Correlation of adenomyosis features to live birth rates after the first IVF/ICSI treatment, when using the revised Morphological Uterus Sonographic Assessment group definitions

Sara Alson, Stefan R. Hansson, Emir Henic, and Povilas Sladkevicius planned the research protocol. Sara Alson performed all the ultrasound examinations and collected and analyzed the data. Sara Alson drafted the first version of the manuscript which was discussed with Stefan R. Hansson, Emir Henic, and Povilas Sladkevicius whereafter appropriate changes were made. Sara Alson wrote the final version of the manuscript, which was approved by all the co‐authors.

The study was approved by the Regional Ethical Review Board of Lund University, Lund, Sweden, on September 11, 2018, with a reference number 2018/555. Informed, written consent was obtained from all participants.

This study was supported by regional research grants from Region Skåne, Sweden.

Out of 1160 women potentially eligible to participate in the study, in total 123 women were excluded. Of these women, in total 103 never started ART and another 20 women were planned for intrauterine insemination instead of IVF/ICSI and therefore were excluded. The remaining 1037 women were included in the study (Figure  1 ). Flow chart demonstrating the process of patient inclusion to the study. TVUS, transvaginal ultrasonography; ART, assisted reproductive treatment; IUI, intrauterine insemination; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection. Numbers are presented as percentage of the total population of 1037 women. In total, 102 (9.8%; 95% CI, 8.3%–11.4%) women had at least one direct feature of adenomyosis. In total, 290 (28.0%; 95% CI, 27.7–30.3) women had at least one indirect feature and 188 (18.1%; 95% CI, 16.2–20.1) women had at least one indirect feature without any direct features. In 747 (72.0%; 95% CI, 69.3–74.8) women, neither direct nor indirect features of adenomyosis were detected. The mean (SD) age was 31.9 (3.9) years. Women with at least one direct feature of adenomyosis were older than women without, p  < 0.001. Background characteristics for women included in the study are presented in Table  1 . Background characteristics of women at inclusion in the study and for women with vs without at least one direct feature of adenomyosis at transvaginal ultrasonography. Note : Values are given as n (%) of each group of women, mean (SD) or median (interquartile range). Abbreviations: ART, assisted reproductive treatment; BMI, body mass index; VUS, transvaginal ultrasonography. Mixed = male and female factors. Including women with polycystic ovarian syndrome. Other = same‐gender couples or single woman. The median, (IQR) serum anti‐Müllerian hormone (pmol/L), and antral follicle count were lower in women with at least one direct or indirect feature of adenomyosis compared to women without any feature of adenomyosis, Table  2 . Ovarian reserve and IVF/ICSI treatment in the total cohort of 1037 women and for women with vs without at least one direct feature of adenomyosis at ultrasonography. Note : Numbers are given as n (%), mean (±SD) or median (interquartile range). The chi‐square test or Fischer's exact test was used to compare proportions between groups. Continuous variables were compared between groups using the Mann–Whitney U ‐test for skewed data and Student's t ‐test for normally distributed data. p  < 0.05 was considered indicating statistically significant difference. Abbreviations: AFC, antral follicle count; AMH, anti‐müllerian hormone; ET, embryo transfer; FET, frozen embryo transfer; FSH, follicle‐stimulating hormone; GnRH, gonadotropin‐releasing hormone; GQE, good quality embryo; ICSI, intracytoplasmic sperm injection; IU, international units; IVF, in vitro fertilization; Nb, number; OPU, ovum pick‐up. No GQE despite normal fertilization. In total, 231 (22.3%, [95% CI, 19.7–24.8]) women had endometriosis. Concomitant endometriosis was more common in women with direct or indirect features of adenomyosis compared to women without any feature, Tables  2 and 4 . In total, 158 (15.2%) women had myomas, with a median (IQR) size of 14 (21) mm, Table  1 . Out of 1037 started IVF/ICSI cycles, in total 822 (79.3%, [95% CI, 76.8–81.7]) cycles resulted in at least one GQE. A fresh ET was performed in 763 (73.6%, [95% CI 70.9–76.3]) women, Table  2 . For 62 (6.0%, [95% CI 4.5–7.4]) women, all embryos were frozen. In total 257/1037 (24.8%, [95% CI 22.2–27.4]) women had 399 cycles of FET. Two IVF cycles were converted to intrauterine insemination due to an inadequate number of developing follicles. Data on the IVF/ICSI treatment is presented in Tables  2 and 3 . Outcome after first IVF/ICSI treatment in the total cohort of 1037 women and for women with vs without direct features of adenomyosis at ultrasonography. Note : Values are given as n (%). A modified Poisson regression analysis was performed to estimate RR, which is presented with 95% confidence intervals. Abbreviations: CI, confidence intervals; CLBR, cumulative live birth rate; ET, embryo transfer; FET, frozen embryo transfer; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilization; LBR, live birth rate; Nb, number; PR, pregnancy rate; RR, relative risk. Adjustments were made for age. Calculated per fresh cycle. Pregnancy loss was defined as extrauterine pregnancies and miscarriages before 22 gestational weeks. Calculated as % of pregnancies after fresh ET or FET. One woman had an intrauterine fetal death at >22 gestational weeks. Natural modified cycle with Letrozole. Of women with stimulated cycle, in total 15 women had several FET: s, first in natural and thereafter in stimulated cycle. Calculated as % of FET cycles. Cumulative PR/ pregnancy loss is calculated per woman. Some women may have had pregnancy or pregnancy loss after fresh as well as frozen ET. The CLBR after the first IVF/ICSI cycle in the total cohort of women was 424 (40.9%, [95% CI, 37.9–43.8]). Women with at least one direct feature of adenomyosis had a lower CLBR, 25/102 (24.5%, [95% CI, 17.5–31‐5]) compared to women without any direct feature, 399/935 (42.7%, [95% CI, 39.5–45.8]), p  < 0.001. The crude RR for CLB after the first IVF/ICSI cycle for women with at least one direct feature of adenomyosis compared to women without any direct feature of adenomyosis was 0.58, (95% CI, 0.41–0.82), p  = 0.002. After adjustments were made for age, the adjusted RR for CLB was 0.62 (95% CI, 0.43–0.88), p  = 0.007, Table  3 . Women with at least one direct feature of adenomyosis had a higher risk for miscarriage after FET than women without, adjusted RR 2.88 (95% CI, 1.49–5.57), p  = 0.002, Table  3 . Data regarding background characteristics and IVF/ICSI treatment for women with or without indirect features, excluding women with direct features, are presented in Supporting information Table  S1 and Table  4 , respectively. In this group, the crude RR for CLB for women with at least one indirect feature was 0.57 (95% CI, 0.45–0.73), p  < 0.001. The adjusted RR for CLB for this group of women was 0.58 (95% CI, 0.45–0.75), p  < 0.001. The results after the first IVF/ICSI treatment for women without direct features and with or without indirect features of adenomyosis are presented in Table  5 . Results for women with or without any feature of adenomyosis are presented in Table  6 . The aRR for CLB per transfer for women with any feature was 0.64 (95% CI, 0.52–0.78), p  < 0.001,. IVF/ICSI treatment outcomes, stratified for women with or without endometriosis, are presented in Table  7 . Ovarian reserve and IVF/ICSI treatment in women without any direct features of adenomyosis and for women with vs without any indirect feature of adenomyosis at ultrasonography. Note : Numbers are given as n (%), mean (±SD) or median (interquartile range). The chi‐square test or Fischer's exact test was used to compare proportions between groups. Continuous variables were compared between groups using the Mann–Whitney U ‐test for skewed data and Student's t ‐test for normally distributed data. p  < 0.05 was considered indicating statistically significant difference. Abbreviations: AFC, antral follicle count; AMH, anti‐müllerian hormone; ET, embryo transfer; FET, frozen embryo