Advances in anticancer treatments have resulted in increasing survival rates among cancer patients. In South Korea, the overall cancer mortality rate decreased by 2.7% annually between 2002 and 2016. A significant improvement was also evident in the 5-year survival rate, which increased to 70.6% for patients diagnosed with cancer between 2012 and 2016, compared with 41.2% for patients diagnosed between 1993 and 1995 [1]. Accordingly, the idea that quality of life after cancer should be actively considered during cancer treatment has become widely accepted. In particular, the preservation of fertility in younger patients receiving cancer treatment has gradually become an essential factor for consideration [2-4]. The main problems associated with chemotherapy in female cancer survivors include early menopause and an increased subfertility rate [5]. For fertility preservation in women, embryo and oocyte cryopreservation are considered to be the standard practice and are widely available [2-4]. However, because embryo and oocyte cryopreservation requires approximately 2 weeks, it is difficult to perform cryopreservation in patients with cancers for which treatment is more urgent due to rapid progression. In addition, cryopreservation may not be an option for all women for economic reasons.

There has been tremendous interest in medical agents that can potentially preserve fertility from the ovarian toxicity of chemotherapy, and gonadotropin-releasing hormone agonists (GnRHa) have been considered to be the most likely category of drugs for this purpose. In 1981, an animal study demonstrated that GnRHa administration protected male mice from gonadal damage caused by cyclophosphamide [6]. An observational long-term follow-up study of children treated with chemotherapy for Hodgkin disease showed that prepubertal administration of chemotherapy agents resulted in less ovarian damage than in similarly treated adult patients, and it was suggested that GnRHa could exert a protective role by creating a temporary “prepubertal” state in women of reproductive age [7]. The first randomized controlled trial (RCT) of GnRHa for preserving fertility during chemotherapy was conducted by Waxman et al. [8] in 1987. The authors reported that buserelin was not effective in preserving fertility. Since this first RCT was reported, several other RCTs and observational studies have been conducted in patients with various diseases, including breast cancer, hematological malignancies, and autoimmune diseases. However, these studies have reported conflicting results [9].

Although the mechanisms of protection have yet to be fully understood, as the results of large-scale RCTs were recently released, researchers’ understanding of the conditions under which GnRHa can preserve fertility has been expanded and clarified. This trend was reflected in the 2018 American Society of Clinical Oncology (ASCO) guidelines for fertility preservation [4]. The 2018 ASCO guidelines stated for the first time that GnRHa may partially help in preserving fertility in patients undergoing chemotherapy. These changes in the ASCO guidelines have also influenced other guidelines released after 2018 [10,11]. Nevertheless, evidence regarding the efficacy of GnRHa for fertility preservation is still regarded as insufficient, and GnRHa use during chemotherapy does not replace established fertility preservation methods. In this review, we describe the scope of utility and limitations of GnRHa in preserving fertility, as well as the mechanisms by which GnRHa protects the ovaries from chemotherapy.

Although controversial [12], it is generally believed that the number of follicles held within the primordial follicle (PMF) pool is finite. PMFs form from 17 weeks of gestation, and consist of an immature oocyte in meiotic arrest surrounded by a single layer of granulosa cells. During the reproductive age, both the oocyte and surrounding granulosa cells undergo stages of sequential growth and development, characterized by granulosa cell proliferation and oocyte growth, until resumption of meiosis by the ovulatory luteinizing hormone (LH) surge. PMFs are continuously being recruited out of the pool of germ cells and activated to grow until few remain. The presence of fewer than 1,000 remaining oocytes has been associated with menopause [13,14].

The risk of developing premature ovarian insufficiency (POI) after chemotherapy can vary depending on various factors. Some chemotherapy regimens are considered more gonadotoxic than others, with particularly strong evidence for the high toxicity of alkylating agents [14]. Dosage of the treatment used is also an important consideration, and previous studies have shown that early menopause occurs in a dose-dependent fashion [15]. In addition, the age of the patient at treatment is pivotal, as older women have a much higher reported incidence of POI during or immediately after treatment [5,16]. Chemotherapy undoubtedly has a detrimental effect directly on follicles [14], and previous studies have shown that chemotherapy acts on the ovary in a wide variety of ways (Table 1).

Although there is still controversy regarding the efficacy of GnRHa on ovarian protection, recent studies have been able to better clarify the efficacy of temporary ovarian suppression with GnRHa during chemotherapy, especially in breast cancer patients (Table 1).

