Does endometrioma diameter impact serum anti-Müllerian hormone levels? A systematic review and meta-analysis

Endometriosis is a globally prevalent gynecological disorder, affecting nearly 1 in 10 women of reproductive age, with an estimated 200 million cases worldwide ( Zondervan et al. , 2020 ). The disease features the presence of ectopic endometrial-like tissue outside the uterine cavity, which is estrogen-dependent and often exhibits resistance to progesterone. This aberrant tissue triggers a chronic inflammatory response, which can have detrimental effects on adjacent organs ( Marquardt et al. , 2019 ; Taylor et al. , 2021 ). A distinct diagnosed type of the disease is the ovarian endometrioma, which may be indicative of advanced-stage endometriosis ( American Society For Reproductive Medicine, 1997 ; Guerriero et al. , 2016 ). Epidemiological data suggest that unilateral and bilateral endometriomas are observed in up to 55% and 28% of affected individuals, respectively ( Jenkins et al. , 1986 ; Vercellini et al. , 1998 ). The mechanisms by which endometrioma impairs fertility and depletes ovarian reserve remain incompletely elucidated. Current evidence points to chronic inflammation, oxidative stress, cortical fibrosis, and follicular attrition or burnout as principal culprits ( Kitajima et al. , 2011 ; Sanchez et al. , 2014 ; Wu et al. , 2019 ; Wyatt et al. , 2023 ). In parallel, a mechanical component such as pressure, stretching, or distortion of the ovarian cortex by the endometrioma itself has long been suspected ( Maneschi et al. , 1993 ; Kitajima et al. , 2011 ; Leone Roberti Maggiore et al. , 2017 ). However, whether these mechanisms increase in proportion to endometrioma size remains unresolved. This is not only of pathophysiological interest but also of direct clinical relevance in reproductive-aged women, especially those facing infertility or delayed childbearing. In this context, choosing between expectant, medical, or surgical management of endometrioma requires weighing potential benefits against the risks to ovarian reserve. Furthermore, although the impact of an intact endometrioma on ovarian reserve remains debated ( Streuli et al. , 2012 ; Benaglia et al. , 2017 ; Kasapoglu et al. , 2018 ; Muzii et al. , 2018 ), growing evidence indicates that endometriotic cystectomy may significantly impair ovarian reserve, with greater loss seen in cases of bilateral disease or repeat surgery ( Raffi et al. , 2012 ; Ferrero et al. , 2015 ; Muzii et al. , 2015 ; Younis et al. , 2019 ). Among available markers, recent data suggest that in the context of endometrioma surgery, serum anti-Müllerian hormone (AMH) is a more sensitive measure than antral follicle count (AFC) for assessing ovarian reserve in women with endometrioma ( Younis et al. , 2022 ). However, whether the size of endometrioma itself influences AMH levels remains a matter of active debate. Over the past decade, studies examining this relationship have produced conflicting results. While some studies have identified a negative correlation between endometrioma size and serum AMH levels ( Ergun et al. , 2015 ; Karadağ et al. , 2020 ), others have found no change ( Hirokawa et al. , 2011 ; Bourdon et al. , 2022 ). Adding further complexity, additional independent studies observed increased serum AMH levels in women with larger endometriomas ( Marcellin et al. , 2019 ; Roman et al. , 2021 ). Taken together, these findings underscore persistent uncertainty regarding the impact of endometrioma size on ovarian reserve. Many previous studies suffer from methodological limitations, including small sample sizes, heterogeneous inclusion criteria, and potential sampling bias. A targeted systematic review and meta-analysis are therefore warranted to clarify this relationship. High-quality evidence would clarify the pathophysiological impact of endometrioma size on ovarian reserve, providing valuable clinical insights. Such evidence would support informed decision-making for women, particularly those experiencing infertility or planning future pregnancies. It could also guide patient counseling regarding treatment timing and oocyte preservation strategies. This systematic review and meta-analysis aims to evaluate the impact of endometrioma diameter on ovarian reserve, as measured by serum AMH levels, in women without prior ovarian surgery.