transfer; FSH, follicle‐stimulating hormone; GnRH, gonadotropin‐releasing hormone; GQE, good quality embryo; ICSI, intracytoplasmic sperm injection; IU, international units; IVF, in vitro fertilization; Nb, number; OPU, ovum pick‐up. No GQE despite normal fertilization. Outcome after first IVF/ICSI treatment in women without direct features of adenomyosis and for women with vs without indirect features of adenomyosis at ultrasonography. Note : Values are given as n (%). A modified Poisson regression analysis was performed to estimate RR, which is presented with 95% confidence intervals. Abbreviations: CI, confidence intervals; CLBR, cumulative Live birth rate; ET, embryo transfer; FET, frozen embryo transfer; ICSI, intracytoplasmic sperm injection; IVF, in vitro Fertilization; LBR, live birth rate; Nb, number; PR, pregnancy rate, RR, relative risk. Adjustments were made for age. Calculated per fresh cycle. Pregnancy loss was defined as extrauterine pregnancies and miscarriages before 22 gestational weeks. Calculated as % of pregnancies after fresh ET or FET. One woman had an intrauterine fetal death at >22 gestational weeks. Calculated as % of FET cycles. Cumulative PR/pregnancy loss is calculated per woman. Some women may have had pregnancy or pregnancy loss after fresh as well as frozen ET. Relative risk for cumulative live birth for women with at least one feature of adenomyosis. Note : Values are given as n (%). A modified Poisson regression analysis was performed to estimate RR, which is presented with 95% confidence intervals. Abbreviations: CI, confidence intervals; CLBR, cumulative live birth rate; ET, embryo transfer; FET, Frozen embryo transfer; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilization; LBR, live birth rate; Nb, number; PR, pregnancy rate, RR, relative risk. Adjustments were made for age. Calculated per fresh cycle. Pregnancy loss was defined as extrauterine pregnancies and miscarriages before 22 gestational weeks. Calculated as % of pregnancies after fresh ET or FET. One woman had an intrauterine fetal death at >22 gestational weeks. Calculated as % of FET cycles. Cumulative PR/ pregnancy loss is calculated per woman. Some women may have had pregnancy or pregnancy loss after fresh as well as frozen ET. Relative risk for cumulative pregnancy, miscarriage, and live birth after IVF/ICSI treatment for women with direct or only indirect features of adenomyosis, stratified for women with or without endometriosis. Note : The relative risk was calculated using a modified Poisson regression analysis, and adjustments were made for age. Abbreviations: aRR, adjusted relative risk; Endo, endometriosis, RR, relative risk. Having at least one feature of adenomyosis located in the JZ lowered the chance for live birth, aRR 0.29 (95% CI; 0.11–0.74), p  = 0.010. Conversely, having any feature of adenomyosis only in the outer myometrium increased the chance of a live birth, aRR 2.61 (95% CI, 1.42–4.8), p  = 0.002, Supporting information Table  S2 . An interrupted JZ was the only individual feature of adenomyosis that had an impact on CLBR. The crude RR for CLB for an interrupted JZ in the coronal plane on 3D TVUS was 0.36 (95% CI, 0.23–0.58), p  < 0.001, and the adjusted RR was 0.34 (95% CI, 0.28–0.51), p  < 0.001. None of the other features showed an individual impact on CLBR. Women with diffuse type of adenomyosis features ( n  = 212) had an aRR for CLB of 0.68 (0.36–1.3), p  = 0.244, compared to women with focal type ( n  = 36), aRR 1.37 (95% CI, 0.68–2.75), p  = 0.43. The RR for CLB for women with mild disease (extent 50%, n  = 7), RR 0.89 (95% CI, 0.27–2.98), p  = 0.85. For the remaining women, the extent was not assessed. The RR for CLB for women with increasing number of features is presented in Supporting information Table  S3 . The aRR for CLB decreased with increasing number of features up to three features. Women with only one feature had an aRR for CLB of 0.45 (95% CI, 0.21–0.95), p  = 0.037, whereas women with three features had an aRR of 0.12 (95% CI, 0.03–0.55), p  = 0.006. For women with four or more features, the aRR was not further significantly reduced.

