Cumulative live birth rates in women with endometriosis undergoing ART treatment

Endometriosis affects approximately 10% of women of reproductive age ( Rowlands et al. , 2021 ), with a significant proportion experiencing difficulties in conceiving ( Barnhart et al. , 2002 , Dunselman et al. , 2014 ). The pathogenesis of endometriosis-related infertility is complex and multifactorial often compounded by other factors that impact fertility such as age and parity. Several proposed mechanisms are thought to be involved, including pelvic anatomical distortion, altered peritoneal function, ovulatory problems, impaired implantation and the negative impact of ovarian endometriosis on sperm, oocytes and embryo quality, leading to attrition of early embryos and reduced implantation ( Practice Committee of the American Society for Reproductive Medicine, 2012 ; Harb et al. , 2013 ; Fenner, 2024 ). Women with endometriosis frequently undergo ART, such as IVF, to achieve pregnancy ( Barnhart et al. , 2002 ). However, the impact of endometriosis on ART outcomes remains a topic of ongoing debate, with varied results. Several studies have reported that endometriosis negatively influences ART success rates, with reduced oocyte quality, lower fertilization and implantation rates, and ultimately lower live birth rates ( Azem et al. , 1999 ; Barnhart et al. , 2002 ; Aboulghar et al. , 2003 ; Omland et al. , 2005 ; Kuroda et al. , 2009 ). Conversely, other studies describe comparable pregnancy and live birth rates between women with endometriosis and those with other infertility factors ( Olivennes et al. , 1995 ; Stern et al. , 2013 ; Saraswat et al. , 2017 ). Cumulative live birth rates (CLBRs) over several ART complete cycles (all fresh and thaw embryo transfers associated with one ovarian stimulation) undertaken by the same women offer a comprehensive measure of ART success, crucial for counselling patients with endometriosis, who may require multiple complete cycles to achieve a live birth. Despite its clinical significance, data on the impact of endometriosis on CLBR are limited and inconsistent ( Feichtinger et al. , 2019 ; Boucret et al. , 2020 ). Additionally, little is known about how the presence of endometriosis combined with other infertility diagnoses influences ART success rates compared to women without a diagnosis of endometriosis. This retrospective cohort study aims to address these gaps by estimating the CLBRs in women with endometriosis undergoing ART treatment, using data from the Australian and New Zealand Assisted Reproduction Database (ANZARD). By comparing women with endometriosis as their sole infertility diagnosis, those with endometriosis and additional diagnoses, and women without endometriosis, we aim to provide valuable insights into how endometriosis influences ART outcomes, particularly live birth rates.

A total of 92 376 women underwent their first autologous fresh ART cycle between 2014 and 2019 in Australia and New Zealand. Of these, 13 058 (14.1%) were excluded from the analysis due to missing infertility data. Among the remaining 79 318 women with available infertility data, 4311 (5%) had endometriosis as the sole cause of infertility ( endometriosis-only ), 6312 (8%) had endometriosis combined with other infertility diagnoses ( endometriosis-plus ) and 68 695 (87%) had no endometriosis but other infertility ( other infertility ). In the endometriosis-plus group, 2068 women (33%) had more than one additional infertility factors (e.g. male factor, tubal factor or other female infertility factor) alongside endometriosis ( Table 1 ). Causes of infertility among women undergoing their initial autologous ovarian stimulation cycle in Australia and New Zealand during 2014–2019 and followed until 2021. The mean age of women at the time of the first ovarian stimulation cycle was slightly higher in the other-infertility group (34.7 years) compared to the endometriosis-only (33.9 years) and endometriosis-plus (34.1 years) groups ( P  < 0.001) ( Table 2 ). The distribution of age groups also varied, with a higher percentage of women aged 40–44 in the other-infertility group (16.9%) compared to the endometriosis-only (10.2%) and endometriosis-plus (11.9%) groups. Demographic characteristics and treatment outcomes of women undergoing their initial autologous ovarian stimulation cycle in Australia and New Zealand during 2014–2019 and followed until 2021, classified by infertility diagnosis. Endometriosis-plus = Endometriosis and one of the concomitant diagnosis, including male infertility, tubal disease, and other female factors. No endometriosis but other infertility = No endometriosis but male infertility or tubal disease or other female factors or unexplained infertility. Women in the other-infertility group had a higher rate of previous pregnancy (12.8%) compared to women with endometriosis-only (8.1%) or endometriosis-plus diagnosis (8.7%) ( P  < 0.001) ( Table 2 ). Cycle cancellation rates (defined as ovarian stimulation initiated but not proceeding to egg retrieval) in the first ovarian stimulation cycle were also lowest in the endometriosis-only group (6.8%) and increased among women with endometriosis-plus (8.4%) and those with other-infertility diagnoses (9.1%) ( P  < 0.001). The mean number of oocytes retrieved was lower in women with endometriosis-only (mean = 8.9) and endometriosis-plus (mean = 8.7) compared to those with other-infertility diagnoses (mean = 9.2) ( P  < 0.001). The mean number of fertilized oocytes was 5.2 for endometriosis-only cases, slightly above the endometriosis-plus group (4.9) and comparable to other-infertility group (5.1; P  < 0.001). Overall, women with endometriosis-only infertility achieved the highest pregnancy (70.5%) and live birth rates (64.5%) compared to the endometriosis-plus group (66.0% pregnancy, 55.9% live births) and other-infertility group (65.0% pregnancy, 57.8% live births) ( P  < 0.001) ( Table 2 ). After adjusting for maternal age, year of treatment, and parity, women with endometriosis-only infertility had a 6% higher likelihood of live birth compared to the other-infertility group (adjusted relative risk [aRR]: 1.06; 95% CI: 1.04–1.08) ( Table 3 ). In contrast, women with endometriosis-plus diagnosis had a 5% lower likelihood of live birth (aRR: 0.95; 95% CI: 0.93–0.97) compared to the other-infertility group. Compared to other-infertility diagnoses, the risk of pregnancy loss was significantly lower in the endometriosis-only group (aRR: 0.88; 95% CI: 0.78–0.99) and significantly higher in the endometriosis-plus group (aRR: 1.46; 95% CI: 1.35–1.59). Association between endometriosis and pregnancy outcomes of women undergoing their initial autologous ovarian stimulation cycle in Australia and New Zealand during 2014–2019 and followed until 2021. Adjusted for year of treatment, maternal age and parity reported at the initial treatment cycle. The CLBRs were calculated for up to six complete ART cycles ( Table 4 ). Following the first complete cycle, the live birth rate was 39.6% for women with endometriosis-only , 27% for women with endometriosis-plus infertility, and 35.3% for those with other-infertility diagnosis. By the sixth cycle, women with endometriosis-only reached a higher CLBR (conservative 64.0%; optimal rate: 83.0%) compared to those with endometriosis-plus infertility (conservative 54.3%; optimal rate: 68.7%) and other-infertility (conservative 57.3%; optimal rate: 76.5%). Cycle-specific and cumulative live birth rates (complete cycle) for women who started their first autologous fresh cycle in Australia and New Zealand during 2014–2019 and followed until 2021 or the first treatment-dependent live birth. ART success rates diminish with successive failed treatment attempts with the largest gains in CLBRs occur within the first two to three cycles across all groups. In the endometriosis-only group, conservative CLBR increases significantly from 39.6% to 53.7%, while the endometriosis-plus group rises from 27.0% to 40.2%, and the other-infertility group from 35.3% to 48.2% ( Table 4 ; Fig. 1 ). As expected, gains beyond the second cycle are modest, with CLBRs increasing by 4–7% between the third and sixth cycle. Cumulative live birth rates (complete cycle) for women who started their first autologous fresh cycle in Australia and New Zealand during 2014–2019 and followed until 2021 or the first treatment-dependent live birth. Conservative CLBR assumes women who discontinued had zero probability of live birth, while optimal CLBR assumes they had the same probability as those continuing ART. Endometriosis-plus group includes women diagnosed with endometriosis along with at least one additional infertility factor such as male, tubal, or other female causes. Discontinuation rates also progressively increased with each successive cycle across all three groups ( Table 4 ). Among women with endometriosis-only infertility, rates rose from 23.7% after the first cycle to 31.2% by the sixth. The endometriosis-plus group exhibited the lowest rates, starting at 14.0% and reaching 26.4% by the sixth cycle. In contrast, the other-infertility group had the highest rates, peaking at 33.4% in the sixth cycle. The CLBRs by age group were calculated for women undergoing their first autologous fresh cycle ( Fig. 2 ). For all age groups, CLBRs were consistently lowest in the endometriosis-plus group compared to the endometriosis-only and other-infertility groups. Across all infertility groups, live birth rates declined with advancing age. Women under 30 achieved the highest rates, with conservative CLBRs after three complete cycles reaching 74.9% for endometriosis-only (optimal rate: 86.6%), 63.5% for endometriosis-plus (optimal rate: 70.5%) and 70.1% for other-infertility groups (81.8%). Significant reductions were observed in women aged 35–39 and even more so in those aged 40 or older, with the most pronounced declines occurring in the endometriosis-plus group. Women over 44 years had negligible live birth rates, regardless of the infertility diagnosis. Cumulative live birth rates by age group for women who initiated their first autologous fresh cycle in Australia and New Zealand during 2014–2019 and followed until 2021 or the first treatment-dependent live birth. ( A ) Endometriosis-only infertility, ( B ) Endometriosis-plus infertility, ( C ) Other infertility. Conservative CLBR assumes women who discontinued had zero probability of live birth, while optimal CLBR assumes they had the same probability as those continuing ART. Endometriosis-plus group includes women diagnosed with endometriosis along with at least one additional infertility factor such as male, tubal, or other female causes.

