Strategies to Improve Assisted Reproductive Technique Outcomes in Women with Adenomyosis: A Narrative Review

The prevalence of adenomyosis varies based on the population studied and the diagnostic modality used. In a population of 1015 women undergoing ART, one or more adenomyosis features were found on 2D TVS in 30% of women over age 40 and in 22% of women aged <40 years. Its prevalence was higher in women with recurrent pregnancy loss when compared to those without (38% vs. 22%; P < 0.005) and in those with previous ART failures versus those without (35% vs. 24%; P < 0.0001).[ 11 ] In another prospective study involving 1228 women undergoing ART, with a mean age of 36.5 years (±0.3), at least one ultrasonography feature of adenomyosis was found in 32% of women and ≥3 features were found in 14% of women.[ 12 ] The mean frequency of adenomyosis as diagnosed by the gold standard of histopathology is 20%–30% in women undergoing hysterectomy for any indication.[ 13 ] All this suggests that while adenomyosis may be present silently, it is prevalent with a higher frequency in older symptomatic women and infertile women, especially in those with a previous history of recurrent miscarriages or recurrent implantation failure.

Overall, women with adenomyosis are at increased risk of various adverse obstetric outcomes, including preeclampsia, preterm delivery (OR 3.09, 95% CI 1.88–5.09), foetal malpresentation, placental attachment abnormalities, post-partum haemorrhage and small for gestational age (OR 3.23, 95% CI 1.71–6.09).[ 16 17 ]

Adenomyosis develops as a result of the ectopic development of endometrial glands and stroma in the underlying myometrium. Surrounding it, there is reactionary fibrosis, hyperplasia and hypertrophy of the surrounding smooth muscle cells.[ 3 ] This inflammatory reaction gives rise to a variety of clinical symptoms including but not limited to dysmenorrhoea, dyspareunia, abnormal uterine bleeding and chronic pelvic pain. The various theories proposed to understand the pathogenesis of adenomyosis are – (A) invagination of the endometrium into the myometrium through an altered or interrupted junctional zone (JZ) and the contribution of the mechanism of tissue injury and repair; (B) metaplasia of intra-myometrial embryonic or adult stem cells and (C) migration and differentiation of adult endometrial and stromal stem cells after retrograde menstruation through the invasion from outside.[ 4 ] Morphologically, adenomyosis is a heterogeneous condition and is defined based on its location, extent and lesional type. Adenomyotic lesions could be located in one or more of the uterine walls: anterior, posterior, left lateral, right lateral or fundal. Some studies suggest further classifying these lesions based on location in inner or outer myometrium, to quantify their individual effects on reproductive outcomes. Within these locations, adenomyosis may be diffuse or focal. To differentiate between diffuse and focal, the lesion is inspected sonographically at its largest diameter in a sagittal plane. In focal adenomyosis, >25% of the lesional circumference is surrounded by a healthy myometrium, and in diffuse adenomyosis, <25% is surrounded by healthy myometrium.[ 5 ] Adenomyoma is a type of focal adenomyosis that is fully circumscribed by surrounding normal myometrium. Rarely, adenomyosis may present as only a large cyst, in which case it is called a cystic adenomyoma. Diagnosis is based on imaging and histopathology. Imaging modalities like transvaginal sonography (TVS) or magnetic resonance imaging (MRI) utilise specific criteria to diagnose adenomyosis. The morphological uterus sonographic assessment consensus published in 2015[ 6 ] provides a list of two-dimensional (2D) and three-dimensional (3D) TVS features associated with adenomyosis – direct signs or features indicating the presence of ectopic endometrial tissue in the myometrium-like myometrial cysts, hyperechogenic islands and echogenic sub-endometrial lines/buds and indirect signs or features reflecting changes secondary to presence of endometrial tissue in the myometrium such as asymmetrical myometrial thickening, globular uterus, fan-shaped shadowing, translesional vascularity and irregular/interrupted JZ. A 2022 DELPHI consensus recommends the use of 3D ultrasound to assess JZ abnormalities.[ 7 ] At least three objective parameters have been identified for MRI diagnosis of adenomyosis, linked to JZ evaluation on T2-weighted sequences:[ 8 ] the thickening of the JZ at least 8–12 mm, the ratio of JZ maximum/total myometrium over 40% and the difference between the maximum and the minimum thickness of the JZ (JZ maximum − JZ minimum) >5 mm. A thickness exceeding 12 mm seems to be highly predictive of adenomyosis.[ 9 ] Both MRI and TVS perform similarly when measured against the gold standard of histopathology. A systematic review included studies using either ultrasound or MRI for adenomyosis diagnosis, before subjecting women to hysterectomy. When measured against histopathology TVS had a sensitivity of 72%–82% and a specificity of 81%–85%, whereas MRI had a sensitivity and specificity of 77% and 89%, respectively.[ 10 ] Because of its low cost and easy availability, TVS remains the preferred imaging modality for diagnosing adenomyosis. MRI use may be restricted to cases with diagnostic dilemma or when there are coexistent fibroids that need exact mapping and distinction from adenomyosis before surgery.

