Associations of tubal ligation and hysterectomy with serum androgen and estrogen metabolites among postmenopausal women in the Women's Health Initiative Observational Study.

Our analysis included women from a nested case-control study of ovarian and endometrial cancers within the Women’s Health Initiative (WHI)-Observational Study (OS) [ 16 – 20 ]. The WHI-OS is a large (n=93,676) prospective cohort of post-menopausal women aged 50 to 79 years [ 21 , 22 ]. Women were recruited from one of 40 clinical centers across the U.S. between 1993 and 1998 [ 21 , 22 ]. At the baseline clinic visit, serum blood samples were collected, and weight, height, and other anthropometric metrics were measured by trained staff. Baseline self-administered questionnaires were used to collect information on participants’ reproductive history, family history of cancer, lifestyle factors, health behaviors (including menopausal hormone therapy [MHT] use), and history of and age at tubal ligation and hysterectomy. Details of the nested endometrial and ovarian cancer case-control studies have been described previously [ 16 – 20 ]. In brief, cases in this study were women with ovarian or endometrial cancer diagnosed between study enrollment (1993–1998) and May 31, 2012. Controls were eligible WHI-OS cohort women who remained cancer-free at the date of the matching case’s diagnosis. Participants were excluded if they had self-reported a history of bilateral oophorectomy or hysterectomy (endometrial cancer controls only) and <1.1 mL of serum available. Baseline serum samples were collected between 45 days and 14.9 years prior to cancer diagnoses (mean time from collection to diagnosis = 6.7 years, standard deviation = 3.6). Analyses were weighted using previously developed inverse probability sampling weights to account for the original selection of participants into the case-control study and to make results more representative of the WHI-OS cohort [ 23 ]. Controls were frequency matched to cases on age at baseline (50–54, 55–59, 60–64, 65–69, 70–74, 75–79), year of blood draw (1993–1996, 1997–1998), self-reported race/ethnicity (Black, Hispanic, White, other), hysterectomy status at baseline (ovarian cancer cases/controls only), and time since last MHT use among never/former users (never or ≤1 year, >1 year). For our analysis, endometrial cancer cases and matched controls were included in the never hysterectomy category given the eligibility criteria for the prior case-control study [ 16 , 18 , 20 ]. The current analytic population included n=920 never/former MHT users (n=473 cases and 447 controls) of which 139 self-reported a history of tubal ligation and 102 a hysterectomy. The WHI study was approved by human subjects review at the Fred Hutchinson Cancer Research Center (WHI Clinical Coordinating Center), as well as at all 40 clinical centers. The current project was reviewed and exempted by the Office of Human Subjects Research at the U.S. National Cancer Institute. Written informed consent was obtained from study participants. The primary exposures of interest were history of tubal ligation or history of hysterectomy, collected via self-report at baseline. The baseline questionnaire ascertained ever/never history of these two gynecologic surgeries. Women reported age at hysterectomy or tubal ligation (25 years, as the median time since procedure was approximately 25 years for both tubal ligation and hysterectomy. The hormones used in this analysis were measured using highly sensitive liquid chromatography–tandem mass spectrometry assays (the Hormone Analysis Unit, Protein Characterization Laboratory, Cancer Research Technology Program, Leidos Biomedical Research, Inc., Frederick, MD) [ 19 , 20 ]. Details of the methods for the androgen [ 16 ] and estrogen [ 24 ] assays have been published previously. The 12 androgens/androgen metabolites were measured as follows (nmol/L): DHEA, DHEAS, A4, testosterone, 5α-androstanedione, dihydrotestosterone (DHT), DHT sulfate (DHTS), androsterone (ADT), ADT-glucuronide (ADT-G), androstanediol glucuronide [found as two isomers: 5α-androstane-3α,17β diol-3-glucuronide (3α-diol-3G) and 5α-androstane-3α,17β diol-17-glucuronide (3α-diol-17G)], and etiocholanolone-glucuronide (Etio-G). The 15 estrogens/estrogen metabolites measured were as follows (pmol/L): estrone, estradiol, metabolites from the 2-hydroxylation pathway (2-hydroxyestrone, 2-hydroxyestradiol, 2-methoxyestrone, 2-methoxyestradiol, 2-hydroxyestrone-3-methyl ether), metabolites from the 4-hydroxylation pathway (4-hydroxyestrone, 4-methoxyestrone, 4-methoxyestradiol), metabolites from the 16α-hydroxylation pathway (16α-hydroxyestrone, estriol, 17-epiestriol, 16-ketostradiol, 16-epiestriol). We additionally measured five of the estrogens (estrone, estradiol, estriol, 