{"paper_id":"a4157a07-f030-49e2-9919-c296ff644949","body_text":"Endometriosis is a chronic inflammatory condition affecting approximately 4%–10% of reproductive-aged women worldwide, characterised by the presence of endometrium-like tissue outside the uterus. 1 3  This condition is associated with various symptoms, primarily including dysmenorrhea, chronic pelvic pain and infertility that notably worsen women’s well-being. 4 6  The pathophysiology of endometriosis remains poorly understood, although there is growing evidence that retrograde menstruation and immune dysregulation may play key roles.\nAmong the manifestations of endometriosis, ovarian function is a significant concern. Women with the condition often experience impaired ovarian reserve mainly due to ovarian endometriomas and their laparoscopic excision, and the inflammatory response of the immune response. 47 10\nAnti-Mullerian hormone (AMH), mainly produced by granulosa cells of small antral and pre-antral ovarian follicles, is a key marker of the size and quality of the primordial follicle pool. 11  AMH levels increase from birth, plateau at approximately age 25 years, and decline with advancing age, becoming undetectable at menopause, showing a strong negative correlation with age. 12 15  This pattern, which remains relatively stable throughout the menstrual cycle and is unaffected by hormonal medications, makes AMH levels for age one of the most reliable markers of functional ovarian reserve, better than age alone or other ovarian markers. 16\nEmerging evidence suggests that women with certain conditions may exhibit distinct patterns of AMH decline. For instance, studies have shown that the decline in AMH levels is slower over time in women with polycystic ovary syndrome (PCOS) compared with normo-ovulatory women. 17 18\nHowever, assessing ovarian reserve and AMH patterns in women with endometriosis is still a matter of debate. The results of studies focusing on the AMH level in women with endometriosis compared with healthy population remain insufficient and inconsistent. Moreover, no studies have explored the association between AMH and endometriosis in relation to age. In this respect, Campos  et al 19  showed that infertile patients with mild endometriosis have a similar follicular fluid AMH level compared with control subjects. Similarly, Inal  et al 20  reported comparable AMH levels in infertile women with endometrioma and infertile controls. In contrast, a prospective study by Uncu  et al  reported that women with endometrioma had lower AMH levels compared with healthy women. 21  Supporting this, Kasapoglu  et al 22  revealed that women with endometrioma experience a progressive decline in serum AMH levels, which is faster than that in healthy women. Interestingly, Marcellin  et al 23  reported an opposite finding, showing that serum AMH levels increased with the size of endometriomas in women with endometriomas. In another study, Feferkorn  et al 24  conducted a study involving 656 women, including 71 with endometriosis, to assess the combined effects of age and endometriosis on ovarian reserve. AMH levels were measured for the entire group as well as stratified by age (35 years or older). The results showed significantly lower levels of AMH in women with endometriosis compared with those without. Notably, the association between lower AMH and endometriosis was significant only in women aged 35 or older.\nAdditionally, considering the unique pattern of AMH levels and its close association with age, there are currently no studies assessing the relationship between AMH levels and age among women with endometriosis. Specifically, it remains unclear whether there are significant critical points in the rate of decline of AMH levels associated with ageing. Therefore, this study aimed to examine the associations between age and AMH levels in women with endometriosis and healthy controls.\n\nThe participants in our study were selected from the Tehran Lipid and Glucose Study (TLGS). The TLGS is an ongoing, long-term, population-based prospective study initiated in 1998 to assess the prevalence and risk factors of non-communicable diseases in a representative population sample of Tehran, Iran. A total of 15 005 individuals, aged ≥3 years, were followed every 3 years to document data on demographic, anthropometric, reproductive and metabolic characteristics, general physical examinations as well as laboratory measurements. More details have previously been published elsewhere. 25 26\nFor the purpose of the current study, we used data collected during the sixth follow-up visit (2015–2018) from all women aged 18–48 years who completed a supplementary reproductive questionnaire. This questionnaire included self-reported endometriosis diagnosis. 25  In cases where participants reported a positive history of endometriosis diagnosis, their medical records were subsequently reviewed to confirm the diagnosis. Additionally, all participants underwent an ultrasound assessment using either a 3.5-MHz transabdominal transducer (for virgin individuals) or a 5-MHz transvaginal transducer for others.