This retrospective cohort study analysed data from ANZARD, an audited clinical quality registry that collects ART treatment and outcome data from all cycles performed in Australia and New Zealand ( Newman et al. , 2021 ). It is a licencing requirement that fertility clinics submit their data to ANZARD and thus complete ascertainment is assumed ( Paul et al. , 2020 ). The study included women who had their first autologous (patient’s own oocytes) ovarian stimulation cycle between 2014 and 2019. Records of all autologous thawed cycles were linked to their corresponding episodes of ovarian stimulation, allowing the identification of patient and treatment characteristics of woman at the time of ovarian stimulation. These women were followed-up through to the end of 2021 or until the first treatment-dependent live birth, with the follow-up period ranging from a minimum of 2 years to a maximum of 8 years. Cycles after the first live birth were excluded from the analysis. The ANZARD registry records the cause of infertility as specified by the treating clinician for each treatment cycle in the following non-mutually exclusive categories: tubal disease, endometriosis, other female factors, male factors and unexplained infertility. A woman or couple may have multiple infertility diagnoses either within a single treatment cycle or throughout the entire treatment period (2014–2021). For this study, women were categorized into three groups: endometriosis-only group comprised women whose sole infertility diagnosis was endometriosis throughout the treatment period; endometriosis-plus comprised women diagnosed with endometriosis along with at least one additional infertility diagnosis during any treatment cycle, comprising a diagnosis male infertility, tubal factor infertility or other female infertility factors; Other-infertility group comprised women without a diagnosis of endometriosis but with other infertility diagnoses, including male infertility, tubal infertility, other female factors or unexplained infertility. Patients whose infertility diagnosis was recorded as ‘ Not recorded ’ across all infertility categories in all treatment cycles were excluded from the analysis. The primary outcome of interest was live birth, defined as the delivery of a live-born infant of at least 20 weeks gestation, or with a birth weight of at least 400 grams. A clinical pregnancy was defined as one that met at least one of the following criteria: ongoing at 20 weeks, ultrasound evidence of an intrauterine sac (with or without a foetal heartbeat), identification of chorionic villi in the products of conception, or confirmation of an ectopic pregnancy via laparoscopy or ultrasound. Pregnancy loss was categorized as a clinical pregnancy without a live birth. We compared the demographic and treatment characteristics across the three groups. Categorical variables were compared using Pearson Chi-squared test, while means of continuous variables across groups were compared using analysis of variance (ANOVA). Separate generalized linear models were used to assess how treatment outcomes—live birth, clinical pregnancy and pregnancy loss—differed in women with endometriosis compared to those without endometriosis. The models were adjusted for treatment year, maternal age and parity at the start of the first treatment cycle. We calculated CLBRs of women up to six complete ART cycles. A complete ART cycle comprises all embryo transfers (including both fresh and thaw embryos) associated with one ovarian stimulation. Given the differing assumptions about the prognosis of women who discontinued ART treatment, we calculated two types of CLBRs: conservative and optimal. The conservative CLBR assumes that women who discontinued treatment would have zero probability of achieving a live birth if they had continued with ART treatment. This was calculated by dividing the number of women who achieved a live birth up to a specific cycle undertaken between 2014 and 2021 by the total number of women who had their first ovarian stimulation between 2014 and 2019. The 95% confidence intervals (CIs) of conservative CLBR were calculated using standard errors based on the binomial distribution. In contrast, the optimal CLBR assumes that women who discontinued treatment had the same probability of achieving a live birth as those who continued. It was estimated (with 95% CI) using the Kaplan–Meier method ( Kaplan and Meier, 1958 ). Analyses were conducted in StataCorp. 2019 Stata Statistical Software: Release 16 (StataCorp LLC, College Station, TX, USA).

