Anti-Müllerian hormone (AMH): a marker for ovarian function
In the evaluation of an infertile patient, ovarian function is one of the most important parameters. The evaluation process is a complex one that involves many different tests and has evolved significantly over the past 40 years.
Forty years ago there were only two parameters for evaluating ovulatory function. These were the basal body temperature chart (BBT) and the patient’s age. We have known for many years that a patient’s age is very strongly correlated with her fertility. One of the best studies was performed 35 years ago in France on women who either had no husband or whose husbands had no sperm and therefore were undergoing donor artificial insemination to achieve pregnancy. The study was therefore looking at the fertility rate in women who had not previously had an opportunity to become pregnant because of the lack of sperm. The following table presents that data:
Patient Age Pregnancy Rate Miscarriage Rate
<30 95% <10%
35 85% 15%
38 65% 20%
40 45% 30%
42 20% 45%
44 10% 65%
>45 <5% >75%
The difficulty with this data is that we did not have the ability to estimate the chance for pregnancy of any individual patient in a specific age category. Which 10% of women in the 44-year-old category are going to achieve pregnancy.
In the past there has been controversy about whether the decrease in fertility with age was related to the age of the woman’s body or the woman’s eggs. Various studies have concluded that it is the age of the eggs that is the critical factor. This can be best illustrated in the following graph. The data was obtained from the 2010 ART Success Rates published by the CDC and generated from national data from hundreds of IVF clinics and thousands of IVF cycles from around the country.
The green line shows IVF success rates using the infertile woman’s own eggs
The blue line shows success rates using donor eggs
In the late 70s and early 80s, radioimmunoassay assays of hormones became available, and very quickly we realized that a patient with a markedly elevated FSH >30 was very unlikely to become pregnant. We also were able to utilize progesterone levels in the luteal phase to estimate the fertility of the ovulatory cycle. Women with luteal progesterone levels less than 15 and 8, performed on days 8 and 11 after ovulation respectively, were unlikely to become pregnant, and if pregnancy occurred, the risk of miscarriage was very high.
In the late 80s, the use of FSH measurements became even more sophisticated. It was noted that relatively subtle elevations in the FSH were associated with a decreased probability of pregnancy. This observation led to the development of the clomiphene challenge test that looked at the estradiol and FSH levels before and after the administration of 100 mg of clomiphene citrate from days five through nine of the cycle. The test is interpreted as follows:
Estradiolideal transitional abnormal FSHideal transitional abnormal
Day 3 <50 50-75 >75 <8 8-12 >12
Day 10 >250 100-250 <100 <8 8-12 >12
(or day after last clomiphene)
For more than 10 years the FSH and clomiphene challenge test remained one of our best means of evaluation of ovarian age/ovarian reserve. Although FSH levels are still of value, AMH levels seem to be replacing FSH as a more important meaningful evaluation of ovarian function.
What is AMH?
Anti‐Müllerian hormone (AMH) is a transforming growth factor‐β family member which is involved in regulation of folliculogenesis. During female fetal development, no ovarian AMH activity can be detected before 36 weeks of gestation. The expression of AMH mRNA in the ovarian granulosa cells coincides with the initiation of primary follicle growth. Studies show a high expression of AMH in the granulosa cells of early growing, pre-antral and small antral follicles, but the signal is lost in large non‐atretic antral follicles and in all atretic follicles.
Figure 1
AMH expression in the ovary. (A) AMH is expressed in granulosa cells of primary (P), preantral (PA) and small antral (SA) follicles. (B) AMH expression disappears in antral (A) and atretic (At) follicles. Expression is lost last in the cummulus granulosa cells that immediately surround the oocytes. AMH expression was detected using a monoclonal antibody that recognises human AMH. Magnification × 200.
Follicle growth and differentiation is a complex process. Two important regulation steps can be identified; the initiation of growth of follicles from the primordial pool of an active follicles (initial recruitment) and rescue of the growing follicles from atresia (cyclic recruitment).
The factors that initiate growth and early differentiation of the follicles from the primordial pool into the active pool are unknown. It does appear that AMH exhibits an inhibitory effect on the initial follicle recruitment from the primordial pool. This prevents a premature loss of follicles from the primordial pool. In an animal model, it can be shown that young animals that have a low AMH level have an accelerated depletion of the primordial follicle pool that results in premature ovarian failure.
Figure 2
Model of AMH action in the ovary. Progressing stages of folliculogenesis are depicted. AMH is produced by the small growing (primary and preantral) follicles in the ovary and has two sites of action. It inhibits initial recruitment of follicles from the primordial pool, and it inhibits FSH-dependent growth and selection of preantral and small antral follicles to become dominant follicles.
FSH is an essential factor in the cyclic recruitment of pre-antral follicles. At later stages of follicle development, growth and differentiation and the selection of the cohort of follicles are largely dependent on FSH activity.
