Functional hypothalamic amenorrhoea and polycystic ovarian morphology: a narrative review about an intriguing association.

Functional hypothalamic amenorrhoea (FHA) is caused by chronic anovulation, which is not due to identifiable organic causes ( Yen et al. , 1973 ; Gordon et al. , 2017 ). A functional reduction in the frequency of the pulsatile secretion of GnRH leads to a decline in the frequency of LH pulses ( Berga et al. , 1989 ) as well as a reduction in serum LH and, sometimes, FSH levels. In detail, slow frequency of GnRH pulses leads to decreased secretion of LH and, to a lesser extent, of FSH, since reduced GnRH pulsatility favours FSH secretion ( Tsutsumi and Webster, 2009 ). Nonetheless, it is unclear whether lower levels of FSH, and/or LH, would imply more severe forms of FHA. However, folliculogenesis and ovulatory function cannot be maintained. Notably, exogenous pulsatile administration of GnRH restores these processes ( Hurley et al. , 1983 ; Miller et al. , 1983 ). Stress and stress sensitivity, vigorous exercise, weight loss, and psychological disorders ( Meczekalski et al. , 2008 ; Gordon et al. , 2017 ; Bonazza et al. , 2023 ) are commonly associated with FHA and are considered as the main factors for the suppression of the hypothalamic-pituitary-ovarian axis. Concerning stress, it has been suggested that kisspeptin neurons would play the bridging role between the hypothalamic-pituitary-adrenal axis, i.e. the stress response system, and the hypothalamic-pituitary-gonadal axis ( Facchinetti et al. , 1993 ; Fioroni et al. , 1994 ). Notably, although data suggest a cause-and-effect relationship between stressful life events and the onset of FHA ( Facchinetti et al. , 1993 ; Fioroni et al. , 1994 ), as well as a causative role for a combination of endocrine and psychogenic dysfunctions and stressful life events ( Fioroni et al. , 1994 ; Gallinelli et al. , 2000 ), there are no clear data about the duration of stress needed to cause amenorrhoea. Of note, secondary amenorrhoea, defined as the cessation of previously regular menstruation for a period of more than three months or previously irregular menstruation longer than 6 months ( Klein et al. , 2019 ), affects about 3–4% of women in the general population ( Bachmann and Kemmann, 1982 ; Munster et al. , 1992 ). FHA is one of the most common underlying conditions together with polycystic ovary syndrome (PCOS), ovarian failure, and hyperprolactinemia. It has been claimed that FHA would be responsible for about 20–35% of cases of secondary amenorrhoea and 3% of primary amenorrhoea cases ( Gordon et al. , 2017 ; Practice Committee of the American Society for Reproductive Medicine, 2004 ). More recent data demonstrate that FHA is the cause for approximately one-third of the cases of secondary amenorrhoea in women of reproductive age ( Saadedine et al. , 2023 ). Polycystic ovarian morphology (PCOM) is a pelvic ultrasonic (U/S) finding and is one of the three items used in the Rotterdam classification for defining PCOS, along with hyperandrogenism and oligo-amenorrhoea ( Rotterdam EA-SPcwg, 2004 ; Dewailly et al. , 2014b ; Teede et al. , 2023 ). At least two out of these three items are required to diagnose a woman as having PCOS, after the exclusion of other diagnoses. Many studies have reported that PCOM may be observed in non-PCOS patients, under various circumstances including women with FHA and even in normo-ovulatory and normo-androgenic women from the general population. However, the prevalence of isolated PCOM varies greatly from one study to another, the main reason being the lack of standardization of methodology and technology used to assess and define PCOM by U/S ( Dewailly et al. , 2014b ). After years of controversies, a consensus now exists to assess and define PCOM by U/S ( Pea et al. , 2024 ). The best criterion is an excessive number (≥20) of 2–9 mm follicles per ovary (follicle number per ovary, FNPO), providing the use of the trans-vaginal route and a transducer frequency >8 MHz. In older studies with frequency 12 ( Dewailly et al. , 2014b ). In case of difficulty in obtaining a reliable estimation of FNPO, an excessive ovarian volume (>10 ml) is considered as an acceptable alternative, but uncertainty remains about the validity of this threshold in adolescents. Notably, it has been suggested that there was an increased FNPO without the PCOS-typical peripheral follicular distribution ( Adams et al. , 1985 ). However, despite this multicystic appearance with less central stroma, PCOM is more typically defined by the FNPO ( Phylactou et al. , 2021 ). One could argue whether the same volume threshold can be used in both pathologies. Because of its close association with follicular excess, excessive serum anti-Müllerian hormone (AMH) has been proposed as a valuable surrogate for elevated FNPO ( Pigny et al. , 2006 ). For a long time, AMH was not included in the PCOM definition due to the lack of assay standardization and the variability of the threshold used in different studies. However, for scientific use, and in studies where in-house normal range has been properly established ( Dewailly et al. , 2011 ), this marker can be used, especially if FNPO is not available. Thus, in the most recent international recommendations for the management of women with PCOS, serum AMH could be used to define PCOM in adults ( Teede et al. , 2023 ). For the aforementioned reasons, the literature is extremely heterogeneous about the prevalence of PCOM in the general population, ranging from 21% to 63% (reviewed in Catteau-Jonard et al. , 2012 ). However, if studies with selection bias and/or inadequate technology and/or inappropriate threshold for FNPO are discarded, there is a general agreement on a prevalence of about 30% (reviewed in Dewailly et al. (2014b )). It is important to take this finding into consideration in order to elucidate whether PCOM in women with FHA is just incidental or not. Indeed, as reviewed below, some women with FHA present PCOM at U/S. Therefore, the combination of amenorrhoea and PCOM can lead to confusion. First, amenorrhoeic patients with PCOM fulfil the revised Rotterdam criteria and, thus, can easily be misdiagnosed with PCOS. Despite the fact that separate diagnostic recommendations and guidelines exist for both diseases ( Gordon et al. , 2017 ; Teede et al. , 2023 ), the diagnoses of both require the exclusion of other causes of menstrual disturbance ( Phylactou et al. , 2021 ). As mentioned previously, the differential diagnosis between FHA and the normo-androgenic PCOS phenotype D in particular might be the main concern ( Beitl et al. , 2022 ). If a woman with FHA is misdiagnosed with PCOS and, thus, incorrectly treated, she will be exposed to the consequences of oestrogen deficiency ( Shufelt et al. , 2017 ). Furthermore, the strategy for ovulation induction will be inappropriate. Moreover, some studies suggest that the two conditions co-exist (reviewed in Phylactou et al. , 2021 ). This is paradoxical from a pathophysiological point of view as ovarian disturbances in FHA are due to slower LH pulsatility while it is the opposite in PCOS. However, studies reporting an association between FHA and PCOS are confusing. This is discussed below. Given these considerations, one could argue that there is uncertainty whether FHA patients with PCOM have been included in many studies about PCOS, especially in the groups of patients with low- to normal weight. This potential for misclassification has already been addressed ( Phylactou et al. , 2021 ). Notably, studies that aim to identify parameters for the differential diagnosis have to rely on an initial categorization. Obviously, that is a contradiction in terms. In addition, the potential for miscategorization challenges this literature field, since some studies might have referred to the FHA-PCOM population as lean women with PCOS. Last but not least, it has been claimed that women with FHA and PCOM differ from women with FHA without PCOM in terms of endocrine regulation and metabolic traits ( Makolle et al. , 2021 ; Hager et al. , 2022 ). These considerations set the stage for a detailed look at the issue of PCOM in women with FHA. The focus is on the prevalence of PCOM in this population, on specific characteristics that go along with PCOM in FHA patients, on therapeutic issues, and many more.

