{"paper_id":"5ade49db-7955-4ca4-b149-69f08d5d12c4","body_text":"Primary infertility affects 8-12% ofreproductive-age couples globally, and the proportion\nis estimated to vary from 4.5 to 30% across countries, with the highest percentages found in\ndeveloping countries ( 1 ). In Indonesia, it was estimated that 12.3% of reproductive-age\ncouples suffered from infertility, whereas another survey estimated a prevalence rate of\n10-15% ( 2 ). As such, the demand for assisted reproductive technology, such as in\n vitro  fertilization (IVF) has risen in recent years.\nAdequate follicle growth, achievable through follicle\nstimulating hormone (FSH) administration in controlled\novarian hyperstimulation protocols, is crucial for the success of an IVF cycle ( 3 ,  4 ). However, the varieties inindividual characteristics and response to FSH stimulation\namong infertile patients havemade it difficult to generate a\ndose cut-off applicable for both low-responders and highresponders while avoiding the risk of ovarian hyperstimulation syndrome. FSH in an IVF cycle is therefore generally administered based on a fixed dose, and the standard\ndose used in our clinic was 225 IU. The two best known\novarian reserve markers to predict ovarian response to\nFSH are mean antral follicle count (AFC) and anti-Mullerian hormone (AMH), although there is a lack of data to\nconclude which of the two markers served better to predict ovarian reserve ( 5 ,  6 ). AFC is the number of follicles\nmeasuring 2-10 mm in size from both ovaries. AMH is\ndetected in the primordial follicle and achieves peak level\nin the small antral follicle. The AMH level indicates the\nnumber of growing follicles, and this level can be used to\ndetermine the prognosis of fertility. The number of oocytes obtained will be probably low if the predicted AMH\nlevel is low, whereas extreme ovarian response complications can be expected when the predicted AMH level is\nexcessive. Currently available ovarian reserve markers, including the AMH and AFC, invariably still show varied results, especially in clinical practice where AMH and\nAFC level could be at odds with each other ( 5 - 7 ).\nIt has been previously suggested that ethnicity may influence ovarian reserve markers ( 8 - 10 ). Furthermore, ethnicity may influence the manner with which ovarian reserve\nmarkers interact with factors such as age and weight. Age\nis negatively correlated with AMH and AFC in Caucasian,\nAfrican-American, Hispanic, and Asian women, however\nBMI was only negatively correlated with serum AMH\nlevel in Caucasian women ( 8 ). A cross-sectional study\ncomparing Indian and Spanish women showed that AFC\nis declined in younger Indian women compared to Spanish women ( 11 ). To date, there are few studies that have\nexamined ovarian reserve markers in Indonesian women,\nmuch less inthose who received controlled ovarian hyperstimulation in an IVF program. Therefore, it is necessary\nto study the association of AMH and AFC to obtain an optimal ovarian response in this population. This study was\nconducted to determine the correlation of AMH and AFC\nwith the number of oocytes in women who hadreceived\ncontrolled ovarian hyperstimulation in an IVF program\nwith an FSH dose of 225 IU, which is the most frequently\nused dose in Indonesian health facilities.\n\nIn this retrospectively study, the data were obtained from the medical records of patients\nwho underwent the IVF program with the FSH dose of 225 IU at Aster Clinic, Hasan Sadikin\nHospital, and Bandung Fertility Center Limijati Hospital, Indonesia. The sample size in this\nstudy was calculated by a sampling formula for unpaired analytic categorical study, set at\nα=0.5 and 1-β=90%. Proportion of the population (P 1  and P 2 ) were\nassumed to be 50 and 10%, respectively. The formula yielded a minimum of 26 samples. The\ninclusion criteria were patients who underwent the IVF program, aged ≤40 years, were given a\nconstant exogenous FSH dose throughout the cycle, and whose medical record included complete\npatient characteristics, physical examination, AMH and AFC levels throughout the cycle. AMH\nand AFC measurements were done on the second or third day