Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Nov 1.
Published in final edited form as: Clin Gastroenterol Hepatol. 2019 Nov 19;18(12):2701–2709.e3. doi: 10.1016/j.cgh.2019.11.030

Association Between Levels of Hormones and Risk of Esophageal Adenocarcinoma and Barrett’s Esophagus

Shao-Hua Xie *,a, Rui Fang ‡,a, Mingtao Huang ‡, Juncheng Dai ‡,§, Aaron P Thrift ∥,, Lesley A Anderson #, Wong-Ho Chow **, Leslie Bernstein ‡‡, Marilie D Gammon §§, Harvey A Risch ∥∥, Nicholas J Shaheen ¶¶, Brian J Reid ##, Anna H Wu ***, Prasad G Iyer ‡‡‡, Geoffrey Liu §§§, Douglas A Corley ∥∥∥, David C Whiteman ¶¶¶, Carlos Caldas ###, Paul D Pharoah ****,‡‡‡‡, Laura J Hardie §§§§, Rebecca C Fitzgerald ∥∥∥∥, Hongbing Shen ‡,§, Thomas L Vaughan ##, Jesper Lagergren *,¶¶¶¶
PMCID: PMC7580878  NIHMSID: NIHMS1622527  PMID: 31756444

Abstract

BACKGROUND & AIMS

Esophageal adenocarcinoma (EAC) occurs most frequently in men. We performed a Mendelian randomization analysis to investigate whether genetic factors that regulate levels of sex hormones are associated with risk of EAC or Barrett’s esophagus (BE).

METHODS

We conducted a Mendelian randomization analysis using data from patients with EAC (n = 2488) or BE (n = 3247) and control participants (n = 2127), included in international consortia of genome-wide association studies in Australia, Europe, and North America. Genetic risk scores or single-nucleotide variants were used as instrumental variables for 9 specific sex hormones. Logistic regression provided odds ratios (ORs) with 95% CIs.

RESULTS

Higher genetically predicted levels of follicle-stimulating hormones were associated with increased risks of EAC and/or BE in men (OR, 1.14 per allele increase; 95% CI, 1.01–1.27) and in women (OR, 1.28; 95% CI, 1.03–1.59). Higher predicted levels of luteinizing hormone were associated with a decreased risk of EAC in men (OR, 0.92 per SD increase; 95% CI, 0.87–0.99) and in women (OR, 0.93; 95% CI, 0.79–1.09), and decreased risks of BE (OR, 0.88; 95% CI, 0.770.99) and EAC and/or BE (OR, 0.89; 95% CI, 0.79–1.00) in women. We found no clear associations for other hormones studied, including sex hormone-binding globulin, dehydroepiandrosterone sulfate, testosterone, dihydrotestosterone, estradiol, progesterone, or free androgen index.

CONCLUSIONS

In a Mendelian randomization analysis of data from patients with EAC or BE, we found an association between genetically predicted levels of follicle-stimulating and luteinizing hormones and risk of BE and EAC.

Keywords: Esophageal Neoplasms, Sex Difference, Gonadal Steroid Hormones, Causality

Esophageal adenocarcinoma (EAC) and its precursor lesion Barrett’s esophagus (BE) are characterized by a strong male predominance, with male-to-female ratios of EAC incidence of 6-to-1 on average in Western countries and up to 8-to-1 in the United States.1–3 The reasons for this striking sex difference are not known, and do not seem to be explained by the 2 major risk factors of EAC and BE of gastroesophageal reflux disease and obesity, given the similar exposure prevalence and strengths of associations with EAC and BE risk between the sexes.1 Abdominal obesity, which is more common in men than in women, may contribute to the male predominance in EAC and BE.1,4 However, in a nationwide Swedish study, the male predominance in EAC was no weaker among lean individuals compared with overweight individuals, arguing against obesity as a factor completely explaining the excess male risk.5 The male predominance in EAC may be attributable to certain biological differences between the sexes. In particular, it has been hypothesized that sex hormonal and reproductive factors may play a role in the etiology of EAC and BE (ie, estrogenic exposures may prevent EAC development, whereas androgens may increase EAC risk). This hypothesis is supported by a 16-year delayed onset of EAC in women than in men.6 Recently, it was proposed that a more rapid age-related immune system decline in males may explain the generally higher cancer risk in males than in females,7 which may be driven by sex hormones.8 However, the existing epidemiologic evidence regarding the role of sex hormone in the development of EAC or BE remains inconclusive.1,2 Recent observational studies have suggested associations between circulating sex hormone levels and the risk of EAC or BE,9–12 but because of possible confounding and other biases inherent in observational studies, no causal relation has been established.

Mendelian randomization analysis provides a useful tool for exploring casual effects of endogenous exposures on disease risk without adding any intervention.13 Inheriting a genetic variant, determined by the random assortment of genes at conception, associated with lifelong changes in endogenous sex hormone levels, can confer altered risk for EAC or BE, which are not confounded by the known risk factors for these diseases. Therefore, the use of genetically predicted sex hormone levels as instrumental variables, based on established sex hormone-associated genetic variants, can facilitate causal inferences about the relationship between sex hormone levels and the risk of EAC or BE.

