Pregnancy Rates after Intrauterine Insemination in Moderate to Severe Endometriosis: A Systematic Review and Meta-analysis of Observational Studies

in higher pregnancy rates compared to IUI alone (8). In moderate to severe endometriosis-related subfertility (ASRM stages 3-4), firm recommendations on performing IUI are lacking in both the European and American guidelines (6, 9). Therefore, IUI as an MAR treatment option is not structurally offered prior to IVF in this subgroup of patients. However, in our experience, IUI (with controlled ovarian hyperstimulation) might be a valuable treatment option in moderate to severe endometriosis patients with normal tub- al-ovarian function. This is supported by recent observational data showing cumulative ongoing pregnancy rates of 41.3% after 6 treatment cycles (10). In addition, it is also suggested that long-term pituitary desensitization with a gonadotropin- releasing hormone (GnRH) agonist prior to the start of IUI should be considered (10), since this may improve pregnancy rates (11, 12). DOI: 10.5301/jeppd.5000299 Pregnancy rates after intrauterine insemination in moderate to severe endometriosis: a systematic review and meta-analysis of observational studies Lisette E.E. van der Houwen1, Anneke M.F. Schreurs1, Roel Schats1, Pam Kaspers2, Cornelis B. Lambalk1, Peter G.A. Hompes1, Velja Mijatovic1 1 Department of Obstetrics and Gynecology, Division of Reproductive Medicine, Academic Endometriosis Center VUMC, VU University Medical Center, Amsterdam - The Netherlands 2 Medical Library, Vrije Universiteit Amsterdam, Amsterdam - The Netherlands

Endometriosis is a benign, estrogen-dependent gyneco- logical disease in which fecundity is commonly impaired (1). Treatment of endometriosis-related subfertility is widely de- bated in literature (2-5). The evidence to perform surgery and/or use medically assisted reproduction (MAR) tech- niques, including intrauterine insemination (IUI) and in vitro fertilization (IVF), was updated in 2014 (6). In women with moderate to severe endometriosis, sur - gery or IVF might be advised, especially when tubal function ABStRAC t To evaluate the efficacy and safety of intrauterine insemination (IUI) in moderate to severe endometrio- sis patients, a systematic review and meta-analysis was conducted since the role of this treatment strategy in these patients is a matter of debate in the literature. Systematic searches were performed in PubMed, EMBASE, Cinahl, and The Cochrane Library from inception to September 1, 2016. Studies including moderate to severe endometriosis patients reporting pregnancy rates after IUI were selected. The primary outcome was live birth after IUI treatment compared to expectant management. Secondary noncomparative outcomes were live birth and clinical pregnancy, which were presented as weighed mean pregnancy rates. Nineteen articles (2 unclear design, 11 retrospective, 6 prospective) were included for the analysis. Our primary out - come measure was only addressed by one study, showing an odds ratio of 1.77 (95% confidence interval [CI], 0.86-3.63) on live birth favoring IUI versus no treatment. The calculated weighed mean live birth and clinical pregnancy rate per patient was 20.3% (95% CI, 11.2-29.4) and 32.7% (95% CI, 21.3- 44.0), respec- tively. This meta-analysis of observational data showed that IUI could be a feasible treatment in moderate to severe endometriosis. Whether this treatment should be structurally offered prior to in vitro fertilization needs to be investigated in a randomized, controlled trial, including time-to-pregnancy, safety, and cost- effectiveness.

Endometriosis, Intrauterine insemination, Meta-analysis, Pregnancy, Recurrence Accepted: August 3, 2017 Published online: August 25, 2017 Corresponding author: Lisette E.E. van der Houwen Department of Reproductive Medicine VU University Medical Center P .O. Box 7057 1007 MB Amsterdam, The Netherlands [email protected] van der Houwen et al 159 © 2017 Wichtig Publishing In view of the current lack of guidance, we aimed to per - form a systematic review and meta-analysis of observational data, investigating the efficacy and safety of IUI (with ovarian stimulation) and the additional effect of preceded long-term pituitary desensitization with a GnRH agonist in patients with moderate to severe endometriosis to endorse future ran- domized trials.

