全文

Abstract

Performing intrauterine insemination (IUI) in moderate-to-severe endometriosis patients is not implemented in international guidelines, as only limited data exist on treatment efficacy and safety. This retrospective study examined the efficacy and safety of two IUI treatment strategies performed between January 2007 and July 2012 in moderate-to-severe endometriosis patients. Eight (40.0%) versus seven (15.6%) ongoing pregnancies were accomplished in patients undergoing IUI with ovarian stimulation (n = 20, 61 cycles) versus IUI without ovarian stimulation in the first three cycles followed by IUI with ovarian stimulation (IUI with natural/ovarian stimulation; n = 45, 184 cycles). Preceding long-term pituitary down-regulation tended to result in a higher ongoing pregnancy rate (adjusted HR 1.8) and a higher chance of endometriosis recurrence (adjusted HR 2.3). Eight (40.0%) versus 16 (35.6%) recurrences of endometriosis complaints were reported in patients receiving IUI with ovarian stimulation versus IUI with natural/ovarian stimulation. IUI might be a valuable treatment in moderate-to-severe endometriosis patients and IUI with ovarian stimulation should be offered over IUI with natural/ovarian stimulation. Preceding long-term pituitary down-regulation might positively influence the ongoing pregnancy rate and can be considered. Whether this treatment strategy can be structurally offered prior to IVF must be investigated in a randomized controlled trial.

Performing intrauterine insemination (IUI) in moderate-to-severe endometriosis patients is not implemented in international guidelines, as only limited data exist on treatment efficacy and safety. This retrospective study examined the efficacy and safety of IUI performed between January 2007 and July 2012 in moderate-to-severe endometriosis patients (ASRM III–IV). Two treatment strategies were compared: IUI with ovarian stimulation (20 patients, 61 cycles and IUI without ovarian stimulation in the first three cycles followed by IUI with ovarian stimulation (45 patients, 184 cycles, IUI with natural/ovarian stimulation). Also, the additional effect of preceding long-term pituitary down-regulation was investigated. Eight (40.0%) and seven (15.6%) ongoing pregnancies were accomplished in patients undergoing IUI with ovarian stimulation and IUI with natural/ovarian stimulation (P = 0.05). Preceding long-term pituitary down-regulation with a gonadotrophin-releasing hormone (GnRH) agonist tended to result in a higher ongoing pregnancy rate (adjusted HR 1.8). Eight (40.0%) and 16 (35.6%) recurrences of endometriosis complaints were reported in patients undergoing IUI with ovarian stimulation and IUI with natural/ovarian stimulation. Preceding long-term pituitary down-regulation with a GnRH agonist tended to result in a higher chance of endometriosis recurrence (adjusted HR 2.3). Although IUI is not implemented in the current guidelines, IUI with ovarian stimulation could be a valuable treatment in moderate-to-severe endometriosis patients. Long-term pituitary down-regulation with a GnRH agonist prior to the first IUI treatment cycle might positively influence the ongoing pregnancy rate and can be considered. Whether this treatment strategy can be structurally offered prior to IVF must be investigated in a randomized controlled trial.

Keywords
complication; endometriosis; GnRH agonist; intrauterine insemination; ongoing pregnancy rate; recurrence

 

Introduction
Fecundity is commonly impaired in patients with endometriosis (Collins et al., 1995). Since its cause is still not unravelled, management of infertility in endometriosis patients is widely debated in literature (de Ziegler et al., 2010).

In subfertile patients with moderate-to-severe endometriosis, the anatomy of the pelvic cavity can be disturbed resulting in impaired ovum retrieval or diminished patency of the Fallopian tubes, making IVF the first choice of fertility treatment (Kennedy et al., 2005). However, in patients with surgically treated endometriosis in which normal functioning ovum retrieval and patency of at least one Fallopian tube has been established, intrauterine insemination (IUI) can be provided prior to IVF.