To test the hypothesis that genetically determined endogenous sex hormone levels influence the risk of EAC and BE, we performed a Mendelian randomization analysis based on merged data from several large genome-wide association studies (GWAS) conducted in Australia, Europe, and North America.

Methods

Study Participants

We analyzed GWAS data from participants in studies included in 3 consortia. First, the Barrett’s and Esophageal Adenocarcinoma Genetic Susceptibility Study within the Barrett’s and Esophageal Adenocarcinoma Consortium (http://beacon.tlvnet.net/), which included 1516 patients with histologically confirmed EAC, 2416 patients with BE, and 2187 control participants from 14 population-based, case-control and cohort studies conducted in Australia, Europe, and North America.14 Second, the Barrett’s Oesophagus Gene Study in the United Kingdom, which included 882 BE patients who were identified at endoscopy and confirmed with histopathology.15 Third, the Stomach and Oesophageal Cancer Study in the United Kingdom, which included 1003 EAC patients with International Classification of Disease coding of esophageal cancer (C15) and a pathologic diagnosis of adenocarcinoma (M8140–8575).15

After GWAS data cleaning, quality control, and imputation procedures, the current study included 2488 EAC patients, 3247 BE patients, and 2127 control participants. The distribution of participants by study is shown in Supplementary Table 1. The individual studies included in this analysis were approved by institutional review boards or research ethics committees. Informed consent was obtained from each participant.

Genotyping and Imputation

Genotyping of DNA from buffy coat or whole-blood samples was performed using the Illumina Omni1M Quad platform (San Diego, CA) in accordance with standard quality-control procedures.15,16 The annotations were based on version H of the Illumina product files and corresponded to the Genome Reference Consortium GRCh37 release. For quality control, genotyped single-nucleotide polymorphisms (SNPs) or samples with a call rate less than 95% were excluded. Based on control participants, SNPs with Hardy-Weinberg equilibrium P values less than 10−4 or minor allele frequency less than 0.01 also were excluded. Imputation was conducted at the study level, based on SHAPE1T2/1MPUTE2 using 1000 Genomes Phase 3 integrated variant set release in NCB1 build 37 (hg19) coordinates.17 Postimputation quality control excluded SNPs with an 1MPUTE2 information score less than 0.8, call rate less than 95%, Hardy-Weinberg equilibrium P value less than 10−4 based on control participants, or minor allele frequency less than 0.01 in control participants.

Genetic Risk Scores or Genetic Variants of Sex Hormones

SNPs associated with sex hormones at the GWAS significance level (P < 5 × 10−8) in populations of European descent were identified from published GWAS indexed in the NHGR1-EB1 GWAS catalog (http://www.ebi.ac.uk/gwas). SNPs predicting the levels of the following 9 sex hormones were found: sex hormone-binding globulin,18–20 dehydroepiandrosterone sulfate,20,21 testosterone,22,23 dihydrotestosterone,23 estradiol,20 follicle-stimulating hormone (FSH),20 luteinizing hormone (LH),20,24 progesterone,20 and free androgen index.20 Multiple SNPs were identified for each of the following 6 hormone measures: sex hormone-binding globulin, dehydroepiandrosterone sulfate, testosterone, dihydrotestosterone, LH, and progesterone. For these 6 hormones, sex hormone-specific genetic risk scores (GRSs) were calculated by summing the number of risk alleles (0, none; 1, heterozygous; 2, homozygous) weighted by the per allele change in the sex hormone level for each participant. For example, we constructed a GRS of sex hormone-binding globulin for each male participant based on 5 SNPs as follows: GRSsex hormone-binding globulin in men = rs12150660-T × 0.110 + rs2411984-A × 0.034 − rs7910927-T × 0.050 − rs293428-A × 0.029 − rs1042522-G × 0.127.18,20

Sex-specific GRSs were constructed for sex hormone-binding globulin, whereas GRSs for testosterone and dihydrotestosterone were constructed in men only because the availability of identified SNPs predicting levels of these sex hormone measures were limited to men. More detailed information about the included genetic variants is presented in Table 1.

Table 1.

Characteristics of Selected Genetic Variant Associated With Sex Hormone Levels From Previous Genome-Wide Association Studies