Search strategy Both MOOSE (13) and PRISMA (14) guidelines were used to conduct this meta-analysis. To identify all relevant publica- tions, we performed systematic searches in the bibliographic databases PubMed, EMBASE.com, Cinahl, and The Cochrane Library (via Wiley) from inception to September 1, 2016. The search terms included controlled terms (e.g., MeSH in PubMed, EMtree in EMBASE.com, Mesh Headings in Cinahl) and free-text terms. We used free-text terms only in The Cochrane Library. Search terms expressing “endometriosis” were used in combination with search terms comprising “intra-uterine insemination (IUI) treatment.” No limitations were used. The Medline search strategy is presented as sup - plementary material (available online at www.j-endometrio- sis.com); the search strategies for the other databases were based on this strategy. The references of the identified arti- cles were also searched for relevant publications. Selection process L.H. and A.S. independently selected articles by title and abstract, which were then coded, but did not show author, year, or journal, to limit selection bias by investigators. To be included, peer reviewed articles, including randomized, controlled trials (RCT) and observational studies with and without a control group of patients receiving no treatment, needed to report on absolute pregnancy numbers per patient and/or per IUI treatment cycle in specifically moderate to se- vere endometriosis patients. To prevent a limited selection of papers, all IUI treatment strategies (i.e., with ovarian hy - perstimulation with clomiphene citrate, recombinant follicle- stimulating hormone [FSH] or human menopausal gonadotro- phins, and unstimulated cycles) were included. Only articles reporting primary data were selected. Studies were excluded if they (i) presented nonprimary data; (ii) were in languages other than English, Dutch, German, French, or Spanish; and (iii) included IUI treatment cycles performed with donor se - men. Through discussion, consensus was formed regarding doubtful articles. Data assessment The full texts of the selected articles were obtained for further review. Two reviewers (L.H. and A.S.) independent - ly evaluated the papers using a standardized record form, checking whether all the necessary data were provided in the article. Disagreement between the reviewers on indi- vidual items were identified and solved during a consensus meeting. Data extraction Details about the following elements were extracted and tabulated independently by two investigators (L.H. and A.S.) from the publications: study design, treatment outcomes, number of patients with moderate to severe endometriosis, number of treatment cycles, IUI treatment characteristics (i.e., medication used for controlled ovarian hyperstimula - tion, use of preceding long-term pituitary desensitization and its duration), and treatment outcomes (definition of pregnan- cy, number of pregnancies, pregnancy rate per cycle and per patient). No unpublished data were used. Outcomes The primary outcome concerned the efficacy of IUI treat- ment defined as live birth rate after IUI compared to expectant management. Secondary outcomes were biochemical; clinical; ongoing pregnancy per patient and per cycle; the efficacy of preceded treatment with a GnRH agonist; and recurrence of endometriosis. Statistical analysis To calculate the odds ratio (OR) of live birth in RCTs and observational (comparative) studies, forest plots were con- ducted with a 95% confidence interval (95% CI) by using Re - view Manager (version 5, The Cochrane Library). In case of observational, noncomparative studies, statistical analysis was performed to calculate and plot the weighed mean pregnancy rate with a 95% CI of the included studies by using Microsoft Excel (15). Plots of weighed mean clinical pregnancy and live birth were composed. Heterogeneity was tested using the Q test and I² test (16).

Search results Figure 1 presents the selection process of the literature search. The search generated a total of 913 references: in PubMed, in EMBASE.com, in Cinahl, and in The Cochrane Li - brary. After removing duplicates of references that were se- lected from more than one database, 646 titles and abstracts were examined. Of these titles and abstracts, 259 full texts were selected for further investigation. After reading the full texts, 18 articles were selected. An additional 2 articles were added to the selection after screening references. One article (17) was excluded due to overlapping data resulting in a final selection of 19 articles. Overview of studies Characteristics of the retrieved articles are presented in Table I. No RCTs were found comparing IUI treatment with expectant management in moderate to severe endometrio- sis patients. One prospective observational study published by Keresztúri et al (18) compared IUI treatment with expect - ant management in surgically treated endometriosis pa- tients. From all other included studies, noncomparative data Efficacy of IUI in moderate to severe endometriosis 160 © 2017 Wichtig Publishing Fig. 1 - Flowchart. concerning