Abstract
Objective
To examine the association between gender of offspring and epithelial ovarian cancer (EOC).
Methods
We compared gender of offspring between 664 incident EOC cases and 1531 controls participating in a population-based study conducted in Pennsylvania, Ohio, and New York from 2003-2008. Multivariable unconditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) adjusting for potential confounders.
Results
Bearing a male offspring was associated with an 8% lower EOC risk; bearing all boys was associated with an 11% lower risk. Compared to bearing all girls, bearing all boys was associated with a 14% decrease risk. Increasing number of male offspring increased the protective effect (adjusted-OR: 0.92, 0.91, 0.84, for 1, 2, and 3+ boys compared to all girls). Results where similar when limiting cases to invasive disease and to the high-grade serous histotype.
Conclusion
Fetal sex, which influences maternal hormonal milieu, may impact EOC risk.
Keywords: Epithelial Ovarian Cancer, case-control study, gender of offspring
Introduction
Epithelial ovarian cancer (EOC) is the most lethal gynecologic malignancy[1]. Bearing children has consistently been shown to protect against the disease[2]. Although the exact mechanism of this protective effect is unknown, it is generally attributed to suppressed ovulation throughout the pregnancy[3]; however, anovulation alone cannot explain the magnitude of the protective effect[4], suggesting that other pregnancy-associated factors may impact EOC risk. One such factor is fetal sex, which may influence the maternal hormonal milieu[5-7] thereby affecting EOC risk. There are few epidemiologic studies of the relationship between gender of offspring and EOC risk, and the results have been inconsistent[8-11].
We used data from a large, case-control study of EOC to assess the association between gender of offspring and EOC.
Methods
Subjects
The Hormone and Ovarian cancer PrEdiction (HOPE) study is a population-based case-control study conducted in the contiguous region of western Pennsylvania, eastern Ohio, and western New York. Details of eligibility criteria for the population and recruitment methods have been published previously[12]. Briefly, cases were women diagnosed with incident epithelial ovarian, peritoneal, or fallopian tube cancer between February 2003 and November 2008. Controls were identified using random digit dialing and frequency matched to cases by 5-year age groups and 3-digit telephone prefix. A total of 902 cases and 1802 controls participated in the study. The study was approved by the University of Pittsburgh Institutional Review Board.
Data Collection an Exposure Assessment
Two-hour, in-person interviews were conducted by trained interviewers to obtain information on reproductive, medical, and demographic data from birth until a reference date. The reference date was calculated as 9 months before diagnosis for cases or interviews for controls to ensure that exposures occurred before ovarian cancer diagnosis in cases and within a similar time frame for cases and controls. A life events calendar with milestones such as marriages, births, and deaths was used to aid recall. For each pregnancy, a woman was asked the outcome, including live or still birth. For each live or still birth, the woman was asked if it was a single birth followed by “Was it a boy or girl?”
Statistical Analyses
Analyses were limited to 1531 controls and 664 cases who reported all singleton, full-term (live or still birth) births and gender of offspring for each birth. Differences in demographic factors between cases and controls were estimated using chi-square and t-tests, as appropriate. Unconditional logistic regression analyses were conducted to estimate odds ratio (OR) 95% confidence intervals (CIs) for the association of gender of offspring and EOC risk. Linear trends were assessed with Wald tests. In the multivariable models, we controlled for age, race, education level, duration of oral contraception use, and number of full-term births. Sensitivity analyses with models including age at first birth, duration of breastfeeding, tubal ligation, and hysterectomy showed no change in results; these variables were excluded from the final models. Because BRCA mutation carriage may skew offspring gender ratios[13] and because it is associated with EOC risk, we performed all analyses limiting to women without a reported family history of breast or ovarian cancers, using family history as a surrogate for BRCA carriage status. These analyses showed no changes in results.
Stata/SE version 15.1 (StataCorp) was used to conduct the analyses. Two-sided P-values less than 0.05 were considered significant.
Results
Characteristics of the 1531 controls and 664 cases are summarized in Table 1. Compared to the controls, cases were older and less likely to be white, highly educated, use oral contraceptives, have had a tubal ligation, and breastfed. They were more likely have had a hysterectomy, use talc, and report a family history of breast or ovarian cancer.
