Ultrasound diagnosed adenomyosis has a negative impact on successful implantation following GnRH antagonist IVF treatment

Transcription

1 Human Reproduction, Vol.27, No.12 pp , 2012 Advanced Access publication on September 20, 2012 doi: /humrep/des305 ORIGINAL ARTICLE Infertility Ultrasound diagnosed adenomyosis has a negative impact on successful implantation following GnRH antagonist IVF treatment V. Thalluri 1 and K.P. Tremellen 2,3, * 1 Department of Gynaecology, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia 2 Repromed, 180 Fullarton Road, Dulwich, South Australia 5065, Australia 3 School of Pharmacy and Medical Science, University of South Australia, Adelaide, South Australia 5000, Australia *Correspondence address. ktremellen@repromed.com.au Submitted on April 27, 2012; resubmitted on July 16, 2012; accepted on July 18, 2012 study question: Does the presence of ultrasound diagnosed adenomyosis interfere with successful implantation in patients undergoing IVF treatment with GnRH antagonist ovarian stimulation? summary answer: The presence of ultrasound diagnosed adenomyosis was associated with a significant reduction in successful implantation of good quality embryos in patients undergoing GnRH antagonist stimulation for IVF treatment (viable clinical pregnancy rate 23.6% versus 44.6%, P ¼ 0.017). what is known and what this paper adds: There is currently no consensus regarding the impact of adenomyosis on implantation potential. Although some studies have identified alterations in the endometrial milieu in adenomyosis patients that may impact implantation, several papers have reported no associated reproductive deficit. However, these pregnancy outcome studies have primarily investigated patients undergoing long down-regulation IVF protocols, where low levels of serum estrogen (before commencing the ovarian stimulation) may inactivate the adenomyosis and potentially negate its effect on implantation. Given that the majority of fertility clinics are now moving towards the more patient-friendly antagonist protocol, where patients are not placed in a hypo-estrogen state before commencing ovarian stimulation, the question of whether adenomyosis has an impact on IVF success rates in GnRH antagonist-stimulated IVF treatment needs to be examined. design: This is a retrospective cohort study of 748 patients who, between April 2010 and March 2012, underwent a screening transvaginal ultrasound to identify possible pelvic pathology before commencing their IVF treatment. From this screening group, 213 patients were eligible to be included in the study as they had no obvious underlying uterine or embryonic factors that could have interfered with successful implantation (aged 39 years, good quality Day 4/5 embryo for single-embryo transfer, no uterine fibroids/hydrosalpinx or endometrial polyps). participants and setting: There were 213 patients in a private IVF unit eligible to be included in the study, with 38 patients (17.84%) having ultrasound diagnosed adenomyosis and 175 patients having no adenomyosis on the scan. Only the first treatment cycle for each patient was included. main results and the role of chance: The adenomyosis group had a viable clinical pregnancy rate of 23.6% compared with 44.6% in the non-adenomyosis group (P ¼ 0.017). However, the median maternal age and duration of infertility of the adenomyosis group was 2 years older and 4 months greater, respectively, than that of the non-adenomyosis group. A logistic regression analysis was performed to account for these differences between the two groups, with the adjusted results still showing a statistically significant decline in viable pregnancy rate in the adenomyosis group (OR ¼ 0.408, CI ¼ , P ¼ when adjusting for maternal age; OR ¼ 0.417, CI ¼ , P ¼ when adjusting for duration of infertility) bias, confounding and other reasons for caution: Given the retrospective nature of this study, there is risk of bias. This risk was minimized by having subjective variables such as embryo quality assessed by individuals not involved in the study, while strictly applying the pre-determined inclusion/exclusion criteria to all study participants. Furthermore, it is acknowledged that ultrasound is & The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oup.com

