Lower antimüllerian hormone is associated with lower oocyte yield but not live-birth rate among women with obesity.

As AMH is used to counsel infertile women regarding treatment options, it is important to consider how a common co-morbidity such as obesity influences the interpretation of this ovarian reserve marker. In this large cohort of infertile women with undergoing IVF, we found that lower AMH is associated with lower oocyte yield among women with obesity defined as BMI ≥ 30 kg/m 2 , but not clinical pregnancy or live birth rate. While obesity has been associated with lower oocyte quality ( 14 ), the present study provides evidence of lower oocyte yield among women with obesity and infertility. Ovarian reserve is measured by day 2–3 FSH and estradiol, antral follicle counts and AMH ( 1 ). We observed lower day 2–3 FSH among obese women, which was more pronounced with increasing obesity. This suggests a direct inhibitory effect of body mass on gonadotropin production and may alter the ability to detect rising FSH levels as evidence of diminished ovarian reserve. The effect of obesity on antral follicle counts has been questioned given that antral follicle counts were no different among obese and normal weight late reproductive age women, despite lower AMH levels observed in women with obesity ( 11 ). Abdominal adiposity can obscure visualization of 2–10 mm antral follicles by transvaginal ultrasound limiting the accuracy of antral follicle counts among obese women. By correlating AMH with an objective marker of ovarian reserve, such as response to ovarian hyperstimulation, we provide evidence that lower AMH is in fact associated with lower ovarian reserve in women with obesity. Notably, the difference in oocyte yield by BMI category was significant among women with an AMH >1.1 ng/ml. However, this effect was not observed among women with low or undetectable AMH levels, suggesting a negligible impact of body mass in women with severely depleted ovarian reserve. We can speculate that the physical challenges of transvaginal oocyte retrieval in women with obesity contributed to the lower oocyte yield among women with higher AMH. It is possible that physician are less likely to retrieve all follicle when a reasonable oocyte yield is obtained in a woman with obesity given that obesity obscures ovarian access and visualization of follicles. It is possible that physicians will make more attempts to retrieval all follicles when the yield is low in a women with obesity despite these physical challenges. We are unable to compare to the number of mature follicles at the final transvaginal monitoring ultrasound and the number of oocytes retrieved in order to provide data to support this speculation. The value of AMH to predict live-birth remains controversial. In this large cohort which excluded women with PCOS, we observed no association with AMH and live-birth rate among obese women. The strength of the association of AMH with live-birth rate appears to be most evident at the extremes, where ultralow levels associated with diminished ovarian reserve and very high levels associated with PCOS are predictive of lower live-birth rate ( 3 , 19 ). The main strength of this study is that the large data set included a large sample of women in each BMI class after excluding women with PCOS. In addition, the dataset enabled analysis of covariates, such as age and smoking, that influence AMH and live birth rates in IVF cycles. A limitation of this study is in the variability in AMH measurements. AMH can vary related to exposures that could not be account for in this dataset, such as recent oral contraceptive use or ovarian surgery. Variations in AMH levels are also reported between laboratories, as different manufacturers and versions of the AMH assay were in use in the US during this time frame ( 20 ). In addition, it is likely that the AMH levels were drawn at variable time points prior to the IVF cycle start. Similarly, the BMI may have been measured at variable time points prior to the IVF cycle start. To minimize the extent of these variations and to avoid counting multiple cycles in the same patient, we limited our analysis to the first autologous IVF cycle. Finally, we are unable to analyze antral follicle count as it is not recorded in SART CORS, nor are we able to comment on cumulative live birth rates from the fresh IVF and frozen embryo transfers that may have resulted. In conclusion, in a large cohort of women with obesity undergoing IVF, AMH was associated with oocyte yield among women with obesity but not clinical pregnancy rate or live-birth rate. Correlating AMH with response to ovarian hyperstimulation provides evidence that lower AMH is in fact associated with lower oocyte yield in women with obesity and is not lower due to a physiologic change associated with obesity. Further research is needed to determine the mechanism by which obesity lowers ovarian reserve and to determine whether weight loss can improve ovarian reserve among obese women.

