Effects of different cycle regimens for frozen embryo transfer on perinatal outcomes of singletons

STUDY QUESTION: Does the endometrial preparation protocol for frozen embryo transfer (FET) have an impact on perinatal outcomes? SUMMARY ANSWER: Singleton newborns from conceptions after an artificial FET cycle had a higher risk of being large for gestational age (LGA). WHAT IS KNOWN ALREADY: Most previous studies have concentrated on the clinical pregnancy, miscarriage and live birth rates of different endometrial preparation protocols for FET. However, the impacts of these cycle regimens on perinatal outcomes including birth-weight, gestational age (GA) and related outcomes require more investigation. STUDY DESIGN, SIZE, DURATION: We retrospectively analysed all singletons conceived by women who underwent non-donor FET cycles between July 2014 and July 2017. The propensity score matching (PSM) method using nearest neighbour matching at a proportion of 1:1 was established to adjust for factors that influence the probability of receiving different FET cycle regimens. The main outcomes of the study included birthweight and its related outcomes, Z-score, low birthweight (LBW, <2500 g), small for gestational age (SGA, <= 10th percentile of referential birthweight), LGA (>= 90th percentile of referential birthweight) and macrosomia (birthweight >4000 g). The study outcomes also included GA at birth, preterm delivery (<37 weeks), very preterm delivery (<32 weeks), very low birthweight (VLBW, <1500 g), term LBW (at 37 weeks of gestation or greater) and preterm LBW (at <37 weeks of gestation). PARTICIPANTS/MATERIALS, SETTING, METHODS: A total of 9267 cycles with live-born singletons were included in the analysis in our centre between July 2014 and July 2017. Of these, 2224, 4299 and 2744 live-born singletons were conceived by natural cycle FET, stimulated cycle FET and artificial cycle FET, respectively. After PSM, 1947 cycles of natural cycle FET versus stimulated cycle FET, 1682 cycles of stimulated cycle FET versus artificial cycle FET and 2333 cycles of natural cycle FET versus artificial cycle FET were included in the analysis. MAIN RESULTS AND THE ROLE OF CHANCE: A higher mean birthweight and Z-score were observed in the artificial cycle FET group than in the stimulated cycle FET group (P = 0.005; P = 0.004, respectively). Singleton newborns conceived after artificial cycle FET were more likely to be LGA than those born after natural cycle FET or stimulated cycle FET (19.92% versus 16.94% and 19.29% versus 16.12%, respectively). The adjusted ORs (95% CIs) were 1.25 (1.05, 1.49) for artificial cycle FET compared with natural cycle FET (P = 0.014) and 1.26 (1.08, 1.46) for artificial cycle FET compared with stimulated cycle FET (P = 0.003). Newborns conceived after stimulated cycle FET had a lower mean GA at birth and a lower mean birthweight than those born after natural cycle FET or artificial cycle FET. The stimulated cycle FET group had lower adjusted odds of being macrosomia than the natural cycle FET group. No significant differences between natural cycle FET and stimulated cycle FET were found for any of the other outcomes. LIMITATIONS, REASONS FOR CAUTION: This study had the disadvantage of being retrospective, and some cases were excluded due to missing data. The original allocation process was not randomized, which may have introduced bias. We have chosen not to account for multiple comparisons in our statistical analysis. WIDER IMPLICATIONS OF THE FINDINGS: LGA can have long-term consequences in terms of risk for disease, which means that the influences of artificial cycle FET are of clinical significance and deserve more attention. Furthermore, these findings are critical for clinicians to be able to make an informed decision when choosing an endometrial preparation method.