We conducted our systematic review and meta-analysis in accordance with the MOOSE statement, the updated PRISMA 2020 guidelines, and adhered to the latest version of the Cochrane Handbook for Systematic Reviews of Interventions ( Stroup et al. , 2000 ; Page et al. , 2021 ; Higgins et al. , 2024 ). The study protocol was registered in advance with PROSPERO (International Prospective Register of Systematic Reviews) under registration number CRD42025632149. We conducted a comprehensive electronic database search using PubMed, ClinicalTrials.gov, the Cochrane Library, as well as Web of Science and EBSCO (as interfaces to databases) to identify research articles published between 1 January 2000 and 30 June 2025. EBSCOhost (short for Elton B. Stephens Company) is a large information platform that provides the ‘digital backbone’ for libraries worldwide. Our focus was on studies that examined the impact of ovarian endometrioma diameter on serum AMH levels in women of reproductive age. The systematic review and meta-analysis aimed to identify studies assessing serum AMH levels in women of reproductive age with varying endometrioma diameters, who had not undergone ovarian surgery, within the same population and setting. The literature search was confined to studies involving human subjects, and we included only full-text articles published in peer-reviewed English-language journals. The inclusion criteria consisted of studies that evaluated serum AMH levels based on endometrioma diameter in women of reproductive age who had not undergone surgery or other minimally invasive procedures. Eligible study designs included randomized controlled trials, case–control studies, cohort studies, and observational studies. To enhance data quality and reduce heterogeneity, we applied several exclusion criteria: studies involving pregnant women, women post-ovarian surgery (such as cystectomy or oophorectomy), ovarian cysts other than endometrioma, those who had undergone ablative or sclerotherapy procedures, or those who had received chemotherapy or radiotherapy. Additionally, we omitted studies that included women with polycystic ovary syndrome and other endocrine disorders, such as hyperprolactinemia and thyroid dysfunction, or women with irregular menstrual cycles, as well as those with a history of ovarian failure or pre-menopausal FSH levels. Studies that involved women taking oral contraceptives or other hormonal therapies at least 3 months before recruitment were also excluded. Furthermore, data presented solely as abstracts or at scientific meetings, as well as case reports, case series, reviews, perspectives, and opinion papers, were not considered. We used the following search terms to find eligible studies on search engines or database interfaces: (Endometrioma OR endometriosis OR endometriotic) AND (cyst OR ovarian OR size OR diameter OR volume OR effect OR diagnosis) AND (AMH OR anti-Müllerian OR antiMüllerian OR Müllerian inhibiting factor OR Müllerian inhibiting substance OR reserve). We carried out the study selection in two stages. In the first stage, we screened the titles and abstracts of the queried articles to determine their suitability. Two review authors, L.G. and N.S., independently selected appropriate abstracts for full-text evaluation. In the second stage, we chose eligible studies for inclusion in the quantitative analysis. Any disagreements regarding study inclusion in both stages were resolved through discussion and consensus or, if necessary, mediation by a third reviewing author (J.S.Y.). We also manually reviewed the references of eligible full-text studies to identify any additional relevant publications. To prevent multiple inclusion of the same data, we selected the latest or largest study and excluded overlapping studies. When data were missing, we emailed the first and last authors of the relevant publications, or the corresponding author, at least twice on separate occasions, with a minimum interval of 2 weeks between attempts, to request additional data. We extracted data from eligible studies using a standardized Excel template prepared by the review group. Our focus included the study’s methodology, population, and design, number of participating women, women’s ages, endometrioma diameter and laterality, serum AMH levels, AMH assay methodology, and data for evaluating risk of bias in each study. Serum AMH levels were standardized to ng/ml, using a conversion factor of 7.14 for values reported in pmol/l. For studies that reported only medians, ranges, and sample sizes, we applied established formulas for meta-analyses to estimate means and SDs ( Hozo et al. , 2005 ). When numerical AMH values were not explicitly provided, data were extracted from published figures. Two authors (L.G. and N.S.) independently collected data from the eligible articles. When disagreements arose, the final decision was reached through discussion and consensus, or through mediation by a third reviewer (J.S.Y.). To assess the risk of bias in the included non-randomized studies, we employed the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool, developed by the Cochrane Bias Methods Group ( Sterne et al. , 2016 ). This tool is recommended as a rigorous method for assessing the risk of bias in conducting systematic reviews ( Schünemann et al. , 2019 ). ROBINS-I is a comprehensive framework designed to evaluate potential bias in observational studies across seven domains: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of the reported result. Each study was independently assessed by two reviewers (N.S. and J.S.Y.) and categorized as having low, moderate, serious, or critical risk of bias in each domain. Discrepancies were resolved through discussion or consultation with a third reviewer (I.I.). This approach enabled a structured and transparent evaluation of internal validity, accounting for the methodological heterogeneity inherent to observational study designs. Continuous variables (age and AMH) were compared between laterality groups (unilateral vs unilateral/bilateral combined) using independent-samples t -tests, based on aggregated study-level data. The primary inferential analysis consisted of a cohort-level multivariable meta-regression including 30 independent, non-overlapping cohorts derived from the 16 eligible studies. Cohort independence was ensured by extracting mutually exclusive subgroups as defined in the original publications. Associations between endometrioma diameter and serum AMH concentrations were evaluated using two complementary modeling strategies: Continuous specification: Mean cyst diameter (cm) was entered as a continuous predictor to assess linear associations. Categorical specification : Diameter was categorized into four prespecified ordinal strata (6 cm) to evaluate potential non-linear or threshold effects. Categories were defined to ensure balanced representation across cohorts. Both models were adjusted a priori for mean maternal age, given its established biological relationship with ovarian reserve and its potential contribution to between-cohort heterogeneity. Random-effects meta-regression models were estimated using Restricted Maximum Likelihood (REML), providing conservative effect estimates while accounting for anticipated clinical and methodological variability across studies. Supportive secondary analyses were conducted using predefined diameter cut-offs (3–8 cm) based on groupings reported in the 16 eligible studies. As these comparisons involve repeated use of overlapping cohorts, they were prespecified as exploratory and hypothesis-generating. To assess the robustness of findings and explore potential sources of heterogeneity, subgroup analyses were performed according to key study characteristics, including study design, disease laterality, study population, AMH assay specification, sample processing strategy, and cyst measurement methodology. Multiplicity arising from repeated comparisons was controlled using the Bonforroni–Holm sequentially rejective procedure, maintaining a stringent family-wise error rate without excessive loss of statistical power. Statistical significance was defined as two-tailed P  < 0.05 after adjustment for multiplicity where applicable. Between-study heterogeneity was evaluated using Cochran’s Q statistic and quantified with the I 2 metric ( Higgins et al. , 2003 ). Given the limited number of studies in certain analyses, interpretation of Q statistics was supplemented by I 2 estimates, recognizing that Q may lack power in small meta-analyses. All analyses were conducted using SPSS Statistics version 29 (IBM Corp., Armonk, NY, USA) and OpenMeta [Analyst], available at www.cebm.brown.edu/openmeta .