According to the guidelines of both the European Society of Human Reproduction and Endocrinology (Kennedy et al., 2005) and American Society for Reproductive Medicine (ASRM; Practice Committee of the American Society for Reproductive Medicine, 2012), IUI is only recommended in subfertile women with minimal-to-mild endometriosis. Recommendations with regard to performing IUI in moderate-to-severe endometriosis patients are not formulated, due to the absence of sufficient data (Macer and Taylor, 2012 and Ozkan et al., 2008). Observational, noncomparative studies showed pregnancy rates of 4–32% per cycle (Dickey et al., 1991, Dickey et al., 1992, Dodson and Haney, 1991, Göker et al., 2002, Lodhi et al., 2004, Tay et al., 2007, Vollenhoven et al., 1996 and Yovich and Matson, 1988). In addition, two randomized controlled trials investigated the effect of long-term pituitary down-regulation with a gonadotrophin-releasing hormone (GnRH) agonist prior to IUI in moderate-to-severe endometriosis patients and showed increased clinical pregnancy rates (Kim et al., 1996 and Rickes et al., 2002).

Treatment decision making, including aspects of efficacy and safety, is hard to make due to the lack of sufficient data. Safety of IUI in moderate-to-severe endometriosis patients is reported in only one retrospective study (D’Hooghe et al., 2006), showing a significantly higher risk of developing endometriosis recurrence after IUI compared with IVF treatment. This observation might be explained by a monthly exposure to ovulation and retrograde menstruation in women undergoing IUI. D’Hooghe et al. (2006) postulated that the exposure to retrograde menstruation might even be increased by ovarian hyperstimulation, negatively affecting the cumulative endometriosis recurrence rate (CERR). Comparison of IUI with and without ovarian stimulation could further clarify this, but, as far as known, it has not been investigated. Besides this, a possible favourable role of long-term pituitary down-regulation with a GnRH agonist in this regard has not been investigated.

Therefore, this study investigated the efficacy and safety of an IUI treatment strategy, comparing IUI with and without ovarian stimulation in patients with moderate-to-severe endometriosis and the effect on efficacy and safety of long-term pituitary down-regulation with a GnRH agonist prior to IUI.

Materials and methods
This study retrospectively analysed patients with surgically confirmed moderate-to-severe endometriosis (ASRM stages III and IV) with at least one patent Fallopian tube receiving IUI treatment. Patients were selected from the electronic patient database of the IVF centre of the VU University Medical Centre, Amsterdam, The Netherlands. Only patients undergoing their first IUI treatment between January 2007 and July 2012 were selected. Up to a maximum of six subsequent IUI treatment cycles were included in the analysis. IUI treatment cycles with donor spermatozoa were excluded. The database was validated and completed by two researchers (AS, LH). This study was formally exempted from ethical approval granted by the Institutional Review Board (reference no. 2013/1).

This fertility clinic is reluctant to perform IUI directly combined with ovarian stimulation due to the fear that ovarian hyperstimulation may lead to an increase in or recurrence of endometriosis complaints and also to prevent multiple pregnancies. Therefore, IUI treatments are usually performed as follows: IUI without ovarian stimulation in the first three cycles followed by IUI with ovarian stimulation (IUI with natural/ovarian stimulation). Data are presented for the total study population and the two treatment strategies (IUI with ovarian stimulation versus IUI with natural/ovarian stimulation). The additional effect of long-term pituitary down-regulation with a GnRH agonist is evaluated for the total population as well as for both groups.

Long-term pituitary down-regulation with a GnRH agonist (Leuprolide 3.75 mg depot injections, Lucrin; Abbott, USA) was administered prior to IUI treatment in a subgroup of patients. Ovarian stimulation was accomplished by subcutaneous administration of highly purified human menopausal gonadotrophin (Menopur; Ferring, Denmark) or recombinant FSH (Gonal-F; Merck Serono, Germany; or Puregon; MSD, USA) from cycle day 3, starting with 75 IU in the first IUI treatment cycle. In case of monofollicularity, the dosage was increased in the next treatment cycle by 37.5 IU per cycle. Follicle growth was measured by transvaginal ultrasound. When a follicle reached the size of ⩾17 mm, 10,000 IU human chorionic gonadotrophin (Pregnyl; Organon, the Netherlands) was administered subcutaneously and 42 h later, insemination was performed. According to this centre’s protocol, after three IUI treatment cycles, a clinical evaluation was performed and, if a pregnancy did not occur, a diagnostic or therapeutic laparoscopy was performed.

