Japanese Translation

Assisted reproduction in endometriosis
(supporting documentation)

Intra-uterine insemination

A

Treatment with intra-uterine insemination (IUI) improves fertility in minimal-mild endometriosis: IUI with ovarian stimulation is effective but the role of unstimulated IUI is uncertain (Tummon et al., 1997).

Evidence Level 1b

IUI with or without controlled ovarian hyperstimulation (COH) is associated with a higher pregnancy rate than expectant management: the RR was 5.6 (95% CI: 1.8-17.4) when women with minimal-mild endometriosis were randomised to IUI + COH (127 cycles, 53 couples) compared to no treatment (184 cycles, 50 couples) (Tummon et al., 1997). Another RCT reported a non-significant RR for pregnancy of 1.6 but a significant increase in cycle fecundity (0.148 vs 0.045) comparing ovulation induction cycles without IUI with expectant management in 49 patients with minimal-mild endometriosis (Fedele et al., 1992). In another RCT, the effect of IUI + COH with gonadotrophins versus IUI alone was compared by alternating treatment cycles in 119 women without endometriosis and 57 women with the disease (Nulsen et al., 1993). Over 127 treatment cycles in the affected women, IUI +COH significantly increased the probability of pregnancy compared to IUI alone (RR 5.1, CI 1.1-22.5).

IUI is often used as treatment for ovulatory infertility which includes unexplained, male and cervical infertility and endometriosis. Most prospective randomized studies on the effectiveness of IUI combine the data of patients with different types of ovulatory infertility including surgically corrected endometriosis.

A systematic review of the treatment of ovulatory infertility with clomiphene citrate (CC) and intrauterine insemination (IUI) identified four randomized trials combining patients with surgically corrected minimal to mild endometriosis and patients with unexplained infertility. Two trials reported significantly improved cycle pregnancy rates comparing CC plus IUI versus no CC (NC) plus timed intercourse (TI), and gonadotrophins plus IUI versus CC plus IUI, respectively. Two trials reported non-significant odds ratios comparing CC plus IUI versus NC plus IUI and CC plus IUI versus CC plus TI (Costello, 2004). Thus, CC and IUI is an effective treatment option resulting in a higher clinical pregnancy rate compared to NC and TI. Treatment with gonadotrophins and IUI results in a higher clinical pregnancy rate compared to CC and IUI.

In patients with unexplained infertility including minimal/mild or surgically treated endometriosis logistic regression analysis of a meta-analysis of 13 trials showed that the likelihood of conception was significantly increased by each of the interventions clomiphene treatment, hMG treatment and IUI, independently by ~ 2 fold (The ESHRE Capri Workshop, 1996).

Another logistic regression model of 5214 IUI cycles from twenty-two randomized studies identified significant adjusted odds ratios for the likelihood of conception for the independent variables FSH-stimulation (OR 2.35) and IUI (OR 2.82). However, the presence of endometriosis reduced treatment effectiveness of IUI significantly by approximately half (OR, 0.45) (Hughes, 1997).

Reduced fecundability associated with the presence of endometriosis (disease severity unclassified) was also shown in prospective cohort studies of donor insemination (Jansen, 1986; Hammond et al., 1986). The negative effect of minimal-mild endometriosis on pregnancy rates after donor insemination was reported to persist after surgical or medical treatment (Toma et al., 1992).

Pregnancy rates following homologous insemination within 6 months of surgical treatment were as high in women with endometriosis as in the control group with unexplained infertility in a case control study (Werbrouck et al., 2006).

In conclusion, a significant improvement of pregnancy rates can be achieved by COH and IUI compared to expectant management, despite the negative impact of endometriosis on treatment effectiveness.

In general, repetitive COH-IUI cycles are characterised by a plateau effect after 3-4 cycles, so the monthly fecundity rate after several unsuccessful cycles could be even lower than that of patients undergoing expectant management (Deaton et al., 1990). Thus, counselling patients to stop treatment or to switch to other treatment options, such as IVF, is advisable after repetitive treatment failures.

In vitro fertilisation

B

In vitro fertilisation (IVF) is appropriate treatment especially if tubal function is compromised, if there is also male factor infertility, and/or other treatments have failed.