Table 1.
Characteristics of HOPE Study Participants, 2003-2008
| Controls (N=1531) |
Cases (N=664) |
P-Value | |
|---|---|---|---|
| Age, years, mean (SD) | 57.37 (12.38) | 60.20 (12.24) | <0.0001 |
| Race | 0.004 | ||
| White | 1490 (97.32) | 630 (94.88) | |
| Non-White | 41 (2.68) | 34 (5.12) | |
| Education | <0.001 | ||
| Less than High School | 74 (4.83) | 69 (10.39) | |
| High School | 476 (31.09) | 241 (36.30) | |
| Post High School Training | 485 (31.68) | 183 (27.56) | |
| College Graduate | 300 (19.60) | 97 (14.61) | |
| Postgraduate | 196 (12.80) | 74 (11.14) | |
| Duration of Oral Contraceptive Use, years | <0.001 | ||
| 0 | 439 (28.67) | 260 (39.16) | |
| <1 | 237 (15.48) | 120 (18.07) | |
| 1-4 | 416 (27.17) | 152 (22.89) | |
| 5-9 | 260 (16.98) | 91 (13.70) | |
| 10+ | 179 (11.69) | 41 (6.17) | |
| Number of Full Births | 0.224 | ||
| 1 | 225 (14.70) | 116 (17.47) | |
| 2 | 576 (37.62) | 249 (37.50) | |
| 3+ | 730 (47.68) | 299 (45.03) | |
| Tubal Ligation | <0.001 | ||
| No | 957 (62.51) | 476 (71.69) | |
| Yes | 574 (37.49) | 188 (28.31) | |
| Hysterectomy | 0.002 | ||
| No | 1252 (81.78) | 505 (76.05) | |
| Yes | 279 (18.22) | 159 (23.95) | |
| Talc Use | 0.001 | ||
| No | 1028 (67.15) | 395 (59.49) | |
| Yes | 503 (32.85) | 269 (40.51) | |
| Family History of Breast or Ovarian Cancer* | 0.038 | ||
| No | 1265 (82.73) | 523 (79.00) | |
| Yes | 264 (17.27) | 139 (21.00) | |
| Duration of breastfeeding (years) | <0.001 | ||
| 0 | 665 (43.46) | 368 (55.51) | |
| <1 | 488 (31.90) | 188 (28.36) | |
| 1 or more s | 377 (24.64) | 107 (16.14) | |
2 missing in controls and 2 missing in cases
among women with at least 1 live birth
The association between gender of offspring and EOC risk is presented in Table 2. Ever bearing a male offspring was associated with an 8% lower EOC risk (adjusted-OR:0.92; 95%CI:0.75-1.24); bearing all boys was associated with a somewhat greater protective effect (adjusted-OR:0.89; 95%CI:0,70-1.13). Compared to giving birth to girls only, giving birth to boys only was associated with a 14% decrease in EOC risk (adjusted-OR:0.86; 95%CI:0.65-1.15). Increasing number of male offspring appeared to increase the protective effect (adjusted-OR: 0.92, 0.91, 0.84, for 1, 2, and 3 or more boys compared to all girl offspring).
Table 2.
