Effects of various controlled ovarian hyperstimulation protocols and surgery on pregnancy outcomes in women with endometriosis

Abstract

Endometriosis is a common gynecological condition in women of childbearing age that causes symptoms such as menstrual changes and dysmenorrhea, and is also a major cause of infertility. Therefore, women with endometriosis usually need to use assisted reproductive technology (ART), such as in vitro fertilization or intracytoplasmic sperm injection, to increase their chances of conceiving. Numerous clinical observations and studies have indicated that endometriosis can affect the success of ART, such that women with endometriosis who use ART have a lower live-birth rate than those without endometriosis who use ART. Therefore, this article reviews the impact of various controlled ovarian hyperstimulation protocols and surgery on the pregnancy outcomes of women with endometriosis using ART to explore the selection of individualized treatment.

Keywords:

• Endometriosis
• pregnancy outcome
• ART
• controlled ovarian hyperstimulation
• surgery

Background

Endometriosis is an estrogen-dependent disease characterized by the presence of endometrial tissue outside the uterine cavity [1]. It is a major cause of infertility, with studies indicating that approximately 30–50% of women with endometriosis experience infertility and 20–50% of women with infertility also have endometriosis [2]. There are various treatments for infertility related to endometriosis, with the most effective being assisted reproductive technology (ART), such as in vitro fertilization – embryo transfer (IVF–ET) [3]. Nonetheless, women with endometriosis have a lower live-birth rate than those with infertility due to other causes, and this difference increases as the severity of endometriosis increases [4]. Adverse pregnancy outcomes in women with endometriosis may be related to impaired sperm function, poor ovarian reserve, a decreased number of retrieved oocytes, low-quality oocytes and embryos, decreased uterine receptivity, and an inflammatory pelvic environment, especially in advanced stages of the disease [5, 6]. Therefore, a determination of the ability of various controlled ovarian hyperstimulation protocols and surgery to improve pregnancy outcomes in women with endometriosis would not only assist clinicians to treat infertile women with endometriosis but also assist such women to achieve fertility.

Mechanism by which endometriosis affects the outcome of IVF–ET

The effect of endometriosis on the fertility of women has been extensively examined in several references. Moreover, the relationship between endometriosis and infertility is multifaceted, as it involves the disruption of the normal anatomical structures of the fallopian tubes and ovaries; a decrease in oocyte and embryo quality due to inflammation and oxidative stress; and changes in the receptivity of the endometrium [7, 8].

Effect of endometriosis on the quality of oocytes and embryos

The quality of oocytes and embryos is an important factor that affects the success of in vitro fertilization (IVF). However, the impact of endometriosis on the quality of oocytes and embryo development remains debated. Current research indicates that endometriosis decreases the quality of oocytes and the potential for embryo development via several pathological mechanisms, such as by increasing the production of reactive oxygen species and other free radicals, inducing immune-system imbalance, and impairing extracellular matrix remodeling [9]. In addition, cell biology research has shown that the general morphology, maturation ability, and organelles of oocytes from women with infertility due to endometriosis differ from those of oocytes from women with infertility caused by other factors. For example, Goud et al. evaluated immature oocytes from infertile women with endometriosis and infertile women with other conditions. They observed that compared with the latter group, the former group showed a greater loss of cortical granules and a harder zona pellucida, which may interfere with fertilization, zona pellucida dissolution, embryo hatching, and implantation. Additionally, they tested the ability of immature oocytes to undergo in vitro maturation (IVM) to the metaphase II (MII) stage. They found that the number of germinal vesicles and oocytes reaching the MII stage was significantly lower in the endometriosis group than in the control group, and during IVM, a higher proportion of abnormal spindle fibers were observed in oocytes from women with endometriosis than in oocytes from women using ART as a treatment for male factor infertility (66.7% vs 16%, p < 0.05) [10]. Moreover, Xu et al. employed transmission electron microscopy to investigate 50 MII stage oocytes from women diagnosed with stage I–II endometriosis and tubal or male factor infertility. They found that compared with oocytes from the control group, those from the endometriosis group had a higher proportion of abnormal mitochondria, such as small or swollen and blurry vesicles, and fewer mitochondria. In addition, real-time quantitative polymerase chain reaction detected lower copy numbers of mitochondrial DNA in the endometriosis group than in the control group [11].

