Authors: Johannes Ott, Rosa Loimer, Rodrig Marculescu, Geoffroy Robin, Didier Dewailly, Marlene Hager
Categories: Research, Hypogonadotropic hypogonadism, Gonadotropin releasing hormone, Luteinizing hormone
Source: Reproductive Biology and Endocrinology : RB&E
Authors: Johannes Ott, Rosa Loimer, Rodrig Marculescu, Geoffroy Robin, Didier Dewailly, Marlene Hager
Almost half of patients with functional hypothalamic amenorrhea (FHA) show polycystic ovarian morphology (PCOM) on the ultrasound, which leads to a diagnostic confusion. Although FHA and polycystic ovarian syndrome (PCOS) have been thought to co-exist and some FHA-patients seem to have had PCOS before developing FHA, respectively, once hypothalamic inhibition proceeds, the FHA phenotype predominates over the PCOS features, except from PCOM. This connection has never been shown longitudinally. Furthermore, it is still not clear if FHA-PCOM is actually related to preexisting PCOS or if these women constitute their very own heterogeneous subgroup. Thus, the aims of this study were to evaluate changes in hormonal parameters and PCOM after remission and to provide further insight into pathophysiological processes of PCOM in FHA.
Monocentric retrospective cohort study. Sixty women with FHA in remission were included. While anti-mullerian hormone (AMH) was the main outcome parameter, we also analyzed total testosterone, luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E2), sex hormone-binding globulin (SHBG) and dehydroepiandrosterone sulfate (DHEAS). PCOM was diagnosed using ultrasound.
At baseline, FHA-PCOM patients revealed higher baseline prolactin (p = 0.029) and AMH levels (p < 0.001). At follow-up, compared to women without PCOM, these women had higher PCOM prevalence (48.1% versus 0%, p < 0.001), higher AMH levels (median 6.49 ng/mL, IQR 4.74–7.95 versus median 2.25 ng/mL, IQR 2.0-2.71; p < 0.001) and higher PCOS prevalence (22.2% versus 0%, p = 0.006). While overall median AMH levels increased significantly, FHA-PCOM patients revealed a significant median decrease in AMH levels (median AMH dynamics − 0.82 ng/mL, IQR – 2.30 - -0.16; p < 0.001).
Our data support the hypothesis that relative FSH deficiency in hypothalamic dysfunction can lead to lower AMH levels. In contrast, the decline in AMH levels and the resolution of PCOM in the FHA-PCOM group may indicate a reversible state of ovarian hyperactivation during FHA.
Not applicable.
Functional hypothalamic amenorrhea (FHA) is among the main causes of non-physiological secondary amenorrhea and is primarily due to stress, reduced caloric intake, excessive physical exercise, or a combination of these factors [1, 2]. On ultrasound, up to 47% of affected women show polycystic ovarian morphology (PCOM) [3–8]. In detail, when selecting only studies with sound definitions of FHA and PCOM, the PCOM prevalence in FHA varied from 41.9% to 46.7% [4, 5, 7, 8]. This creates clinically relevant confusion with a potential for misdiagnoses [9, 10].
As recently reviewed, the high prevalence of PCOM in women with FHA is considerably surprising [3]. It exceeds the generally agreed prevalence of about 30% in the general female population of reproductive age [11]. Among these women, not all reveal polycystic ovary syndrome (PCOS). Rather, a general prevalence of “silent PCOM” (PCOM without PCOS) of about 10% was reported [3, 12].
Several differences between FHA patients with and without PCOM were found. For example, women with FHA and PCOM revealed a higher body mass index (BMI) [4, 5], higher Anti-Mullerian hormone (AMH) [4, 5, 7, 13] and testosterone levels [13], a higher homeostasis model assessment of insulin resistance (HOMA) [13] as well as lower sex hormone binding globulin (SHBG) [4] and follicle stimulating hormone (FSH) levels [5]. Some of these findings can be considered typical for PCOS regarding both the hypothalamic-pituitary-ovarian axis and the metabolic situation and, thus, suggest that at least some of the FHA patients with PCOM originally had PCOS before developing FHA [3]. Recently it has been discussed that PCOS and FHA might co-exist in a woman initially [3]. However, once hypothalamic inhibition proceeds, the FHA phenotype predominates and features of PCOS are attenuated, except PCOM [3].
