In this observational study of 836 infertile patients (age <40 y) with stage Ⅲ–Ⅳ endometriosis undergoing ART, there was an overall pregnancy rate of 65.67% and 60.65% delivered a live birth. cLBR per cycle and per ET cycle reached 48.38% and 43.71% after four ART cycles, respectively. This study showed that ovarian responsiveness was significantly reduced in women with recurrent endometriomas undergoing an IVF/ICSI cycle after surgery. Indeed, the AFC, the estradiol level on hCG-day, and the number of oocytes retrieved were significantly lower, compared to the primary group. However, there were similar outcomes in the cumulative cPR/LBR between the two groups. Lastly, according to multivariate logistic regression analysis, an independent significant relationship with poor prognosis on cLBR in patients with stage Ⅲ and Ⅳ ovarian endometriosis for age (≥35), time interval between surgery and ART (≥2 years), number of surgeries (≥2), and AFC (<10) was found.
Surgical treatment of endometriosis could create a more favorable environment for successful conception . On the other hand, surgical intervention for endometrioma may increase the risk of infertility by reducing the ovarian reserve [12, 13]. Previous reports on the impact of endometriosis severity on ART outcome have drawn conﬂicting conclusions. In previous studies, the IVF outcomes in patients with minimal/mild endometriosis were similar to those in patients for whom IVF was performed for other indications, while the outcomes were inferior in infertile patients with severe endometriosis. In cases with stage Ⅲ/Ⅳ endometriosis, fewer oocytes were retrieved, and lower implantation rates and cPR were reported [18, 19]. On the contrary, one study, which reviewed 3930 endometriosis cases, showed no difference in pregnancy outcome according to disease stage . In the current study, cumulative cPR per cycle and per ET cycle reached 54.20% and 48.97% after four ART cycles. Maignien et al  recently conducted a retrospective observational cohort study showing that cumulative pregnancy rates reached 65.8% in a series of 359 endometriosis patients after four ART cycles. The higher cumulative PR than found in our study may be owing to an indistinctive description of disease stage. Other published studies show that the cumulative cPR can be from 38.6% to 43% in the rAFS stage Ⅲ and Ⅳ endometriosis patient, although these sample sizes were relatively small [25, 26]. Such discrepancies could result from uncertainty about the exact endometriosis phenotype. An additional explanation may arise from exacerbating surgical approaches among patients with advanced pathology, which may impair ovarian reserve in cases of ovarian involvement.
Ovarian endometriosis is the most common type, and the recurrence rate following surgical intervention remains high. Since endometrioma is a pseudocyst, the risk of removing normal tissue during surgery is high. Therefore, there are concerns about reducing fertility and IVF outcome. Findings show that endometriomas, especially those that are recurrent endometriomas after surgical treatment, have a negative effect on the ovarian reserve [27, 28]. Likewise, in this study, it was observed that women with recurrent endometrioma after previous surgery for endometrioma had a lower AFC, estradiol level on hCG-day, and a lower number of total oocytes collected than the primary group. However, other studies record that cumulative cPR and LBR per started cycle in recurrent and primary endometrioma are similar [29, 30]. The current results also provided support for the conclusion that no significant differences were observed between recurrent and primary endometrioma in terms of cLBR and cPR in women with rAFS stage Ⅲ and Ⅳ endometriosis. The reasons possibly due to the young age (<40 y) in the recurrent group, and these patients have a relatively good ovarian reserve and small endometrioma cyst (<4 cm) in ART procedures. However, the current results found that advanced maternal age (≥35 y) had a higher risk of poor IVF outcome. This corresponded with results of retrospective-analyses . The outcomes are likely based on the decline of both ovarian reserve and oocyte competence with advancing age. A study estimated that in women aged 35–37, 38–40, 41–42, and >42 y it was necessary to collect ∼5, 7, 10, and 20 oocytes, respectively, to identify at least one euploid embryo . Therefore, women with severe endometriosis associated infertility should achieve pregnancy as soon as possible, and in patients >35 y who fail to get pregnant, IVF should be the treatment of choice.
Multivariate analysis indicated that a second surgery was a negative risk for cLBR. Clinicians are often faced with the decision of whether to undertake a second surgical procedure or to treat recurrent ovarian endometrioma with medical therapy. Current guidelines on the management of recurrent endometriosis suggests that clinicians should avoid repeated surgery in women who want to conceive when endometriosis has recurred after a first surgery . However, the effect of secondary surgery on ART outcomes is still debated, with one study  reporting a similar PR after primary and secondary surgery while another study  reports poorer results after surgery for recurrence. In the present study, the lower cLBR was suggested to be linked with the second surgery. The reason for this result may be owing to the lower ovarian reserve in the recurrent endometrioma patients. The excised cyst wall in the specimens from patients with recurrent endometriomas were signiﬁcantly thicker than in the specimens from patients undergoing surgery for the ﬁrst time . Furthermore, ovarian tissue was more abundant in the cyst wall of recurrent endometriomas than in the cyst wall after primary surgery . As a result, AFC and ovarian volumes for the operated ovaries were signiﬁcantly decreased in the recurrent endometrioma group. Therefore, if possible, clinicians should avoid a second surgery for recurrent endometriosis in women who plan to have further pregnancies. Clinicians may try postoperative medical treatment for recurrent endometriosis between IVF cycles.
