Is the endometriosis recurrence rate increased after ovarian hyperstimulation?

Is the endometriosis recurrence rate increased after ovarian hyperstimulation? Thomas M. D Hooghe, M.D., Ph.D., Bénédicte Denys, M.D., Carl Spiessens, Ph.D., Christel Meuleman, M.D., and Sophie Debrock, Ph.D. Leuven University Fertility Center, Department of Obstetrics and Gynecology, University Hospital Gasthuisberg, Leuven, Belgium Objective: To test the hypothesis that the cumulative endometriosis recurrence rate (CERR) after fertility surgery for endometriosis stage III or IV is increased in women exposed to very high E 2 levels during ovarian hyperstimulation (OH) for IVF when compared with women exposed to less high E 2 levels during OH for intrauterine insemination (IUI). Design: Retrospective cohort study including infertility patients with endometriosis stage III or IV. Setting: Leuven University Fertility Center, between 1990 and 2001. Patient(s): Patients (n 67) with endometriosis stage III (n 45) or IV (n 22) who underwent pelvic reconstructive surgery and subsequently started fertility treatment with either IVF only (n 39), both IVF and IUI in different cycles (n 11), or IUI only (n 17). Intervention(s): Life table analysis was used to calculate the CERR. Main Outcome Measure(s): The CERR based on histologic or cytologic proof of disease recurrence. Result(s): At 21 months after the start of OH the overall CERR was 31% and was significantly lower in patients treated with IVF only (7%) or women treated with both IVF and IUI in different cycles (43 %) than in those treated with IUI only (70%). At 36 months after the start of OH, the overall CERR was 63%. Conclusion(s): In contrast to our hypothesis, the results from this study showed that the CERR is lower after ovarian hyperstimulation for IVF than after lower-dose ovarian stimulation for IUI, suggesting that temporary exposure to very high E 2 levels in women during OH for IVF is not a major risk factor for endometriosis recurrence in women treated with assisted reproductive technology. (Fertil Steril 2006;86:283 90. 2006 by American Society for Reproductive Medicine.) Key Words: Endometriosis recurrence, ovarian hyperstimulation, moderate to severe endometriosis Received January 10, 2006; revised and accepted January 17, 2006. Reprint requests: Thomas M. D Hooghe, M.D., Ph.D., Department of Obstetrics and Gynecology, University Hospital Gasthuisberg, Herestraat 6, 3000 Leuven, Belgium (FAX: 32-16-344368; E-mail: thomas. dhooghe@uz.kuleuven.ac.be). Endometriosis can be considered as an estrogen-dependent disease, because it is rarely observed before menarche and usually disappears after menopause. It is well known that moderate to severe (1, 2) endometriosis can be a recurrent disease, both after cessation of medical suppressive treatment and after surgical treatment (3 5). Recurrences of endometriosis after surgery can be explained by incomplete surgery, persistence and growth of microscopic endometriosis, the development of new lesions, or a combination of these factors. However, it is not known whether temporary and repeated exposure to high E 2 levels during controlled ovarian hyperstimulation (COH) for IVF contributes to the recurrence of endometriosis. Some case reports suggest indeed that OH may lead to a higher recurrence rate of endometriosis. About 10 years ago (6), a rare coincidence of ureteral endometriosis and ovarian stimulation with hmg was described in a patient with a history of a surgically removed endometriotic cyst. Symptoms occurred 10 days after her first IVF attempt. Furthermore, four cases of severe digestive complications due to the rapid growth of sigmoid endometriosis were reported under OH with hmg (7). In that study, all four patients underwent a laparoscopy where endometriosis of revised American Fertility Society/American Society for Reproductive Medicine (rafs) stage IV was found and treated before the onset of COH. Symptoms occurred after one to seven cycles (mean 3.5 cycles) of COH for IVF. The mean E 2 concentration was 2,383.75 187.94 pg/ml when the bowel symptoms started. These case reports suggest that repeated ovarian hyperstimulation during IVF may stimulate the growth of endometriotic lesions and lead to recurrence of endometriosis. However, no prospective controlled cohort studies are available to support this hypothesis. In this retrospective cohort study we tested the hypothesis that the cumulative endometriosis recurrence rate (CERR) after fertility surgery for stage III and IV endometriosis is increased in women exposed to very high E 2 levels during OH for IVF when compared with a control group of women exposed to less high E 2 levels during OH for intrauterine insemination (IUI). As a second outcome variable, we measured the CERR after fertility surgery for moderate or deep 0015-0282/06/$32.00 Fertility and Sterility Vol. 86, No. 2, August 2006 doi:10.1016/j.fertnstert.2006.01.016 Copyright 2006 American Society for Reproductive Medicine, Published by Elsevier Inc. 283

