Background
Functional hypothalamic amenorrhea is a disorder characterized by cessation of menstrual cycles in the absence of organic disease. In most patients, it occurs in adult life after a stressful event and may be related to a condition of mild chronic energy deprivation. The endocrine pattern is characterized by low estrogen levels with an absent response to a progestogen challenge test and low-normal gonadotropin levels. A few studies have shown that some of these women may have some features of polycystic ovary syndrome; these features include an increased androgen response to gonadotropins, increased anti-Mullerian hormone levels, and altered ovarian morphology or increased ovarian size. These findings suggest a link between these 2 completely different disorders: functional hypothalamic amenorrhea and polycystic ovary syndrome. The importance of the possible coexistence of these disorders in some women is important for follow-up of these women and in their treatment if they desire to become pregnant.
Objective
To determine whether a subgroup of well-characterized women with functional hypothalamic amenorrhea may have the coexistence of polycystic ovary syndrome.
Study Design
Retrospective analysis of women with functional hypothalamic amenorrhea. Forty consecutive patients and 28 normal age-matched control patients were studied. Blood was obtained for serum anti-Mullerian hormone, androgens, and other hormone levels and all women had ovarian ultrasonographic measurements.
Results
In the entire group of women with functional hypothalamic amenorrhea, anti-Mullerian hormone and ovarian volume were greater than in control patients. In 13 patients (32.5%), anti-Mullerian hormone was elevated (>4.7 ng/mL, levels consistent with polycystic ovary syndrome) and in this group, ovarian volume was significantly greater than in the remaining patients with functional hypothalamic amenorrhea. Four of the 13 women with functional hypothalamic amenorrhea who had elevated anti-Mullerian hormone levels (10%), also had ovarian volume ≥10 cc (consistent with polycystic ovarian syndrome). In these patients all studied androgens were in the upper normal range or slightly elevated despite low-normal gonadotropins; mean total testosterone was significantly greater than in the other patients with increased anti-Mullerian hormone values with normal ovarian size ( P <.05.) Six other women with functional hypothalamic amenorrhea who had increased anti-Mullerian hormone also had isolated elevations of some androgen levels, but mean testosterone and ovarian size were normal.
Conclusions
As many as 10% of women with functional hypothalamic amenorrhea may have the coexistence of polycystic ovary syndrome. Because no signs or symptoms of this disorder were reported by these women before the appearance of the amenorrhea, it does not seem to be a coincidental relationship. The possibility that functional hypothalamic amenorrhea favors the appearance of polycystic ovary syndrome or more likely, that a mild (ovulatory) phenotype of polycystic ovary syndrome predisposes to the development of functional hypothalamic amenorrhea should be considered. Possible mechanisms are unclear and need to be investigated but may involve common vulnerabilities such as psychologic and mood disturbances.
Functional hypothalamic amenorrhea (FHA) is a disorder characterized by dysfunction of the hypothalamic–pituitary–ovarian axis, leading to anovulation and cessation of menstrual cycles in the absence of organic disease. In most instances, FHA occurs in adult life after a stressful event and may be related to a condition of mild chronic energy deprivation from excess energy expenditure, stress, and/or insufficient nutritional intake. In many cases, there is no obvious reason found, and the FHA is considered idiopathic.
The endocrine pattern is characterized by low estrogen levels with an absent response to a progestogen challenge test and low−normal gonadotropin levels. It has been suggested that increased cortisol, decreased kisspeptin, and low leptin levels may participate in the pathogenesis of this functional form of hypogonadotropic hypogonadism.
It is not generally appreciated that some women with FHA may have features of polycystic ovary syndrome (PCOS) and that these disorders may coexist. It has been reported that some patients with FHA may have subtle increases in serum androgens during gonadotropin administration, similar to the response seen in women with PCOS, and in addition may have ovarian morphology similar to women with PCOS. Similar studies have also reported an evolving picture of PCOS in women with FHA who were using pulsatile gonadotrophin-releasing hormone therapy. We also have observed that some women with FHA exhibit increased androgen responses to controlled ovarian stimulation and in long-term follow-up of the women described have found that some of these women developed features of PCOS as recovery of the hypothalamic–pituitary–ovarian axis occurred.
