P-659 Natural cycle versus medically programmed cycle in frozen embryo transfer: a retrospective analysis

Abstract

Abstract Study question Does frozen embryo transfer (FET) in natural cycle (NC) and programmed cycle (PC) differ in achieving live birth, and what are the confounding factors? Summary answer The probability of live birth was significantly higher in natural cycle FET compared to programmed cycle. What is known already FET can be performed in natural or programmed (hormonally modified) cycle but the data regarding the most optimal preparation of the endometrium remains conflicting and inconclusive. The latest Cochrane review (Glujovsky, 2020) including very low- to moderate-quality evidence could not demonstrate superiority of either protocol on life birth rate or other pregnancy outcomes. Study design, size, duration This is a retrospective single center registry study including all frozen embryo transfers performed at Helsinki University Hospital Reproductive Medicine Unit during 2016-2020. Main outcome measure was live birth and secondary outcomes were biochemical pregnancy, ectopic pregnancy and miscarriage. The study cohort included 5435 attempted frozen embryo transfers to 1913 women, of which 1735 cycles were cancelled, mostly due a poor cycle or a weekend in a 5-day clinic, and 3700 were performed. Participants/materials, setting, methods Study included 2581 FETs in NC and 1119 in PC. The collected data included patient and male partner age, BMI, smoking, underlying health conditions, causes of infertility, endometrial thickness, embryo quality classification and the pregnancy outcome. Embryos from donor oocytes were excluded. To assess the effect of different confounding factors, we performed a regression analysis with a generalized linear mixed model. Main results and the role of chance Mean age of the women in the PC group was slightly younger (33.5, 33.0-33.9 years) than in the NC group (34.5, 34.1-34.5 years; P < 0.001) years. Mean BMI was not statistically different (NC 24.2 vs PC 25.1 kg/m2), but significantly more women with BMI above 30 (PC 19.5% vs NC 10.2%; P < 0.001), as well as women with ovulation disorder (PC 67.2% vs NC 5.2%) were observed in the PC group. Smoking was more common in the NC group compared to the PC group (5.9 vs 4.0%, P < 0.001). Mean endometrial thickness at check-up visit prior to FET was 8.3 (8.1-8.5) mm in the PC group compared to 7.7 (7.6-7.8; P < 0.001) mm in the NC group. We observed a lower live birth rate in the PC group (23.3%) compared to the NC group (29.1%, OR 0.74, 95% CI 0.63-0.87). However, when adjusted with most significant confounding factors (female age, embryo quality score and tubal factor) the effect of the FET protocol on LBR was diminished. Incidence of miscarriage was higher in the PC group compared to the NC group, but no statistically significant difference was observed in the other secondary outcomes of biochemical pregnancy or ectopic pregnancy. Limitations, reasons for caution Retrospective design may include a selection bias between the study groups, although the demographic characteristics of the groups were quite comparable. Wider implications of the findings Confounding factors seem to explain much of the inferiority of programmed cycle compared to the natural cycle in terms of achieving live birth. However, prospective, randomized controlled trials including women with normal, ovulatory cycle would be needed to conclusively assess the optimal FET protocol. Trial registration number N/A