The mechanisms of action of adenomyosis in lowering reproductive outcomes are twofold: aberrant mechanical factors and deranged inflammatory processes at the implantation surface. Aberrant uterine contractility at the endometrial–myometrial interface interferes with favourable implantation. Strategies to improve ART outcomes aim to mitigate the effects of these aberrations. These strategies can be divided into ovarian stimulation protocols, ET strategies and pre-IVF treatments aimed at reducing adenomyosis burden. The use of long-acting or continuous gonadotropin-releasing hormone agonist (GnRH-a) in ovarian stimulation protocols, originally used for suppressing premature luteinising hormone surges and preventing spontaneous ovulation, has been adopted in women with adenomyosis due to its hypoestrogenic and antiproliferative effects.[ 18 ] The ultra-long GnRH-a protocol starting 3–6 months before controlled ovarian stimulation has been commonly employed and compared with the long agonist, the short agonist and the antagonist protocols in women with adenomyosis. Relevant outcomes in protocol selection would be its effect on oocyte numbers, utilisable embryo numbers, clinical pregnancies, miscarriages and live births after fresh transfer and cumulative LBRs [ Table 1 ]. Fresh embryo transfer and cumulative-assisted reproductive treatment outcomes with different interventions versus controls in women with adenomyosis *Studies employing DNG as a pre-treatment specifically in adenomyosis patients not found, Endometriosis pathology taken as a surrogate for adenomyosis. ART=Assisted reproductive treatments, CLBR=Cumulative live birth rates, DNG=Dienogest, ET=Embryo transfer, GNRH=Gonadotropin-releasing hormone, IVF= In vitro fertilisation, ICSI=Intracytoplasmic sperm injection, OR=Odds ratio, CI=Confidence interval, NS=Non significant Ge et al . analysed the effects of different ovarian stimulation protocols on IVF/ICSI in a meta-analysis of 23 observational studies that included 4872 cycles in adenomyosis population and 6289 cycles in non-adenomyosis population.[ 19 ] In the adenomyosis population, the CPRs: 42.47% vs. 55.89%, OR: 0.74, 95% CI: 0.66–0.82 and LBRs: 30.72% vs. 47.77%, OR: 0.65, 95% CI: 0.58–0.73 were lower, but the miscarriage rates were higher (MR: 27.82% vs. 13.9%, OR: 1.90, 95% CI: 1.56–2.31) than in the non-adenomyosis population. In addition, in the adenomyosis population undergoing fresh ET cycles, the ultra-long protocol led to a higher CPR than the other three protocols (476/900 vs. 264/646; OR = 1.33; 95% CI = 1.06–1.66, I 2 = 40%); similar MRs (99/450 vs. 54/256; OR = 1.06; 95% CI = 0.71–1.58, I 2 = 31%) and similar LBRs (325/724 vs. 174/468, OR = 1.08; 95% CI = 0.65–1.81, I 2 = 72%). These results could suggest three things: first, that non-randomisation in all of the included studies could have led to a strong selection bias when choosing protocols, that is the severity of disease might have determined the choice of ultra-long versus other protocols, thus not reflecting the true effect of ultra-long protocol on fresh transfers. As it is, most of the included studies fail to mention distribution of adenomyosis severity in their cohorts. The other is that, while clinical pregnancies show a significant increase with ultra-long protocol, this effect is mitigated due to high miscarriages in adenomyosis, which could not be saved by the ultra-long protocol, thus showing no influence on live births. Lan et al . in a retrospective analysis, compared the effectiveness of the ultra-long GnRH-a protocol that utilised 2–3 months of depot injections, with that of the long GnRH-a protocol[ 20 ] in a total of 371 women with adenomyosis. While total gonadotropin dose ( P = 0.98), mean age ( P = 0.32) and anti-Müllerian hormone (AMH) (2.4 vs. 1.9; P = 0.11) were similar in the two arms, the oocytes recovered were fewer (8.3 ± 5.4 vs. 9.9 ± 7.0; P = 0.001) in the ultra-long than in the long agonist protocol. Although no overall significant differences were found in implantation, clinical pregnancy or LBRs, the ultra-long protocol showed significantly reduced incidence of early miscarriages (12.0% vs. 26.5%, P = 0.045) compared to the long protocol. Subgroup analysis revealed that women with diffuse adenomyosis had significantly higher clinical pregnancy (55.3% vs. 37.9%, P = 0.025) and LBRs (43.4% vs. 25.9%, P = 0.019) with the ultra-long versus the long GnRH-a protocol. In contrast, no significant differences in pregnancy outcomes were observed between the two protocols in women with focal adenomyosis. The authors of another retrospective study found that 3–6 months of GnRH-a pre-treatment before controlled ovarian stimulation actually reduced cumulative live birth rates (CLBR) versus when no pre-treatment was used (27.9 vs. 40.5%, P = 0.019). Their mean retrieved oocytes (7 vs. 9; P = 0.003) and mean good-quality embryos (2 vs. 1; P = 0.02) were significantly lower with GnRH-a pre-treatment than without.