Some studies have reported normal levels of serum anti-Mullerian hormone (AMH) in FHA, but in most studies, serum AMH has been reported to be increased. Increased AMH levels have been considered to be an important diagnostic marker for the altered ovarian morphology in women with PCOS. Further confusing the picture in FHA, altered cystic morphology (not characteristic of polycystic ovaries) has been reported, which has been described to be multicystic or multifollicular, which may or may not result in slightly enlarged ovaries, but usually is not confused with polycystic ovaries. In 1 report, as many as 41% of patients with FHA were reported to have enlarged ovaries (>10 cc).
In this study, we wished to further characterize the possible relationship between FHA and PCOS and to estimate how frequently the coexistence may occur. In the hypoestrogenic state of FHA, we used serum AMH and ovarian parameters on ultrasound as possible markers. In so doing, we hoped to provide further insight into the possible connection between FHA and PCOS. A retrospective analysis of women with well-characterized FHA was performed.
Materials and Methods
We wished to compare well-characterized patients with FHA with a normal, age-matched control population and then determine features of these women with FHA (namely AMH levels and ovarian ultrasound findings) that resemble those of women with PCOS. Forty consecutive women with a diagnosis of FHA, aged 17−33 years (mean age 23.7 ± 5 years), were evaluated retrospectively. These patients were referred between 2012 and 2014 to the Endocrine Unit of the University of Palermo and the Department of Obstetrics and Gynecology of the University of Pisa, Pisa, Italy, because of secondary amenorrhea. All of these women reported fairly regular cycles before their complaint of amenorrhea. None of the women had severe weight loss or known eating disorders, and diagnosis of FHA was considered idiopathic. Some of these women had been treated previously with various therapies for amenorrhea but had not received any treatment for ≥3 months before evaluation in this study.
The diagnosis of FHA was based on the finding of secondary amenorrhea not associated with any organic disease, which was confirmed by the finding of hypoestrogenism (low estradiol levels <30 pg/mL; and no response to a progestogen challenge test). The women had normal levels of circulating prolactin and thyroid-stimulating hormone, and uterine disease was excluded by ultrasonography. In all patients magnetic resonance imaging of the brain was performed to exclude hypothalamic or pituitary disease.
For controls, we selected a group of 28 healthy women from Palermo who were age matched (mean age 23.4 ± 5 years) with the group of women with FHA. The controls were recruited from family members of hospital coworkers and had to have regular menses, no symptoms of hyperandrogenism (acne or hirsutism), and normal androgen levels. Normal menses were defined as cycles lasting 25–34 days. Patients and controls were all white Italian women.
All subjects underwent a complete history and physical examination, biochemical analyses, and transvaginal ultrasonography. Height and weight were recorded and body mass index (BMI) was calculated as kg/m 2 . None of patients or controls was taking medications for at least 3 months before entering the study. The procedures were in accordance with the Helsinki Declaration of 1975 as revised in 1983 and this study was approved by the local Ethic Council.
Laboratory analyses
In all patients and controls serum levels of luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol, total testosterone (T), dehydroepiandrosterone sulfate (DHEAS), androstenedione, cortisol, and AMH were evaluated. In controls, serum hormones were determined on days 3−5 of the cycle. Serum hormone levels were quantified by well-established methods that had been validated previously in our laboratories. LH, FSH, and estradiol were measured by traditional assays. Total T concentrations were determined by the use of a competitive immunoassay (Johnson & Johnson – Ortho Clinical Inc., Rochester, NY). DHEAS concentrations were determined by the use of radioimmunoassay (Orion Diagnostics, Espoo, Finland). Androstenedione was measured by an enzyme-linked immunosorbent assay from DRG Instruments Gmbh, Marburg, Germany. AMH was measured by a commercial enzyme-linked immunosorbent assay, the AMH Gen II (Beckman Coulter, Brea, CA). The conversion of AMH in ng/mL to pmol/L requires that values be multiplied by 7.143. In all assays, intraassay and interassay coefficients of variation did not exceed 6% and 15%, respectively.
Biochemical hyperandrogenism was defined as serum T ≥60 ng/dL (≥2.08 nmol/L) and/or serum DHEAS ≥3 μg/mL (≥7.8 mmol/L) and/or serum androstenedione ≥3.2 ng/mL. These values for hyperandrogenism have been validated previously in adult women with the use of the previously described assays. Increased AMH levels were defined as serum AMH >4.7 ng/mL. This represented values above the upper 95% confidence intervals of our control population, and is the value shown by us (by ROC analyses) and others by meta-analyses to suggest the finding of polycystic ovaries.
