Impact of antim€ullerian hormoneassays on the outcomes of in vitrofertilization: a prospectivecontrolled study

Mohamed Aboulghar, M.D.,a,b Walid Saber, M.D.,a,b Yahia Amin, M.D.,b Mona M. Aboulghar, M.D.,a,b

Gamal Serour, F.R.C.O.G.,b,c and Ragaa Mansour, M.D., Ph.D.b

a Department ofObstetrics andGynecology, CairoUniversity; b The Egyptian IVF Center; and c Department ofObstetrics andGynecology, Al-Azhar University, Cairo, Egypt

Objective: To assess the value of routine antim€ullerian hormone (AMH) assays in patients considered high risk for cancellation.Design: Prospective controlled study.Setting: A private IVF center, Cairo, Egypt.Patient(s): In total 4,917 patients received counseling before starting IVF/intracytoplasmic sperm injection (ICSI). They werecomprised of group A1 (n ¼ 1,335), who were considered to be at risk for cancellation after ovarian stimulation, and group A2 (n ¼3,582), who were considered low risk for cancellation. A control group, B (n ¼ 4,639), included group B1 (n ¼ 1,248) and group B2(n ¼ 3,391) based on the same criteria as groups A1 and A2.Intervention(s): An AMH assessment was performed for group A1. All of the patients were stimulated using the long GnRH agonistprotocol. Patients with low AMH levels received the flare-up protocol.Main Outcome Measure(s): The cancellation of IVF/ICSI cycles before or after stimulation, as well as the pregnancy rates (PR) in rela-tion to AMH levels.Result(s): The group A1 patients (6.4%) did not start IVF due to low AMH, and some (6.6%) had their cycles canceled due to poor re-sponses, compared with 2.6% in group A2 and 13.2% in group B1. The clinical PR was 42% in patients with normal AMH and 20% in

Use your smartphone

patients with low AMH. The differences among these three groups were highly significant.Conclusion(s): The AMH assays reduced the cancellations, cost, and stress experienced by cou-ples. (Fertil Steril� 2014;101:134–7. �2014 by American Society for Reproductive Medicine.)Key Words: AMH, poor responder, IVF, cancellation rate

Discuss: You can discuss this article with its authors and with other ASRM members at http://fertstertforum.com/aboulgharm-amh-ivf-poor-responder/

to scan this QR codeand connect to thediscussion forum forthis article now.*

* Download a free QR code scanner by searching for “QRscanner” in your smartphone’s app store or app marketplace.

A ntim€ullerian hormone (AMH) isa dimeric glycoprotein, a mem-ber of the transforming growth

factor (TGF)-b superfamily (1), whichacts on tissue growth and differentia-tion. In women, AMH is produced bythe granulosa cells (GC) in the preantraland antral follicles (2).

Because serum AMH levels reflectthe ovarian follicular pool, any reduction

Received June 29, 2013; revised September 19, 2013;October 23, 2013.

M.A. has nothing to disclose. W.S. has nothing to dinothing to disclose. G.S. has nothing to disclose

Reprint requests: Mohamed Aboulghar, M.D., The EMaadi, Cairo 11431, Egypt (E-mail: gharmd@gm

Fertility and Sterility® Vol. 101, No. 1, January 2014Copyright ©2014 American Society for Reproductivehttp://dx.doi.org/10.1016/j.fertnstert.2013.09.028

134

in the number of small, growing folliclesmay be followed by a reduction in circu-lating AMH. Recently, AMH has beenevaluated by several groups of investiga-tors as a potential novel clinical markerof the ovarian reserve and of a patient'sresponse to gonadotropins (3, 4).

Several studies have demonstratedthat AMH is a better marker of theovarian reserve than age alone and is

accepted September 20, 2013; published online

sclose. Y.A. has nothing to disclose. M.M.A. has. R.M. has nothing to disclose.gyptian IVF Center, 3 St. 161 Hadayek El Maadi,ail.com).

0015-0282/$36.00Medicine, Published by Elsevier Inc.

superior to other markers described inthe literature, such as basal FSH, E2,and inhibin B (5, 6). Age clearlyremains the primary determinant ofthe probability of a live birth afterassisted conception, but for any givenage, women with higher ovarianreserves and therefore highercirculating AMH levels have a highersuccess rate than their counterpartswith lower AMH (7).