of the menstrual cycle and this\nwas done consistently. The exclusion criteria were the presence of a history of ovarian\nsurgery, polycystic ovary syndrome, endometriosis, ovarian cyst, orendocrine disease. We\nalso recorded the patients’ identity, characteristics, previous medical history, previous\nmedical therapy, levels of ovarian reserve markers (AFC and AMH), and the number of oocytes\nproduced (the oocyte numbers in this study represent numbers for all oocytes aspirated). In\nthis study, we selected patients as a whole, which means that all patients underwent the\nsame treatment regimen using a short protocol with recombinant FSH and human chorionic\ngonadotropin, and all sperm used had normal parameters.\nThis study protocol was approved by Faculty of Medicine, Universitas Padjadjaran, Ethics Committee Review\nBoard (LB.04.01/ACS/TC/066/III/2018) and all study\nparticipants gave informed consent, patients consent to\nparticipate was written. All authors hereby declare that\nall patients have been examined in accordance with the\nethical standards laid down in the 1964 Declaration of\nHelsinki.\nNumerical data arepresented as mean, SD, median and\nrange. Data normality was assessed using the ShapiroWilk or Kolmogorov-Smirnov test. The subjects’ characteristics were compared using an unpaired t test or a\nMann-Whitney test, as appropriate. Correlation between\nthe variables was assessed using Pearson or Spearman\ncorrelation test. A P≤0.05 was considered statistically significant. SPSS v24.0 (IBM Corporation, USA) was used\nto perform statistical analysis.\n\nOf 356 patients enrolled during 2005-2017, 42 patients met the study criteria. Mean age\nof the patients investigated in this study was 34.8 ± 2.8 years (range: 29-39 years, median:\n35 years) and the body mass index (BMI) was 24.4 ± 4.1 kg/m 2  (range: 14.3-33.3\nkg/m 2 , median: 23.6 kg/m 2 ). Baseline demographics of the study\nsubjects are shown in Table 1. Normality test results showed that AFC and AMH-AFC levels\nwere normally distributed, which were then analysed using Pearson correlation test, whereas\nAMH levels and oocyte amount were not normally distributed and analysed by Spearman\ncorrelation test.\nBaseline demographics and study subject characteristics\nBMI; Body mass index, AMH; Anti-mullerian hormone, and AFC; Antral follicle\ncount\nThe largest number of oocytes ( 4 - 15 ) was produced at\na range of AMH of 1.2-4 ng/ml and AFC 4-15 ( Table 2 ).\nThere were 37 patients who produced 4-15 oocytes and\nthey were classified as normo-responders. One patient\nwas a hyper-responder due to the production of >15 oocytes; this excessive response was predicted as she had\nan AFC of >15. Furthermore, four patients produced <4\noocytes and they were classified as hypo-responders.\nSerum AMH and AFC according to ovarian stimulation response groups\nAMH; Anti-mullerian hormone and AFC; Antral follicle count.\nAMH levels were analysed using Spearman correlation test, AFC and AMH-AFC were analysed by Pearson\ncorrelation test. A significant positive correlation was\nfound between AMH (r=0.530, P≤0.001), AFC (r=0.687,\nP≤0.001), and AMH-AFC combination (r=0.652,\nP≤0.001) and the number of oocytes. To reduce bias from\npossible confounding by age, Spearman’s correlation\nwas used to determine the correlation between age and\nnumber of oocytes. There was an insignificant negative\ncorrelation between age and number of oocytes produced\n(P=0.129 and r=-0.179). Pearson’s correlation test was\nused to determine the correlation between BMI and AMH,\nand between BMI and the number of oocytes retrieved;\nan insignificant positive correlation was found between\nthe two variables (P=0.216, r=0.123, and P=0.452, and\nr=0.19, respectively).\n\nOur study showed that AMH, AFC, and AMH-AFC\nare significantly correlated with the number of oocytes\nproduced. This is in agreement with previous studies\nthat investigated the relationship between AMH levels\nand the number of oocytes. Asada et a.l ( 12 ) observed\na positive correlation between AMH level and oocyte\ncount among Japanese women. AMH can effectively\npredict ovarian responses and allow clinicians to avoid\niatrogenic complications and choose optimal stimulation\nstrategies ( 13 ). However, AMH levels showed variations\nwhen examined by different examination kits and among\ndifferent populations. Although we found a relatively\nstrong and significant correlation between AMH levels\nand the number of oocytes with the FSH dose of 225 IU\nin this study, these potentially confounding factors should\nbe considered.