IUI treatment in moderate to severe endome - triosis patients were extracted. One RCT compared an ultra- long GnRH agonist protocol with 2 weeks of GnRH agonist treatment prior to ovarian hyperstimulation with IUI (11). In 5 studies, inclusion was limited from moderate to severe endometriosis patients (10, 19-22). In all other studies, mod- erate to severe endometriosis patients were a subgroup of included patients (18, 23-34). Primary outcome Keresztúri et al (18) performed the only prospective nonran- domized study that was included in our meta-analysis, which in- vestigated the benefit of surgery followed by IUI with controlled ovarian hyperstimulation over surgery followed by no treat - ment during a follow-up period of 12 months. No significant difference was shown between the groups with an OR of 1.77 (95% CI, 0.86-3.63) (Fig. 2). In this study, no long-term pituitary desensitization with a GnRH agonist was performed. Secondary outcomes Treatment outcomes are presented in Table II. Twelve studies reported on clinical pregnancies (11, 18, 19, 21, 23, 25, 26, 28-31, 33). The calculated weighed mean clinical pregnancy rate was 13.4% per treatment cycle and 32.7% per patient with a 95% CI of 7.4-19.4 and 21.3-44.0, respectively (Fig. 3). The test for heterogeneity showed a Q of 11.5 and 7.8 per cycle and per patient, respectively, and an I 2 of 39.0% and 0% on pregnancy rate per cycle and per patient, respectively, using the random effects model for both values. Nine studies reported on live births (10, 11, 18, 19, 23, 25, 26, 28, 29). The calculated weighed mean live birth rate per cycle and per patient was 5.6% (95% CI, 3.0-8.2) and 20.3% (95% CI, 11.2-29.4), respectively (Fig. 4). The test for hetero- geneity showed a Q of 8.1 and 7.4 per cycle and per patient, respectively, and an I 2 of 25.9% and 5.4% on pregnancy rate per cycle and per patient, respectively, using the random ef - fects model for both values. Seven studies included patients undergoing long-term pitu- itary desensitization using GnRH agonists prior to IUI treatment (10, 11, 21, 22, 24, 28, 33). However, no studies included a con- trol group of IUI treatment with no preceding GnRH agonist use. One trial (11), which used a randomized, controlled set - ting to compare ultra-long treatment of at least 6 weeks with 2 weeks of GnRH agonist pre-treatment, showed an OR of 3.53 (95% CI, 0.34-37.15; p = 0.29). The rate of recurrence of endometriosis was reported in 1 study (10). Recurrence of endometriosis, which was defined as recurrence of or increase in patient’s complaints within 12 months after the last IUI treatment attempt, was seen in 24 out of 65 patients (36.9%), requiring no intervention (n = 2), medical intervention (n = 13), and interference with the con- tinuation of IUI treatment or laparoscopic surgery (n = 9).

In our meta-analysis, observational (noncomparative) data were plotted by calculation of the weighed mean pregnancy rate, showing favorable pregnancy rates after IUI in this group of endometriosis patients. To our knowledge, this is the first systematic search and meta-analysis that has investigated the pregnancy rates of patients with moderate to severe endo- metriosis undergoing IUI treatment. We were able to com- bine a reasonable amount of articles for this specific research goal. Unfortunately, no RCTs could be identified. Therefore, our primary outcome measure could only be addressed by 1 study, showing a nonsignificant difference favoring IUI with controlled ovarian hyperstimulation compared with expect - ant management (18). Interpretation of the study results is difficult, since the in- cluded studies investigated a variety of treatment strategies and populations as presented in Table I. The best available evidence was provided by Keresztúri and colleagues (18) in - vestigating a treatment strategy of laparoscopic surgery fol - lowed by IUI with controlled ovarian hyperstimulation in all stages of endometriosis in a prospective cohort compared to no treatment after surgery. For the total group of endome- triosis patients (all ASRM stages), surgery followed by IUI with controlled ovarian hyperstimulation was more beneficial than van der Houwen et al 161 © 2017 Wichtig Publishing tABLE I - Characteristics of included studies

Design Population Purpose IUI strategy GnRH analogue Abuzeid et al (19) Retrospective cohort Moderate to severe endo - metriosis patients receiving surgery for infertility To investigate the effect of unilat- eral versus bilateral adnexal involve - ment on pregnancy Not specified; IUI was performed in a subgroup of patients after surgery in male factor infertility, or ovulatory disorder (resistant to CC) or after 6 months of natural conception. Some patients opted to directly start with IUI with COH. Not specified Alborzi et al (20) Prospective RCT Endometriosis with uni- or bilateral cysts receiving differ - ent methods of laparoscopic surgery To compare the response to COH after (i) fenestration and coagula- tion of an unilateral endometrioma; (ii) cystectomy of an unilateral en - dometrioma; (iii) cystectomy at one side