It has been mentioned that women with exercise-induced FHA would reveal changes in cardiovascular function, including endothelial dysfunction, despite their regular training. This might lead to development of atherosclerosis ( O’Donnell et al. , 2011 ). It has been demonstrated that cycle irregularities increase the subsequent risk for coronary heart disease ( Solomon et al. , 2002 ). The negative impact of oestrogen deficiency due to premature ovarian insufficiency on cardiovascular health is well-established ( Shufelt et al. , 2017 ). However, there is a lack of long-term studies in women with FHA. Notably, a recent study used the reactive hyperaemia index (RHI), a well-established technique for non-invasive assessment of peripheral microvascular function and a predictor of all-cause and cardiovascular morbidity and mortality ( Rosenberry and Nelson, 2020 ). About one-third of FHA patients revealed a decreased RHI, consistent with endothelial dysfunction ( Shufelt et al. , 2023 ). Thus, a negative effect of FHA on cardiovascular health can be assumed. Given that women with FHA and PCOM reveal a few metabolic characteristics of PCOS, as summarized above, one could hypothesize that FHA-PCOM patients would carry an increased metabolic risk. However, it is still unclear whether lean PCOS patients are at a clinically relevant increased risk for metabolic complications, although lower antioxidant capacity and a higher risk for insulin resistance were found in comparison to controls, as reviewed by Goyal and Dawood ( Goyal and Dawood, 2017 ). Some data even showed no metabolic differences between lean PCOS patients and matched healthy controls ( Faloia et al. , 2004 ). This is supported by recent Mendelian randomization studies showing that PCOS by itself, without the presence of obesity increased testosterone levels and fasting insulin (among other factors) and would not have a direct causal effect on type 2 diabetes, coronary heart disease, or stroke ( Zhu and Goodarzi, 2022 ). Following our hypothesis that only some FHA-PCOM patients originally had PCOS before acquiring FHA, a potentially PCOS-specific metabolic risk would only affect this minority of patients. However, more clinical data for the comparison of FHA patients with and without PCOM are needed. Most of the data on obstetrical complications in women with FHA are extrapolated from studies examining the association between low BMI and pregnancy complications. Thus, as reviewed by Gordon et al. (2017) , low BMI (a very common situation in FHA women) and anorexia nervosa (an extreme form of FHA) are associated with an increased risk of miscarriage, hyperemesis, prematurity (3–4 times increased risk of preterm birth), and cesarean section for severe foetal growth restriction. Malnutrition is also associated with small for gestational age babies. These data specifically in women with anorexia nervosa were recently confirmed, first in a meta-analysis including 21 studies ( Pan et al. , 2022 ), then in a large prospective study comparing 241 pregnant women diagnosed with anorexia nervosa and 6 418 236 pregnant women without an eating disorder ( Baer et al. , 2023 ). As far as FHA-PCOM is concerned, it might be questioned whether pregnancy is exposed to the same risks associated to PCOS, in addition to the risks of FHA. Indeed, PCOS is known to be a risk factor for a number of obstetric complications (3–4-fold increased risk of pregnancy-induced hypertension and pre-eclampsia, 3-fold increased risk of gestational diabetes, and 2-fold increased risk of preterm delivery), as highlighted in a number of studies reviewed by Palomba et al. (2015) and confirmed by other more recent large studies ( Mills et al. , 2020 ; Farland et al. , 2022 ). However, there are no specific data on pregnancy outcomes in women with FHA-PCOM compared to FHA-non-PCOM, PCOS, and/or a control group of normo-ovulatory and non-hyperandrogenic women. In any case, given the current state of knowledge, it seems logical to consider obstetrical and neonatal risks related to FHA in this particular group. Therefore, before considering any ovulation-inducing treatment, it is essential to ensure not only that the energy deficit has been corrected (which often translates into an increase in BMI), but also that any pre-existing nutritional deficiencies have been corrected and that the patient is psychologically well balanced. Setting the BMI threshold at 18.5 kg/m 2 , as suggested by the experts from the Endocrine Society ( Gordon et al. , 2017 ), remains controversial. Further studies are needed to determine whether or not to include this criterion in the decision to initiate ovulation induction in women with FHA or FHA-PCOM. One of the long-term detrimental effects of FHA is on bone health. FHA is associated with oestrogen deficiency and additional neuroendocrine and metabolic changes that lead to disruption of bone homeostasis. The clinical consequence is a 2-fold increased risk of fractures in women with FHA compared to healthy eumenorrheic women, as recently reviewed by Behary and Comninos (2022) . In patients with FHA-PCOM, Gordon et al. (2017) recommend a baseline bone mass density (BMD) measurement by dual-energy X-ray absorptiometry (DEXA) in those with at least 6 months of amenorrhoea and earlier in those with risk factors for osteoporosis (history or suspicion of severe nutritional deficiency, other energy deficit states, and/or skeletal fragility). These data are based on a retrospective study comparing clinical and hormonal characteristics of 41 women with FHA-PCOM and 41 women with FHA-non-PCOM women of comparable age ( Sum and Warren, 2009 ). This study confirms similar bone mass density-related clinical findings: Of the 15 patients in the FHA-PCOM group whose BMD were evaluated by DEXA, 13.3% (2/15) had osteoporosis, 20% (3/15) had hip osteopenia, and 26.6% (4/15) had spine osteopenia. Of the 27 patients in the FHA-non-PCOM group whose BMD were evaluated by DEXA, 11% (3/27) had osteoporosis, 18.5% (5/27) had hip osteopenia, and 51.8% (14/27) had spine osteopenia. In this small series, although the prevalence of osteoporosis and osteopenia appeared comparable in the two groups, and there was no significant difference in the prevalence of osteoporosis and osteopenia, bone mineral density appeared to be significantly higher in the FHA-PCOM group than in the FHA-non-PCOM group (spine BMD (T-score): −1.1 ± 1.0 vs −1.9 ± 0.5; P  < 0.05, respectively, and hip BMD (T-score): −0.5 ± 0.8 vs. −1.5 ± 0.3; P  = 0.002)) ( Sum and Warren, 2009 ). These data need to be confirmed in prospective studies involving larger numbers of patients FHA-non-PCOM and FHA-PCOM. Gordon et al. (2017) note that bone health in FHA women may not be protected with oestrogen replacement therapy if nutritional and/or energy deficiencies persist. In the same guidelines, experts from the Endocrine Society favour the use of hormone replacement therapy over combined oral contraception to prevent the risk of reduced bone mineral density and osteoporosis ( Gordon et al. , 2017 ). Thus, short-term use of transdermal oestradiol therapy with cyclic oral progestin in adolescents and women who have not had a return of menses after an adequate trial of nutritional, psychological, and/or modified exercise intervention is recommended. The benefits of transdermal oestradiol on bone homeostasis in women with FHA have been highlighted in a recent review of the literature, which also included other more recent studies ( Behary and Comninos, 2022 ). Combined oral contraceptives inhibits IGF-1 (insulin-like growth factor-1) production via first-pass hepatic metabolism. Indeed, the reduction in IGF-1 levels would induce a decrease in osteoblastic activity ( Behary and Comninos, 2022 ). On the other hand, transdermal oestradiol is exempt of first-pass hepatic metabolism, so has no impact on IGF-1 secretion. Thus, combined oral contraceptives should only be used for its contraceptive action (when ovulatory menstrual cycles return) or for other non-contraceptive benefits in specific medical situations (menorrhagia, dysmenorrhea, acne, etc.) ( Gordon et al. , 2017 ). In addition, the use of oral combined contraception masks the return of spontaneous menstruation ( Gordon et al. , 2017 ). LH also plays a critical role in the expression of ovarian hyperandrogenism in PCOS, amplifying the intrinsic dysfunction of the inner theca cells ( Dewailly et al. , 2016 ). In FHA, the very significant decrease in LH leads to a profound reduction in ovarian androgen secretion, which can significantly attenuate the clinical, biological, and ultrasound signs of PCOS. Thus, after recovery of FHA, the critical question is whether restoration of gonadotropic function risks unmasking an underlying PCOS in FHA-PCOM patients. In a subset of 41 women with FHA-PCOM, Sum and Warren (2009) reported in a retrospective study that seven of these women developed signs of hyperandrogenism (acne, hirsutism, and/or alopecia) after significant weight gain. Wang and Lobo (2008) described the uncontrolled follow-up of five women with FHA-PCOM who experienced increases in BMI (5–18%) during 1–3 years. All of them developed oligomenorrhea during follow-up, whereas three of them developed evidence of hyperandrogenism. In the current state of knowledge, there is only one study that has prospectively looked at the outcome of women with FHA-PCOM ( Carmina et al. , 2018 ). In this small prospective observational study of 28 women with FHA, 43% (n = 12/28) had PCOM or ‘PCO-like abnormalities’. Surprisingly, 1 year after recovery from FHA, serum AMH, androgens, and ovarian volume decreased in all women with FHA-PCOM and only one of them developed true PCOS according to the Rotterdam criteria. As previously suggested ( Robin et al. , 2012 ), these data seem to confirm that not all women with FHA-PCOM will develop PCOS in the event of FHA recovery. In fact, the presence of PCOM (and/or PCO-like abnormalities, such as high serum AMH levels) in women with FHA may reflect either: the presence of pre-existing PCOS, which has been masked by the acquired gonadotropic deficiency, likely a minority of patients; or pre-existing asymptomatic PCOM in women with initially normal ovulatory function and no clinical or biological hyperandrogenism, probably the majority. Nevertheless, prospective data from larger series would allow us to better define possible clinical, hormonal, metabolic and/or ultrasound risk factors for the subsequent development of PCOS in the event of recovery from FHA.

An overview of our conclusions is provided in the graphical abstract. PCOM is common in women with FHA. Presumably, PCOM in these patients might be a mixture of pre-existing asymptomatic/‘silent’ PCOM, which likely is the case for the majority of cases of FHA-PCOM, and pre-existing ‘real’ PCOS. FHA-PCOM is not a distinct entity, just a variant of FHA. It results from the well-known causes for FHA, which include excessive exercise, caloric deficiency and/or psychological stress. Moreover, the typical characteristics of FHA can be found, namely amenorrhoea and oestrogen deficiency. It is also likely that these women are at risk for the typical consequences of oestrogen deficiency, which include a loss in bone mass density and changes in cardiovascular function. The management of anovulation in women with FHA-PCOM appears to be also very similar to that recommended for FHA-non-PCOM. Due to the high FNPO and AMH levels, this variant can easily be misdiagnosed as PCOS, especially PCOS phenotype D. Future studies are warranted to clarify the exact pathophysiologic mechanisms underlying FHA-PCOM as well as specific long-term consequences, which might differ from FHA-non-PCOM patients.

The differential diagnosis of FHA and PCOS is of high clinical relevance, given that the two diseases are the most common causes for secondary oligo-/amenorrhoea ( Gordon et al. , 2017 ; Phylactou et al. , 2021 ). The situation is further aggravated by the fact that, for the diagnosis of both exclusion of other causes of menstrual disturbance is warranted ( Gordon et al. , 2017 ; Teede et al. , 2023 ). PCOS is usually defined using the revised Rotterdam criteria, where a minimum of two out of three criteria are needed; (i) PCOM, (ii) oligo-/anovulation, and (iii) hyperandrogenism ( Rotterdam EA-SPcwg, 2004 ). It is conceivable that FHA patients with PCOM fulfil these criteria and can easily be mis-classified as PCOS phenotype D, which consists of PCOM and oligo-/anovulation only ( Beitl et al. , 2022 ). Although physicians are recommended to also take history of weight loss, vigorous exercise or stress as well as a negative progesterone challenge test into account when defining FHA ( Gordon et al. , 2017 ), the differential diagnosis can be challenging empirically. Moreover, women with PCOS also have an increased odds of bulimia ( Cesta et al. , 2016 ), suffer from more stress than healthy controls ( Lansdown and Rees, 2012 ), and tend to exercise heavily in order to achieve weight loss, which makes the differential diagnosis even more difficult. Notably, recent data from our study group demonstrate that in a cohort of women referred for PCOS, the minimum rate of misdiagnosed FHA cases was about 2% ( Holzer et al. , 2024 ). As reviewed recently, patients with PCOS present with oligomenorrhea more often, whereas amenorrhoea is more typical in patients with FHA ( Phylactou et al. , 2021 ). The majority of women with oligomenorrhea have PCOS (80–90%), whereas this is the case in only about 40% of amenorrhoeic patients ( Teede et al. , 2010 ). Concerning the clinical approach, the underlying causes for FHA are also of utmost importance. In detail, about half of FHA patients revealed eating disorders, which included dieting, bulimia, and food/overweight preoccupation, in one study; however, binge-eating was more frequent in PCOS ( Tay et al. , 2019 ). Interestingly, in a large Swedish cohort study about psychiatric comorbidity in patients with PCOS, there was an increased risk for bulimia but a decreased odds for anorexia nervosa ( Cesta et al. , 2016 ). It