Study question: What is the relationship between the anti-Müllerian hormone (AMH), gonadotropin and androgen concentrations within a single follicle and live birth after ICSI and a transfer of an embryo developed from the matched oocyte?

Summary answer: Among the analysed markers on the day of oocyte retrieval, AMH concentration in follicular fluid (FF) is a predictor of live birth after single embryo transfer (SET).

What is known already: High serum concentrations of AMH and low FSH concentrations have been associated with a high chance of pregnancy after ART. Whether there are differences in the hormonal milieu for individual follicles and whether this impacts the laboratory and clinical outcomes for the individual oocyte developing within that follicle are unknown.

Study design, size, duration: This prospective cohort study included 322 individual FF samples from 199 infertile women scheduled for ICSI/SET over an 18-month period. Of these women, 76 provided a single FF sample, while 123 women contributed two FF samples taken from two different follicles.

Participants/materials, setting, methods: The first follicle aspirated in each ovary on the day of oocyte retrieval had the FF aspirated; the individual cumulus-oocyte complex (COC) was tracked, and the associated FF was stored at -80°C. FF AMH, FSH, LH, testosterone (T) and androstenedione (A2) levels were measured by mass spectrometry (androgens) and immunoassays. The laboratory and clinical outcomes for each individual oocyte were related to their unique follicle hormone concentrations.

Main results and the role of chance: Of the 322 oocytes with paired FF samples, 70 (21.7%) oocytes did not fertilise. From the remaining 252 2PN embryos, 88 (34.9%) were transferred as single embryos on Day 3; of the remaining 164, 78 developed into blastocysts, and 18 single blastocyst transfers were performed. Thus, a total of 106 transferred embryos had matching FF samples. An analysis of these individual FF concentrations revealed that AMH concentrations were higher in follicles in which the oocyte developed into a top quality (TQ) blastocyst (6.33 ± 5.52 ng/ml) and whose transfer led to live birth (7.49 ± 5.03 ng/ml) than those in which there was a failure of fertilisation (3.34 ± 2.21 ng/ml). In contrast, follicular FSH concentrations were the lower for oocytes that resulted in a TQ blastocyst (5.36 ± 2.20 mIU/ml) and live birth (5.60 ± 1.41 mIU/ml) than for oocytes that failed to fertilise (9.06 ± 3.36 mIU/ml). FF AMH was the only studied marker that increased the chance of live birth (odds ratio: 1.93 [95% CI: 1.40-2.67], P < 0.001). The receiver operating characteristic analysis showed that FF AMH levels predicted live birth with a very high sensitivity (91.2%), specificity (91.7%) and an excellent AUC value of 0.954, whereas serum AMH level only had a fair (AUC = 0.711) significance as a predictor for live birth after ICSI/SET. The predictive capabilities of the interfollicular markers were not limited to the TQ embryos or blastocysts; they applied to all SET cycles.

Limitations, reasons for caution: Whether an altered intrafollicular hormonal environment reflects the developmental capacity of the oocyte or defines cannot be determined from this cross-sectional analysis. Inclusion of 21 subjects with polycystic ovary syndrome (PCOS) may have biased the findings due to a unique intrafollicular milieu associated with PCOS.

Wider implications of the findings: Our results suggest that highly competent human oocytes have an FF composition of AMH, FSH, T and A2 that is close to that in a natural cycle. Also, the relationships between intrafollicular AMH, gonadotropin and androgen levels in the same follicle support the hypothesis that FF AMH concentration may reflect granulosa cell proliferation during gonadotropin-stimulated follicle growth. Finally, the serum AMH concentration is markedly lower than the FF AMH concentration, with a moderate correlation between serum and FF AMH, implying ovarian follicle autonomy with regards to its secretory products.

Study funding/competing interest(s): The National Science Centre of Poland supported this work (grant number: N N407 217 040). The authors declare that there is no conflict of interest regarding the publication of this article.