'Age specific variations in ovarian reserves in healthy fertile and infertile women: A cross sectional study.

Use of ART techniques have been on a rise with increase in infertility [ 1 ]. Over recent years, testing of ovarian reserve to predict future reproductive life has become crucial while women postpone childbearing globally [ 2 ]. Testing for ovarian reserve is fundamental in planning the course of infertility treatment and offering ART to those with borderline reserves besides predicting outcomes of treatment [ 3 ]. There are varieties of ovarian reserve tests that include ultrasound and biochemical parameters [ 3 ]. Antral follicle count (AFC) and anti-Mullerian hormone (AMH) have been shown to be the best markers of ovarian reserve [ 4 , 5 ]. Anti-Mullerian hormone (AMH) is a dimeric glycoprotein which is a member of the transforming growth factor beta (TGF-b) superfamily, produced by granulosa cells of primordial follicles that have undergone initial recruitment, and is thought to reflect the size and quality of the ovarian reserve [ 6 ]. In addition, AMH inhibits follicular sensitivity to follicle stimulating hormone (FSH) and plays a role in the process of follicle selection and dominance. Currently, AMH is considered as an optimal biomarker that reflects ovarian reserve, convenient in detection and more effective than other biomarkers, including FSH and estradiol (E2). Measuring AMH during an initial fertility workup is crucial to plan ART cycles, including protocols, dosing of gonadotropins and also predicting the response to ovarian stimulation. These results can be used to determine the recombinant human FSH starting dose and predict the final oocyte yields and develop a nomogram that could predict oocyte yield. Lee et al. investigated the cutoff level of serum AMH for predicting poor (number of oocytes retrieved, ≤3), normal (4–19), and high responders (≥20). Especially for predicting poor responders, the cutoff level was 1.08 ng/mL, with 85.8% sensitivity and 78.6% sensitivity [ 7 ]. In a multicentric trial by Nelson et al, Anti-mullerian hormone proved to be a stronger predictor of ovarian response to gonadotropin therapy compared to AFC utilizing GnRH agonist and GnRH antagonist protocols. Antral follicle count provided no added predictive value beyond AMH [ 8 ]. AMH has also found usefulness in even predicting rate of euploid blastocyst besides live birth, after In-vitro fertilization/ intracytoplasmic cycles [ 9 ]. Further beyond infertility it is utilized by clinicians to assess the ovarian reserve and predict menopause age [ 10 ]. Therefore, the reference range of AMH in normal healthy women is needed to provide useful information on ovarian function that can be of potential clinical benefit. Anti-Mullerian hormone is measured using enzyme linked immunosorbent assays (ELISAs). Various assays are available: Gen II (Beckman Coulter); pico AMH (Ansh Labs); and Elecsys (Roche) [ 11 ]. Previous studies have mostly measured AMH levels using ELISAs, the Diagnostics Systems Laboratory (DSL 10–14400) assay or the Immunotech (IOT, A11893 IVD, EU) assay [ 12 ]. These assays utilize different primary antibodies against AMH and different standards, and consequently, the crude values reported differed substantially. With the consolidation of these 2 companies by Beckman Coulter in 2011, as sole ownership of the patent to measure mammalian AMH, there is finally a single commercially available assay: the AMH Gen II assay (A79765), which has fully replaced the DSL and IOT assays [ 13 ]. Therefore, using the Beckman Coulter Gen II assay to estimate the reference range for AMH is more practical. However, only a few studies established the reference range for AMH using the new Gen II assay, especially for Indian women [ 14 ]. Antral follicle count (AFC) also serves as an objective ovarian reserve test helping decide treatment protocols and anticipating response. AFC is the most commonly used ultrasound marker to identify ovarian reserve and is reliable and easy to measure [ 15 ]. It consists of counting of all follicles in the range of 2–10 mm visualized through high resolution transvaginal ultrasound examination. Assessment of AFC has been a simple method of predicting the occurrence of menopause and thus the duration of the reproductive lifespan [ 16 ]. It is well established that the female reproductive function deteriorates with age because of reduction of the ovarian follicle pool [ 17 ]. Several studies of autopsy and surgical specimens showed that the number of antral follicles is related to the number of primordial follicles within the ovaries [ 18 ]. In particular, the number of small antral follicles decreases with age similarly to the number of primordial follicles [ 17 , 19 ]. Indeed, several studies have demonstrated that AFC declines with chronologic age in women, although it is still not clear whether this decline has a biphasic [ 20 ] or a linear pattern [ 21 , 22 ]. Ovarian reserve markers including AMH and AFC are prone to variations because of race, ethnicity, use of hormones including oral contraceptives [ 23 ]. Various studies have been published on nomograms based on AMH and AFC for fertile and infertile population [ 24 , 25 ]. As the values are affected by race, ethnicity, it becomes prudent for each population to have a nomogram to be used as reference. Several AMH–age nomograms have been reported before; however, most of the studies used hospital inpatient samples, particularly samples from infertile women instead of healthy females. With improved access to ART services, more and more Indian women are seeking treatment and using American or European standards may not seem justifiable. There is also evidence that ovarian reserves among Indian women decline earlier compared to women from Europe [ 3 , 26 , 27 ]. However, this data comes from women who had undergone ART cycle and may not reflect the nomograms as would come from women other that the infertile cohort. The reference range for AMH levels obtained from a large, population-based sample of healthy women was believed to be more practical, reliable, and accurate in representing the normal population, therefore, additional studies were warranted to determine that. Currently there are no age-related centile charts for AFC and AMH from Indian women. The present study was aimed to derive nomograms for AMH and AFC among fertile and infertile women of Indian origin, indicating the trends in decline and the relationship between two ovarian reserve markers. Additionally, we evaluated the relationship of AMH and AFC with age to provide an optimal model of decline in AMH and AFC with age.