\nParticipants were classified as endometriosis cases if they self-reported a diagnosis, which was then verified through their medical records. Furthermore, women with a positive ultrasound diagnosis of endometriosis (with or without symptoms such as chronic pelvic pain, dysmenorrhea or dyspareunia) were included in the endometriosis group. The ultrasound criteria for endometriosis were assessed using the systematic approach developed by the International Deep Endometriosis Analysis group, which offers standardised definition and examination protocol. 27\nIndividuals experiencing chronic pelvic pain, dysmenorrhoea or dyspareunia but with negative ultrasound findings were excluded, as a negative result may not conclusively rule out endometriosis, especially superficial peritoneal disease, in these individuals.\nParticipants were classified as controls if they presented no symptoms of endometriosis and had negative ultrasound results. This classification aligns with the 2022 ESHRE (European Society of Human Reproduction and Embryology) guidelines, 28  which outline the diagnostic process for endometriosis and reassess the role of laparoscopy and histology as gold standard diagnostic tests. The guidelines suggest that laparoscopy should now be reserved for symptomatic patients with negative imaging results and/or those for whom empirical treatment has been ineffective or inappropriate. We recognise that some asymptomatic individuals with negative imaging results might be diagnosed with endometriosis via laparoscopy, although such cases are rare. Therefore, routine laparoscopic screening is not recommended for these participants. Consequently, the risk of including undiagnosed cases in the control group is minimal, and any misclassification is likely to have a minimal effect on the validity of our study’s findings.\nIn addition, women with a history of endocrine disorders, PCOS, hysterectomy, oophorectomy or any other type of ovarian surgery that could potentially influence AMH levels were excluded from participation in both study groups. Details of the PCOS diagnosis have been published elsewhere. 29\nOvernight fasting blood samples were collected for each participant. AMH was assessed using a two-site enzyme immunoassay technique with the Gen II kit (Beckman Coulter, Fullerton, California, USA) and the Sunrise ELISA reader (Tecan, Salzburg, Austria). AMH Gen II controls A79766 were used at two different concentration levels to assess the accuracy of the assay. The coefficients of variation for intra-assay and inter-assay were 1.9% and 2.0%, respectively.\nBaseline demographic and clinical characteristics of participants are described and compared based on endometriosis and control women. Continuous variables are presented as mean (SD) and compared using the Student’s t-test if they had normal distribution and compared with the Mann-Whitney test for those with skewed distribution. One-sample Kolmogorov-Smirnov test was used for checking the normality of continuous variables. Categorical variables are presented as numbers (n) and percentages (%) and compared between groups by χ 2  test or Fisher’s exact test (for tabs with sparse data).\nIn our study, the primary outcome variable was AMH levels, while the main independent variable was age. We aimed to explore the relationship between these two variables, with the additional consideration of case/control status (endometriosis vs control) as a covariate. Specifically, we employed two modelling approaches: a quadratic regression model and a segmented regression model.\nIn comparing the linear and quadratic models for evaluating the relationship between age and AMH, we selected the quadratic model based on its higher R² values in both the healthy (0.448) and endometriosis (0.358) groups compared with the linear model (0.415 and 0.334, respectively). To further validate the choice of the quadratic model, we calculated the AIC (Akaike Information Criterion) for both models: AICQuadratic=3218.2 and AICLinear=3246.1, which confirmed our model selection. A scatterplot with quadratic regression fitted lines was used to visualise the relationship between age and AMH in two study groups. We explored a more exact association by fitting segmented regression to evaluate potential breakpoints in the relationship between age and AMH in both endometriosis and control women. In this respect, segmented linear regression was performed to find if there was a threshold at which point a significant change in the slope of the relation between AMH and age ( A M H = β 0 + β 1 ( A g e ) + ε ) could be identified. The breakpoint is identified where the slope of the linear relationship changes, using the least squares method to minimise the sum of squared differences. More precisely, in this model, the independent variable is partitioned into intervals based on breakpoints where its relationship with the dependent variable changes, and fits multiple linear regression models to different segments of the data by a distinct function (fi( X )). This statistical method has been detailed in previous studies. 