P–682 Serum progesterone level as prognostic factor in frozen-thawed embryo transfer cycles: effect of selected threshold on gestational results. Systematic review, stratified meta-analysis and meta-regression

Study question Is there an optimum progesterone threshold level below which gestational results are significantly worse in frozen embryo transfer cycles (FET) with hormone replacement therapy (HRT)? Summary answer Low serum progesterone during luteal phase of HRT-FET cycles impairs substantially its gestational outcomes, regardless of threshold level, origin of oocytes and euploidy of embryos. What is known already HRT for endometrial preparation in FET or oocyte donation cycles is widely used. Oestrogen doses are usually patient-tailored varying upon endometrial thickness, whereas the optimal level of progesterone exposure has not been defined. Various studies have found a negative association between serum progesterone levels measured during luteal phase and FET results in terms of pregnancy and miscarriage rates. Most likely there is an optimal level below which results are worse but a standard threshold level is yet to be established, as in almost every study a different threshold has been found. Study design, size, duration Systematic review and stratified meta-analysis with meta-regression following PRISMA guidelines. An electronic search of MEDLINE, EMBASE, Web of Science, Cochrane Gynaecology and Fertility Specialised Register of Controlled Trials and ClinicalTrials.gov was conducted from inception to January 2021. The aim was to identify prospective or retrospective cohort studies measuring serum progesterone levels around frozen embryo transfer date in HRT cycles. A combination of the following key search terms was used: “progesterone”, “serum”, “frozen embryo”, “transfer”, “frozen-thawed”. Participants/materials, setting, methods Studies analyzing association of luteal serum progesterone with FET-HRT outcomes were included. Risk of bias within studies was assessed using the Newcastle-Ottawa Scale (NOS). Clinical/ongoing pregnancy and miscarriage rates (C/OPR,MR) were considered as primary and secondary outcomes respectively. Odds Ratios with 95% Confidence Interval (OR,95%CI) were calculated applying a random effects model meta-analysis. Heterogeneity was assessed using the I2 statistic. A meta-regression was conducted to examine the association of the effect with the threshold level. Main results and the role of chance The systematic search retrieved 792 studies, 494 after duplicates removal of which 343 were screened and 51 assessed for eligibility. 12 studies, reporting 14 threshold levels, were included in the meta-analysis involving 5009 HRT-FET cycles. Two of them were prospective cohort studies while the rest were retrospective. 10 of them have been published in peer review journals and two were conference abstracts. Quality of studies assessed with NOS varied between 5 and 9. The progesterone threshold ranged from 5.0 to 21.94 ng/ml. Low progesterone levels were associated with less C/OPR (OR: 0.52; 95% CI: 0.40 to 0.66; 11 studies, 5009 cycles). Low progesterone was also associated with high MR (OR: 2.01; 95% CI: 1.57 to 2.58; 9 studies, 2560 pregnancies). These effects showed remarkable consistency in specific sub-analyses considering separately studies with progesterone thresholds up to or above 10 mg/mL, and studies carried out in cycles using oocyte donation, autologous oocytes and embryo aneuploidies screening. Meta-regression did not identify significant association between size effect and progesterone threshold, regarding neither C/OPR (regression coefficient: 0.02; CI 95%: –0.02 to 0.06; p: 0.28) nor MR (regression coefficient: 0.11; CI 95%: –0.13 to 0.36; p: 0.32). Limitations, reasons for caution High degree of clinical and statistical heterogeneity was found due to different routes and doses of progesterone administration, date of progesterone analyses and variety of thresholds as well as high diversity of embryo origin. Despite sensibility analysis by embryo origin any of these sources of heterogeneity can preclude the results. Wider implications of the findings: Despite low progesterone levels are significantly associated to lower gestational results, and a threshold of 10 ng/ml constitutes the median value of our distribution, high quality prospective studies are needed to validate the prognostic value of progesterone levels and to establish an standardised threshold level for clinical application. Trial registration number not required