Our search identified 3404 studies, of which 1870 were duplicates; 853 were excluded after reviewing the titles, and 970 after reading the abstracts ( Fig. 1 ). Of the final 47 full-text articles assessed for eligibility, 31 were excluded from further evaluation because they did not meet the inclusion and exclusion criteria. In 21 studies, the impact of endometrioma size on serum AMH levels was not assessed ( Ercan et al. , 2010 , 2011 ; Tsolakidis et al. , 2010 ; Lee et al. , 2011 ; Litta et al. , 2013 ; Uncu et al. , 2013 ; Urman et al. , 2013 ; Aboul Gheit et al. , 2014 ; Chen et al. , 2014 ; Kwon et al. , 2014 ; Saito et al. , 2014 , 2018 ; Tanprasertkul et al. , 2014 ; Ottolina et al. , 2017 ; Candiani et al. , 2018 ; Kasapoglu et al. , 2018 ; Kostrzewa et al. , 2019 ; Muzii et al. , 2019 ; Sweed et al. , 2019 ; Weng et al. , 2023 ; Azizova et al. , 2024 ). Flow diagram of identified studies . Five studies examined endometrioma diameter cutoffs in relation to serum AMH levels, but essential data were missing, preventing their inclusion in the quantitative review ( Abd Elhameid et al. , 2018 ; Dong et al. , 2019 ; Yilmaz et al. , 2021 ; Grigalashvili, 2022 ; Mansouri et al. , 2022 ). In four studies, serum AMH levels were not clearly reported ( Ding et al. , 2015 ; Marcellin et al. , 2019 ; Elizur et al. , 2023 ; Kang et al. , 2023 ). One publication was retracted ( Alborzi et al. , 2014 ). The authors of 17 publications were contacted by email at least twice to request missing data; only one responded and provided the data, which were included in the quantitative analysis ( Bourdon et al. , 2022 ). A total of 16 studies ( Table 1 ), involving 1484 women of reproductive age were included in the quantitative analysis, with a weighted mean age and weighted SD of 31.54 ± 2.18 years ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Sandya and Kumar, 2016 ; Mehdizadeh Kashi et al. , 2017 ; Wang et al. , 2019 ; Karadağ et al. , 2020 ; Yoon et al. , 2020 ; Roman et al. , 2021 ; Akgul et al. , 2022 ; Bourdon et al. , 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ; Sun et al. , 2025 ). Eligible studies for meta-analysis assessing the association between non-operated endometrioma diameters and serum AMH levels. AMH, anti-Müllerian hormone; A.A.AMH, automated access AMH; AFC, antral follicular count; DSL, Diagnostic Systems Laboratories; E2, estradiol; EIA, enzyme immunoassay; ELICA, Electro Chemi Luminescence Immuno Assay; ELISA, enzyme-linked immunoabsorbent assay; GEN II, second generation ELISA; IOT, Immunotech; MIS, Müllerian inhibiting substance; ND, not disclosed. Age range (mean was not disclosed). All eligible studies compared intact (non-operated) endometrioma diameters with serum AMH levels within the same setting. Ten eligible studies were prospective ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Sandya and Kumar, 2016 ; Mehdizadeh Kashi et al. , 2017 ; Wang et al. , 2019 ; Karadağ et al. , 2020 ; Roman et al. , 2021 ), and six were retrospective in design ( Javedani Masroor et al. , 2015 ; Mehdizadeh Kashi et al. , 2017 ; Yoon et al. , 2020 ; Bourdon et al. , 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ; Sun et al. , 2025 ). Six studies targeted women with unilateral endometriomas ( Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Mehdizadeh Kashi et al. , 2017 ; Karadağ et al. , 2020 ; Akgul et al. , 2022 ; Zareii et al. , 2023 ), whereas the others included both unilateral and bilateral endometriomas. The risk of bias of all eligible studies was evaluated using the ROBINS-I tool and is summarized in Table 2 . All seven domains were assessed, and the overall judgment was determined by the highest level of bias identified across them. As all included studies were observational, the domain of bias due to deviations from intended interventions was considered not applicable. Overall, two studies were rated as having a low risk of bias and 14 as moderate, indicating that the overall body of evidence provides a moderate to good level of certainty. ROBINS-I risk of bias assessment in eligible studies: the impact of Endometrioma diameter on serum anti-Müllerian hormone levels. No measurement formula for E size No stratification of E size by laterality No stratification of E size by laterality No measurement formula for E size Reproductive age not specified in selection criteria, but age range was 25–38. No stratification of E size by laterality No measurement formula for E size, AMH kit ND, Exclusion was not stated for previous hormone treatment No measurement formula for E size, 15 were excluded due to pregnancy, loss to follow-up, or lack of histological confirmation. No analysis of whether these exclusions introduced bias. Exclusion was not stated for previous hormone treatment The mean age was ND, only age range. Retrospective