Qiu M, Shields CL. Relationship Between Female Reproductive Factors and Choroidal Nevus in US WomenAnalysis of Data From the 2005-2008 National Health and Nutrition Examination Survey. JAMA Ophthalmol. 2015;133(11):1287–1294. doi:10.1001/jamaophthalmol.2015.3178
Copyright 2015 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Choroidal nevus is a precursor for uveal melanoma and there are no known risk factors besides white race. Female sex hormones have been hypothesized to play a role in the pathogenesis of uveal melanoma.
To explore the association between female reproductive factors and choroidal nevus in the US adult female population.
Design, Setting, and Participants
Cross-sectional, population-based study of 2505 US women aged 40 years or older from the 2005-2008 National Health and Nutrition Examination Survey. The women completed the reproductive health questionnaire and underwent retinal imaging. The analysis was conducted in April 2015.
Age at menarche, oral contraceptive use, pregnancy, parity, age at first and last births, age at menopause, hysterectomy, oophorectomy, hormone therapy use, and body mass index (calculated as weight in kilograms divided by height in meters squared).
Main Outcomes and Measures
Choroidal nevus on retinal imaging.
The mean age of participants was 56.5 years, and the racial/ethnic distribution was 76.9% white, 10.1% African American, 8.4% Hispanic, and 4.8% other. The weighted prevalence of choroidal nevus was 4.5%. Premenopausal women who first gave birth before age 20 years had more than 4 times higher odds of choroidal nevus than those who first gave birth after age 35 years (odds ratio [OR], 4.16; 95% CI, 1.29-13.45; P = .02), and premenopausal women who gave birth to their last child before age 25 years had nearly 5 times higher odds of choroidal nevus than those who gave birth to their last child after age 35 years (OR, 4.89; 95% CI, 1.15-20.74; P = .03). These relationships were independent of total parity and years between the first and last births. The odds of choroidal nevus in postmenopausal women who were overweight and obese were 2 times higher than in postmenopausal women with normal body mass index (OR, 2.11; 95% CI, 1.19-3.76; P = .01 for overweight and OR, 1.92; 95% CI, 1.13-3.25; P = .02 for obese).
Conclusions and Relevance
The association between choroidal nevus and earlier start and end to childbearing in premenopausal women and obesity in postmenopausal women suggests that this relationship could be partially mediated via increased total lifetime unopposed estrogen. Clinicians should have an increased index of suspicion for choroidal nevus and choroidal melanoma in this population.
Choroidal nevus, the most common clinically documented intraocular mass,1 has an estimated 1 in 8845 risk for transformation into malignant melanoma.2 Besides white race, there are no widely accepted risk factors for choroidal nevus,3- 8 and population-based studies have demonstrated that choroidal nevus and choroidal melanoma affect both sexes approximately equally.4- 8 There is growing interest regarding the topic of sex differences in cancer susceptibility,9 and it has been documented that female sex hormones could be involved in the pathogenesis of cutaneous melanoma10- 14 and possibly choroidal melanoma.15- 21 There is an abundance of literature concerning the relationship between sex hormones and cutaneous nevus and melanoma10- 14 but a relative paucity of studies investigating the role of sex hormones in choroidal melanoma.15- 21
Cutaneous nevi have been documented to arise during puberty22 and change shape and color during pregnancy,23 suggesting that cutaneous melanocytes could be estrogen-receptive. A report from the Nurses’ Health Study found that more cutaneous nevi was associated with higher free estradiol and higher risk for estrogen receptor–positive breast cancer.24 In the ophthalmology literature, there are several reports regarding choroidal melanoma in pregnant women15- 18 but evidence regarding estrogen receptors in choroidal melanoma has been inconsistent.19- 21 There have been several studies investigating the relationship between female sex hormones and cutaneous nevi, cutaneous melanoma, and choroidal melanoma10- 23 but no published reports about female reproductive factors and choroidal nevus. A meta-analysis found an association between uveal melanoma and cutaneous nevi and freckles.25 Knowing the relationship between female sex hormones and cutaneous nevus, we herein explore the relationship of female sex hormones with choroidal nevus.
