P-753 Obstetric and neonatal risks in 7,378 singletons following fresh and frozen-thawed embryo transfers in Japan: An analysis utilizing doubly robust estimation

Abstract

Abstract Study question How do obstetric and neonatal risks differ between frozen-thawed embryo transfer (FET) and fresh embryo transfer (FreshET), as well as other non-assisted-reproductive-technology (non-ART) infertility treatments? Summary answer FET demonstrated a significantly higher incidence of obstetric complications, with a discernable but not statistically significant trend toward increased neonatal anomalies. What is known already Several studies indicate that FET is associated with an increased incidence of obstetric complications, encompassing gestational hypertension, placenta accreta, and gestational diabetes. Additionally, reports suggest that ART may elevate the risk of neonatal anomalies, although discerning whether this risk is more prevalent in FET or FreshET remains uncertain. Study design, size, duration This single-center, retrospective cohort analysis was conducted, encompassing 7,378 singletons delivered between January 2013 and December 2022 following fertility treatment within our institution. Within this cohort, cases conceived through controlled ovarian stimulation, timed intercourse, intrauterine insemination, or routine infertility checkups were categorized as the non-ART group. We compared the obstetric complications and neonatal outcomes between the FET and FreshET groups, using the non-ART group as the control. Participants/materials, setting, methods The evaluated outcomes included placenta accreta, placenta previa, postpartum hemorrhage, gestational hypertension, gestational diabetes, preterm birth, placental abruption, neonatal anomalies, high birth weight (≥4,000 g), and low birth weight (<2,500 g). For the statistical analysis, we employed inverse probability weighted regression adjustment (IPWRA). IPWRA integrates outcome regression with a model for exposure (propensity score), ensuring unbiased estimates even if one of the models is misspecified (a doubly robust property). Main results and the role of chance Among the 7,378 singleton births, 2,643 were in the non-ART group, 4,219 in the FET group, and 516 in the FreshET group. Comparative analysis with the non-ART group revealed that the FET group exhibited a higher risk across various obstetric complications, including placenta accreta (adjusted risk difference [ARD] 3.61%, 95% confidence interval [CI] 2.95 to 4.28), placenta previa (ARD 0.55%, 95% CI 0.14 to 0.96), postpartum hemorrhage (ARD 7.08%, 95% CI 6.03 to 8.13), gestational hypertension (ARD 3.57%, 95% CI 2.47 to 4.68), gestational diabetes (ARD 0.96%, 95% CI 0.17 to 1.75), preterm birth (ARD 2.13%, 95% CI 1.23 to 3.02), and high birth weight (ARD 0.97%, 95% CI 0.42 to 1.52). The analysis of neonatal anomalies initially showed a significant risk difference between the FET and non-ART groups in the crude analysis (crude risk difference 0.74%, 95% CI 0.09 to 1.38), although this significance diminished after adjustment (ARD 0.68%, 95% CI -0.02 to 1.39). The FreshET group did not demonstrate an increased risk of obstetric complications or neonatal anomalies. Limitations, reasons for caution The retrospective design of this study, conducted at a single institution, limits the generalizability of its findings to a broader population. Factors inherent to the specific facility, practices, and patient demographics may introduce variability in outcomes. Moreover, the potential for bias from unmeasured confounders cannot be entirely ruled out. Wider implications of the findings The mitigation of the risk of obstetric complications may be achieved through careful selection for FreshET, as opposed to a predominant inclination towards FET. It is crucial to develop a collaborative framework that enables the seamless exchange of information on maternal and neonatal prognoses between ART and perinatal facilities. Trial registration number not applicable