2 3488 Thalluri and Tremellen not a perfect test for the diagnosis of adenomyosis and, therefore, may underestimate the incidence of adenomyosis by misclassifying some patients with mild adenomyosis as not affected. wider implications of the findings: The results of this study should be representative of outcomes for any patient undergoing a GnRH antagonist ovarian stimulation cycle for IVF since standard IVF treatment protocols were used. study funding/competing interest: MSD Australia have provided us with a small amount of funding to cover our costs (including a travel grant for Dr Thalluri to present this work at a conference). Key words: adenomyosis / infertility / implantation / GnRH antagonist / in-vitro fertilization Introduction Adenomyosis is defined as the benign invasion of ectopic endometrium into the myometrium with hyperplasia of the adjacent smooth muscle (Ferenczy, 1998). Adenomyosis has traditionally been thought to present with heavy, painful menstrual bleeding in the fourth and fifth decades of life. Histological analysis of hysterectomy samples taken from women who no longer wish to retain their fertility suggests that adenomyosis affects around one in five women by their forties (Larsen et al., 2011), with risk factors for the development of adenomyosis including advancing age, multiparity, early menarche, obesity and previous uterine surgery (Templeman et al., 2008). In patients wishing to preserve their fertility, histological diagnosis of adenomyosis following hysterectomy is clearly not an option. Given this difficulty, and since multiparity has been established as a risk factor for adenomyosis, it has been generally believed that adenomyosis is not a cause of infertility. Furthermore, adenomyosis is currently not considered a widely accepted cause for implantation failure during IVF treatment (Maheshwari et al., 2012), with the vast majority of gynaecologists not screening or treating the condition before embarking on IVF/ICSI cycles (Tan et al., 2005). However, advancements in magnetic resonance imaging and high-quality transvaginal ultrasound now allow for an accurate, non-invasive diagnosis of adenomyosis to be made. A recent meta-analysis of 14 trials and 1985 participants reported the sensitivity and specificity of ultrasound diagnosed adenomyosis to be as high as 82.5 and 84.6%, respectively (Meredith et al., 2009). This has resulted in a number of recent studies examining the effect of adenomyosis on IVF implantation success rates. Some observational studies have shown a negative effect on implantation rates and increase in miscarriage risk (Kunz et al., 2005; Piver, 2005; Maubon et al., 2010; Martinez-Conejero et al., 2011; Tremellen and Russell, 2011; Youm et al., 2011; Ballester et al., 2012), whereas others have shown no reproductive deficit at all (Camargo et al., 2001; Shimizu et al., 2002; Mijatovic et al., 2010; Costello et al., 2011). A recent systematic review concluded that further studies are still required to determine the true impact of adenomyosis on IVF success rates (Maheshwari et al., 2012). There are good biological reasons to suspect that adenomyosis may have the potential to impair the implantation of good quality embryos transferred during IVF treatment. A recent observational study clearly linked adenomyosis diagnosed on magnetic resonance imaging (MRI) with an increase in macrophage and natural killer cells in the endometrium of women experiencing infertility (Tremellen and Russell, 2012). Macrophages are known to release chemicals which may be detrimental to embryos, such as cytokines tumour necrosis factor a (TNFa) and interferon g (IFNg), plus harmful reactive oxygen species ( free radicals ). Earlier studies have linked the presence of adenomyosis with increased free radical concentrations in the endometrium (Yang et al., 2004), providing a possible mechanism by which adenomyosis may impair implantation and cause miscarriage. Previous studies examining the effect of adenomyosis on IVF outcomes have all used long down-regulation protocols, not the GnRH antagonist protocol, which is now becoming the standard IVF treatment due to its improved patient side-effect profile (Verberg et al., 2008). Long down-regulation IVF produces a period of estrogen deficiency that may temporarily inactivate adenomyosis and normalize some of the distorted endometrial functions seen in the condition. This may explain the previously