Our study showed that the presence of direct or indirect MUSA features of adenomyosis was associated with a reduced chance of having a live birth after the first IVF/ICSI treatment. Women with direct features had an increased risk of miscarriage after FET. Features of adenomyosis located in the inner myometrium, particularly an interrupted JZ, reduced the chance of live birth the most, whereas features located merely in the outer myometrium were associated with a favorable ART outcome. The number of features present correlated negatively to the chance of live birth. Our results agree with recent metanalyses concluding that adenomyosis negatively affects IVF treatment outcomes. 3 , 8 , 23 However, various diagnostic criteria for adenomyosis have been used in different studies. According to the revised MUSA definitions, only direct features are pathognomonic and therefore diagnostic of adenomyosis. Indirect features are merely suggestive of the disease. 11 Therefore, the adenomyosis prevalence and hence the association with ART results may differ between studies. Differences in the adenomyosis prevalence compared to other studies could also be explained by different methods used to establish a diagnosis, including the use of 3D TVUS. 10 Our results dispute those of a recent prospective study of 228 women undergoing oocyte donation treatment. 24 No difference in LBR between women with or without direct features was found. However, the miscarriage risk was significantly increased in women with direct features, which is similar to our results. In the present study, the presence of adenomyosis was strongly associated with pregnancy loss after FET, which is also like results reported by others. 25 It has been suggested that an impaired uterine environment, with hyperperistaltic contractions, inflammation, progesterone resistance, and an altered immune response as well as reduced endometrial receptivity, hampers gamete function and implantation in women with adenomyosis. 26 , 27 A previous study did not find any significant difference in LBR in a large prospective study of 1228 women with or without ultrasonographic features of adenomyosis that underwent IVF treatment, 6 which is in line with others. 5 , 7 However, the presence of endometriosis was only 11% in women with adenomyosis features, which is significantly lower than the 40% prevalence of endometriosis in women with any direct feature found in the present study. This is interesting, as concurrent endometriosis in women with adenomyosis negatively affects LBR after IVF treatment. 8 However, a higher reported prevalence of endometriosis among women with adenomyosis in our study could possibly be explained by the fact that all women in our cohort were actively evaluated for endometriosis. 14 Three‐quarters of women with endometriosis on TVUS were previously unaware of having the disease, which would thus have remained unknown without systematic TVUS. We found that an interrupted JZ was the only individual feature of adenomyosis that was associated with lower chances of a live birth. This supports previous findings of a disrupted endomyometrial junction being the single feature associated with poor reproductive performance. 4 There is evidence that the JZ is of vital importance for fertility and proper embryo implantation. 28 An incomplete remodeling of the JZ may negatively impact uterine peristalsis, vascular plasticity of the spiral arteries as well as local inflammatory factors. 29 A disrupted endometrial‐myometrial interface may affect sperm transport, endometrial receptivity and decidualization, trophoblast invasion, and angiogenesis. 27 Altogether, this may compromise successful embryo implantation and pregnancy. In the present study, there was a trend towards lower chances of having a live birth with an increasing number of adenomyosis features. This is in line with a previous study, 4 and could possibly be explained by a lower risk of misdiagnosis with more ultrasound features of the disease. Moreover, it is likely that an increasing number of features is associated with more severe disease, 30 which in turn could impact implantation and CLBR. However, this trend was not significant for four or more features. It is possible, that as the most important single feature associated with a poor outcome was an interrupted JZ, even the presence of only one feature impacted the outcome. With three features, the adenomyosis may already be severe enough, why an additional number of features would not reduce the chances further. Women in our study with direct features of adenomyosis without endometriosis had lower chances of live birth compared to women with direct features and concomitant endometriosis, which is somewhat surprising. Possibly, the presence of direct features in the outer myometrial layer is more associated with DE, which in turn was not associated with lower CLBR in a previous study by this group. 