No abstract

STUDY QUESTION What is the association between serum progesterone levels on the day of frozen embryo transfer (FET) and the probability of live birth in women undergoing different FET regimens? SUMMARY ANSWER Overall, serum progesterone levels <7.8 ng/ml were associated with reduced odds of live birth, although the association between serum progesterone levels and the probability of live birth appeared to vary according to the route of progesterone administration. WHAT IS KNOWN ALREADY Progesterone is essential for pregnancy success. A recent systematic review showed that in FET cycles using vaginal progesterone for endometrial preparation, lower serum progesterone levels (<10 ng/ml) were associated with a reduction in live birth rates and higher chance of miscarriage. However, there was uncertainty about the association between serum progesterone levels and treatment outcomes in natural cycle FET (NC-FET) and hormone replacement therapy FET (HRT-FET) using non-vaginal routes of progesterone administration. STUDY DESIGN, SIZE, DURATION This was a multicentre (n = 8) prospective cohort study conducted in the United Kingdom between January 2020 and February 2021. PARTICIPANTS/MATERIALS, SETTING, METHODS We included women having NC-FET or HRT-FET treatment with progesterone administration by any available route. Women underwent venepuncture on the day of embryo transfer. Participants and clinical personnel were blinded to the serum progesterone levels. We conducted unadjusted and multivariable logistic regression analyses to investigate the association between serum progesterone levels on the day of FET and treatment outcomes according to the type of cycle and route of exogenous progesterone administration. Our primary outcome was the live birth rate per participant. MAIN RESULTS AND THE ROLE OF CHANCE We studied a total of 402 women. The mean (standard deviation) serum progesterone level was 14.9 (7.5) ng/ml. Overall, the mean adjusted probability of live birth increased non-linearly from 37.6% (95% confidence interval [CI] 26.3%-48.9%) to 45.5% (95% CI 32.1%-58.9%) as serum progesterone rose between the 10th (7.8 ng/ml) and 90th (24.0 ng/ml) centiles. In comparison to participants whose serum progesterone level was ≥7.8 ng/ml, those with lower progesterone (<7.8 ng/ml, 10th centile) experienced fewer live births (28.2% versus 40.0%, adjusted odds ratio [aOR] 0.41, 95% CI 0.18-0.91, P = 0.028), lower odds of clinical pregnancy (30.8% versus 45.1%, aOR 0.36, 95% CI 0.16-0.79, P = 0.011) and a trend towards increased odds of miscarriage (42.1% versus 28.7%, aOR 2.58, 95% CI 0.88-7.62, P = 0.086). In women receiving vaginal progesterone, the mean adjusted probability of live birth increased as serum progesterone levels rose, whereas women having exclusively subcutaneous progesterone experienced a reduction in the mean probability of live birth as progesterone levels rose beyond 16.3 ng/ml. The combination of vaginal and subcutaneous routes appeared to exert little impact upon the mean probability of live birth in relation to serum progesterone levels. LIMITATIONS, REASONS FOR CAUTION The final sample size was smaller than originally planned, although our study was adequately powered to confidently identify a difference in live birth between optimal and inadequate progesterone levels. Further, our cohort did not include women receiving oral or rectal progestogens. WIDER IMPLICATIONS OF THE FINDINGS Our results corroborate existing evidence suggesting that lower serum progesterone levels hinder FET success. However, the relationship between serum progesterone and the probability of live birth appears to be non-linear in women receiving exclusively subcutaneous progesterone, suggesting that in this subgroup of women, high serum progesterone may also be detrimental to treatment success. STUDY FUNDING/COMPETING INTERESTS This work was supported by CARE Fertility and a doctoral research fellowship (awarded to P.M.) by the Tommy’s Charity and the University of Birmingham. M.J.P. is supported by the NIHR Birmingham Biomedical Research Centre. S.F. is a minor shareholder of CARE Fertility, but has no financial or other interest with progesterone testing or manufacturing companies. P.L. reports personal fees from Pharmasure, outside the submitted work. G.P. reports personal fees from Besins Healthcare, outside the submitted work. M.W. reports personal fees from Ferring Pharmaceuticals, outside the submitted work. The remaining authors have no conflict of interest to declare. TRIAL REGISTRATION NUMBER ClinicalTrials.gov: NCT04170517

Background: Various factors, including treatment protocols, can influence the outcomes of frozen embryo transfers (FETs). The study objectives were to compare different endometrial preparation protocols of FET cycles and to evaluate the factors, including the endometrial thickness (ET), that affect outcomes. Methods: This observational cohort study involved 5037 women undergoing FETs at eight tertiary clinics in the UK between January 2016 and March 2019. The endometrial preparation protocols used were natural cycle (NC-FETs), artificial hormone support cycle with oestradiol valerate but without pituitary downregulation (AC-FETs) and artificial hormone support cycle with agonist downregulation (ACDR-FETs). Results: The mean (±SD) ages across NC-FET, AC-FET and ACDR-FET groups were 36.5 (±4.2), 35.9 (±5.0) and 36.4(±4.9) years, respectively. LBRs were comparable (40.7%, 175/430; 36.8%, 986/2658; and 36.7%, 716/1949, respectively) across the three groups. Clinical pregnancy, implantation, multiple pregnancies, miscarriage and ectopic pregnancy rates were also similar. In the regression analysis of variables including age, duration of infertility, number of embryos transferred, protocol type and endometrial thickness, age was the only significant predictor of LBRs, although its predictive ability was poor (AUC: 0.55). With the overall LBR of the study population being 37.1%, the post-test probability of a live birth at an ET of <5 mm was 0%, and at 5–5.9, 6–6.9, 7–7.9 and 8–8.9 mm, the probabilities were 16.7%, 33.8%, 36.7% and 37.7%, respectively. The LBR remained above 35% up to the 14–14.9 mm range and then declined gradually to 23% for the 17–25 mm range. Conclusions: The FET outcomes were similar for the three protocols used for endometrial preparation. The protocol type and endometrial thickness were not predictive of FET outcomes; age was the only predictive variable, despite its low predictive ability.