published observations of no effect of adenomyosis on embryo implantation. This period of potentially therapeutic estrogen deficiency does not occur in GnRH antagonist cycles. The question of whether adenomyosis impairs embryo implantation in the GnRH antagonist cycle setting has yet to be investigated. Thus, the purpose of this retrospective cohort study was to examine the differences in IVF outcomes between patients with or without ultrasound diagnosed adenomyosis undergoing a GnRH antagonist stimulated IVF cycle, which is the primary default mode of IVF treatment used in our fertility unit. Materials and Methods This retrospective cohort study included women who had a baseline pelvic ultrasound before their first IVF cycle and fit our strict inclusion/exclusion criteria. It is the protocol in our unit for all women undergoing IVF for the first time to undergo a baseline pelvic ultrasound within 6 months of commencing treatment in order to exclude pelvic pathology that may impact on IVF success (endometrial polyps, fibroids, hydrosalpinx and endometrioma). Patients were classified as having adenomyosis or not based on traditional radiological criteria such as the presence of an enlarged globular uterus in the absence of fibroids, asymmetric thickening of the anterior or posterior myometrial wall, heterogeneous poorly circumscribed areas within the myometrium, anechoic myometrial blood filled lacunae or cysts of varying sizes, increased echo-texture of the endometrium and subendometrial linear striations (Meredith et al., 2009). Inclusion criteria included the transfer of a solitary good quality embryo on Day 4/5 (Grade 1 or 2). Embryos were graded according to the Gardner and Schoolcraft (1999) classification of embryos, with the results then being condensed to a summary grade of excellent (Grade 1), good (Grade 2), poor (Grade 3) or very poor (Grade 4). It is rare in our unit for Grade 4 embryos to successfully implant, yet many Grade 3 embryos do result in viable pregnancy. However, studies have suggested that high-quality embryos have the lowest aneuploidy rates (Alfarawati et al., 2011) and therefore the best implantation potential. As such, we limited our analysis to only include Grade 1 and 2 embryos. This corresponded to the

3 Effects of adenomyosis on implantation 3489 development of at least a compact morula on Day 4 and a well-formed blastocyst on Day 5 of culture. Embryologists grading the embryos were unaware of the patient s adenomyosis status. Further inclusion criteria were the use of GnRH antagonist for ovarian stimulation, single-embryo transfer, maternal age 39 years and the patient having had a high-quality transvaginal ultrasound in the 6 months prior to the index IVF cycle. Exclusion criteria included maternal age 40 years, patients who underwent long down-regulation ovarian stimulation, transfer of a poor quality embryo on Day 4/5 (Grade 3 or 4), patients who had more than one embryo transferred, women who were the recipient of donor oocytes or donor embryos, women on their second or subsequent IVF cycle and women who had a pelvic ultrasound that showed the presence of uterine fibroids, hydrosalpinx or endometrial polyps (all known causes of implantation failure). The GnRH antagonist treatment protocol has been the primary mode of ovarian stimulation for IVF treatment used in our clinic since 2007 and now accounts for more than 95% of all IVF cycles. The standard gonadotrophin starting dose is 150 IU of recombinant FSH (Puregon, MSD or Gonal F, Merck Serono) in women under 36 years of age and IU in the year-old age group. All stimulation starts on Day 2 or 3 of the menstrual cycle and a GnRH antagonist is commenced on Day 5 of stimulation (ganarelix 0.25 mg, MSD). A pelvic ultrasound and serum hormone assessment is performed on Day 8 or 9 of stimulation and oocyte maturation is triggered when three or more follicles of 17 mm of diameter are present. A delay or advance in the day of trigger of 1 day from ideal is used to avoid weekend oocyte collections, as has been previously published by our group (Tremellen and Lane, 2010). The traditional trigger used for oocyte maturation in the vast majority of patients attending our unit is 250 mg of rhcg (Ovidril, Merck Serono), with an agonist trigger only being used in women deemed at a high risk of developing ovarian hyperstimulation