The objective of this work was tostudy the impact of routine AMH as-says on infertile women considered tobe at high risk for poor responses inthe context of a large IVF center andduring a 1-year period. Next, we

VOL. 101 NO. 1 / JANUARY 2014

Fertility and Sterility®

compared the results of these women with those of a controlgroup from a previous year in the same IVF center before theintroduction of the AMH assay.

MATERIALS AND METHODSThis is a prospective controlled study for subfertile women(aged%40 years) who were referred to the Egyptian IVF cen-ter for IVF/intracytoplasmic sperm injection (ICSI) in 2011(n ¼ 4,960).

After routine investigations, 4,917 couples (group A) werecounseled to begin stimulation for IVF/ICSI. The patients weredivided into group A1 (n ¼ 1,335), who were considered tohave a high risk for a poor ovarian reserve, and group A2(n ¼ 3,582), who were considered to have a low risk for apoor ovarian reserve.

We hypothesized that women with a diagnosis of unex-plained infertility or endometriosis, women with a historyof ovarian surgery for any condition other than polycysticovary syndrome (PCOS) and women between the ages of 37and 40 years would be at a higher risk of having a poorovarian reserve (group A1).

Routine AMH was performed for all patients in group A1before starting IVF (Beckman Coulter kit, Beckman Coulter).Patients with AMH levels lower than 1 mg/L were recounseledconcerning their IVF prognosis and offered the option of notstarting the IVF stimulation protocol if their chances ofachieving a successful pregnancy were low due to a poorovarian reserve. The assessment of ovarian response wasbased on the AMH levels and the age of the patient. Womenwith AMH levels less than 0.5 mg/L were considered to havea very poor prognosis, particularly if they were more than37 years old. The decision not to start IVF was made by thecouple after we explained the possible cancellation of the cy-cle due to poor ovarian response.

All patients in groups A and B, apart from patients withAMH levels below 1 mg/l, received our routine GnRH agonist(GnRH-a) (Decapeptil, Ferring Pharmaceuticals) long protocol(8). We used the long GnRH-a protocol in the form of daily 0.1mg SC injections, starting from the midluteal phase for 2weeks. Down-regulation was confirmed by E2 levels lessthan 50 pg/mL. Daily injections of FSH or hMG 150–300 IUwere given depending on the age, weight, and AMH levels. In-jections of hCG 10,000 IU were administered when two tothree lead follicles reached 17 mm. Vaginal P was used forluteal phase support. Patients with AMH levels less than 1mg/L, who decided to undergo IVF/ICSI (n ¼ 425), received450 IU of FSH using the ultrashort flare-up protocol.

Flare-up Protocol

Subcutaneous GnRH-a (0.1 mg) was given daily from first dayof the cycle, and 450 IU of FSH or hMGwas injected daily fromday 3 of the cycle. Human chorionic gonadotropin (10,000 IU)was given when the lead follicles reached 17 mm. Luteal phasesupport in the form of vaginal P was provided. The criteria forcycle cancellation were the presence of less than three follicleswith an E2 level less than 500 pg/mL.

Low-risk patients for poor ovarian responses (group A2,N ¼ 3,582) included patients less than age 37 years with no

VOL. 101 NO. 1 / JANUARY 2014

endometriosis and no previous history of surgery on theovaries. These patients were not tested for AMH levels.

Of a control group of 4,680 patients who were referred forIVF in 2008, 41 patients did not join the program for personalreasons, and 4,639 patients (group B) underwent the GnRH-along protocol. These patients were divided into group B1 (n¼1,248), which included high-risk patients for cycle cancella-tion after ovarian stimulation using the same criteria asused for groups A1 and B2 (n¼ 3,391), which consisted of pa-tients with a low risk for cycle cancellation. No AMHwas per-formed on any of these patients.

The ethical committee provided approval before thebeginning of the study. Consent was provided by all patientsincluded in the prospective portion of the study.

Statistical Analysis

Data were statistically described in terms of frequencies(number of patients) and percentages. Comparisons betweenthe study groups were performed using the c2 test. P valuesless than .05 were considered statistically significant. All sta-tistical calculations were performed using the computer pro-gram SPSS (Statistical Package for the Social Science; SPSSInc.) version 15 for Microsoft Windows.

RESULTSThe mean age of group A patients was 34 � 4.1 years,compared with 33.8 � 3.8 years for group B patients (P ¼not significant [NS]).