It is thought that AMH reflects the pool of oocytes that have progressed from the primordial pool into the pool of pre-antral and small antral follicles. It is from this group of follicles that cyclic recruitment occurs and a cohort of multiple follicles begin the maturation process towards ovulation. In the natural, spontaneous ovulation cycle, a single follicle becomes dominant and completes maturation and goes on to ovulate. It is from this cohort of maturing follicles that we are able to rescue from atresia multiple follicles for the super ovulation needed with in vitro fertilization. High doses of FSH are utilized to accomplish this rescue process. When there are very few antral follicles present, the stimulation process is unsuccessful and very few follicles are matured for ovulation or egg retrieval.
Pearls about ovarian reserve and AMH levels.
1. AMH is produced in the greatest amount by pre-antral follicles measuring <6 mm. Antral follicles >8 mm in diameter and atretic follicles have virtually undetectable levels of AMH.
2. High AMH levels are often times seen in patients with polycystic ovary syndrome (PCOS). Levels >3.0 ng/ML are generally seen in PCOS patients.
3. AMH levels decline with age. It appears that a decline in AMH may be the earliest marker of ovarian aging. A study by de Vet showed that patients who had their AMH levels studied at two time points approximately 2.6 years apart, showed a significant decline in their AMH levels. This decline was present at all ages.
Figure 3
Serum AMH levels in normoovulatory women. (A) Box and whiskers plots show declining AMH levels with increasing age. AMH levels were measured at two time points with a 2.6 ± 1.7-year interval. (B) AMH levels positively correlate with numbers of antral follicles at visit 1 (closed circles, solid line) and visit 2 (open circles, dotted line). Reproduced from de Vet et al.(2002).
This data also shows that at any AMH level the average basal antral follicle count declines with age. In other words, a 30-year-old woman with an AMH of 4.0 is likely to have a significantly higher basal antral follicle count and therefore likely to produce more follicles with super ovulation than is a 40-year-old woman with an AMH of 4.0.
4. AMH levels become undetectable at least 5 years before menopause. The rise in FSH levels usually occurs later and therefore AMH appears to be a more sensitive indicator of ovarian reserve.
5. There are patients with discordance between AMH and FSH levels with reassuring AMH levels but concerning FSH levels. A recent study of women with elevated FSH levels (>10 mIU/ML) showed that those with an AMH level >0.6 ng/ML had twice the number of eggs retrieved, a greater number of embryos, and a higher clinical pregnancy rate when compared with women with an AMH level <0.6 ng/ML (Buyuk et al., Fertil Steril 95:2369-72, 2011).
6. Our current thoughts are that the number of eggs retrieved from IVF patients decreases in the following order: a) reassuring FSH and AMH, b) concerning FSH but reassuring AMH, c) reassuring FSH but concerning AMH and d) both concerning FSH and AMH.
7. Obese women have 65% lower mean AMH levels compared to age controlled nonobese women. This suggests that obesity might be associated with the managed ovarian reserve. There are no studies in women who have lost weight to see if there AMH levels rise.
8. AMH levels do not vary a lot throughout the menstrual cycle. This is in contradistinction to FSH which needs to be measured on cycle day two, three or four because it varies dramatically depending upon the phase of the menstrual cycle. AMH levels can be measured on any day of the cycle. Improve. However, there are some variations across the menstrual cycle (Sowers et al., Fertil Steril 94:1482-6, 2010). In women with values less than 1.5 ng/ML there was minimal variation throughout the menstrual cycle. In patients with higher AMH levels, there was significant variation with rising AMH levels in the follicular phase. Therefore, in patients who have normal AMH levels, there may be some variation but this is not of any practical significance as their values are always normal. In patients with abnormal AMH levels, there is minimal variation in the AMH level through the menstrual cycle. In summary, if the AMH level is <1 ng/ML, this likely indicates decreased ovarian reserve and this interpretation can be made independent of the day of collection across the menstrual cycle.
9. AMH levels are influenced by birth control pills. A recent study measured AMH levels in 42 healthy young women over nine weeks (Kallio et al., Fertil Steril, 99:1305-1310, 2013). The levels had not changed much at five weeks but were significantly suppressed at nine weeks. The average AMH levels were 3.88 ng/ML at baseline and 3.34 at five weeks but only 1.91 at nine weeks. Women who are on birth control pills but concerned about their potential future fertility should not be tested while continuing to take the birth control pills. In addition, egg donors should not have their AMH levels tested if they are taking birth control pills. Erroneous results may be obtained and therefore erroneous interpretation and recommendations may be made. Women need to be off of oral contraceptive pills for at least one month before the AMH level is tested.
10. AMH levels predict ovarian response in IVF. Higher AMH levels are associated with a greater number of eggs retrieved. And low AMH levels (<0.5 ng/ML) predict poor ovarian response is a high degree of sensitivity and specificity.
11. AMH levels can be used to screen potential egg donors. Egg donors are chosen to the highly fertile and therefore are always <32 years of age and have AMH levels >1.5 ng/ML.
12. AMH levels >3.5 ng/ML are associated with an increased risk of ovarian hyperstimulation syndrome. AMH levels can therefore be used as a guide when deciding on the dose of fertility drugs for ovulation induction. In a patient with a high AMH level, the dose can be lowered and the opposite is true for the patient with a low AMH level.