syndrome. Fertilization was undertaken using the standard protocols and all embryos were cultured for 4 5 days before being transferred or cryopreserved. Luteal support (1500 IU hcg twice or estradiol valerate plus vaginal progesterone Crinone) Table I Clinical characteristics of study participants. was continued until the onset of menses or a positive pregnancy test result. Biochemical pregnancy was defined as positive b-hcg.40 IU on Day 19 post-oocyte retrieval. Clinical pregnancy was defined as the presence of at least one intrauterine gestational sac on the 8-week gestation ultrasound. Viable clinical pregnancy was defined as the presence of a fetal heart beat at the 8-week ultrasound. Statistical analysis was performed using the Graphpad Prism 5 software (San Diego, USA). None of the continuous variables were normally distributed, and therefore, results were expressed as medians with interquartile (25th 75th percentile) ranges and were analysed by the non-parametric Mann Whitney test. Categorical variables such as infertility diagnosis were analysed by a Fisher exact test. Logistic regression analysis was conducted by an independent, professional statistician using the SAS version 9.2 (2008, NC, USA) software. The differences between comparison groups were considered to be statistically significant at P, Results Between the period of April 2010 and March 2012, 748 baseline ultrasound scans were performed prior to the commencement of IVF. A total of 213 patients met the criteria to be included in the study, with the majority of patients being excluded due to advanced maternal age, embryo quality restrictions, having had previous IVF treatment cycles and the presence of fibroids on ultrasound. Of the 213 women who met the study entry criteria, 38 individuals (17.84%) had ultrasound evidence of adenomyosis. The baseline characteristics of the study subjects are outlined in Table I. The two characteristics that were significantly different between the study groups were maternal age and duration of infertility. The adenomyosis group had a median maternal age of 35 years compared with 33 years in the control group. The duration of infertility was also greater by 4 months in the adenomyosis Adenomyosis positive Adenomyosis negative P-value... Number Maternal age (years) 35 ( ) 33 (30 36) Duration of infertility (months) 18 (12 48) 14 ( ) Gravida 0 (0 1) 0 (0 1) Parity 0 (0 1) 0 (0 0) 0.39 BMI 23.8 ( ) 24 ( ) Antral follicle count (2 10 mm) 15 (11 26) 18 ( ) Uterine volume (ml) 45 ( ) 43.5 ( ) Aetiology of infertility [n (%)] Male 17 (44.7) 72 (41.1) Tubal 3 (7.9) 13 (7.4) Anovulation 2 (5.3) 12 (6.9) Endometriosis 1 (2.6) 4 (2.3) Mixed 9 (23.7) 47 (26.9) Unexplained 3 (7.9) 19 (10.8) Other 3 (7.9) 8 (4.6) All data are expressed as medians (interquartile range).

4 3490 Thalluri and Tremellen group than in the non-adenomyosis group. The IVF outcomes are shown in Table II. There were no significant differences in the IVF stimulation outcomes (oocytes retrieved, estrogen and progesterone levels on Day 9 of stimulation), nor the embryology outcomes of successful fertilization or embryo quality between the two groups. In the adenomyosis group, 12 of the 38 patients (31.6%) showed biochemical evidence of pregnancy, with 9 of these pregnancies (23.6%) being viable at the 8-week gestation ultrasound, 1 being non-viable and the remaining 2 miscarrying at the biochemical pregnancy stage. In the control non-adenomyosis group, 87 women (49.7%) had biochemical evidence of implantation, with 78 having a viable singleton pregnancy (44.6%) on the 8-week ultrasound, 3 had a non-viable intrauterine pregnancy and 1 had an ectopic pregnancy, and the remaining 5 experienced an early biochemical pregnancy loss. During the period in which the study subjects had IVF treatment, the IVF success rates in our clinic for female patients aged under 38, with no exclusions for poor embryo quality or repeat IVF cycle number and an average of 1.17 embryos transfered, were a biochemical pregnancy rate of 47% and a clinical pregnancy rate of 39%. Given that age is an important prognostic variable in IVF, a logistic regression analysis was used to account for the difference in maternal age between the adenomyosis and non-adenomyosis groups. The unadjusted odds ratio for clinical pregnancy showed that the adenomyosis group had significantly