The National Health and Nutrition Examination Survey (NHANES) is a population-based survey conducted annually in the United States by the Centers for Disease Control and Prevention.26 The NHANES uses a stratified multistage sampling design to select participants for a series of comprehensive health-related interviews and examinations, including a reproductive health questionnaire and retinal imaging. The purpose of this study was to characterize the relationship between female reproductive factors and choroidal nevus in a representative sample of US women.
The purpose of this cross-sectional population-based study was to explore the relationship between female reproductive factors and choroidal nevus, a precursor to uveal melanoma, in the US adult female population.
In premenopausal women, younger age at first and last births (before ages 20 and 25 years, respectively) was associated with more than 4 times higher odds of choroidal nevus compared with older age at first and last births (after age 35 years).
In postmenopausal women, higher body mass index (>30, calculated as weight in kilograms divided by height in meters squared) was associated with 2 times higher odds of choroidal nevus compared with normal body mass index (18.5-25).
Conducted in April 2015, our analysis included women aged 40 years or older from the 2005-2008 NHANES database (n = 3575) who successfully completed the retinal imaging examination (n = 2790) and participated in the reproductive health questionnaire (n = 2684). Retinal imaging was only administered by the NHANES to this age group during these years. Women without information about menstrual periods, who reported lack of regular periods for any reason besides menopause, and with history of breast, ovarian, or uterine cancer were excluded because these conditions can independently affect estrogen exposure.27 The final study population contained 2505 women. The NHANES was approved by the Centers for Disease Control and Prevention’s National Center for Health Statistics Ethics Review Board; NHANES data are subsequently deidentified and publically available online, thus, institutional review board approval is not needed to access and analyze these data. The analysis of the relationship between female reproductive factors and choroidal nevus was planned prior to analyzing the data.
The reproductive health survey included questions about age at menarche, oral contraceptive (OCP) use, number of pregnancy and live births, menopause status, hysterectomy, oophorectomy, age at natural menopause, age at hysterectomy, age at oophorectomy, and hormone replacement therapy (HRT) use. Menopause was assessed by self-report, and all women in natural menopause or who had hysterectomy or bilateral oophorectomy were considered postmenopausal. Other variables available in the reproductive health survey that were not included in this analysis were reason for no periods, number of pregnancies, breastfeeding, number of preterm births, age at hysterectomy, age at oophorectomy, endometriosis, age at endometriosis, fibroids, age at fibroids, vaginal prolapse, type of birth control, duration of birth control, reason for using HRT, and vaginal symptoms (discharge, odor, or itching). Body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) was also evaluated as a predictor variable because high BMI can lead to deposition of adipose tissue that can aromatize body androgen into estrogens, thus, obesity is associated with higher estrogen levels.28
Retinal imaging was offered to all participants 40 years and older unless they were unable to see light with both eyes open or had an eye infection. Two 45° nonmydriatic digital images were obtained from each eye using the Canon Non-Mydriatic Retinal Camera CR6-45NM (Canon). The first image was centered on the macula and the second was centered on the optic nerve. The images were graded at the University of Wisconsin retinal reading center using standardized methods.
The primary outcome variable for this study was the presence of choroidal nevus in one or both eyes. The standard definition of choroidal nevus used in population-based studies5- 8 is a melanocytic choroidal lesion of at least 500-µm diameter and of brown, slate gray, or yellow color. Images graded as “questionable choroidal nevus” were categorized into the “no choroidal nevus” group for data analysis.