endometriosis in infertile patients before the start of IVF or IUI. MATERIALS AND METHODS Patient Recruitment The database of the Leuven University Fertility Center was searched in June 2001 for all patients with a previous history of surgically treated endometriosis stage III or IV, who underwent IUI, IVF, or a combination of both between February 1990 and June 2001. Institutional Review Board approval was not needed, because this was a retrospective study based on electronic patient files. Staging of endometriosis followed the standard rafs classification system (1, 2). We identified 74 women who had received pelvic reconstructive surgery for endometriosis before entering our program of medically assisted reproduction using artificial reproductive techniques (ARTs) including IUI and/or IVF. Controlled Ovarian Stimulation In our IUI program, patients were stimulated with either clomiphene citrate (50 100 mg per day on days 3 7 of the cycle) or gonadotropins (75 to 150 IU FSH/LH per day, starting on day 2 of the cycle) and monitored using gynecologic ultrasound and hormonal analysis (E 2, P, LH). An injection of 10,000 IU hcg was given when one or two mature follicles of at least 17 mm diameter were documented, at a median E 2 level of 331 (range 114 1,241) pg/ml (Table 1). When three or more follicles were present, the patient was offered a selective aspiration of some follicles, with the goal to conserve one (at most two) follicles to prevent higher-order multiple pregnancies. In our IVF program, COH was performed as follows. In the previous cycle an oral contraceptive pill was given for 25 days (Cilest; Janssen-Cilag, Berchem, Belgium), followed by pituitary down-regulation with busereline acetate (600 g daily; Suprefact; Hoechst, Frankfurt, Germany) from day 21 onwards. Daily injections of hmg (Humegon [Organon, Oss, The Netherlands] or Menopur [Ferring, Copenhagen, Denmark]) were started on the second day after the subsequent menstruation, followed by monitoring using ultrasound and hormonal analysis as mentioned above. Final follicular maturation was induced with 10,000 IU hcg (Pregnyl; Organon) when at least three follicles of 17 mm or more were present. Ultrasound-guided oocyte aspiration was performed 36 hours later. Luteal supplementation was given either by 1,500 IU hcg, every three days, starting on the third day after oocyte retrieval, or by vaginal progesterone (600 mg/ day; Utrogestan; Piette International, Drogenbos, Belgium) when there was an increased risk for hyperstimulation syndrome (E 2 level 3500 pg/ml at the time of hcg injection, or patients with polycystic ovarian syndrome). In freeze-thaw cycles, the endometrium was prepared using COH with a low dose (75 IU) of hmg (Humegon or Menopur). Monitoring and hcg injection was performed as described above for IUI cycles. Post-ART Recurrence Study The study group included patients exposed to at least one cycle of IVF and consisted of two subgroups: women treated with IVF alone (n 39) and women treated with both IVF and IUI in different cycles (n 11) (Table 1). The control group (IUI only; n 17) consisted of women treated with IUI after minimal ovarian hyperstimulation with gonadotropins or clomiphene citrate, with a median E 2 level of 331 pg/ml at the time of hcg injection (Table 1). The CERR was calculated from the first cycle of OH for IUI or IVF. Patients entered the study at the time of their first TABLE 1 Recurrence of endometriosis after treatment with ART according to type of treatment and to median E 2 levels per cycle (post-art recurrence study). Study group Control group IVF alone Combination of IUI and IVF IUI alone IVF OPU IVF cryo IUI IVF OPU IVF cryo IUI Patients 39 (7) a 11 (1) b 17 Total cycles 101 11 28 21 1 50 Median E 2 level/cycle (pg/ml) 1838 691 454 1140 463 331 Range 532 5479 441 1,333 53 1,097 52 2,807 114 1,241 Note: OPU ovum pick up. a Seven out of 39 patients undergoing IVF ovum pick up cycles additionally received one or several cryo cycles. b One out of 11 patients undergoing IVF ovum pick up cycles additionally received a cryo cycle. D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. 284 D Hooghe et al. Endometriosis recurrence after ovarian hyperstimulation Vol. 86, No. 2, August 2006