Ovarian ultrasound
In all patients and in control subjects (between days 3 and 5 of the cycle), ovarian morphology was assessed by transvaginal ultrasound. In both centers, the same machine (MyLab 50 Xvision; Esaote SpA, Genoa, Italy) was used but transducer frequency changed from 4−6 mHz to 8−10 mHz over the time the subjects were assessed, and this varied at the 2 different centers.
Ovarian volume was reported in all patients and was calculated by the formula π/6 (D 1 × D 2 × D 3 ) where the dimensions (D) of length, width, and thickness were used. The size of both ovaries was assessed, and mean ovarian size was calculated. In no instance was there the need to repeat ovarian ultrasound because of the finding of a dominant follicle. Increased ovarian size was defined as an ovarian volume ≥8.8 cc, which represented the upper 95% confidence intervals of our control population and according to our own ROC analyses for the ultrasound diagnosis of polycystic ovaries. We could not rely on follicular or antral follicle counts because of the variability of this measure over time and the use of 2 different ultrasound transducers.
The addition of a PCOS control group was not considered necessary and the criteria for normal androgens, AMH levels, and ovarian volume have been described previously in this article. Such data on well-characterized patients may be found in our recent paper.
Statistical analyses
Statistical analyses were performed with the use of Statview 5.0 (SAS Institute, Cary, NC). Univariate analyses were performed with the unpaired t test for the numeric variables, whereas the differences in the prevalence for the nominal variables were analyzed by the χ 2 test. Group means were compared using analysis of variance with post hoc least squares means pairwise comparisons (after log transformation of the values). Correlations were analyzed by Pearson test. We took the position that for women with FHA to possibly have coexisting polycystic ovaries or PCOS, AMH had to be elevated as well as having an increased ovarian volume suggestive of PCOS. All results are expressed as mean ± SD.
Results
In Tables 1 and 2 , the main clinical and hormonal data of patients with FHA are depicted. In comparison with normal control patients, the entire group of women with FHA had lower BMI; ( P < .001) and lower LH and estradiol levels ( P < .001) but greater AMH ( P < .05) and androstenedione ( P < .01) values and larger ovarian volumes ( P < .05). FSH, T, DHEAS, cortisol, and thyroid hormone values were similar in the 2 groups. The difference in AMH values between controls and FHA remained significant ( P < .01) after we controlled for BMI. None of the controls had increased ovarian volume (mean 4.3 ± 2 cc) or high AMH values (mean: 2.8 ± 1 ng/mL).
| Age, y | BMI, kg/m 2 | LH, mIU/mL | FSH, mIU/mL | E2, pg/mL | Cortisol, ng/mL | Ovarian size, cc | |
|---|---|---|---|---|---|---|---|
| FHA, n = 40 | 23.7 ± 5.4 | 19.6 ± 2.4 | 2.1 ± 1.4 | 5.6 ± 2.2 | 22 ± 8 | 16.1 ± 4.2 | 6.1 ± 3 |
| Ovulatory controls, n = 28 | 23.4 ± 5 | 23.1 ± 4 | 5.6 ± 1.5 | 5.4 ± 1.6 | 43 ± 15 | 15.8 ± 4 | 4.3 ± 2 |
| NS | P < .001 | P < .001 | NS | P < .001 | NS | P < .05 |
| AMH, ng/mL | T, ng/dL | DHEAS, μg/mL | A, ng/mL | |
|---|---|---|---|---|
| FHA | 3.9 ± 2.7 | 38 ± 18 | 1.9±1 | 2.9 ± 1.1 |
| Ovulatory controls | 2.8 ± 1 | 31 ± 13 | 1.9 ± 0.6 | 2.1 ± 0.5 |
| P < .05 | NS | NS | P < .01 |
In the entire group of patients, AMH levels correlated significantly with ovarian volume (r = 0.47, P < .01) and with T (r = 0.35, P < .05) and DHEAS (r = 0.34, P < .05); however, there was no correlation between AMH and age, BMI, LH, FSH, estradiol, androstenedione, and cortisol.
Thirteen of the 40 women with FHA (32.5%) had elevated AMH levels. In Tables 3 and 4 , the hormonal profiles and ovarian size of the subgroups of women with FHA with increased or normal AMH are compared. The 2 subgroups had similar hormone profiles but ovarian size was significantly greater in FHA patients with increased AMH values.