However, although the above-mentioned results of basic experiments suggest that endometriosis affects the quality of oocytes and embryos, the results of clinical assisted reproductive treatments do not all support this. For example, an early study on oocyte donation found that women with and without endometriosis had the same implantation and pregnancy rates after they had received oocytes from healthy donors, with these rates decreasing after they had received oocytes from donors with endometriosis [12]. However, this study had a small sample size. In 2022, Kamath et al. published a study reporting the live birth rates of 758 women with endometriosis who had undergone cycles of IVF with donated oocytes and those of 12,856 women with endometriosis who had undergone cycles of autologous IVF. After adjusting for confounding factors, they found that the live birth rates of the women who had received donated oocytes for fresh and frozen embryo transfers were not significantly different from those of the women who had undergone autologous IVF. Therefore, they concluded that oocyte quality may have a limited effect on IVF outcomes in women with endometriosis [13]. However, they did not report the severity of the women’s endometriosis or the number of oocytes that had been retrieved. It is also unclear whether the donors had been screened for endometriosis or if some women with endometriosis had coexisting adenomyosis, both of which might have affected the results. A retrospective study published in the same year found that compared with infertile women without endometriosis, infertile women with endometriosis had a significantly reduced ovarian reserve and response to stimulation but did not have significantly worse oocyte quality or clinical outcomes, such as the clinical pregnancy rate and cumulative live birth rate [14]. However, a retrospective study by Wu et al. published in 2021 suggested that endometriosis does have a negative impact on the quality and quantity of oocytes but not on overall pregnancy outcomes [15]. In addition, a retrospective analysis by Sanchez et al. found that endometriosis does not affect the fertilization rate, quality of cleavage-stage embryos, number of blastocysts, and blastocyst rate, but does decrease ongoing pregnancy rates [16]. The differences between the results of the above-described studies demonstrate that further clinical research is needed to clarify how endometriosis affects oocyte and embryo quality.

Effect of endometriosis on endometrial receptivity

Embryo implantation requires both high-quality embryos and a well-prepared uterine lining [17]. Endometriosis can affect the receptivity of the uterine lining and is primarily associated with imbalances in steroid-hormone signaling (such as upregulation of estrogen function and resistance to progesterone), differential gene expression in the uterine lining, immune abnormalities, and abnormal expression of cell adhesion molecules [17–19]. Moreover, compared with women without endometriosis, those with endometriosis have lower levels of expression of uterine endometrial hormone receptors and, in the mid-secretory phase of the endometrium, higher levels of expression of estrogen receptor 1 [20]. Additionally, a comparative transcriptomic analysis of eutopic and ectopic endometria between women with and without endometriosis revealed that women with endometriosis exhibit dysregulation of selected genes related to implantation, with decreased expression of homeobox A10 (HOXA10) and HOXA11 in the luteal phase. This decreases the level of transcription of empty spiracles homeobox 2 and directly affects the proliferation and function of endometrial cells before and after implantation, resulting in abnormal implantation [6].

The impact of the immune system on the endometrial microenvironment and implantation window is currently unclear. Compared with women without endometriosis, those with endometriosis have significantly higher concentrations of pro-inflammatory cytokines (interleukin [IL]-1α, IL-1β, and IL-6) in their endometrial microenvironment. In addition, unlike the former group, the latter group has activated class I macrophages (which secrete pro-inflammatory factors) as the main cell population in the endometrium throughout the menstrual cycle. This dominance of a pro-inflammatory phenotype may affect embryo implantation [7, 19].

However, the extent to which the above-described genetic and immunological factors affect the clinical outcomes of assisted reproduction in women with endometriosis remains debated. In one example, it was reported that endometriosis negatively affects endometrial receptivity, leading to a decrease in implantation and ongoing pregnancy rates [21]. Moreover, considering that hormone concentrations greater than normal physiological concentrations during conventional stimulation may impair endometrial receptivity, regulated thawed embryo transfer is often employed to restore optimal receptivity in women with endometriosis and thus improve pregnancy rates. A large retrospective study found that in women with advanced endometriosis, implantation rates, clinical pregnancy rates, and live birth rates were significantly higher after frozen embryo transfer than after fresh embryo transfer [22]. A retrospective matched-cohort study of 135 women with endometriosis who had undergone either fresh or frozen embryo transfer revealed that the cumulative clinical pregnancy rates and cumulative ongoing pregnancy rates were significantly higher in the frozen embryo transfer group than in the fresh embryo transfer group [23]. In contrast, another retrospective matched cohort study found that compared with women with tubal factor infertility, women with endometriosis had lower clinical pregnancy rates and live birth rates (p = 0.028 and p = 0.008, respectively) in frozen–thawed embryo transfer cycles, leading to a decrease in cumulative live birth rates (p = 0.001) [24]. Similarly, a retrospective analysis conducted by Peking Union Medical College Hospital found no significant differences in clinical pregnancy, implantation, live birth, miscarriage, or multiple pregnancy rates between women with endometriosis who had received high-quality embryo transfer and women with tubal factor infertility who had received the same treatment, suggesting that endometriosis does not alter endometrial receptivity [25]. Moreover, Pirtea et al. posited that reduced endometrial receptivity may play a minimal or no role in ART failures [26]. Therefore, although the above-described studies have confirmed that there are differences between the endometrial characteristics of women with endometriosis and those of women without endometriosis, there remains a need for more clinical data to validate the impact of changes in endometrial receptivity on pregnancy outcomes.