and fenestration and coagula - tion at the other side in bilateral endometrioma COH with CC and hMG Ovulation: 10.000 IU hCG Insemination: 34-36 h after hCG Luteal support: not specified Not specified el Amrani et al (21) Retrospective cohort Patients with endometriosis ASRM III or IV receiving sur- gery for infertility To define the best surgical strategy in moderate to severe endometrio - sis patients with infertility Not specified; IUI was performed in a subgroup of patients after 2 months of GnRH analogue treatment after surgery. Other patients waited on a spontaneous conception or received IVF. Treatment allocation dependent on age, the presence of male factor infertility, and surgical findings. Two months Burwinkel et al (23) Retrospective cohort Infertile couples receiving IUI To determine a relationship of basal FSH and age to ovarian responsive - ness and pregnancy COH with hMG Ovulation: ? IU hCG Insemination: 36-40 h after hCG Luteal support: not specified Not specified Daru et al (24) a Unclear Infertile patients with endo - metriosis receiving surgery and GnRH agonist treatment To investigate if IUI with COH follow- ing surgery and GnRH agonist treat - ment increases pregnancy rate in infertile patients with endometriosis COH with FSH and hMG. Ovulation: 10.000 IU hCG Insemination: 36 h after hCG Luteal support: not specified GnRH agonist for 6 months Dickey et al (25) b Prospective cohort Couples receiving IUI with hMG To determine characteristics associ - ated with pregnancy and multiple gestation after IUI with hMG COH with hMG alone, with or after CC Ovulation: 10.000 IU hCG or LH surge Insemination: 24-36 h after hCG or LH surge Luteal support: not specified Not specified Dickey et al (26) Prospective cohort Couples receiving IUI with CC (anovulatory cycles or luteal insufficiency) To determine characteristics associ - ated with pregnancy and multiple gestation after IUI with CC COH with CC cycle day 3-7 Ovulation: 5.000-10.000 IU hCG or LH surge Insemination: 2-40 h after hCG, 2-28 h after LH surge. Luteal support: Not specified Not specified To be continued Efficacy of IUI in moderate to severe endometriosis 162 © 2017 Wichtig Publishing

Design Population Purpose IUI strategy GnRH analogue Dodson and Haney (27) Review including ret- rospective cohort Infertile couples receiving IUI To determine cycle fecundity of IUI with hMG in couples without anatomic pelvic distortion COH with hMG Ovulation: 5.000 IU hCG Insemination: 36-40 h after hCG Luteal support: 2.500 IU hCG day 3 and 6 Not specified van der Houwen et al (10) Retrospective cohort Moderate to severe endome - triosis patients receiving IUI To investigate the efficacy and safety of 2 different IUI treatment strategies; (i) 3 times IUI in the natural cycle followed by up to 3 cycles IUI with COH; (ii) IUI directly combined with COH COH with hMG or FSH Ovulation: 10.000 IU hCG Insemination: 42 h after hCG Luteal support: not specified A subgroup of patients re - ceived a GnRH agonist for at least three months van der Houwen et al (22) Prospective cohort Moderate to severe endo - metriosis patients receiving IUI, IVF or IVF with preced - ing long-term GnRH agonist desensitization To investigate patient satisfaction concerning one ART treatment cycle COH with or without hMG or FSH Ovulation: 10.000 IU hCG Insemination: 42 h after hCG Luteal support: not specified A subgroup of patients received GnRH agonist for at least three months Keresztúri et al (18) Prospective cohort Patients with endometriosis receiving IUI versus no treat- ment following surgery (non - random allocation) To investigate if IUI has a significant effect on pregnancy in infertile patients with endometriosis after laparoscopic surgery COH with CC and hMG Ovulation: 10.000 IU hCG Insemination: 36 h after hCG and sub- sequent day (double insemination) Luteal support: no luteal support No GnRH agonist pre- treatment Kim et al (11) RCT ultra-long versus short GnRH agonist desensitization Patients with endometriosis (ASRM I-IV) undergoing IUI To compare the ULP and LP with GnRHa desensitization for ovulation induction with IUI COH with FSH and hMG Ovulation: 10.000 IU hCG Insemination: 36-40 h after hCG Luteal support: 50 mg progesterone daily ULP: ≥6 weeks GnRHa desensitization LP: 2 weeks GnRHa desensitization Lodhi et al (28) Retrospective cohort Endometriosis patients receiv - ing GIFT or IUI To compare the effectiveness of GIFT and IUI with COH in endome - triosis patients COH with FSH or hMG Ovulation: 10.000 IU hCG Insemination: 42 h after hCG Luteal support: not specified Mid-luteal pituitary de - sensitization with a GnRH agonist Stepniewska et al (29) Retrospective cohort Infertile patients with bowel endometriosis To determine the influence of bowel endometriosis on fertility in patients who (i) underwent colorectal seg - mental resection; (ii) had evidence of bowel endometriosis without bowel resection; and (iii) underwent surgery for moderate to severe en - dometriosis with DIE without bowel involvement Not specified; A subgroup of patients tried to conceive after surgery; those patients tried to conceive spontane - ously, received IUI or IVF. Not specified tABLE I - Continued To be continued van der Houwen et al 163 © 2017 Wichtig Publishing

