Reduced FSH and LH action: implications for medically assisted reproduction.

GnRH agonists and antagonists cause transient LH and FSH deficiency and are used during OS to prevent premature ovulation, thus enabling retrieval of more oocytes ( Zegers-Hochschild et al. , 2017 ). GnRH agonists and antagonists have different modes of action. GnRH agonists, after an initial increase in LH and FSH secretion (flare up), induce downregulation of the GnRH receptor ( Lambalk et al. , 2017 ). Conversely, GnRH antagonists competitively block the GnRH receptor thereby preserving pituitary gland responsiveness, so that gonadotropin levels are restored within a few hours of discontinuing suppression ( Duijkers et al. , 1998 ). Usually, the residual circulating levels of LH are sufficient to support steroidogenesis in theca cells, and recombinant human FSH (r-hFSH) is sufficient for OS. However, LH levels much lower than baseline can negatively affect MAR outcomes in some women ( Westergaard et al. , 2000 ; Huirne et al. , 2005 ; Lahoud et al. , 2006 ; Kol, 2014 ; Chen et al. , 2016 ; Benmachiche et al. , 2019 ). In a study of 200 normo-gonadotropic women treated with a standard long GnRH agonist protocol and OS with r-hFSH, Westergaard et al. (2000) found that, in 48% of women, LH levels remained profoundly suppressed (< 0.5 IU/l) on stimulation Day 8, and that the early pregnancy loss was significantly higher in this group than in women whose LH levels were not profoundly suppressed (45% versus 9%, P  < 0.005). Similarly, Lahoud et al. (2006) found that a 50% decrease in the LH level from the early to mid-follicular phase after GnRH agonist downregulation was associated with a lower live birth rate than in women with a less sharp decrease in LH levels. It can be speculated that an abrupt LH drop during OS results in a decrease in the conversion of androgen precursors to oestrogens thereby resulting in insufficient E 2 production by the growing follicles and a drop in circulating E 2 levels ( Kol and Homburg, 2008 ). However, other studies did not find that low serum LH levels after pituitary downregulation with GnRH agonists negatively affected reproductive outcomes ( Balasch et al. , 2001 ; Humaidan et al. , 2002 ). Thus, it is not clear to which extent serum LH levels are predictive of reproductive outcome in GnRH agonist cycles. However, it should be noted that different LH threshold values were used in different studies to define low LH groups (i.e. <1.5–0.5 IU/l LH) ( Westergaard et al. , 2000 ; Balasch et al. , 2001 ; Humaidan et al. , 2002 ). Profound LH suppression has also been reported in some women after the administration of GnRH antagonists. Huirne et al. (2005) found that if the change in serum LH during GnRH antagonist administration was −2.2 IU/l below baseline values, the chance of achieving a clinical pregnancy under standard OS was almost nil. Kol (2014) reported that, after treatment with a GnRH antagonist, LH was over suppressed in 26% of women and these patients had a significantly lower increase in E 2 during the first 24 h after antagonist administration compared to women who were not over suppressed. The effect was reversed in these patients after the addition of recombinant human LH (r-hLH). It remains unclear if this strong suppression of LH levels in response to GnRH antagonists is observed only in some subgroups of patients. It is conceivable that it may be associated with the individual pharmacodynamic response to the GnRH antagonist. However, it has been suggested that the severity of the LH deficiency caused by GnRH antagonist treatment is not linked to absolute LH serum levels but rather to the magnitude of suppression over time versus the baseline ( Kol and Homburg, 2008 ). Several studies have shown that r-hFSH and r-hLH co-treatment can improve OS in women with transient LH and FSH