Endometriosis and oocyte quality: an analysis of 13 614 donor oocyte recipient and autologous IVF cycles

Endometriosis is a chronic gynecological condition which affects 1 in 10 women in the reproductive age group ( Macer and Taylor, 2012 ). A substantially higher prevalence of endometriosis, of up to 50%, has been reported in women with infertility ( den Hartog et al. , 2008 ; Georgiou et al. , 2019 ). ART is often considered as a treatment option for women with endometriosis and infertility ( Dunselman et al. , 2014 ; Georgiou et al. , 2019 ). There has been much debate as to whether the poorer outcome in women with endometriosis is secondary to impairment of either oocyte quality or endometrial receptivity. A systematic review including 22 studies reported lower implantation and pregnancy rates following ART in women with endometriosis compared to women who had tubal factor infertility ( Barnhart et al. , 2002 ). The authors suggested that endometriosis has a negative impact on endometrial receptivity as well as the oocyte quality. Another systematic review that included 27 studies, exploring the association between endometriosis and ART outcomes, reported lower implantation and pregnancy rates in women with severe endometriosis compared to the non-endometriosis group ( Harb et al. , 2013 ). A study involving the analysis of the Society of Assisted Reproductive Technology (SART) data reported a lower live birth rate (LBR) following autologous fresh and frozen ART, when endometriosis was associated with other causes of infertility compared to infertility causes without an additional diagnosis of endometriosis ( Senapati et al. , 2016 ). The evidence for endometriosis significantly affecting the LBR compared to other causes of fertility, is further corroborated by data from our previous study showing that the live birth after IVF cycles with fresh and frozen embryo transfers in women with endometriosis was significantly lower compared to male factor infertility (23.7% vs 25.9%; odds ratio (OR) 0.89, 95% CI 0.86 to 0.91) ( Sunkara et al. , 2021 ). However, it is unclear how endometriosis affects ART outcomes, whether it is due to the generation of poorer quality oocytes and resulting embryos or due to the endometriosis negatively impacting on endometrial receptivity ( Georgiou et al. , 2019 ; Simon et al. , 1994 ). Studies evaluating endometrial receptivity have reported altered gene expression and aberrant cell signaling pathways seen in the endometrium of women with endometriosis when compared with healthy controls ( Wei et al. , 2009 ; Lessey and Kim, 2017 ; Focarelli et al. , 2018 ). Other studies have suggested poor oocyte and embryo quality as the main reason for the lower implantation rate with endometriosis ( Simon et al. , 1994 ; Diaz et al. , 2000 ; Orazov et al. , 2019 ). A recent study evaluated treatment outcomes following transfer of frozen-thawed euploid embryos in women with and without endometriosis and reported no difference in LBR between the two groups ( Bishop et al., 2021 ). They inferred that impaired endometrial receptivity may not be the key factor that causes poorer ART outcomes for women with endometriosis. We aim to address whether the poorer outcomes with endometriosis are due to the effect on the oocyte quality or endometrial receptivity. We compared treatment outcomes following donor oocyte recipient IVF ± ICSI cycles where the recipient had endometriosis versus autologous IVF ± ICSI in women with endometriosis as a sole cause of infertility.

A total of 1 176 357 IVF cycles were performed during the period 1996–2016 as shown in Fig. 1 . For the study purpose, 1 162 743 cycles were excluded for the following reasons: cycles with more than one cause of infertility, cycles other than endometriosis, autologous IVF cycles with women's age more than 34 years, surrogacy cycles, cycles where embryos were developed for reasons other than infertility treatment, cycles with missing information on oocyte source and the lack of information on fresh or frozen treatment. A total of 13 614 donor oocyte recipient and autologous IVF cycles with endometriosis and no other cause of infertility were analyzed. There were 758 donor oocyte recipient cycles and 12 856 autologous IVF cycles for final analysis. Data selection process. Baseline characteristics for