Sex Hormone SNP Chromosome Position Gene Minor/major allele Effect/other allele βa Pb Call rate
Male Sex hormone-binding globulin rs12150660 17 7521915 SHBG T/G T/G .110 4 × 10−80 0.99
Male Sex hormone-binding globulin rs2411984 17 47445751 ZNF652 A/G A/G .034 2 × 10−10 0.98
Male Sex hormone-binding globulin rs7910927 10 65138910 JMJD1C G/T T/G −.050 1 × 10−25 1.00
Male Sex hormone-binding globulin rs293428 4 69591782 UGT2B15 G/A A/G −.029 3 × 10−8 1.00
Female Sex hormone-binding globulin rs12150660 17 7521915 SHBG T/G T/G .087 6 × 10−30 0.99
Female Sex hormone-binding globulin rs7910927 10 65138910 JMJD1C G/T T/G −.046 2 × 10−13 1.00
Female Sex hormone-binding globulin rs780093 2 27742603 GCKR T/C T/C −.041 9 × 10−11 1.00
Female Sex hormone-binding globulin rs727428 17 7537792 FXR2/SHBG/SAT2/ATP1B2 T/C T/C −.126 2 × 10−16 1.00
Both Sex hormone-binding globulin rs 1042522c 17 7520197 TP53 G/C G/C −.127 1 × 10−15 1.00
Both Dehydroepiandrosterone sulfate rs78900934 1 101738121 PPIAP7 A/G A/C .050 6 × 10−12 1.00
Both Dehydroepiandrosterone sulfate rs2911280 16 81591313 CMIP A/G A/G .090 6 × 10−10 0.99
Both Dehydroepiandrosterone sulfate rs148982377 7 99075038 ZNF789 C/T C/T −.255 2 × 10−14 1.00
Male Testosterone rs12150660 17 7521915 SHBG T/G T/G 1.103 1 × 10−41 0.99
Male Testosterone rs6258 17 7534678 SHBG T/C T/C −2.856 2 × 10−22 1.00
Male Testosterone rs10822184 10 65337153 JMJD1C C/T T/C −.058 1 × 10−8 0.99
Male Testosterone rs727428 17 7537792 SHBG T/C T/C −.073 1 × 10−12 1.00
Male Dihydrotestosterone rs72829446 17 7552123 SHBG T/C T/C .164 9 × 10−10 0.98
Male Dihydrotestosterone rs727428 17 7537792 SHBG T/C T/C −.103 1 × 10−11 1.00
Both Progesterone rs112295236 11 62915346 SLC22A9 G/C G/C 255 8 × 10−12 0.99
Both Progesterone rs34670419 7 99130834 ZKSCAN5 T/G T/G −.346 6 × 10−14 0.99
Both Estradiol rs117585797 12 6011490 ANO2 A/C A/C Single variant 2 × 10−8 0.98
Both Follicle-stimulating hormone rs11031005 11 30226356 FSHB C/T C/T Single variant 2 × 10−8 0.99
Both Luteinizing hormone rs11031002 11 30215261 FSHB A/T A/T 221 4 × 10−9 1.00
Both Luteinizing hormone rs139643250 19 49517146 RUVBL2 T/C T/C −.68 3 × 10−50  
Both Free androgen index rs117145500 16 52947630 LOC643714 C/A C/A Single variant 2 × 10−8 0.99
Open in a new tab

SNP. single-nucleotide polymorphism.

a

Changes per effect allele are given in μmol/L for dehydroepiandrosterone sulfate, in U/L for luteinizing hormone, and in nmol/L for other hormones.

b

P value for the association between the single nucleotide variant and the specific sex hormone measure as reported in the original genome-wide association study.

c

Replacing rs1641549 of high linkage disequilibrium (r2 = 0.95) due to low call rate (0.44).

Covariates

We assumed that genetically predicted hormone levels were not associated with any risk factor for EAC or BE, and thus act as confounders. However, in the analyses (see the Statistical Analysis section) we still considered the potential influence of the main risk factors of recurrent gastroesophageal reflux symptoms,25 body mass index (BM1),26 and tobacco smoking.27 Information on these 4 covariates was retrieved from written questionnaires or personal interviews. Data were harmonized across studies and merged into a single data set. Recurrent reflux symptoms were defined as symptoms of heartburn or regurgitation occurring at least weekly. BM1 was calculated as the body weight divided by the square of height (kg/m2). Adult weight before any disease-related weight loss was used when available. Otherwise, we used weight at 1 year, 5 years, or 20 years before the data collection, depending on the varying data collection in the individual studies. Participants who had ever smoked at least 100 cigarettes or smoked regularly were defined as ever smokers. 1n the Barrett’s and Esophageal Adenocarcinoma Genetic Susceptibility Study, the missing data on covariates was low, but among the 1885 patients with EAC or BE from the United Kingdom, information regarding reflux symptoms was missing in 776 (41%) participants and BM1 data were missing in 1209 (64%) participants.

Statistical Analysis

Logistic regression was used to estimate odds ratios (OR) and 95% CIs for the associations between sex hormone-specific GRSs or single SNPs and the risk of EAC, BE, as well as a combined outcome of EAC or BE (hereafter referred to as EAC/BE), and separately in men and women. The ORs were adjusted for age (continuous) and the first 4 principal components that reflected the population structure to control for population stratification. The GRSs and single SNPs were included in the models as continuous variables and the ORs and 95% C1s were calculated per SD increase in GRS and per allele increase when single SNPs were used.

To ensure the exclusion restriction assumption of an instrumental variable analysis that the instrumental variables (GRSs or single SNPs) were independent of the 4 covariates, we assessed the associations between the instrumental variables and these covariates, that is, recurrent reflux symptoms (yes or no), BMI (continuous), and tobacco smoking (yes or no) among the control participants, using analysis of variance or chi-square test, whichever was appropriate. Participants with missing data were excluded in each of these analyses.