\nFertility begins to decrease at the age of 30 years and\nfurther decreases significantly after the age of 35 years\n( 14 ,  15 ), which made our subjects’ age (34.8 ± 2.8)\na significant potential confounder inour study. In the\ncorrelational analysis, however, we did not observe a\nsignificant interaction between age and the number of\noocytes produced.\nHow BMI influences ovarian reserve markers and the\nnumber of oocytes retrieved, is still unclear. In ametaanalysis, Moslehi et al. ( 16 ) concluded that AMH is\nsignificantly lower in obese women. On the other hand,\na study of 402 women in Turkey, categorized based\non ovarian reserve patterns (poor, <7 baseline AFC;\nadequate, ≥7 baseline AFC, and high ovarian reserve) and\nBMI group, revealed that serum AMH and FSH levels\nwere similar across all categories ( 17 ). Another study\nof women receiving controlled ovarian stimulation for\nassisted reproductive technology reported that BMI did\nnot negatively affect the number of oocytes retrieved\n( 18 ). In our study, we did not find a significant correlation\nbetween BMI and AMH or BMI and oocyte count.\nWe observed a relatively strong positive correlation\nbetween AMH-AFC combination and the number of\noocytes. In a sequential order, it can be observed that AFC\ncorrelates best with the number of oocytes, followed by\nAMH–AFC combination and AMH. The results of this\nstudy are in contrast to the result of a study conducted by\nNelson et al. ( 19 ), which compared the predictive value\nof live births that indirectly represents the association\nbetween the number of oocytes and AMH, AFC, and\nAMH-AFC combination only with age, in the UK. The\nauthors reported that AMH showed the best predictive\nvalue, followed by the combination of AMH-AFC and\nAFC only. In the present study, the strongest relationship\nwas observed between the number of oocytes and AFC,\nand the results were not much different from those of the\ncombination of AMH-AFC. This suggests that, in women\nwithout discordant ovarian marker, AFC may be a better\nchoice compared to AMH in predicting ovarian reserve.\nThis agrees with the results of Jayaprakasan et al. ( 20 ),\nwho found that AFC predicts ovarian response better than\nAMH or a combination of AFC and AMH. This is further\nreinforced by the results of Liao et al. ( 21 ) whose study\non 8269 women undergoing IVF/intracytoplasmic sperm\ninjection (ICSI) treatment showed a strong association\nbetween AFC, number of oocytes retrieved, and clinical\npregnancy rate. These data imply that in the absence\nof AMH examination, AFC may suffice, as it is wellcorrelated with the number of oocytes in clinical practice.\nAFC is easier, and relatively inexpensive, and offers\nalmost immediate results.\nIn the present study, we found a relatively strong positive\ncorrelation between AFC and the number of oocytes at the\nFSH dose of 225 IU. The AFC measurement performed using ultrasound was effective, easy to use, safe, and\nnon-invasive. Therefore, estimating the number of\nantral follicles can be used as a predictive test of ovarian\nfunction, ovarian reserve, and ovarian response.\n\nSignificant positive correlations of AMH levels, AFC,\nand AMH-AFC with the number of oocytes were found\nin this study. These correlations werestrong enough at\nthe FSH dose of 225 IU. AFC is a better ovarian reserve\nmarker compared to AMH and the combination of AMHAFC in predicting the number of oocytes. Existing data\non variations in infertility causes and longevity suggest\nthat an analysis free of infertility including confounding\nvariables and duration, would be preferable. Our study\nlimitations could be overcome by multivariable analysis,\nbut a larger sample size is needed.","source_license":"CC-BY-4.0","license_restricted":false}