Optimizing embryo transfer timing based on serum progesterone levels from a known implantation cohort: A...

True natural thawed embryo transfer (tNC-FET) is increasingly popular due to favorable obstetric and perinatal outcomes. Accurate ovulation timing is essential for synchronizing the embryo with the endometrium and determining the optimal ET timing. While serum LH surge and ultrasound-detected follicle collapse are common methods, they often require frequent monitoring, causing inconvenience and higher cancellation rates. This study aimed to assess serum progesterone (P4) levels in a known implantation protocol for tNC-FET and to compare monitoring visits and clinical outcomes between the serum progesterone-based and alternative timing approaches.

A retrospective analysis of tNC-FET leading to successful implantation at National Taiwan University Hospital (Jan 2000-Dec 2015). In these cycles, follicular growth and serum hormone levels were closely monitored daily or every other day via ultrasound examination, beginning 3-4 days before the anticipated ovulation. A serum LH level exceeding 30 IU/L was defined as an LH surge, and the day of dominant follicle collapse was designated as the ovulation day. We reviewed each protocol and collected serum progesterone (P4) levels according to the follow-up schedule. A subsequent retrospective analysis of tNC-FET cycles from January 2016 to December 2022 categorized cycles into two groups based on embryo transfer (ET) timing. Group 1 (P4-based protocol): ET timing determined by serum progesterone (P) levels. Day 1 was defined as 1.43 ≤ P < 3.16 ng/mL, day 2 as 3.16 ≤ P < 6.55 ng/mL, day 3 as 6.55 ≤ P < 9.26 ng/mL, and day 4 as P ≥ 9.26 ng/mL. These thresholds were based on P levels observed from January 2000 to December 2015. Group 2 (LH/follilce-collapse-based protocol): ET timing based on the day after the LH surge (day -1) or the day of follicle collapse (day 0). Thawed blastocysts were all transferred on day 5. Statistical comparisons were made using T-tests, Pearson's χ² tests, logistic regression and a linear mixed model (LMM) to adjust for operator variability.

In the know implantation cohort (Jan 2000-Dec 2015), serum progesterone levels on day -1 were 0.78 ± 0.49 ng/mL (range: 0.2-6.41, Q1-Q3: 0.47-0.95 ng/mL), 1.28 ± 0.56 ng/mL on day 0 (range: 0.2-4, Q1-Q3: 0.92-0.56 ng/mL), 2.27 ± 1.2 ng/mL on day 1 (range: 0.36-0.89, Q1-Q3: 1.43-2.8 ng/mL), 3.98 ± 1.19 ng/mL on day 2 (range: 0.67-8.03, Q1-Q3: 3.16-4.58 ng/mL) and 7.87 ± 3.05 ng/mL on day 3 (range: 1.12-19.1, Q1-Q3: 6.55-9.27 ng/mL). The baseline characteristics and clinical pregnancy outcomes were similar between the group 1 and 2. After adjusting for operator variability, no significant difference in clinical pregnancy outcomes was observed. Group 1 had significantly fewer monitoring visits (2.69 ± 0.74 vs. 3.45 ± 1.31, p < 0.0001) and cycle cancellations (3.6 % vs. 12.5 %, p < 0.0001). Most cancellations in Group 2 were due to ovulation before the implantation window. Adjusting for operator variability using a linear mixed model, Group 1 still required fewer monitoring visits (b = -1.00, p = 0.0008).

Using center-specific P4 thresholds derived from a known implantation cohort, our protocol enables accurate ET timing with fewer monitoring visits and lower cancellation rates. This flexible approach allows scheduling even after ovulation, maintaining comparable pregnancy outcomes.

This study was retrospectively registered with the Institutional Review Board of National Taiwan University Hospital (202309092RINA).

not applicable.