P-669 Not the oocyte yield but the number of embryos predicts live birth at high response: analysis of 39,136 fresh IVF/ ICSI cycles from National Database

Abstract

Abstract Study question Oocyte or embryo yield- which is a better predictor of live birth (LB) in fresh cycle IVF/ ICSI? Summary answer The embryo yield is better than oocyte in predicting fresh-cycle LB and positively correlates with LB at high ovarian response when the oocytes yield doesn’t. What is known already As the LB rates plateau after a certain oocyte yield, the number of oocytes collected no longer can predict the live birth in IVF/ ICSI cycles with fresh embryo transfer.Since pregnancy outcomes also depend on the number of mature oocytes obtained, fertilisation rate and subsequent embryo development, it can be postulated that the number of total or transferrable embryos could be a better predictor of LB. Study design, size, duration Retrospective analysis of 39,136 fresh IVF/ ICSI cycles that met the inclusion criteria from the national database published by Human Fertilisation and Embryology Authority (HFEA) in the United Kingdom. We analysed 2 years of published data between January 2015 and December 2016. Participants/materials, setting, methods Only the first cycles of IVF/ ISCI that had oocyte retrieval procedure with intended fresh embryo transfer were included. Couples with any cause of infertility and female partners aged up to 44 years were included. Very few cycles with >25 oocytes had fresh transfer, therefore excluded. The ovarian response has been graded by 0-5, 6-10, 11-15, 16-20, 21-25 collected oocytes. Significant confounders found in univariate analysis were adjusted for in the multiple regression model. Main results and the role of chance Women’s age, method of insemination (IVF or ICSI) and number of transferred embryos and stage of transfer significantly associated with LB in univariate analysis (p < 0.0001 each). The oocyte yield predicted fresh-cycle LB only in low response group (0-5 oocytes), but not in higher response groups. Total number of embryos, the number of transferable embryos and embryo utilisation rate remained positively correlated with LB in all response groups. The number of transferable embryos found to be the best predictor of LB with the highest adjusted OR 1.51 (CI:1.4-1.65, p < 0.0001) in the lowest response (0-5 oocyte) group, that gradually declined to 1.06 (CI:1.03-1.10, p = 0.0003) in 21-25 oocytes group. Followed next by the total number of embryos with the highest adjusted OR of 1.20 (CI:1.13-1.27, p < 0.0001) in the lowest response (0-5 oocytes) group, down to 1.04 (CI:1.02-1.06, p = 0.0003) in the 21-25 oocytes group. The number of oocytes was the weakest in predicting LB (adjusted OR 1.07 (CI 1.02-1.12; p = 0.003) in the 0-5 oocytes group only. The inference remained the same when ICSI cycles with supposedly sperm factor was excluded. Limitations, reasons for caution Information on patient’s BMI or stimulation dose is not available in the HFEA database. However, it might have little impact in our findings, given large sample size. Ages are given in groups, so the groups taken as ordinal (dummy) variables in multiple regression analysis. Cumulative LB data were not available. Wider implications of the findings Having been influenced by both oocyte and sperm competency, the number of transferable embryos may have a stronger relation with the pregnancy outcomes than merely the oocyte number. On theory, this may be applicable for cumulative LB too, but awaiting confirmation by data evidence. Trial registration number Not applicable