In group A1 (n ¼1,335), AMH assays were performed onall patients. The mean AMH level in goup A1 patients was 2.1� 1.1 mg/L, and 425 patients (32%) had low AMH levels (<1mg/L). Among these, 87 patients (6.4%) opted not to startstimulation because of their expected poor response basedon the results of their AMH assessment, whereas 1,248 pa-tients underwent ovarian stimulation. After stimulation, 82(6.6%) cycles were canceled due to poor response comparedwith a cancellation of 92 cycles (2.6%) in 3,582 patients ingroup A2 (P< .0001). The total cycle cancellation for groupA1 was 169 cycles (12.6%) compared with 92 patients(2.6%) in group A2 (P< .0001).

In group A1, 425 patients had AMH levels less than 1 mg/L. In this group 82 patients (19.3%) did not reach oocyte aspi-ration due to a poor ovarian response.

There was a highly significant difference in the cancella-tion rate in group A1 (routine AMH) (6.1%) and group B1 (noAMH assay) (13.2%) (P< .0001).

Table 1 shows the details of all cycles in group A and Bpatients. Table 2 shows the clinical pregnancy rate (PR) inall groups. The clinical PR was 42% in patients with normalAMH levels compared with 20% in patients with low AMHlevels (Table 2).

DISCUSSIONThis is a large prospective study of patients who were enrolledfor IVF/ICSI. We show that the routine use of AMH levelsbefore IVF/ICSI in patients considered at a higher risk for cy-cle cancellation significantly reduced the incidence of cycle

135

TABLE 1

Details of cycle cancellation in groups A and B.

Group A Group B P value

Total cycles counseled for stimulation 4,917 4,639No. of high-risk patients 1,335 (group A1) 1,248 (group B1)Total cycles canceled after AMH before stimulation 87 (6.4%) (group A1)No. of high-risk patients stimulated 1,248 (group A1) 1,248 (group B1)Canceled cycles after stimulation 82 (6.6%) (group A1) 165 (13.2%) (group B1) >.0001No. of low-risk patients 3,582 (group A2) 3,391 (group B2)Cancellations in the low-risk group 92 (2.6%) (group A2) 89 (2.6%) (group B2) NSNote: AMH ¼ antim€ullerian hormone; NS ¼ not significant.

Aboulghar. Routine AMH in IVF reduces cancellation rates. Fertil Steril 2014.

ORIGINAL ARTICLE: ASSISTED REPRODUCTION

cancellation, compared with a control group with no AMHassays.

These results could be explained by two factors. First, the87 patients who had poor AMH levels opted not to start IVF,and it is likely that many of these patients would have hadcanceled cycles or poor PRs as a result of poor responses ifthey had started the stimulation. Second, the ultrashortflare-up protocol uses a higher dose of FSH in patients withlow AMH levels, and this may have prevented cancellationin many cycles. However, the literature remains controversialconcerning the value of increasing the dose of FSH and usingflare-up protocols to improve PRs in patients expected to bepoor responders (9, 10).

Our assumption that patients with unexplained infertility,endometriosis, previous ovarian surgery, or aged 37–40 yearswould be at higher risk of cycle cancellation in IVF proved tobe correct. There was a highly significant rate of cancellationdue to poor response in groups A1 and B1, compared withgroups A2 and B2 (Table 1; P>.0001).

Patients with expected poor responses due to low AMHlevels were counseled to ensure that they understood thelow possibility of achieving a successful pregnancy. A poorAMH result provided patients with the opportunity to avoidattempting IVF and instead to revert to egg donation or adop-

TABLE 2

Pregnancy rates in all groups.

Group A Group B P value

No. of patients 4,917 4,639Pregnancies 1,926 (39.2%) 1,901 (40.9%) NS

Group A1 Group B1

No. of patients 1,335 1,248Pregnancies 430 (32.2%) 416 (33.3%) NS

Group A2 Group B2

No. of patients 3,582 3,391Pregnancies 1,496 (41.7%) 1,485 (43.8%) NS

Normal AMH Low AMH (<1 mL)

No. of patients 823 425Pregnancies 345 (41.9%) 85 (20%) < .0001Note: AMH ¼ antim€ullerian hormone; NS ¼ not significant.

Aboulghar. Routine AMH in IVF reduces cancellation rates. Fertil Steril 2014.

136

tion. In countries where IVF is not covered by the nationalhealth scheme, it is more important for patients to recognizetheir chances of a successful pregnancy before proceedingto IVF.

Some of the patients in group A1 with low AMH levelspreferred not to undergo IVF, particularly those close to theage of 40 years. Thus, this group of patients was able to avoidthe loss of a large sum of money, as well as the stress of theIVF procedure or its failure, the latter outcome of which isexpected to be more likely in this group of patients.