has been concluded that although both FHA and PCOS patients might reveal bulimic traits, restrictive eating patterns would be more characteristic for FHA ( Phylactou et al. , 2021 ). The second main aetiological factor is psychological stress ( Gordon et al. , 2017 ). It has been demonstrated that women with FHA had suffered from more life events (59.8%) compared to patients with PCOS (26.6%). At the same time, a higher number of undesirable events, uncontrolled events, and events with an objective negative impact were found in patients with FHA ( Fioroni et al. , 1994 ). Moreover, when exposed to stressors, women with FHA seemed to be more susceptible to stress than women with PCOS with an altered autonomic response ( Gallinelli et al. , 2000 ). The last main factor causal for FHA is excessive exercise. It has been recommended that clinicians should ask patients about recent exercise also in association with dietary habits. This is relevant to the assessment of energy availability, which is the energy remaining after subtracting exercise energy expenditure from ingested energy ( Gordon et al. , 2017 ). Several hormonal parameters differ between PCOS and FHA patients. The latter reveal lower serum levels of oestradiol, androgens, LH, and AMH, whereas SHBG tends to be higher in FHA than in PCOS. Although none of these parameters are completely reliable, an overview about previously evaluated thresholds have been summarized in the review by Phylactou et al. (2021) . Notably, one could argue that the labelling of all women with low oestradiol to have FHA might be a little simplistic, especially as there is a wide range for the arbitrary thresholds for low oestradiol in guidelines. For example, 200 pmol/l (54.5 pg/ml) was recently suggested by the Society for Endocrinology ( Jayasena et al. , 2024 ), while the Endocrine Society position statement commented that ‘current direct oestradiol assays, used to measure levels in normal cycling women in the follicular phase, are insensitive below 20 pg/ml, making the diagnosis and treatment of oestradiol deficiency quite difficult and did not propose any threshold for low oestradiol’ ( Rosner et al. , 2013 ). For the discrimination between the non-hyperandrogenemic PCOS phenotype D and FHA-PCOM, the following formula was calculated based on a retrospective dataset: (7.05 * testosterone ng/ml) – (0.005 * SHBG nmol/l) + (0.117 * LH mIU/ml) − 2.463. A result of ≤0 would mean that the patients would belong to the FHA-PCOM group, whereas a result of >0 suggests PCOS phenotype D. Sensitivity and specificity of this formula were 87.9% and 89.7%, respectively ( Beitl et al. , 2022 ). While lean PCOS patients reveal a higher body fat percentage than weight-matched controls ( Kirchengast and Huber, 2001 ), the opposite was observed for FHA women ( Couzinet et al. , 1999 ). FHA patients often reveal low insulin levels and normal insulin sensitivity ( Laughlin et al. , 1998 ; Andrico et al. , 2002 ), which is in contrast to PCOS patients where insulin resistance is common, even in non-obese PCOS patients ( Behboudi-Gandevani et al. , 2016 ). Since insulin inhibits hepatic SHBG synthesis and SHBG is involved in metabolic processes, including the regulation of lipogenic enzymes, effects on the PI3K/AKT pathway, expression of the glucose transporter 3 and 4 among others ( Qu and Donnelly, 2020 ), the fact that lower SHBG levels are found in PCOS patients than in FHA patients can also be seen from a metabolic perspective. So far, there is a lack of direct comparisons between FHA-PCOM and PCOS patients concerning insulin metabolism and metabolic data in general. In the absence of parameters that allow a highly reliable distinction between FHA-PCOM and PCOS, one must draw on personal experience and make individual decisions based on the presence of several factors. The progestin challenge test has been recommended to rule out chronic oestrogen exposure ( Gordon et al. , 2017 ). Although this test is suspected to be negative for women with FHA-PCOM and positive for women with PCOS ( Phylactou et al. , 2021 ), up to nearly 60% of women with FHA have a withdrawal bleed after progesterone withdrawal ( Kletzky et al. , 1975 ; Nakamura et al. , 1996 ). Given the fact that endometrial thickness is a good indicator for the response to progesterone withdrawal ( Nakamura et al. , 1996 ), the additional use of the progesterone challenge test has been questioned ( Phylactou et al. , 2021 ). Notably, ultra-sensitive oestradiol assays have been mentioned as useful tools in hypogonadism and PCOS ( Ketha et al. , 2015 ), but they are not available in many centres. Based on many of these assumptions, Schlaff and Coddington (2020) suggest that a focused history and examination, pelvic ultrasound, and focused laboratory evaluation at the initial visit should be the standard approach. The focus should also be on whether a well-known cause for FHA is present; i.e. excessive exercise, underweight, caloric deficiency and/or stress. Low levels of gonadotropins, especially LH, are often seen as characteristic for FHA. We suggest that in the presence of typical causes of FHA and clear signs of oestrogen deficiency, especially low endometrial thickness, FHA can be safely assumed. We believe that with the use of parameters, the majority of patients can be correctly assigned to either the FHA or the PCOS group. However, more studies are needed to ultimately solve this complex issue. One additional diagnostic tool that might be relevant in the future is the LH to FSH ratio that is often elevated and, moreover, exceeds 2, in PCOS patients ( Yang and Chen, 2024 ), since recent data demonstrated a LH to FSH ratio <1 in about 82% of FHA patients ( Boegl et al. , 2024 ).

To answer the main questions about (i) the prevalence of PCOM in FHA and (ii) regulatory mechanisms in metabolic pathways and the hypothalamic-pituitary-ovarian axis, we included only studies about FHA, which differentiated between PCOM and non-PCOM in well-defined patient populations as follows: (i) patients’ age ≥18 years; (ii) year of publication >1980, when transvaginal ultrasound (TVUS) became available; (iii) original studies; (iv) validated diagnosis of FHA (secondary amenorrhoea for more than 6 months after exclusion of pregnancy in a context of recent weight loss and/or insufficient caloric intake and/or intense physical activity and/or notion of recent psychological stress); negative progestin challenge test, reflecting chronic hypoestrogenism; normal pituitary MRI excluding a mass lesion of the sella and/or suprasellar region; exclusion of hyperprolactinemia, defined as plasma levels >25 ng/ml (i.e. 530 mIU/l) on two occasions; exclusion of premature ovarian insufficiency and history of drug abuse or hormone treatment in the 3 months before the study); and (v) validated definition of PCOM (exclusion of patients in whom vaginal ultrasound was not possible (virgin or refusal)); only trans-vaginal ultrasound (TVUS); assessment of follicles number per ovary, FNPO, all follicles <10 mm in diameter counted by slow scanning through the ovary, with appropriate threshold, i.e. ≥12 or 20 according to the max frequency of the probe 8 MHz, respectively; ovarian volume >10 ml as an acceptable surrogate to excessive FNPO; no follicle > 10 mm). For studies that evaluated the presence of PCOM and included a non-FHA control group, we applied the following criteria for the control population: normo-ovulatory women, where PCOM was not an exclusion criterion and was defined as above. Using these criteria, seven relevant articles were identified ( Dumont et al. , 2016a , b ; Beitl et al. , 2022 ; Hager et al. , 2022 ; Mayrhofer et al. , 2022 ; Hager et al. , 2023 ; Makolle et al. , 2023 ) and one more article was found through the reference lists ( Makolle et al. , 2021 ). Table 1 shows an overview of all selected articles. Overview on characteristics of the main reviewed studies. 