deficiency after GnRH agonist or antagonist treatment ( Garcia-Velasco et al. , 2007 ; Pezzuto et al. , 2010 ), whereas two systematic reviews did not observe such an effect in the general population ( Griesinger and Shapiro, 2011 ; Alviggi et al. , 2018a ). To evaluate the benefits of r-hFSH:r-hLH treatment, studies should include subgroups of patients who are more prone to develop LH and FSH deficiency after GnRH analogues, such as women of AMA and hypo-responders ( Alviggi et al. , 2018a , 2018b ). The impaired functioning of the LH and FSH systems in AMA women, reviewed above, could be exacerbated by the transient gonadotropin deficiency induced by GnRH analogue regimens during OS. Co-treatment with r-hFSH:r-hLH has been proposed in AMA women undergoing MAR ( Humaidan et al. , 2004 ; Alviggi et al. , 2006 ). A meta-analysis of studies in women between the ages of 35 and 40 years found that r-hFSH:r-hLH co-treatment resulted in higher clinical pregnancy rates than r-hFSH monotherapy ( Hill et al. , 2012 ). Younis et al. (2017) demonstrated that r-hLH addition could compensate for the LH deficiency observed in women above 35 years, and three randomised controlled trials, which were included in the Hill meta-analysis ( Hill et al. , 2012 ), found that co-treatment with r-hLH and r-hFSH in women between the ages of 35 and 39 years improved both live birth and implantation rates ( Matorras et al. , 2009 ; Bosch et al. , 2011 ). This effect was not found in studies of women over the age of 40 years ( Barrenetxea et al. , 2008 ; König et al. , 2013 ; Vuong et al. , 2015 ). The latter finding is not surprising as there is a dramatic reduction in the proportion of normal euploid embryos in women over the age of 40 years ( Ata et al. , 2012 ), and currently there are no data to suggest that this effect can be compensated by r-hLH. The benefit of r-hFSH:r-hLH treatment for AMA women may also be related to the effect of LH on oocyte maturation and embryo implantation. In fact, LH exerts an anti-apoptotic effect on cumulus cells and promotes the paracrine signalling involved in cell expansion and oocyte maturation during folliculogenesis ( Ruvolo et al. , 2007 ; Huang et al. , 2015 ). In addition, LH modulates various signalling molecules involved in endometrial implantation namely, leukaemia inhibiting factor, colony-stimulating factor-1, interleukin-1, integrins, glycodelin and mucin 1, and may improve endometrial receptivity ( Herrler et al. , 2003 ; Tesarik et al. , 2003 ). LH and FSH action may be reduced by receptor and post-receptor defects that cause a reduced response to OS. In clinical trials, about 10–15% of normo-ovulatory and normo-gonadotropic women undergoing MAR had a hypo-response to OS with exogenous gonadotropins ( Conforti et al. , 2019a ). Ovarian hypo-response can be defined as an unexpected slow response or stagnated follicle growth during OS with standard-dose FSH monotherapy, or that may require a higher than expected dose of gonadotropins depending on age, body mass index and ovarian reserve ( Alviggi et al. , 2018b ). A meta-analysis on hypo-response found that OS with r-hFSH:r-hLH resulted in significantly higher clinical pregnancy and implantation rates versus r-hFSH alone ( Conforti et al. , 2019b ). Recently, two indices based on individual ovarian reserve have been proposed to measure the ovarian response to OS: the follicle output rate (FORT) ( Genro et al. , 2011 ) and the follicle-to-oocyte index (FOI) ( Alviggi et al. , 2018b ). FORT measures the ratio of the number of pre-ovulatory follicles on the day of hCG trigger × 100 over the number of small antral follicles at baseline prior to OS, while FOI is the ratio between the number of oocytes collected at ovum pick up and the number of antral follicles at the beginning of OS. These two indices are a direct measure of the individual