The lack of a standardized definition of chemotherapy-associated POI has been identified in several MAs as a limitation to the interpretation of findings in the literature regarding GnRHa. All MAs covered in this review analyzed the amenorrhea rate or resumption of menses as the primary outcome and the rate of spontaneous pregnancy as the secondary outcome, except for two studies [88,89]. Three studies assessed the results of the primary endpoint without setting a specific time point [90-92], one study analyzed the primary outcome 2 years after completion of treatment [88], and the remaining studies included analyses of the primary outcome at 1 year after the end of chemotherapy [89,93-96]. The rate of amenorrhea 1 year after the end of chemotherapy is a commonly adopted primary endpoint in several RCTs [79,80,82,86,87] because menopause is clinically defined as the absence of menstruation for a year, and resumption of menses is a clinically relevant and reproducible outcome [93].

However, resumption of menses does not necessarily translate into restoration of fertility [93]. Oktay and Turan [97] emphasized that menstruation or amenorrhea should not be considered surrogate markers of GnRHa efficacy in preserving the ovarian reserve. However, owing to the lack of data related to pregnancy, controversy remains as to whether GnRHa administration will really help future pregnancies in premenopausal women receiving chemotherapy. Ten SRs and MAs have been published in the past 5 years, seven of which have reported results on pregnancy outcomes. Five of the seven MAs found significantly higher rates of pregnancy in patients receiving chemotherapy and GnRHa than in those receiving chemotherapy alone [90,91,93-95]. In particular, studies that included only breast cancer patients showed clearer results for efficacy of GnRHa on the spontaneous pregnancy rate after chemotherapy [90,93-95]. However, as Munhoz et al. [94] noted in their MA, these analyses of pregnancy rates are closer to exploratory analyses than to a valid endpoint for a main analysis. The results of pregnancy outcomes in MAs might be affected by several biases. First, although pregnancy may be the most specific indicator of gonadal function, subfertility can be multifactorial, and inability to conceive does not necessarily indicate hypogonadism [88]. Second, limited information is available on the number of patients interested in future pregnancies at the time of randomization and on those who attempted to become pregnant [91]. Only one of the RCTs reported the number of women attempting pregnancy and calculated the pregnancy rate for this subgroup [83]. Therefore, most MAs used the number of the entire randomized population as the denominator for comparing pregnancy outcomes, rather than the number of women who actually attempted pregnancy. However, multiple biases are also possible when using the number of patients attempting pregnancy as a denominator for the pregnancy rate. For example, without blinding, intervention assignment can affect the likelihood of attempting pregnancy. In addition, because a significantly lower rate of participants developed POI in the group administered GnRHa, the intervention itself had an effect on the size of the group in which attempting pregnancy was possible. Furthermore, as observed in the POEMS-SWOG/S0230 study, some pregnancies might occur in women who did not attempt pregnancy [83]. Finally, the study period of the RCTs included in the MAs varied. A short median follow-up of the studies might be a possible explanation of the limited number of pregnancies observed, especially in trials including patients with HR-positive disease who received adjuvant endocrine therapy for at least 5 years [78,82,85], thus delaying attempts for pregnancy.

GnRHa was recently reported to be effective in RCTs including breast cancer patients [74,75,79,81-83,85,87], whereas GnRHa was found to be ineffective in preserving fertility in a large-scale RCT targeting only lymphoma patients [84]. Conflicting results have been reported in MAs. Elgindy et al. [80] observed that GnRHa administration during chemotherapy did not protect the ovaries from gonadal toxicity regardless of cancer type. Senra et al. [91] reported that GnRHa had a significant benefit on the risk of POI in breast cancer patients (relative risk [RR], 0.57; 95% CI, 0.43–0.77), but not in lymphoma patients (RR, 0.70; 95% CI, 0.20–2.47). Meanwhile, Sofiyeva et al. [89] observed a gonadoprotective effect of GnRHa in both hematological malignancies (RR, 1.77; 95% CI, 1.15–2.74) and breast cancer (RR, 1.31; 95% CI, 1.05–1.62) .

In fact, the difference in the efficacy of GnRHa may not be attributable to the type of cancer, but rather to the regimen of chemotherapy used for each cancer type. It is well known that gonadotoxicity varies depending on the type of chemotherapeutic agent and the duration of administration [2]. To overcome this heterogeneity, some early RCTs enrolled only those who received the same anticancer regimen to evaluate the efficacy of GnRHa [74-76]. However, unlike in RCTs, it is difficult to control the different anticancer regimens and different administration periods of the studies included in MAs. Several MAs have reported inconsistent results for risk of POI after administering different regimens and agents. A recent MA found that women exposed to taxanes demonstrated a lower rate of menstrual recovery than those treated with chemotherapy regimens that did not contain taxanes (OR, 0.49; 95% CI, 0.30–0.80; p=0.004) [98]. In another MA that included only early breast cancer patients who were administered GnRHa for the purpose of fertility preservation, whether or not a taxane was included in the anticancer regimen was not an independent variable associated with the risk of developing chemotherapy-induced POI [95]. In the same MA, the duration of chemotherapy (>4 months) was also not associated with the risk of POI [95].