Data from 16,579 women with obesity, defined as BMI ≥30 kg/m 2 , and 13,316 normal weight women, defined as BMI 18.5–24.9 kg/m 2 were included. There were 8,121 women with class I obesity (, 3,530 women with Class II obesity and 1,665 women with Class III obesity. Women with obesity were older (36.0 ± 4.8 vs. 35.5 ± 4.8. P<0.001) and more likely to smoke (7.0% vs. 4.5% P<0.001) compared to women with a normal weight ( Table 1 ). There was no difference in the frequency of women who were attempted IVF due to diminished ovarian reserve, as defined in SART CORS, among women who were normal weight and women who were obese (P=0.179), nor was the diagnosis more common with increasing BMI (P=0.174). Women with obesity had lower maximum follicular FSH levels (7.6 ± 3.7 vs. 8.3 ± 4.5, P<0.001) compared to women with normal weight. Further, women with class III obesity had lower day 2–3 maximum FSH levels than women with class I and II obesity. Women with obesity had lower AMH levels (1.8 ± 2.0 vs. 2.1 ± 2.0, P<0.001) than women with normal BMIs and there was a statistically significant progressive decline in AMH with increasing BMI (P=0.002). Women with obesity, defined as BMI ≥30 kg/m 2 , were more likely to undergo ovarian stimulation using an antagonist protocol (P<0.001) and agonist flare protocols (P=0.02) and were less likely to receive agonist suppression (P<0.001) compared to normal weight women ( Table 2 ). Women with obesity were more likely to receive higher cumulative FSH dose for stimulation (P<0.001), undergo longer stimulation (P<0.001), experience a higher rate of cancellation (P<0.001) for low response and have fewer oocytes retrieved (11.9 ± 7.3 vs. 12.8 ± 7.7, P<0.001) compared to women with normal BMIs. Women with obesity had a similar frequency of undergoing an embryo of transfer (P=0.424) and had a similar number of embryos transferred (2.1 ± 0.8 vs. 2.1 ± 0.8, P=0.401), yet women with obesity had fewer embryos cryopreserved (2.1± 3.6 vs. 2.5±4.1, P 1.1 ng/ml (P<0.001) ( Table 3 ). There was no difference in the number of oocytes retrieved by BMI category in women with undetectable AMH levels (P=0.108) and low AMH levels (P=0.574). Women with obesity, defined as BMI ≥30 kg/m 2 , had lower clinical pregnancy rates (42.7% vs. 49.5%, P<0.001), higher spontaneous abortion rates (8.0% vs 7.2%, P=0.042), and lower live birth rates (34.3% vs 41.8%, P<0.001) compared to normal weight women ( Table 2 ). After adjusting for covariates, AMH was associated with number of oocytes retrieved (β=0.069, SE=0.005, P<0.001), in women with a BMI ≥ 30 kg/m 2 implying that lower AMH was associated with decreased oocyte yield in this population ( Table 4 ). AMH was not associated with clinical pregnancy rate (OR 0.98 (95% CI 0.93–1.04), P=0.54) or live birth rate (OR 0.98, 95% CI 0.93–1.04, P=0.57) among women with a BMI ≥ 30 kg/m 2 .