Nonetheless, also given the fact that the PCOM prevalence in FHA of nearly 50% exceeds the prevalence of PCOM in the general female population [14], it has been suggested that FHA-PCOM women might in fact constitute a heterogeneous subgroup. In the study of Robin et al. [15] using cluster analysis, PCOM was considered incidental in the majority of their cases while a minority shared some features with PCOS (respectively, 38% and 10% of their total FHA population). Notably, no longitudinal data have been presented so far, which could prove or refute the hypothesis that in some FHA-PCOM women, PCOS preceded FHA and might be present again, once FHA resolves. Ideally, such studies should be prospective, including PCOS women before they develop FHA and after FHA resolves, with sequential assessments of the follicle number per ovary (FNPO) at ultrasound and hormonal parameters. However, obtaining such data is almost impossible for obvious reasons. Nonetheless, providing that PCOM is strictly defined at baseline, the follow up of FHA-PCOM women who recovered menses, compared to FHA-non PCOM patients, may afford relevant information. In the present study, we report longitudinal data about such women. We believe that this retrospective study will provide further insights in the pathophysiological meaning of PCOM in FHA patients. Moreover, the data might help in counselling affected women.
In this monocentric retrospective cohort study, which was conducted at the Clinical Division of Gynecologic Endocrinology and Reproductive Medicine of the Medical University of Vienna, Austria, data from 60 women with FHA in remission were included. Recovery of menses was defined as three consecutive spontaneous menstrual bleeding within six months, which is in accordance with previous reports [16, 17] as also reviewed recently [18]. All patients had been seen at the department from January 2017 to December 2024. PCOM was defined as follicle number per ovary (FNPO) >12 and/or an ovarian area ≥ 5.5 cm^2^, identified through the utilization of an ultrasound machine with a frequency range less than 8 MHz according to the recommendations of the international expert panel “Task Force” [11]. FHA was considered if the patients fulfilled the following secondary amenorrhea for at least three consecutive months; a negative progestogen challenge test; with context of disordered eating, insufficient caloric intake, excessive physical activity and/or psychological stress [19, 20]. Pregnancy, hypothyroidism and any organ-related pituitary dysfunction (on magnetic resonance imaging) had to be excluded. Women were included when their cycles returned back to normal or to oligomenorrhea after the cause of FHA dissolved within the time span stated above.
The AKIM-software (SAP-based patient management system at the Medical University of Vienna) was used for data acquisition. The main outcome parameter was AMH. In addition, serum levels of total testosterone, luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E2), sex hormone-binding globulin (SHBG) and dehydroepiandrosterone sulfate (DHEAS) were analyzed. At baseline, the bloodwork was obtained in the amenorrheic state. In contrast, after recovery of menses, blood was drawn during the early follicular phase (cycle days 2–5). The hormonal workup was determined at the Department of Laboratory Medicine, Medical University of Vienna, according to ISO 15,189 quality standards. As reported previously Cobas electrochemiluminescence immunoassays (ECLIA) were performed on Cobas e 602 analyzers (Roche, Mannheim, Germany) for the determination of serum estradiol, FSH, LH, AMH, testosterone, DHEA-S, thyroid-stimulating hormone (TSH) and SHBG [21, 22]. The corresponding maximal coefficients of variation of these assays were 10.6, 4.5, 2.2, 3.5, 14.5, 2.7, 11.9 and 4.0%, respectively.
The patients’ age and body mass index (BMI) at the time of the initial diagnosis and at follow-up after recovery of menses were also included. For the evaluation of PCOM, a transvaginal sonography was performed using an Aloka Prosound 6 ultrasound machine (Wiener Neudorf, Austria; frequency range 3.0–7.5 MHz).