Effect of various controlled ovarian hyperstimulation protocols on the outcomes of IVF in women with endometriosis

The inflammatory reaction caused by endometriosis within the pelvic cavity can disrupt the interaction between sperm and oocytes, reducing the chances of fertilization. However, this inflammatory reaction does not affect IVF, which means that IVF is an excellent option for women with endometriosis-related infertility [27]. Controlled ovarian hyperstimulation is implemented in an IVF cycle to obtain a sufficient number of oocytes to increase the cultivation and quality of embryos and thus optimize the chance of pregnancy [28]. Commonly used ovulation hyperstimulation protocols include ultralong, long, short, antagonist, and progesterone protocols. Which of these protocols is best for women with endometriosis-related infertility remains to be determined, as there is insufficient clinical research in this area. The results of recent research are summarized in Table 1.

Table 1. Effect of various controlled ovarian hyperstimulation protocols on the outcomes of IVF in women with endometriosis.

Article	Type	Subgroup	Experimental group	Control group	Retrieved oocytes	Fertilization rate	Implantation rate	Clinical pregnancy rate	Live birth rate
Liu et al. [30]	Meta-analysis	/	Ultralong protocol	Long protocol	MD = 0.70,95%CI: (-0.67–2.07), p = 0.31	RR = 0.97,95%CI: [0.93, 1.01], p = 0.19	RR = 1.37,95%CI:(0.78–2.38), p = 0.27	RR = 1.33,95%CI:(1.13–1.56), p = 0.002	RR = 2.07,95%CI:[1.00,4.27], p = 0.05
Cao et al. [31]	Meta-analysis	RCTs	Ultralong protocol	Long protocol	MD=-0.2,95%CI:(−0.50–0.10), p > 0.05	RR = 0.97,95%CI:(0.93–1.01), p > 0.05	RR = 1.37,95%CI:(0.78–2.38), p > 0.05	RR = 1.44,95%CI:(1.21–1.72), p < 0.05 stages I–II RR = 0.99,95%CI:(0.64–1.55), p > 0.05 stages III–IV RR = 2.04,95%CI:(1.37–3.04), p < 0.05	/
		Non-RCTs	Ultralong protocol	Long protocol	MD=-0.09,95%CI:(−0.39–0.22), p > 0.05	RR = 1.02,95%CI:(0.85–1.22), p > 0.05	RR = 1.18,95%CI:(0.78–2.38), p > 0.05	RR = 1.05,95%CI:(0.93–1.20), p > 0.05 stages I–IIRR= 1.05,95%CI:(0.80–1.37), p > 0.05 stages III–IVRR= 1.16,95%CI:(0.93–1.44), p > 0.05	/
		Non-RCTs	Ultralong protocol	Short protocol	MD = 0.50,95%CI:(0.27–0.74), p < 0.05	RR = 1.19,95%CI:(1.01–1.40), p < 0.05	RR = 1.85,95%CI:(0.58–5.90), p > 0.05	RR = 1.78,95%CI:(1.07–2.97), p < 0.05 stages I–II RR = 1.21,95%CI:(0.58–2.53), p > 0.05 stages III–IV RR = 2.28,95%CI:(0.88–5.91), p > 0.05	/
Pabuccu [33]	RCT	Stages of I–II endometriosis	GnRH antagonist	Long protocol	/	/	15.4% vs 18.2%, p = 0.90	30% vs 31.2%, p = 1.00	/
		History of ovarian surgery for endometrioma	GnRH antagonist	long protocol	4.3 ± 2.6 vs. 8.8 ± 4.6, p = 0.0001	/	15.9% vs 22.6%, p = 0.60	27.5% vs 39%, p = 0.38	/