The data for this study were obtained from the SART CORS online database, which captures more than 90% of all ART cycles in the United States. Data were collected and verified by SART and reported to the Centers for Disease Control and Prevention, in compliance with the Fertility Clinic Success Rate and Certification Act of 1992 (Public Law 102– 493). Since 2012, serum AMH levels have been recorded in SART CORS. This study was deemed exempt from Institutional Review Board (IRB) approval at by the University of Rochester Medical Center IRB prior to the obtaining and analyzing the data. Cycles reported to the SART CORS from January 1, 2012 through December 31, 2014 involving women undergoing their first autologous IVF with a fresh transfer and recorded pretreatment AMH and BMI levels, were included in this study. Cycles that used gestational carriers, pre-implantation genetic diagnosis or screening, or were designated as minimal stimulation cycles, unstimulated cycles, research cycles, or frozen embryo transfers were not included. A cycle classified as a research cycle was excluded because the research protocol would not be available to the investigators performing a secondary analysis. The SART CORS category which captures PCOS is titled “ovulation disorders/polycystic ovaries,” includes women with PCOS as well as women with multiple ovarian cysts affecting fertility; oligo-ovulation (<6 cycles per year); and anovulation (of hypothalamic or nonhypothalamic causes). Women with ovulatory disorders and thus PCOS were excluded when evaluating AMH. The SART CORS category which captures the reason for assisted reproductive technology includes the category diminished ovarian reserve which includes women with premature ovarian failure, Turner Syndrome and other diminished ovarian reserve which is defined as high FSH, high estradiol measured in the early follicular phase or during a clomiphene challenge test. The independent variable for this analysis is serum AMH, categorized as normal >1.1 ng/ml, low 0.16–1.1 ng/ml, and undetectable <0.16 ng/ml. ( 3 ) BMI was treated as a categorical variable based on the World Health Organization (WHO) BMI guidelines with normal weight defined as BMI of 18.5–24.9 kg/m 2 and obesity was defined as BMI ≥30 kg/m 2 ( 13 ). Obese women were further stratified as class I obesity (BMI 30.0–34.9 kg/m 2 ), class II obesity (BMI 35.0–39.9 kg/m 2 ) and class III obesity (BMI ≥40 kg/m 2 ). Covariates included age, ethnicity, smoking, gravidity and parity, reasons for infertility, day 2–3 FSH, IVF protocol and number of embryos transferred. The primary outcome was oocyte yield. The secondary outcomes were clinical pregnancy rates per transfer (presence of a gestational sac by transvaginal ultrasound/number of cycles with a transfer), loss/abortion rates (nonviable clinical pregnancy/number of cycles with a transfer) and live birth rates per transfer (occurrence of a live birth/number of cycles with a transfer). Baseline characteristics were compared using t-tests or Chi-square tests, as appropriate, with continuous variables reported as mean ± SD and categorical variables reported as percentages. Multivariable modeling was used to assess oocyte yield using linear regression with a logarithmic transformation and odds of clinical pregnancy and live birth using logistic regression. Age, BMI and number of embryos transferred were kept in the model regardless of significance. A P value <.05 was considered to be statistically significant. Statistical analysis was performed using SAS V9.3 (SAS Institute, Cary, NC).

Anti-Mullerian hormone (AMH) is produced by the granulosa cells of small antral follicles and reflects ovarian reserve. AMH declines with reproductive aging and is used clinically to predict response to ovarian stimulation and possibly live-birth rate in women undergoing in vitro fertilization (IVF) ( 1 , 2 ). A serum AMH level <0.8 ng/mL has a sensitivity of 40–97% and specificity of 78–92% for poor response to ovarian stimulation, defined as <4 oocytes retrieved. An ultralow serum AMH level 1–1.2 ng/mL) ( 3 ). Several factors in addition to age influence the interpretation of AMH levels. AMH is increased in women with polycystic ovary syndrome (PCOS) likely due to arrested development of antral follicles, as well as increased AMH production per PCOS granulosa cell ( 4 , 5 ). AMH is decreased by oral contraceptive use possibly due to altered follicular recruitment. AMH is also decreased by tobacco use and chemotherapy due to follicular atresia ( 6 ). Obesity is associated with lower AMH though the etiology and implications are unclear ( 7 – 9 ). In a cross-sectional study of serum AMH levels in a population based cohort of healthy late reproductive age women, AMH levels were 65% lower in women with obesity compared to women with normal BMIs and BMI remained significantly associated with AMH levels after adjusting for age and race ( 10 ). Despite evidence of lower levels of AMH among obese women in this cohort, antral follicle count did not differ between normal weight and obese women ( 11 ). In addition, women with obesity do not appear to enter menopause at an earlier age ( 12 ). These observations raise the question of whether lower AMH in women with obesity is reflective of diminished ovarian reserve or is lower due to physiologic changes associated with obesity. Counseling women with obesity and low AMH is challenging given that the predictive value of AMH for oocyte yield and live birth rates has not been validated in a large cohort of women with obesity. Our objective in this study is to determine whether lower AMH among women with obesity undergoing IVF is associated with oocyte yield and live birth rate based on national data from the Society for Assisted Reproductive Technology Clinic Outcome Reporting System (SART CORS).