No stratification of E size by laterality Exclusion was not stated for previous hormone treatment No stratification of E size by laterality Retrospective study Exclusion was not stated for previous hormone treatment Retrospective study Exclusion was not stated for previous hormone treatment No stratification of E size by laterality Retrospective No stratification of E size by laterality AMH kit ND Retrospective No stratification of E size by laterality Exclusion was not stated for previous hormone treatment Retrospective No stratification of E size by laterality No measurement formula for E size AMH kit ND AMH, anti-Müllerian hormone; E, endometrioma; L, low; M, medium; NA, not applicable; ND, not disclosed. In all eligible studies, ultrasonography (US) was used to diagnose endometriomas and measure their sizes ( Table 3 ). In most cases, transvaginal US was employed; however, transabdominal US, transrectal US, computed tomography (CT) scan, and MRI were used in some cases. As outlined in Table 3 , in most studies, the ultrasonographic diagnosis of endometrioma was accomplished using standardized criteria ( Moore et al. , 2002 ; Kinkel et al. , 2006 ; Van Holsbeke et al. , 2010 ). Imaging and measurement approaches for endometrioma diagnosis and diameter evaluation in eligible studies. CT scan, computed tomography scan; MRI, magnetic resonance imaging; ND, not disclosed; TAUS, transabdominal ultrasound; TRUS, transrectal ultrasound; TVUS, transvaginal ultrasound; US, ultrasound. Two main methods were used to measure endometrioma diameter in the included studies. The first method involved averaging two or three perpendicular measurements ( Celik et al. , 2012 ; Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Wang et al. , 2019 ; Karadağ et al. , 2020 ; Yoon et al. , 2020 ), while the second method recorded the largest single dimension ( Roman et al. , 2021 ; Akgul et al. , 2022 ; Bourdon et al. , 2022 ; Zeng et al. , 2022 ). As shown in Table 3 , six studies did not disclose their measurement methodology. The AMH assays used across the included studies varied, likely reflecting differences in commercial availability and the timing of the study conduct. Details on assay types and their technical performance are summarized in Table 4 . Five studies adhered to best-practice protocols for analyte measurement by analyzing all AMH samples in a single batch (frozen samples), thereby minimizing inter-assay variability ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Giampaolino et al. , 2015 ; Wang et al. , 2019 ). In contrast, the other studies appeared to perform AMH measurements on a rolling basis as samples became available (fresh samples). There were no discernible differences in the direction or magnitude of findings between studies employing batch analyses and those using rolling measurements. AMH assay kits employed by the eligible studies AMH, anti-Müllerian hormone; EIA, enzyme immunoassay; ELICA, Electro Chemi Luminescence Immuno Assay; ELISA, enzyme-linked immunosorbent assay;  DSL, Diagnostic Systems Laboratories; GEN II, second generation ELISA; IOT, Immunotech; MIS, Müllerian inhibiting substance; ND, not disclosed. Frozen was selected as it was explicitly noted in the article’s methods. In five studies, the AMH assay methodology was not explicitly described ( Javedani Masroor et al. , 2015 ; Bourdon et al. , 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ; Sun et al. , 2025 ). In one study, multiple assays were used, all employing high-sensitivity ELISA techniques ( Roman et al. , 2021 ). The primary inferential analysis was a multivariable cohort-level meta-regression incorporating 30 independent cohorts derived from the 16 eligible studies ( Supplementary Table S1 ). This modeling strategy preserved statistical independence across cohorts and enabled adjustment for prespecified covariates. Endometrioma diameter was not independently associated with serum AMH concentrations. When modeled as a continuous variable, cyst diameter was not a significant predictor of AMH levels ( β  = 0.122, SE = 0.128, P  = 0.35). Similarly, when categorized into predefined diameter strata, including 6 cm (n = 206), no graded or threshold association with AMH depletion was observed ( β  = 0.245, SE = 0.220, P  = 0.28). In contrast, maternal age emerged as the dominant determinant of AMH variability, accounting for 58.2% of the between-cohort heterogeneity ( R 2 = 58.2%). Endometrioma size contributed negligibly to the explained variance in either modeling strategy. These findings were consistent across both continuous and categorical specifications of diameter, indicating that the absence of association between cyst size and ovarian reserve is independent of age and robust to alternative model parameterizations. Three studies focused on