In the current study, a statistically significant correlation was observed with the presence of adenomyosis, and this was associated with lower cLBR according to univariate analysis, but this did not remain a negative prognostic factor of IVF outcome after multivariate analysis. This agrees with some previous studies that show no impact of the disease on pregnancy rates [24, 35, 36]. Yet, data regarding the effect of adenomyosis on ART outcomes are still inconsistent. For instance, in a meta-analysis, Vercellini et al  show decreased clinical PR in patients with adenomyosis, as compared to controls. The negative results are probably owing to the presence of DIE, which is suggested to be a major negative predictive factor of ART results, decreasing cLBR from 51.9 to 19% . The current study showed that the concurrent rate of adenomyosis was not a poor risk for cumulative LBR after excluding the factor of DIE. Therefore, correct identification of coexisting pathological conditions for DIE and adenomyosis is necessary for the development of effective IVF protocols.
GnRH-a is commonly administered for at least six months post-surgery to prevent the recurrence of endometriosis. In the current study, using the GnRH-a treatment after surgery was not associated with a poorer outcome of cLBR in infertile patients. These results are also supported by recent works, which indicate that postoperative ovarian suppression by GnRH-a is not helpful in enhancing PR . In a recent prospective trial, women with mild endometriosis who received GnRH-a for three months before IVF improved their fertilization rate but not cPR . The current data indicated that the GnRH-a ultra-long protocol was more frequently used in recurrent patients with severe endometriosis who achieved a similar cLBR as the primary group, which suggests that the GnRH-a ultra-long protocol can improve the cPR and cLBR of patients with recurrent endometrioma. These results may be explained by GnRH-a reducing inflammation, blood flow, and adhesions in endometriosis . An additional explanation may arise from the lower levels of inflammatory cytokines that can reduce the toxic effects of cytokines on oocytes or embryos . Despite these favorable results, some data show that long-term administration of GnRH-a could suppress the expression of implantation factors, which could lead to decreased endometrial receptivity . In addition, long-term use of GnRH-a could induce side effects such as hot flashes, vaginal dryness, and decreased bone mineral density. Therefore, the potential benefits of GnRH-a pretreatment must be weighed against the ovarian reserve, additional costs, delays in the initiation of IVF, and the possibility of decreased response to ovarian stimulation.
Previous analyses demonstrate that there may be no effect of the interval from surgical management of endometriosis and IVF on PR throughout a 5-y evaluation period [44, 45]. However, an inverse result was observed in the current study; that is, the interval between surgical management of endometriosis and IVF had a signiﬁcant effect on ART outcome for patients with advanced-stage endometriosis. We found that the cLBR was significantly higher when IVF was performed <2 y after surgery for endometriosis. Similarly, Nesbitt-Hawesetal et al  report a 13-month median time of surgery and IVF among patients with stage Ⅲ to Ⅳ endometriosis who conceived by ART. Other studies indicate that the highest ongoing PR can be achieved in patients undergoing their IVF cycle 6 to 25 months after their endometriosis surgery . The reduced LBR after 2 y may be explained by either age factors, endometriosis recurrence, and/or ovarian reserve. Therefore, if IVF is planned after surgery for endometriosis, IVF delay may be considered to around 6 months but at no longer than 2 y.
The strength of this study was in the methodological design. First, the large number of patients with stage Ⅲ and Ⅳ ovarian endometriosis enrolled (836 women undergoing 1048 ART cycles), which will have increased the statistical power of the study. Second, although previous series exploring the relationship between endometriosis and ART exist in the literature, only a few focused on endometrioma. Third, only patients with a diagnosis of endometrioma based on histological conﬁrmation after surgery were included. Nevertheless, the current study still had some limitations. The major limitation was its retrospective nature based on a single center, which may therefore contain bias regarding patient characteristics. Second, AMH was examined over 3 recent years in our hospital, therefore these results were not included in this study. The third limitation arose from the recurrent endometriomas being defined by ultrasound or MRI as the presence of a persistent ovarian cyst. This practice depended heavily on the skill and experience of radiologists. Moreover, the study population was represented by women <40 y, non-obese, with cleavage stage embryo transfer at day 2 or 3, so the data from this study can only be extrapolated to patients with a similar profile. Consequently, these limitations may have resulted in an under or over estimation of the associations in the results.
In conclusion, the current study suggests that postoperative recurrent endometriosis has no impact on ART outcome, whereas it did reduce the size of the ovarian reserve and the number of oocytes retrieved. Lower AFC (<10), age (≥35), number of surgeries (≥2), and time interval between IVF and surgery (≥2 y) were associated with a lower cLBR for women with rAFS III and IV endometriosis. These results might be useful in daily clinical practice to inform and counsel couples with rAFS Ⅲ and Ⅳ endometriosis before undertaking ART. No more than a 2-y interval between IVF and surgery should be recommended for these patients with advanced stage. These findings could also facilitate the identification of patients with poor chances of success with IVF-ICSI, thus avoiding unnecessary treatments and allowing the guidance of couples regarding alternative approaches. Further multi-center prospective studies are needed to confirm the results of this study and to support the management of infertile women with endometriosis.