30 31  We tested whether the slopes were significantly different above and below the threshold by Davies test. If this difference was significant, subgroup analysis (linear regression) of each study was completed to determine if it affected the relationship between AMH and age below and above threshold, and analyses were completed separately for endometriosis and control women. Models were fitted crudely and adjusted for body mass index (BMI), parity and smoking status as potential influential variables on the association between AMH reduction and age.\nStatistical analysis was performed using SPSS software V.26 and the ‘Segmented’ package of R software V.4.2.2 (R Foundation for Statistical Computing,  www.R-project.org ). The two-tailed p values <0.05 were considered statistically significant.\n\nA total of 1005 women aged 18–48 years participated in the study, including 305 (30.3%) women with endometriosis and 700 (69.7%) healthy controls. The mean age and BMI of women with endometriosis were comparable to those of healthy controls (age: 33.31 vs 33.61 years, p=0.254; BMI: 26.14 vs 26.68 kg/m², p=0.189). Women with endometriosis had significantly lower AMH levels compared with healthy controls (1.99±1.42 vs 2.30±1.61 ng/mL; p=0.029). The characteristics of the study participants, both overall and according to endometriosis status, are presented in  table 1 .\nBold values indicate statistical significance.\nAMH, anti-Mullerian hormone; BMI, body mass index; n, number of participants.\nFigure 1  shows the point-by-point relationship between AMH levels and age in women with endometriosis compared with healthy controls, as illustrated by the scatter plot, and compares their respective quadratic regressions. The model revealed an R² value of 0.448 for the endometriosis group and 0.355 for the control group, indicating that approximately 44.8% and 35.5% of the variance in AMH levels can be explained by the model for each group, respectively, showing a good fit for both groups. Although AMH levels are lower in women with endometriosis compared with healthy controls, both groups show a decline in AMH levels with advancing age.\nThe relationship between AMH levels and age was examined using both linear and segmented regression models for endometriosis and healthy participants, as presented in  table 2 .\nAdjusted for BMI (kg/m2), parity and smoking status.\nAMH, anti-Mullerian hormone; BMI, body mass index.\nIn healthy controls, linear regression analysis demonstrated a significant negative association between AMH levels and age. Each 1 year increase in age was associated with a 0.15 ng/mL decrease in AMH levels (95% CI: −0.17 to −0.14; p<0.001). Segmented regression analysis identified a significant threshold at age 27 years (1.92), where the slope of the AMH-age relationship significantly changed, as determined by Davies’ test (p<0.001). Below this age, AMH levels declined by 0.35 ng/mL per year (95% CI: −0.47 to −0.23; p<0.001), while after 27 years, the decline was 0.12 ng/mL per year (95% CI: −0.14 to −0.10; p<0.001) ( figure 2 ).\nThe findings from the adjusted models, accounting for BMI, parity and smoking status, align with those of the unadjusted models, revealing a significant negative association between AMH levels and age in healthy controls, as indicated by linear regression analysis. Each 1 year increase in age was associated with a 0.15 ng/mL decrease in AMH levels (95% CI: −0.18 to −0.13; p<0.001). Segmented regression analysis revealed a significant breakpoint at age 27 years (1.78), where the slope of the relationship between AMH and age underwent a notable change, as confirmed by Davies’ test (p<0.001). Below this threshold, each additional year of age was associated with a decrease of 0.38 ng/mL in AMH levels (95% CI: −0.53 to −0.22; p<0.001), compared with a decrease of 0.12 ng/mL per year above the threshold (95% CI: −0.14 to −0.10; p<0.001).\nIn individuals with endometriosis, linear regression analysis showed a decline in AMH levels with increasing age, where each 1 year increase in age was associated with a 0.16 ng/mL decrease in AMH (95% CI: −0.17 to −0.14; p<0.001). Segmented regression analysis identified a threshold at 31 years (4.28) in this group, but Davies’ test did not reveal a statistically significant change in the slope (p=0.984), indicating a consistent decline in AMH levels across the age range in this group ( figure 2 ).\nFigure 3  was created to merge the two relationships into a single graph for a more precise comparison and enhance the clarity of the differences in AMH decrease by ageing between the two groups. This visualisation highlights that the AMH levels in the endometriosis group consistently fall below those of the healthy control group across all age ranges.