unilateral endometrioma ( Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Akgul et al. , 2022 ), three additional provided data on unilateral and bilateral cases separately ( Mehdizadeh Kashi et al. , 2017 ; Karadağ et al. , 2020 ; Zareii et al. , 2023 ), while 10 studies presented their data for unilateral and bilateral endometrioma combined ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Sandya and Kumar, 2016 ; Wang et al. , 2019 ; Karadağ et al. , 2020 ; Roman et al. , 2021 ; Bourdon et al. , 2022 ; Zeng et al. , 2022 ; Sun et al. , 2025 ). To examine whether endometrioma laterality impacted age and serum AMH levels, we compared six studies with data on unilateral endometrioma to those with data on bilateral cases and those with data on both unilateral and bilateral endometriomas combined. The two independent sample t -test indicated that there was no significant age difference between the two groups (unilateral compared to uni- and bilateral combined), with a mean difference of 1.517 years (95% CI: −4.30 to 2.20, t 13 = −0.70, P  = 0.48). In addition, no significant effect of endometrioma laterality on serum AMH levels was detected, with a mean difference of 0.529 ng/ml (95% CI: −1.05 to 2.11, t 14 = −0.373, P  = 0.74). To evaluate the robustness of the primary findings and explore potential non-linear or threshold effects, predefined diameter cut-offs (3, 4, 5, 6, 7, and 8 cm) were examined. As these comparisons repeatedly draw on overlapping cohorts, they were prespecified as supportive and exploratory. Multiplicity was controlled using the Holm–Bonferroni sequentially rejective procedure to maintain a stringent family-wise error rate. To reduce between-study heterogeneity and assess the consistency of findings across clinically and methodologically relevant strata, subgroup analyses were conducted according to key study characteristics derived from the 16 eligible studies. Analyses were restricted to prospective designs to evaluate methodological rigor; to unilateral endometriomas to minimize confounding by bilateral ovarian involvement; and to distinct study populations (infertility-only vs unselected endometrioma cohorts) to account for potential baseline differences in ovarian reserve. Methodological sources of variability were further examined by stratifying according to AMH assay specification, sample processing strategy (single-batch frozen sera vs fresh samples analyzed on a rolling basis), and endometrioma measurement approach (largest single diameter vs averaged orthogonal dimensions). These analyses allowed assessment of whether laboratory methodology or cyst measurement technique contributed materially to heterogeneity in reported associations. Across all prespecified subgroup domains (prospective designs, unilateral disease, infertility-only cohorts vs unselected endometrioma populations, AMH assay disclosure, frozen vs fresh sample analysis, and endometrioma measurement approach), the direction and magnitude of effects were generally small and inconsistent across thresholds, and none of the subgroup findings materially altered the primary inference that endometrioma diameter is not independently associated with AMH. After Holm–Bonferroni adjustment within each subgroup family, all pooled comparisons remained non-significant, including prospective-only analyses (smallest unadjusted P  = 0.24; Holm–Bonferroni-adjusted P  = 0.96), unilateral-only analyses (adjusted P  = 1.00 for all feasible thresholds), infertility-only analyses (adjusted P  = 0.75 for all feasible thresholds), AMH assay-disclosed analyses (adjusted P  = 1.00), and frozen versus fresh/rolling analyses (adjusted P  = 0.48 and 1.00, respectively, across feasible thresholds). One isolated signal was observed in the measurement-method stratum (largest single-diameter studies at the 3 cm cut-off: WMD −0.587 ng/ml; unadjusted P  = 0.009; Holm–Bonferroni-adjusted P  = 0.027), but this was not reproduced across other thresholds, occurred in sparse data, and was not supported by the primary independent-cohort meta-regression; accordingly, it is most consistent with methodological heterogeneity or instability of small subgroup analyses rather than a clinically meaningful diameter threshold (full numeric outputs in Table 5 ). Supportive subgroup and threshold meta-analyses of serum AMH levels by endometrioma diameter (random-effects), with heterogeneity metrics and Holm–Bonferroni-adjusted inference. Adjusted P -values were calculated using the Holm–Bonferroni step-down procedure to control for the family-wise error rate (FWER) across the meta-analytic outcomes. This method accounts for the increased risk of Type I errors arising from multiple comparisons within the same study population. CI, confidence interval; WMD, weighted mean difference; S, significant ( P   0.05).