In younger patients with low AMH levels, the chance ofpregnancy is reasonable, even if few oocytes are retrieved.These young patients were encouraged to proceed with IVFafter receiving counseling that their results would not beoptimal but would be reasonable. The determination of lowAMH levels resulted in a change in the stimulation protocolto the flare-up protocol, using high doses of FSH. This resultedin a decent clinical PR (20%). It is evident that the routine useof AMH screening in this subgroup of patients improved theoutcomes of IVF and excluded a number of patients whopreferred not to try IVF.

In the low-risk group, the percentage of patients who hadcanceled cycles due to a poor response was 2%. It remains un-clear whether it is cost effective to measure the AMH level forthese patients. A focused cost-effective study should be per-formed. Finally, improved counseling for the prediction ofpoor responsesmay ameliorate disappointment and distress (11).

It is widely accepted that the oocyte yield, independent ofage, predicts live births in IVF cycles. This was demonstratedby an analysis of 400,135 IVF cycles, which showed a linearrelationship between oocyte yield and live births up to 15oocytes (12). The AMH assay is itself strongly associatedwith oocyte yield (5).

As women get older, the criteria for nurturing a successfulembryo become more important. Therefore, a restrictedovarian reserve represents a handicap. Assessing the ovarianreserve represents critical, straightforward information thatshould be imparted to our patients (7). Although cliniciansmay be tempted to exclude patients with a low AMH levelfrom IVF programs on the grounds of cost effectiveness(13), this process will inevitably deny treatment to patientswith low/moderate chances of a successful live birth (14).

The AMH level fails to predict which patients will becomepregnant (15). Neither AMH level nor other ovarian reservetests can accurately predict the quality of an oocyte (16).

VOL. 101 NO. 1 / JANUARY 2014

Fertility and Sterility®

The preexisting issues among the different AMH assayshave been resolved because Beckman Coulter purchased allpatents for all previous versions of the assay to establish theAMH Generation II assay. Although this action establisheda monopoly (17), the use of distinct assays among differentresearchers has been standardized. Still, it was recently re-ported that the AMH levels in serum from 48 subjects incu-bated at room temperature for up to 7 days increasedprogressively inmost samples (58% increase overall). The pre-dilution of serum before the assay provided AMH levels up totwice those found in the corresponding undiluted samples.The premixing of serum with assay buffer before their addi-tion to the microtiter plate gave higher readings (72% overall).Storage at �20�C for 5 days increased AMH levels by 23%compared with fresh samples. Antim€ullerian hormone maynot be stable under some storage or assay conditions (18).

Our study proved that routine AMH assays in a subgroupof patients can effectively identify a higher percentage of pa-tients who would not respond effectively to ovarian stimula-tion, as well as those patients who would have a significantlylower PR (Table 2) (19). Also, we noticed that the predictivevalue of AMH levels for poor responses was not absolute,based on consequent false-positive and false-negative results.In particular, false-positive results may have negative conse-quences on the couple's lives because this type of result mightincorrectly prohibit these women from undergoing IVF. Inparticular, younger poor responders have a different prog-nosis compared with older poor responders (20, 21). Hence,before proposing AMH measurements in ovarian reservetesting, we should better define the aim of the testingprocess itself.

In conclusion, the routine assay of AMH levels for higherrisk groups of poor responders proved to be of value in coun-seling patients to decide whether to proceed to IVF cycles.These data also helped reduce the cycle cancellation rate.

REFERENCES1. Cate RL, Mattaliano RJ, Hession C, Tizard R, Farber NM, Cheung A, et al.

Isolation of the bovine and human genes for M€ullerian inhibiting substanceand expression of the human gene in animal cells. Cell 1986;45:685–98.

2. Weenen C, Laven JS, von Bergh AR, Cranfield M, Groome NP, Visser JA,et al. Anti-Mullerian hormone expression pattern in the human ovary: po-tential implications for initial and cyclic follicle recruitment. Mol HumanitRep 2004;10:77–83.

3. Van Rooij IA, Broekmans FJ, te Velde ER, Fauser BC, Bancsi LF, Jong FH, et al.Serum anti-Mullerian hormone levels: a novel measure of ovarian reserve.Hum Reprod 2002;17:3065–71.

4. La Marca A, Giulini S, Tirelli A, Bertucci E, Marsella T, Xella S, et al. Anti-Mullerian hormonemeasurement on any day of the menstrual cycle stronglypredicts ovarian response in assisted reproductive technology. Hum Reprod2007;22:766–71.