5–7 MHz (<2008) 5–13 MHz (≥2008) PCOM was an inclusion criterion Pulsatile GnRH treatment versus gonadotropin stimulation 5–7 MHz (<2008) 5–13 MHz (≥2008) ≥12 (<2008) ≥19 (≥2008) <8 MHz (<2008) ≥8 MHz (≥2008) PCOM was an inclusion criterion for FHA patients PCOS phenotype D patients were included as a second group <8 MHz (<2008) ≥8 MHz (≥2008) ≥12 (5.5 cm 2 5–7 MHz (<2008) 5–13 MHz (≥2008) FHA, functional hypothalamic amenorrhoea; FNPO, follicle number per ovary; HOMA-IR; homeostasis model assessment of insulin resistance; PCOM, polycystic ovarian morphology; PCOS, polycystic ovary syndrome; TVUS, transvaginal ultrasound. When analysing the available studies critically, in four studies ( Makolle et al. , 2021 , 2023 ; Hager et al. , 2022 ; Selzer et al. , 2024 ), a high FNPO was used as the sole criterion to define PCOM, with a threshold of ≥12 or ≥20 according to the frequency probe of either < or ≥8 MHz, respectively ( Table 1 ). Prevalence of PCOM in FHA varied from 41.9% to 46.7%. In the study of Robin et al. (2012) the prevalence of PCOM was slightly greater (48.3%). This can be explained by the fact that, using cluster analysis, no pre-defined threshold was used for markers of PCOM. The strongest variables allowing non-subjective generation of the FHA-PCOM group were first AMH and then ovarian area (FNPO was not used as a continuous variable because of the change of U/S probe during the inclusion period). Therefore, the combination of these two variables might be more efficient to detect PCOM, as recently suggested by Fraissinet et al. (2017) . Interestingly, cluster analysis identified two sub-groups of FHA-PCOM patients, suggesting that this entity might be heterogeneous. This is discussed below. Carmina et al. (2018) used the serum AMH level as a marker of PCOM, with a threshold of ≥4.7 ng/ml that was based on a meta-analysis including studies with second generation AMH assays ( Iliodromiti et al. , 2013 ). Notably, this cut-off was higher than AMH thresholds that were reported more recently with automated assays ( Dietz de Loos et al. , 2021 ). The prevalence of PCOM using AMH ≥4.7 ng/ml in their patients with FHA was similar to those found by others with FNPO ( Table 1 ). In a previous study, Carmina et al. (2016) found that 32.5% of women with FHA had an increased serum AMH level (≥4.7 ng/ml). Intriguingly, in both studies, all control patients had normal AMH levels and ovarian volume. We can therefore suspect a selection bias for controls, since the prevalence of PCOM in the general population is about 30% (14–33% in studies with a sound FNPO definition; see section ‘Tentative estimation of the prevalence of PCOM in FHA’). This precludes any statistical comparison between controls and FHA patients. As shown in Table 1 , our literature search found only two studies ( Robin et al. , 2012 ; Selzer et al. , 2024 ) that reported the PCOM prevalence in both control and FHA groups. Notably, Robin et al. used a different definition for PCOM. In both studies, the prevalence of PCOM was significantly greater in patients with FHA than in the control group. However, the two studies are discordant about the rate of PCOM in controls (26.7% and 11.4%, respectively). In the study of Robin et al. (2012) , the control group was large and included women who were referred for infertility due to male factor and/or tubal abnormality. However, they were neither matched for age nor for BMI. In Selzer et al. (2024) , the control group consisted of healthy, normally ovulating women, who had been recruited for previous studies. However, the control group was smaller than the FHA group but was matched for age. Notably, one might consider that the prevalence of PCOM in controls reported by Selzer at al. (2024) (11.4%) is lower than generally reported as previous studies in Europe had evaluated the PCOM prevalence in general female populations to be 14–33% in studies with a reliable FNPO definition. However, such previous studies included women with PCOS and women with silent PCOM ( Polson et al. , 1988 ; Clayton et al. , 1992 ; Koivunen et al. , 1999 ; Michelmore et al. , 1999 ). In contrast, in the study by Selzer et al. (2024) , no PCOS patients were included in the control group. Only one other previous epidemiologic studies also provided the prevalence of silent PCOM in a population without PCOS ( Michelmore et al. , 1999 ); the rate of 9.1% there was quite similar to the data of Selzer et al. (2024) . One study compared the prevalence of PCOM between 83 controls and 159 patients with so-called ‘hypothalamic hypogonadism’ Alemyar et al. , 2020 ). Although not stated, some of these patients most likely had FHA, at least those with amenorrhoea, the others being oligomenorrheic. Alemyar reported a prevalence of PCOM of 31.4% vs 24.1% in controls, the difference being non-significant. However, the authors found significantly higher AMH levels than in controls. Dumont et al. (2016b ) reported a prevalence of PCOM higher than in other studies (59.7%), most likely because of a selection bias since the FHA patients included in the population were all referred for ovulation induction. The presence of PCOM could have favoured the referral to an academic centre. To conclude, given the very small number of studies that included controls, and the differences in their origins, we cannot provide a precise estimate of silent PCOM in control populations, but estimate a range of prevalence between 10% and about 25%. Nonetheless, this is much lower than in FHA, where the rates of PCOM are quite homogeneous between studies. Of note, all selected studies reporting the prevalence of PCOM in patients with FHA yielded homogeneous results, ranging from 41.9% to 46.7% ( Table 1 ). This is higher than the general population (range 14% to 33%) according to reliable studies with a sound FNPO definition ( Polson et al. , 1988 ; Clayton et al. , 1992 ; Farquhar et al. , 1994 ; Koivunen et al. , 1999 ; Michelmore et al. , 1999 ). However, unfortunately, only two studies included statistical comparisons. Given the homogeneous distribution, it is reasonable to assume that FHA patients have a prevalence of PCOM higher than the general population.