This systematic review and meta-analysis, including 16 studies and 1484 women, demonstrates that the diameter of non-operated endometriomas is not independently associated with ovarian reserve, as assessed by serum AMH levels. In the primary multivariable cohort-level meta-regression incorporating 30 independent non-overlapping cohorts, endometrioma size was not a significant predictor of AMH concentrations, whether modeled as a continuous variable or categorized into predefined diameter strata (6 cm). In contrast, age consistently emerged as the dominant determinant of AMH variability, accounting for the majority of between-cohort heterogeneity. These findings were robust across all prespecified secondary analyses. No association between endometrioma diameter and AMH was observed in analyses restricted to prospective studies, unilateral disease, defined AMH assay methodologies, single-batch frozen versus fresh sample analysis, infertility-only populations, or according to the method used for cyst size measurement. After adjustment for multiplicity using the Bonferroni–Holm procedure, none of the exploratory threshold comparisons altered the primary inference. Collectively, the evidence indicates that any reduction in ovarian reserve observed in women with endometrioma appears to be independent of cyst diameter. These findings argue against a clinically meaningful mechanical or ‘mass-effect’ contribution to diminished ovarian reserve. Instead, they support the concept that any ovarian reserve impairment in endometriosis reflects underlying biological mechanisms, including chronic inflammation, oxidative stress, cortical fibrosis, and local cytokine-mediated disruption of the ovarian microenvironment ( Kitajima et al. , 2011 ; Sanchez et al. , 2014 ; Wu et al. , 2019 ; Wyatt et al. , 2023 ), rather than cyst size per se . Furthermore, these results are conceptually consistent with a previous systematic review and meta-analysis, which reported similar preoperative AMH levels in women with unilateral versus bilateral endometriomas ( Younis et al. , 2019 ). This concordance suggests that if any detrimental impact of endometriomas on ovarian reserve, it is not determined by size or laterality, but rather reflects a broader pathophysiological effect intrinsic to the disease. Our systematic review and meta-analysis have several notable strengths. To our knowledge, this is the first comprehensive quantitative synthesis specifically designed to evaluate whether endometrioma diameter influences ovarian reserve, employing strict methodology and serum AMH as the primary outcome. AMH is considered the most sensitive and clinically reliable biomarker of ovarian reserve in the context of endometrioma surgery ( Younis et al. , 2022 ). A key strength of our analysis is its deliberate focus on studies assessing AMH levels in non-operated ovaries, thereby avoiding the confounding effects of surgical intervention on ovarian reserve. Furthermore, we included only studies that evaluated serum AMH concentrations across different endometrioma diameters within the same setting, allowing for internally controlled comparisons. Our literature search was extensive and systematic, employing three search engines and two major database platforms. This thorough approach enabled the identification of 16 eligible studies comprising a total of 1484 cases. Notably, quality assessment revealed a low to moderate risk of bias in eligible studies, further strengthening the reliability and validity of our findings ( Schünemann et al. , 2019 ). While one prior review reported a negative association between endometrioma size and AMH levels, it was a narrative review and did not quantitatively assess the impact of cyst diameter ( Muzii et al. , 2023 ). Previous attempts to address this clinical question have been confined to individual observational studies, many of which have reported inconsistent or conflicting results ( Kasapoglu et al. , 2018 ; Marcellin et al. , 2019 ; Somigliana et al. , 2020 ; Roman et al. , 2021 ). In contrast, our analysis integrates data from multiple studies across diverse populations and methodological designs, providing a more robust and generalizable assessment of this potential relationship. Our systematic review and meta-analysis have several limitations. First, only six of the included studies examined women with unilateral endometriomas, whereas the remaining studies combined unilateral and bilateral cases. Notably, the mean age and serum AMH levels were comparable between these two groups. Furthermore, our subgroup analysis of studies limited to unilateral endometriomas yielded consistent findings, showing that endometrioma size, above or below various diameter thresholds, did not significantly impact AMH levels. These results are further supported by a previous systematic review and meta-analysis, which demonstrated that the laterality, as an indirect measure of endometrioma size burden, does not influence preoperative AMH concentrations ( Younis et al. , 2019 ). Therefore, while this heterogeneity in laterality represents a methodological limitation, it does not appear to compromise the validity of our overall conclusion that endometrioma size does not significantly affect ovarian reserve as measured by serum AMH levels. Second, 6 of the 16 included studies employed retrospective designs, which are inherently more vulnerable to selection bias, incomplete data capture, and unmeasured confounding compared with prospective investigations. The inclusion of retrospective data may therefore have contributed to methodological heterogeneity and residual bias. However, when analyses were restricted to the 10 prospective studies (n = 812), the pooled estimates remained directionally and quantitatively consistent with the overall findings. This concordance suggests that the primary conclusions were not materially driven by retrospective study inclusion and supports the robustness of the overall inference despite design-related limitations. Third, considerable between-study heterogeneity was observed in several supportive threshold-based subgroup analyses. This variability likely reflects unmeasured methodological and clinical differences across studies, including inter-assay variability in AMH measurement, e.g. automated versus manual platforms ( Nelson et al. , 2015 ; Li et al. , 2021 ), differences in endometrioma size assessment methodology, operator-dependent variability in ultrasound measurement, and inconsistency in sample processing and storage. Residual confounding