VOL. 101 NO. 1 / JANUARY 2014

5. LaMarca A, Nelson SM, SighinolfiG, MannoM, Baraldi E, Roli L, et al. Anti-Mullerian hormone (AMH) based prediction model for the live birth inassisted reproductive technology (ART). Reprod Biomed Online 2010;22:341–9.

6. Anchaert E, Smitz J, Schiettecatte J, Klein BM, Arce JC. The value of anti-Mullerian hormone measurement in the long GnRH agonist protocol: asso-ciation with ovarian response and gonadotrophin dose adjustments. HumReprod 2012;27:1829–39.

7. Nelson SM, Anderson RA, Broekmans FJ, Raine-Fenning N, Fleming R, LaMarca A. Anti-M€ullerian hormone: clairvoyance or crystal clear? Hum Re-prod 2012;27:631–6.

8. Aboulghar MA, Mansour RT, Serour GI, Amin YM. The prognostic value ofsuccessful in vitro fertilization in subsequent trials. Hum Reprod 1994;9:1932–4.

9. Pandian Z, McTavish AR, Aucott L, Hamilton MP, Bhattacharya S. Interven-tion for poor responders to controlled ovarian hyperstimulation (COH) inin vitro fertilization (IVF). Cochrane Detabase Syst Rev 2010:CD004379.

10. Fleming R, Broekmans F, Calhaz-Jorge C, Dracea L, Alexander H, NyboeAndersen A, et al. Can anti-Mullerian hormone concentrations be used todetermine gonadotrophin dose and treatment protocol for ovarian stimula-tion? Reprod Biomed Online 2013;26:431–9.

11. La Marca A, Sighinolfi G, Radi D, Argento C, Baraldi E, Carducci A, et al.Anti-Mullerian hormone (AMH) as a predictive marker in assisted reproduc-tive technology (ART). Hum Reprod Update 2010;16:113–30.

12. Sunkara SK, Rittenber V, Raine-Fenning N, Bhattacharya S, Zamora J,Coomarasamy A. Association between the number of eggs and live birthin IVF treatment: an analysis of 400135 treatment cycles. Hum Reprod2011;26:1768–74.

13. Yates AP, Rustamov O, Roberts SA, Lim HYN, Pemberton PW, Smith A, et al.Anti-Mullerian hormone-tailored stimulation protocols improve outcomeswhilst reducing adverse effects and costs of IVF. Hum Reprod 2011;26:2353–62.

14. Weghofer A, DietrichW, Barad DH, Gleicher N. Live birth chances in womenwith extremely low-serum anti-Mullerian hormone levels. Hum Reprod2011;26:1905–9.

15. Lamazou F, Genro V, Fuchs F, Grynberg M, Gallot V, Achour-Frydman N,et al. L'AMH seique n'a pas de valeur predictive en fecundation in-vitro encycle natural modifie: analyse de 342 cycles The AMH assay is not of predic-tive value in modified natural cycle IVF: analysis of 342 cycles [in French].J Gynecol Obstet Biol Reprod 2011;40:205–10.

16. Loh JS, Maheshwari A. Anti-Mullerian hormone—is it a crystal ball for pre-dicting ovarian ageing? Hum Reprod 2011;26:2925–32.

17. Nelson SM, La Marca A. The journey from the old to the new AMH assay:how to avoid getting lost in the values. Reprod Biomed Online 2011;23:411–20.

18. Rustamov O, Smith A, Roberts SA, Yates AP, Fitzgerald C, Krishnan M, et al.Anti-Mullerian hormone: poor assay reproducibility in a large cohort of sub-jects suggests sample instability. Hum Reprod 2012;27:3085–91.

19. Brodin T, Hadziosmanovic N, Berglind L, Olovsson M, Holte J. Antimullerianhormone levels are strongly associated with live-birth rates after assistedreproduction. J Clin Endocrinol Metab 2013;98:107–14.

20. Lashen H, Ledger W, Lopez-Bernal A, Barlow D. Poor responders to ovula-tion induction: is proceeding to in-vitro fertilization worthwhile? HumReprod 1999;14:964–9.

21. Ulu U, Ben-Shlomo I, Turan E, Erden HF, AkmanMA, Bahceci M. Conceptionrates following assisted reproduction in poor responder patients: a retro-spective study in 300 consecutive cycles. Reprod Biomed Online 2003;6:439–43.

137