2-methoxyestrone, and 2-methyoxyestrdaiol) in their conjugated forms in circulation. We also evaluated four hormone ratios: unconjugated estrone to A4, unconjugated estradiol to testosterone, DHT to testosterone, and 5α-androstanedione to A4. For both hormone assays, reliability was monitored using masked technical replicates within and across batches, coefficients of variation (CVs) and intraclass correlation coefficients (ICCs) were calculated from masked technical replicates and duplicate study samples randomly distributed within and across batches [ 17 , 20 ]. For the androgens/androgen metabolites assay, laboratory CVs (within and across batches) were <11.0% and ICCs ranged from 0.77–0.997 with a mean value of 0.94 (median=0.99) [ 17 ]. For the estrogens/estrogen metabolites assay, laboratory CVs were <6.0% and ICCs ranged from 0.93 to 0.996 with a median=0.98 [ 20 ]. No samples in the current study had undetectable levels for any of the androgens or estrogens measured. We used weighted multivariable-adjusted linear regression to estimate geometric means (GMs) of androgens and estrogens by history of tubal ligation and/or hysterectomy status, overall and stratified by time since each procedure. Mean hormone levels were obtained for both women with and without a history of each procedure via least-squares estimates. We assessed differences in the geometric means by surgery status using a Wald Test. We adjusted for the following potential confounders, which were selected a priori based on the literature and directed acyclic graphs: age at blood draw (<60, 60–69, 70+ years old), year of blood draw (1993–1996 and 1997–1998), race and ethnicity (dichotomized to non-Hispanic White (NHW) women and all other racial/ethnic groups due to small numbers), parity (never pregnant, 1, and 2+ pregnancies), oral contraceptive use history (never, ever), BMI (<25, 25–29.9, ≥30 kg/m 2 ), and time since menopause (<10, 10–20, ≥ 20 years). Because subclinical or underlying diseases may have influenced associations, we performed the following sensitivity analyses: 1) excluding women with a history of diabetes at baselines (n=60), 2) excluding endometrial and ovarian cancer cases (n=473), and 3) limiting the comparison group for each analysis to those with neither a hysterectomy or tubal ligation (n=693) versus those that had tubal ligation only (n=125) and those with hysterectomy only (n=88). Among our sample, there were 14 women that had both a hysterectomy and tubal ligation procedure. All statistical tests were two-sided with 5% type I error. Q-values reflecting the false discovery rates (FDR) were calculated to account for multiple comparisons (43 tests per exposure). All statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC).

The mean age at cohort enrollment and blood draw of the women included in our study was 64.4 years (± SD 7.2, Table 1 ). The study population included predominantly NHW women (87.7%). In our study population of women with no history of bilateral oophorectomy at blood draw, 15.1% of women self-reported a history of tubal ligation and 16.2% of women self-reported a history of hysterectomy. Circulating postmenopausal androgen and estrogen levels were comparable between women who reported having a previous tubal ligation compared with women who did not; we did not observe any consistent pattern in higher or lower hormone levels associated with a history of this surgery ( Table 2 ). We did not observe any differences in circulating hormone levels for those with a longer time since tubal ligation (>25 years) compared to women with a shorter time since procedure (≤ 25 years) ( Table 3 ). We observed a tendency toward lower levels of postmenopausal adrenal androgens (DHEA, DHEAS, A4, testosterone, and ADT) and higher levels of many postmenopausal estrogens/estrogen metabolites among women reporting prior hysterectomy compared those without a history of hysterectomy ( Table 4 ). After considering multiple comparisons, women with hysterectomy had lower levels of testosterone (GM yes 0.46 [95% CI: 0.37–0.57] vs. GM no 0.62 [95% CI: 0.53–0.72]) and higher levels of both 2-hydroxyestrone-3-methyl ether (GM yes 11.1 [95% CI: 8.95–13.9] vs. GM no 8.70 [95% CI: 7.38–10.3]) and 4-methoxyestrone (GM yes 6.50 [95% CI: 5.05–8.37] vs. GM no 4.92 [95% CI: 4.00–6.05]). Among women with hysterectomy, circulating hormone levels did not substantially vary by time since procedure (>25 years vs. ≤ 25 years) ( Table 5 ). The results of the following sensitivity analyses were largely consistent with the overall analysis: 1) excluding diabetics ( Supplemental Table 1 – 2 ), 2) limited to controls only ( Supplemental Tables 3 – 4 ), and 3) limiting the comparison group to those who did not have a history of hysterectomy or tubal ligation ( Supplemental Table 5 ).