In another MA including breast cancer and lymphoma patients [91], when all patients were subdivided according to chemotherapy regimen into high toxicity and low to medium toxicity groups, the benefit of GnRHa therapy in preventing POI appeared to be of a greater magnitude among women treated with low- to medium-toxicity chemotherapy (RR, 0.49; 95% CI, 0.29–0.84); this is in comparison to women treated with highly toxic chemotherapy (RR, 0.66; 95% CI, 0.45–0.96). The major mechanism of ovarian protection by GnRHa is believed to be the downregulation of pituitary GnRH receptors with a drastic reduction of serum gonadotropin levels and blocking of follicular recruitment [91,99]. The authors assumed that the reason for a weaker protection of GnRHa against highly toxic chemotherapy was that these agents cause ovarian damage through multiple mechanisms beyond follicle destruction, such as cortical fibrosis, vascular lesions, and accelerated atresia [91,100].

To date, there is a paucity of data available on the appropriate follow-up period to observe recovery of ovarian function after anticancer treatment. The difference in follow-up duration between groups [101] and between studies is a another possible source of heterogeneity and bias in the studies selected for MAs, which in turn could impact the chance of POI diagnosis and pregnancy outcomes [93]. Although there are several RCTs with a follow-up duration of 1 year after the end of chemotherapy, recent large-scale RCTs tended to have a follow-up duration of 2 years or more [75,78,82-84]. Recently, as RCTs have included HR-positive breast cancer patients, the study period has also been growing longer. In a trial by Lambertini et al. [82], because patients with HR-positive breast cancer were included, patients were followed up for more than 5 years to avoid endocrine treatment masking resumption of menses (mean follow-up, 7.3 years) . Moreover, an RCT showed different results within 2 years of follow-up compared with follow-up for only 1 year. For example, Sverrisdottir et al. [75] have observed that the proportion of menstruating women in the GnRHa group showed a statistically significant increase between 24 and 36 months, in contrast to all the other groups in which menses were unchanged or decreasing in the study. A short follow-up period may be responsible for discrepancies between studies, leading to premature conclusions [102].

The time at which ovarian function was evaluated has varied widely [88]. Therefore, several MAs have analyzed the outcomes at various time points, except for three studies that did not set a specific time point for assessing outcomes [90-92]. Eight MAs out of the nine studies in listed in Table 3 found a positive impact on preservation of ovarian function with GnRHa treatment compared with chemotherapy alone. Meanwhile, three of the eight MAs reported that the benefits of GnRHa administration may change over time [88,95,96]. Hickman et al. [88] reported that the effect of GnRHa treatment on ovarian protection was not valid after 2 years (OR, 0.53; 95% CI, 0.22–1.30). Chen et al. [96] also reported no difference during a follow-up period longer than 12 months between the GnRHa group and controls (RR, 1.08; 95% CI, 0.95–1.22). Meanwhile, the incidence of menstrual recovery was significantly higher in the GnRHa group during a follow-up period no longer than 12 months, with an overall effect favoring treatment with GnRHa (RR, 1.60; 95% CI, 1.14–2.24; p=0.006). However, an MA by Lambertini et al. [95] presented a non-significantly reduced risk of amenorrhea at 1 year after chemotherapy (OR, 0.92; 95% CI, 0.66–1.28; p=0.623). However, they found a significantly reduced risk at 2 years (OR, 0.51; 95% CI, 0.31–0.85; p=0.009) . Thus far, the duration of benefits from GnRHa cotreatment appears to be unclear and requires further study.

Age has been identified as a major determinant of the risk of POI after chemotherapy [98,103]. This age-related difference is most likely a result of the reduction of the PMF pool with aging, resulting in an increase in the risk of developing ovarian failure and infertility in older women after cytotoxic treatment, even at smaller doses [104]. According to recent MAs, this tendency is prominent in patients 40 years and older [89,90,95,98]. Patient age for eligibility varies considerably across trials, and only four RCTs set age limitations for eligibility to 40 years [74,76,80,85]. In the studies by Song et al. [79] and Munster et al. [78], the mean patient age was 41 and 45 years, respectively, which may have contributed to diminished recovery. Future clinical trials verifying the efficacy of GnRHa for ovarian protection should be designed considering the patient’s age.

