Does exogenous hormonal therapy affect the risk of glioma among females: A systematic review and meta-analysis

Abstract Background The effect of exogenous hormone replacement therapy (HRT) and oral contraceptive pills (OCPs) on glioma risk in females is unclear despite numerous studies; hence, we conducted a meta-analysis to evaluate this relationship. Methods Studies investigating the impact of exogenous female hormones on glioma risk were retrieved by searching 4 databases from inception until September 2022. Articles of any design, such as case–control and cohort studies, proving the relative risk (RR), odds ratio (OR), or hazard ratio were included. Summary OR values were calculated using a random effects model. Results Both HRT and OCP use of any duration decreased the risk of developing glioma [HRT OR = 0.78, 95% CI 0.66–0.91, P = .00; OCP: OR = 0.80, 95% CI 0.67–0.96, P = .02]. When stratified by duration of use, HRT use >1 year significantly reduced glioma risk (<1 year: OR = 0.82, 95% CI 0.63–1.07, P = 0.15; 1–5 years: OR = 0.79, 95% CI 0.67–0.92, P = .00; 5–10 years: OR = 0.80, 95% CI 0.66–0.97, P = .02; >10 years: OR = 0.69, 95% CI 0.54–0.88, P = .00). In contrast, only OCP use for >10 years significantly reduced glioma risk (<1 year: OR = 0.72, 95% CI 0.49–1.05, P = .09; 1–5 years: OR = 0.88, 95% CI 0.72–1.02, P = .09; 5–10 years: OR = 0.85, 95% CI 0.65–1.1, P = 0.21; >10 years: OR = 0.58, 95% CI 0.45–0.74, P = .00). Conclusions Our pooled results strongly suggest that sustained HRT and OCP use is associated with reduced risk of glioma development.

ionizing radiation and hereditary syndromes such as neurofibromatosis 1 and 2, tuberous sclerosis, Lynch syndrome, and von Hippel-Lindau syndrome. 2The incidence of glioma is also higher among males, suggesting that development may be influenced by hormones. 3Consistent with this notion, glioma cells express steroid hormone receptors 4 and factors such as duration of exogenous hormone use, age at first childbirth, number of births, age at menarche, age at menopause, and type of menopause (natural or medically induced), and duration of hormone alter glioma incidence. 5There are many important indications for hormone replacement therapy (HRT), including treatment of menopause symptoms and prevention of cardiovascular disease or osteoporosis. 6Hot flashes and urogenital atrophy are common examples of postmenopausal symptoms that are frequently managed by HRT. 7 It was reported that 44% of postmenopausal females have used HRT at least once, most often in pill form (40%). 8 While numerous studies have addressed the effects of HRT and oral contraceptive pills (OCPs) on glioma risk, many of the results are contradictory.For instance, Benson et al. reported an increased risk of developing glioma and meningioma, 9 while Yang et al. found that risks of glioma and meningioma were dependent on the duration of OCP use. 10 Others have found that factors such as old age at menarche increase the risk of developing glioma. 11,12onversely, Lan et al. reported that HRT reduced the risk of developing glioma, although they did not stratify by duration of use. 13In this meta-analysis, we examined the relationship between glioma risk and the use of HRT or OCP with duration of use stratification.

Search Strategy
This study was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. 14Studies on the effects of HRT on glioma risk in females were retrieved by searching Medline, Cochrane, Embase, and CENTRAL as well as the references lists of included papers and previous meta-analyses.Searches were conducted in September 2022 and were restricted to English language literature.The search string used for all databases was as follows: [(Brain Glioma OR high-grade tumor OR glial cell neoplasm OR glioblastoma multiforme OR GBM OR diffuse glioma OR glial tumors OR anaplastic glioma) AND (hormone replacement therapy OR contraceptives OR exogenous hormones OR exogenous estrogen OR estrogen OR HRT OR OCP) AND (risk OR health risk assessment OR risk factor)].

Study Selection
Inclusion criteria were: (i) studies describing the relationship between glioma incidence and current or past use of female exogenous hormones using a case-control or cohort study design, and (ii) providing the relative risk (RR), odds ratio (OR), or hazard ratio.No randomized controlled trials were identified through our search.Retrieved studies conducted in animal models, presented as conference abstracts, that did not classify CNS tumor subtypes or did not include glioma as the outcome of interest were excluded.In addition, reviews and previous meta-analyses were excluded.Two groups of authors independently performed the primary survey according to our preset inclusion criteria, and conflicts were resolved by senior authors through discussion and consensus.

Data Extraction and Quality Assessment
The following parameters were extracted from each study and entered into an Excel sheet: first author, year of publication, country where the study was conducted, mean or median age, sample size, study design, follow-up duration, exposure (HRT, OCPs, or both), risk estimate, duration of use, and Newcastle-Ottawa Scale (NOS).The data were then reviewed by a third author.
Study quality was assessed using the NOS, a wellvalidated metric for evaluating observational and nonrandomized studies according to participant selection criteria, comparability, and exposure or outcome.Comparability points were given whenever the age at glioma diagnosis and duration of hormone use were available.Additionally, the adequacy of the follow-up duration was determined by the senior authors.The NOS score ranges from 0 to 9 stars, and studies with≥6 stars are considered to be of relatively higher quality. 15We searched for the source of funding and reported it as yes (provided), no (not provided), or not mentioned (Table 1).