In this study of gynecologic surgeries and postmenopausal circulating hormone levels, we observed that circulating hormone levels did not differ by prior tubal ligation history. In contrast, we observed lower postmenopausal testosterone levels and higher levels of some estrogen metabolites among women who had a hysterectomy compared to women who had not. Our findings of no association between circulating hormones and tubal ligation history are consistent with previous studies reporting no short-term [ 8 , 9 ] or long-term [ 1 ] changes in circulating hormone levels (i.e., DHEAS, estradiol, and androstenedione). Our study is most comparable in size and time since procedure to the study by Nelson et al. that evaluated hormones among women stratified into groups of less than 15 years since procedure and 15 years or more since tubal ligation and reported no difference in DHEAS, testosterone, or estradiol levels [ 1 ]. Our study is consistent with previous studies reporting reduced adrenal androgen levels (i.e., testosterone, androstenedione, DHEA, and DHEAS) among women with a history of hysterectomy compared to women with no hysterectomy [ 10 , 11 ]. To our knowledge, our study may be the first to suggest lower ADT levels as well among postmenopausal women with a prior hysterectomy. Interestingly, Laughlin et al. [ 11 ] found that the highest testosterone concentrations in their study were observed among women with no surgery (i.e., intact uterus and ovaries) and lower levels were noted among women with hysterectomy but no oophorectomy—comparable to the current study population. They also observed that the lowest levels of testosterone were among women with both, hysterectomy and bilateral oophorectomy [ 11 ]. Our study is also consistent with these prior studies in reporting no association between hysterectomy and estrone or estradiol levels, but we do report an association between history of hysterectomy and higher levels of several estrogen metabolites, which have not been measured in other studies. There is a substantial body of research reporting associations between tubal ligation and hysterectomy on later life disease risks, including cancer risk. A prior meta-analysis of 54 studies demonstrated that both tubal ligation (n=30 studies) and hysterectomy (n=24 studies) were associated with lower ovarian cancer risk [ 25 ]. In contrast, there was no association between breast cancer risk and tubal ligation [ 26 – 28 ], but breast cancer risk may be reduced among women who underwent a hysterectomy without oophorectomy or with partial ovary removal [ 26 , 29 , 30 ]. Both, ovarian cancer risk [ 31 ] and breast cancer [ 32 ] risk have been shown to rise with increasing postmenopausal circulating testosterone levels. Given the limited evidence of an association between tubal ligation and postmenopausal levels of circulating hormones, the decreased ovarian cancer risk with prior tubal ligation is likely explained by other non-hormonal mechanisms. Previous research has examined the association between hysterectomy procedures and cardiovascular disease (CVD) risk and mortality [ 33 – 37 ]. Hysterectomy is associated with increased CVD risk in many population-based studies [ 33 – 37 ]. From our findings and those reported by other researchers, differences in hormone levels associated with hysterectomy may contribute to hormone-related disease risks in later life, though these mechanisms are not yet clear. Strengths of our study include a large sample size and use of reliable and sensitive LC-MS based assays that quantified a comprehensive list of circulating androgens and estrogens and their major metabolites in postmenopausal women. Our study is limited in that it used a single serum measurement to quantify circulating hormones in postmenopausal women; however, the temporal stability of circulating hormones is relatively high in postmenopausal women (2-year ICC ranges for estrogen metabolites: 0.93–0.99) [ 38 ]. Although our study is cross-sectional in design, with hormones being measured in serum samples collected at study enrollment, participants were asked to recall historical surgical procedures, so the temporal sequence of the surgical exposure of interest preceded sample collection and hormone measurement. There are established differences in hysterectomy rates by race and ethnicity, but whether the potential changes in hormone metabolism associated with this procedure also differ among other racial/ethnic groups is something we were unable to assess due to small sample sizes; approximately 90% of our study population identified as non-Hispanic White [ 39 ]. Finally, our laboratory assay was not designed to measure sex hormone binding globulin (SHBG) levels. Thus, we were not able to calculate the bioavailable fraction of the unconjugated estradiol or testosterone levels. This limitation is unlikely to influence our findings, as Laughlin et al. reported similar reductions in total and bioavailable testosterone by hysterectomy status and no differences in SHBG levels (or bioavailable estradiol) by hysterectomy status [ 11 ]. In summary, our work suggests that circulating hormone levels among postmenopausal women are not appreciably different when comparing women with a history of tubal ligation to those with no such history. However, our study supports other research indicating that history of hysterectomy may be associated with lower circulating levels of testosterone and higher levels of some estrogen metabolites among postmenopausal women, which may have implications for disease risk in postmenopausal women.

