Gynecologic and reproductive outcomes in fibrous dysplasia/McCune-Albright syndrome

McCune-Albright syndrome (MAS)(ORPHA:562) is a rare disorder arising from somatic gain-of-function mutations in Gα_s [1]. Disease presents along a broad spectrum that includes a variable combination of fibrous dysplasia of bone (FD), hyperpigmented skin macules, and hyperfunctioning endocrinopathies [2]. A hallmark feature of MAS is autonomous estrogen-secreting ovarian cysts [3]. Girls typically present in early childhood with signs of episodic estrogen exposure, including breast development, growth acceleration, and vaginal bleeding, which resolve in the interval between cysts [4, 5]. Gonadotropin levels are typically suppressed when estradiol levels are elevated; however, prolonged exposure to high estradiol levels can mature the hypothalamic pituitary axis, leading to secondary gonadotropin-dependent precocious puberty. Treatment with aromatase inhibitors, alone or in combination with gonadotropin releasing hormone agonists, is typically effective at preventing progressive pubertal development during childhood.

While the presentation, natural history, and clinical management of MAS-associated precocious puberty has been well-characterized, ovarian function in adulthood is poorly understood. It is unknown if women with a history of MAS-associated precocious puberty are at risk for gynecologic disease. Effects of persistent ovarian activation on fertility and childbearing, as well as the effects of pregnancy on skeletal outcomes, have not been determined. The obstetrical literature in FD/MAS is limited to case reports of skeletal complications during pregnancy, including increased FD-related bone pain [6, 7], aneurysmal bone cysts [6, 8, 9], and malignant transformation of FD lesions [10]. This has led some to speculate that pregnancy increases FD activity, placing women at risk for poor skeletal outcomes [6, 7, 11]. These knowledge gaps in gynecologic and reproductive outcomes are a significant source of concern for patients and families affected by FD/MAS.

The purpose of this study was to evaluate gynecologic and reproductive outcomes in a cohort of adult women with FD/MAS.

Subjects were evaluated at the NIH Clinical Center between 1998 and 2015 as part of a longstanding FD/MAS natural history study. The protocol was approved by the NIDCR Institutional Review Board, and informed consent/assent was obtained from all subjects and/or their guardians. Subjects underwent history and physical exam, pelvic ultrasonography, and biochemistries to include LH, FSH, and estradiol. The diagnosis of FD/MAS was made on clinical grounds based on the presence of 2 or more characteristic features, with molecular testing as needed, according to previously published guidelines [2].

Attempts were made to contact all 90 women who had been seen in the natural history protocol to obtain detailed gynecologic and reproductive histories, and responses were received from 39 subjects. Five interviews were conducted at the NIH Clinical Center, and the remainder were performed by telephone. Subject characteristics are reported in Table 1. All subjects met criteria for clinical diagnosis of FD/MAS based on the presence of 2 or more characteristic features (Table 1). In addition, GNAS mutation testing from FD tissue was available for 15 subjects; in 7 subjects a mutation was identified at the R201C position, and in 8 subjects a mutation was identified at the R201H position. The mean total length of follow-up was 11.5 years (standard deviation 5.1 years, range 0–16 years).

Pelvic ultrasounds were reviewed by a single reader (DOC). Biochemistries were drawn within 48 h of pelvic ultrasound. Biochemical assays were carried out by the Department of Laboratory Medicine at the NIH Clinical Center. From 2008 to 2015, LH and FSH were measured using Immulite (Diagnostic Products Corporation, Los Angeles, CA) competitive immunoassay. For LH, intra-assay coefficient of variation was 3.7%, inter-assay coefficient of variation was 6.7%, and lower limit of detection was 0.1 mIU/mL. For FSH, intra-assay coefficient of variation was 3.2%, inter-assay coefficient of variation was 4.8%, and lower limit of detection was 0.1 mIU/mL. Estradiol was measured using Roche Cobas e601 analyzer (Roche Diagnostics, Indianapolis, IN) electrochemiluminescence immunoassay. Intra-assay coefficient of variation was 1.9%, inter-assay coefficient of variation was 3.1%, and lower limit of detection was 4 pg/mL. For assays run prior to 2008, LH and FSH were measured using a Microparticle Enzyme Immunoassay (Abbott Laboratories, Abbott Park, IL). For FSH, intra-assay coefficient of variation was 4%, inter-assay coefficient of variation was 3%, and lower limit of detection was 0.37 mIU/mL. For LH, intra-assay coefficient of variation was 5%, inter-assay coefficient of variation was 4%, and lower limit of detection was 0.5 mIU/mL. Estradiol was measured using Immulite (Diagnostic Products Corporation, Los Angeles, CA) competitive immunoassay. Intra-assay coefficient of variation was 6.6%, and lower limit of detection was 15 pg/mL. In pre-menopausal women, levels vary widely according to menstrual cycle, but are typically in the range of 15–350 pg/mL for estradiol, 1–24 U/L for FSH, and 1–100 U/L for LH.

Statistical tests were performed using GraphPad Prism 7.03 (San Diego, CA). Comparisons between groups were performed using t-tests and Mann-Whitney tests, depending on the normality of the distribution. Results are reported using mean (standard deviation) or median (inter-quartile range)(IQR) as described, depending on the normality of the distribution.