We used the MR-Egger method, which was adapted from the Egger regression used in meta-analysis, to assess the possible pleiotropic effects (in which a SNP might affect more than 1 phenotypic characteristic) of the SNPs included in the GRSs. In the MR-Egger regression, an intercept differing from zero suggests the existence of directional horizontal pleiotropy. The MR-Egger regression was performed for the GRSs based on 3 or more SNPs only (ie, those for sex hormone-binding globulin, dehydroepiandrosterone sulfate, and testosterone). Some SNPs were included in 2 or more GRSs of different sex hormones (ie, rs12150660 and rs727428), predicting both sex hormone-binding globulin and testosterone. Thus, we re-estimated the associations between the corresponding GRSs and the risk of EAC and BE after excluding these SNPs, to assess the robustness of the estimates.

All statistical tests were 2-sided. The statistical software packages R 3.4.2 (R Foundation for Statistical Computing, Vienna, Austria) and SAS 9.4 (SAS Institute, Cary, NC) were used for the analyses.

Power Estimation

We estimated the statistical power using the web tool mRnd for power calculations in Mendelian randomization analysis (http://cnsgenomics.com/shiny/mRnd).28 Assuming the predicting SNPs explain 20% of the variance in levels of a sex hormone, with the given sample size at the significance level of 0.05, our study had 78% and 29% power to detect an OR of 1.2 in men and in women, respectively, per SD change in sex hormone levels.

Results

Participants

Selected characteristics of the study participants are shown in Table 2. The mean age was 65.1 years. (SD, 10.4 y) in EAC patients, 63.1 years (SD, 12.0 y) in BE patients, and 61.7 years (SD, 11.2 y) in control participants. There were more male than female participants in all groups. Compared with control participants, more patients with EAC and BE had recurrent reflux symptoms, higher BMI, and were ever smokers.

Table 2.

Characteristics in Study Participants

Characteristic Control participants (N = 2127), n (%) Esophageal adenocarcinoma patients (N = 2488), n (%) Barrett’s esophagus patients (N = 3247), n (%)
Age, y      
 <50 301 (14.2) 185 (7.4) 438 (13.5)
 50-59 533 (25.1) 540 (21.7) 766 (23.6)
 60-69 736 (34.6) 878 (35.3) 1002 (30.9)
 70-79 521 (24.5) 688 (27.7) 827 (25.5)
 ≥80 36 (1.7) 177 (7.1) 207 (6.4)
 Missing 0(0) 20 (0.8) 7 (0.2)
 Means ± SD 61.7 ± 11.2 65.1 ± 10.4 63.1 ± 12.0
Sex      
 Male 1670 (78.5) 2173 (87.3) 2454 (75.6)
 Female 457 (21.5) 315 (12.7) 793 (24.4)
Recurrent reflux symptoms    
 No 1411 (66.3) 956 (38.4) 1042 (32.1)
 Yes 344 (16.2) 845 (34.0) 1164 (35.9)
 Missing 384 (18.1) 687 (27.6) 1041 (32.1)
Body mass      
 index      
 <25 772 (36.3) 241 (9.7) 596 (18.4)
 25-29.9 918 (4.2) 442 (17.8) 1178 (36.3)
 ≥30 420 (19.7) 295 (11.9) 919 (28.3)
 Missing 17 (0.8) 1510 (60.7) 554 (17.1)
 Means ± SD 27.0 ± 4.7 28.4 ± 5.2 28.7 ± 5.1
 Tobacco      
  smoking      
 Never 866 (40.7) 563 (22.6) 1065 (32.8)
 Ever 1249 (58.7) 1664 (66.9) 1964 (60.5)
 Missing 12 (0.6) 261 (10.5) 219 (6.7)
Open in a new tab

Genetically Predicted Sex Hormone Levels and Esophageal Adenocarcinoma/Barrett’s Esophagus Risk

Table 3 presents the sex-specific ORs and 95% CIs for GRSs or single SNPs predicting sex hormone levels in relation to the risk of EAC, BE, and EAC/BE. Higher genetically predicted FSH levels were associated with increased risks of EAC, BE, and EAC/BE in both men and women. The point estimates (ORs) for the per allele increase in FSH level were numerically higher in women (OR, 1.28; 95% CI, 1.03–1.59 for EAC/BE) than in men (OR, 1.14; 95% CI, 1.01–1.27 for EAC/BE).

Table 3.