The distribution of each predictor variable was compared between participants with and without choroidal nevus using the design-adjusted Rao-Scott Pearson-type χ2 and Wald tests for categorical and continuous variables, respectively. Table 1 and Table 2 include weighted percentages rather than total numbers because NHANES methodology involves oversampling certain groups, for example, Hispanic and African American individuals, and then assigns a corresponding weight to each study participant to accurately represent the true prevalence of disease in the US population. Thus, the total number divided by the sample size does not actually match with the weighted percentage. Next, a series of multivariate logistic regressions adjusted for age, race/ethnicity, and BMI were performed to identify independent reproductive predictors of choroidal nevus in premenopausal and postmenopausal women. The analysis was stratified by menopausal status because the hormonal milieu of premenopausal and postmenopausal women are markedly different, and some factors may not be as influential in premenopausal women because the systemic levels of estrogen are so high, whereas these factors may become influential after menopause, when ovarian estrogen levels drop and could potentially unmask the effects of some other influential factors. Body mass index was included in the multivariate logistic regression because BMI is associated with elevated estrogen levels as well as higher prevalence of many cancers, thus, obesity could confound the relationship between estrogen and choroidal nevus. Odds ratios (ORs) and 95% CIs were calculated. Once the association between choroidal nevus and first and last births was identified, an additional series of multivariate logistic regression analyses were performed. Sequentially, more reproductive factors were added to the regression to examine whether the findings could be associated with any of the other factors. By doing so, we identified that the association between choroidal nevus and early age at first and last births remained significant even after adjusting for multiple other factors. However, when adjusting for age at first birth and age at last birth in the same regression, the findings were no longer significant because early first birth and early last birth are colinear and cancel each other out. In an effort to most accurately calculate confidence intervals around estimates for the US population, all data analysis was performed (Stata version 12.0; Stata Statistical Software) using weighted data, calculating standard errors of population estimates using Taylor linearization methods.
The 2005-2008 NHANES yielded 2505 women aged 40 years and older who completed the retinal imaging examination; participated in the reproductive health questionnaire; did not have breast, ovarian, or uterine cancer; and were not lacking regular periods for any reason other than menopause and hysterectomy. There were 97 women with choroidal nevus in at least 1 eye. After implementing the NHANES’ weight-adjustment methods to account for the purposeful oversampling of some populations, the weighted prevalence of choroidal nevus in this study population was 4.5%; the mean age was 56.5 years and the racial/ethnic distribution was 76.9% white, 10.1% African American, 8.4% Hispanic, and 4.8% other. White race was more prevalent among women with choroidal nevus than women without choroidal nevus (90.8% vs 76.2%; P < .001), and there were no statistically significant differences in any of the other unadjusted reproductive variables between women with and without choroidal nevus (Table 1).
Multivariate logistic regressions were performed to investigate the relationship between choroidal nevus and each female reproductive factor, adjusting for age, race/ethnicity, and BMI (Table 2). Premenopausal women who first gave birth before age 20 years had 4 times higher odds of choroidal nevus than those who first gave birth after age 35 years (OR, 4.16; 95% CI, 1.29-13.45; P = .02). Women who gave birth to their last child before age 25 years also had 4 times higher odds of choroidal nevus than those who gave birth to their last child after age 35 years (OR, 4.89; 95% CI, 1.15-20.74; P = .03). Postmenopausal women with 5 to 9 years between their first and last births had 2.6 times higher odds of choroidal nevus than those with 10 or more years between their first and last births (OR, 2.63; 95% CI, 1.24-5.55; P = .01). The odds of choroidal nevus in postmenopausal women who were overweight or obese were twice as high compared with their counterparts with normal BMIs (OR, 2.11; 95% CI, 1.19-3.76; P = .01 for overweight and OR, 1.92; 95% CI, 1.13-3.25; P = .02 for obese).
The complex relationship between choroidal nevus and age at first birth, age at last birth, total parity, and duration of time between first and last births was explored via a series of multivariate logistic regressions, each adjusting for progressively more potential confounders (data not shown). Among premenopausal women, the association between choroidal nevus and younger age at first birth remained statistically significant even after further adjusting for age at menarche, total parity, and duration of time between first and last births. Similarly, the association between choroidal nevus and younger age at last birth also remained statistically significant even after further adjusting for total parity and duration of time between first and last births. In a linear regression, age at first birth and age at last birth were colinear with each other, with a coefficient of 0.56, 95% CI of 0.50 to 0.61, and P < .001.
This analysis of a representative sample of US women identified a possible association between choroidal nevus and several reproductive factors that may be surrogates for higher lifetime unopposed estrogen levels. Premenopausal women who started having children before age 20 years or finished having children before age 25 years were more than 4 times as likely to have choroidal nevus than those who started having children after age 35 years or finished having children after age 35 years, respectively. There are several plausible explanations for these findings, which largely support the hypothesis that unopposed estrogen could play a role in the pathogenesis of choroidal nevus, a risk for eventual choroidal melanoma.