response to OS and can help to identify a hypo-response to OS better than oocyte number alone. The pathophysiological mechanisms of hypo-response are not yet fully understood. In some cases, a hypo-response may be associated with LH and FSH single nucleotide polymorphisms (SNPs) and their receptors ( Perez Mayorga et al. , 2000 ; Alviggi et al. , 2013 ; La Marca et al. , 2013 ), but the exact prevalence of these SNPs in MAR is unknown because they are not routinely tested. Some SNPs related to LH, FSH, and their receptors that reduce the activity of endogenous gonadotropins or induce ovarian resistance to OS due to receptor and post-receptor defects have been associated with reduced fertility ( Simoni et al. , 2002 ; Desai et al. , 2013 ; Conforti et al. , 2019d ). A polymorphism of the FSH molecule that affects the FSH beta chain promoter (c.-221 G > T) seems to affect gonadotropin function ( Grigorova et al. , 2008 ; Schüring et al. , 2013 ). Indeed, in a study of Finnish women of reproductive age, the T allele carriers were characterised by increased baseline serum LH and FSH levels, and by idiopathic infertility ( Rull et al. , 2018 ). In another study of 63,350 women, the T allele was associated with a longer menstrual cycle, later age at menopause and a reduced incidence of endometriosis ( Ruth et al. , 2016 ). Also FSHR SNPs may be associated with a hypo-response to OS ( Conforti et al. , 2019d ). Some variants of FSHR 680 c.2039G>A (rs6166) are related to higher FSH levels and to higher FSH consumption ( Overbeek et al. , 2009 ; Laisk-Podar et al. , 2015 ; Alviggi et al. , 2018c ). FSHR -29 G > A (rs1394205) allele A carriers were found to have fewer FSHRs than allele G carriers, which translated in a higher FSH consumption during OS and a lower ovarian sensitivity in A carriers ( Casarini et al. , 2014 ; Riccetti et al. , 2017 ). Also the FSHR c. G919A (rs6165) variant may affect ovarian response but not ovarian reserve ( Achrekar et al. , 2009 ; Trevisan et al. , 2014 ; Alviggi et al. , 2018c ). Interestingly, different combinations of genetic variants of the FSH beta chain and of FSHR affect menstrual Day 3 serum FSH levels in women of reproductive age ( La Marca et al. , 2013 ). Variants of the LH beta chain and of LHCGRs that affect OS have also been identified. A common variant of the LH beta chain (rs1800447) affects the LH system during OS ( Haavisto et al. , 1995 ; Alviggi et al. , 2009a , 2013 ). Clinically, carriers of this LH variant are characterised by a less active form of LH that does not adequately support FSH activity during follicle stimulation, and consequently results in a reduced response to OS with FSH ( Alviggi et al. , 2009a , 2013 ). The LHCGR c.872A>G (rs12470652) polymorphism could affect ovarian response when associated with other polymorphisms. Indeed, the ratio between the cumulative FSH dose and the total number of oocytes retrieved was higher in carriers of allele C of FSHR -29 rs1394205, FSHR rs6166, and LHCGR 291 rs12470652 than in women who did not have allele C in at least one of these SNPs ( Alviggi et al. , 2016b ). In addition, LHCGR S312N (rs2293275) carriers required a higher FSH dosage during OS versus asparagine carriers ( Lindgren et al. , 2016 ). Thus, polymorphisms of the LH beta chain and of LHCGRs could reduce ovarian sensitivity to FSH, which suggests that the function of both gonadotropins is crucial during OS. In addition, the presence of serine instead of asparagine in Position 312 might lead to reduced sensitivity to LH action and may explain why carriers of this polymorphism required more r-hLH during OS than non-carriers ( Ramaraju et al. , 2018 ). Moreover, in the same study, the pregnancy rate was significantly higher in serine carriers who received r-hFSH:r-hLH than in those receiving r-hFSH alone ( P  = 0.04) ( Ramaraju et al. , 2018 ).