		Endometrioma and no history of ovarian surgery	GnRH antagonist	Long protocol	6.7 ± 2.6 vs 8.2 ± 5.5, p = 0.002	/	12.5% vs 14.8%, p = 0.70	20.5% vs 24.2%, p = 0.90	/
Kolanska et al. [34]	Retrospective analysis	Fresh ET	GnRH-a	GnRH antagonist	/	/	/	25% vs 13%, p = 0.02	18% vs 8%, p = 0.04
		FET	GnRH-a	GnRH antagonist	/	/	/	5% vs 7%, p = 0.70	2% vs 7%, p = 0.09
		Fresh ET+FET	GnRH-a	GnRH antagonist	/	/	/	29% vs 18%, p = 0.06	21% vs 14%, p = 0.2,
Drakopoulos et al. [36]	Retrospective cohort study	Endometriosis stages I–II	GnRH-a	GnRH antagonist	Median(IQR) 9 (6–13) vs 7 (5–12), p = 0.09	/	/	50% vs 36%; p = 0.14	42.8% vs 26.7%, p = 0.07
		Endometriosis stages III–IV	GnRH-a	GnRH antagonist	Median(IQR) 8 (5–11) vs 7 (5–11), p = 0.33	/	/	34.3% vs 32.5%; p = 0.70	27.3% vs 23.8%, p = 0.50
Guo et al. [41]	RCT	/	MPA + HMG	Ultralong protocol	Mean ± SD:9.30 ± 5.73 vs 9.33 ± 5.36 p = 0.959	65.23% (910/1395) vs 65.35% (915/1400), p = 0.945	MPA + HMG-FET vs ultralong-ET vs ultralong-FET: 34.27% (98/286) vs 33.85% (65/192) vs 39.67% (48/121), p = 0.517	MPA + HMG-FET vs ultralong-ET vs ultralong-FET: 50.31% (79/157) vs 55 %(55/100) vs 48.53% (33/68), p = 0.67	MPA + HMG-FET vs ultralong-ET vs ultralong-FET:43.95 %(69/157) vs 49.00 %(49/100) vs39.71% (27/68), p = 0.48
Guo et al. [40]	RCT	/	/	/	Mean ± SD:MPA + hMG vs dihydrogesterone + hMG vs progesterone + hMG :9.30 ± 5.7 vs 8.00 ± 4.5 vs 7.60 ± 5.2, p = 0.021	/	MPA + hMG vs dihydrogesterone + hMG vs progesterone + hMG :33.8% vs 34.2% vs 38.9%, p > 0.05	MPA + hMG vs dihydrogesterone + hMG vs progesterone + hMG :49.6% vs 57.9% vs 56.2%, p > 0.05	/
Zhao et al. [37]	Retrospective study	/	/	/	/	/	Ultralong protocol vs GnRH antagonist vs long protocol :25.16% vs 18.01% vs 17.16% p > 0.05	Ultralong protocol vs GnRH antagonist vs long protocol :45.24% vs 33.33% vs 28.99% p > 0.05	Ultralong protocol vs GnRH antagonist vs long protocol :32.14% vs 19.54% vs 24.64% p > 0.05
Chen et al. [35]	Retrospective study	Fresh ET	GnRH antagonist	GnRH-a	Mean ± SD:6.17 ± 4.49 vs 6.99 ± 4.86 p = 0.213	77.87% (461/592) vs 81.22% (545/671) p = 0.140	18.87% (20/106) vs 35.90% (56/156) p = 0.003	28.57% (18/63) vs 50.57% (44/87) p = 0.007	19.05%(12/63) vs 41.38%(36/87) p = 0.004
		FET	GnRH antagonist	GnRH-a	Mean ± SD:6.17 ± 4.49 vs 6.99 ± 4.86 p = 0.213	77.87% (461/592) vs 81.22% (545/671) p = 0.140	33.75% (27/80) vs 28.57% (14/49) p = 0.540	47.27% (26/55) vs 37.5% (12/32) p = 0.375	41.82%(23/55)vs 34.38% (11/32) p = 0.493

Note. MPA = medroxyprogesterone acetate; hMG = human menopausal gonadotropin; GnRH-a = gonadotropin-releasing hormone agonist; FET = frozen–thawed embryo transfer; ET = embryo transfer.