Semen was prepared by this centre’s protocol. In summary, after liquefaction, semen was centrifuged through a PureSperm 70% gradient (Nidacon, Mölndal, Sweden) HEPES-human tubal fluid (HTF; Lonza, Basel, Switserland) medium supplemented with human serum albumin (HSA; Albuman, Sanquin, Amsterdam, The Netherlands) at 710g for 15 min. The pellet was washed with HTF/HSA medium at 270g for 7 min. After removal of the supernatant, a maximum of 50 × 106 spermatozoa were resuspended with 0.5–5 ml HTF/HSA medium and incubated at room temperature (5% CO2). Prior to insemination, one last wash step was performed (200g for 7 min). The pellet was concentrated to a volume of 250 μl. Insemination was performed with a total motile sperm count of 0.5 × 106 up to a maximum of 50 × 106 spermatozoa.

The primary outcome measure was the ongoing pregnancy rate (OPR), which was calculated as the total number of ongoing pregnancies divided by the total number of patients included as well as a cumulative OPR calculated by life-table analysis. Secondary outcomes included total number of complications and recurrences as well as the CERR and time to pregnancy including all ongoing pregnancies after treatment (including IUI and IVF) and natural conception up to 1 year after start of IUI treatment. An ongoing pregnancy was defined as a vital intrauterine pregnancy visualized by transvaginal ultrasound at 12 weeks of gestation. Ovarian hyperstimulation syndrome, infection or hospital admittance occurring within 1 month after treatment were defined as complications. Recurrence of or increase in a patient’s complaint within 12 months after the last IUI treatment suspected to be based on endometriosis was stated as endometriosis recurrence.

Baseline measurements included female age, body mass index, type of infertility (primary/secondary), duration of infertility, duration since last therapeutic procedure and the presence of deep infiltrating endometriosis. The presence of deep infiltrating endometriosis was based on MRI findings and included rectovaginal or colorectal endometriosis or endometriosis of the urinary tract.

Statistical analysis
Statistical analysis was performed using SPSS version 20.0 (IBM SPSS, USA). Data were expressed as mean ± standard deviation (SD) or n (%). Nonparametric data were expressed as medians with range. Differences between the two groups were evaluated using the Students’ t-test (continuous data) or Mann–Whitney U-test in case of nonparametric data and chi-squared or Fishers’ exact tests for categorical data. A P-value < 0.05 was considered significant.

Life-table analysis was used to calculate the cumulative ongoing pregnancy rate and CERR. Both the number of cycles and duration in the months since first IUI treatment cycle were used as time parameters. The date of entry was the date of the first IUI treatment cycle. Up to a maximum of six subsequent cycles were included. Patients were followed up to 1 year after finishing their last IUI treatment cycle. For measuring the CERR, the endpoint of the study was 12 months after the last IUI treatment cycle, the date of endometriosis recurrence, the date of positive pregnancy test (in ongoing pregnant patients) or the date of patient’s last visit (in case of loss to follow up). To plot the time to pregnancy, the endpoint was stated at 12 months after start of IUI treatment. The log rank test was used to compare both groups in cumulative ongoing pregnancy rate and the CERR. Cox regression analysis was performed to estimate hazard ratios (HR) for treatment strategy and long-term pituitary down-regulation on ongoing pregnancy and endometriosis recurrence rates.

Results
This retrospective study selected 65 patients receiving 245 IUI treatment cycles from the electronic database who met the inclusion criteria (Figure 1). Forty-five patients (69.2%) were assigned to the combined IUI with natural/ovarian stimulation treatment strategy. In this group, 125 natural and 59 stimulated IUI treatment cycles were performed. The remaining 20 patients (30.8%) were assigned immediately to IUI with ovarian stimulation and received 61 stimulated IUI treatment cycles.

At baseline, all patient characteristics were not statistically different between the treatment groups (Table 1). For IUI with natural/ovarian stimulation versus IUI with ovarian stimulation, 31 and 42 patients were surgically treated within 6 and 12 months, respectively, before commencing IUI treatment (not significant).