The results from this study contribute to a growing body of literature about the relationship between female reproductive factors and uveal melanoma. This analysis found an association of earlier first and last births with choroidal nevus that was dependent on each other but entirely independent of total parity or time between first and last births. It can be postulated that an earlier childbearing window may result in a longer postchildbearing duration of unopposed estrogen lasting from childbearing years to the start of menopause, thereby supporting the hypothesis that a longer unopposed duration of estrogen could be a factor in the formation of choroidal nevus. This analysis also identified a possible relationship between lower total parity and choroidal nevus in postmenopausal women. Although this association was not statistically significant, it is consistent with prior studies reporting a decreased risk for uveal melanoma in women with higher parity29 and improved survival in uveal melanoma among women with children compared with those without children.30,31 However, another study reported increased uveal melanoma risk among women with children.32
The lack of association with OCPs is consistent with other studies reporting no apparent relationship between OCPs and uveal melanoma metastases.29,30,33 This analysis did not find any association between choroidal nevus and ever using any type of postmenopausal HRT, which is consistent with studies in the literature reporting no association between HRT and uveal melanoma.29,33 There were subtle associations between choroidal nevus and hysterectomy (OR, 1.33; 95% CI, 0.83-2.12) and bilateral oophorectomy (OR, 1.38; 95% CI, 0.79-2.41), but neither reached statistical significance. There are case reports of uterine and ovarian metastases from uveal melanoma34- 36 but, to our knowledge, there are no large-scale analyses on this topic.
There is abundant literature describing the role of estrogen in regulating cell proliferation and apoptosis in hormone-sensitive tissues.37 The most biologically active type of estrogen is 17-β-estradiol (E2). In premenopausal women, most E2 is produced by the ovaries and circulates in the blood. In postmenopausal women, circulating estrogen levels are low, and most E2 is produced at extragonadal sites such as the breast, brain, liver, bone, and fat. The 2 most important pathways for estrogen production in target tissues are the sulfatase pathway, where inactive steroid sulfates are the precursor for E2, and the aromatase pathway, where androgens are the precursor for E2. Adipose tissue is the source for the aromatase pathway in the production of extraovarian estrogen.38
Basic science studies have demonstrated melanocyte growth in response to estrogen stimulation.38Estrogen receptors have been identified in normal skin cells, cutaneous melanocytic nevi, and cutaneous melanomas,39,40 as well as in their counterpart tissues in the eye including normal conjunctiva, conjunctival melanocytic nevi, and conjunctival melanoma.41- 43 Clinically, girls often manifest pigmentation of cutaneous nevi during puberty,22 women using OCP or HRT can develop facial hyperpigmentation, and pregnant women can develop abdominal hyperpigmentation,23 all of which represent increased pigmentation in response to female sex hormones.44 Furthermore, cutaneous melanoma has been reported to be associated with earlier menarche, later menopause, lower parity, and older age at first birth, which support unopposed estrogen as a possible risk factor; however, the data about the association between OCP or HRT and cutaneous melanoma remain inconsistent.22,45
The complex relationship between estrogen and hormone-sensitive cancers has been well described. In the breast, ovaries, and endometrium, unopposed excess estrogen has generally been linked to increased cancer risk.46- 49 Risk factors for breast cancer include early menarche, lower parity, older age at first birth, OCP use, older age at menopause, postmenopausal estrogen and progesterone HRT, hysterectomy and oophorectomy if associated with HRT use, and obesity. Reproductive risk factors for ovarian cancer include nulliparity, lower parity, older age at first birth, postmenopausal estrogen-only HRT, and obesity, whereas OCP use is protective. Reproductive risk factors for endometrial cancer include early menarche, late menopause, nulliparity, lower parity, postmenopausal estrogen-only HRT, tamoxifen use, overweight status, and obesity, whereas OCP is protective. Conversely, in the stomach and colon, estrogen has been shown to exert protective antitumor effects by activating apoptotic signaling mediated by estrogen receptor β, inhibition of inflammatory signals, and modulation of the tumor microenvironment.50,51
Similar to the pattern seen in breast cancer and cutaneous melanoma, this analysis identified a trend toward association between choroidal nevus and earlier menarche, lower parity, OCP use, and postmenopausal estrogen-only HRT, supporting the hypothesis that estrogen could play a role in the pathogenesis of choroidal nevus. In contrast to the literature indicating that younger age at first birth is protective against cutaneous melanoma and breast and ovarian cancers, this analysis found that younger age at first and last births was associated with a higher incidence of choroidal nevus, independent of total parity, likely related to the longer total exposure to unopposed estrogen from the early childbearing window to menopause.
