Female reproductive function gradually changes throughout puberty, the growth period, the perimenopausal period and menopause, so the detection of ovarian function reserve is very important for clinicians to guide patients to choose fertility treatments and predict their clinical prognosis. Traditional ovarian function assessment methods include the determination of basic hormone levels (FSH, LH, E2), inhibin B (InhB), and antral follicles (AFC)[12]. Due to the complex interactions in the internal environment, FSH, InhB and AFC may not be able to predict the ovarian reserve function in women who do not have changes in ovulation. However, even for these women, AMH can still indicate changes in ovarian function, which is detectable earlier and is more accurate than traditional methods[13].
AMH is secreted by granulosa cells of preantral follicles and small antral follicles, inhibits the recruitment of primordial follicles, regulates the growth and development of follicles[14] and is not regulated by the hypothalamus-pituitary-ovarian axis. It stimulates the continuous growth of antral follicles, and it can be detected on any day during the menstrual cycle. AMH is produced by follicles entering the cycle and also by other follicles not entering the cycle, which can predict ovarian responsiveness and is not affected by FSH[15–16].
In a study of age in China, it was found that fertility tends to be static from the establishment of menarche to the age of 30 and it begins to decline after the age of 30. After the age of 50, the weight and volume of the ovaries decrease significantly, and ovarian function shows an exponential decay[17]. AMH could be reduced to 1.43 µg/L at age 37 and start to decline at 30 years old, especially after 35 years old. The level of AMH decreases linearly, which is positively correlated with AFC and the ovum capture rate[18].
During the process of assisted reproductive technology, the level of ovarian reserve function can affect the outcomes of ovulation and pregnancy. If a deficiency of ovarian reserve function can be detected before the implementation of assisted reproductive technology, we could deduce the results of drugs from the level of AMH and explore how to improve the clinical outcome. The sensitivity and specificity of AMH in evaluating ovarian hyporesponsiveness were reported to be 80% and 85%, respectively[19].
Through this study, we can see that during the process of assisted reproductive technology, compared with low AMH (AMH < 1.1), the women with higher AMH (AMH > 1.1) used less Gn (bottle) (35 (25 ~ 44) vs. 42 (36 ~ 50.8)), had a lower starting Gn (IU) (225(150 ~ 300) vs. 300(225 ~ 300)), and during the oocyte acquisition and embryo culture, the numbers of oocytes obtained (10 (8 ~ 13) vs. 6 (4 ~ 7)), the numbers of MII (9 (7 ~ 12) vs. 5 (3 ~ 6.5)), the numbers of cleavages (8 (5 ~ 11) vs. 5 (3 ~ 6)), the numbers of 2PN(6 (4 ~ 9) vs. 4 (2 ~ 5), the numbers of high-quality embryos(4 (2 ~ 6) vs. 3 (2 ~ 4)) were all better (P < 0.001). This is similar to most other studies that found that serum AMH can better predict ovarian responsiveness and the numbers of oocytes retrieved during controlled ovarian hyperstimulation[20].
Compared with basic FSH, oestrogen and inhibin B and other indicators to predict ovarian reserve function in patients, the stability of AMH, not affected by the detection time and with other advantages, makes it very useful in the field of assisted reproduction to evaluate ovarian reserve function and predict ovarian responsiveness and it has received a great deal of attention[21]. Is there any correlation between AMH and the clinical outcomes of assisted reproductive technologies? Interestingly, in our study, through ROC curves, we found that AMH may not have a good predictive value for pregnancy outcomes of IVF (P = 0.980, 0.093, 0.447, 0.146, 0.526, > 0.05). Although the AUC for miscarriage was equal to 0.642, indicating that the accuracy of AMH for predicting miscarriage as the IVF outcome could reach 64.2%, it was not statistically significant (P = 0.093, < 0.05).
Some researchers found that AMH could not only promote the development of oocytes but also increase the development potential of fertilized eggs[22]. However, there are researchers who found the opposite, that there is no correlation between basic AMH levels and pregnancy outcomes[23]. In this study, we included 518 patients who were divided into a pregnancy group (n = 221) and a nonpregnancy group (n = 297). We found a significant difference in the starting gonadotropin dose (start Gn) (225 (150 ~ 300) vs. 225 (187.5 ~ 300), P = 0.038 < 0.05). Other variables, including total gonadotropin dose, total days of use of gonadotropin, numbers of oocytes retrieved, numbers of MII, numbers of cleavages, numbers of 2PN and numbers of good-quality embryos, showed no differences between the two groups. We deduced that pregnancy outcome might be related to other variables that may improve the pregnancy outcome and increase the pregnancy rate in IVF patients, such as BMI, endometrium, and hysteroscopy status, which can be collected for analysis in future studies while increasing the sample size to reduce selection bias in data analysis.