lower odds of viable pregnancy (OR ¼ 0.386, CI ¼ , P ¼ 0.039). When accounting for age in the model, the negative effect of adenomyosis on viable pregnancy remained, with the adenomyosis group still showing a lower odds of a viable pregnancy relative to the non-adenomyosis group (OR ¼ 0.408, CI ¼ , P ¼ 0.031). The same analysis was performed with duration of infertility as a co-variable, and the results again showed a statistically significant reproductive deficit in the adenomyosis group (OR ¼ 0.417, CI ¼ , P ¼ 0.047). Thus, even after accounting for the small differences in age and duration of infertility, the adjusted odds ratio showed that the non-adenomyosis group had twice the chance of a single-embryo transfer producing a viable pregnancy than did the adenomyosis group. Similar results were obtained when looking at biochemical evidence of pregnancy, with the adenomyosis group showing less than half the odds of a biochemical pregnancy compared with the non-adenomyosis group (OR ¼ 0.456, CI ¼ , P ¼ 0.039). The results remained statistically significant when adjusted for age (OR ¼ 0.447, CI ¼ , P ¼ 0.037) or duration of infertility (OR ¼ 0.436, CI ¼ , P ¼ 0.045). Discussion While our group has previously described an adenomyosis-related implantation deficit in a small case series of women undergoing GnRH antagonist stimulation for IVF treatment (Tremellen and Russell, 2011), this is the first study to clearly describe an implantation problem in a relatively large cohort exclusively undergoing GnRH antagonist cycles in women with ultrasound diagnosed adenomyosis. This deficit is clearly apparent with the almost halving of the viable pregnancy rate (44.6% versus 23.6%, P ¼ 0.017), and the trend towards an increased biochemical and clinical miscarriage rate. The results of this study are in direct contrast to some previously published studies, but agree with others (Camargo et al., 2001; Shimizu et al., 2002; Piver, 2005; Maubon et al., 2010; Mijatovic et al., 2010; Costello et al., 2011; Martinez-Conejero et al., 2011; Tremellen and Russell, 2011; Youm et al., 2011; Ballester et al., 2012). Maubon et al. (2010) was the first to conclusively link MRI defined adenomyosis with a significant decrease in IVF pregnancy rates. A recently published multicentre study by Ballester et al. (2012) looking at pregnancy rates in patients with colorectal endometriosis reported that the additional presence of adenomyosis diagnosed by MRI scanning was a major negative factor on pregnancy rates. The study found that the cumulative pregnancy rate following three cycles of IVF/ICSI in the group with adenomyosis was 19% compared with 82.4% in the group without adenomyosis (Ballester et al., 2012). In previous studies that have showed no negative effect of adenomyosis on implantation rates, a long downregulation ovarian stimulation protocol rather than a GnRH antagonist protocol was used as the primary mode of treatment. Costello et al. (2011) performed a retrospective cohort study of 201 patients undergoing long down-regulation ovarian stimulation and reported no difference in pregnancy outcomes between those women with or without ultrasound defined adenomyosis. These authors acknowledged that their failure to observe an implantation deficit may have been because a long down-regulation-related hypo-estrogenic state leading up to the ovarian stimulation could have normalized any endometrial disturbance present in the adenomyosis group. Interestingly, their reported incidence of adenomyosis (18.4%) was very similar to ours (17.8%), which also corresponds similarly to the figure reported by Larsen et al. Table II Clinical outcomes for study participants. Adenomyosis positive (n 5 38) Adenomyosis negative (n 5 175) P-value... Number of oocytes retrieved (mature) 8 (4 10) 8 (5 12) 0.18 Fertilization rate (%) 66.7 ( ) 66.7 ( ) Embryo grade (G1 excellent G4 very poor) 1 (1 2) 1 (1 2) 0.37 Estradiol day 9/10 of cycle (pmol/l) 2100 ( ) 3200 ( ) Progesterone day 9/10 of cycle (nmol/l) 1.8 ( ) 2.1 ( ) Biochemical evidence of pregnancy 12/38 (31.6%) 87/175 (49.7%) Clinical (viable) pregnancy rate 9/38 (23.6%) 78/175 (44.6%) Non-viable pregnancy rate (miscarriage and ectopic) 3/12 (25%) 9/87 (10.3%) All data are expressed as median (interquartile range) or as percentages. Viable pregnancy is assessed as fetal heart seen on scanning.