related to differences in baseline ovarian reserve, within-cohort age distribution, bilaterality classification, and ovarian volume may also have contributed, despite adjustment for mean age at the cohort level. Importantly, cyst diameter may not fully capture underlying disease burden, such as cortical fibrosis, inflammatory activity, or concomitant deep infiltrating endometriosis, which could vary across populations. These sources of heterogeneity could not be harmonized at the study level and warrant cautious interpretation of secondary pooled estimates. Standardized prospective studies incorporating uniform AMH assays, harmonized ultrasound protocols, and detailed phenotyping of disease severity are needed to further refine these findings. Fourth, stratification according to the method of endometrioma size measurement revealed a statistically significant association in the subgroup of studies defining cyst size by the largest single diameter at the 3 cm threshold (WMD −0.587 ng/ml; 95% CI −1.028 to −0.147; unadjusted P  = 0.009; I 2 = 0%). This finding, however, was derived from only two studies and was not reproduced at alternative thresholds (4 or 6 cm) within the same measurement stratum. Importantly, it was also not supported by the primary multivariable cohort-level meta-regression, which incorporated all independent cohorts and adjusted for age. Accordingly, this isolated signal should be interpreted cautiously and is most consistent with sparse-data instability or methodological heterogeneity rather than evidence of a biologically meaningful diameter threshold. Fifth, this systematic review was restricted to non-operated endometriomas and therefore does not address whether preoperative cyst diameter influences the magnitude of postoperative AMH decline. Prior systematic reviews and meta-analyses have consistently shown that endometriotic cystectomy, particularly with the stripping technique, is associated with a significant reduction in serum AMH levels, reflecting postoperative impairment of ovarian reserve ( Raffi et al. , 2012 ; Somigliana et al. , 2012 ; Younis et al. , 2019 ). A recent critical appraisal of nine systematic reviews concluded that moderate- to high-quality evidence supports a sustained negative impact of cystectomy on ovarian reserve for up to 9–18 months, with a more pronounced decline in bilateral compared with unilateral cases (∼57% vs 39.5%) ( Younis and Taylor, 2024 ). However, whether endometrioma size modifies the extent of postoperative ovarian reserve loss remains uncertain. This unresolved question represents an important knowledge gap and warrants prospective studies specifically designed to evaluate the interaction between preoperative cyst diameter and postoperative AMH decline. In general, ovarian reserve can be assessed in cases with endometrioma using ovarian reserve biomarkers, such as serum AMH levels, as in our systematic review, or by evaluating ovarian response to controlled ovarian stimulation (COS) in the IVF setting. Over the last decade, multiple studies have investigated ovarian response to COS, yielding inconsistent results ( Coccia et al. , 2014 ; Benaglia et al. , 2017 ; Ferrero et al. , 2017 ; González-Foruria et al. , 2020 ; Somigliana et al. , 2020 ; Bourdon et al. , 2022 ). Some reports demonstrated an inverse association between endometrioma diameter and both the number of mature follicles and retrieved oocytes ( Coccia et al. , 2014 ; Ferrero et al. , 2017 ; González-Foruria et al. , 2020 ; Somigliana et al. , 2020 ), whereas others did not ( Benaglia et al. , 2017 ; Bourdon et al. , 2022 ). Importantly, the threshold for defining endometrioma diameter varied among the studies. Furthermore, the majority of studies were retrospective in design and included relatively modest cohorts of 26–101 women ( Benaglia et al. , 2017 ; Ferrero et al. , 2017 ; González-Foruria et al. , 2020 ; Somigliana et al. , 2020 ). Some focused specifically on non-operated ovaries with unilateral endometrioma ( Coccia et al. , 2014 ; Ferrero et al. , 2017 ; Somigliana et al. , 2020 ), while others did not. Notably, Bourdon et al. (2022) conducted a recent single-center, well-designed study with the largest cohort to date (n = 326), stratifying women into five groups based on the largest endometrioma diameter (<2, 2–4, 4–6, 6–8, and ≥8 cm) ( Bourdon et al. , 2022 ). They found no significant difference in the number of oocytes retrieved across these size categories, suggesting that endometrioma size does not affect ovarian reserve and supporting our conclusion. In parallel, several studies have evaluated ovarian reserve through serum AMH levels, but findings have also been contradictory ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Sandya and Kumar, 2016 ; Mehdizadeh Kashi et al. , 2017 ; Abd Elhameid et al. , 2018 ; Marcellin et al. , 2019 ; Wang et al. , 2019 ; Karadağ et al. , 2020 ; Yoon et al. , 2020 ; Roman et al. , 2021 ; Akgul et al. , 2022 ; Bourdon et al. , 2022 ; Grigalashvili, 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ; Sun et al. , 2025 ). Some studies reported a negative association between increasing endometrioma size and serum AMH levels ( Ergun et al. , 2015 ; Sandya and Kumar, 2016 ; Mehdizadeh Kashi et al. , 2017 ; Karadağ et al. , 2020 ; Sun et al. , 2025 ). Others observed no significant effect ( Hirokawa et al. , 2011 ; Javedani Masroor et al. , 2015 ; Wang et al. , 2019 ; Yoon et al. , 2020 ; Akgul et al. , 2022 ; Bourdon et al. , 2022 ; Grigalashvili, 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ), while some reported paradoxically higher serum AMH levels with increasing cyst diameter ( Celik et al. , 2012 ; Giampaolino et al. , 2015 ; Abd Elhameid et al. , 2018 ; Marcellin et al. , 2019 ; Roman et al. , 2021 ). Again, the threshold of endometrioma diameter was inconsistent across all studies, and most were limited by a relatively modest sample size (n = 26–104) ( Hirokawa et al. , 2011 ; Celik et al. , 2012 ; Ergun et al. , 2015 ; Giampaolino et al. , 2015 ; Javedani Masroor et al. , 2015 ; Sandya and Kumar, 2016 ; Mehdizadeh Kashi et al. , 2017 ; Karadağ et al. , 2020 ; Akgul et al. , 2022 ; Grigalashvili, 2022 ; Zeng et al. , 2022 ; Zareii et al. , 2023 ; Sun et al. , 2025 ). A previous systematic review and meta-analysis evaluating the effect of endometrioma size on ovarian reserve levels, including seven studies with a total of 603 cases, reported no significant change in serum AMH levels between small and large endometriomas ( Liu et al. , 2025 ). However, the review