IUI or IVF cycle. Patients left the study when they became pregnant (regardless of outcome of pregnancy, including patients with miscarriage, ectopic pregnancy, and premature birth), when they developed symptoms or signs of recurrence of endometriosis, or when they were lost to follow-up. Out of 74 patients included in our study, 67 entered the post-art recurrence study. The seven patients excluded from the post-art recurrence analysis included three women who developed recurrence of endometriosis at an early stage during the first ART cycle and who were included in the pre-art recurrence analysis, as discussed below. These women did not continue any infertility treatment in our center after being reoperated. The other four patients had a previous history of endometriosis of rafs III or IV but developed symptoms or signs of recurrence before the planned start of ART. At the time of reoperation, rafs I or II endometriosis was found. Because these patients did not have rafs III or IV endometriosis at this reoperation (defined as the last complete surgery) they could not enter the post-art recurrence study and also were included in the pre-art recurrence analysis. The mean age of this population at the time of their initial participation in the ART program was 30 4 years. At the time of entrance in the study, 45 (67%) and 22 (33%) patients had a previous history of rafs III and rafs IV endometriosis, respectively. All patients had undergone macroscopically complete resection of endometriosis, either by laparoscopy (94%) or by laparotomy (6%). In 55 patients (82%), the surgery had been performed at the University Hospital Gasthuisberg, whereas 12 women (18%) had undergone surgery at another center. All patients operated at the University Hospital Gasthuisberg had received laser surgery. In other centers, mostly unipolar coagulation and/or excision of the lesions had been performed. Six patients (9%) received additional treatment with GnRH analogs immediately after surgery. The mean time between the surgical treatment of endometriosis and the start of ART was 8 9 months. At the time of entrance in the study, 56 women (84%) were primary infertile and 11 (16%) were secondary infertile. The mean duration of infertility at the moment of the first ART cycle was 33 10 months. Seventeen patients (25%) were treated with IUI only, 39 patients (58%) underwent IVF treatment only, and 11 patients (17%) were treated with both IVF and IUI in different cycles. All but one patient underwent the ART treatment in the University Hospital Gasthuisberg. Pre-ART Recurrence Study Retrospective analysis revealed that three patients already had developed recurrence of endometriosis at the time of their first ultrasound monitoring during OH for their first IUI or IVF cycle. These patients also had symptoms of dysmenorrhea, dyspareunia, or chronic pelvic pain during the months preceding the first IUI or IVF treatment. Because it is unlikely that these early recurrences were related to IUI or IVF treatment, they were classified as pre-art recurrences. In addition, all other patients included in the post-art analysis were included in a study evaluating the cumulative recurrence rate before the onset of ART (pre-art recurrence analysis). In the main analysis (pre-art recurrence analysis part 1), patients entered the study at the time of their primary surgical treatment of endometriosis. When they had received additional treatment with GnRH analogs immediately after the surgery, they only entered the study 1 month after the last injection of GnRH analogs, because they were very unlikely to have recurrent endometriosis during treatment with GnRH analogs. Patients left the study when they became pregnant (regardless of outcome of pregnancy, including patients with miscarriage, ectopic pregnancy, premature birth, etc.), when they developed recurrence of endometriosis, or when they started ART treatment. In an additional analysis (pre-art recurrence analysis 2), patients only left the study when they had an ongoing pregnancy of more than 20 weeks, when they developed recurrence of endometriosis, or when they started ART treatment. In patients with a pregnancy duration of less than 20 weeks, the duration of this pregnancy was subtracted from the total time of follow-up. This was done because it is well accepted that pregnancy, like GnRH treatment, is a period of amenorrhea during which the risk of endometriosis recurrence can be considered to be very low. All the 74 patients entered the pre-art study at the time of their primary complete surgical treatment of endometriosis. The mean age of this population at the time of primary surgery was 29 4 years. At the time of entrance in the study, 63 patients (85%) were primary infertile and 11 (15%) were