The focus was on whether PCOS had developed after recovery of menses. PCOS was defined based on the revised Rotterdam criteria, where presence of presence of two of the following three findings are signs of clinical or biochemical hyperandrogenism; chronic ovulatory dysfunction; and PCOM [23]. PCOS phenotypes were also phenotype A (PCOM + hyperandrogenism + oligo-/anovulation), phenotype B (hyperandrogenism + oligo-/anovulation), phenotype C (PCOM + hyperandrogenism) and phenotype D (PCOM + oligo-/anovulation) [24].
Continuous and categorical parameters are presented as median values (interquartile ranges, IQR) and numbers (%), respectively. Mann-Whitney U tests for independent variables were used to compare continuous variables, and chi square or Fisher’s exact tests were used to compare categorical variables between two groups. To compare numerical parameters obtained longitudinally within a group of patients, Mann-Whitney U tests for dependent variables were applied. The IBM Statistical Package for Social Science software (SPSS 25.0) was used for all statistical tests. P-values < 0.05 were considered statistically significant.
Basal patient characteristics at the time of initial diagnosis of FHA are provided in Table 1. At baseline (i.e., before remission), 27 women (45%) showed FHA in combination with a PCOM, whereas 33 patients (55%) did not show any signs of PCOM. FHA patients with PCOM revealed significantly higher baseline prolactin (p = 0.029) and AMH levels (p < 0.001; Table 2).Table 1Basic patient characteristics and results of hormonal testing in FHA patients with and without PCOM at the time of FHA diagnosisPCOM (n = 27)nonPCOM (n = 33)pAge (years)^a^25 (22;28)25 (22;31)0.602BMI (kg/m^2^)^a^22.0 (18.8;23.3)20.0 (18.1;23.1)0.298Causes for FHAWeight loss/underweight^b, c^6 (22.2)13 (39.4)0.176Eating disorder^b, c^6 (22.2)9 (27.3)0.768Excessive exercise^b, c^7 (25.9)13 (39.4)0.409Stress^b, c^15 (55.6)11 (33.3)0.117Duration since last menstrual bleeding (months)^a^13 (9;17)12 (9;17)0.812Data are provided as ^a^median (IQR) or ^b^n (%); ^c^multiple mentions for causes of FHA possible
Table 2Patient characteristics and results of hormonal testing in FHA patients with and without initial PCOM at baseline and after recovery of mensesBaseline PCOM(n = 27)Baseline non PCOM (n = 33) p BMI (kg/m^2^)^a^Baseline22.0 (18.8;23.3)20.0 (18.1;23.1)0.602Follow-up22.0 (19.8;23.1)20.2 (19.2;23.3)0.988p0.051< 0.001BMI dynamics (kg/m^2^)From baseline to follow-up0 (0;1.2)0.8 (0;1.4)0.236PCOM prevalence^b^Baseline27 (100)0-Follow-up13 (48.1)0< 0.001TSH (IU/mL)^a^Baseline1.70 (1.36;2.36)1.54 (1.35;2.00)0.345Follow-up1.72 (1.45;2.14)1.70 (1.34;1.86)0.226p0.6750.580FSH (mIU/mL)^a^Baseline4.6 (3.6;6.5)4.7 (2.6;6.8)0.661Follow-up7.0 (6.2;8.2)7.4 (6.3;8.4)0.475p< 0.001< 0.001LH (mIU/mL)^a^Baseline3.4 (1.7;4.0)2.7 (1.5;4.7)0.471Follow-up7.6 (6.4;8.8)6.8 (5.8;8.0)0.072p< 0.001< 0.001Prolactin (ng/mL)^a^Baseline10.2 (7.9;13.1)7.4 (6.1;11.2)0.029Follow-up12.4 (10.7;15.7)11.4 (9.7;14.0)0.206p0.004< 0.001Estradiol (pg/mL)^a^Baseline16.0 (5.0;24.0)20.0 (5.5;24.5)0.964Follow-up40.0 (32.0;55.0)37.0 (31.0;53.5)0.603p< 0.001< 0.001Testosterone (pg/mL)^a^Baseline0.21 (0.16;0.31)0.22 (0.13;0.29)0.744Follow-up0.25 (0.20;0.35)0.24 (0.19;0.28)0.360p< 0.0010.028DHEAS (µg/mL)^a^Baseline1.97 (1.48;3.05)2.03 (1.60;2.97)0.705Follow-up2.39 (1.80;2.96)2.17 (1.69;2.74)0.603p0.1270.386SHBG (nmol/L)^a^Baseline78.6 (49.1;100.6)86.7 64.6;120.9)0.1156Follow-up73.3 (51.4;95.2)81.7 (70.7;92.6)0.147p0.1600.066AMH (ng/mL)^a^Baseline6.38 (4.95;10.52)2.16 (1.83;2.62)< 0.001Follow-up6.49 (4.74;7.95)2.25 (2.0;2.71)< 0.001p0.0050.004AMH dynamics (ng/mL)^a^From baseline to follow-up−0.82 (−2.30;−0.16)0.09 (−0.02;0.30)< 0.001PCOS prevalence^b^Baseline---Follow-up6 (22.2)00.006Data are provided as ^a^median (IQR) or ^b^n (%)p: follow up vs. baseline