Associations Between Genetic Risk Scores or Single-Nucleotide Variants of Sex Hormones and the Risk of EAC and BE

      EAC BE EAC/BE
Sex Hormone Missinga OR (95% CI)b OR (95% CI)b OR (95% CI)b
Single-nucleotide variants        
Male Follicle-stimulating hormone 3 1.17 (1.03-1.34) 1.12 (0.99-1.27) 1.14 (1.01-1.27)
Female Follicle-stimulating hormone 0 1.29 (0.96-1.73) 1.26 (1.00-1.59) 1.28 (1.03-1.59)
Male Estradiol 101 1.23 (0.69-2.18) 0.74 (0.40-1.35) 0.95 (0.57-1.60)
Female Estradiol 29 0.28 (0.03-2.43) 0.49 (0.14-1.68) 0.42 (0.13-1.35)
Male Free androgen index 73 1.09 (0.91-1.31) 1.08 (0.90-1.29) 1.09 (0.93-1.28)
Female Free androgen index 20 0.82 (0.52-1.30) 0.74 (0.52-1.05) 0.78 (0.56-1.08)
Genetic risk scores        
Male Luteinizing hormone 91 0.92 (0.87-0.99) 0.99 (0.93-1.06) 0.96 (0.91-1.02)
Female Luteinizing hormone 22 0.93 (0.79-1.09) 0.88 (0.77-0.99) 0.89 (0.79-1.00)
Male Sex hormone-binding globulin 171 0.96 (0.91-1.04) 0.99 (0.93-1.05) 0.98 (0.93-1.04)
Male Sex hormone-binding globulinc 102 0.97 (0.91-1.04) 1.00 (0.94-1.07) 0.99 (0.94-1.05)
Female Sex hormone-binding globulin 14 1.04 (0.89-1.21) 0.97 (0.86-1.09) 0.99 (0.89-1.11)
Female Sex hormone-binding globulinc 0 0.97 (0.84-1.13) 1.00 (0.90-1.13) 1.00 (0.90-1.12)
Male Dehydroepiandrosterone sulfate 81 0.98 (0.92-1.05) 0.98 (0.92-1.05) 0.98 (0.92-1.04)
Female Dehydroepiandrosterone sulfate 24 1.15 (0.97-1.36) 0.98 (0.88-1.11) 1.02 (0.91-1.14)
Male Progesterone 72 0.99 (0.93-1.06) 0.97 (0.91-1.04) 0.98 (0.92-1.04)
Female Progesterone 19 0.95 (0.81-1.11) 0.97 (0.86-1.09) 0.96 (0.86-1.07)
Male Testosterone 89 0.95 (0.89-1.02) 0.97 (0.91-1.03) 0.96 (0.91-1.02)
Male Testosteronec 20 0.97 (0.90-1.03) 0.99 (0.93-1.06) 0.98 (0.92-1.04)
Male Dihydrotestosterone 129 1.03 (0.97-1.10) 0.97 (0.91-1.04) 1.00 (0.95-1.06)
Open in a new tab

BE, Barrett’s esophagus; EAC, esophageal adenocarcinoma; OR, odds ratio.

a

Number of missing values of genetic risk score or single-nucleotide variant.

b

ORs (95% CIs) of per allele increase in estradiol, follicle-stimulating hormone, and free androgen index and ORs of per SD increase in genetic risk score forthe remaining, adjusted for age (continuous) and the first 4 genetic principal components.

c

Excluding SNPs rs12150660 and rs727428.

Higher genetically predicted LH levels were associated with a reduced risk of EAC in men (OR per SD increase, 0.92; 95% CI, 0.87–0.99), and we observed a similar association in women (OR, 0.93; 95% CI, 0.79–1.09). Higher genetically predicted LH levels also were associated with reduced risks of BE (OR per SD increase, 0.88; 95% CI, 0.77–0.99) and EAC/BE (OR, 0.89; 95% CI, 0.79–1.00) in women, but no such associations were observed in men.

No statistically significant associations were found between single SNPs of estradiol or free androgen index and the risk of EAC, BE, or EAC/BE in any of the sexes. No associations were observed between GRSs of sex hormone-binding globulin, dehydroepiandrosterone sulfate, or progesterone and the risk of EAC, BE, or EAC/ BE in men or women. No associations were found for GRS of testosterone or dihydrotestosterone in men. All ORs for the per SD increase in GRSs of these sex hormones were close to 1 (range, 0.95–1.15).

Pleiotropy Assessment

The MR-Egger regressions found no evidence for pleiotropy for GRSs of sex hormone-binding globulin (MR-Egger intercept, 0.013; P = .724 for EAC/BE in men; intercept, −0.012; P = .219 in women) or testosterone (intercept, −0.028; P = .352) (Supplementary Table 2). The ORs for sex hormone-binding globulin and testosterone remained unchanged after excluding the SNPs predicting levels of multiple sex hormones from the GRSs (Table 3). On the other hand, pleiotropy was indicated for GRS of testosterone (MR-Egger intercept, 0.111; P = .045 for EAC/BE in men; intercept, 0.231; P = .055 in women) (Supplementary Table 2).

Independence of Instrumental Variables With Covariates

As expected, the genetic variants for FSH and LH were not associated with any of the 4 covariates: recurrent reflux symptoms, BMI, or tobacco smoking in control participants (P < .05 for all comparisons) (Supplementary Tables 3–5).