Design Population Purpose IUI strategy GnRH analogue Göker et al (30) Retrospective cohort Patients with endometriosis, male factor, tubal factor or unexplained infertility To evaluate the efficacy of IUI with COH in endometriosis COH with FSH or hMG Ovulation: 10.000 IU hCG Insemination: 24 h after hCG Luteal support: not specified Not specified Tay et al (31) Retrospective cohort Infertile couples receiving IUI To analyze the influence of patient and treatment characteristics on the cumulative pregnancy rate COH with CC and FSH Ovulation: 5.000 IU hCG Insemination: 36 h after hCG Luteal support: 200 mg twice daily vaginally administered progesterone Not specified Vollenhoven et al (32) Retrospective cohort Infertile couples receiving IUI To determine the effectiveness of IUI with hMG and to identify prog - nostic factors for pregnancy COH with hMG Ovulation: 10.000 IU hCG Insemination: 18-24 h and/or 36-48 h after hCG Luteal support: 25 mg twice daily progesterone vaginally Not specified Wu et al (33) Retrospective cohort Endometriosis patients receiv - ing IUI after laparoscopic treatment of endometriosis To investigate if the presence and amount of peritoneal fluid is cor- related to severity of endometriosis and pregnancy outcome of IUI treatment COH with FSH Ovulation: 250 ugr hCG Insemination: 36 h after hCG Luteal support: progesterone 600 mg/ day during 2 weeks and 1500 IU hCG on day 6 2 months Yovich and Matson (34) Unclear Nontubal infertility couples receiving IUI To determine any relationship be- tween pregnancy rate after IUI and the underlying disorder COH with hMG and/or CC Ovulation: 5.000 IU hCG or LH surge Insemination next three mornings Luteal support: non specified Not specified a Overlap of patient with Keresztúri et al (18) cannot be excluded. b Overlap of patients with Dickey et al (26) cannot be excluded. ART = assisted reproductive technology; ASRM = American Society of Reproductive Medicine; CC = clomiphene citrate; COH = controlled ovarian hyperstimulation; DIE = deeply infiltrating endometriosis; FSH = follicle-stimulating hormone; GIFT = gamete intrafallopian transfer; GnRH = gonadotropin-releasing hormone; hCG = human chorionic gonadotrophin; hMG = human menopausal gonadotropin; IU = international units; IUI = intrauterine insemination; LH = luteinizing hormone; LP = long protocol; RCT = randomized controlled trial; ULP = ultra-long protocol. tABLE I - Continued Efficacy of IUI in moderate to severe endometriosis 164 © 2017 Wichtig Publishing tABLE II - Pregnancy rates

p atients c ycles Biochemical pregnancy Clinical pregnancy Ongoing pregnancy Live birth n n n per pt per cycle n %/pt %/cycle n %/pt %/cycle n %/pt %/cycle Abuzeid et al (19) 29 51 - - - 14 48.3% 27.5% - - - 11 37.9% 21.6% Alborzi et al (20)a 81 149 48 59.3% 32.2% - - - - - - - - - el Amrani et al (21) 22 - - - - 15 68.2% - - - - - - - Burwinkel et al (23) 15 37 - - - 3 20.0% 8.1% - - - 3 20.0% 8.1% Daru et al (24)a 38 - 17 44.7% - - - - - - - - - - Dickey et al (25) - 114 - - - 6 - 5.3% - - - 5 - 4.4% Dickey et al (26) 82 213 12 14.6% 5.6% 10 12.2% 4.7% - - - 7 8.5% 3.3% Dodson and Haney (27) - 65 9 - 13.8% - - - - - - - - - van der Houwen et al (10) 65 245 - - - - - - 15 23.1% 6.1% 12 18.5% 4.9% van der Houwen et al (22) 25 25 4 16.0% 16.0% - - - 2 8.0% 8.0% - - - Keresztúri et al (18) 68 - - - - 31 45.6% - - - - 28 41.2% - Kim et al (11) 41 41 - - - 14 34.1% 34.1% 10 24.4% 24.4% 4 9.8% 9.8% Lodhi et al (28) 14 16 - - - 5 35.7% 31.3% - - - 2 14.3% 12.5% Stepniewska et al (29) 6 - - - - 4 66.7% - - - - 3 50.0% - Göker et al (30) 17 - - - - 4 23.5% - - - - - - - Tay et al (31) - 48 - - - 5 - 10.4% - - - - - - Vollenhoven et al (32) 8 15 3 37.5% 20.0% - - - - - - - - - Wu et al (33) 47 47 - - - 11 23.4% 23.4% - - - - - - Yovich and Matson (34) 23 49 2 8.7% 4.1% - - - - - - - - - a Unclear definition of pregnancy used, stated in this table as biochemical pregnancy. N = number; pt = patients. Fig. 2 - Forest plot on live birth rate. surgery alone in terms of live birth (OR 1.85; 95% CI, 1.09- 1.34; p = 0.02). For patients with stage III/IV endometriosis, this effect was less pronounced (41.2% vs. 28.4%; p = 0.29). This difference, favoring IUI treatment, did not reach statisti- cal significance, which was possibly due to a lack of power. Surgery in moderate to severe endometriosis patients seeking for natural conception is recommended by the The European Society of Human Reproduction and Embryology (ESHRE) guideline (6), since surgery compared to expectant management improves spontaneous pregnancy rates (level B evidence). However, prior to IVF/intracytoplasmic sperm injection (ICSI) surgical treatment of moderate to severe en- dometriosis (i.e., performing cystectomy) should be carefully considered, taking into account the lack of evidence that per- forming a cystectomy improves pregnancy rates; the possible loss of functional ovarian tissue (especially in repetitive ovar- ian surgery); the