Table 1 presents the baseline values of age, BMI, AMH and AFC values in the fertile and infertile groups. *- Mean, BMI- Body Mass Index, AMH- Antimullerian Hormone, AFC- Antral Follicle count Age segregated number were comparable in the fertile and infertile groups across different sub groups, with lower number of women between ages 40 and beyond ( Table 1 and Fig 1 ). Tables 2 and 3 provides the age segregated subgroup analysis of AMH and AFC (Nomograms). The mean AMH and AFC values were higher in the fertile group compared to the infertile with the differences being significant in the age group of 20–25 and 25–30 years sub-groups. The centiles of AMH and AFC were computed at each age point using the linear specification of the CG-LMS method (Figs 2 and 3 ). Age-related nomograms for AMH and AFC respectively were generated. Decline in AMH and AFC was linear in the infertile cohort while the fertile cohort had a non-linear pattern in decline. The rate of decline OF AMH was found to be 0.192 (± 0.05) in fertile group and 0.172(± 0.04) in the infertile group. The rate of decline of AFC was found to be 0.65 (± 0.41) in fertile group and 0.56(± 0.05) in the infertile group. The estimated centiles for AMH and AFC for the fertile and infertile groups were generated (Figs 4A, 4B , 5A and 5B ). (a) Centiles of AFC computed at each age point (Fertile population). (b) Centiles of AFC computed at each age point (Infertile population). (a) The estimated centiles for AMH for the fertile and infertile groups. (b) The estimated centiles for AFC for the fertile and infertile groups. CG-LMS method using more complex models did not result in a significant improvement in the data-fit. Age cut off for decline in AMH and AFC in fertile women came out to be approximately 31 years using ROC analysis. Age cut off for decline in AMH and AFC in infertile women was found approximately 34 years. (Figs 6 – 9 ). The correlation between AMH and AFC was estimated for both infertile and fertile populations and it was found that both the parameters were significantly correlated ( Fig 10 )

The study provides nomograms to be used as reference levels for Indian females while assessing their ovarian reserve markers and reduce time to pregnancy by formulating an individualized plan. It gives insight on the age at which North Indian women suggest declining fertility and may be useful to advise on options available while women plan to defer child-bearing.

