Assessment of interindividual follicular size variation at ovulation in natural‐cycle frozen embryo transfer

Ovulation is mediated by a complex interplay of estradiol (E2), progesterone (P4) and luteinizing hormone (LH). Correct identification of ovulation, with a subsequent rise in P4 that creates a receptive endometrium, is critical for successful embryo implantation following natural-cycle (NC) frozen embryo transfer (FET). While it is known that NC-FET reduces the risk of pre-eclampsia in comparison with programmed-cycle FET1, this approach is still under-represented in clinical practice, partially because of the high cost of monitoring with frequent ultrasonography and hormonal testing. Women with regular menses have a mean ± SD follicular phase length of 16.9 ± 5.3 days, and ovulate at a maximum follicle size (FS) of 19.8–21.9 mm and a peak E2 level of 200–300 pg/mL2-4. However, variation between cycles is well described, and planning monitoring intervals and embryo-transfer timing based purely on any of these parameters may be unreliable. We conducted a retrospective study to investigate if variation in these parameters influences the timing of ovulation. The study included 262 cycles in 246 women who underwent NC-FET and had obtained a measurement of the dominant follicle on the day of ovulation, defined as day 0 (D0). Ethical approval was obtained from ART Fertility Clinics, Abu Dhabi, United Arab Emirates (REFA117/117a). The first ultrasound scan and serum hormonal assessment of E2, P4 and LH were performed on cycle day 10. Women were followed up in intervals of 1 to 3 days, defined by the trajectory of the serum levels of LH (rise), E2 (initial rise/decline after ovulation) and P4 (rise) (Table S1). FS was measured using transvaginal ultrasound and calculated as the mean of the two greatest diameters of the follicle. Ovulation (D0) was defined by a serum P4 level of ≥ 0.9 ng/mL, followed by a further P4 rise. Embryo transfer was performed on day 5 and sufficient serum P4 levels (> 8–10 ng/mL) were confirmed on day 4. In the 262 cycles, ovulation occurred at a median FS of 20.0 (interquartile range (IQR), 18.0–22.0; range, 12.0–26.0) mm, on median cycle day 15.0 (IQR, day 12.3–17.0; range, day 10.0–28.0) and a median E2 level of 222.0 (IQR, 131.4–292.1; range, 50.2–698.5) pg/mL (Table 1, Figure 1). There was no significant association between the women's age or body mass index and FS, E2 level or cycle day at ovulation. Only the E2 level on D0 was significantly associated with FS (P = 0.002), and women with a higher anti-Müllerian hormone (AMH) level appeared to have a slightly higher E2 level on D0 (8.8 (95% CI, 2.4–15.3) pg/mL higher per 1 ng/mL rise in AMH, P = 0.007) (Figure 2). However, the overall model fit was quite poor (pseudo-R2 = 0.033). Despite significant variation between women in FS, cycle day and E2 level at ovulation, women who underwent repeat NC-FET cycles (n = 16) did not show significant intercycle changes (Figure S1, Table S2). While our results are consistent with published data regarding average FS, cycle day and E2 level at ovulation3, 4, it is important to note that there was considerable variation in FS at ovulation (range, 12–26 mm). Delaying hormonal monitoring until FS reaches a predetermined threshold (i.e. > 16 mm) may mean that the initial P4 rise is missed, resulting in embryo–endometrial asynchrony and implantation failure. Moreover, the wide range of results for E2 level and cycle day at ovulation render these parameters unreliable alone as a tool for establishing monitoring intervals and/or prediction of ovulation in the first NC-FET cycle. However, they may have increased utility in repeat cycles of the same woman. Interestingly, LH, the most widely used surrogate parameter for determining ovulation, did not correlate with FS at ovulation. This observation is consistent with the recently described distinct pattern of LH rise prior to ovulation, corroborating the existence of cycle variability5. To conclude, ultrasonographic assessment of follicle growth and endometrial features is crucial for NC-FET monitoring; however, frequent monitoring may strain the patient and be viewed as a limitation of the NC method. Due to ovulation occurring at considerably different FS, ovulation can be confirmed finally by P4 rise. A small preovulatory P4 peak, to approximately 0.5 ng/mL, precedes the postovulatory P4 rise4. Sole P4 measurements could be used for identification of ovulation and the planning of the embryo transfer, thus reducing the costs of hormonal monitoring. The data that support the findings of this study are available from the corresponding author upon reasonable request. Table S1 Monitoring protocol for endometrial preparation for natural-cycle frozen embryo transfer Table S2 Mixed-effects regression analysis of effect of cycle number on estradiol (E2) level, cycle day and follicle size at ovulation in 16 repeat cycles Figure S1 Intercycle changes in cycle day, estradiol (E2) level and follicle size at ovulation in 16 women who underwent repeat natural-cycle frozen embryo transfer Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.