MD = mean difference; CI = confidence interval; RR = risk ratio; IQR = interquartile range; SD = standard deviation.

Gonadotropin-releasing hormone agonist protocols

Gonadotropin-releasing hormone agonists (GnRH-a) significantly reduce secretion of luteinizing hormone (LH) by blocking the production of the LH-β subunit, and maintain responsiveness to exogenous GnRH injections by increasing the production of the LH-α subunit. Furthermore, GnRH-a can prevent premature luteinization of follicles, alleviate inflammation, and improve the microenvironment of the pelvic cavity, thereby aiding the growth and development of follicles [29]. Consequently, GnRH-a are used to downregulate production of LH in women with endometriosis-related infertility and are typically employed in ultralong, long, or short ovulation hyperstimulation protocols.

A 2021 meta-analysis by Shuai Liu et al. [30] examined the effect of ultralong protocols versus long protocols on the outcomes of IVF/intracytoplasmic sperm injection – embryo transfer (IVF/ICSI-ET) in women with endometriosis-related infertility. This meta-analysis included nine randomized controlled trials (RCTs) involving a total of 943 participants and found that compared with a long protocol, an ultralong protocol with a 3-month suppression period increased the clinical pregnancy rate and ongoing pregnancy rate (relative risk RR = 1.90, 95% CI: 1.39–2.61, p < 0.0001) in infertile women with endometriosis. However, there was no difference in the clinical pregnancy rate between the ultralong and long protocols for women with various subtypes of endometriosis, nor between a 6-month suppression period and a 3-month suppression period. Furthermore, there were no significant differences between the ultralong and long protocols in terms of other reproductive outcomes (implantation rate, miscarriage rate, ectopic pregnancy rate, multiple pregnancy rate, live birth rate) and ovulation hyperstimulation outcomes (duration of ovarian stimulation, dosage of gonadotropins, number of retrieved mature oocytes, fertilization rate, number of obtained embryos, and number of transferred embryos). However, this meta-analysis did not compare the protocols’ respective effects on the cumulative pregnancy rate and cumulative live birth rate.

Another meta-analysis [31] that compared RCT studies and non-RCT studies found that in non-RCT studies, ultralong protocols typically improved implantation rates, while in RCT studies, the clinical pregnancy rate of women with stage III–IV endometriosis in ultralong protocol groups was significantly higher than that of such women in long protocol groups. However, the RCT studies found that the ultralong and long protocols resulted in no significant differences in the implantation rate, fertilization rate, oocytes quantity, and clinical pregnancy rate in women with stage I–II endometriosis. These findings differ from the findings of the meta-analysis by Shuai Liu et al. [30], i.e. there was no difference in efficacy between ultralong and long protocols in women with different subtypes of endometriosis. This suggests that the use of an ultralong protocol in women with stage III–IV endometriosis is of unclear utility, perhaps as most published RCTs or non-RCT studies have been conducted in a small number of women. Therefore, more clinical research is needed to confirm the effectiveness of ultralong protocols in women with different stages of endometriosis, and a systematic analysis of RCTs is needed to obtain reliable results. Nevertheless, it appears that compared with long protocols, ultralong protocols result in higher pregnancy rates.

This meta-analysis [31] also compared the effectiveness of ultralong and short protocols in treating endometriosis-related IVF-ET and included non-RCTs. The results showed that the fertilization rate, duration of controlled ovarian hyperstimulation (COH), and number of oocytes retrieved were significantly higher in an ultralong protocol group than in a short protocol group. However, there were no significant differences between the ultralong and short protocol groups in terms of the implantation rate, clinical pregnancy rates in women with different subtypes of endometriosis, baseline follicle-stimulating hormone (FSH) concentrations, and dosage of gonadotropins. Additionally, this meta-analysis did not compare the ultralong and short protocols in terms of their effects on cumulative pregnancy rates and cumulative live birth rates.