[ 21 ] In conclusion, the ultra-long GnRH-a protocol may positively influence fresh ET outcomes in women with diffuse adenomyosis by possibly correcting implantation defects. However, the finding that this protocol leads to fewer retrieved oocytes, which negatively affects cumulative live births, should prompt clinicians to seek other strategies for improving outcomes. Frozen ET (FET) along with GnRH-a suppression has been employed as a strategy to improve live births in women with adenomyosis since it possibly plays a role in reducing endometrial alterations linked to ovarian hyperstimulation and in improving immunological environment and EM junction contractility in adenomyosis. Bourdon et al . retrospectively compared the outcomes of ART between the fresh ( n = 111) and freeze-all strategies ( n = 195) in infertile women with adenomyosis.[ 22 ] Both groups had a similar distribution of adenomyosis severity, but the group receiving freeze-all treatment had a higher AMH, and the predominant indication for freeze-all in this group was ovarian hyperstimulation syndrome. The protocols used for controlled ovarian stimulation and for uterine preparation in FET were equally distributed within the two groups. The freeze-all group demonstrated significantly higher CLBR (44.1% vs. 30.6%) and cumulative ongoing pregnancy rates (45.1% vs. 32.4%) compared to the fresh ET group. Multivariate logistic regression analysis revealed that the freeze-all strategy was independently associated with increased odds of achieving a live birth in women with adenomyosis (OR = 1.80; 95% CI = 1.02–3.16). In response to the research question of whether long-term GnRH-a pre-treatment can improve pregnancy outcomes in patients with adenomyosis undergoing FET after endometrial preparation with hormone replacement therapy (HRT), Niu et al . retrospectively compared 339 patients with adenomyosis, divided into two groups: 194 receiving long-term GnRH-a plus HRT and 145 receiving only HRT. The two groups had similar baseline and embryo characteristics. In GnRH-a + HRT group, clinical pregnancy, implantation and ongoing pregnancy rates (51.35%, 32.56% and 48.91%, respectively) were significantly higher than that of HRT group (24.83%, 16.07% and 21.38%, respectively).[ 23 ] In a similar retrospective analysis, Zhang et al . analysed the results of 263 FET cycles in women with adenomyosis and uterine volumes between 56 and 100 cm 3 .[ 24 ] The intervention involved administering GnRH-a pre-treatment before FET, while the comparison group underwent FET without GnRH-a pre-treatment. The pre-treatment group showed a significantly higher LBR (46.7% vs. 24.8%, P = 0.009) and a significantly lower miscarriage rate (12.5% vs. 37.2%, P = 0.044) compared to the group without GnRH-a. Although the CPR was higher in the GnRH-a pre-treatment group (53.3% vs. 39.4%), the difference was not statistically significant ( P = 0.098). These findings suggest that GnRH-a pre-treatment before FET may improve reproductive outcomes by reducing miscarriage and increasing LBRs in adenomyosis. These include the use of treatments to reduce the adenomyosis burden for variable lengths of time, ranging from 1 to 6 months, before initiating IVF treatments. These include oral progesterone DNG, levonorgestrel intrauterine device and low-dose letrozole. Surgery to excise adenomyoma is also considered a pre-treatment strategy. Levonorgestrel intrauterine system (LNG-IUS) acts by maintaining high concentrations of levonorgestrel in the uterine cavity compared to its serum levels.[ 25 ] The pseudo-decidualisation of endometrium thus promoted, reduces endometrial glandular activity and indirectly, the inflammatory response. Liang et al . retrospectively evaluated the effect of up to 3 months of pre-treatment with an LNG-IUS on FET outcomes in women with adenomyosis ( n = 358).