Findings from this large series demonstrate that autonomous ovarian activation in MAS persists into adulthood, with significant effects on gynecologic function and reproduction. In our cohort, abnormal uterine bleeding resulted in significant morbidity, including the need for blood transfusions in multiple subjects. This is consistent with previous case reports and small series reporting menometrorrhagia [14‐17]. Treatment for abnormal uterine bleeding in this cohort included primarily oral contraceptive pills and surgical management. A striking finding was the unusually young age at which hysterectomies were performed. Most subjects were under age 35 years at the time of surgery, likely reflecting the severity of abnormal uterine bleeding. Subjects reported only infrequent use of levonorgestrel intrauterine devices, now considered first-line management for abnormal uterine bleeding in women who do not desire pregnancy [18]. It is possible that increased use of medical therapies that act locally at the level of the uterine endometrium, such as intrauterine devices or ulipristal acetate for women co-morbid leiomyomas, may improve gynecologic morbidity in women with MAS, potentially decreasing or preventing the need for surgical intervention [19].

Ovarian cysts persisted into adulthood in this cohort, resulting in hyperestrogenism and suppression of pituitary gonadotropin production in some patients. Implications of hyperestrogenism in women with MAS are unclear. MAS has recently been associated with an increased risk of breast cancer, particularly in younger women [20]. Precocious puberty has been identified as a risk factor for breast cancer within the MAS population, potentially reflecting the effects of long-term exposure to increased circulating estrogen.

This is the largest series to report data on fertility in women with MAS. Previous reports have been limited to single cases and small series, which have reported both spontaneous pregnancies [6, 8‐10, 21‐24] and infertility [16, 17, 25, 26]. Findings from this cohort included an infertility prevalence of 43%, which is substantially increased over the U.S. national average of 10.9% [27]. Thirty-five percent of pregnancies in this cohort were unplanned, which is only mildly decreased in comparison to the 49% U.S. national average [28]. Taken together, these findings suggest that fertility is impaired in women with MAS, however the possibility remains for spontaneous conception. The mechanism of infertility is likely related in part to anovulatory cycles resulting from autonomous ovarian activity [29], and it is possible that conception may not be impaired during intervals without ovarian activation. This could potentially account for the relatively robust prevalence of unplanned pregnancies in this cohort, highlighting the importance of contraceptive care in the MAS population. It should be noted that none of the women in this study sought reproductive assistance. Subjects were not specifically questioned about their decision-making, however concerns about skeletal effects of pregnancy and reproductive interventions may have played a role. It is likely that reproductive assistance may have increased pregnancy rates in this cohort, and should be recommended at an appropriate point to women with MAS who desire pregnancy.

If anovulation is the primary cause of infertility in MAS, ovarian cyst frequency may predict an individual’s degree of fertility impairment. There are 4 reports of women with frequent ovarian activation who underwent unilateral oophorectomy in attempts to improve contralateral ovarian function and fertility [16, 17, 25, 26]. All women reported improvement in menses, and 2 achieved spontaneous pregnancy. It is important to note that success with this approach may be predicated on the laterality of ovarian involvement, and that a unilateral oophorectomy may be less likely to improve contralateral function in women with bilateral disease. In addition, decreased ovarian reserve is an established risk of pelvic surgery [30], which is particularly relevant given the frequent and often inappropriate pelvic surgeries performed in women with MAS [31]. Two women in our series developed primary ovarian insufficiency after ovarian surgery, including one who had resolution of precocious puberty after oophorectomy, suggesting she had unilateral involvement. These findings suggest that unilateral oophorectomy may be helpful for treatment of MAS-associated infertility in select cases, however this approach should be undertaken with extreme caution with consideration of the potential for harm to fertility, and only under the supervision of an experienced reproductive specialist.

The effects of menstruation and pregnancy on FD has been an area of much speculation and concern for patients. Case reports have described aneurysmal bone cysts and malignant transformation in pregnant women, fueling conjecture that pregnancy hormones may affect adversely affect the metabolic activity of FD lesions [6‐11]. Osada et al. reported increased bone turnover during pregnancy in women with FD, however these changes were consistent with the bone turnover variations typical of pregnancy and the post-partum period [21]. Skeletal complications are an established morbidity in FD, and it is possible that their reported association with pregnancy is mediated in part by reporting bias, due to the current literature dominated by case reports. In this cohort study, there were no consistent effects of pregnancy or menstruation on bone pain, and no association with skeletal morbidity. Currently there is no clear evidence to establish a causative relationship between pregnancy or menstruation and skeletal outcomes in women with FD/MAS.

Strengths of this study include the large number of subjects and extensive follow-up. This is the first cohort study to establish gynecologic disease as a highly prevalent source of morbidity in MAS, to evaluate fertility and reproductive outcomes in women with MAS, and to report on skeletal outcomes in pregnancy. This is also the first report of primary ovarian insufficiency as a potential consequence of pelvic surgery in MAS, a novel finding that may directly inform gynecologic management. Limitations include the retrospective nature of the study design. Clinical outcomes data were not collected at same time as radiographic data and biochemical data, and are subjected to recall bias. Interpretation of ultrasound and biochemical data is limited, because studies were obtained in a cross-sectional manner and were not standardized according to menstrual cycles. Additional, prospective studies are needed to characterize the prevalence and frequency of ovarian cysts, which will provide important context for these clinical findings. Prospective studies should correlate gynecologic data with skeletal outcomes and surrogate endpoints, such as bone turnover markers, to more fully investigate the effects of these hormonal changes on mineral metabolism. Selection bias likely resulted in over-representation of more severely affected patients among respondents. The high prevalence of precocious puberty among respondents likely reflects this selection bias. This data should therefore be generalized only to patients with known MAS ovarian involvement.

Abnormal uterine bleeding is common in MAS and may result in significant morbidity, including severe anemia and hysterectomies at an early age. In this series, the prevalence of infertility was substantially increased over the national average, however the prevalence of unplanned pregnancies was only mildly reduced. Menses and pregnancy were not consistently associated with increased FD-related bone pain or skeletal complications. These findings will inform family planning for patients with FD/MAS by highlighting both the potential for spontaneous pregnancy, as well as possibility of impaired fertility and need for reproductive assistance.