secondary infertile. The mean duration of infertility was 23 18 months. At the time of entrance in the pre-art study, there were 17 patients with a history of at least one previous diagnostic or minimal therapeutic surgery at a median of 22 (range 5 94) months before complete surgical treatment. Three of them additionally received a treatment with gonadotropic hormones for IUI or IVF before the complete surgical treatment (Table 2). At the time of diagnosis of endometriosis, there were 26 patients (35%) with endometriosis stage IV and 48 (65%) with stage III. The presence of endometriosis was confirmed by histopathology (ectopic endometrial glands and stroma), except in four patients who had only a visual diagnosis of endometriosis without biopsy confirmation. All patients were treated by laparoscopy except for five (7%) who underwent laparotomy. The day of complete surgical treatment was considered as the time of entry in the pre-art study. Out of 74 patients, 43 (58%) received complete surgery in one step, and 31 (42%) had complete surgery in two steps (initial surgery and com- Fertility and Sterility 285

TABLE 2 Clinically relevant previous history before complete primary surgery in patients analyzed for recurrence of endometriosis before the start of treatment with ART (pre-art recurrence study). Patient number Diagnosis of endometriosis without any treatment 15 Diagnosis of endometriosis with minimal therapeutic surgery 20, 22, 63 Diagnosis of endometriosis followed by treatment with GnRH analogues 3, 56, 57, 65, 66, 68 Diagnosis and treatment of umbilical endometriosis 14, 18 Diagnosis of endometriosis with or without minimal therapeutic surgery followed 19, 27, 48 or not followed by treatment with GnRH analogues n cycles IUI and/or IVF Diagnosis of other lesions (PID, adhesions, benign cysts, etc.) 25, 50, 57 D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. plete second-look surgery). In the interval period (median duration 4 months, range 1 14 months) between the two surgical procedures, 27 of them received a treatment with GnRH analogs. In the 31 patients who received second-look surgery, the date of the second and complete operation was chosen as the time of entry in the study. In 60 patients (81%) the surgery was performed in the University Hospital Gasthuisberg, and 14 (19%) received the surgery in a regional hospital. Eight patients (11%) received additional treatment with GnRH analogs immediately after complete surgical treatment. Definition of Recurrence of Endometriosis Proven recurrence of endometriosis was defined as the presence of clinical and/or biopsy-proven endometriosis at laparoscopy or as the presence of an endometriotic cyst on ultrasound, confirmed by cytologic examination (presence of endometrial glands and stroma) of the cyst fluid aspirated under ultrasound guidance. Suspected recurrence based only on ultrasound appearance of ovarian cysts was not considered to be sufficient evidence to consider the patient as having recurrence of endometriosis. Life Table Analysis Life table analysis (8) was used to calculate the CERR. Differences in CERR between groups were analyzed with the Kaplan-Meier test. Life table analysis is a method that allows taking into account a variable time of follow-up and a variable number of treatment cycles for each patient, as described below. In the post-art recurrence study, the date of entry for each patient was the start of ART. The end point was dependent on the clinical follow-up. For patients with recurrent endometriosis, the end point was the date of reoperation. For pregnant patients, the end point was the date of their positive pregnancy test. For the other patients who became lost to follow-up, the end point was the date of their last visit. Time intervals between end point and start of the study were calculated by subtraction of dates ([reoperation date] [start ART date] for patients with recurrent endometriosis, [date last ART before pregnancy] [start ART date] for patients who became pregnant, and [date last visit] [start ART date] for patients lost to follow-up). The time interval used for life table analysis was 1 month for the first year and 3 months for the subsequent years. In the pre-art recurrence study, the date of entry for each patient was the date of complete surgery for endometriosis. Each patient s end point was again determined in one of several ways. In the main analysis (pre-art recurrence analysis 1), the end point for reoperated patients was the date of reoperation. For patients who became pregnant spontaneously, the end point was the date of diagnosis of pregnancy. For the majority of patients who were not reoperated and who