5 Effects of adenomyosis on implantation 3491 (19.4%). Given that it is the protocol in our unit for all women to have a screening baseline pelvic ultrasound prior to commencing IVF treatment, not just those women reporting symptoms suggestive of pelvic pathology, our finding of an incidence of ultrasound diagnosed adenomyosis of 17.8% is a true, non-biased assessment of how common adenomyosis is in a general infertile population. Mijatovic et al. has also reported no impact of adenomyosis on IVF pregnancy outcomes; however, this study used an ultra long down-regulation stimulation protocol. All patients had at least 3 months of GnRH agonist use prior to IVF/ICSI treatment, with the mean period of down-regulation being 5.35 months (Mijatovic et al., 2010). It has already been well established that a period of hypo-estrogenism associated with long downregulation may normalize the endometrial pathology associated with adenomyosis (Khan et al., 2010) and potentially return implantation rates to normal (Tremellen and Russell, 2011). This of course does not happen in a GnRH antagonist cycle when stimulation starts immediately following the commencement of menses. A 3-month period of down-regulation has previously been reported to be a successful treatment for adenomyosis-related implantation failure by our group (Tremellen and Russell, 2011). Furthermore, both the Maubon et al. (2010) and Costello et al. (2011) studies showed increased rates of adenomyosis in the unexplained infertility subgroups, supporting the possibility that adenomyosis may impair successful conception by impeding implantation in those individuals with no obvious male factor, ovulatory or tubal cause for infertility. Finally, the reported observations of increased macrophages and natural killer cells and reactive oxygen species in the endometrium of adenomyosis patients perhaps helps explain the biochemical mechanisms behind the observed implantation deficit (Yang et al., 2004; Tremellen and Russell, 2012). The strengths of our study are the relatively large sample size and the use of high-quality transvaginal ultrasound scans performed by sonographers and reported by radiologists who have an interest and experience in gynaecological imaging. It has been our experience that the ultrasound diagnosis of adenomyosis is often missed by sonographers and radiologists who do not have an interest in fertility and are possibly unaware of its potential importance to the treating gynaecologist. A further strength of our study was that all patients underwent fertility treatment at a single IVF unit, with the strict inclusion criteria of only including patients with a good IVF prognosis (young, good quality embryos), thus allowing a better assessment of true endometrial implantation potential. The limitations of our study include the fact that it is a retrospective study, and therefore, open to bias. Only 213 of the 748 patients screened were found to be eligible for the study raising the possibility of selection bias. However, before commencing the study, we developed a very strict inclusion and exclusion criteria that was rigorously applied to the available study cohort to minimize bias. Furthermore, the assessment of embryo quality was performed by embryologists not directly involved in this study, nor aware of the clinical diagnosis of adenomyosis, making exclusion based on poor embryo quality relatively easy and not subject to bias. In addition, it was found that the adenomyosis group was 2 years older than the control group. This is not surprising given that adenomyosis becomes more prevalent with advancing age (Kunz et al., 2005, 2007). As it is well recognized that maternal age has a negative effect on IVF pregnancy rates, it is possible that the decrease in pregnancy rates related to adenomyosis may simply reflect the older age of this group. However, even when controlling for this age confounder using regression analysis, the differences in clinical pregnancy rates between the adenomyosis and the non-adenomyosis group was still significant (P ¼ 0.031). Furthermore, this small difference in age did not translate into any difference in the number of oocytes retrieved or the quality of the embryos transferred, thereby making the difference in age of unlikely