Recovery of menses was diagnosed after a median time of 31 months (IQR 24–38) and 26 months (IQR 14–39) after initial FHA diagnosis in the PCOM (n = 27) and the non PCOM (n = 33) groups, respectively (p = 0.587). At follow-up, the median age was 28 years (IQR 25–31). Compared to women who had no PCOM initially, those who had initially revealed PCOM still had a higher PCOM prevalence (48.1% versus 0%, p < 0.001) and higher AMH levels (median 6.49 ng/mL, IQR 4.74–7.95 versus median 2.25 ng/mL, IQR 2.0–2.71; p < 0.001) as well as a higher PCOS prevalence (22.2% versus 0%, p = 0.006; Table 2).
When focusing on the six women who fulfilled the diagnostic criteria for PCOS after recovery of menses, there were three cases of PCOS phenotype A (PCOM + hyperandrogenism + oligo-/anovulation), two cases of PCOS phenotype C (PCOM + hyperandrogenism) and one case of PCOS phenotype D (PCOM + oligo-/anovulation).
The focus was also on the hormonal dynamics in both groups from baseline to follow-up after recovery of menses (Table 2). In both the FHA-PCOM and the FHA-non PCOM groups, there were significant increases in FSH, LH, estradiol, and prolactin levels (p < 0.001 for all comparisons, except prolactin where p = 0.004). In addition, total testosterone had increased slightly but not significantly, in the same manner in both groups. Concerning the median AMH levels, there were significant differences from baseline to follow-up in both the FHA-PCOM and the FHA-nonPCOM group (p = 0.005 and p = 0.004, respectively). In the FHA-nonPCOM, there was a rise in median levels. The same was found in the FHA-PCOM group. However, the AMH’s IQR decreased in these patients. Notably, the FHA-non PCOM group revealed a slight median non-significant AMH increase of 0.09 ng/mL (IQR (−0.02–0.30), whereas FHA-PCOM women showed a significant median decrease in AMH levels (median AMH dynamics − 0.82 ng/mL, IQR − 2.30 - −0.16; p < 0.001). The individual AMH courses can be seen in Fig. 1. None of the patients revealed exactly the same AMH value at baseline and at follow-up. In the FHA-non PCOM group, there were eight decreases (minimum − 0.37 ng/mL to maximum − 0.06 ng/mL) and 25 increases in AMH (minimum 0.02ng/mL to maximum 1.12 ng/mL), while in the FHA-PCOM group, there were 22 decreases (minimum − 5.19 ng/mL to maximum − 0.14 ng/ml) and five increases in AMH (minimum 0.12 ng/mL to maximum 3.12 ng/mL; p < 0.001 in Fisher’s exact test). Four out of the six patients who fulfilled the PCOS criteria at follow-up, revealed an increase in AMH from baseline to follow-up (median AMH increase 1.82 ng/mL, IQR 0.82–3.06; time interval from initial diagnosis to recovery of menses 14–41 months), whereas a decrease was found for two women (−0.35 ng/mL, time interval from baseline to follow-up 30 months, and − 2.56 ng/mL, time interval from baseline to follow-up 14 months).Fig. 1Individual courses of AMH from baseline to follow-up after recovery of menses in FHA-PCOM and FHA-nonPCOM patients