Evidence Level 2b

Impaired tubal function and disturbed interaction between the fallopian tubes and the ovary may result in reduced fimbrial efficiency to pick up eggs from the ovarian surface and in impaired tubal transport of eggs, sperm and embryos. IVF treatment appears therefore to be appropriate particularly in patients with advanced disease which is frequently associated with adhesions, ovarian endometriomas and tubal obstruction.

A

IVF pregnancy rates are lower in patients with endometriosis than in those with tubal infertility (Barnhart et al., 2002).

Evidence Level 1a

The recommendation above is based on a systematic review but the working group noted that endometriosis does not adversely affect pregnancy rates in some large databases (e.g SART and HFEA) (Templeton et al., 1996).

A systematic review indicated that pregnancy rates are lower in women undergoing IVF treatment with endometriosis than in women with tubal infertility (Barnhart et al., 2002). The review included 22 studies, consisting of 2,377 cycles in women with endometriosis and 4,383 in women without the disease. After adjusting for confounding variables, there was a 35% reduction in the chance of achieving pregnancy with IVF in women with endometriosis (OR, 0.63; CI, 0.51-0.77). Other outcome parameters, e.g. fertilization rate, implantation rate, mean number of oocytes retrieved and peak oestradiol concentration were also significantly lower in women with endometriosis compared to those with tubal factor infertility. The data therefore suggest that the presence of endometriosis affects multiple factors determining reproductive success during IVF.

It has to be noted that endometriosis does not adversely affect pregnancy rates in some large databases (e.g. SART and HFEA) (Templeton et al., 1996). However, since the inclusion of confounding factors in the systematic review strengthened the negative association between endometriosis and IVF outcome, this might explain the findings in the large databases which are not controlled for confounders. In an observational study on IVF/ICSI outcome, women with minimal to mild endometriosis had a better live-birth rate than women with moderate to severe endometriosis or women with tubal factor infertility (Kuivassari et al., 2005).

Only one prospective randomized study is available to address the impact of assisted embryo hatching on ART outcome in women with endometriosis. No significant effects on pregnancy rate and implantation rate could be detected. However, the study comprising 90 cycles randomized in a 2:1 ratio favouring the assisted hatching study group was underpowered to detect clinically meaningful differences (Nadir et al., 2005).
.
To investigate possible mechanisms for reduced pregnancy rates in endometriosis patients, data from donor egg IVF-programs can be analyzed to dissect out the influence on embryo implantation and pregnancy of ovarian/oocyte and uterine factors, depending on whether endometriosis is present in the donor or recipient. Only retrospective studies are available addressing this issue. Similar implantation and pregnancy rates have been reported in two cohorts of recipients with stage III-IV endometriosis (n=25) and recipients without endometriosis (n=33), who received randomly distributed donor eggs from healthy donors. In contrast, another small retrospective study demonstrated that oocytes donated by women with stage III-IV endometriosis (n=14) to women with a normal pelvis gave rise to embryos with reduced quality, and reduced implantation rates per embryo were also observed (Simon et al., 1994; Garrido et al., 2002). A study analyzing a cohort of 170 oocyte donors reported no significant effects but a trend for reduced pregnancy rates in recipient cycles if the donor had endometriosis and a trend for reduced implantation rates in recipients with endometriosis, suggesting a potential mild effect of endometriosis on both the uterine environment and the quality of the oocyte (Katsoff et al., 2006). In conclusion more studies are needed to provide adequate power to address the question whether endometriosis-associated subfertility is related to reduced oocyte quality or to reduced endometrial receptivity.

A

Treatment with a GnRH agonist for 3-6 months before IVF or ICSI should be considered in women with endometriosis as it increases the odds of clinical pregnancy fourfold. However the authors of the Cochrane review stressed that the recommendation is based on only one properly randomised study and called for further research, particularly on the mechanism of action (Sallam et al., 2006).

Evidence Level 1b

In a prospective randomized study, a significantly higher cumulative pregnancy rate over up to 3 cycles of IVF/ICSI therapy was achieved in patients with stage III-IV endometriosis after ultralong (5-6 months) GnRH-agonist treatment (82%) compared to patients with ovarian stimulation using a long protocol starting the agonist on day 18 of the previous cycle (40%) (Rickes et al., 2002).