GnRH-a and GnRH-antagonist protocols

GnRH antagonists competitively bind to pituitary GnRH receptors and thus immediately suppress the secretion of gonadotropins and prevent early peaks in LH concentrations during ovarian stimulation [32]. In 2007, Pabuccu et al. [33] reported a comparison of the effects of GnRH antagonists and GnRH-a in COH followed by ICSI cycles in women with mild-to-moderate endometriosis and ovarian endometrioma. The women were divided into groups comprising women with confirmed stage I–II endometriosis by laparoscopic examination, women with a history of ovarian surgery and ovarian endometrioma, and women with unilateral or bilateral ovarian endometrioma without a history of ovarian surgery. They were randomly assigned to receive either a GnRH-a long protocol or a GnRH-antagonist protocol, and the resulting embryos were transferred in fresh cycles. The results showed that the GnRH-a and GnRH-antagonist protocols led to similar pregnancy outcomes in women with stage I–II endometriosis, with no statistically significant differences in implantation and clinical pregnancy rates. In women with a history of ovarian surgery for endometrioma, the GnRH-antagonist protocol resulted in a decrease in the number of MII stage oocytes, number of available embryos, and fertilization rate compared with the GnRH-a protocol. However, there was no statistically significant difference between the implantation rates and clinical pregnancy rates resulting from the two protocols. In women with no history of surgery for endometrioma, the use of a GnRH-antagonist protocol significantly reduced human chorionic gonadotropin-day estradiol concentrations, the number of follicles larger than 17 mm, the total number of retrieved oocytes, and the number of MII stage oocytes compared with the use of a GnRH-a protocol. However, there was no statistically significant difference between the implantation and clinical pregnancy rates resulting from the two protocols. In summary, in women with stage I–II endometriosis, the outcomes of IVF using GnRH-a and GnRH-antagonist protocols are similar. However, in women with endometrioma, a GnRH-a protocol can yield more MII oocytes and viable embryos than a GnRH-antagonist protocol, and thus, the cumulative pregnancy rate subsequent to a GnRH-a protocol may be higher than that subsequent to a GnRH-antagonist protocol.

A recent retrospective analysis by Kolanska et al. [34] compared the pregnancy outcomes of infertile women with endometriosis who used a GnRH-a protocol and those whose used a GnRH-antagonist protocol for COH. The results showed that subsequent to fresh embryo transfer, the GnRH-a group had higher pregnancy rates and live birth rates than the GnRH antagonist group. However, subsequent to frozen embryo transfer using the same artificial cycle protocol, there was no statistically significant difference between the groups in terms of clinical pregnancy rates and live birth rates. This finding was similar to that of Chen et al. [35] who retrospectively analyzed 639 infertile women with endometriosis, some of whom were treated with a GnRH-a protocol and others were treated with a GnRH-antagonist protocol for ovulation hyperstimulation. They found that the clinical pregnancy rate, implantation rate, and live birth rate were significantly lower in the GnRH-antagonist group than in the GnRH-a group (all p < 0.05) in fresh cycles, but that there was no significant between-group difference in pregnancy outcomes in subsequent frozen cycles with the same number of retrieved oocytes. Moreover, the cumulative clinical pregnancy rate (42.71% vs 55.2%, p = 0.083) and cumulative live birth rate (36.46% vs 47.92%, p = 0.108) were similar between the GnRH antagonist and GnRH-a protocols. These results may be attributable to the fact that GnRH-a reduce the synthesis and release of nitric oxide in the endometrium, restore the normal expression of integrin in the endometrium, and improve the receptivity of the endometrium. In contrast, GnRH antagonists alter the level of expression of HOXA10 protein in the endometrial stromal cells during ovarian stimulation treatment, thereby affecting the receptivity of the endometrium.

In 2018, Drakopoulos et al. [36] reported a retrospective cohort analysis that compared the effect of a long GnRH-a protocol and an antagonist protocol on the live birth rate of women with endometriosis undergoing IVF/ICSI treatment with fresh or frozen embryo transfer. The results showed that in women with stage I and stage II endometriosis, the use of GnRH-a was associated an increase in live birth rates and the quantity of frozen embryos, but this increase was not statistically significant. In 2020, Zhao et al. [37] reported a retrospective analysis of 342 women whose ovarian reserve function had declined after laparoscopic excision of endometriomas. The women were divided into three groups: an ultralong GnRH-a protocol group (n = 113), an GnRH-antagonist protocol group (n = 121), and a long GnRH-a protocol group (n = 108), and no statistically significant between-group differences were found in pregnancy, live birth, and abortion rates.