This study provides a comprehensive age-stratified analysis of CLBRs in women with endometriosis undergoing ART treatment, highlighting the impact of endometriosis on treatment success. Our results demonstrated that women with endometriosis as their sole infertility diagnosis had higher live birth rates compared to women with endometriosis combined with additional infertility factors or other infertility diagnosis. Specifically, the adjusted relative risk for live birth was 6% higher in the endometriosis-only group compared to the other-infertility group (aRR: 1.06, 95% CI: 1.04–1.08), suggesting that isolated endometriosis does not adversely affect ART outcomes. Women with endometriosis-plus infertility had the poorest outcomes. After six complete ART cycles, women with endometriosis-only infertility achieved the highest CLBRs (64.0% conservative, 83.0% optimal), exceeding those with endometriosis-plus infertility (54.3%, 68.7%) and other-infertility (57.3%, 76.5%). These results are reassuring for patients with endometriosis and aid in counselling regarding ART success rates. Our findings are consistent with previous studies reporting comparable or even improved ART outcomes in women undergoing ART with endometriosis as the sole infertility factor ( Harb et al. , 2013 ; Senapati et al. , 2016 ; Maggiore et al. , 2024 ; Zhang et al. , 2025 ). The favourable outcomes in the endometriosis-only group are supported by growing evidence that endometriosis may not substantively compromise oocyte or embryo quality. Several studies have shown comparable fertilization, blastulation, implantation and euploidy rates between women with endometriosis and those with other infertility diagnoses ( Juneau et al. , 2017 ; Ata and Telek, 2021 ; Bishop et al. , 2021 ; Benlioglu et al. , 2025 ). These findings support the hypothesis that oocyte competence may be maintained in women with endometriosis, particularly in the absence of other reproductive pathologies or advanced-stage disease. Although the exact mechanisms of how endometriosis impacts reproductive outcomes are not yet fully understood ( Macer and Taylor, 2012 ; Tanbo and Fedorcsak, 2017 ), it is known that endometriosis can lead to heightened pelvic inflammation, oxidative stress, and reduced gamete quality, potentially hindering the success of ART ( Gupta et al. , 2006 ; Harb et al. , 2013 ). These effects are likely compounded in women with endometriosis who have additional infertility factors. For example, diminished ovarian reserve or male infertility significantly decrease the likelihood of conception ( Harb et al. , 2013 ; Dunselman et al. , 2014 ; Tanbo and Fedorcsak, 2017 ). In our study, women with endometriosis-plus infertility had lower oocyte retrieval and fertilization rates than those with endometriosis-only or other-infertility diagnoses, suggesting the combined effects of inflammation, oxidative stress and gamete quality in multiple factor infertility are likely to contribute to these poorer outcomes. A population-based retrospective cohort study of 291 244 embryo transfer cycles reported higher cancellation rates among women with endometriosis and concomitant diagnosis (11.3%) compared to women with isolated endometriosis (8.5%) and women with other infertility diagnoses (8.9% for tubal infertility and 8.1% for unexplained infertility) ( Senapati et al. , 2016 ), indicating a poorer prognosis. In our analysis, the cancellation rate for the endometriosis-plus group (8.4%) was significantly higher than in the endometriosis-only group (6.8%) but lower than in the other-infertility group (9.1%). Few studies are available for direct comparison with our results that specifically estimate CLBRs among women with endometriosis after a defined number of complete ART cycles. Most prior research has calculated CLBRs either on a per-patient basis or per complete cycle, rather than following a specific number of complete cycles ( Feichtinger et al. , 2019 ; Boucret et al. , 2020 ; Zhou et al. , 2022 ; Zimmermann et al. , 2023 ). However, these approaches do not provide insights into success probabilities after a specific number of completed ART cycles, a key area for patient-centred decision-making. Given that success rates are generally highest in the first cycle and progressively decline in subsequent cycles, having cycle-specific success data becomes critically important ( McLernon et al. , 2016 ; Chambers et al. , 2017 ). Such detailed data not only provide a clearer picture of how outcomes evolve over time but also serve as a valuable tool for shared decision-making between clinicians and patients ( Maheshwari et al. , 2015 ). It helps clinicians provide realistic guidance while empowering patients with cycle-specific success rates to make informed decisions, balancing parenthood aspirations with the emotional, physical and financial demands of continued treatment. In our study, we reported two types of CLBR estimates—conservative and optimal. The conservative CLBR assumes that women who discontinued treatment had a zero probability of achieving a live birth if they had continued, although this assumption is overly pessimistic since not all patients who discontinue treatment would have zero chance of success. Conversely, the optimal estimate assumes that women who discontinue treatment have the same probability of live birth as those who continue, which is overly optimistic, as patients who discontinue often do so due to a poor prognosis. The range between these two estimates offers a reasonable assessment of the probability of live births after a specified number of complete cycles. This range allows clinicians to tailor their counselling to individual