[ 26 ] The device was removed 3 months after insertion, and ET was performed in the subsequent cycle. The uterine preparation methods were either natural, HRT or a stimulated cycle and were equally distributed in the two groups. The mean time to ET after LNG-IUS removal was 50–68 days. The mean endometrial thickness was 10.64 ± 2.31 versus 10.14 ± 2.07 ( P = 0.0436) at transfer. The ongoing pregnancy rate per transfer was significantly higher with LNG-IUS than in the control group (41.8% vs. 29.5%; P = 0.017). A logistic regression analysis that controlled for age, body mass index, infertility duration and endometrial thickness showed that the ongoing pregnancy rate was significantly higher with LNG-IUS use (adjusted OR [aOR], 1.63; 95% CI, 1.01–2.62). A number of small retrospective studies have analysed the post-ART reproductive outcomes after use of DNG in women with endometriosis. They do not specifically segregate the adenomyosis-only cohorts from the endometriosis-only cohorts. However, assuming that endometriosis and adenomyosis have similar pathogenesis, it is worthwhile examining these studies. Barra et al .[ 27 ] included 151 women with endometriosis and failed multiple IVF attempts. Sixty-three women received 2 mg DNG daily for 3 months, while 88 women received no pre-IVF hormonal treatment. The implantation, clinical pregnancy and LBRs were significantly higher with DNG pre-treatment (39.7%, 33.3% and 28.6%, respectively) than without (23.9%, 18.2% and 14.8%; P = 0.049, 0.037 and 0.043, respectively). This is in contrast to a small randomised controlled trial (RCT), which found reduced rates of clinical pregnancy after IVF (34% vs. 68%; P < 0.05) in the DNG arm ( n = 33) than in the no pre-treatment arm ( n = 38).[ 28 ] The authors attributed this reduction in reproductive outcome to a decrease in retrieved oocyte numbers in the DNG arm (5 ± 3.6 vs. 7.5 ± 4.2; P < 0.05). This, they surmised, was a result of increased follicular atresia or reduced antral follicular induction, caused by prolonged pre-IVF DNG therapy. In a systematic review that pooled results from five studies of both the RCT and observational cohort designs, comprising 538 women with endometriosis, the authors found a weighted mean difference in mature oocyte recovery of -0.99 (95% CI= -2.26 to 0.28, P >0.05) in the DNG arm versus controls.[ 29 ] Alternatively, it is possible that DNG may be a helpful adjunct before FET, where its adverse effect on antral follicular induction can be bypassed; however, in the absence of larger clinical data, no conclusions can be made yet. In women with adenomyosis awaiting IVF, a low dose of letrozole (2.5 mg three times a week) given for 12 weeks has been compared in a randomised control trial, with monthly goserelin 3.6 mg for 3 months, to assess differences in pre- and post-treatment lesional sizes, bleeding and pain. There was significant improvement in sonographic features of diffuse adenomyosis of the myometrium (letrozole P = 0.015, GnRH-a P = 0.039); diffuse adenomyosis of the JZ (letrozole P = 0.025, GnRH-a P = 0.001) and adenomyoma size (letrozole P = 0.049, GnRH-a P = 0.024), suggesting that low-dose letrozole could be an effective, safe and low-cost alternative to GnRH-a prolonged downregulation without causing amenorrhoea or significant hypoestrogenic effects.[ 30 ] Based on the hypothesis that lesional size reduction should improve the molecular environment around implantation, further trials are needed to study its reproductive outcomes after ART. Pre FET interventions are listed in Table 2 . Assisted reproductive treatments outcomes in women with adenomyosis using different interventions prior to frozen embryo transfer cycles ART=Assisted reproductive treatments, FET=Frozen embryo transfer, GNRH=Gonadotropin-releasing hormone, HRT=Hormone replacement therapy, LNG-IUS=Levonorgestrel intrauterine system, IVF= In vitro fertilisation Cytoreductive surgery has been suggested for partial or full removal of adenomyotic lesions in women with painful cystic adenomyosis not responding to medical therapy or in infertile women with previous failed transfers with good-quality embryos. The goal of surgery for adenomyosis is to optimally reduce the adenomyosis burden by excising it wholly or partially, while maintaining endometrial cavity integrity and the molecular environment at the endo-myometrial interface, reconstructing the uterus adequately and preserving the functionality of the ovaries and tubes. Various surgical approaches, resection and reconstruction techniques have been described for adenomyosis.