Treatment outcomes, recurrences and complications are presented in Table 2. Significantly fewer treatment cycles were performed in patients assigned to IUI with ovarian stimulation (P = 0.045). In 13 (28.9%) patients assigned to IUI with natural/ovarian stimulation, completing six IUI treatment cycles was considered to be ineffective: IVF was offered in 10 (22.2%) patients and treatment was completely discontinued in three (6.7%). In six (30.0%) patients assigned to IUI with ovarian stimulation, completing six IUI treatment cycles was considered to be ineffective. In all of them, IVF was offered. Four patients discontinued IUI treatment for personal reasons (IUI with natural/ovarian stimulation: n = 2 (4.4%); IUI with ovarian stimulation: n = 2 (10.0%)).

Primary outcome
In total, 15 ongoing pregnancies (OPR 6.1% per cycle and 23.1% per patient) were accomplished after IUI treatment: seven (OPR 3.8% per cycle and 15.6% per patient) for IUI with natural/ovarian stimulation and eight (OPR 13.1% per cycle and 40.0% per patient) for IUI with ovarian stimulation, respectively (P = 0.014 per cycle).

The cumulative ongoing pregnancy rate after six cycles in the total population is shown in Figure 2A. Significantly higher cumulative ongoing pregnancy rates were found in patients assigned to IUI with ovarian stimulation (45.8%) compared with patients assigned to IUI with natural/ovarian stimulation (19.5%, P = 0.016; Figure 2B). Univariate Cox regression analysis showed a HR of 3.2 (95% CI 1.1–8.8, P = 0.027), reflecting a significantly higher chance of ongoing pregnancy in patients receiving IUI with ovarian stimulation. Multivariate Cox regression analysis showed that long-term pituitary down-regulation with a GnRH agonist could not be identified as a confounding variable in this regard. In patients pretreated with long-term pituitary down-regulation (n = 31), the cumulative ongoing pregnancy rate was 35.0% compared with 21.3% in women (n = 34) who directly started with IUI treatment ( Figure 2C). Univariate Cox regression analysis showed a HR of 2.0 (95% CI 0.7–5.5). Adjusted for treatment strategy (IUI with natural/ovarian stimulation or IUI with ovarian stimulation), the HR was 1.8 (95% CI 0.6–5.1).

Secondary outcomes
The time to pregnancy at 1 year after start of IUI treatment is presented in Figure 3. The cumulative ongoing pregnancy rate, including only ongoing pregnancies accomplished by IUI treatment, was significantly higher in the patients assigned to IUI with ovarian stimulation (41.3%) compared with IUI with natural/ovarian stimulation (16.8%, P = 0.014; Figure 3A). In patients assigned to IUI with natural/ovarian stimulation, two naturally conceived pregnancies were achieved up to 1 year after start of IUI treatment. Sixteen (35.6%) patients underwent at least one IVF attempt, resulting in four ongoing pregnancies. With IUI with ovarian stimulation, no naturally conceived pregnancies were achieved, and up to 1 year after start of IUI treatment, seven (35.0%) patients underwent at least one IVF attempt. Three ongoing pregnancies after IVF were described. The cumulative ongoing pregnancy rate 1 year after start of IUI treatment, including naturally conceived and IVF, ongoing pregnancies was significantly higher in patients assigned to IUI with ovarian stimulation (58.9%) compared to IUI with natural/ovarian stimulation (31.8%, P = 0.008; Figure 3B).

As shown in Table 2, 24 patients developed a recurrence of endometriosis complaints. Two (3.1%) patients reported recurrence of their endometriosis complaints, but did not receive an intervention. Five (7.7%) patients started on oral contraception, eight (12.3%) patients started with a GnRH agonist and discontinued IUI treatment and four (6.2%) patients underwent surgery for their endometriosis. Eleven patients had a laparoscopy after three IUI treatment cycles. Cystectomy of endometrioma or extensive adhesiolysis was performed in five (7.7%) and those were considered as recurrences of endometriosis.