16 The present study was underpowered to detect differences between small subgroups, which could be an alternative explanation to this finding. Indirect features impacted CLBR regardless of the presence of endometriosis. Possibly, this could be explained by the fact that direct features were more often focal, and focal distribution of adenomyosis features was in turn associated with a more favorable outcome. Another possible explanation is that direct features, such as lines and buds, are more often small, and may be associated with early‐stage disease, whereas indirect features might be more associated with later‐stage disease, with fibrosis causing uterine enlargement and fan‐shaped shadowing. When evaluated with 3D TVUS, a previous study found that the JZ was more often infiltrated in women with endometriosis than in women without. 31 Possibly, some indirect features are more associated with endometriosis than with adenomyosis. A potential confounding effect of endometriosis on IVF outcomes in studies evaluating the effect of adenomyosis has been suggested. 2 , 8 , 9 Endometriosis is believed to negatively affect oocyte quality. 32 Women with severe endometriosis have been shown to have similar live birth rates as women without endometriosis when using donor oocytes from healthy women. 33 Our group recently showed that the presence of endometriosis is associated with lower LBR in women undergoing IVF treatment. 16 If the JZ changes are more correlated to endometriosis than to adenomyosis, oocyte donation may overcome a potential impact on oocyte quality caused by endometriosis. This would in part explain the results of two studies that did not find any correlation between LBR and the presence of MUSA features, when using donor oocyte embryos. 24 , 34 Further, women in our study with direct features of adenomyosis were older than women without. The presence of adenomyosis increases with age. 22 The ovarian reserve declines with age, not only quantitatively but also qualitatively and with an increasing rate of aneuploidy. 35 This may partly explain the decrease in LBR in women with direct features. This effect would be overcome when using donor oocyte embryos, as was done in previous studies. 24 , 34 One strength of our study is its prospective design, with systematic TVUS examinations adhering to strictly defined ultrasonographic diagnostic criteria, prior to ART start. This way, any concurrent endometriosis could also be diagnosed. Another strength is using the same experienced examiner for all examinations, excluding any interobserver variability. There are some weaknesses in our study. At first, we did not have any histopathological confirmation of the diagnosis, which is a limitation shared by all ultrasound studies on adenomyosis in women wishing to preserve their fertility. Another potential limitation is using different IVF/ICSI protocols for women in the study. Using the ultralong protocol for some women may have compensated for possible effects of adenomyosis on IVF/ICSI outcome. 25 However, treatments were planned to maximize the chances of a successful outcome for all women, based on what was known at the time the treatment was planned. It would not have been ethical to treat all women with the exact same protocol. However, over the course of the study, evidence for the benefits of an ultralong GnRH treatment was questioned, which may have affected its use. A consequence of strictly adhering to the MUSA definitions of adenomyosis, with only one direct feature of adenomyosis necessary for a diagnosis, is that some women without or with only mild disease may be wrongly classified as having the disease. This could be considered a weakness of the MUSA definitions and hence of this study. Another consequence of using these definitions is that the need for 3D TVUS to detect features of adenomyosis in the JZ may impose difficulties in establishing a diagnosis in a pragmatic real‐life clinical setting, without access to 3D TVUS and/or ultrasonographic expert knowledge. Most women with direct or indirect features of adenomyosis in this study simultaneously had endometriosis, which potentially may affect ART outcomes. A limitation of this study is therefore a potential confounding effect of endometriosis on LBR. To account for this, the cumulative outcomes for women with features of adenomyosis and with or without endometriosis were compared. However, endometriosis and adenomyosis may share the same pathophysiology and have by some been considered the same disease. 26 In addition, endometriosis, particularly superficial peritoneal disease, may be present even if the ultrasound scan is normal. Therefore, we cannot exclude that some women without visible endometriosis on TVUS in fact had endometriosis. This study was underpowered to detect differences in CLBR for subgroups of women, and type II error due to small sample size cannot be excluded. Larger studies should be conducted to account for this.

We found that direct and indirect features of adenomyosis were associated with reduced CLBR after the first IVF/ICSI treatment. Features of adenomyosis located in the inner myometrium, particularly an interrupted JZ, appear to impact the chances of live birth the most. For adequate counseling, systematic ultrasound examinations using the revised MUSA definitions should be considered for all women undergoing ART. Women with features located in the outer myometrium can be reassured, whereas those with several features and/or an interrupted JZ can be informed about a higher risk of treatment failure. The importance of indirect features, such as an interrupted JZ, in the diagnostics and evaluation of adenomyosis in women with subfertility, should be investigated further.