Association between gender of offspring and risk of epithelial ovarian cancer: The HOPE Study, 2003-2008
| All Cancer | Invasive | HGSOC | |||||
|---|---|---|---|---|---|---|---|
| Controls | Cases | Adjusted OR* |
Cases | Adjusted OR* |
Cases | Adjusted OR* |
|
| N(%) | N(%) | 95% CI | N(%) | 95% CI | N(%) | 95% CI | |
| Gave birth to a boy | |||||||
| Never | 316 (20.64) | 150 (22.59) | ref | 126 (21.95) | ref | 76 (22.16) | ref |
| Ever | 1215 (79.36) | 514 (77.41) | 0.92 (0.72, 1.16) | 448 (78.05) | 0.93 (0.73, 1.20) | 267 (77.84) | 0.89 (0.65, 1.21) |
| All Boys | |||||||
| No | 1199 (78.31) | 525 (79.07) | ref | 458 (79.79) | ref | 277 (80.76) | ref |
| Yes | 332 (21.69) | 139 (20.93) | 0.89 (0.70, 1.13) | 119 (20.21) | 0.85 (0.66, 1.10) | 66 (19.24) | 0.83 (0.61, 1.13) |
| Gender of offspring | |||||||
| All girls | 316 (20.64) | 150 (22.59) | ref | 116 (20.21) | ref | 66 (19.24) | ref |
| All boys | 332 (21.69) | 139 (20.93) | 0.86 (0.65, 1.15) | 126 (21.95) | 0.85 (0.63, 1.15) | 76 (22.16) | 0.80 (0.55, 1.17) |
| Mixed | 883 (57.67) | 375 (56.48) | 0.95 (0.74, 1.24) | 332 (57.84) | 0.99 (0.76, 1.31) | 201 (58.60) | 0.95 (0.68, 1.32) |
| Number of boys | |||||||
| No boy | 316 (20.64) | 150 (22.59) | ref | 126 (21.95) | ref | 76 (22.16) | ref |
| 1 boy | 602 (39.32) | 261 (39.31) | 0.92 (0.72, 1.19) | 227 (39.55) | 0.94 (0.72, 1.23) | 141 (41.11) | 0.96 (0.70, 1.33) |
| 2 boys | 403 (26.32) | 171 (25.75) | 0.91 (0.68, 1.22) | 150 (26.13) | 0.93 (0.68, 1.26) | 79 (23.03) | 0.76 (0.52, 1.11) |
| 3 or more boys | 210 (13.72) | 82 (12.35) | 0.84 (0.56, 1.25) | 71 (12.37) | 0.82 (0.54, 1.25) | 47 (13.70) | 0.79 (0.48, 1.30) |
| P for trend | 0.41 | 0.36 | 0.24 | ||||
| Among women with one birth† | |||||||
| No boy | 117 (52.00) | 65 (56.03) | ref | 51 (53.13) | ref | 31 (51.67) | ref |
| 1 boy | 108 (48.00) | 51 (43.97) | 0.85 (0.53, 1.36) | 45 (46.88) | 0.95 (0.58, 1.56) | 29 (48.33) | 1.04 (0.57, 1.88) |
| Among women with two births† | |||||||
| No boy | 139 (24.13) | 66 (26.51) | ref | 58 (26.98) | ref | 35 (29.91) | ref |
| 1 boy | 300 (52.08) | 124 (49.80) | 0.84 (0.58, 1.22) | 108 (50.23) | 0.84 (0.57, 1.23) | 60 (51.28) | 0.78 (0.49, 1.26) |
| 2 boys | 137 (23.78) | 59 (23.69) | 0.88 (0.57, 1.36) | 49 (22.79) | 0.84 (0.53, 1.32) | 22 (18.80) | 0.62 (0.34, 1.12) |
| P for trend | 0.56 | 0.44 | 0.12 | ||||
Adjusted for age, race, education, duration of OC use, number of full-term births; further adjusting for duration of breast feeding, tubal ligation and hysterectomy did not change the results (data not shown)
Not adjusted for number of full-term pregnancies in the analysis by number of births.
Stratified analyses by number of full-term births provided additional support for reduced risk associated with male offspring. Compared to women bearing only girls, among women with exactly one full-term birth as well as among women with exactly two full-term births, bearing a male offspring was consistently associated with reduced risk (Table 2).
Results were similar when limiting cases to invasive disease and to the high-grade serous histotype, the most common EOC subtype (Table 2). Small number of cases of the other main histotypes precluded meaningful analyses.
Discussion
We investigated the association between gender of offspring and EOC risk. In general, bearing male offspring was associated with reduced EOC risk. Although our findings failed to reach statistical significance, the consistent protective effect of a male offspring and the increasing protective effect with greater number of male offspring found herein lend support to the hypothesis that the influence of fetal sex on the maternal hormonal milieu during pregnancy may impact EOC risk.