A similar, prospective, randomized study showed that 25 patients who received a GnRH agonist for 3 months prior to IVF had a significantly higher ongoing pregnancy rate per cycle initiated compared to 26 patients undergoing the regular mid-luteal start long protocol. The number of oocytes retrieved, fertilisation rates, and implantation rates were not significantly different between the groups. However, randomization was not successful in achieving similar baseline characteristics between the groups: in the ultralong protocol group a significantly higher portion of patients were classified as moderate to severe endometriosis, compared to the long protocol group (Surrey et al., 2002).

B

Risk for recurrence is no reason to withhold IVF therapy after surgery for endometriosis stage III or IV since cumulative endometriosis recurrence rates are not increased after ovarian hyperstimulation for IVF (D´Hooghe et al., 2006).

Evidence Level 2a

Since endometriosis is an oestrogen-dependant disease, there is concern about the negative impact of supra-physiological oestradiol levels during COH. A retrospective cohort study including 67 infertility patients after surgery for endometriosis stage III or IV showed a significantly lower cumulative endometriosis recurrence rate (CERR) in patients treated with IVF only compared to patients treated with IUI. Moreover, CERR before and after assisted reproductive technology were similar, suggesting that cumulative exposure to high levels of oestradiol during ovarian hyperstimulation is not a risk factor for endometriosis recurrence (D'Hooghe et al., 2006). However, rare case reports have described increased growth and recurrence of endometriotic lesions during COH and the onset of severe symptoms coincided with high levels of plasma oestradiol (Renier et al., 1995; Govaerts et al., 1998; Anaf et al., 2000).

A

Laparoscopic ovarian cystectomy in patients with unilateral endometriomas between 3 and 6 cm in diameter before IVF/ICSI can decrease ovarian response without improving cycle outcome (Demirol et al., 2006).

Evidence Level 1b

There is one RCT (Demirol et al., 2006) and two systematic reviews of retrospective studies (Gupta et al., 2006; Somigliana et al., 2006a) on the impact of ovarian endometriomas and their removal on the outcome of assisted reproduction. In the prospective randomized trial 49 patients underwent conservative ovarian surgery before the ICSI cycle and 50 patients underwent the ICSI cycle directly. Ovarian stimulation parameters for those who underwent ovarian endometrioma cystectomy were significantly reduced and fewer mature oocytes were retrieved in the cystectomy group. No difference in implantation and clinical pregnancy rates were detected (Demirol et al., 2006). Studies evaluating the response to ovarian stimulation in patients previously operated for endometriomas have led to controversial results in terms of ovarian response and cycle outcome. In patients with unilateral disease a significantly reduced number of follicles in the operated ovary compared to the intact side were reported in several but not all studies. The authors of one systematic review conclude that overall evidence suggests that surgery does not benefit asymptomatic women preparing to undergo IVF-ICSI who are found to have an endometrioma (Somigliana et al., 2006a). The other metaanalysis indicates that ovarian endometrioma have adverse effects on follicle number and oocytes retrieved but not on embryo quality or pregnancy outcomes. Surgery may decrease the number of retrieved oocytes, but the overall fertility outcome is not affected (Gupta et al., 2006). Most publications do not report the size of endometriomas or only endometriomas > 3 cm were considered for data analysis.

The observation of an impaired ovarian response in women with endometriomas does not clarify whether the damage is consequent to surgery or antecedent to the intervention. An observational study in women with unilateral endometriomas who did not undergo previous ovarian surgery showed a significant mean reduction in follicles in the affected ovaries, suggesting that the presence of ovarian endometriomas is associated with a reduced responsiveness to gonadotrophins (Somigliana et al., 2006b).

GPP

Laparoscopic ovarian cystectomy is recommended if an ovarian endometrioma ≥ 4 cm in diameter is present to confirm the diagnosis histologically; reduce the risk of infection; improve access to follicles and possibly improve ovarian response. The woman should be counselled regarding the risks of reduced ovarian function after surgery and the loss of the ovary. The decision should be reconsidered if she has had previous ovarian surgery.

In summary, since no indication exists that even in experienced hands laparoscopic surgery of endometriomas improves ovarian function or enhances IVF outcome substantially, practical considerations have to be taken into account and patients have to be counselled on an individual basis. Surgical removal of endometriomas gives histological information and can rule out malignancy, can relieve discomfort and pain, may reduce the risk of rupture and adnexal torsion and may facilitate transvaginal access to ovarian follicles. On the other hand, the possibility of ovarian failure due to destruction of normal ovarian tissue during surgery, especially in patients who have already had repetitive surgery for endometriomas, has to be considered.

Concise