There are very few data in the scientific literature that have addressed the specific health consequences of oestrogen deficiency in women with FHA-PCOM compared to women with FHA-non-PCOM. Only Sum and Warren (2009) had addressed this issue in a small retrospective series with 41 women, of comparable age, in each group. These authors were interested in the long-term effects on the risk of osteopenia and osteoporosis. In 2017, recommendations from the Endocrine Society stated that the presence of PCOM should not alter the management strategy for women with FHA, particularly with regard to preventing the short-, medium-, and long-term consequences of chronic oestrogen deficiency ( Gordon et al. , 2017 ). Oestrogen replacement is beneficial and recommended for long-term health in young women with hypogonadism, regardless of the cause. Thus, an oestrogen replacement therapy should be offered to all women with FHA, with or without PCOM. The advantages and disadvantages of the various oestrogen replacement modalities for FHA are detailed in these same recommendations ( Gordon et al. , 2017 ). If pregnancy is desired in a women with FHA (with or without PCOM), it is strongly recommended to perform an infertility work-up to look for another aetiology of female (tubal obstruction, endometriosis, etc.) and/or male infertility ( Gordon et al. , 2017 ). This step is necessary before starting ovulation induction cycles combined with timed intercourse. Energy deficit is a relevant and frequent cause for the development of FHA. Up to 89% of women with anorexia nervosa and up to 60% of female high-performance athletes are affected ( Sanborn et al. , 1982 ; Warren and Perlroth, 2001 ; Watson and Andersen, 2003 ; Andersen and Ryan, 2009 ; Roupas and Georgopoulos, 2011 ). Recent guidelines from the International Olympic Committee have highlighted the importance of detecting low energy availability in high-level athletes ( Mountjoy et al. , 2024 ). A recent comprehensive review focused on the association between body weight, energy availability and return of menses in women with eating disorders or excessive exercise. It was reported that anthropometric characteristics, which included BMI, body weight or body composition, were associated with resumption of menses ( Pape et al. , 2021 ). Although correction of the energy deficit and weight gain lead to recovery of menses, the individual BMI threshold differed considerably ( Pape et al. , 2021 ). It has been mentioned that for recovery of menses, approximately 2 kg more than the patient’s weight at the time when menses were lost would be needed ( Golden et al. , 1997 ). According to Pape et al. (2021) , recovery of menstrual cycles requires attainment of 18–28% body fat in anorexic patients, although not all women achieve this even when 36% body fat is reached. This variation in required energy balance could reflect a genetic predisposition to FHA, as suggested by Caronia et al. (2011) . Thus, apart from BMI, the overall caloric intake might be of major relevance, as it has been shown to differ between women with persisting FHA who had achieved and maintained a normal weight for at least 1 year and age-, weight-, and body fat-matched normally cycling controls ( Miller et al. , 1998 ). No data exist about differences in the recovery of menses between FHA-PCOM and FHA-non-PCOM patients. However, FHA-PCOM women revealed a higher median BMI in two studies ( Makolle et al. , 2021 , 2023 ). A higher BMI at the onset of amenorrhoea was associated with a higher BMI needed for return of menses ( Pitts et al. , 2014 ; Berner et al. , 2017 ); therefore, one could assume that this might be the case for women with FHA and PCOM. Ovulation induction is justified only in FHA women (with or without PCOM) with persistent amenorrhoea despite (at least partial) correction of energy deficit and moderate weight gain. However, in this situation, the Endocrine Society recommends that women with FHA should have a BMI higher or equal to 18.5 kg/m 2 before ovulation induction is offered ( Gordon et al. , 2017 ). Once spontaneous menstrual cycles of normal duration have resumed, it should be recommended that ovulation induction treatments be postponed and that the couple be allowed to conceive spontaneously if the initial infertility workup is compatible with spontaneous fertilization. Oral ovulation inducers are simple methods of inducing ovulation in infertile women, particularly in cases of anovulation secondary to PCOS. Clomiphene citrate is a selective oestrogen-receptor modulator that can stimulate pituitary gonadotropin secretion by antagonizing hypothalamic-pituitary oestrogen receptors. There are no randomized clinical trials evaluating the use of clomiphene citrate for the treatment of infertility in women with FHA. Many, non-randomized studies do not support its use in this indication as reviewed by Gordon et al. (2017) . Two studies suggested that prolonged uses of clomiphene citrate would be more effective in inducing a return of menses ( Djurovic et al. , 2004 ; Borges et al. , 2007 ). Nevertheless, no studies have clearly studied the impact on pregnancy rates. The Endocrine Society’s 2017 recommendations for managing women with FHA suggest a possible test of treatment with clomiphene citrate only for women with a sufficient endogenous oestrogen level (FHA being recovered) ( Gordon et al. , 2017 ). However, the oestradiol levels indicating recovery of FHA are unclear and the use of the progestin challenge test has been questioned, as outlined above. In any case, the chances of success of clomiphene citrate treatment remain uncertain. Therefore, it cannot be recommended as first line treatment for FHA. Aromatase inhibitors, such as letrozole, induce a significant decrease in circulating oestradiol levels. This eliminates the negative feedback of oestrogens on the gonadotropic axis and stimulates the secretion of pituitary gonadotropins. While it is currently considered the first line treatment for ovulation induction in women suffering from PCOS ( Teede et al. , 2023 ), to our knowledge there are no studies concerning the management of women with FHA, with or without PCOM, with the use of aromatase inhibitors. Gonadotropins combining both FSH and LH activity ( Messinis, 2005 ; Li and Ng, 2012 ) and pulsatile administration of exogenous GnRH ( Gompel and Mauvais-Jarvis, 1988 ; Schivardi et al. , 1989 ) are the two recommended ovulation inducers in women with hypogonadotropic hypogonadism due to hypothalamic anovulation, and also in women with FHA ( Gordon et al. , 2017 ). A meta-analysis by Tranoulis et al. (2018) , which included just over 1000 women with hypothalamic amenorrhoea (functional or not), confirmed the high efficacy of pulsatile GnRH administration in women with FHA: high monofollicular growth and ovulation rates associated with high single pregnancy rates and a low risk of ovarian hyper-response and multiple pregnancy. There are very few studies comparing the efficacy of pulsatile GnRH therapy with that of gonadotropins (FSH + LH activity). The first prospective randomized controlled trial to compare a first cycle of pulsatile GnRH therapy (15 patients) vs recombinant gonadotropins (15 patients) in patients with FHA-PCOM ( Dubourdieu et al. , 2013 ) showed significantly higher ongoing pregnancy rates in the pulsatile GnRH therapy group (46.6% vs 0%, P =  0.02). There are several limitations to this study: the small size of the two groups compared, the fact that only the first cycle was analysed, unclear PCOM definition, as well as the hormonal profile of the patients included in this study. In fact, the mean serum LH level was relatively high in both groups (close to 5 IU/l), suggesting that those patients were already recovering from FHA. A larger retrospective study comparing 55 FHA-PCOM patients, treated either with pulsatile GnRH therapy (38 patients, 93 cycles) or with gonadotropins (17 patients, 53 cycles) confirmed that ovulation rates were significantly lower with gonadotropins (56.6% vs 