had several key methodological limitations. The largest included study (n = 193) was subsequently retracted from publication ( Alborzi et al. , 2014 ). Another study was published exclusively in Chinese without an accessible English translation ( Jie-Han et al. , 2022 ), which limited its appraisal and reproducibility. Moreover, the review did not exclude participants with a history of prior ovarian surgery and failed to account for laterality, both important confounders in the assessment of ovarian reserve. Most critically, the definitions of ‘large’ versus ‘small’ endometriomas varied across the included studies, with no consistent threshold applied. This lack of standardization introduced substantial methodological heterogeneity, thereby reducing the reliability of the pooled estimates and weakening the overall level of evidence. Taken together, available studies assessing the effect of endometrioma diameter on ovarian reserve, through ovarian response to COS and serum AMH levels, produced inconsistent and often contradictory results. Due to methodological heterogeneity and limited study quality, there has been no reliable evidence to date supporting a definitive link between endometrioma size and ovarian reserve. Our present systematic review, rigorously conducted in accordance with the updated PRISMA 2020 guidelines and the Cochrane Handbook for Systematic Reviews, showing a low to moderate risk of bias employing the ROBINS-I tool, clearly demonstrates that any detrimental effect of endometrioma on ovarian reserve is not related to cyst diameter. The observation that endometrioma diameter does not significantly correlate with serum AMH levels aligns with emerging histopathological evidence. While mechanical stretching and size-related factors have been traditionally blamed for ovarian reserve depletion, recent studies suggest that the extent of cortical fibrosis in the ovarian tissue neighboring the endometrioma is a more critical determinant of follicular loss. Specifically, research has shown that, along with age, the degree of cortical fibrosis is significantly and negatively correlated with serum AMH levels, whereas lesional fibrosis on the cyst itself is not ( Nie et al. , 2022 ). This reinforces the notion that if any detrimental impact of endometriomas on the ovary is driven by intrinsic pathophysiological processes, such as chronic inflammation and localized tissue remodeling, rather than the physical dimensions of the cyst. Persistent uncertainty regarding the clinical relevance of endometrioma diameter has historically shaped management guidelines, although evolving evidence has prompted a substantial paradigm shift. Earlier recommendations from professional societies, including the American Society for Reproductive Medicine (ASRM) and ESHRE, frequently endorsed a size-based threshold for surgical intervention. A widely cited position suggested that endometriomas larger than 4 cm should be excised to alleviate pain or potentially improve fertility outcomes ( Practice Committee of the American Society for Reproductive Medicine, 2012 ; Dunselman et al. , 2014 ). More recent guidance from these organizations has moved away from rigid diameter-based criteria, particularly in asymptomatic women and those pursuing fertility preservation or ART. The most recent ESHRE guideline ( Becker et al. , 2022 ) emphasizes individualized decision-making, recommending that surgical intervention be guided primarily by pain severity, patient age, ovarian reserve, prior treatment history, and reproductive goals, rather than cyst size alone. Notably, ESHRE no longer supports routine excision of endometriomas before ART in the absence of specific clinical indications. Similarly, contemporary ASRM committee opinions acknowledge that although larger cysts (e.g. >4 cm) may occasionally justify surgery for diagnostic clarification or to facilitate oocyte retrieval, the potential detrimental impact of cystectomy on ovarian reserve must be carefully balanced against anticipated benefits. In this context, our findings, demonstrating no independent association between endometrioma diameter and serum AMH levels, provide quantitative support for this modern, individualized management paradigm. Taken together, the evidence suggests that clinical decision-making should not rely on cyst size as a solitary determinant, but rather integrate symptomatology, reproductive planning, and baseline ovarian reserve within a patient-centered framework. This systematic review and multivariable cohort-level meta-analysis provide robust evidence that the diameter of non-operated endometriomas is not independently associated with ovarian reserve, as reflected by serum AMH levels. The consistency of findings across prespecified exploratory subgroup analyses further supports the conclusion that any endometrioma-related impairment in ovarian reserve is unlikely to be mediated by cyst size or a simple ‘mass-effect’. Rather, reduced ovarian reserve in affected women may likely reflects intrinsic pathophysiological mechanisms of endometriosis, including chronic inflammation, oxidative stress, and cortical remodeling. From a clinical perspective, these findings support an individualized management approach. The presence of a large endometrioma alone should not constitute an indication for surgical intervention in reproductive-age women, particularly when preservation of ovarian reserve is a priority. Instead, decision-making should integrate symptom burden, reproductive goals, age, baseline ovarian reserve markers, and informed patient preference. These data are relevant across several clinical contexts. In ART, cyst diameter alone should not be considered a determinant of ovarian reserve assessment or pre-ART surgical planning in asymptomatic women. Similarly, in the setting of fertility preservation, the presence of an asymptomatic endometrioma does not appear to imply a size-dependent reduction in ovarian reserve, reinforcing the primacy of age and baseline ovarian reserve parameters in counseling. For patients presenting primarily with pain, the decision to proceed with surgery should be guided by symptom severity and overall clinical context rather than adherence to a fixed diameter threshold. Future prospective studies are needed to determine whether preoperative endometrioma diameter modifies the magnitude of postoperative AMH decline following cystectomy. Clarifying this relationship will be essential for refining fertility-preserving surgical strategies and optimizing individualized care.