accessibility of follicles; and patients’ pain complaints. Whether surgery of moderate to severe endo- metriosis prior to IUI affects pregnancy rates has not been investigated. However, it can be hypothesized that laparo- scopically treating peritoneal disease in moderate to severe endometriosis patients (without performing cystectomy in the absence of endometriosis-related pain complaints), may be more beneficial in improving pregnancy rates with IUI compared to IUI alone, as it is already shown for natural conception (35) (level A evidence) and pregnancy after IVF (36) (level C evidence). Unfortunately, the studies included in our meta-analysis reported no data on surgical interventions. Performing surgery also results in the ability to calculate the endometriosis fertility index (EFI) score as described by van der Houwen et al 165 © 2017 Wichtig Publishing Fig. 3 - Weighed mean clinical pregnancy rate. Fig. 4 - Weighed mean live birth rate. Adamson and Pasta to predict pregnancy rates in patients attempting nonIVF conception (37). This fertility index is al- ready externally validated (38), and recent studies focus on the possibility to use this tool to select patients who should be offered IVF/ICSI (39, 40). Whether IUI can be performed prior to IVF/ICSI at specific cut-off values of the EFI score has not yet been investigate. Next to the variety of strategies and populations, different IUI treatment strategies have been described in the included studies, such as those concerning IUI in a natural cycle and IUI with controlled ovarian hyperstimulation, including anti- estrogens (clomifenecitrate), recombinant follicle stimulat - ing hormone (FSH) and/or human menopausal gonadotropin (hMG). It is worth noting that in subfertile couples, gonado - trophins are superior to antiestrogens regarding pregnancy rates after IUI (41), but this has never been evaluated in mod- erate to severe endometriosis patients. It has been shown that prior long-term pituitary desen- sitization promotes pregnancy rates after IUI in endometrio- sis patients ASRM stage 2-4 (12). Besides this, the beneficial effect in this specific subgroup of ASRM 2-4 endometriosis patients (100% vs. 70% pregnancy rate) was not significant - ly different, which is possibly due to a lack of power. This is in line with the results published by Kim and colleagues (11), investigating a modified ultra-long desensitization with a GnRH agonist for 6 weeks in an RCT showing an OR on live birth of 3.53 (95% CI, 0.34-37.15). Besides this, our recently published retrospective study on IUI in moder - ate to severe endometriosis patients showed a hazard ra- tio of 1.8 (95% CI, 0.6-5.1) in favor of long-term pituitary desensitization (nonsignificantly) for at least 3 months (10). Those results ask for more research regarding the possible beneficial effect of long-term pituitary desensitization with a GnRH agonist. Differences in stimulation protocols, timing of surgical in- terventions, and the use of preceding treatment with a GnRH agonist might have influenced the outcome of our meta-anal- ysis, and most of the included studies were not designed to specifically investigate the effect of IUI in moderate to severe endometriosis patients. However, despite this clinical hetero- geneity, calculation of the weighed mean pregnancy rate was statistically the best method to combine observational data and calculate the effect of intrauterine insemination on preg- nancy rates, resulting in a level 4 evidence (grade C) recom- mendation. Next to efficacy, the safety of treatment in terms of com- plications and the recurrence of endometriosis should be taken into account, especially in moderate to severe endome- triosis patients. Recurrence of endometriosis complaints may lead to discontinuation of fertility treatment due to the need to start hormonal suppression therapy or surgical interven- tion. This directly results in postponement or cancellation of the possibility to conceive (10), which negatively affects time to pregnancy. D’Hooghe and colleagues (42) have shown that the endometriosis recurrence rate is higher in patients un - dergoing IUI than in IVF, which is explained by the result of a more frequent exposure to ovulation and subsequent retro- grade menstruation in IUI treatment. By limiting the amount of IUI treatment cycles to 3 per patient, this negative effect might be negligible (10). Efficacy of IUI in moderate to severe endometriosis 166 © 2017 Wichtig Publishing

This meta-analysis of observational data has shown that IUI could be a feasible treatment in moderate to severe endometriosis. Due to the lack of RCTs, no clear recommen- dation can be made in offering IUI treatment in moderate to severe endometriosis patients. Whether this treatment should be structurally performed prior to IVF must be investigated in a RCT, including time-to-pregnancy, safety, and cost effectiveness. Disclosures Financial support: None. Conflict of interest: None. Meeting presentation: The work described in this manuscript was presented at the 13 th World Congress on Endometriosis (2017), in Vancouver, Canada.