Progestin protocols

Progestin protocols are used to suppress pelvic inflammation and thus alleviate pelvic pain associated with endometriosis, as progestins create a low-estrogen-concentration environment that slows the growth of endometrial tissue outside the uterus. Progestins protocols are also effective for preventing early LH surges in women undergoing ovulation hyperstimulation [38, 39]. However, there has been little research on whether progestin protocols can effectively improve oocyte quality, embryo quality, and pregnancy outcomes. It also remains unknown whether progestin protocols can be used as an alternative protocol for women with severe endometriosis who wish to receive IVF/ICSI treatment. In 2020, Guo et al. [40] reported the results of an RCT of three progestin protocols in 450 women with severe endometriosis who were undergoing IVF/ICSI and had normal ovarian function. The women were divided into three groups: a medroxyprogesterone acetate + human menopausal gonadotropin (hMG) group, a dydrogesterone + hMG group, and a progesterone + hMG group. The embryos obtained were thawed and transferred. No significant between-group differences were found in fertilization and pregnancy outcomes, which suggests that these three progestin protocols achieve the same pregnancy outcomes in women with late-stage endometriosis. However, this finding applies only to women with severe endometriosis and normal ovarian function; further research is needed to determine if it can be generalized to women with decreased ovarian reserve. Additionally, the RCT did not compare the effectiveness of GnRH-a with that of GnRH antagonists.

In 2022, Guo et al. reported the results of an RCT [41] that investigated the effectiveness and safety of an MPA + hMG protocol vs a GnRH-a ultralong protocol during IVF in women with severe endometriosis and normal ovarian reserve. They enrolled 300 women with late-stage endometriosis and who were undergoing IVF and divided them into two groups: an MPA + HMG group and an ultralong protocol group. The MPA + HMG group underwent frozen embryo transfer, while the ultralong protocol group had fresh embryo transfer as the preferred option. There were no significant differences between the two groups in terms of the numbers of retrieved oocytes, mature oocytes, high-quality embryos, and viable embryos. However, there were a larger number of follicles with a diameter greater than 10 mm or 14 mm in the ultralong protocol group than in the MPA + HMG group, and the fertilization rate with ICSI was higher in the MPA + HMG group than in the ultralong protocol group. Nevertheless, both groups had similar implantation rates in terms of pregnancy outcomes, as well as similar clinical pregnancy miscarriage, multiple pregnancy, ongoing pregnancy, cumulative pregnancy, and live birth rates. Furthermore, there were no statistically significant differences between the two groups in terms of pregnancy complications, postpartum complications, and the birth defect rate. Therefore, in women with late-stage endometriosis undergoing IVF/ICSI, the administration of MPA during COH may result in a similar number of oocytes and pregnancy and live birth outcomes as the administration of an ultralong protocol. This suggests that the use of MPA in COH could be a new alternative to the standard protocol for women with endometriosis.

Effect of surgery on the outcomes of IVF in women with endometriosis

Surgical treatment is another way to treat endometriosis, the benefit of surgery on pain and quality of life is well known, but the benefit on fertility remains controversial [42]. A few studies show that surgery can improve the fertility rate in DIE patients for both spontaneous pregnancy and IVF, but there are studies have also confirmed that surgery can damage ovarian tissue, reduce ovarian reserve, and thus lower pregnancy rates [43, 44]. Current researches on the effect of surgery on endometriosis mostly relies on retrospective studies with low levels of evidence, with only a few ongoing RCT studies being identified [45]. Some studies suggest that surgery for endometriosis can be beneficial for reproductive outcomes. For example, a retrospective review by Ekine et al. in 2020 [46] examined the fertility performance of women after combined hysterolaparoscopic surgical management of endometriosis have shown that the combined hysterolaparoscopy treatment significantly improves reproductive performance and is even more effective when combined with ART, the pregnancy rate improved considerably after the surgery, and they have also found that the different stages of endometriosis do not affect fertility. Another retrospective comparative cohort study by Ferrier et al. [47] compared first-line surgery with first-line assisted reproductive techniques (ART) in infertile women with deep infiltrating endometriosis (DIE) without colorectal involvement. Their results support that in patients with DIE without colorectal involvement, the first-line surgery offer higher pregnancy rates, cumulative pregnancy rates, live birth rates, and cumulative live birth rates, with statistical differences compared to first-line ART. This difference was observed even in women aged >35, AMH <2, and with concomitant adenomyosis, they also found that in the surgery group, 17 cases of spontaneous pregnancy could be observed, while no spontaneous pregnancies were found in the ART group.