[ 31 ] The approach includes laparoscopic, open or robotic surgery. The common resection techniques include classic wedge, single incision and enucleation; and the open triple flap method. The classic wedge resection is used to debulk diffuse adenomyosis. It involves giving two vertical incisions on the serosa slicing through the myometrium, converging at the endomyometrial junction, excising the tissue in between and resuturing the seromuscular layer in two to three layers. Adenomyomectomy, a term used for the excision of focal adenomyoma, usually employs a single incision and resection of adenomyoma by sharp or blunt dissection, much like that employed for enucleation of fibroid. Subsequent suturing can be in 2–3 layers or using the double-breasting technique to resuture overhanging flaps and restore myometrial integrity. The triple flap is an open method, requiring a midline uterine bisection down to the endometrial cavity, identification of adenomyotic tissue using digital palpation, complete resection until healthy myometrium is palpated and re-suturing of the bed using the triple flap method, with separate sutures for the endometrium, myometrium and serosa. Bischiniotis et al .[ 31 ] have pooled results from six studies to report on reproductive outcomes after adenomyomectomy. Totally, of 181 women wishing to conceive post-surgery, total conceptions, miscarriages and delivery rates were 58.1%, 15% and 46.9%, respectively, over 18–120 months of follow-up. In a single-arm study assessing the determinants of pregnancy after laparoscopic adenomyomectomy, the authors achieved a CPR of 32% (32/101) within 2 years after the procedure. These pregnancies were either conceived spontaneously ( n = 16) or through IVF ( n = 16). Positive determinants of clinical pregnancy were a history of IVF treatment (aOR: 6.22, 95% CI = 1.9–20.33) and negative determinants were posterior uterine wall involvement (aOR: 0.18, 95% CI = 0.09–0.63) and age ≥40 years (aOR: 0.77, 95% CI = 0.67–0.88). The authors concluded that a laparoscopic adenomyomectomy was a reasonable option in young infertile women with diffuse posterior wall involvement.[ 32 ] A study aimed to compare the effectiveness of conservative surgery against medical therapy alone (GnRH-a for 6 months) in 65 women with histopathology-proven extensive adenomyosis.[ 33 ] The authors found improved symptom control and significantly higher CPRs (46.4% vs. 10.8%; P = 0.002) and LBRs (32.1% vs. 8.1%; P = 0.020) over 36 months of follow-up with surgery compared to medical therapy. The same authors in a different publication have evaluated relapse rates after laparoscopic adenomyomectomy at 2 years and found it to be 49%. However, they observed that with GnRH-a therapy given for 6 months post-surgery, the symptom relapse rate could be reduced to 29%.[ 34 ] A systematic review and meta-analysis pooling results from uncontrolled studies, that looked amongst other things, at the adverse effects of adenomyomectomy, reported uterine rupture and preterm birth rates of 6.8% (3/44) and 4.5% (2/44) in pregnant patients with diffuse adenomyosis and 0% (0/35) and 10.9% (12/110) in patients with focal adenomyosis.[ 35 ] In addition, reports of abnormal placentation after adenomyomectomy exist.[ 36 ] Hence, women who conceive after adenomyomectomy should be put in a high risk pregnancy category throughout the three trimesters, and while vaginal delivery is possible in such cases, a caesarean section appears to be the safest. Two different non-surgical ablative technologies have been used to thermally ablate adenomyotic lesions: high-intensity focused ultrasound (HIFU) ablation and radiofrequency (RF) ablation. MRI-guided HIFU has been studied clinically in a few studies. HIFU ablation relies on the creation of ultrasound waves by an external transducer. Zhou et al . did a follow-up of 68 HIFU-treated adenomyosis patients who wished to conceive. Following treatment, 54 of the 68 women conceived within a median of 10 months. Among those who conceived, 21 successfully delivered healthy babies without any reported cases of uterine rupture or neonatal complications.