In a cross-sectional study, conducted at a tertiary referral hospital in North India, both infertile and healthy fertile women who had borne one child conceived spontaneously were invited for testing ovarian reserves at one time. The study was approved by the ethics committee of the Institute (IEC-491/07.10.2016) and all participants gave informed written consents to be part of the study. The study was conducted from January 2017 until December 2022, with recruitment affected by the COVID pandemic, enrolling 3240 women; 1902 infertile and 1338 healthy women as controls. The participating women including infertile cohort attended fertility and ART clinics while fertile cohort consisted of those attending the family planning and contraception clinics or healthy oocyte donors with confirmed fertility. The inclusion criteria for infertile women were age 21–40 years, with infertility of more than one year, while healthy age and BMI matched women were selected as controls. Women with conditions that could affect ovarian reserves including endometriosis, PCOS, previous ovarian surgery, immunological disorders and those on gonadotoxic therapy were excluded from both cohorts. Also, women with co-morbidities such as diabetes mellitus, morbid obesity were excluded. Informed consent was obtained from all women in both the cohorts, informing them that these tests will be confidential and not relevant on future child bearing especially in the fertile cohort. Testing for both biomarkers of ovarian reserve was done early in the follicular phase between day 2 to day 6 of menstrual cycle ensuring no intake of oral contraceptive pills prior to testing in either group. Blood samples for AMH were centrifuged after collection were aliquoted and stored at -80 degrees for future use within 4 weeks of blood collection. Serum Anti Mullerian Hormone was measured in all samples using the Beckman Coulter (M/S Immunotech) AMH Gen II ELISA kits. This kit uses the two-site ‘sandwich’ assay and has an analytical sensitivity of 0.08ng/ml. This method has got a validated correlation with the automated access AMH assay [ 28 ] The limit of detection for Gen II was 0.18 nmg/ml and the measurement range varied between 0.16 to 22.6 ngm/ml. Antral follicle counts were done using high frequency transvaginal transducer (8.5 MHz; S-6, GE 3600 Healthcare, USA)when follicles between 2–8 mm were measures in three planes in each ovary, number totaled for both ovaries All scans were done the same personnel (NM, PG) who were trained sufficiently to ensure consistency. All data recorded on the proforma was entered in EXCEL and further statistical analysis was conducted using R software environment. Baseline characteristics such as age, BMI, AFC, AMH were measure for both the fertile as well as infertile cohorts, which were expressed as mean and standard deviations, percentages whichever was most appropriate. Both cohorts were divided into six sub groups based on age. Age specific centile charts for AMH and AFC were generated as per the CG-LMS method using GAMLSS (Generalized Additive Models for Location Scale and Shape) package available in R Statistical Computing Language. This method produces a model that specifies the centiles in terms of age-specific curves called L (skewness), M (median) and S (coefficient of variation). At each age point, M curve corresponds to the median AFC/AMH level, S curve corresponds to the coefficient of variation and L curve allows for age dependent skewness of distribution of AMH/AFC. The value of any centile at a given age can be computed from the values of L,M and S. Six empirical centiles, including the 3rd, 10th, 25th, 50th, 75 th and 90th centiles, and nomogram tables were constructed. Receiver Operator Characteristic(ROC) curve analysis was conducted to analyze the predictive accuracy of age with respect of AFC and AMH and to determine the cut-off values using the Area Under the Curve (AUC) analysis. The cut off values for AMH of <1.2 ngm/ml and AFC of < 5 was taken as per the POSEDION criteria in describing diminished ovarian reserves while generating the ROC [ 29 ].

(XLSX)