13. AMH levels may be a marker of ovarian injury in cancer patients who have undergone chemotherapy or radiation.
14. While low AMH levels to suggest poor ovarian reserve, this does not mean that these patients cannot get pregnant with her own eggs; however, many IVF specialists tell patients with low AMH levels that their only chance for pregnancy is to use donor eggs. Without question the chance for pregnancy with donor eggs will be much higher than attempting fertility treatments using the patient’s own eggs, but donor eggs should not be given to patients as their first and only treatment choice.
15. Low AMH levels in young women (<32 years old) does not correlate as well with poor fertility as and a low AMH level in older women (>38 years old). A recent Danish study (Hagen et al., Low concentration of circulating antimullerian hormone is not predictive of reduced fecundability in young healthy women: a prospective cohort study, Fertil Steril 98:1602-8, 2012) followed 186 young couples and calculated the fecundability ratio (FR, the monthly probability of conceiving) and time to pregnancy (TTP, the number of cycles from stopping birth control to pregnancy). AMH levels were measured and were divided into low, medium and high groups. 59% of couples conceived during the study period. There was no difference in the FR in women with low or medium AMH levels (FR 0.81); however, there were very few women who had a very low AMH level (<0.16 ng/ML). In contrast, women with high AMH levels had reduced FR (FR 0.62; however, high AMH levels are often seen in women with polycystic ovary syndrome (PCOS) who may be an ovulatory and therefore infertile.
The conclusions of this study were that a low AMH level in a young healthy woman does not seem to be a predictor of reduced fecundity. This is consistent with high egg quality and niche young women despite a reduced ovarian reserve. Therefore, it appears as though AMH may be a better marker of follicle quantity than necessarily egg quality. Age may be the very best indicator of age quality.
16. It is worth trying alternative medicine to try to improve your ovarian reserve. While some of these treatments have been shown to be effective in some studies others are unproven; however, unless they are likely to cause harm, they may be worth the effort. (A forthcoming blog will address some of these treatments.) There is never a rush to go to donor eggs. The success rate is not dependent upon your age, FSH or AMH levels. It is dependent upon the age and AMH level of your donor.
Do AMH levels predict IVF pregnancy rates?
Anti-mullerian hormone (AMH) levels are now routinely used clinically to assess ovarian reserve. AMH levels are believed to reflect the active pool of oocytes in the ovary not the total number. Primordial follicles (the inactive pool) do not produce AMH. Low AMH level are associated with a reduced number of antral follicles that are available for recruitment and therefore often predicts low ovarian response during an IVF stimulation cycle. AMH levels are therefore believed to reflect the quantity of eggs that are available. The question is: do AMH levels also reflect egg quality ?
A recent study performed in Sweden looked at a group of 892 women undergoing in vitro fertilization (Brodin et al., Antimullerian hormone levels are strongly associated with live-birth rates after assisted reproduction. J Clin Endocrinol Metab 98:1107-14, 2013). They stratified the AMH values into four groups (25th, 50th and 75th percentiles of the observations). The AMH levels ranged from 0.06 to 26.3 ng/ML. The median level was 1.6 and the mean level was 2.3 ng/ML. The 25th and 75th percentiles were 0.84 and 2.94 ng/ml, respectively. The pregnancy rate was positively correlated with AMH levels up to 5 ng/ML and then leveled off. This correlation was true even after adjusting for age. Although there was no AMH level below which no pregnancies occurred, in patients with very low AMH levels (<0.2 ng/ML), only 18 of 53 (34%, includes all ages) resulted in embryo transfer and only three cycles (3% of cycle starts and 17% of embryo transfers) resulted in pregnancy and live birth.
AMH clearly was able to predict pregnancy rate and live birth rate with the chances of the live birth increasing 42% for each AMH stratum.
Overtreating patients with low AMH levels.
AMH is a relatively new blood test for evaluating the ovarian reserve. An increasingly large body of evidence suggests that AMH allows us to estimate the quantity and possibly the quality of eggs left in your ovaries. It certainly gives us a good idea of how well a woman’s ovaries will respond to super ovulation for intrauterine inseminations and/or IVF. It seems to be much better than FSH levels and the clomiphene challenge test.
However, AMH testing does have limitations. As with all biological tests, there is black or white with lots of shades of gray in between. Not only does the tests need to be interpreted by a skilled doctor, it is important to put the test result into the right clinical context. While laying normal AMH level is reassuring many reproductive endocrinologist may not be advised seen their patients well when the AMH level is low. A low AMH level does suggest poor ovarian reserve, but this does not mean that these patients cannot get pregnant with their own eggs. Some IVF specialists reflexively recommend donor eggs as the first treatment of choice for patients with low AMH levels.
Both doctors and patients need to remember that we should not treat lab results – we should be treating the whole patient. Sadly, many reproductive endocrinologist find it easier to advise the “donor eggs fix” to patients with a low AMH rather than talk to the patient about other treatments and options that they can pursue.