\n\nIn this population-based study, we investigated the association between AMH levels and age in women with endometriosis. Our findings revealed that (1) AMH levels were lower in women with endometriosis compared with healthy controls; (2) both groups exhibited a significant negative association between AMH levels and age, indicating a decline in AMH levels with increasing age and (3) while healthy controls exhibited a significant threshold at 27 years, characterised by a change in the rate of decline of AMH levels, no such threshold was observed in women with endometriosis, indicating a more consistent decline within this group.\nEndometriosis is one of the common oestrogen-dependent inflammatory disorders in reproductive-aged women, which causes pelvic scarring, pain and infertility. 4  It is shown that retrograde menstruation, benign metastasis, immune dysregulation, coelomic metaplasia, hormonal imbalance, involvement of stem cells, and alterations in epigenetic regulation contribute to the pathogenesis of this disorder. 32 33\nThe assessment of ovarian reserve in patients with endometriomas remains a subject of controversy. While some studies have reported that women with endometriosis exhibit reduced levels of AMH compared with control subjects, 21 22  other research has indicated the contrary, demonstrating that serum AMH levels are similar to or even increased with the size of endometriomas in women affected by this condition. 19 20\nOur study aligns with some previous research showing that AMH levels in women with endometriosis are lower than in the healthy population. In this respect, two systematic reviews and recent meta-analyses 34 35  reported that AMH levels were significantly lower in patients with endometriomas compared with those without endometriomas. In this respect, it is shown that endometriosis could reduce follicular density and increase atresia. 36  High levels of reactive oxygen species, free iron and proteolytic enzymes in endometrioma fluid may infiltrate surrounding tissue, causing fibrosis and follicular loss, thereby accelerating ovarian reserve decline. This damage, regardless of surgical excision, may explain why AMH levels are lower in women with endometriomas compared with those with non-endometriotic cysts or healthy ovaries. 3436 39\nHowever, findings provide new insights by demonstrating the distinct decline pattern of AMH levels associated with ageing in women with and without endometriosis. Our findings indicate that the rate of AMH decline with age differs between the control and endometriosis groups. Specifically, in the control group, the slope of AMH decline decreases after age 27, whereas in the endometriosis group, the slope remains constant. This non-linear pattern of AMH decline with age is consistent with observations from previous studies, which have reported varying patterns. 1 3  For instance, Kelsey  et al 40  found that AMH peaks at approximately 24.5 years, while Lie Fong  et al 14  reported a peak at around 15.8 years and de Kat  et al 41  in the Doetinchem Cohort Study, reported that the acceleration of AMH decline occurs after age 40. Notably, visual inspection of the data presented by Kelsey  et al  suggests that the rate of AMH decline slows around age 30 compared with earlier ages.\nOur observation of a threshold at age 27 in the control group aligns with evidence suggesting early ovarian ageing dynamics. Specifically, the rate of follicular depletion may accelerate in the late 20s to early 30s in women, transitioning from a rapid loss of smaller follicles to a slower decline as the reserve diminishes. Our finding of a steeper decline before age 27 (slope: −0.35, 95% CI: −0.47 to −0.23; p<0.001), followed by a more gradual decrease afterwards (slope: −0.12, 95% CI: −0.14 to −0.10; p<0.001), may capture this shift in follicular dynamics. This pattern supports the notion that the late 20s represent a critical period in ovarian ageing, during which changes in follicular depletion rates may influence AMH levels.\nAccordingly, this may hypothetically raise an important clinical consideration, as while AMH is a valuable marker for assessing ovarian reserve in the general population, it may not be the best marker for ovarian reserve in women with endometriosis, as its behaviour may differ in the context of this condition.\nIn line with this hypothesis, Inal  et al 20  found that women with endometrioma had a significantly lower number of oocytes retrieved than the controls despite the same AMH levels in both groups and suggested that antral follicle count (AFC) may be a better marker of ovarian response than AMH in women with endometrioma. In another study, Ersahin  et al 42  assessed the clinical value of AFC and AMH for the prediction of ovarian response in women with endometrioma. They reported that AFC is more sensitive than AMH in detecting ovarian response in women with ovarian endometrioma.