Adenomyosis is a benign, estrogen‐dependent disease, characterized by ectopic endometrial glands and stroma within the myometrium, often coexisting with endometriosis. 1 Data suggest that adenomyosis has a negative impact on fertility and assisted reproductive treatment (ART) outcome in terms of reduced pregnancy and live birth rates and increased miscarriage rates. 2 , 3 , 4 However, evidence is conflicting, and others have not found such correlations. 5 , 6 , 7 The presence of concurrent endometriosis, which potentially may confound the effect on ART results, is not always well described. 2 , 8 , 9 Moreover, the reported prevalence of adenomyosis in women with subfertility varies, 10 which may affect the interpretation of studies that evaluate the outcome after In vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) treatment. Some of these differences may be attributed to heterogeneity regarding the methods and definitions used to diagnose adenomyosis. Therefore, studies that evaluate the IVF/ICSI outcome in women with adenomyosis, using well‐defined and standardized criteria for a diagnosis, are essential. 8 Over the last years, transvaginal ultrasonography (TVUS) has emerged as the first‐line noninvasive method to diagnose adenomyosis. 11 The Morphological Uterus Sonographic Assessment (MUSA) group has suggested a uniform system to describe different features of adenomyosis at transvaginal ultrasonography. 11 , 12 , 13 Indirect features are only suggestive of adenomyosis whereas direct features are pathognomonic of the disease. 11 Improved ultrasonography combined with the revised MUSA definitions has enabled the diagnostics of adenomyosis in subfertile women, as well as comparison between studies. However, the impact of direct or indirect MUSA features on ART outcomes needs to be further investigated. 10 The aim of this study was to examine the cumulative live birth rates (CLBR) after the first IVF/ICSI treatment in women with or without revised MUSA features of adenomyosis. 11

The authors declare no conflict of interest.

This was a prospective, observational cohort study at the Reproductive Medical Center (RMC) at Skåne University Hospital, Malmö, Sweden, a tertiary care clinic treating subfertile couples from the Swedish Southern Healthcare region. Women scheduled for their first ART were consecutively included between December 2018 and May 2021, as previously described. 14 The same cohort of women that is used in the present study has been examined for the prevalence of endometriosis and features of adenomyosis and for IVF/ICSI outcomes for women with or without endometriosis. 14 , 15 , 16 Publicly funded ART is offered to subfertile, non‐smoking women between 25 and ≤39 years, with a Body Mass Index (BMI) ≤30 kg/m 2 , or a >10% weight loss in case of initial BMI ≤35 kg/m 2 . No common children with the present partner are allowed. Women with previous surgical destruction of superficial endometriotic lesions were excluded, as they may have wrongly been classified as not having endometriosis, although you cannot operate the disease per se, only the morphological changes. Women with current hormonal treatment or previous uterine surgery were excluded as this may alter the sonographic appearance of the myometrium. All women underwent a systematic 2D and 3D TVUS examination by an experienced examiner. All women were examined by the first author in the lithotomy position, using a Voluson 10 Expert (GE Medical systems, Zipf, Austria) high‐resolution ultrasound machine equipped with a 5–9 MHz transvaginal transducer (RIC5‐9D). All ultrasound images, volumes, and video clips were stored on the imaging and information management systems Syngo® Dynamics (Siemens Medical Solutions Health Services, Malvern, PA, USA) and ViewPoint (GE Healthcare, München, Germany) for offline analysis. Features of adenomyosis were assessed during the examination and documentation and measurements were made in connection with this. The systematic and dynamic 2D and 3D TVUS examination has been described previously. 15 The myometrium and the junctional zone (JZ) were assessed for features of adenomyosis, as defined by the MUSA group. 11 , 12 , 13 All ultrasonographic features were classified as either indirect or direct. Indirect features include globular uterus, asymmetry, fan‐shaped shadowing, translesional vascularity, and interrupted or irregular JZ, whereas direct features are myometrial cysts, hyperechogenic islands, and lines and buds. The localization in the myometrium, uterine layer involvement, extent (mild 50%), and type of adenomyosis feature (focal/diffuse/mixed type/adenomyoma) were also documented. The pelvis and adnexa were assessed for signs of endometriosis. 14 Unilocular cysts with ground glass echogenicity were considered as endometriomas. 17 Deep endometriosis was reported using the International Deep Endometriosis Assessment group definitions. 