78.6%, P  = 0.005) ( Dumont et al. , 2016a ). Moreover, ongoing pregnancy rates were significantly higher with GnRH therapy, either per initiated cycle (26.9% versus 7.6%, P  = 0.005) or per patient (65.8% versus 23.5%, P  = 0.007). Although retrospective and non-randomized, this study looked at cumulative pregnancy rates pregnancy after multiple subsequent cycles of treatment, and the diagnostic criteria for FHA-PCOM were consistent, with mean serum LH levels below 2 IU/l in both groups. The results of these two studies suggest that pulsatile GnRH therapy could be considered a first-line treatment for this population, as it is for FHA women without PCOM ( Gordon et al. , 2017 ). However, these results should be confirmed by larger prospective studies. If pulsatile GnRH therapy appears to be more effective than gonadotropins in inducing ovulation in women with FHA-PCOM, it is interesting to check whether the presence of PCOM influences the response to this treatment. One study suggested that pulsatile GnRH administration can ‘wake up’ underlying PCOS. Mattle et al. (2008) reported that in a cohort of 120 patients with FHA (no evidence of prior PCOS) receiving pulsatile GnRH therapy for more than 100 days, 6 patients developed a hormonal profile suggestive of PCOS. The occurrence of possible PCOS during pulsatile GnRH therapy in women with FHA with signs suggestive of PCOM has also been reported in two small retrospective studies ( Adams et al. , 1985 ; Schachter et al. , 1996 ). Dumont et al. (2016b ) published a retrospective cohort study including 27 patients with FHA-non-PCOM and 40 patients with FHA-PCOM treated by pulsatile GnRH therapy for ovulation induction (85 and 104 initiated cycles, respectively). With equivalent doses of GnRH per pulse, the authors reported similar ovulation rates (80.8% vs 77.7%; P  > 0.05). There was no significant difference in ongoing pregnancy rates (26.9% vs 20% per initiated cycle; P  > 0.05 and 70% vs 63% per patient; P  > 0.05, respectively). The aim of this retrospective study was to evaluate the efficacy of pulsatile GnRH administration as a treatment for infertility secondary to anovulation (in terms of ovulation and ongoing pregnancy rates). Thus, this study cannot attest to the revelation of true PCOS in the FHA-PCOM group. To do so, it would be interesting to perform a prospective study and systematically monitor for the revelation of features of PCOS during treatment with pulsatile GnRH therapy (persistent anovulation, clinical signs of hyperandrogenism, serum androgen, and AMH levels, number of antral follicles on ultrasound, etc.). Regarding ovulation induction by pulsatile GnRH therapy in FHA, Dumont et al. (2016a ) showed that having an associated PCOM does not increase the risk of ovarian hyper-response at an equivalent dose of GnRH per pulse (12.5% vs 10.6%; P  > 0.05). In the retrospective study of Dumont et al. (2016a ) highlighted significantly higher cancellation rates due to excessive ovarian response in women receiving gonadotropins (human menopausal gonadotrophin (HMG) or recombinant FSH + recombinant LH) versus pulsatile GnRH therapy (respectively 34% vs 14% per initiated cycle, P  < 0.005). This difference was not found in the small prospective randomized controlled trial by Dubourdieu et al. (2013) , which analysed only the first cycle. In the case of ovulation induction with gonadotropins, a prospective observational study of 20 FHA patients (half of whom had PCOM) compared to 59 PCOS patients treated with 75 IU of HMG per day showed a similar ovarian response between women with FHA-PCOM and those with PCOS, particularly in terms of the number of follicles selected, oestradiol levels, and cancellation rates for excessive ovarian response ( Shoham et al. , 1992 ). Wang and Lobo administered 150 IU recombinant FSH alone to women with PCOS (n = 10), FHA-PCOM (n = 6), and normo-ovulating women (controls; n = 20), finding an increase in serum androgen levels per follicle, which was significantly greater in women with PCOS and FHA-PCOM than in the control group. On the other hand, there was no difference in the number of follicles selected and variations in oestradiol levels between the three groups. A limitation of this study is that only FSH (and not LH) was used for this ovarian stimulation test in the three groups. It is possible to speculate that the results of stimulation in the FHA-PCOM group were probably suboptimal due to the lack of correction for LH deficiency ( Wang and Lobo, 2008 ). Other therapeutic approaches may one day be proposed, such as the administration of kisspeptin ( Jayasena et al. , 2009 , 2010 ) or L-acetyl-carnithine ( Genazzani et al. , 2011 , 2017 ) in the treatment of gonadotropic deficiency in FHA women. However, these data are still experimental, and the potential of these treatments needs to be further explored in other studies. In addition, the value of these therapies in women with FHA-PCOM will need to be specifically evaluated.

A few studies differentiated between FHA-PCOM and FHA-non-PCOM and focused on differences in hormonal profiles and metabolic parameters ( Makolle et al. , 2021 , 2023 ; Hager et al. , 2022 ; Mayrhofer et al. , 2022 ). An overview is provided in Table 2 . One finding, which could be reproduced in all four studies, was that women with FHA with PCOM revealed higher AMH levels than those without PCOM. In detail, median AMH levels were 33.2 pmol/l (interquartile range, IQR, 11.3–59.0) versus 16.6 pmol/l (IQR 8.0–31.0; P <  0.001) ( Makolle et al. , 2021 ), 46.4 pmol/l (IQR, 16.2–166.4) versus 21.1 pmol/l (IQR 7.5–55.7; P =  0.001) ( Makolle et al. , 2023 ), 6.38 ng/ml (IQR, 4.34–10.10) versus 2.03 ng/ml (IQR, 1.40–2.50; P <  0.001) ( Hager et al. , 2022 ), and 4.77 ng/ml (IQR, 3.61–6.12) vs 2.00 ng/ml (IQR, 1.58–2.55, P  < 0.001) ( Mayrhofer et al. , 2022 ). This is not surprising, given the fact that the FNPO and AMH are positively correlated (reviewed in Dewailly et al. (2014a )). However, one could suggest that due to the deficit in gonadotropin-controlled ovarian stimulation, women with FHA might reveal a lower FNPO and AMH level than PCOS patients. Only one study compared women with FHA-PCOM to patients with PCOS phenotype D; although this study revealed higher AMH levels in PCOS-D patients (8.4 ± 5.1 vs 6.9 ± 3.8 ng/ml), this difference failed to reach statistical significance ( P =  0.071) ( Beitl et al. , 2022 ). No comparative study about correlation between antral follicle count (AFC) and AMH in FHA-PCOM vs PCOS is available in the literature. Differences in patient characteristics and serum parameters between FHA patients with and without PCOM. Makolle et al. , 2021 Makolle et al. , 2023 Makolle et al. , 2021 Makolle et al. , 2023 Hager et al. , 2022 Makolle et al. , 2021 Makolle et al. , 2023 Mayrhofer et al. , 2022 Hager et al. , 2022 Makolle et al. , 2021 Makolle et al. , 2023 Mayrhofer et al. , 2022 Hager et al. , 2022 Makolle et al. , 2021 Mayrhofer et al. , 2022 Compared to FHA patients without PCOM. Compared to FHA patients with PCOM. Compared to the levels before treatment. AMH, anti-Müllerian hormone; FHA, functional hypothalamic amenorrhoea; HOMA-IR, homeostasis model assessment of insulin resistance; PCOM, polycystic ovarian morphology; SHBG sex hormone binding globulin. There were several findings that might support the suggestion that at least some women with FHA with PCOM would have underlying PCOS. In detail, PCOM was associated with higher testosterone levels and higher free androgen indices in one study ( Mayrhofer et al. , 2022 ). Moreover, FHA-PCOM patients also revealed a higher median BMI ( Makolle et al. , 2021 , 2023 ), lower sex hormone binding globulin (SHBG) levels ( Makolle et al. , 2021 ) and a higher median homeostasis model assessment of insulin resistance (HOMA-IR) than those without FHA-non-PCOM patients, which seems similar to metabolic characteristics of PCOS patients ( Behboudi-Gandevani et al. , 2016 ). This is in line with the positive correlation between BMI and HOMA-IR