Folliculogenesis is controlled by a complex interaction among hormones in the hypothalamus, anterior pituitary gland and the ovaries. Luteinizing hormone (LH) and follicle-stimulating hormone (FSH) are secreted from the anterior pituitary gland in response to GnRH ( Conn and Crowley, 1994 ; Kaiser et al. , 1997 ) and play a complementary role in follicle development and ovulation ( Casarini et al. , 2018a ). In ovarian theca cells, LH stimulates the secretion of androgens that are transferred to granulosa cells to be converted to oestradiol (E 2 ) by aromatase. In granulosa cells, FSH stimulates the development of ovarian follicles, while LH action is involved in follicle development and maturation. A deficiency in LH and FSH production or action compromises gametogenesis and gonadal steroid production thereby reducing female fertility ( Santoro et al. , 1986 ; Huhtaniemi and Themmen, 2005 ; Wide et al. , 2009 ; Casarini et al. , 2018a ). Recently, the International Committee for Monitoring Assisted Reproductive Technologies (ICMART) highlighted the importance of gonadotropin action in determining a deficiency of LH and FSH, and provided a broader definition of hypogonadotropic hypogonadism that encompasses this notion, namely ‘Gonadal failure associated with reduced gametogenesis and reduced gonadal steroid production due to reduced gonadotropin production or action’ ( Zegers-Hochschild et al. , 2017 ). While the causes of reduced LH and FSH production are well described ( Bianco and Kaiser, 2009 ), deficiency of LH and FSH action has received much less attention, even though it is now recognised that both production and action are relevant for human fertility and are of clinical interest for medically assisted reproduction (MAR). The action of LH and FSH is determined by a variety of factors, i.e. the frequency and amplitude of GnRH peaks, the different isoforms of LH and FSH, polymorphisms of FSH and LH and their receptors, and intracellular signalling. Furthermore, in MAR, inter-individual demographic, clinical and treatment factors, such as ageing, comorbidities, and the effect of oral contraceptives and GnRH analogue protocols, can influence gonadotropin action and the response to exogenous gonadotropins ( Anobile et al. , 1998 ; Davison et al. , 2005 ; Alviggi et al. , 2009b ; Santi et al. , 2017 ). In this review, we focus on the determinants of reduced LH and FSH action that are associated with a reduced quantitative and qualitative response to ovarian stimulation (OS) in the attempt to gain insights that may help to improve the clinical management of women with LH and FSH deficiency undergoing MAR.

Deficiency of LH and FSH is caused by reduced gonadotropin production or action due to internal and external factors ( Fig. 3 ). These conditions are generally characterised by low E 2 and low-normal gonadotropins ( Yasmin et al. , 2013 ; Gordon et al. , 2017 ). Congenital abnormalities are well described rare conditions that usually present with deficient GnRH secretion occurring in isolation or in association with anosmia (Kallmann syndrome) ( Kallmann et al. , 1944 ; Boehm et al. , 2015 ). Among acquired conditions, intensive exercise and eating disorders are widely recognised as life-style factors that could suppress the HPG axis ( Boyar et al. , 1974 ; Shangold, 1985 ). Furthermore, poorly controlled diabetes and thyroid disorders could significantly affect gonadotropin secretion and action. To restore reproductive function, stress habits should be corrected and the underlying endocrine disorder be treated ( Strotmeyer et al. , 2003 ; Gordon, 2010 ). Other acquired conditions of reduced LH and FSH action could be linked to pituitary tumours (e.g. prolactinomas) or pituitary infarct (e.g. Sheehan’s syndrome) that are usually characterised by specific symptoms due to pituitary dysfunction or compression of tissues surrounding the pituitary or that can severely affect pituitary function beyond reproductive function (e.g. panhypopituitarism) ( Li and Ng, 2012 ). All these conditions are in line with the ICMART definition of hypogonadotropic hypogonadism being associated with reduced gametogenesis and steroidogenesis due to reduced gonadotropin production or action. In MAR, a combination of factors such as AMA and genetic variants of gonadotropins, or their receptors that impair gonadotropin action, may further exacerbate the transient reduced LH and FSH production caused by GnRH analogues, and result in a low or sub-optimal response to OS ( Alviggi et al. , 2016a ; Conforti et al. , 2019a , 2019c ). LH and FSH deficiency is a spectrum of conditions with various aetiologies. Women with LH and FSH deficiency have low/normal gonadotropin levels and low oestradiol levels.