The CERR was 60.4% for the total population (Figure 4A). Twelve months after start of first IUI treatment, the CERR was 36.5% for IUI with natural/ovarian stimulation versus 72.3% for IUI with ovarian stimulation (Figure 4B). Univariate Cox regression analysis showed a HR of 2.6 (95% CI 1.1–6.3, P = 0.03), reflecting a significantly higher chance of endometriosis recurrence in patients receiving IUI with ovarian stimulation versus patients receiving IUI with natural/ovarian stimulation. Adjusted for the confounder long-term pituitary down-regulation with a GnRH agonist, the HR became 2.2 (95% CI 0.9–5.3).

A significantly higher 12-month CERR was found in patients who were treated with a GnRH agonist prior to the first IUI treatment cycle (n = 31) versus patients without a GnRH pretreatment ((n = 34; 68.0% versus 26.6%; Figure 4C). Univariate Cox regression analysis showed a HR of 2.6 (95% CI 1.1–6.0, P = 0.024), reflecting a significantly higher chance of endometriosis recurrence in patients treated with long-term pituitary down/regulation with a GnRH agonist. Adjusted for the confounder treatment strategy, the HR became 2.3 (95% CI 0.98–5.3).

Two complications were reported (3.1%): one patient had to be admitted to the hospital for an infected endometriosis cyst and another patient had to be admitted because of an infection of unknown origin (Table 2).

Discussion
This study demonstrates that IUI treatment in the total population of moderate-to-severe endometriosis patients results in a cumulative ongoing pregnancy rate calculated by life-table analysis of 28% after six subsequent IUI treatment cycles with low complication and modest cumulative endometriosis recurrence rates. A significantly higher cumulative ongoing pregnancy rate calculated by life-table analysis is found in patients receiving IUI with ovarian stimulation versus IUI with natural/ovarian stimulation. Although, as far as is known, this is the first study comparing both IUI treatment strategies in moderate-to-severe endometriosis patients, this observation was not contrary to expectations, since similar results have been reported for minimal-to-mild endometriosis (Nulsen et al., 1993) and unexplained infertility (Veltman-Verhulst et al., 2012). At the same time, the CERR is significantly higher in patients receiving IUI with ovarian stimulation versus IUI with natural/ovarian stimulation. This result supports the hypothesis, stated by D’Hooghe et al. (2006), that a monthly exposure to ovulation and retrograde menstruation is the basis of an increased risk of endometriosis recurrence, which might be facilitated by ovarian hyperstimulation.

This study’s 1-year CERR after IUI with ovarian stimulation is in line with the reported 1-year CERR of 70% by D’Hooghe et al.(2006); however that study used a different definition of endometriosis recurrence (i.e. surgically confirmed or cytologically proven endometriosis). In total, nine of the current study’s patients (13.8%) were surgically treated with confirmation of recurrence of endometriosis. Since the development of endometriosis complaints (presumably based on recurrence of endometriosis) is an important factor in the possibility of continuing IUI treatment, this study used this broad definition of recurrence of endometriosis to calculate the CERR. In a future prospective study, the use of validated questionnaires would be recommended (Jones et al., 2006).

An important advantage of calculating the CERR by using life-table analysis is achieved by the opportunity to take the variable ‘time of follow up’ into account (Olive and Lee, 1986). However, one assumption in life-table analysis is that censored patients have the same risk of endometriosis recurrence as noncensored patients (Daya, 2005). A discrepancy between cumulative and absolute number can be explained by a large number of early censored patients. In this study, the high discrepancy between the CERR and absolute number of endometriosis recurrence can be explained by, in particular, early censored pregnant patients. For them, the endpoint was stated as the date of a positive pregnancy test. This conforms with the definition used by D’Hooghe et al. (2006). However, these patients presumably have a lower risk of endometriosis recurrence (Busacca et al., 2006), meaning that the CERR is an overestimation of the endometriosis recurrence risk.