[ 37 ] RF ablation is an alternative ablation technique that involves ultrasound-guided implantation of an electrode, which generates heat by causing the molecules within cells to resonate, resulting in lesion ablation. One study has evaluated the outcomes after RF ablation in 74 infertile patients with adenomyosis. Twenty-nine (35.8%) patients achieved pregnancies, with 22 (27.8%) patients achieving a live birth. 33.3% (12/36) of all pregnancies ended in spontaneous abortions. No case of uterine rupture occurred.[ 38 ] Until safety data from further high-quality trials is available, non-surgical ablative techniques in women who wish to conceive should be adopted with caution. This is a narrative review and may be unintentionally limited by its literature search strategy and the author biases towards study selection.

Over the past 40 years, the number of assisted reproductive technique (ART) cycles has increased exponentially across the world. A recent study estimates an ART-conceived infant is born every 35 s, making up to 0.5% of all global births and 5%–10% of births in some high-income countries.[ 1 ] This increase has been attributed to a global trend towards delayed childbirth as well as to an increasing availability, efficiency and acceptance of ART procedures. ART success rates depend upon a myriad of patient factors ranging from female age, infertility duration, previous childbirth history, ovarian reserve factors, semen characteristics and uterine pathologies affecting the endometrium or the myometrium. Adenomyosis, a condition characterised by the presence of endometrial stroma and glands within the myometrium, is one such uterine condition that can adversely affect embryo implantation.[ 2 ] This review describes adenomyosis in brief, covers the burden of adenomyosis in women undergoing ART and critically appraises the strategies employed to improve ART outcomes in such women.

This narrative review employs a literature search on PubMed using three keyword search strings with the AND operator and employing ALL fields comprising the title, abstract or text. The keyword strings were Fertilisation in vitro or intracytoplasmic sperm injection or assisted reproductive technology or embryo transfer (ET) or controlled ovarian stimulation or Controlled ovarian hyperstimulation AND Adenomyosis or Adenomyoma AND pre-treatment or protocol or agonist or antagonist or ultra-long protocol or progesterone or levonorgestrel or dienogest (DNG) or surgery or adenomyomectomy or resection or non-medical interventions. When looking at ‘prevalence’ and ‘diagnosis,’ controlled studies were chosen that described the prevalence or diagnostic accuracy against a control or reference standard. When, discussing ‘interventions’, controlled studies reporting on clinical pregnancies were chosen. When available, systematic reviews that incorporated most past studies were chosen and described, and all subsequently done studies were picked up and described.

Adenomyosis is a heterogeneous condition morphologically, and there is evidence that its severity negatively affects reproductive outcomes in ART. As of now, published strategies to improve ART outcomes in adenomyosis are derived from observational data which suffer from the limitation of selection bias, missing data and heterogeneity which makes valid, generalisable conclusions difficult. As of now, ultra-long GnRH agonist protocol is an option to improve outcomes in fresh ET but has limited role in improving cumulative outcomes because of its negative effects on oocyte recovery. Therefore, freeze-all strategy followed by medical or surgical therapy to reduce adenomyosis burden and subsequent FET appears to be a reasonable strategy. Surgery appears to reduce adenomyosis burden more effectively than medical therapy, but this option should be judiciously exercised in the wake of known pregnancy complications of adenomyomectomy. Further, there is a need to homogenise the reporting of adenomyosis features on ultrasonography and stratify adenomyosis populations based on severity when determining the efficacy of various interventions. RS was involved in conceptualisation, structure, literature search and overall writing and review of the manuscript. NV and V were involved in searching literature and writing a primary draft. There are no conflicts of interest. No original data used.