as well as with the negative correlation between SHBG and HOMA-IR ( Mayrhofer et al. , 2022 ). In addition, when the results of GnRH stimulation tests were analysed, the median LH increase was much higher in FHA-PCOM than in FHA-non-PCOM patients (604.9% versus 240.0%, respectively, P  < 0.001) ( Hager et al. , 2022 ). Unfortunately, there are no data about LH pulse profiles in women with FHA according to the presence or absence of PCOM. Some analyses were specific for the FHA-non-PCOM group: a positive correlation between FSH and AMH ( Hager et al. , 2022 ) as well as a negative correlation between AMH levels and age ( Makolle et al. , 2021 ). Lastly, when hormonal changes were analysed after three months of pulsatile GnRH treatment, FSH and AMH increased only in FHA patients without PCOM, whereas LH and oestradiol had increased in both groups ( Hager et al. , 2022 ). Given the above-mentioned studies, it seems reasonable to suggest that at least some of the FHA patients with PCOM originally had PCOS before developing FHA. It seems important that some findings were typical for PCOS regarding both the hypothalamic–pituitary–ovarian axis and the metabolic situation. Although women with FHA-PCOM do not reveal baseline gonadotropin levels like patients with PCOS, the GnRH stimulation test seems to unmask a more PCOS-typical pattern, as mentioned above. Concerning the metabolic aspects, however, it needs to be stated that despite the higher HOMA-IR levels in FHA-PCOM women, only 10% exceeded the threshold level for insulin resistance of 2.5 with a median level of 1.55 ( Mayrhofer et al. , 2022 ). Compared to data about non-obese PCOS patients, this is low, with a meta-analysis demonstrating a pooled HOMA-IR of 2.68 in these women ( Behboudi-Gandevani et al. , 2016 ). Moreover, although there were strong LH increases in FHA-PCOM patients in the course of a GnRH stimulation test (stimulated LH levels 17.2 ± 13.4 IU/l) ( Hager et al. , 2022 ), LH responses in FHA-PCOM women were not as high as those seen in PCOS women (stimulated LH levels in PCOS: 35.48 ± 31.4 IU/l) ( Lewandowski et al. , 2011 ). Last, one could argue that PCOM might ‘disappear’, though this has not been demonstrated so far, when the gonadotropin levels become low because of FHA. However, if we refer to women with PCOS who are taking oral contraceptive pills, completely suppressing FSH and LH levels, the follicle count decreases by 30–40% but remains in the range of PCOM ( Bentzen et al. , 2012 ; Park and Chun, 2020 ). This is in accordance with recent findings of a 17–23% decline in AMH levels during the use of combined oral contraception ( Hariton et al. , 2021 ; Nelson et al. , 2023 ). Therefore, one could postulate that in a woman with PCOS, the number of follicles would not decrease significantly when she develops FHA. Since there are no longitudinal studies, this hypothesis is still to be tested. Lastly, what supports the assumption of FHAPCOM being a sequel of PCOS, at least in some patients, is the observation of Mattle et al. (2008) that in a subgroup of women with FHA, pulsatile GnRH therapy had unmasked some cases of PCOS. This issue has been raised in the review by Phylactou et al. (2021) . As mentioned above, this association is paradoxical. We therefore revisit the articles that tried to document this association. In the study of Alemyar et al. (2020) , the authors reported that 36% of their patients met the Rotterdam criteria. However, they clearly stated that those patients had hypothalamic hypogonadism (HH) instead of PCOS, as their LH and FSH levels were both <2 IU/l. Since, this might not exclude PCOS completely, all HH patients had to reveal oestradiol levels <100µmol/l. However, the authors emphasized the risk of false positive if the Rotterdam criteria were used without any previous differential diagnosis. Bradbury et al. (2017) made a clear distinction between FHA and PCOS in a group of patients with an elevated AMH and a BMI <20 kg/m 2 . In this study, all FHA patients had to reveal a typical cause (eating disorder, pre-occupation with weight, excessive exercise) and an endometrial thickness <4 mm. Sum and Warren (2009) reported that some of their patients with FHA had an ‘underlying’ PCOS. However, the FHA definition was loose (‘low to- normal FSH and LH levels’). Given the quite high LH levels in the FHA patients with an ‘underlying’ PCOS and the presence of hyperandrogenism, it might be suspected that they had, in fact, an amenorrhoeic PCOS rather than an association of FHA and PCOS. Likewise, Koltun et al. , reported that 17% of their exercising women had a biological hyperandrogenism. They concluded that oligo/amenorrhoea in these women should not automatically be presumed to be associated with hypothalamic inhibition, in as much as the average exercise volume was quite moderate (6 h/week) in their group. Presumably, they had rather a mild PCOS. Unfortunately, there was no ultrasound nor AMH data in their series ( Koltun et al. , 2020 ). To summarize, initially, PCOS and FHA might co-exist in the same woman; however, when they do so, the FHA phenotype predominates as a result of hypothalamic inhibition and features of PCOS are blunted, except PCOM. This could be seen as a phenotypic conversion rather than ‘co-existence’. It can thus be concluded that FHA cannot co-exist with a symptomatic PCOS. The confusion about this issue comes mainly from inappropriate definition of FHA and/or PCOS. This major diagnostic issue is discussed in the section ‘Diagnostic issues: how to differentiate PCOM-FHA from PCOS’. One could also hypothesize that PCOM in FHA patients is merely a normal variant since it can also be found in the general population with a prevalence of 14–33%, as mentioned above. However, PCOM seems to be overrepresented with a prevalence of up to 48% in FHA. Notably, in cluster analysis, it has been stated previously that FHA-PCOM women might in fact constitute a heterogeneous subgroup consisting of a majority of cases with incidental PCOM (38%; 22/58) and a minority of women with some features of PCOS (10%; 6/58) ( Robin et al. , 2012 ). As reviewed above, the PCOM rate of up to 48% in women with FHA is considerably high. In addition to the above-mentioned idea of a mixture of silent PCOM and hidden PCOS, it has been suggested that PCOS patients would be somehow prone to develop FHA or that the development of PCOM would be more likely in FHA patients. One hypothesis, which was evaluated recently ( Hager et al. , 2023 ) is based on the fact that women with PCOS reveal a higher generalized sympathetic nerve activity, which suggests that they suffer from increased stress and/or less adequate response to stressors ( Lansdown and Rees, 2012 ). It is well known that stress and sensitivity to stress are relevant causes for the development of FHA ( Gordon et al. , 2017 ). In the mentioned study, 38 stress-induced women with FHA were compared to 38 cases of FHA due to excessive exercise. The stress group showed a higher prevalence of PCOM (57.9% versus 31.6%, P =  0.019). Although there is no direct pathway for a stress-induced development of PCOM, it was concluded based on the mentioned data that due to their well-known stress sensitivity, women with PCOS might be prone to develop FHA. However, the results could only provide a background for this hypothesis, especially due to the lack of longitudinal data ( Hager et al. , 2023 ). In addition, one could also assume that women with PCOS might often reduce their weight with the aim of improving their metabolic health and/or their ovulatory fertility. Substantial weight loss could put these patients at an increased risk for developing FHA. Lastly, PCOM in women with FHA might reflect yet unknown alterations in the intra-ovarian regulation of folliculogenesis, independent from gonadotropins whose serum levels are specifically low in FHA. No studies concerning these two hypotheses have been published so far. As already stated, longitudinal observations would be most helpful to clarify these issues.