Certainly, the opposite conclusion also exists. A retrospective analysis by Maignien et al. [48] examined the impact of previous surgery for endometriosis on ART cumulative live-birth rates in DIE patients, their study suggests that in a population of DIE patients, previous surgery for any type of endometriosis may be associated with less favorable ART outcomes, patients with a history of surgical treatment for endometriosis have significantly lower clinical pregnancy rates, live birth rates, and cumulative live birth rates, all of which are statistically significant. While the cumulative live birth rates are satisfactory in bowel endometriosis patients undergoing first-line ART, with a low risk of complications. Study by Frangež et al. [49] showed that for women with endometriosis who need to undergo ART treatment, surgery may lead to impaired ovarian reserve, reducing fertility chances. However, pregnancy rates, implantation rates, fertilization rates, and live birth rates remain unaffected.

In systematic reviews and meta-analyses with higher levels of evidence, most studies demonstrate that surgical treatment for endometriosis does not improve the outcomes of IVF. A meta-analysis by Hamdan et al. in 2015 [50] showed that compared to patients with endometrioma who did not undergo surgical intervention, surgical treatment for endometrioma does not alter the results of IVF/ICSI treatment. Two groups of live birth rates, clinical pregnancy rates, miscarriage rates, mean number of oocytes retrieved, and cancellation rates of cycles were similar, however, patients undergoing surgical treatment for endometrioma had lower AFC, required higher doses of FSH. Daniilidis et al. [51] showed that there is no evidence to suggest the surgical removal of deep endometriosis prior to ART in infertile women with endometriosis to improve reproductive outcomes. Another meta-analysis by Bourdon et al. in 2023 [52] compared ongoing pregnancy rates and live birth rates in patients who underwent endometriosis surgery before ART in comparison with patients who underwent first-line ART, the study showed that no statistically significant differences in live birth rates, ongoing pregnancy rates, and early pregnancy loss rate per cycle were found when comparing patients who underwent endometriosis surgery before IVF/ICSI and those who did not. After the exclusion of the studies with high risks of bias, the live birth rates per cycle was significantly reduced in the case of surgical treatment before IVF/ICSI. Therefore, surgical management of endometriosis should not be routinely performed before ART.

Discussion

Endometriosis is a common cause of infertility, and ART is widely used to treat infertility in women with endometriosis [3]. According to current research, the use of GnRH-a and GnRH-antagonist protocols for IVF treatment of infertile women with endometriosis results in similar clinical pregnancy and live birth rates. However, multiple studies have shown that the pregnancy outcomes of ultralong protocols are better than those of other protocols, although some data did not reach statistical significance. GnRH-a protocols have been used for the longest time in the clinic because of their ability to precisely downregulate, which may help to enhance the pelvic microenvironment and endometrial receptivity. After binding to the corresponding receptors, GnRH-a causes desensitization of the pituitary receptors, leading to the inhibition of endogenous LH surge [53]. However, prolonged suppression of the pituitary gland and low concentrations of endogenous FSH may lead to a decrease in the number and size of follicles. This could potentially increase the concentrations of gonadotropins required and decrease the number of retrieved oocytes, thereby increasing the difficulty of oocyte retrieval, especially from women with diminished ovarian reserve who have undergone ovarian surgery [54]. Compared with GnRH-a, GnRH antagonists have several advantages, such as a flexible timing of initiation, a lower average dosage, a shorter duration of treatment, a lower cost, and a lower risk of causing ovarian hyperstimulation syndrome. This improves patient compliance and has resulted in the increasing use of GnRH antagonists in clinical practice [55, 56]. GnRH antagonists also can be combined with embryo cryopreservation protocols to achieve cumulative clinical pregnancy and cumulative live birth rates similar to those of ultralong protocols, thereby avoiding the longer pretreatment time required with the use of GnRH-a in ultralong protocols [35]. Therefore, in clinical practice, ovulation hyperstimulation protocols can be selected to suit the needs of the patient.

Overall, there have been few studies on ovulation hyperstimulation protocols for infertility related to endometriosis, and most have been small-sample studies and retrospective analyses. Therefore, more RCT studies are needed to comprehensively compare the efficacies of various ovulation hyperstimulation protocols.

As for surgery treatment, according to the current research, for patients with evident pain symptoms and complications such as hydronephrosis, pyelonephritis, intestinal obstruction, and pelvic abscess, surgery is recommended firstly [57]. However, for asymptomatic patients, current data is insufficient to recommend surgical intervention as a first-line treatment. The choice of treatment between IVF and surgery as a first-line treatment remains questionable, more RCTs are needed to compare the differences between surgery and IVF. Therefore, individualized treatment should be considered for endometriosis patients before IVF/ICSI.

Disclosure statement

The authors have no conflict of interest to declare.

Data availability statement

The data in this paper is available.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.