Reduced gonadotropin production or action may cause clinically significant LH and FSH deficiency associated with reduced gametogenesis and steroidogenesis. This may explain why some women treated with MAR have an ‘unexpected’ hypo-response to standard OS with r-hFSH alone, notwithstanding their normal gonadotropin levels and normal ovarian reserve, and may contribute to a reduced ovarian response in women of AMA. It may also pave the way to precision medicine solutions for fertility patients ( Mol et al. , 2018 ), and thus improve reproductive outcomes, particularly in AMA and hypo-responder patients who may benefit from OS with r-hFSH and r-hLH.

LH and FSH deficiency is characterised by reduced LH and FSH production or action in the presence of normal-low LH and FSH and low E 2 levels. In specific subgroups of women (AMA women and women with a hypo-response to OS), LH and FSH deficiency may be the underlying cause of an unfavourable prognosis. Indeed, OS with r-hFSH:r-hLH has been reported to improve reproductive outcomes in these subgroups ( Alviggi et al. , 2018a ). In addition, preliminary findings suggest that serine carriers of the LHCGR variant (rs 2293275) might require higher amounts of r-hLH during OS together with r-hFSH than do aspargine homozygous carriers ( Ramaraju et al. , 2018 ). Should these data be confirmed in larger studies, they will facilitate the identification of subpopulations of women with potential LH and FSH deficiency and consequently a poor MAR prognosis. Recently, the POSEIDON (Patient-Oriented Strategies Encompassing IndividualizeD Oocyte Number) group proposed a classification of low prognosis based on a confirmed or expected inappropriate ovarian response to gonadotropin stimulation ( Alviggi et al. , 2016a ; Humaidan et al. , 2016 ). The POSEIDON concept contributes to identifying more homogeneous populations for clinical studies. Indeed, a large retrospective study of 18 455 cycles found that, while the cumulative live birth rate was similar in younger patients with normal ovarian parameters and a poor prognosis (POSEIDON Groups 1 and 3, respectively), it was significantly lower in women of advanced age with a normal ovarian reserve (POSEIDON Group 2 vs. Group 1) ( Shi et al. , 2019 ). Interestingly, according to systematic reviews, pregnancy rate and implantation rate were higher in hypo-responders treated with r-FSH and r-hLH versus those treated with r-hFSH alone ( Alviggi et al. , 2018a ; Conforti et al. , 2019b ). Nevertheless, to evaluate the effect of gonadotropins in women with potential LH and FSH deficiency, studies should include information on endocrinological and clinical outcomes. Relevant endocrine endpoints, including serum testosterone, serum E 2 values and the E 2 /oocyte ratio could elucidate the stimulatory effect exerted by LH during OS on steroidogenesis in both theca and granulosa cells, and may lead to biomarkers for the diagnosis of LH and FSH deficiency before and during OS. Moreover, clinical end points such as FORT and FOI can directly measure the downstream clinical effect of gonadotropins in relation to the baseline antral follicle count. Ultimately, individualised treatment aimed at optimising FORT and FOI will reveal the highest number of oocytes attainable for each cycle and each patient. This in turn could increase the probability of live birth because oocyte number is positively correlated with the number of good quality embryos ( Vermey et al. , 2019 ). Lastly, given that the probability of live birth is also related to endometrial receptivity, it is important to evaluate if and how different gonadotropin regimens affect endometrial receptivity, how the luteal phase can be effectively supported in women with potential LH and FSH deficiency ( Revel, 2012 ), and when a freeze all strategy with subsequent frozen embryo transfer cycles is superior to embryo transfer in a fresh cycle with respect to efficacy and safety.

No new data were generated or analysed in support of this research.