the two cohorts are described in Table I . The majority of the IVF cycles in both the cohorts were performed after 2011. For both the cohorts, the majority of cycles were fresh embryo transfers and day of embryo transfer was <5 days. The mean number of embryos transferred was 1.63 (0.7) for donor oocyte recipient and 1.54 (0.8) in autologous IVF cycles. Baseline characteristics of donor oocyte recipient and autologous ART cycles in women with endometriosis (n = 13 614). All are under 35 years of age and exact age is not available in the database. There was no significant difference in LBR in women with endometriosis undergoing donor oocyte recipient cycles compared to women undergoing autologous IVF cycles (28.0% vs 29.8%; OR 0.92, 99.5% CI 0.72 to 1.16). After adjusting for potential confounders (number of previous IVF cycles, previous live birth, period of treatment, day of embryo transfer, number of embryos transferred, fresh and frozen cycle), there was no significant difference in LBR in women undergoing donor oocyte recipient cycles versus autologous IVF cycles (adjusted OR (aOR) 1.0, 99.5% CI 0.78 to 1.28) ( Table II and Supplementary Table SI ). Live birth rate in women with endometriosis following donor oocyte recipient versus autologous ART cycles. Adjustment for confounders (number of previous IVF cycles, previous live birth, year of treatment, day of embryo transfer, number of embryo transferred, fresh and frozen cycle for combined fresh/frozen). LBR, live birth rate; OR, odds ratio. The post hoc subgroup analysis found no significant difference in LBR in women with endometriosis undergoing donor recipient cycles versus autologous cycles for the treatment period between 1996–2005 (aOR 1.08, 99.5% CI 0.66 to 1.79) and 2006–2016 (aOR 0.96, 99.5% CI 0.72 to 1.28) ( Table III ). Live birth rate in women with endometriosis following donor oocyte recipient versus autologous ART cycles (subgroup analysis for period of treatment—1996–2005 and 2006–2016). Adjustment for confounders (number of previous IVF cycles, previous live birth, day of embryo transfer, number of embryo transferred, fresh and frozen cycle for combined fresh/frozen). LBR, live birth rate; OR, odds ratio. There was no significant difference in the LBR in women with endometriosis undergoing donor oocyte recipient fresh embryo transfer cycles compared to women undergoing autologous IVF fresh embryo transfer cycles (31.6% vs 31.0%; OR 1.03, 99.5% CI 0.79 to 1.35). After adjusting for potential confounders (number of previous IVF cycles, previous live birth, period of treatment, day of embryo transfer and number of embryos transferred), there was no significant difference in LBR in women undergoing donor oocyte recipient fresh embryo transfer cycles versus autologous IVF fresh embryo transfer cycles (aOR 1.05, 99.5% CI 0.79 to 1.41) ( Table II ). There was no significant difference in the LBR in women with endometriosis undergoing donor oocyte recipient frozen embryo transfer cycles compared to women undergoing autologous IVF frozen embryo transfer cycles (19.6% vs 24.0%; OR 0.77, 99.5% CI 0.47 to 1.25). After adjusting for potential confounders, there was no significant difference in the LBR in women undergoing donor oocyte recipient frozen embryo transfer cycles compared with autologous IVF frozen embryo transfer cycles (aOR 0.85, 99.5% CI 0.51 to 1.41) ( Table II ). The LBR in women with endometriosis undergoing donor oocyte recipient cycles was lower compared to women undergoing donor oocyte recipient cycles without endometriosis (28.0% vs 30.7%; aOR 0.86, 99.5% CI 0.67 to 1.10) but the difference did not reach statistical significance. The trend toward lower LBR was observed when donor recipient fresh embryo cycles (31.6% vs 34.3%; aOR 0.86, 99.5% CI 0.65 to 1.15) and donor recipient frozen embryo transfer cycles were compared separately (19.6% vs 22.6%; aOR 0.82, 99.5% CI 0.50 to 1.35) ( Table IV and Supplementary Table SII ). Live birth rate in women with endometriosis versus women without endometriosis (male factor) following donor oocyte recipient cycles. Adjustment for confounders (number of previous IVF cycles, previous live birth, period of treatment, day of embryo transfer, number of embryo transferred, fresh and frozen cycle for combined fresh/frozen). LBR, live birth rate; OR, odds ratio.