significance. Logistic regression was also used to account for the difference in duration of infertility between the two groups. Again, even when accounting for this variable, a significant difference in the reproductive outcome was found between the two groups (P ¼ 0.047). A final weakness of the study is that ultrasound is not a perfect test for the presence or the absence of adenomyosis. When using histological diagnosis at hysterectomy as the gold standard for diagnosis, a recent meta-analysis of 14 studies by Meredith et al. (2009) showed that transvaginal ultrasound had 82.5% sensitivity and 84.6% specificity in diagnosing adenomyosis. Given that our radiologists and sonographers used the standard criteria for the diagnosis of adenomyosis, we would expect our sensitivity and specificity to be in this range. MRI may have some advantages over ultrasound in terms of sensitivity and specificity in picking up adenomyosis (Valentini et al., 2011); however, given the availability and low cost of ultrasound, we agree with Meredith et al. that ultrasound likely represents the most practical diagnostic test for adenomyosis. Our surprising observation of no significant difference in uterine volume between the adenomyosis and non-adenomyosis groups may be explained by the fact that any woman with adenomyosis and co-existing fibroids was excluded from our study. Given that adenomyosis and fibroids often co-exist, the exclusion of these patients with definite enlarged uteri may have artificially reduced the overall uterine volume in the adenomyosis group. It is generally well accepted that uterine volume on its own is a relatively insensitive and nonspecific marker for the presence of adenomyosis. Brosens et al. (1995) was the first to report that myometrial heterogeneity, not uterine size, is the most sensitive and specific marker for adenomyosis. In conclusion, this study is the first to show a clear implantation deficit in a cohort of ultrasound diagnosed adenomyosis patients exclusively undergoing GnRH antagonist ovarian stimulation for IVF cycles. Although we must wait for the outcome of prospective randomized control trials comparing GnRH antagonist with long-down-regulation stimulation, we would suggest that it may be prudent to consider using long down-regulation ovarian stimulation as the preferred IVF treatment in adenomyosis patients, especially if they have already failed several GnRH antagonist cycles. We base this advice on the fact that the majority of reported studies have shown no adenomyosis-related implantation deficit when long downregulation IVF is used. We hope this study will encourage the conduct of prospective randomized controlled studies comparing pregnancy outcomes between GnRH antagonist and long down-regulation ovarian stimulation protocols for IVF treatment in an adenomyosis infertile population. Acknowledgements We would like to thank Nancy E. Briggs, Senior Statistician, Data Management and Analysis Centre, Discipline of Public Health, University of Adelaide, for performing the logistic regression analysis for this study.

6 3492 Thalluri and Tremellen Authors roles V.T. and K.P.T. were involved in the design of the study and acquisition of data for the study. K.P.T. performed the statistical analysis with interpretation of the data by both V.T. and K.P.T. V.T. wrote the manuscript with revision and critical appraisal by K.P.T. Funding This project was supported by an educational grant from MSD Australia. Conflict of interest None declared. References Alfarawati S, Fragouli E, Colls P, Stevens J, Gutierrez-Mateo C, Schoolcraft WB, Katz-Jaffe MG, Wells D. The relationship between blastocyst morphology, chromosomal abnormality, and embryo gender. Fertil Steril 2011;95: Ballester M, d Argent EM, Morcel K, Belaisch-Allart J, Nisolle M, Daraï E. Cumulative pregnancy rate after ICSI-IVF in patients with colorectal endometriosis: results of a multicentre study. Hum Reprod 2012; 27: Brosens JJ, de Souza NM, Barker FG, Paraschos T, Winston RM. Endovaginal ultrasonography in the diagnosis of adenomyosis uteri: identifying the predictive characteristics. Br J Obstet Gynaecol 1995; 102: Camargo F, Gaytan J, Caligara C, Simon C, Pellicer A, Remohi J. Impact of ultrasound diagnosis of adenomyosis on