Tubal ligation is considered one of the most effective forms of pregnancy prevention and is the most common form of surgical sterilization in women [ 1 , 2 ]. More than 18% of U.S. women ages 15 to 49 years report having undergone tubal ligation based on data from the National Center for Health Statistics [ 3 ]. Hysterectomy is also a common gynecologic surgical procedure in U.S. women [ 4 ]; 31.7% of women 50 years of age and older reported having had a hysterectomy in recent National Health Interview Survey data [ 5 ]. Women undergo hysterectomies for many indications (e.g. fibroids, cancer, prolapse, endometriosis) and roughly half of these hysterectomies include simultaneous bilateral oophorectomy [ 6 , 7 ]. The high prevalence of these two gynecologic surgeries underscores the importance of conducting research on the long-term health effects among women who have undergone these procedures. Studies examining circulating sex steroid hormones following tubal ligation [ 1 , 8 , 9 ] and hysterectomy [ 10 , 11 ] are limited. For tubal ligation, most studies were limited by small sample sizes (n<15) and report no short-term changes in select hormones measured in days to 3 months after surgery [ 8 , 9 ]. The only study to date that evaluated longer-term changes (average 15 years) in hormones following tubal ligation (n=134) reported no differences in hormone concentrations comparing women with and without tubal ligation and who were primarily pre- or peri-menopausal at time of blood draw [ 1 ]. Studies on hormone levels following hysterectomy report reductions in circulating androgen levels (e.g., testosterone) and no change in circulating estrogens, among postmenopausal women [ 10 , 11 ]. McTiernan et al. [ 10 ], reported lower geometric mean testosterone and dehydroepiandrosterone (DHEA) levels among individuals with (n=65) and without hysterectomy in a subsample of postmenopausal women from the WHI Dietary Modification trial. Similar differences in testosterone and androstenedione levels were reported in a cohort of postmenopausal women living in Rancho Bernardo California when comparing women with a hysterectomy (n=246 with or without bilateral oophorectomy) to women without hysterectomy or oophorectomy [ 11 ]. These prior studies focused primarily on comparing a limited number of hormones and few hormone metabolites. As sex steroid hormone levels in women decrease with aging and menopause, the balance between circulating levels of androgens and estrogens changes and may contribute to a slightly pro-inflammatory milieu [ 12 ]. Most androgens in circulation are the relatively inactive precursors: androstenedione (A4), DHEA, DHEA sulfate (DHEAS). These molecules serve as precursors to both androgens and estrogens (via the aromatization of testosterone and A4). Tissue-based metabolism of these androgen precursors provides the major source for active androgen metabolites [ 13 ]. Glucuronidated products of peripheral androgen synthesis can be found in circulation and are reflective of this tissue-level androgenic activity [ 13 – 15 ]. To date, no study has examined an extensive panel of androgens and estrogens, including these conjugated forms, in relation to tubal ligation or hysterectomy status in postmenopausal women. The goal of the current analysis was to evaluate whether there were long-term differences in circulating levels of androgens, estrogens, and their major metabolites in postmenopausal women by tubal ligation or hysterectomy status.