did not became pregnant, the end point was the date of starting ART. Time intervals between end point and start were again calculated by subtraction of dates ([reoperation date] [primary surgery date] for patients with recurrent endometriosis, [date of diagnosis of pregnancy] [primary surgery date] for patients who became pregnant, [start ART date] [primary surgery date] for patients who started ART. The life table intervals were 1 month during the first 2 years of follow-up and 3 months during the further follow-up. In the additional analysis (pre-art recurrence analysis 2), two patients who became pregnant but miscarried within 20 weeks of gestational age changed their end point. They were both reoperated, so the end point for these two patients became the date of reoperation. The other two patients who miscarried started ART a few months later, and that date was their end point of the study. The duration of the pregnancy from the four patients who miscarried was subtracted from the length of follow-up. 286 D Hooghe et al. Endometriosis recurrence after ovarian hyperstimulation Vol. 86, No. 2, August 2006

FIGURE 1 Post-ART recurrence study: CERR after treatment with IUI, IVF, or a combination of IUI and IVF. D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. Estimated Cumulative Peak Exposure to E 2 During Ovarian Hyperstimulation Because we were interested in the relation between the exposure to high levels of E 2 and the recurrence of endometriosis, we calculated the total exposure to E 2 for each patient during their participation to the study. For each patient, we added the peak E 2 level (at the time of hcg injection) from each stimulated cycle (IUI or IVF), to a standard value of 250 pg/ml for each natural cycle during their period of participation to the study. We considered the value of 250 pg/ml to be an acceptable presentation of the preovulatory level of E 2 in a natural cycle. As a result, we obtained for each patient a value that represented the total peak E 2 exposure during the time of the study. In only four patients the E 2 levels during OH were not known. In those patients, we used the median peak E 2 levels from all patients during all IUI cycles (340 pg/ml), all IVF cycles (1,728.5 pg/ml), and all cryo-ivf cycles (674 pg/ ml) as a substitute for the missing data. Statistics Life table analysis was used to calculate the CERR. The Kaplan-Meier test was used to compare the CERR between groups. The Kruskal-Wallis test was used to compare E 2 levels in the three different subgroups (IUI only, IVF IUI, IVF only). Finally, the Mann-Whitney test was used to find out whether there was a correlation between the total peak E 2 exposure per patient (independent of the kind of ART) and the presence or absence of endometriosis recurrence. Ten of the 11 reoperations, performed in patients with recurrence of endometriosis, were performed at the University Hospital Gasthuisberg. When comparing the revised AFS endometriosis score (2) at primary surgery with the score at 11 reoperations, the score was either unchanged (n 5), decreased (n 3), or increased (n 3). In one patient, deep endometriosis was present at reoperation but was not present or observed at primary surgery. In another patient a nodule at the sigmoid was found and considered as new disease. One patient became pregnant after IVF, started 4 years later again with ART to have a second child, and then developed recurrence of endometriosis. Another patient became pregnant after IUI and developed recurrence of endometriosis after her pregnancy. The recurrences of these two patients were not included in the life table analysis. After a follow-up of 21 months, the overall post-art CERR was 31% (Fig. 1). At that time, the CERR was 84% in the IUI-only group, 7% in the IVF-only group, and 43% in the IVF IUI group. The CERR after 21 months was significantly higher in the IUI group (84%) compared with the group treated with IVF alone (7%) (P.0002) and when compared with all patients treated with IVF alone or with both IVF and IUI in different cycles (17%) (P.0002) (Figs. 1 and 2). Pre-ART Recurrence Study Main Analysis (Pre-ART Recurrence Study Analysis 1). In the first analysis, the median follow-up of patients was 11 (1 38) months after the primary surgical treatment of endometriosis. Of 74 patients with proven endometriosis, 4 patients became pregnant (all miscarried before 20 weeks of gestation), 55 patients started ART, and 15 patients (20%) developed recurrence of endometriosis. Additional Analysis (Pre-ART Recurrence Study Analysis 2). In the second analysis, the median follow-up of patients was 11 (1 38) months after the primary surgical treatment of endometriosis. Of 74 patients with proven endometriosis, none had an ongoing pregnancy of more than 20 weeks, 57 pa- FIGURE 2 Post-ART recurrence study: CERR after treatment with IUI, IVF, or a combination of IUI and IVF. RESULTS Post-ART Recurrence Study In the post-art recurrence study, the median follow-up was 10 (0 46) months after the start of ART. Of 67 patients who received ART, 11 patients (16%) developed recurrence of endometriosis, 38 patients (57%) became pregnant, and 18 patients (27%) were lost to follow-up. D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. Fertility and Sterility 287