In this longitudinal study, FHA patients experienced recovery menses after about two to three years after the initial diagnosis. Notably, according to Table 2, it seems that the median BMI increased neither in the FHA-PCOM nor in the FHA-non PCOM group. However, this seems to be due to the lack of a normal distribution. The non-Gaussian distribution did not allow us to use the paired t-test. When having a look at the raw data, one can see that most patients had a higher BMI at the time of follow-up. Moreover, when focusing on BMI dynamics from baseline to follow-up, it becomes evident that there was a median increase of about 1 kg/m^2^ in the FHA-nonPCOM group. This was not the case in the FHA-PCOM group, where the majority of patients had acquired FHA due to stress and excessive exercise rather than eating disorders or weight loss/underweight. This is in line with our previous finding that FHA-PCOM was associated with psychological stress as main cause for FHA [6]. However, this did not reach statistical significance in the present data set (Table 1). Moreover, the previously reported differences in BMI, testosterone and SHBG levels between the FHA-PCOM and the FHA-nonPCOM groups [4, 5, 13] could neither be proven in the present data set, which is likely due to the smaller sample size. In contrast, higher AMH levels were found for women with FHA-PCOM, which had also been reported in previous publications [4, 6].
One of the key findings is the observation that in women with FHA-nonPCOM, the AMH levels increased in the majority of affected women (25/33, 75.8%), despite the fact that this rise in median AMH levels did not reach statistical significance, which is likely due to the sample size and the use of the Mann-Whitney U test. However, the observation that AMH tends to increase after recovery of menses, is in accordance with previous data, which showed a similar trend in AMH for FHA-non PCOM women after three months of pulsatile gonadotropin releasing hormone (GnRH) therapy [7]. As hypothesized previously [4, 5, 7] and according to the two-triangle hypothesis [25], low AMH levels in FHA-non PCOM patients are due to the relative FSH deficiency, which leads to a decrease in the pool of growing follicles and therefore to a decrease in ovarian AMH production. In the present data set, the observed rise in FSH levels after recovery of menses (Table 2) supports the relative FSH deficiency during FHA.
An interesting outcome is that in the FHA-PCOM group, there was a median decline in AMH levels from baseline to follow-up. Notably, such a decrease was found in 22/27 women (81.5%). Notably, PCOM had resolved in about 52% of FHA-PCOM patients after recovery of menses (Table 2). This might be of major relevance. In previous studies, the baseline PCOM prevalence in FHA varied from 41.9% to 46.7% [4, 5, 7, 8] and was 45% (27/60) in the present data set. This seems to exceed the generally agreed on PCOM prevalence of about 30% in the female population of reproductive age, which includes both cases of PCOS and women with “silent PCOM” [3, 11]. Notably, after recovery of menses, six women fulfilled the criteria of PCOS, all of them having belonged to the FHA-PCOM group at baseline (6/27, 22.2%). Notably, four of these six patients revealed an increase in AMH from baseline to follow-up. Taken these observations together, our data seem to support the previous hypothesis that the finding of PCOM in women with FHA is only partially caused by pre-existing PCOS. Robin et al. had reported that about 20% of FHA-PCOM patients might be affected by some PCOS features [15].