There have been safety concerns regarding the potential negative effect of the concurrent use of tamoxifen with chemotherapy based on preclinical and clinical evidence [93,105-107]. However, GnRHa has been found to have different pharmacodynamic properties from those of tamoxifen. Since the 1990s, there have been randomized studies that investigated the impact of concurrent ovarian function suppression (obtained pharmacologically or with surgery or radiotherapy) with chemotherapy, and these findings did not demonstrate differences in patients’ prognoses [108-110]. Moreover, in recently published trials, excellent survival outcomes were reported with the use of GnRHa concomitantly with chemotherapy in patients, including HR-positive patients [82,83,86,111]. In an MA by Lambertini et al. [82], no difference was noted in the prognosis of breast cancer according to HR status when GnRHa was co-administered. An analysis according to HR status showed no significant interaction (p_interaction=0.762); and the adjusted hazard ratios for disease-free survival were 0.79 (95% CI, 0.24–2.59) and 0.65 (95% CI, 0.39–1.07) in patients with HR-positive and HR-negative disease, respectively. According to these favorable results, the OPTION trial amended its protocol to allow enrollment of women with HR-positive tumors [85]. The original protocol of the OPTION trial restricted enrollment to only those with HR-negative tumors.

It has also been speculated that GnRHa will not have an ovarian-protective effect on HR-positive breast cancers [112]. However, recent trials that included HR-positive breast cancer patients demonstrated favorable outcomes in preventing POI through the concurrent use of GnRHa with chemotherapy [75,79,82,85,86]. In an MA by Senra et al. [91], HR status was determined not to influence the effect of GnRHa on the risk for POI, which was comparable among women with receptor-positive (RR, 0.69; 95% CI, 0.48–1.00) and receptor-negative (RR, 0.62; 95% CI, 0.33–1.14) tumors.

However, opinions still vary among guidelines. Although the St. Gallen International Expert Consensus [113] and the National Comprehensive Cancer Network Guidelines [112] have recommended the use of GnRHa only for patients with HR-negative breast cancer, the Breast Cancer in Young Women-2 Panel concluded that a protective effect was likely in both HR-negative and HR-positive patients, without an increased risk of breast cancer recurrence [114]. Recent findings further support the position that GnRHa has an ovarian protection effect in breast cancer patients regardless of HR status.

Several studies have reported that AMH was an accurate biomarker for assessing the extent of ovarian damage after chemotherapy [115-120]. Dunlop and Anderson [115] found that pretreatment AMH can predict POI or ongoing ovarian activity after chemotherapy. Despite the fact that the values of AMH fluctuate before 25 years of age in women, there are reports that AMH may be of value for assessing ovarian function in prepubertal girls after cancer treatment [116,117]. Thus, some trials have analyzed AMH as a secondary outcome to examine the ovary-protective effect of GnRHa [72,76,77,80,85].

However, some researchers are skeptical regarding the role of AMH in predicting chemotherapy-associated ovarian dysfunction. It is well known that AMH does not predict spontaneous pregnancy. A recent prospective time-to-pregnancy cohort study including 750 women between the ages of 30 and 44 found no association between AMH levels and rates of spontaneous conception [121]. Similarly, a retrospective study showed that AMH levels did not accurately predict pregnancy in breast cancer patients who underwent chemotherapy [122]. As in the study of Demeestere et al. [84], in which five patients with protocol-defined POI became pregnant during follow-up, low AMH levels do not necessarily indicate infertility. In addition, patients who have undergone ovarian tissue cryopreservation often have nearly undetectable AMH levels, but decent pregnancy rates have been reported [123].

Currently, no MAs have analyzed AMH levels as an outcome. The RCTs by Giuseppe et al. [72], Elgindy et al. [80], and Leonard et al. [85] reported that there was no difference in changes of AMH levels after chemotherapy between the GnRHa group and controls, and Gerber et al. [77] reported incomplete data regarding AMH. Meanwhile, an RCT by Leonard et al. [85] reported that AMH levels before chemotherapy were a predictor of posttreatment amenorrhea (p<0.001). However, after adjustment for age, the effect of pretreatment AMH was no longer significant.

In the early days, the use of GnRHa for preserving fertility showed conflicting results. However, recent trials have reported that the use of GnRHa is helpful in preventing POI following chemotherapy, at least in breast cancer patients, owing to the overall increase in the number of patients enrolled in trials and the longer follow-up periods. In our opinion, all young women (even patients with high-risk leukemia) facing gonadotoxic chemotherapy should be counselled about and offered various options for fertility preservation, including both GnRHa and cryopreservation of embryos, oocytes, and ovarian tissue.

However, it remains unclear to what extent GnRHa coadministration could provide benefits for achieving pregnancy after chemotherapy. In general, attempting pregnancy after cancer treatment must be delayed until chemotherapy is out of the patient’s system and the patient is at a low risk of recurrence. Experts suggest that the timing of pregnancy should be individualized [88,124]. In patients with HR-positive breast cancer who wish to become pregnant, it may be more difficult to decide when to attempt pregnancy, because these patients are typically treated with tamoxifen for 5 to 10 years [88,124]. The POSITIVE study is currently underway to determine the safety of an interruption in endocrine therapy to allow childbearing, and the results of this ongoing research have attracted significant attention.