patients, particularly those with better prognosis (e.g. relative short period of infertility and no co-morbidities), who are more likely to approach the optimal estimates ( Maheshwari et al. , 2015 ). While these population-based tables provide guidance on age-specific ART success rates for women with endometriosis, there are a number of unmeasured factors (e.g. BMI and duration of infertility) which influence ART success rates but are not explicitly included in this analysis. Furthermore, what is considered an acceptable ART success rate for patient and clinicians depends on a number of non-medical factors (e.g. cost, phycological burden) which should be considered to make shared decisions about fertility treatment options. A key strength of this study is the large, complete, population-based dataset, which allows for robust comparisons of ART outcomes across multiple infertility diagnoses. However, there are limitations. First, we were unable to account for the severity and phenotype of endometriosis, which may influence ART outcomes. Advanced-stage endometriosis (Stage III/IV) is often associated with reduced ovarian reserve and impaired endometrial receptivity, leading to poorer outcomes, while mild cases (Stage I/II) may show comparable results to other single-factor infertility diagnoses ( Kuivasaari et al. , 2005 ; Harb et al. , 2013 ). Furthermore, different endometriosis phenotypes, such as superficial peritoneal disease, deep infiltrating endometriosis or ovarian endometriomas, variably affect fertility and ART success ( Vercellini et al. , 2023 ). Second, the diagnosis of endometriosis in this study was made by treating physicians in IVF clinics, based on clinical assessments and thus not uniformly made across patients. This diagnosis will often be reliant on a previous laparoscopy, imaging findings (such as endometrioma or deeply infiltrating disease) or clinical history. Endometriosis is often underdiagnosed or misdiagnosed, especially in the setting of infertility, due to its heterogeneous presentation, and reliance on clinical evaluation alone may lead to inconsistencies ( Kuligowska et al. , 2005 ). Furthermore, a registry-based study cannot account for surgical variability in accurately diagnosing endometriosis or determining whether tubal disease may be attributable to endometriosis. Third, the endometriosis-only group in our study may reflect a subset of women with milder disease, as those with co-existing gynaecological conditions were classified into the endometriosis-plus group. Endometriosis is increasingly understood as part of a broader ‘uterine syndrome’, overlapping with conditions such as adenomyosis, fibroids and endometrial polyps ( Brosens and Benagiano, 2011 ). Adenomyosis, in particular, is so closely linked to endometriosis that it is sometimes considered a related manifestation of the disease, with emerging evidence highlighting its role in the development of endometriosis-related symptoms, including pain and infertility ( Vercellini et al. , 2023 ). However, due to the lack of data on these conditions in ANZARD, we were unable to classify women with overlapping pathology separately. As a result, the endometriosis-only group may represent a more favourable clinical profile, introducing the potential for selection bias and more favourable ART outcomes. This limitation should be considered when interpreting the relatively higher ART outcomes observed in this group. Fourth, in our study, the heterogeneity of other-infertility group, which includes a large proportion of women with 'other female factors’ (28%) and ‘unexplained infertility’ (24%) recorded in ANZARD, poses a challenge in interpreting comparative ART outcomes. The 'other female factors’ category likely encompasses diverse conditions including anovulation, diminished ovarian reserve and uterine abnormalities. Some of these conditions, particularly low ovarian reserve, are known to negatively influence ART outcomes ( Busnelli et al. , 2021 ; Zhu et al. , 2024 ). As such, the apparently higher success rates in the endometriosis-only group may partly reflect poorer prognosis in parts of the other-infertility group rather than a truly superior outcome profile. This potential confounding should be considered when interpreting the relative differences in CLBRs. Finally, although statistically significant differences in certain baseline characteristics, such as maternal age, were observed across groups, these differences were generally small and may not be clinically meaningful. In large cohorts, even minor variations can reach statistical significance; therefore, such findings should be interpreted with caution. This study provides valuable insights into ART success rates in women with endometriosis as a sole cause of infertility, those with endometriosis combined with other infertility factors, and those with infertility due to non-endometriosis causes. Women with isolated endometriosis achieved success rates comparable to those with other causes of infertility; however, women with endometriosis combined with additional infertility factors had lower success rates and higher pregnancy loss. Success rates declined significantly with advancing age and increasing numbers of ART cycles, with the highest success rates observed within the first three complete egg retrieval cycles. These findings are critical for patient counselling and highlight the importance of early intervention and tailored treatment strategies in women undergoing ART. Further research is needed to explore the role of endometriosis severity in influencing ART outcomes and to develop strategies for improving success rates in women with multiple infertility diagnoses.