18 The ovarian antral follicle count was the sum of all follicles sized 2–10 mm, assessed manually during TVUS. Women with submucous myomas expected to affect ART outcomes were referred for extirpation of myomas and therefore not included in the study. The ovarian stimulation has been previously described in detail. 16 All women were treated according to the gonadotropin‐releasing hormone (GnRH) agonist or antagonist protocol, depending on individual patient characteristics or preferences. The agonist protocol was primarily recommended for women assessed to be low responders or those with either large endometriomas or endometriosis‐associated severe pain suggestive of inflammation. Some women with endometriosis went through an ultralong downregulation 19 with GnRH agonist (Enanton Depot, Orion Pharma AB, Danderyd, Sweden or Synarela, Pfizer AB, Stockholm, Sweden) for 3–6 months prior to ART start. The agonist protocol was initiated with GnRH agonists (Synarela, Pfizer AB, Stockholm, Sweden). For ovarian stimulation, the FSH used was either GONAL‐f, (Merck‐Serono, Darmstadt, Germany), Bemfola, (Gedeon Richter, Stockholm, Sweden), or Menopur, (Ferring, GmbH, Kiel, Germany) with an individualized starting dose. Subcutaneous Fyremadel (SUN Pharmaceutical, Hoofddorp, Netherlands) was started on day five or six in the antagonist protocol. The development of follicles was monitored with TVUS, and ovulation was induced with either subcutaneous human chorionic gonadotropin (Ovitrelle, Merck, KGsA, Darmstadt, Germany) when three or more follicles reached 17 mm. Ovulation induction with GnRH agonist was performed in patients at risk for ovarian hyperstimulation syndrome (and total freezing was performed). After 35–36 h, transvaginal follicle aspiration was carried out. If fewer than three follicles matured, the treatment was either converted to intrauterine insemination or canceled. Depending on semen quality, mature oocytes were either inseminated or injected with sperm. The fertilization rate, which is the number of normally fertilized oocytes divided by the total number of mature oocytes retrieved, was recorded. Embryos were assessed according to the Gardner blastocyst grading scale 20 and embryo transfer was done either two or three (cleavage stage) or 5 days (blastocyst stage) after ovum pickup. Single embryo transfer is a clinical routine. In case of surplus good quality embryos (GQE):s, or if fresh embryo transfer (ET) was not possible, these were cryopreserved on Day 5–6. Frozen thawed ET (FET) was carried out in natural or hormone replacement cycles. Progesterone vagitories (Lutinus, Ferring, Lausanne, Switzerland) were given as luteal phase support for 2 weeks after ovum pickup. All embryos from the index treatment cycle were used until live birth was achieved or no embryos remained. The primary outcome of this study was the CLBR after the first IVF/ICSI treatment cycle in subfertile women with or without direct MUSA features of adenomyosis. Secondary outcomes were pregnancy and miscarriage rates, IVF/ICSI outcomes in women with or without indirect features of adenomyosis and PR or CLBR stratified for concomitant endometriosis. For statistical analyses, we used the statistical package IBM Corp. released 2020. IBM SPSS Statistics for Windows, Version 29.0. Armonk, NY, USA. A previous publication has suggested that women with ultrasound‐diagnosed adenomyosis have a pregnancy rate of 23.6% after IVF compared with 44.6% in the non‐adenomyosis group. 21 For an adequate 80% power with a two‐sided alpha level of 0.05, we needed 158 patients with adenomyosis to detect an equivalent difference. Assuming a prevalence of 15% in our population (lower than the estimated 20% prevalence in the general population 22 as our population is younger), we would need a total of 1053 patients. Normally distributed values were presented as mean with standard deviation (SD) and non‐normally distributed values as median with interquartile range (IQR). Student's t ‐test was used to compare normally distributed data between groups and the Mann–Whitney U ‐test to compare median values. The chi‐square test or Fisher's exact test was used to compare proportions between groups. Proportions were calculated as percentages. A modified, robust error variance, Poisson regression analysis was used to assess the relative risk (RR) for cumulative live birth CLB in women with direct and/or indirect features of adenomyosis, before and after adjusting for age as a potentially confounding factor, serum anti‐müllerian hormone (s‐AMH), BMI, antral follicle count, protocol, FSH dose, stimulation days and ET day as well as endometriosis as potentially confounding factors. In addition, a modified Poisson regression analysis was performed to assess if any of the individual features of adenomyosis were negatively associated with LBR. Crude as well as adjusted RR are presented with 95% confidence intervals (CI). A p ‐value <0.05 was considered indicating a statistically significant association.

Table S1. Table S2. Table S3.