The association between postmenopausal obesity and choroidal nevus mirrors the known association between obesity and breast, ovarian, and endometrial cancers.52- 55 Postmenopausal women have reduced ovarian-produced estrogen, and estrogen levels after menopause depend on androgen aromatization in adipose tissue. Obese postmenopausal women have higher serum estrogen than nonobese postmenopausal women, likely related to adipose production of estrogen.56 Thus, the association between higher postmenopausal BMI and choroidal nevus could be mediated through estrogen. In the breast cancer literature, obesity is additionally implicated in the insulin and inflammatory pathways. Abdominal obesity is associated with elevated insulin that is mitogenic, antiapoptotic, and proangiogenic, and it has been associated with worse breast cancer prognosis.53 Proinflammatory mediators produced by activated macrophages in adipose tissue have also been linked to breast cancer progression.54 The relationship between postmenopausal obesity and choroidal nevus could be mediated through these insulin or inflammatory mechanisms, and this question warrants further investigation.
The limitations of our study included the NHANES’ cross-sectional design, which allows identification of associations but not causation. A Bonferroni correction was not used because some of the reproductive factors are known to be colinear with each other; thus, they are not truly being examined as independent variables. By reporting each P value, as well as demonstrating in Table 1 and Table 2 that a large number of variables were analyzed, it is acknowledged that these associations could have arisen by chance and additional studies are needed to confirm these findings. Furthermore, the NHANES does not contain data about iris nevus, conjunctival nevus, and cutaneous nevus, and it would have been useful to evaluate whether there was any association between female reproductive factors and these other melanocytic lesions. Additionally, the retinal image graders did not report any descriptive details about the nevus that could help to stratify the lesion as low or high risk. Furthermore, there is no long-term follow-up on NHANES participants to characterize whether and how these lesions evolve over time.
In summary, this large cross-sectional study of women in the United States identified associations between choroidal nevus and younger age at first birth and last birth in premenopausal women and an association between choroidal nevus and obesity in postmenopausal women. The additional possible associations with earlier menarche, lower parity, OCP use, and estrogen-only HRT did not reach statistical significance and require additional investigation. Further studies are needed to better characterize the relationship between obesity and choroidal nevus in both sexes and identify possible mechanisms that may be mediated through gonadal or extragonadal estrogen or pathways such as insulin or inflammation. Overall, the results of this study support the hypothesis that estrogen could play a role in the pathogenesis of choroidal nevus, and it may be prudent for ophthalmologists to have a higher index of suspicion for choroidal nevus and possible choroidal melanoma if they notice a pigmented fundus lesion in female patients who have no children, have had children at a younger age, or manifest postmenopausal obesity.
Corresponding Author: Mary Qiu, MD, Wilmer Eye Institute, 600 N Wolfe St, Wilmer B-29, Baltimore, MD 21287 (email@example.com).
Submitted for Publication: May 12, 2015; final revision received July 16, 2015; accepted July 21, 2015.
Published Online: September 24, 2015. doi:10.1001/jamaophthalmol.2015.3178.
Author Contributions: Dr Qiu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Qiu.
Acquisition, analysis, or interpretation of data: Both authors.
Drafting of the manuscript: Qiu.
Critical revision of the manuscript for important intellectual content: Both authors.
Statistical analysis: Qiu.
Obtained funding: Shields.
Administrative, technical, or material support: Both authors.
Study supervision: Shields.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: Support for this study was provided by the Eye Tumor Research Foundation, Philadelphia, Pennsylvania (Dr Shields).
Role of the Funder/Sponsor: The Eye Tumor Research Foundation had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.