Four studies have investigated the association between offspring gender and ovarian cancer risk[8-11]. Our previous population-based study conducted in eastern Pennsylvania reported findings similar to the current study – bearing male offspring was associated with a non-significant decrease in EOC risk[8]. Notably, the point estimate for bearing all boys compared to all girls was 0.84, which is similar to what we report herein (0.86). A nested case-control study in Sweden supported these findings reporting that bearing a male child was significantly associated with reduced EOC risk, and increasing number of male offspring was associated with increasing protection (adjusted ORs: 0.92, 0.87, 0.82, for 1, 2 or 3+ boys, compared to all girls)[9]. Again, the magnitude of the point estimates were similar to what we found (0,92, 0.91, 0.84, respectively). In contrast, a population-based study in Australia reported no association between offspring gender and EOC in general but a 2-fold increase risk of the mucinous histotype associated with bearing only male offspring[11]. A cohort study in Norway reported no EOC-offspring gender association in general, but an increased risk of endometrioid tumors among women who gave birth to girls only[10]. Although our sample size was limited for examining mucinous and endometroid subtypes, we found bearing male offspring was associated with reduced risk of high-grade serous EOC, the most common and fatal subtype. Prior studies did not separate low from high-grade serous EOC, which are now believed to have distinct etiologies and be different disease[14]. This may explain, in part, the disparate findings between prior work and what we report herein.
Pregnancy has consistently been shown to protect against EOC[2]. Although the exact mechanism for this association remains unknown, an altered maternal hormonal milieu may play a role[15]. During pregnancy, maternal hormone concentrations may differ by fetal sex. Carriage of a male fetus is associated with lower maternal levels of estradiol and hCG[5, 6] and higher maternal levels of progesterone[7]. While the role of hCG in EOC etiology is unclear, progesterone is believed to protect against EOC while estrogens may increase risk[15]. Thus, it is biologically plausible that bearing male offspring can impact EOC risk differently than bearing female offspring.
A strength of this study is its population-basis and collection of data through standardized, structured, in-person interviews administered by trained personnel, ensuring consistent and high-quality exposure measurements. The major weakness is sample size, which enabled us to detect a minimum OR of only 0.75 (80% power, alpha=0.05). A greater sample size is needed to determine whether the actual estimate is more in the range of what we and others have reported (0.84-0.86). In addition, a larger sample size would enable us to explore histotype-specific associations, as EOC is now believed to be several etiologically-distinct diseases[16]. Another weakness is that this data set did not have BRCA or other germline mutation status recorded. Emerging data suggest that among BRCA mutation carriers, gender is skewed towards female offspring[13]. Thus, carriage of a BRCA mutation may confound the relationship between offspring gender and EOC risk. However, we controlled for family history of breast or ovarian cancer (a surrogate for BRCA status) in our analyses. Moreover, we found similar associations between offspring gender and EOC when limiting analyses to women without a family history of breast or ovarian cancers.
In summary, we report that bearing male offspring may be associated with a decreased risk of EOC compared to bearing female offspring. Although our results were not statistically significant, the consistency of findings in our analyses, the similarity of our estimates to previous studies, and the biologic plausibility of the association support our conclusions. As no one study may be powered to address the question adequately, pooled analyses of existing studies to examine the EOC-offspring gender association are warranted. The larger sample size will enable histotype-specific analyses, which could shed light on disease etiology and pave the way for new avenues of prevention research for this often-fatal disease.
Highlights.
Pregnancy protects against ovarian cancer and hormones impact risk for the disease
During pregnancy, hormone levels vary based on fetal sex
Results imply that bearing male offspring reduces risk more than female offspring
Acknowledgments
Funding: This work supported by National Cancer Institute (K07-CA80668, R01CA095023); the Department of Defense (DAMD17-02-1-0669); and the University of Pittsburgh School of Medicine Dean’s Faculty Advancement Fund.
Footnotes
Conflict of Interest: The authors report no conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Zhuxuan FU, Department of Epidemiology, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA USA.
Kirsten MOYSICH, Roswell Park Cancer Institute, Buffalo, NY USA.