Analysis
Descriptive statistics, including mean and frequency, were calculated using IBM SPSS version 2, while the meta-analysis was conducted using Comprehensive Meta-Analysis software version 3. Summary ORs) and RR with 95% CI of developing glioma were calculated separately.Due to the rarity of glioma, ORs were considered equivalent to RRs.For simplicity, therefore, pooled results are expressed as ORs.The influences of oral contraceptives and HRT on glioma risk were also examined separately.Additional subgroup analyses were performed on treatment groups stratified by duration of use (when

Importance of the Study
This updated meta-analysis and systematic review reveals a significant association between hormonal therapy and reduced glioma risk among adult females.
These findings may warrant further evaluation of the role of female hormones as preventative therapies for glioma.CC = case-control, C = cohort, NOS = New castle-Ottawa scale.

Alfuridy et al.: Exogenous hormonal therapy and glioma risk
available) as follows: <1 year, 1-5 years, 5-10 years, and >10 years.A study design influences the risk of bias, this assessment was conducted separately for case-control and cohort studies.The possibility of heterogeneity was evaluated using the I-squared statistic, with <25% considered low, 25-50% moderate, and 50-75% as high heterogeneity.Due to the heterogeneity among studies, a random effects model was for pooled analysis.Sensitivity analysis was performed by omitting 1 study at a time and assessing the stability of the result and by omitting studies with NOS scores less than 6.Publication bias was assessed using Begg's funnel plot and Egger's test.

Search Results and Study Characteristics
A total of 386 studies were retrieved from Midline, Cochrane, Embase, or CENTRAL using the indicated search string.Among these, 12 were excluded as duplicates and 356 due to irrelevance after reviewing the title and abstract.The full texts of the remaining 18 studies were carefully examined, and 5 was excluded as reviews.However, 4 studies found by searching the reference lists of included studies (n = 2) and previous meta-analysis (n = 2) were included.Finally, 17 valid observational studies were enrolled, 12 population-based case-control stud ies 5,9,[16][17][18][19][20][21]26,27,29,30 and 5 cohort studies [22][23][24][25]28 (Figure 1). The basic fatures of the enrolled studies are summarized in Table 1. Among th 17 observational studies included, 4 examined the effect of OCPs on glioma risk, 3 examined the effect of HRT, and 10 examined the effects of both HRT and OCPs.

Descriptive Statistics and Participant Demographics
The secondary aim of this study was to provide updated descriptive statistics on glioma and associations with OCP and HRT use.The 17 studies included in this metaanalysis were conducted in 5 different countries, of which the United States of America was the site of the greatest number.Most studies were conducted between 1990 and 2015 (inclusive) and included a total of 2 995 082 glioma cases.The median patient age was 52.More than or less than9.063years, and the mean duration of follow-up was 8.76 ± 4.433 years.

Quantitative Synthesis
The primary aim of this study was to provide updated estimates of glioma risk among females using OCPs or receiving HRT.
HRT and Glioma Risk.
OCPs and Glioma Risk.

Quality Assessment and Bias
Risk of Bias.
Quality assessment was conducted using the NOS scale.Two studies were given a score of 5 stars, 4 studies a score of 6 stars, 5 a score of 7 stars, and the rest a score of 8 stars (all out of 9).Based on a score of 6 or higher, 15 studies (88%) were classified as high quality.
Sensitivity Analysis and Publication Bias.
-Omitting each study separately yielded no significant changes in OR, indicating that the results were stable and robust.Construction of a Begg's funnel plot and Egger's test also yielded no evidence of publication bias (Figure 2B and D).We also examined the effect of omitting the 2 studies with high risk of bias (NOS scores of 5), one a casecontrol study on the effects of HRT and one a case-control study examining the effects of OCPs on glioma risk, 16,17 but again significant protection was maintained (OR = 0.76, 95% CI 0.63-0.91,P = .000,I 2 = 55.54 and OR = 0.72, 95% CI 0.65-0.80,P = .000,I 2 = 00.00,respectively).