This study involves retrospective analysis of the Human Fertilization and Embryology Authority (HFEA) dataset. The HFEA is the government regulator for all assisted conception treatment in the UK and as part of its role, data are prospectively collected and validated for all ART cycles conducted in the UK. As this was a retrospective analysis of anonymized data ( https://www.hfea.gov.uk/about-us/our-data/ ), ethics approval was not necessary. Data were extrapolated between the years 1996 and 2016 for the following two cohorts consisting of donor oocyte recipient IVF cycles, where the recipient had a diagnosis of endometriosis and a second cohort of all autologous IVF cycles with endometriosis as a sole cause of infertility. This consisted of a total of 758 IVF cycles for donor oocyte recipient and a total of 12 856 IVF cycles for autologous. The two cohorts were evaluated against each other to compare LBR per cycle. We also performed a post hoc subgroup analysis to explore any differences in the result due to the different period of treatment (1996–2005 and 2006–2016). Additionally, as a control, we included a cohort of recipient women who underwent treatment for male factor (women without endometriosis) and compared them against donor oocyte recipients with endometriosis. Live birth was defined as delivery of one or more neonates with signs of life at ≥24 weeks gestation. IVF ± ICSI for each cohort was referred as IVF in this article. Information was obtained for the donor oocyte age group (≤20, 21–25, 26–30, 31–35 years) and women undergoing autologous IVF (<35 years), period of treatment, cause of infertility, previous live birth, day of embryo transfer, number of embryos transferred, IVF or ICSI and fresh or frozen embryo transfer. A subgroup analysis was performed from the data acquired, comparing outcomes for fresh versus frozen embryo transfer cycles. Absolute and relative frequencies with 99.5% CIs were used to interpret each cohort with a primary outcome of LBR per cycle. A decision was made to use 99.5% CIs to facilitate analysis of cycle multiplicity. Distributions of the two cohorts were characterized and stratified by female age, period of treatment, previous live birth, IVF or ICSI, fresh or frozen embryo transfers, day of embryo transfer and number of embryos transferred. Data on all donor oocyte recipient IVF cycles, where the recipients were women diagnosed with endometriosis and all autologous IVF cycles, where endometriosis was the sole cause of infertility in these women, were analyzed to compare the LBR per cycle. Adjusted logistic regression analysis was performed to compare the LBR between the two cohorts, accommodating for the following confounding factors: number of previous IVF cycles, previous live birth, IVF or ICSI, period of treatment, day of embryo transfer, number of embryos transferred and fresh or frozen embryo transfer cycles. The confounders identified were already known to affect clinical outcomes or potentially were capable of this. Further subgroup analyses for fresh and frozen embryo transfer cycles were also performed. The statistical package Stata, IC version 16 (StataCorp, College Station, TX, USA) was used to analyze the data and a P -value of <0.05 was considered statistically significant.

The current study did not find any significant difference in the LBR in women with endometriosis undergoing donor oocyte recipient embryo transfer cycles compared to autologous IVF embryo transfer cycles. Furthermore, no significant difference was observed in the LBR in the subgroup of women with endometriosis undergoing donor oocyte recipient fresh embryo transfer cycles versus fresh autologous IVF fresh embryo transfer cycles. Similarly, the LBR did not differ significantly following donor oocyte recipient frozen embryo transfer cycles versus autologous IVF frozen embryo transfer cycles. Taken together, these findings suggest that there may be a limited or no effect of oocyte quality on IVF outcomes in women with endometriosis. In a study by Simon et al. (1994) , the treatment outcomes following oocyte donation according to origin of the oocyte (n = 141) was evaluated. The results of this study reported a significantly lower implantation rate in recipient women who received donated oocytes from women with endometriosis compared to other indications. The authors suggested that poor oocyte quality could have been a reason for the worse treatment outcomes following IVF in women with endometriosis, although the causative role of an endometrial factor was not excluded. However, the number of treatment cycles performed using donor oocyte from women with endometriosis was small (n = 11). Furthermore, other donor oocyte recipient characteristics as well as specific treatment related factors may also have influenced the outcomes. In an interesting matched control study by the same group, 25 women with severe endometriosis (Stages III–IV) and 33 women without endometriosis underwent oocyte donation treatment ( Diaz et al. , 2000 ). A single donor donated oocytes to recipient women from both the groups (the sibling oocyte model). The implantation rate (14.8% vs 16%) and LBR (28% vs 27.2%) did not show any significant difference between the two groups. The authors suggested that the likely reason for the poorer outcomes in women with endometriosis is due to suboptimal oocyte/embryo quality and any negative impact on endometrium was not evident in this study. In a matched case controlled study, the authors reported that the donor oocyte IVF treatment outcome was not negatively influenced by the presence of endometriosis in the recipient ( Bodri et al. , 2007 ). The findings from these studies are not in agreement with the current study results. The possible reasons could be due to small sample size and differences in the study population. While the study by Diaz et al. (2000) , included only women with severe endometriosis, the current study analyzed a nationwide dataset in which women with varying severity of endometriosis were included. Increasing severity of endometriosis has been shown to result in a lower LBR following IVF in women with endometriosis ( Muteshi et al. , 2018 ). A recent single-center retrospective cohort study compared treatment outcomes following frozen embryo transfer of euploid blastocysts in women with endometriosis (n = 39) versus two control groups, male factor (n = 253) and preimplantation testing for monogenic disease (PGT-M) (n = 36) ( Bishop et al., 2021 ). The authors found no significant difference in LBR in women with endometriosis (61.1%) compared to male factor (49.6%) and the PGT-M group (52.1%). Since the investigators controlled for embryo quality and reported no significant difference in treatment outcomes in women with endometriosis versus controls, their findings indicated that endometrial receptivity has a limited role in poorer outcomes following IVF in women with endometriosis which is in contrast to the current study finding. While the authors included only surgically confirmed endometriosis, the information relating to the severity of endometriosis was lacking. Furthermore, a smaller sample size and the potential selection bias due to the inclusion of only euploid blastocyst transfers, which are more likely to comprise of women with a better prognosis, could be some of the possible reasons for disagreement with the current study results. The current study findings suggest impaired endometrial receptivity could be an important reason for poorer IVF outcomes in women with endometriosis. While many non-clinical studies have highlighted the role of aberrant signaling pathways, altered gene expression, chronic inflammation, progesterone resistance and estrogen dominance in making the endometrium less receptive to embryo implantation in women with endometriosis, only a few clinical studies have corroborated these findings ( Bulun et al. , 2006 ; Aghajanova et al. , 2010 ; Lessey and Kim, 2017 ). A large prospective study by Prapas et al. (2012) investigated whether endometriosis affects endometrial receptivity using the donor oocyte recipient model. The study population consisted of 240 menopausal women, divided into two groups of those with a history of laparoscopically diagnosed endometriosis and those without. Oocytes derived from a single donor who was diagnosed as free from endometriosis, were randomly divided between the two groups on the day of oocyte retrieval. The LBR was significantly lower in the group of women diagnosed with endometriosis compared to the group without endometriosis (35% vs 50.83%; aOR 0.19, 95% CI 0.09–0.38). The authors attributed these findings to the negative impact of endometriosis on endometrial receptivity itself. However, the current study reported no significant difference in LBR for donor oocyte recipient women with endometriosis compared with the donor oocyte recipient women who do not have endometriosis. This is in disagreement with the study by Prapas et al. (2012) , and the possible reasons for this could be due to the differences in the study population, use of oocytes from the same donor versus different donor and data from single center versus nationwide dataset. In a retrospective cohort study, Juneau et al. (2017) investigated whether oocytes in endometriosis are more susceptible to meiotic error. The authors compared aneuploidy rates in women with endometriosis (n = 305) versus controls (n = 3798) who underwent preimplantation genetic screening cycles. The aneuploidy rates did not differ significantly between both the two groups. The authors suggested that the reason for lower success rate after IVF in women with endometriosis could be multifactorial. The study findings suggest a lesser role of oocyte quality in poor IVF outcomes in endometriosis. The strengths of the current study include that it is one of the largest studies that has analyzed a nationwide validated dataset, using a donor oocyte recipient versus autologous IVF model. The analysis exclusively included treatment cycles with endometriosis alone as an indication while those cycles with more than one cause of infertility were excluded to avoid confounding variables. To minimize the impact of age-related decline in LBR in the autologous group, we included only women aged <35 years in the autologous IVF cohort whilst in the donor oocyte recipient cohort, the recipients received oocytes from women aged ≤35 years. The results have also been adjusted for other potential known confounders as described in the Results section. We also performed a subgroup analysis to explore possible differences in the results due to the variation in laboratory practices during the two-decade-long study period. Additionally, the inclusion of data from multiple ART centers across the UK increases the generalizability of the current study findings. However, limitations of this study include a lack of data on the method of diagnosis of endometriosis and its severity as well as the presence of coexisting adenomyosis. Furthermore, there was no information as to whether the women donating oocytes were free from endometriosis. There was also a lack of information on oocyte number for the donor recipient cycle, which was an important limitation as it is considered a prognostic factor ( Boucret et al. , 2020 ). Limitations of a retrospective design also apply to current study findings. Finally, the anonymized nature of the dataset precluded reporting of cumulative LBR. In conclusion, the current study findings add further weight to the body of evidence that the reduced LBR seen in women with endometriosis may actually be predominantly due to impaired endometrial receptivity. Whilst we cannot completely rule out the negative impact of oocyte quality on IVF outcomes in women with endometriosis, it may have much smaller role than indicated by earlier studies ( Simon et al. , 1994 ; Diaz et al. , 2000 ). Furthermore, this study could not explore the impact of oocyte number on the IVF treatment outcome in women with endometriosis. The current study finding needs to be further validated by similar large dataset studies which have information on the oocyte number and cumulative LBR. The definitive knowledge about the reason for poorer IVF outcomes in endometriosis would help with optimizing treatment outcomes.

The anonymized dataset underlying this article is available from HFEA ( https://www.hfea.gov.uk/about-us/our-data/ ).

Click here for additional data file. Click here for additional data file.