\nThe identification of a breakpoint at 27 years in healthy controls further substantiates the non-linear relationship between AMH levels and age in women without endometriosis. This finding is consistent with prior studies indicating that the decline of AMH levels accelerates during the late 20s. 13 14 40  Conversely, the absence of a comparable breakpoint in women with endometriosis suggests that the condition may supersede age-related physiological changes in AMH decline. This uniform decline across all ages may reflect cumulative damage to ovarian tissue resulting from the chronic and progressive nature of endometriosis. 43\nThe primary strengths of the present study lie in its methodological framework, characterised by a large population-based design and the application of advanced statistical techniques. The utilisation of segmented regression analysis in our investigation enabled the precise identification of breakpoints within the relationship between AMH levels and age. This sophisticated statistical approach enhances the robustness of our findings by effectively accommodating potential non-linear patterns in the data, which may have been overlooked if a traditional linear model had been employed.\nFurthermore, the reliability of AMH measurements was ensured through minimal intra-assay and inter-assay variability, achieved by employing a unique AMH assay kit and conducting all measurements within a single laboratory setting. This methodological rigour not only strengthens the validity of our results but also contributes to the reproducibility of our findings in future research.\nIt is essential to acknowledge the limitations of our study. The primary limitation is the lack of comprehensive assessment of all endometriosis cases through laparoscopy, which is considered the gold standard for definitive diagnosis of this condition. Additionally, our dataset lacks sufficient reliability for the precise identification of endometriosis severity. We also did not include AFC measurements, which could have provided a more comprehensive evaluation of ovarian reserve. Moreover, while our analysis accounted for age-related changes in AMH levels, we do not have adequate statistical power to further categorise our study participants based on variables such as disease duration, severity of endometriosis and prior medical or surgical interventions. Future studies should strive to incorporate these variables to enhance understanding of their influence on AMH decline within this population. For AMH measurements, we used stored serum samples that were not collected on specific days of the menstrual cycle. However, we contend that this limitation has minimal impact on our findings, as serum AMH levels are generally regarded as independent of the menstrual cycle, 44 45  and previous research indicates that long-term storage does not significantly affect sample integrity. 46 47  Furthermore, we excluded individuals with mild chronic pelvic pain, dysmenorrhoea or dyspareunia if their ultrasound imaging showed no signs of endometriosis. Since negative ultrasound findings do not rule out endometriosis, particularly in cases of superficial peritoneal disease, our findings may not fully represent individuals with the milder forms of endometriosis. Additionally, the majority of our participants had both ovarian and peritoneal endometriosis, while cases of isolated ovarian or peritoneal endometriosis were limited. As a result, conducting a robust subgroup analysis based on endometriosis location was not feasible, and we were unable to perform subgroup analyses based on the type or location of endometriosis. This limitation prevented us from investigating whether the observed differences in AMH levels vary between these subtypes.\nAdditionally, our study population was confined to a specific geographic region, which may limit the generalisability of our findings to broader populations. Further, the cross-sectional design of the study limits our ability to establish causality. Finally, the cross-sectional design of this study precludes any determination regarding whether observed changes in AMH levels are attributable to the progression of endometriosis progression or the effects of treatment interventions.\n\nIn conclusion, this study provided novel insights into the relationship between AMH levels and age in women with endometriosis. Our study showed that women with endometriosis had significantly lower AMH levels compared with healthy controls and did not exhibit the age-related threshold observed in the control group, where a sharper decline in AMH occurred before 27 years of age. These findings suggest that endometriosis may disrupt the typical pattern of AMH, which may be important clinical considerations for assessing ovarian reserve in this population. Clinicians should exercise caution when interpreting AMH levels in women with endometriosis, as conventional age-related patterns of ovarian reserve may not apply. Further research is needed to validate these findings in diverse populations and to explore whether alternative biomarkers or assessment strategies might offer a more accurate evaluation of ovarian reserve in this patient group.","source_license":"public-domain-us","license_restricted":false}