer duration of breast feeding was a protective factor for meningioma but a risk factor for glioma. Additionally, no modifying effect of other reproductive factors including parity, age at first birth, menopausal status, and age at menopause was observed in relation to the risk of brain tumors.

Meningiomas are mainly benign tumors derived from meningothelial cells of the arachnoid membrane [39]. Although meningiomas are generally encapsulated and grow slowly, they develop extensively. The etiology of meningioma remains largely unknown. It has been shown that the presence of progesterone receptors is a favorable prognostic factor for meningioma [40, 41]. Besides, progesterone together with estrogens can promote the proliferation of meningioma cells [42, 43]. Taken together, sex hormones may play a crucial role in meningioma development and progression. Given significant gender bias in the incidence of meningioma between men and women, many epidemiological studies have evaluated roles of female reproductive factors in the development of meningioma [6, 11, 15, 16, 19, 31]. A previous meta-analysis by Qi et al. suggested that exposure to female sex hormones, menopausal status, and parity was associated with an elevated risk of meningioma [31]. Besides, no such association was observed when evaluating roles of oral contraceptive use, age at menarche, age at menopause, and age at first birth in meningioma risk [31]. On the contrary, there was no appreciable association between parity and meningioma risk demonstrated by Anic et al. [11]. Additionally, parity, age at first birth, and oral contraceptive use were found to exert no effects on the risk of meningioma [20]. Interestingly, the present meta-analysis showed that older age at menarche was related to an increased risk of total brain tumors, but not meningioma. Moreover, we found that longer breast feeding played a protective role in the development of meningioma. The conflicting findings might be attributed to different study designs, sample size, matching criteria, and adjusted factors. More studies are warranted to shed some light on those contradictory findings.

Glioma is the most common type of central nervous system tumor, which represents more than 80 % of adult brain tumors [44]. High dose of ionizing radiation exposure is a well-known risk factor for glioma. Nonetheless, not all exposed individuals develop glioma, implicating that other factors, for example female reproductive factors, are involved in glioma pathogenesis. A recent case-control study showed the evidence that older age at menarche was associated with a higher risk of glioma, with a stronger association observed in premenopausal than postmenopausal women [11]. Increasing parity was not related to the risk of glioma [11]. Similarly, Benson et al. demonstrated that parity, age at first birth, and oral contraceptive use were not associated with the risk of glioma [20]. Reversely, oral contraceptive use reduced the risk of glioma in females, which suggested a protective role of oral contraceptive use in the development of glioma [13]. Nonetheless, current users of oral contraceptives were at an increased risk of meningioma demonstrated by Michaud et al. [16]. Taken together, findings among independent studies are inconsistent and inconclusive. Besides, individual studies have insufficient statistical power to detect significant association due to a relative small number of subjects. Meta-analysis with large sample size can provide more precise estimates. The meta-analysis by Qi et al. was based on 11 eligible studies with a total of 4,860 cases and 14,740 controls [32]. They demonstrated that older age at menarche but not menopause status, parous status, age at menopause, or age at first birth could confer a risk effect on glioma [32]. Nevertheless, several eligible publications were not included into this meta-analysis. In our meta-analysis of 14 included studies with 6,863 cases and 730,641 controls, we found an increased risk of glioma related to older age at menarche and longer duration of breast feeding. In addition, the risk of glioma did not seem to be influenced by parity, age at first birth, menopausal status, and age at menopause, as suggested by the pooled results.

The role of breast feeding in the development of brain tumors remains to be not fully understood. Michaud et al. reported that the duration of breast feeding did not modify the risk of glioma and meningioma [16]. Similar findings were demonstrated in another case-control study [24]. On the contrary, parous women with breast feeding for more than 36 months were at an increased risk of glioma compared with women with breast feeding less than 3 months [19]. In this meta-analysis, we found that the longer duration of breast feeding was associated with the risk of meningioma negatively and glioma positively. The effect of breast feeding duration on glioma and meningioma risk seemed to be opposite. It was worthy to be mentioned that only two independent studies with 838 cases and 2,210 controls were included into our study investigating the role of breast feeding in glioma risk. Thus, explanations for this discrepancy and underlying mechanisms need to be elucidated by more studies.

Some limitations should be considered when interpreting the pooled results in the present meta-analysis. First, stratified analyses by study design, ethnicity, smoking status, etc., could be performed to estimate their influence in the risk of glioma and meningioma. We failed to do this for a lack of sufficient eligible data. Second, no significant association of common reproductive factors with glioblastoma risk was observed, including age at first live birth, parity, age at menopause, type of menopause, and exogenous hormone use [14]. The effect of female reproductive factors on brain tumorigenesis might differ with different histological types and tumor stages of brain tumors. More future studies are warranted for further investigation. Third, an interaction between parity and breast feeding was observed in the risk of glioma in a previous study [19]. Combined effects of reproductive factors in brain tumorigenesis should be fully elucidated by more studies with high quality, which may provide a better understanding of the etiology of brain tumors.

In conclusion, our study shows that older age at menarche is a risk factor of brain tumors, particularly glioma. The effect of breast feeding duration on brain tumor risk needs further elucidation by increasing the sample size and considering its interaction with other reproductive factors, such as parity, age at first birth, menopausal status, and age at menopause.

Conflicts of interest

None

Copyright information

© International Society of Oncology and BioMarkers (ISOBM) 2014

Authors and Affiliations

  • Hailiang Zong
    • 1
  • Hongsheng Xu
    • 1
  • Zhongqun Geng
    • 2
  • Chong Ma
    • 1
  • Xing Ming
    • 1
  • Ming Shang
    • 1
  • Kai Li
    • 1
  • Xiaoguang He
    • 1
  • Hai Du
    • 1
  • Jianping Zhao
    • 1
  • Lei Cao
    • 1
  1. 1.Department of Neurosurgery, Central Hospital of XuzhouAffiliated Hospital of Southeast UniversityXuzhouChina
  2. 2.Department of Gynecology and ObstetricsThe Xuzhou Maternity and Children’s Health Care HospitalXuzhouChina

Personalised recommendations