Discussion

This Mendelian randomization analysis indicated that higher genetically predicted FSH levels increase the risk of EAC and BE and higher LH levels decrease the risk in both sexes. No associations were found for the other 7 sex hormones under study.

The strong male predominance in EAC and BE has prompted the hypothesis that sex hormonal and reproductive factors may be involved in the etiology of these conditions. However, the existing evidence is limited and inconclusive.1,2 A few studies have directly investigated the associations between circulating sex hormone levels and the risk of EAC or BE.9–12 However, these studies all were restricted to men because of the low incidence of EAC in women, and most had a crosssectional design (ie, the sex hormone levels were tested at the time of the cancer diagnosis, which is why the temporal relation could not be established). A recent case-control study nested in prospective cohorts found inverse associations between higher circulating levels of dehydroepiandrosterone and estradiol and the risk of EAC or gastric cardia adenocarcinoma in men,12 but these findings were not supported by the results of the present study. No previous study has specifically assessed the association between endogenous FSH or LH levels and the risk of EAC or BE.

FSH and LH are essential gonadotropins, stimulating the secretion of sex steroids in both sexes.29–31 Increased levels of these hormones have been associated with some health problems, for example, increased FSH levels may cause infertility in women,32 and higher LH levels may contribute to cognitive deficits in Alzheimer’s disease.33 The increased risk of EAC and BE associated with higher FSH levels observed in this study is in line with previous findings of a decreased risk of EAC associated with more childbearing and breastfeeding.34 Interestingly, the receptors of both FSH and LH are highly expressed in the human lower esophagus (ie, where EAC and BE arise). According to the Bgee dataBase for Gene Expression Evolution, a database to retrieve and compare gene expression patterns in multiple species, the expression levels of FSH and LH receptors are in fact highest in the lower esophagus among all anatomic entities, with available expression data in human beings.35 However, the specific downstream mechanisms after binding to their receptors in EAC development are unclear. Notably, the FSH receptor has been found to be expressed selectively on the endothelial surface of the blood vessels of a wide range of tumors,36 indicating an angiogenesis-related mechanism for the potential involvement of FSH in tumor development. The specificity of associations observed only with FSH and LH in the present study suggests that these 2 hormones may be involved in the development of EAC through pathways independent of other sex hormones. It should be noted that 2 genetic variants used for predicting FSH and LH levels (rs11031005 and rs11031002) in this study are in linkage disequilibrium with a functional polymorphism in the promoter of the FSHB gene, which codes for the β polypeptide of FSH.20 In addition, previous studies have generated conflicting findings regarding the direction of effect of these variants on FSH and LH levels,37–39 although we assumed the minor allele would be associated negatively with FSH levels and associated positively with LH levels based on the results of the only relevant GWAS.20 Overall, the specific etiologic roles and mechanisms of FSH and LH in EAC development remain to be identified.

EAC has a poor prognosis, with an overall 5-year survival rate of less than 20% in Western populations.2 Clarifying the role of sex hormones in the development of this cancer may unravel novel targets for prevention and treatment. If an important role of FSH and LH in EAC development is confirmed in future research, it may be worth evaluating potential therapeutic targets (eg, blocking FSH-receptor signaling in the prevention of EAC among high-risk individuals and as adjuvant therapy to counteract tumor recurrences in patients who have undergone curatively intended treatment).

This study was a Mendelian randomization analysis of associations between endogenous sex hormone levels and the risk of EAC and BE. We used data from many high-quality GWASs, which have been merged and analyzed in collaboration through large consortia. A weakness was the lack of a replication analysis in an independent sample, but the availability of such a sample collection will depend on future large-scale collaborative endeavors because of the relatively low incidence of EAC. However, the observed associations, particularly for FSH levels, are less likely to be the result of chance, considering the consistent findings in separate analyses of EAC, BE, and EAC/BE, as well as in both sexes. In a Mendelian randomization analysis, the genetic variants ideally are associated strongly with the endogenous exposure of interest to avoid weak instrument problems, that is, biased results if the “exclusion restriction” is violated or lowered statistical power,40 which might have been a limitation in the present study. Because only a limited number of genetic variants predicting sex hormone levels have been identified from existing GWAS, the instrumental variables used in this study were based on no more than 5 genetic variants or even single variants only. This could have reduced the statistical power, particularly in the analyses with relatively weak instruments. Specifically, only 1 or 2 SNPs have been used for predicting endogenous FSH and LH levels, and these SNPs only account for a small proportion of the variations in FSH and LH levels (Supplementary Table 6). Thus, the estimated associations between genetically predicted sex hormone levels and the risk of EAC or BE were probably biased toward the null in this Mendelian randomization analysis. Potential pleiotropy of the SNPs used for predicting sex hormone levels could not be ruled out. Notably, a few SNPs used in this study correlated moderately or strongly, including the pair of rs11031005 predicting FSH levels and rs11031002 predicting LH levels (r2 of linkage disequilibrium = 0.79). Thus, the observed specific genetic instrument outcome associations might be partially attributable to correlations between sex hormones. In addition, although all selected sex hormone-associated SNPs have been confirmed by GWAS in populations of European descent,18–23 we were unable to verify the validity of the instrumental variables in the study owing to the unavailability of directly measured sex hormone levels. Taken together, the findings of the present study need to be interpreted with caution when making causal inferences.