recipients of sibling oocytes. Fertil Steril 2001;76:S150. Costello M, Lindsay K, McNally G. The effect of adenomyosis on in vitro fertilisation and intra-cytoplasmic sperm injection treatment outcome. Eur J Obstet Gynaecol Reprod Biol 2011;158: Ferenczy A. Pathophysiology of adenomyosis. Hum Reprod Update 1998; 4: Gardner DK, Schoolcraft WB. Vitro culture of the human blastocyst. In: Jansen R, Mortimer D (eds). Towards Reproductive Certainty: Infertility and Genetics Beyond Carnforth, UK: Parthenon Publishing, 1999, Khan KN, Kitajima M, Hiraki K, Fujishita A, Sekine I, Ishimaru T, Masuzaki H. Changes in tissue inflammation, angiogenesis and apoptosis in endometriosis, adenomyosis and uterine myoma after GnRH agonist therapy. Hum Reprod 2010;25: Kunz G, Beil D, Huppert P, Noe M, Kissler S, Leyendecker G. Adenomyosis in endometriosis-prevalence and impact on fertility. Evidence from magnetic resonance imaging. Hum Reprod 2005;20: Kunz G, Herbertz M, Biel D, Huppert P, Leyendecker G. Adenomyosis as a disorder of the early and late human reproductive period. Reprod Biomed Online 2007;15: Larsen SB, Lundorf E, Forman A, Dueholm M. Adenomyosis and junctional zone changes in patients with endometriosis. Eur J Obstet Gynecol Reprod Biol 2011;157: Maheshwari A, Gurunath S, Fatima F, Bhattacharya S. Adenomyosis and subfertility: a systematic review of prevalence, diagnosis, treatment and fertility outcomes. Hum Reprod 2012;18: Martinez-Conejero J, Morgan M, Monstesinos M, Fortuño S, Meseguer M, Simón C, Horcajadas JA, Pellicer A. Adenomyosis does not affect implantation, but is associated with miscarriage in patients undergoing oocyte donation. Fertil Steril 2011;96: Maubon A, Faury A, Kapella M, Pouquet M, Piver P. Uterine junctional zone at magnetic resonance imaging: a predictor of in vitro fertilization implantation failure. J Obstet Gynaecol Res 2010;36: Meredith SM, Sanchez-Ramos L, Kaunitz AM. Diagnostic accuracy of transvaginal sonography for the diagnosis of adenomyosis: systematic review and metaanalysis. AM J Obstet Gynecol 2009;201:107.e1 107.e6. Mijatovic V, Florijn E, Halim N, Schats R, Hompes P. Adenomyosis has no adverse effects on IVF/ICSI outcomes in women with endometriosis treated with long term pituitary down regulation before IVF/ICSI. Eur J Obstet Gynecol Reprod Biol 2010;151: Piver P. Uterine factors limiting ART coverage. J Gynecol Obstet Biol Reprod (Paris) 2005;34:5S30 5S33. Shimizu Y, Fukuda J, Kawamura K, Tanaka T. Retrospectively analysis of the fertility of adenomyosis and the outcome of adenomyosis complicated pregnancy in in vitro fetrilization patients. Fertil Steril 2002;77:S50. Tan BK, Vandekerckhove P, Kennedy R, Keay SD. Investigation and current management of recurrent IVF treatment failure in the UK. BJOG 2005;112: Templeman C, Marshall SF, Ursin G, Horn-Ross PL, Clarke CA, Allen M, Deapen D, Ziogas A, Reynolds P, Cress R et al. Adenomyosis and endometriosis in the California teachers study. Fertil Steril 2008; 90: Tremellen KP, Lane M. Avoidance of weekend oocyte retrievals during GnRH antagonist treatment by simple advancement or delay of hcg administration does not adversely affect IVF live birth outcomes. Hum Reprod 2010;25: Tremellen K, Russell P. Adenomyosis is a potential cause of recurrent implantation failure during IVF treatment. Aust N Z J Obstet Gynaecol 2011;51: Tremellen KP, Russell P. The distribution of immune cells and macrophages in the endometrium of women with recurrent reproductive failure. II: adenomyosis and macrophages. J Reprod Immunol 2012;93: Valentini AL, Speca S, Gui B, Soglia G, Micco M, Bonomo L. Adenomyosis: from the sign to the diagnosis. Imaging, diagnostic pitfalls and differential diagnosis: a pictorial review. Radiol Med 2011; 116: Verberg MF, Eijkemans MJ, Heijnen EM, Broekmans FJ, de Klerk C, Fauser BC, Macklon NS. Why do couples drop-out from IVF treatment? A prospective cohort study. Hum Reprod. 2008;23: Yang JH, Chen MJ, Chen HF, Lee TH, Ho HN, Yang YS. Decreased expression of killer cell inhibitory receptors on natural killer cells in eutopic endometrium in women with adenomyosis. Hum Reprod. 2004;19: Youm HS, Choi YS, Han HD. In vitro fertilization and embryo transfer outcomes in relation to myometrial thickness. J Assist Reprod Genet 2011;28: