Aiming to comprehensively update the clinical data regarding the quality of the oocytes in patients affected by endometriosis, we have systematically retrieved meta-analyses describing data on number of MII oocytes and fertilization rates in IVF/ICSI cycles of women affected by endometriosis. Relevant clinical studies conducted after the publication of the last meta-analysis were also individually retrieved. In addition, studies from oocyte donation experiences of patients with endometriosis will be also discussed.
Four meta-analyses provided clinical insights on the effect of endometriosis on oocyte quality (Table
1). The first, published by Barnhart and colleagues (2002), compared the IVF outcome of women with endometriosis, taking into account the different stages of the disease, with women with other causes of infertility (tubal factor, male factor, ovulatory dysfunction). Overall, the authors included data from 22 non-randomized studies for a total of
n = 2377 IVF cycles of women with endometriosis and
n = 4383 IVF cycles from non-affected women, even if the number of studies included in each sub-analysis was not reported. Their adjusted analysis reported a lower fertilization rate in endometriosis patients [odds ratio (OR), 0.81; 95% confidence interval (CI), 0.79–0.83,
p < 0.001)] supporting a deleterious impact on oocyte quality. Then, they separately compared women with minimal/mild disease and those with moderate/severe disease [
25], with women with tubal factor infertility. The fertilization rate in women with severe endometriosis was higher than that in women with tubal factor infertility (OR 1.54; 95%CI, 1.39–1.70,
p < 0.001), or women with minimal/mild endometriosis (OR 1.11; 95%CI, 1.09–1.13,
p < 0.001) [
62]. Unfortunately this meta-analysis did not distinguish between women who received previous medical and/or surgical treatments, thus potentially weakening the associations identified [
62].
Table 1
Results from meta-analysis studies providing clinical insights on the effect of endometriosis disease on oocyte quality.
| Overall Stage I-II Stage III-IV | | | Unknown Unknown Unknown | OR 0.81 (0.79 to 0.83)c
OR 0.94 (0.93 to 0.96)c
OR 1.54 (1.39 to 1.70)c
|
| Untreated Stage I-II Untreated Stage III-IV | | | 7 3 | RR 0.93 (0.87 to 0.99)c
RR 1.01 (0.93 to 1.10) |
| Untreated endometrioma | 2 | MD −3.61 (−4.44 to −2.78)c
| 2 | OR 1.06 (0.71 to 1.60)b
|
| Overall Stage I-II Stage III-IV Treated disease (surgery) Untreated disease Endometrioma | 4 2 3 3 1 2 | OR −1.22 (−2.38 to −0.06)c
OR −0.55 (−1.34 to 0.25) OR −0.83 (−1.73 to 0.08) OR −1.62 (−3.31 to 0.07) OR −0.50 (−1.59 to 0.59) OR −2.48 (−4.43 to −0.53) c
| | |
Another and more recent meta-analysis from Harb and colleagues (2013), where studies of women who had received medical or surgical treatment for endometriosis before the IVF cycles were excluded, found a 7% reduction in fertilization rate from seven studies including only minimal/mild endometriosis [relative risk (RR) 0.93, 95%CI 0.87–0.99
, p = 0.03]. However, this effect was small, mainly due to the low number of studies included (
n = 7) and the significant heterogeneity between them - as indicated by an I
2 value of 66% (
p = 0.008). In contrast, no significant reduction in fertilization rates was found for moderate/severe disease compared to controls (RR 1.01, 95%CI 0.93–1.10,
p = 0.84) even if this sub-analysis was only based on three, heterogeneous studies (I
2 value = 70%,
p = 0.03) [
63].
An independent and crucial discussion remains, whether the presence of a non-surgically treated ovarian endometrioma alone may adversely affect the oocyte quality. In this case, only one meta-analysis including only 9 studies has been published [
64]. In this review, Yang and collaborators reported that the total number of oocytes retrieved were 1.5 fewer in women with ovarian endometrioma, compared to those without [weighted mean difference (WMD) -1.5; 95%CI -2.84 to −0.15,
p = 0.03)]. Moreover, the number of MII oocytes retrieved in the ovarian endometrioma group was 3.61 fewer (WMD - 3.61; 95%CI -4.44 to −2.78,
p < 0.000001). In terms of fertilization competence, only 2 studies that compared the fertilization rates between the ovary with endometrioma and the contra-lateral healthy ovary were included in the meta-analysis, showing no difference (OR 1.06, 95% CI 0.71 to 1.60,
p = 0.77) [
64]. The main limitations of this meta-analysis are the clinical heterogeneity of the studies and the low sample size.
Another recent meta-analysis reported a decreased number of mature oocytes retrieved in women with endometriosis (treated/untreated) compared to women with other causes of infertility (OR -1.22, 95%CI -2.38 to −0.06). Again the total number of retrieved oocytes was lower (OR -1.93, 95%CI -3.67 to −0.18) and fertilization rates were not included in the analysis. It is thus difficult to determine whether this result should be related to a qualitative or rather a quantitative impairment in ovarian reserve, possibly also due to previous surgical treatments. When providing subgroup analyses restricted to different stages of endometriosis or previous treatments (See Table
1), the same authors did not find any significant effect on number of mature oocytes retrieved compared to controls except for patients with ovarian endometrioma (OR -2.48, 95% CI 4.43 to −0.53). Unfortunately, whether any treatment was received for the ovarian endometrioma was not specified [
65]. In addition, only one to three studies were included in the above-mentioned sub-analyses and the retrieval of the specific included studies included is not possible, hindering further interpretation of these observations [
65].
Other recent meta-analyses on the outcomes of ART in patients with endometriosis unfortunately do not provide data on the quality of oocytes as indicated by number of MII oocytes retrieved or fertilized [
66,
67]. For this reason, recent publications providing additional evidence were individually retrieved and are herein discussed.
Three retrospective studies have compared women with ovarian endometriosis with controls with tubal factor infertility. In two of the studies including patients who were surgically treated for moderate/severe endometriosis, a slight reduction in fertilization rate was observed compared to controls [61.6% vs 64.0%,
p = 0.03 and 64.8% vs 70.2%,
p = 0.04 in the study by Dong et al. (2013) and Singh et al. (2014), respectively] [
68,
69]. In a third, very small study by Luca et al. including patients with known ovarian endometriosis and without previous surgery, no differences in number of MII oocytes retrieved were reported compared to controls. However, sample size calculation was lacking and a lack of statistical power is very likely [
70]. Harada and colleagues compared the outcomes of IVF in the ovary following an excision of endometrioma with the contralateral ovary, demonstrating a lower number of oocytes retrieved (
p = 0.009) with comparable fertilization rates. Again, sample size calculation analysis was not reported [
71]. Results from two other small studies by Pop-Trajkovic and colleagues, reporting similar fertilization rates for minimal/mild endometriosis and controls, with higher fertilization rates in moderate/severe disease, are also difficult to interpret because the fertilization rates were unexpectedly as low as 54% for controls with tubal factor infertility, and again sample size calculation was not available [
72,
73].
More recently, two prospective case-control studies found similar results and confirmed significantly lower number of mature oocytes and lower fertilization rates in patients with endometriosis compared to patients with other causes of infertility [
24,
74]. In both studies, unfortunately, no information regarding previous surgical treatments received is provided. In particular, in the study by Shebl and colleagues, all stages of endometriosis were included and mature MII oocytes were also screened for morphological anomalies (vacuoles, refractile bodies, perivitelline space anomalies, aggregation of the smooth endoplasmic reticulum, central granulation, brownish discoloration, and ovoid shape). Fewer morphologically normal oocytes were observed in endometriosis patients compared to controls (
p < 0.001) and fertilization was significantly reduced in endometriosis patients compared to controls in conventional IVF (44.9% vs 54.4% respectively,
p < 0.03) but not in ICSI (74.9% vs 76.9% respectively,
p = 0.38). In addition, severe endometriosis was associated with significantly worse-quality oocytes than less severe stages (
p < 0.01) [
74].
Lastly, a large retrospective study based on the register from the American Society of Reproductive Technology, was also recently carried out [
1]. The study compared the outcomes of
n = 33458 cycles where endometriosis was present either alone (
n = 12335) or with other concomitant infertility-related diagnoses (
n = 21123) to cycles due to tubal factor infertility (
n = 22778), unexplained infertility (
n = 38713) and other causes of infertility (
n = 196295). Among the results, we deem relevant with respect to the present review that a significant reduction in fertilization rate compared to women with tubal factor infertility was found both in women with endometriosis only (RR 0.97, 95% CI 0.96 to 0.98,
p = 0.0001) and in women with endometriosis and concomitant diagnoses (RR 0.98, 95% CI 0.97 to 0.99,
p = 0.0001) [
1]. However, the retrospective, registry-based nature of the data, as well as the absence of information about previous endometriosis treatment(s) represent reasons for caution in the interpretation of the results.
Oocyte donation experiences
Oocyte donation programs have also been studied in the past to provide interesting clinical insights into the main causes of endometriosis related-infertility [
75]. One of the first studies that tried to gain clinical knowledge of the factors involved in the aetiology of the endometriosis-associated infertility was the study by Simon and colleagues [
76]. They demonstrated that patients with endometriosis have the same chances of implantation and pregnancy as other recipients when the oocytes came from healthy donors. In contrast, patients who received embryos derived from endometriotic ovaries showed a significantly reduced implantation rate as compared to the remaining groups (
p < 0.05) and hypothesized that this observation was related to oocyte quality.
Later, the same group published a prospective study, in which three groups were established in order to eliminate the inherent bias of a retrospective nature: group 1, contained donors and recipients without endometriosis (
n = 44); group 2, donors with endometriosis that provided oocytes to recipients without (
n = 14) and group 3 donors and recipients with endometriosis (
n = 16). The impairment of pregnancy rate per transfer observed in group 2 confirmed that the embryos derived from oocytes from women with endometriosis may have lower ability to implant due to alterations within the oocyte [
77].
The fact that endometrial receptivity is spared in oocyte donation recipients affected by endometriosis, was later confirmed by Sung et al. (1997) in a retrospective analysis on 239 consecutive oocyte recipients (patients with endometriosis
n = 55, patient without
n = 184), where no difference in pregnancy rates between the two groups was found [
78].
Furthermore, in a prospective matched case-control study by Diaz et al.
, 2000, IVF outcomes of women with or without endometriosis that received ‘siblings’ oocytes from the same “healthy” donor were evaluated in an attempt to avoid the bias of assigning oocytes of different quality to the different groups. Pregnancy, implantation, and miscarriage rates were not affected by moderate/severe endometriosis when compared with the control group [
79].
Overall, all this studies support the idea that infertility of endometriosis patients could be not related to endometrial environment but rather to a diminished oocyte quality. More recently, a small retrospective study by Katsoff et al. did not find any differences in clinical pregnancy and live birth rates among recipients receiving eggs from donors with (
n = 21) or without (
n = 133) laparoscopically diagnosed endometriosis. However, power calculation analysis was lacking and results are thus difficult to interpret [
80].