O-067 Does the measurement of different AMH isoforms in poor responder patients improve the prediction of ovarian response?

Abstract

Abstract Study question Does the measurement of AMH isoforms in low ovarian reserve patients improve the prediction of the number of oocytes after ovarian stimulation (OS)? Summary answer AMH assays measuring a broader range of AMH isoforms combined with antral follicle count (AFC) improve the prediction of oocytes collected after OS. What is known already Granulosa cells secrete AMH as a non-active prohormone (proAMH). After cleavage, it forms a complex (AMHN,C) which activates the receptor (AMHR2). Circulating AMH is a mixture of isoforms (proAMH, AMHN,C and other sub-fragments) after proteolysis, targeted by the monoclonal antibodies included in the AMH assays commonly used in clinical practice, potentially affecting the quantification of the hormone. Novel AMH assays include antibodies directed towards a broader range of the molecule. They may provide more accurate and reliable results, especially important in poor responders, as results may involve clinical decisions on whether or not to proceed with OS. Study design, size, duration Prospective observational study measuring AMH isoforms using five different AMH assays, including 72 women with low ovarian reserve performing OS for IVF/ICSI at a tertiary referral fertility center, from February 2019 to December 2021. Low ovarian reserve was defined as AMH serum levels <1.1ng/ml (Elecsys, Roche), following Bologna criteria. All patients performed OS for IVF/ICSI with antagonist protocol. Participants/materials, setting, methods On day 2/3 of cycle, prior to initiating the OS, AFC, serum FSH, LH, estradiol, progesterone and AMH levels were measured using Elecsys assay (Roche). Extra serum samples were frozen at -20C for subsequent analysis, using four different AMH assays (AnshLabs, Texas): AL-196 (PCOCheck ELISA), AL-124 (PicoAMH ELISA), AL-105 (US-AMH ELISA) and AL-133 (Total Mature-AMH ELISA). Ethical approval was obtained (Research Ethics Committee-REFA033c) and informed consent was signed by all participants. Main results and the role of chance Patient characteristics were [Median(IQR)]: age:39(36-42)years, BMI:28.51(26.1-30.1) Kg/m2, AFC:5.5(4-7), FSH:8.23(6.4-12.5) mIU/mL, LH 6.54(4.51-8.91) mIU/mL, Estradiol: 41.18(28.47-56.44)pg/mL, AMH-Elecsys: 0.64(0.29-0.81)ng/mL, AL-196:0.64(0.39-0.99)ng/mL, AL-124:0.79(0.49-1.24ng/mL, AL-105: 0.89(0.57-1.39)ng/mL, AL-133:1.06(0.63-1.59)ng/mL. Stimulation outcomes were: follicle number on the day of trigger (Fdot):5(3-7), cumulus-oocyte-complexes (COC’s):3(2-4.5) and metaphase II oocytes (MII):3(1-4). Spearman correlation was performed between AMH assays, AFC and OS outcomes. Elecsys revealed a good correlation with the other AMH assays (rs = 0.60-0.63,p<0.001), however, it was higher among ELISA assays (rs = 0.954-0.993,p<0.001). All AMH assays showed a significant positive correlation with AFC, Fdot, COC’s and MII. However, AL-196 showed the highest correlation (rs = 0.524, rs = 0.615, rs = 0.594, rs = 0.595, respectively;p<0.001). Different models combining one AMH assay plus AFC were created to predict COC’s and MII: AFC+AL-196 revealed the best prediction (Adjusted R2=0.474[p < 0.001] and 0.485[p < 0.001], respectively) when compared to the same AMH assay alone. AUROC analysis was performed to investigate which model predicted better < =3COC’s after OS. Although all models demonstrated fair values, AUROC was better for AFC+AL-196 model (AUROC:0.796), yet no significant differences were seen when compared to AUROC for AFC+AMH-Elecsys (p = 0.39). AFC and AMH were good predictors for COC’s collected [RR(95%CI):1.49(1.11-2.04),p<0.01; RR(95%CI):1.26(1.13-1.49),p<0.001, respectively). AFC+AMH-Elecsys predicted COC’s better than AFC alone (R2=0.47 vs 0.29, respectively;Performance-Score:42.56% vs 13.21%, respectively). AFC+AL-196 had the highest performance (R2=0.599;Performance-Score:90.12%). Limitations, reasons for caution Although clinical assessment was performed in the same centre following the same methodology, inter-observer variability for AFC is a limitation. Besides, fresh AMH serum samples were analysed using Elecsys assay, whereas frozen serum samples were shipped to Ansh Lab (Texas) for batched analysis. Wider implications of the findings For patients with serum AMH<1.1ng/mL, a wide range of oocytes might be expected after OS. Novel AMH assays, targeting a wider broad of the molecule, together with AFC, can be used for these patients to obtain a better prediction for clinical outcomes and anticipate very poor response. Trial registration number Not applicable