tients started ART, and 17 patients (23%) developed recurrence of endometriosis. In six patients who developed recurrence of endometriosis, the diagnosis was suspected at the start of the first ART cycle and confirmed by reoperation. Sixteen of the 17 reoperations in the patients with recurrence of endometriosis were performed at the University Hospital Gasthuisberg, Leuven. The revised AFS endometriosis score (2) at reoperation was unchanged (n 11) or decreased (n 6) compared with the score obtained at the primary surgery. One patient had umbilical endometriosis at reoperation which was not present at primary surgery, but she had a history of umbilical endometriosis 3 years earlier (Table 2). Another patient had a nodule at the sigmoid and at the rectovaginal septum and these were considered to be new disease. This patient also developed a new recurrence 8 months later. A third patient developed two more recurrences after the first recurrence, with intervals of 6 and 13 months after the operation for the first recurrence. At the last reoperation a nodule at the rectosigmoid was found and resected. After a follow-up of 24 months, the pre-art CERR was 52% according to the main and 56% according to the additional analysis (Fig. 3). The CERR after 21 months and 3 years of follow-up was similar in the pre-art group without OH (56% and 72%, respectively) and the post-art group with OH (31% and 63%, respectively) (Fig. 4). Relationship Between E 2 Levels and Recurrence of Endometriosis The median peak E 2 values were significantly lower in the IUI group than in the groups treated with IVF alone or with both IVF and IUI in different cycles (P.0001; Table 1). FIGURE 3 Pre-ART recurrence study: CERR before treatment with ART. End points were development of endometriosis recurrence, start of treatment with ART, and clinical pregnancy before 20 weeks (analysis 1) or ongoing pregnancy after 20 weeks (analysis 2). D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. FIGURE 4 Comparison of pre-art and post-art recurrence studies: CERR before (pre-art) and after (post- ART) treatment with assisted reproductive technology. For information regarding analysis 2, see Figure 3. D Hooghe. Endometriosis recurrence after ovarian hyperstimulation. Fertil Steril 2006. There was no correlation between the cumulative peak E 2 per patient and the recurrence of endometriosis (P.2230). DISCUSSION Possible pathophysiologic mechanisms that can explain the recurrence of endometriosis include incomplete surgical excision of endometriotic tissue (owing to technical problems, lack of surgical skills/experience, or presence of microscopic endometriosis), the development of new endometriotic lesions, and the growth of microscopic or small subclinical endometriotic lesions, undetectable during surgery, to larger symptomatic and detectable endometriotic disease after surgery, possibly favored by higher exposure to estrogens following ovarian stimulation for ART. To the best of our knowledge, this is the first study evaluating the postoperative endometriosis recurrence rate after surgical treatment for moderate to severe endometriosis in patients who received ovarian hyperstimulation for IUI or IVF. In this study, we could not confirm our hypothesis that the cumulative recurrence rate of endometriosis is higher after cumulative exposure to very high E 2 levels following OH for IVF than after cumulative exposure to not as high E 2 levels following OH for IUI. In contrast to our hypothesis, the cumulative recurrence rate after 21 months of follow-up was significantly lower in the group of patients who had received at least one IVF cycle (17%) compared with the group treated with IUI only (84%). Furthermore, after a follow-up of 36 months, there was little difference between the CERR pre-art (72%) and post- ART (63%). It is possible that OH for IVF does not have an adverse effect on the recurrence of endometriosis. However, it is also possible that women with blocked or dysfunctional tubes due to endometriotic adhesions who are treated with IVF are less 288 D Hooghe et al. Endometriosis recurrence after ovarian hyperstimulation Vol. 86, No. 2, August 2006