Discussion
This updated meta-analysis aimed to determine the effects of HRT and OCP on glioma risk among adult females.
The pooled dataset included 12 case-control and 5 cohort studies with an overall total of 2 995 082 glioma patients.1][12][13] Similarly, HRT reduced the risk of developing glioma, also consistent with previous studies, 11,13 but this protective effect required only 1 year or more of treatment.Further, sensitivity analysis in which studies with NOS score < 6 were removed   (leaving only studies deemed high quality) yielded qualitatively similar results.Additional subgroup analysis revealed that the protective effects of both treatments were only significant in case-control studies.However, it is well known that case-control studies carry a higher risk of bias due to potential improper control group selection,

Alfuridy et al.: Exogenous hormonal therapy and glioma risk
especially for rare diseases such as glioma.For instance, using interviews or registries to identify participants with equivalent exposure can be a challenge, and in some of these case-control studies, exposure risk was taken from a proxy interviewer due to death or disability.Therefore, caution is warranted in interpreting these results, and future large-scale prospective studies are essential for confirmation.
The protective effect of HRT against glioma development is likely related to direct hormonal effects as glioma cells express steroid hormone receptors.However, Benson et al. found an increased risk of glioma among patients receiving HRT for any length of time (ever use subgroup). 9his contradictory finding suggests that the relationship between HRT and glioma is influenced by other factors, such as the timing, dose, type, and duration of HRT, and possibly also by individual differences in hormone metabolism.Anderson et al. also reported a significant increase in glioma risk among OCP users, particularly females taking progesterone-only therapy, and this enhanced risk was specific for glioblastoma multiforme, the most aggressive and deadly form of glioma. 16Several potential confounders may account for these discrepancies.Progesterone-only pills are usually prescribed for overweight women, and obesity alone has been identified as a risk factor for CNS tumors. 17Further, data on OCP were collected from a prescription registry initiated in 1995, and so may exclude longer-term use by older females (i.e. the sample included a disproportionate number of females <50 years old). 16Therefore, this result may not be applicable to older females.In fact, Hatch et al. found that OCPs reduced overall glioma risk, but stratification by age at diagnosis based on a cutoff of 50 years revealed that the protective effect was significant only in the older age group, possibly because older patients are more likely to have used more potent preparations before the 1970s. 18ormone replacement therapy is prescribed more often for females with higher education and socioeconomic status.For instance, Hatch et al. found that HRT cases were better educated than controls. 18Similarly, Felini et al. found a greater number of low-income participants among controls in their study, although there were equal numbers of high-income earners among cases and controls. 26owever, no stratified analysis based on income was conducted in either study.Alternatively, Benson et al. found that socioeconomic status had no effect on CNS tumor incidence, including glioma and meningioma incidence. 28onetheless, we acknowledge that an association between HRT and income or education could influence glioma incidence and thus should be included in future studies.
A previous meta-analysis by Zong et al. also found that older age at menarche was associated with a higher risk of brain tumors and glioma in particular.In addition, a longer duration of breastfeeding was associated with higher glioma risk, although with lower meningioma risk.In contrast, other reproductive factors such as menopausal status, parity, age at first birth, and age at menopause exhibited no significant association. 12 The meta-analysis by Benson et al. also examined the influence of HRT type on CNS tumor risk and found enhanced risk among estrogenonly users, amounting to an absolute excess risk of 2/10 000 users over 5 years, while no difference in risk was found for estrogen-progesterone users. 9Therefore, the HRT type should also be included in future studies.
The associations of HRT and OCP exposure with lower glioma incidence both became stronger as the duration of use increased, but significant protection required only 1 year for HRT but 10 years for OCPs.These findings are in partial accord with the results of Yang et al., who found that only OCPs used for 7.5 years or more substantially reduced the risk of glioma. 10This difference in the effect of treatment duration between OCPs and HRT may be explained by age, as OCPs are used by premenopausal females while HRT tends to be prescribed for older females already at increased risk of glioma.
One important factor missing from some of the included studies was the particular type of glioma.This lack of specificity is concerning because glioma types may be differentially sensitive to OCP exposure.This gap may lead to false perceptions regarding risks for specific glioma types.However, gliomas are rare tumors, so stratification according to type is challenging.Other limitations of this meta-analysis include the absence of age stratification in some studies.While the majority of studies found reduced glioma risk among exogenous hormone users, especially after prolonged use, the pooled result is inconsistent with some individual studies.Thus, larger-scale prospective studies considering possible confounders such as age at menarche, age at menopause, parity, breastfeeding history, age during treatment, hormone type(s), and dose among others are required to establish more accurate associations with glioma risk.
A funnel plot revealed no signs of publication bias.However, publication bias is a potential limitation of all meta-analyses as it is well known that negative results are often not published.Finally, the source of funding can be a potential source of bias, and 2 studies did not mention the source of funding.

Conclusion
This meta-analysis suggests an association between HRT for at least 1 year and OCP for at least 10 years and a reduction in the overall risk of glioma among adult females.However, additional research is needed to elucidate the mechanisms underlying this protective effect.Such information could help in the development of therapeutic applications for the prevention or treatment of glioma.

Figure 1 .
Figure 1.PRISMA flow chart for the search strategy.

Figure 2 .
Figure 2. (A) Forest plots for the OR of developing Glioma after HRT regardless of the duration of use, (B) funnel plot for HRT use and glioma, (C) forest plots for the OR of developing glioma after OCP regardless of the duration of use, (D) funnel plot for OCP use.

Figure 4 .
Figure 4. Forest plots for the OR of developing Glioma after HRT regardless of the duration of use, stratified by study type, C = cohort study, CC = case-control study.

Table 1 .
Summary of Included Studies