In summary, this Mendelian randomization analysis based on GWAS data from high-quality studies provides evidence of a role of endogenous FSH and LH levels in the etiology of EAC and BE. Whether the observed associations are causal remains to be confirmed in independent samples with valid instruments or in randomized controlled trials, if ethical and feasible.

Supplementary Material

1

What You Need to Know.

Background

Esophageal adenocarcinoma (EAC) occurs most frequently in men. Studies are needed to determine whether levels of sex hormones are associated with risk of EAC or Barrett’s esophagus (BE).

Findings

In a Mendelian randomization analysis of data from patients with EAC or BE, we found an association between genetically predicted levels of follicle-stimulating and luteinizing hormones and risk of BE and EAC.

Implications for patient care

Monitoring levels of follicle-stimulating and luteinizing hormones might identify patients at risk for BE and EAC.

Acknowledgments

Funding

Supported by the Bengt Ihres Foundation (SLS-78016), the Ruth and Richard Julin Foundation (2018-00137), the Swedish Research Council (521-2014-2536 and 2015-06275), the Swedish Cancer Society (CAN 2015/460), and the National Natural Science Foundation of China (8151101160).

Abbreviations used in this paper

BE

Barrett’s esophagus

BMI

body mass index

EAC

esophageal adenocarcinoma

FSH

follicle-stimulating hormone

GRS

genetic risk score

GWAS

genome-wide association studies

LH

luteinizing hormone

OR

odds ratio

SNP

single-nucleotide polymorphism

Footnotes

Conflicts of interest

The authors disclose no conflicts.

Supplementary Material

Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at https://doi.org/10.1016/jxgh.2019.11.030.