likely to develop new endometriotic lesions than patients with open fallopian tubes who are exposed to retrograde menstruation every month. It also cannot be excluded that hyperestrogenism following COH may result in a more pronounced proliferation of functional endometrium, result in a higher volume of shedded endometrium during retrograde menstruation, and represent an independent risk factor for the recurrence of endometriosis in women with open fallopian tubes. Finally, short-term exposure to LHRH analogs during our routine long protocol for IVF may also protect against the recurrence of endometriosis. Indeed, the reversible medical castration achieved by busereline acetate has been shown in open studies to be an effective treatment for endometriosis (9, 10). However, it is well known that recurrence of endometriosis is also common after the cessation of treatment with LHRH analogs: Waller and Shaw (11) reported a cumulative endometriosis recurrence rate of 53.4% after the cessation of LHRH analog treatment. In our pre-art recurrence study, we only included patients with persistence of infertility after initial surgery for endometriosis who therefore started ART. Patients with the same history who became pregnant spontaneously were not recruited in this study. For this reason the study population recruited in our pre-art recurrence study is biased, i.e., not representative of all patients operated upon for moderate to severe endometriosis but rather a selection of patients with moderate to severe endometriosis who failed to conceive spontaneously after surgery and who chose treatment with ART. This bias may partly explain why the maximum pre-art cumulative recurrence rate of 72% at 3 years is higher in our study than according to published data. In other studies, the endometriosis recurrence rates following conservative surgical treatment varied between 2% (12) and 47% (13, 14). In a more recent study (15), among 135 out of 216 women responding to a postoperative questionnaire after laparoscopic excision of minimal to severe endometriosis, 33% required further surgery for suspicion of recurrent disease after primary surgery. During this further surgery, endometriosis recurrence was confirmed in 68% of the patients, but no visual or histologic evidence of resident or recurrent endometriosis was observed in 32% (15). Similary, among 67 of 84 patients responding to a postoperative questionnaire after laparoscopic treatment of complete obliteration of the cul-de-sac associated with endometriosis (16), 23 patients (34%) reported a new surgical (n 13) or hormonal (n 10) treatment. In another study (17), painful symptoms had recurred in 74% of patients within a median time of 20 (range 5 60) months after laser laparoscopy for endometriosis-associated pain. The main limitation of the above-mentioned studies evaluating recurrence of endometriosis is the fact that in none of these studies was histologic or cytologic proof of recurrence obtained, nor was life table analysis performed, and therefore it is impossible to compare these data with the data from our study. Reports that do not provide cytologic or histologic proof of recurrence may overestimate (adhesions without endometriosis) or underestimate (ovarian endometriotic cyst wrongly interpreted to be a functional ovarian cyst) the true endometriosis recurrence rate. Reports that are uncorrected for patients lost to follow-up can underestimate the recurrence rates and therefore it is well accepted that life table analysis is necessary for the calculation of endometriosis recurrence (4, 18, 19). Furthermore, the fact that some patients who did not answer postoperative questionnaires had developed disease recurrence, were not satisfied with the treatment received, and therefore did not reply to these questionnaires cannot be excluded. Only a limited number of studies have assessed the CERR using life table analysis, mostly in patients with rafs I and rafs II endometriosis. Not surprisingly, the CERR in these patients was 19.5% (17) or 19% (18) after 5 years, considerably lower than in our study evaluating recurrence in women with rafs III and rafs IV endometriosis. However, when the authors of the previously mentioned study (15) used life table analysis for the extrapolation of their data, they estimated that the cumulative likelihood for repeat surgery within 5 years was significantly higher in women with rafs scores (2) higher than 70 than in women with rafs scores lower than 70 (about 55% vs. 30%, respectively). Similarly, long-term follow-up after conservative surgery for