1. Collins JA, Burrows EA, Wilan AR. The prognosis for live birth among untreated infertile couples. Fertil Steril. 1995;64(1): 22-28. 2. Hamdan M, Dunselman G, Li TC, Cheong Y . The impact of en- dometrioma on IVF/ICSI outcomes: a systematic review and meta-analysis. Hum Reprod Update. 2015;21(6):809-825. 3. Johnson NP , Hummelshoj L. World Endometriosis Society Montpellier Consortium. Consensus on current management of endometriosis. Hum Reprod. 2013;28(6):1552-1568. 4. Surrey ES. Endometriosis-related infertility: the role of the assisted reproductive technologies. Biomed Res Int. 2015, Article ID 482959, doi:10.1155/2015/482959. http://dx.doi. org/10.1155/2015/482959. Accessed August 9, 2017. 5. de Ziegler D, Borghese B, Chapron C. Endometriosis and infertility: pathophysiology and management. Lancet. 2010;376(9742):730-738. 6. Dunselman GAJ, Vermeulen N, Becker C, et al. European Soci - ety of Human Reproduction and Embryology. ESHRE guideline: management of women with endometriosis. Hum Reprod. 2014;29(3):400-412. 7. Tummon IS, Asher LJ, Martin JS, Tulandi T. Randomized con - trolled trial of superovulation and insemination for infertility associated with minimal or mild endometriosis. Fertil Steril. 1997;68(1):8-12. 8. Nulsen JC, Walsh S, Dumez S, Metzger DA. A randomized and longitudinal study of human menopausal gonadotropin with intrauterine insemination in the treatment of infertility. Obstet Gynecol. 1993;82(5):780-786. 9. Practice Committee of the American Society for Reproductive Medicine. Endometriosis and infertility: a committee opinion. Fertil Steril. 2012;98(3):591-598. 10. van der Houwen LEE, Schreurs AMF, Schats R, et al. Efficacy and safety of intrauterine insemination in patients with mod - erate-to-severe endometriosis. Reprod Biomed Online. 2014; 28(5):590-598. 11. Kim CH, Cho YK, Mok JE. Simplified ultralong protocol of gonad- otrophin-releasing hormone agonist for ovulation induction with intrauterine insemination in patients with endometriosis. Hum Reprod. 1996;11(2):398-402. 12. Rickes D, Nickel I, Kropf S, Kleinstein J. Increased pregnancy rates after ultralong postoperative therapy with gonadotropin- releasing hormone analogs in patients with endometriosis. Fertil Steril. 2002;78(4):757-762. 13. Stroup DF, Berlin JA, Morton SC, et al. For the meta-analysis of observational studies in epidemiology (MOOSE) group. Meta- analysis of observational studies in epidemiology – a proposal for reporting. JAMA. 2000;283:2008-2012. 14. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Pre- ferred reporting items for systematic reviews and meta-analy- ses: the PRISMA statement. BMJ. 2009;339:b2535. 15. Neyeloff JL, Fuchs SC, Moreira LB. Meta-analyses and Forest plots using a microsoft excel spreadsheet: step-by-step guide fo- cusing on descriptive data analysis. BMC Res Notes. 2012;5:52. 16. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring in- consistency in meta-analyses. BMJ. 2003;327(7414):557-560. 17. Dodson WC, Whitesides DB, Hughes CL, Jr, Easley HA III, Haney AF. Superovulation with intrauterine insemination in the treatment of infertility: a possible alternative to gamete intrafallopian transfer and in vitro fertilization. Fertil Steril. 1987;48(3):441-445. 18. Keresztúri A, Kozinszky Z, Daru J, et al. Pregnancy rate after con- trolled ovarian hyperstimulation and intrauterine insemination for the treatment of endometriosis following surgery. Biomed Res Int. 2015, Article ID 282301, doi:10.1155/2015/282301. https:// pdfs.semanticscholar.org/f89e/ec4a25d7d6ddf5feae33e1e075f- 63c88241b.pdf. Accessed August 9, 2017. 19. Abuzeid M, Ahmed A, Sakhel K, et al. Unilateral versus bilateral adnexal disease in stage III and stage IV endometriosis does not affect pregnancy outcome after operative laparoscopy. Gy- necol Surg. 2009;6:39-44. 20. Alborzi S, Ravanbakhsh R, Parsanezhad ME, Alborzi M, Alborzi S, Dehbashi S. A comparison of follicular response of ovaries to ovulation induction after laparoscopic ovarian cystectomy or fenestration and coagulation versus normal ovaries in patients with endometrioma. Fertil Steril. 2007;88(2):507-509. 21. el Amrani R, Henry-Suchet J, Cornier E, et al. Comparaison de deux stratégies thérapeutiques dans les endométrioses sévères, chez des femmes jeunes consultant pour stérilité ou douleurs. II. En cas dinfertilité, intérêt des stimulations ovari- ennes avec inséminations intra-utérines après la chirurgie. Gy- necol Obstet Fertil. 2001;29(3):192-199. 