References

  • 1.Xie SH, Lagergren J. The male predominance in esophageal adenocarcinoma. Clin Gastroenterol Hepatol 2016;14:338–347 e1. [DOI] [PubMed] [Google Scholar]
  • 2.Coleman HG, Xie SH, Lagergren J. The epidemiology of esophageal adenocarcinoma. Gastroenterology 2018; 154:390–405. [DOI] [PubMed] [Google Scholar]
  • 3.Xie SH, Lagergren J. Social group disparities in the incidence and prognosis of oesophageal cancer. United European Gastroenterol J 2018;6:343–348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Singh S, Sharma AN, Murad MH, et al. Central adiposity is associated with increased risk of esophageal inflammation, metaplasia, and adenocarcinoma: a systematic review and meta-analysis. Clin Gastroenterol Hepatol 2013;11:1399–1412 e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Lagergren K, Mattsson F, Lagergren J. Abdominal fat and male excess of esophageal adenocarcinoma. Epidemiology 2013; 24:465–466. [DOI] [PubMed] [Google Scholar]
  • 6.Derakhshan MH, Liptrot S, Paul J, et al. Oesophageal and gastric intestinal-type adenocarcinomas show the same male predominance due to a 17 year delayed development in females. Gut 2009;58:16–23. [DOI] [PubMed] [Google Scholar]
  • 7.Palmer S, Albergante L, Blackburn CC, et al. Thymic involution and rising disease incidence with age. Proc Natl Acad Sci U S A 2018;115:1883–1888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Taneja V Sex hormones determine immune response. Front Immunol 2018;9:1931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Cook MB, Wood S, Hyland PL, et al. Sex steroid hormones in relation to Barrett’s esophagus: an analysis of the FINBAR study. Andrology 2017;5:240–247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Petrick JL, Falk RT, Hyland PL, et al. Association between circulating levels of sex steroid hormones and esophageal adenocarcinoma in the FINBAR Study. PLoS One 2018;13: e0190325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Cook MB, Wood SN, Cash BD, et al. Association between circulating levels of sex steroid hormones and Barrett’s esophagus in men: a case-control analysis. Clin Gastroenterol Hepatol 2015;13:673–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Petrick JL, Hyland PL, Caron P, et al. Associations between prediagnostic concentrations of circulating sex steroid hormones and esophageal/gastric cardia adenocarcinoma among men. J Natl Cancer Inst 2019;111:34–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Boef AG, Dekkers OM, le Cessie S. Mendelian randomization studies: a review of the approaches used and the quality of reporting. Int J Epidemiol 2015;44:496–511. [DOI] [PubMed] [Google Scholar]
  • 14.Levine DM, Ek WE, Zhang R, et al. A genome-wide association study identifies new susceptibility loci for esophageal adenocarcinoma and Barrett’s esophagus. Nat Genet 2013;45:1487–1493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Gharahkhani P, Fitzgerald RC, Vaughan TL, et al. Genome-wide association studies in oesophageal adenocarcinoma and Barrett’s oesophagus: a large-scale meta-analysis. Lancet Oncol 2016;17:1363–1373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Laurie CC, Doheny KF, Mirel DB, et al. Quality control and quality assurance in genotypic data for genome-wide association studies. Genet Epidemiol 2010;34:591–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet 2009;5:e1000529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Coviello AD, Haring R, Wellons M, et al. A genome-wide association meta-analysis of circulating sex hormone-binding globulin reveals multiple loci implicated in sex steroid hormone regulation. PLoS Genet 2012;8:e1002805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Prescott J, Thompson DJ, Kraft P, et al. Genome-wide association study of circulating estradiol, testosterone, and sex hormone-binding globulin in postmenopausal women. PLoS One 2012;7:e37815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ruth KS, Campbell PJ, Chew S, et al. Genome-wide association study with 1000 genomes imputation identifies signals for nine sex hormone-related phenotypes. Eur J Hum Genet 2016; 24:284–290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Prins BP, Kuchenbaecker KB, Bao Y, et al. Genome-wide analysis of health-related biomarkers in the UK Household Longitudinal Study reveals novel associations. Sci Rep 2017;7:11008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Ohlsson C, Wallaschofski H, Lunetta KL, et al. Genetic determinants of serum testosterone concentrations in men. PLoS Genet 2011;7:e1002313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Jin G, Sun J, Kim ST, et al. Genome-wide association study identifies a new locus JMJD1C at 10q21 that may influence serum androgen levels in men. Hum Mol Genet 2012; 21:5222–5228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Sun BB, Maranville JC, Peters JE, et al. Genomic atlas of the human plasma proteome. Nature 2018;558:73–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Cook MB, Corley DA, Murray LJ, et al. Gastroesophageal reflux in relation to adenocarcinomas of the esophagus: a pooled analysis from the Barrett’s and Esophageal Adenocarcinoma Consortium (BEACON). PLoS One 2014;9:e103508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Hoyo C, Cook MB, Kamangar F, et al. Body mass index in relation to oesophageal and oesophagogastric junction adenocarcinomas: a pooled analysis from the International BEACON Consortium. Int J Epidemiol 2012;41:1706–1718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Cook MB, Kamangar F, Whiteman DC, et al. Cigarette smoking and adenocarcinomas of the esophagus and esophagogastric junction: a pooled analysis from the international BEACON consortium. J Natl Cancer Inst 2010;102:1344–1353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Brion MJ, Shakhbazov K, Visscher PM. Calculating statistical power in Mendelian randomization studies. Int J Epidemiol 2013;42:1497–1501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Filicori M The role of luteinizing hormone in folliculogenesis and ovulation induction. Fertil Steril 1999;71:405–414. [DOI] [PubMed] [Google Scholar]
  • 30.Howles CM. Role of LH and FSH in ovarian function. Mol Cell Endocrinol 2000;161:25–30. [DOI] [PubMed] [Google Scholar]
  • 31.Araujo AB, Wittert GA. Endocrinology of the aging male. Best Pract Res Clin Endocrinol Metab 2011;25:303–319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Baird DT, Collins J, Egozcue J, et al. Fertility and ageing. Hum Reprod Update 2005;11:261–276. [DOI] [PubMed] [Google Scholar]
  • 33.Burnham VL, Thornton JE. Luteinizing hormone as a key player in the cognitive decline of Alzheimer’s disease. Horm Behav 2015;76:48–56. [DOI] [PubMed] [Google Scholar]
  • 34.Cronin-Fenton DP, Murray LJ, Whiteman DC, et al. Reproductive and sex hormonal factors and oesophageal and gastric junction adenocarcinoma: a pooled analysis. Eur J Cancer 2010; 46:2067–2076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Bastian F, Parmentier G, Roux J, et al. Bgee: integrating and comparing heterogeneous transcriptome data among species. DILS: data integration in life sciences. In: Lecture notes in computer science. Volume 5109 2008:124–131. [Google Scholar]
  • 36.Radu A, Pichon C, Camparo P, et al. Expression of follicle-stimulating hormone receptor in tumor blood vessels. N Engl J Med 2010;363:1621–1630. [DOI] [PubMed] [Google Scholar]
  • 37.Schuring AN, Busch AS, Bogdanova N, et al. Effects of the FSH-beta-subunit promoter polymorphism −211G->T on the hypothalamic-pituitary-ovarian axis in normally cycling women indicate a gender-specific regulation of gonadotropin secretion. J Clin Endocrinol Metab 2013;98:E82–E86. [DOI] [PubMed] [Google Scholar]
  • 38.Grigorova M, Punab M, Ausmees K, et al. FSHB promoter polymorphism within evolutionary conserved element is associated with serum FSH level in men. Hum Reprod 2008; 23:2160–2166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Grigorova M, Punab M, Zilaitiene B, et al. Genetically determined dosage of follicle-stimulating hormone (FSH) affects male reproductive parameters. J Clin Endocrinol Metab 2011;96: E1534–E1441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Davies NM, von Hinke Kessler Scholder S, Farbmacher H, et al. The many weak instruments problem and Mendelian randomization. Stat Med 2015;34:454–468. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1622527-supplement-1.pdf (108.8KB, pdf)

RESOURCES