rectovaginal endometriosis in patients with moderate to severe disease (2) showed a cumulative clinical or sonographic recurrence rate of 34% within 3 years after life table analysis (20). In contrast, the cumulative ultrasonographic recurrence rate of ovarian endometriotic cysts in women with moderate to severe endometriosis (21) was only 12% over 4 years. Taken together, these data suggest that the cumulative recurrence rate of moderate to severe endometriosis can vary between 30% and 60% within 3 to 5 years. In conclusion, the data from our study suggest that patients with moderate to severe endometriosis who do not conceive spontaneously after laparoscopic treatment in a referral center for endometriosis have a high CERR. In patients subsequently treated with ART, the CERR was higher in those treated with IUI only than in those treated with IVF only or both IVF and IUI in different cycles. It is possible that OH for IVF does not have an adverse effect on the recurrence of endometriosis. However, it is also possible that women with blocked tubes, due to endometriotic adhesions, who are treated with IVF are less likely to develop new endometriotic lesions than patients with open fallopian tubes who are treated with IUI and who are exposed to retrograde menstruation every month, and perhaps more so after treatment with OH. Finally, short-term exposure to LHRH analog during our routine long protocol for IVF may also protect against recurrence of endometriosis. Fertility and Sterility 289

REFERENCES 1. American Fertility Society. Revised American Fertility Society classification of endometriosis. Fertil Steril 1985;43:351 2. 2. American Fertility Society. Revised American Society for Reproductive Medicine classification of endometriosis. Fertil Steril 1997;67: 817 21. 3. Punnonen R, Klemi P, Nikkanen V. Recurrent endometriosis. Gynecol Obstet Invest 1980;11:307 12. 4. Wheeler JM, Malinak LR. Recurrent endometriosis: incidence, management, and prognosis. Am J Obstet Gyn 1983;146:247 53. 5. Fedele L, Bianchi S, Di Nola G, Candiani M, Busacca M, Vignali M. The recurrence of endometriosis. Ann N Y Acad Sci 1994;734:358 64. 6. Renier M, Verheyden B, Termote L. An unusual coincidence of endometriosis and ovarian hyperstimulation. Eur J Obstet Gynecol Reprod Biol 1995;63:187 9. 7. Anaf V, El Nakadi I, Simon P, Englert Y, Peny MO, Fayt I, et al. Sigmoid endometriosis and ovarian stimulation. Hum Reprod 2000;15: 790 4. 8. Olive DL. Analysis of clinical fertility trials: a methodologic review. Fertil Steril 1986;45:157 71. 9. Shaw RW, Fraser HM, Boyle H. Intranasal treatment with luteinizing hormone-releasing hormone agonist in women with endometriosis. Br Med J 1983;287:1667 9. 10. Lemay A, Maheux R, Faure N, Jean C, Fazekas ATA. Reversible hypogonadism induced by a luteinizing hormone-releasing hormone (LH-RH) agonist (buserelin) as a new therapeutic approach for endometriosis. Fertil Steril 1984;41:863 71. 11. Waller KG, Shaw RW. Gonadotropin-releasing hormone analogues for the treatment of endometriosis: long-term follow-up. Fertil Steril 1993; 59:511 5. 12. Green RH. Conservative surgical treatment of endometriosis. Clin Obstet Gynecol 1966;9:293 308. 13. Andrews WC. Medical versus surgical treatment of endometriosis. Clin Obstet Gynecol 1980;23:917 24. 14. Andrews WC, Larsen GD. Endometriosis. Treatment with hormonal pseudopregnancy and/or operation. Am J Obstet Gynecol 1974;118: 643 51. 15. Abbott, JA, Hawe J, Clayton RD, Garry R. The effects and effectiveness of laparoscopic excision of endometriosis: a prospective study with 2 5 year follow-up. Human Reprod 2003;18:1922 7. 16. Redwine DB, Wright JT. Laparoscopic treatment of complete obliteration of the cul-de-sac associated with endometriosis: long-term follow-up of en bloc resection. Fertil Steril 2001;76:358 65. 17. Jones KD, Haines P, Sutton CJ. Long-term follow-up of a controlled trial of laser laparoscopy for pelvic pain. JSLS 2001;5:111 5. 18. Wheeler JM, Malinak LR. Recurrent endometriosis. Contrib Gynecol Obstet 1987;16:13 21. 19. Redwine DB. Conservative laparoscopic excision of endometriosis by sharp dissection: life table analysis of reoperation and persistent or recurrent disease. Fertil Steril 1991;56:628 34. 20. Fedele LG, Bianchi S, Zanconato G, Bettoni G, Gotsch F. Long-term follow-up after conservative surgery for rectovaginal endometriosis. Am J Obstet Gynecol 2004;190:1020 1024. 21. Busacca M, Marana R, Caruana P, Candiani M, Muzii L, Calia C, et al. Recurrence of ovarian endometrioma after laparoscopic excision. Am J Obstet Gynecol 1999;180:519 23. 290 D Hooghe et al. Endometriosis recurrence after ovarian hyperstimulation Vol. 86, No. 2, August 2006