This study also investigated the additional effect of long-term pituitary down-regulation with a GnRH agonist prior to the first IUI treatment cycle, which resulted in nonsignificantly higher odds of achieving an ongoing pregnancy. This observation is in line with two randomized controlled trials investigating the effect of GnRH administration prior to IUI. Rickes et al. (2002) investigated the effect of long-term pituitary down-regulation with a GnRH agonist for 6 months after surgery before commencing IUI treatment in patients with endometriosis ASRM stages II–IV. They found a significantly higher pregnancy rate in the long-term pituitary down-regulation group (89%) versus no prior GnRH treatment group (61%). Kim et al. (1996) compared standard down-regulated IUI cycles (2 weeks GnRH agonist) versus an ultralong down-regulated IUI cycle (6 weeks GnRH agonist) and found a significantly increased clinical pregnancy rate. Before long-term pituitary down-regulation with a GnRH agonist for 3–6 months prior to IUI can be routinely offered, a well-powered randomized controlled trial should be performed.

Although positive effects of long-term pituitary down-regulation with a gonadotrophin-releasing hormone (GnRH) agonist on pregnancy rate after IUI (Kim et al., 1996 and Rickes et al., 2002) and IVF (Sallam et al., 2006) have been described, data on endometriosis recurrence is limited. D’Hooghe et al. (2006) postulated that short-term administration of GnRH analogues might protect against recurrence of endometriosis. Against all odds, in the current study, long-term pituitary down-regulation with a GnRH agonist resulted in a nonsignificant 2.3-fold increase in the adjusted odds on endometriosis recurrence. It is inconceivable that GnRH agonists directly increase the risk on endometriosis recurrence, since GnRH agonists have positive effects on endometriosis complaints by lowering the oestrogen concentrations (Brown et al., 2010). Indirectly, cessation of GnRH-agonist suppression therapy can result in a quick relapse of endometriosis symptoms (Waller and Shaw, 1993), which can explain the higher amount of recurrences in this group of patients. Besides this, patients receiving long-term pituitary down-regulation with a GnRH agonist probably had more severe complaints of endometriosis prior to start of GnRH-agonist treatment.

Due to the retrospective design of this study, a selection bias might be achieved by the uncertainty about the assignment of patients to one of the two treatment groups. However, both groups had comparable baseline patient characteristics. Still, it is possible that the included patients represent a negative selected group, because patients with a strong suspicion of endometriosis without a surgical confirmation were not included. Those patients may have started IUI with or without ovarian stimulation before they are scheduled for a laparoscopic therapeutic procedure. In case of an ongoing pregnancy, those patients were missing in the database.

This study showed that patients assigned to IUI with ovarian stimulation not only conceived more but also had a shorter time to pregnancy compared with patients receiving IUI with natural/ovarian stimulation. Although the number of patients who discontinued IUI treatment and started IVF was not different between the groups, patients receiving IUI with ovarian stimulation switched to IVF earlier, which resulted in a higher cumulative ongoing pregnancy rate at 1 year after start of IUI treatment. This might be explained by this centre’s standard evaluation after three IUI with ovarian stimulation cycles of whether or not to stop treatment, continue IUI with ovarian stimulation or switch to IVF.

From these data, it is concluded that IUI might be a valuable treatment for achieving an ongoing pregnancy in this group of patients. IUI with ovarian stimulation should be offered over IUI with natural/ovarian stimulation, since this study found a significantly higher cumulative ongoing pregnancy rate in patients immediately starting ovarian stimulation. Taking into account an acceptable cumulative recurrence rate at 6 months after start of treatment, IUI with ovarian stimulation should be offered with a maximum of three cycles, providing that IUI treatment is completed within half a year after start of treatment. Long-term pituitary down-regulation with a GnRH agonist prior to the first IUI treatment cycle might positively influence the ongoing pregnancy rate and can be considered. Although conclusions can only be made with some degree of uncertainty, the results highlight the need for prospective trials with regard to the efficacy and safety of IUI in moderate-to-severe endometriosis. Since IUI is less radical and less expensive than an IVF procedure, the possibility to offer IUI with ovarian stimulation in patients with at least one patent Fallopian tube, for a maximum of three cycles prior to start of IVF treatment, should be further investigated in a randomized controlled trial, taking into account the efficacy, safety and cost-effectiveness of both treatment strategies.

Acknowledgements
The authors would like to thank Jan-Willem Lens for his assistance in using the electronic patient database of the IVF centre of the VU University Medical Centre.