Roberta B. NESS, University of Texas School of Public Health, Houston, TX USA.
Francesmary MODUGNO, Womens Cancer Research Program, Magee-Womens Research Institute and UPMC Hillman Cancer Center, and Department of Obstetrics, Gynecology and Reproductive Sciences, University of Pittsburgh School of Medicine, Pittsburgh, PA USA.
References
- 1.Howlader N, N.A., Krapcho M, Miller D, Brest A, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA (eds), SEER Cancer Statistics Review, 1975-2016, National Cancer Institute Bethesda, MD, https://seer.cancer.gov/csr/1975_2016/, based on November 2018 SEER data submission, posted to the SEER web site, April 2019. [Google Scholar]
- 2.Sung HK, et al. , The Effect of Breastfeeding Duration and Parity on the Risk of Epithelial Ovarian Cancer: A Systematic Review and Meta-analysis. J Prev Med Public Health, 2016. 49(6): p. 349–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fathalla MF, Incessant ovulation--a factor in ovarian neoplasia? Lancet, 1971. 2(7716): p. 163. [DOI] [PubMed] [Google Scholar]
- 4.Risch HA, et al. , Events of reproductive life and the incidence of epithelial ovarian cancer. Am J Epidemiol, 1983. 117(2): p. 128–39. [DOI] [PubMed] [Google Scholar]
- 5.Obiekwe BC and Chard T, Human chorionic gonadotropin levels in maternal blood in late pregnancy: relation to birthweight, sex and condition of the infant at birth. Br J Obstet Gynaecol, 1982. 89(7): p. 543–6. [DOI] [PubMed] [Google Scholar]
- 6.Adamcova K, et al. , Steroid hormone levels in the peripartum period - differences caused by fetal sex and delivery type. Physiol Res, 2018. 67(Supplementum 3): p. S489–s497. [DOI] [PubMed] [Google Scholar]
- 7.Boroditsky RS, et al. , Serum human chorionic gonadotropin and progesterone patterns in the last trimester of pregnancy: relationship to fetal sex. Am J Obstet Gynecol, 1975. 121(2): p. 238–41. [DOI] [PubMed] [Google Scholar]
- 8.Gierach GL, Modugno F, and Ness RB, Gender of offspring and maternal ovarian cancer risk. Gynecol Oncol, 2006. 101(3): p. 476–80. [DOI] [PubMed] [Google Scholar]
- 9.Baik I, et al. , Gender of offspring and maternal risk of invasive epithelial ovarian cancer. Cancer Epidemiol Biomarkers Prev, 2007. 16(11): p. 2314–20. [DOI] [PubMed] [Google Scholar]
- 10.Albrektsen G, et al. , Twin births, sex of children and maternal risk of ovarian cancer: a cohort study in Norway. Br J Cancer, 2007. 96(9): p. 1433–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jordan SJ, et al. , Beyond parity: association of ovarian cancer with length of gestation and offspring characteristics. Am J Epidemiol, 2009. 170(5): p. 607–14. [DOI] [PubMed] [Google Scholar]
- 12.Lo-Ciganic WH, et al. , Aspirin, nonaspirin nonsteroidal anti-inflammatory drugs, or acetaminophen and risk of ovarian cancer. Epidemiology, 2012. 23(2): p. 311–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Moslehi R, et al. , Impact of BRCA mutations on female fertility and offspring sex ratio. Am J Hum Biol, 2010. 22(2): p. 201–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Vang R, Shih Ie M, and Kurman RJ, Ovarian low-grade and high-grade serous carcinoma: pathogenesis, clinicopathologic and molecular biologic features, and diagnostic problems. Adv Anat Pathol, 2009. 16(5): p. 267–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Risch HA, Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst, 1998. 90(23): p. 1774–86. [DOI] [PubMed] [Google Scholar]
- 16.Kurman RJ and Shih Ie M, The Dualistic Model of Ovarian Carcinogenesis: Revisited, Revised, and Expanded. Am J Pathol, 2016. 186(4): p. 733–47. [DOI] [PMC free article] [PubMed] [Google Scholar]