22. van der Houwen LE, Schreurs AM, Schats R, Lambalk CB, Hompes PG, Mijatovic V. Patient satisfaction concerning assist- ed reproductive technology treatments in moderate to severe endometriosis. Gynecol Endocrinol. 2014;30(11):798-803. 23. Burwinkel TH, Buster JE, Scoggan JL, Carson SA. Basal follicle stimulating hormone (FSH) predicts response to controlled ovarian hyperstimulation (COH)-intrauterine insemination (IUI) therapy. J Assist Reprod Genet. 1994;11(1):24-27. 24. Daru J, Keresztúri A, Szöllősi J, Koloszár S, Pál A. The role of intrauterine insemination in the therapy of infertile women with endometriosis. J Endometr Pelvic Pain Disord. 2011;3(2): 93-98. 25. Dickey RP , Olar TT, Taylor SN, Curole DN, Rye PH, Matulich EM. Relationship of follicle number, serum estradiol, and other fac- tors to birth rate and multiparity in human menopausal gonad- otropin-induced intrauterine insemination cycles. Fertil Steril. 1991;56(1):89-92. 26. Dickey RP , Olar TT, Taylor SN, Curole DN, Rye PH. Relationship of follicle number and other factors to fecundability and mul - tiple pregnancy in clomiphene citrate-induced intrauterine in- semination cycles. Fertil Steril. 1992;57(3):613-619. 27. Dodson WC, Haney AF. Controlled ovarian hyperstimulation and intrauterine insemination for treatment of infertility. Fertil Steril. 1991;55(3):457-467. 28. Lodhi S, Abdel Fattah A, Abozaid T, et al. Gamete intra-fallopian transfer or intrauterine insemination after controlled ovarian hyperstimulation for treatment of infertility due to endome- triosis. Gynecol Endocrinol. 2004;19(3):152-159. van der Houwen et al 167 © 2017 Wichtig Publishing 29. Stepniewska A, Pomini P , Bruni F, et al. Laparoscopic treat - ment of bowel endometriosis in infertile women. Hum Reprod. 2009;24(7):1619-1625. 30. Göker EN, Ozçakir HT, Terek MC, Levi R, Adakan S, Tavmergen E. Controlled ovarian hyperstimulation and intrauterine insemi - nation for infertility associated with endometriosis: a retro- spective analysis. Arch Gynecol Obstet. 2002;266(1):21-24. 31. Tay PY , Raj VR, Kulenthran A, Sitizawiah O. Prognostic factors influencing pregnancy rate after stimulated intrauterine in - semination. Med J Malaysia. 2007;62(4):286-289. 32. Vollenhoven B, Selub M, Davidson O, et al. Treating infertility. Controlled ovarian hyperstimulation using human menopausal gonadotropin in combination with intrauterine insemination. J Reprod Med. 1996;41(9):658-664. 33. Wu HM, Tzeng CR, Chen CH, Chen PH. Pelvic endometriosis with peritoneal fluid reduces pregnancy rates in women un - dergoing intrauterine insemination. Taiwan J Obstet Gynecol. 2013;52(4):512-515. 34. Yovich JL, Matson PL. The treatment of infertility by the high intrauterine insemination of husbands washed spermatozoa. Hum Reprod. 1988;3(8):939-943. 35. Jacobson TZ, Duffy JM, Barlow D, Farquhar C, Koninckx PR, Olive D. Laparoscopic surgery for subfertility associated with endometriosis. Cochrane Database Syst Rev. 2010;(1): CD001398. 36. Opøien HK, Fedorcsak P , Byholm T, Tanbo T. Complete surgical removal of minimal and mild endometriosis improves outcome of subsequent IVF/ICSI treatment. Reprod Biomed Online. 2011;23(3):389-395. 37. Adamson GD, Pasta DJ. Endometriosis fertility index: the new, validated endometriosis staging system. Fertil Steril. 2010;94(5):1609-1615. 38. Tomassetti C, Geysenbergh B, Meuleman C, Timmerman D, Fieuws S, DHooghe T. External validation of the endome- triosis fertility index (EFI) staging system for predicting non- ART pregnancy after endometriosis surgery. Hum Reprod. 2013;28(5):1280-1288. 39. Garavaglia E, Pagliardini L, Tandoi I, et al. External validation of the endometriosis fertility index (EFI) for predicting spon - taneous pregnancy after surgery: further considerations on its validity. Gynecol Obstet Invest. 2015;79(2):113-118. 40. Zeng C, Xu JN, Zhou Y , Zhou YF, Zhu SN, Xue Q. Reproduc - tive performance after surgery for endometriosis: predictive value of the revised American Fertility Society classification and the endometriosis fertility index. Gynecol Obstet Invest. 2014;77(3):180-185. 41. Cantineau AE, Cohlen BJ, Heineman MJ. Ovarian stimulation protocols (anti-oestrogens, gonadotrophins with and with - out GnRH agonists/antagonists) for intrauterine insemination (IUI) in women with subfertility. Cochrane Database Syst Rev. 2007;(2):CD005356. 42. D’Hooghe TM, Denys B, Spiessens C, Meuleman C, Debrock S. Is the endometriosis recurrence rate increased after ovarian hyperstimulation? Fertil Steril. 2006;86(2):283-290.