When focusing on FHA patients with PCOM at baseline only, it becomes evident that at follow-up after recovery of menses, 6/27 women (22.2%) revealed PCOM in the context of PCOS, silent PCOM was found in 7/27 patients (25.9%), whereas in 14/27 cases (51.9%), PCOM could not be detected on ultrasound any more. This raises the question, whether a relevant proportion of the PCOM is attributable to an intermediate ovarian stimulation state. We find it hard to provide a sound rationale for this phenomenon. Hypothetically, the low-but-normal FSH levels, which are often found in FHA women [26], might be sufficient for follicular recruitment from the follicle pool but not for follicular growth and ovulation. This could lead to an accumulation of smaller follicles. Moreover, a previous study showed that chronic psychological stress was associated with PCOM in FHA patients [6]. Unfortunately, no such data is available for women with silent PCOM in the general population. It has been mentioned that in the development of PCOM, local intraovarian factors might also be relevant contributors to PCOM etiology. Among these, the nerve growth factor (NGF), a sympathetic neurotrophin and marker for sympathetic nerve activity, has been reported to facilitate follicular development. In PCOS rats, ovarian sympathetic activation was accompanied by increased intra-ovarian synthesis of NGF [27]. Thus, there might be a link between the stress system and ovarian activity. Another connection could be cortisol-induced increased ovarian follicle atresia, which may be due to the inhibited expression of growth and differentiation factor 9. As a result, the number of atretic antral follicles might increase [28].
Regardless of these considerations, our data can help to provide adequate information to patients. They show that only a small percentage of FHA-PCOM women will develop PCOS after recovery of hypothalamic function. In relation to the entire cohort (FHA-PCOM + FHA-non PCOM patients, n = 60), the prevalence of PCOS and silent PCOM was 10.0% and 12.7% after recovery of menses, respectively. These rates seem to correspond to what can be expected in a female population of early reproductive age [8, 12, 14].
Notably, only FHA patients who had experienced recovery of menses were included in this analysis. Thus, patients who had recovered from FHA but remained amenorrheic would have been overlooked. Moreover, during the phase of hypothalamic insufficiency, the FHA-patients became older and, thus, one could argue that some PCOS cases might have been missed due to the age-related decline in the FNPO. However, recently it has been demonstrated that age-specific stratifications separating adult studies with mean PCOS age of ≥ 30years old and those with mean age < 30years old did not improve diagnostic measures for FNPO [29]. Notably, in our data set, the median age at follow-up was 28 years (with an upper IQR of 31 years). However, the PCOS prevalence might be have been slightly underestimated in the present data set, which should be considered a minor study limitation.
Moreover, it should be noted that recovery of menses was defined as three spontaneous bleedings within six months. Thus, women with oligomenorrhea were also included in our data set. Given the fact that both PCOS and FHA are exclusion diagnoses, it is hard to say whether patients with oligomenorrhea and PCOM suffered from a remaining FHA component or from PCOS phenotype D [24]. It has already been claimed that PCOS phenotype D would be the one phenotype carrying the highest risk for a misdiagnosis with FHA [10, 20]. However, since there was only one patient with exactly this constellation, this seems of minor relevance.
Last not least, the sample size and the retrospective design must also be mentioned as study limitations.
In conclusion, in the majority of women with FHA-nonPCOM, increases in AMH were found after recover of menses. This finding supports the hypothesis that the relative FSH deficiency is responsible for lower AMH levels in these patients as long as hypothalamic dysfunction is active. It should be pointed out that these women should not be not diagnosed with a low ovarian reserve in FHA patients too quickly on the basis of decreased AMH level alone as mentioned previously [4, 7]. The majority of FHA-PCOM patients did not reveal PCOS after recovery of hypothalamic function (about 78%). In many of these women, AMH declined and PCOM resolved after recovery of menses. One could suggest that in these cases, PCOM might reflect a reversible state of ovarian hyperactivation during FHA.