Birth weight in singletons after autologous fresh transfer according to the ovarian hyperstimulation protocol used

Abstract

OBJECTIVE: To evaluate whether the type of ovarian hyperstimulation protocol is associated with birth weight (BW) among singleton births from fresh autologous embryo transfer cycles. DESIGN: Cohort study. MATERIALS AND METHODS: The Society for Assisted Reproductive Medicine (SART) registry was used, including data collected from fresh autologous cycles that resulted in a singleton birth between years 2008- 2013. ANOVA and chi square tests were used as appropriate to compare BW categories and gestational age characteristics by protocol. Modified Poisson regression with robust error variance was used to estimate risk ratios (RR) and 95% confidence intervals for low birth weight (LBW), very low birth weight (VLBW) by ovarian hyperstimulation protocol (luteal agonist, agonist flare, and antagonist). Models were adjusted for age, body mass index, race, previous full term birth, previous preterm birth, infertility diagnosis, oocytes retrieved, embryos transferred, embryo stage, vanishing twin and infant gender. Interactions between protocol and infertility diagnosis were also explored. RESULTS: There were 54,041 births in the luteal agonist group, 10,943 in the agonist flare group and 47,886 in the antagonist group. There were no significant differences in BW, except in VLBW (BW less than1,500 g), which was 2% in the antagonist protocol group vs 1.8% in the luteal agonist and 1.6% in the agonist flare (P=0.0095). After adjusting for covariates, this difference remained significant (RR 1.16, 95% CI 1.02 to 1.31). A significant interaction was observed for LBW and VLBW between endometriosis and protocol. We found that among those with endometriosis, the agonist flare protocol was associated with a higher risk of LBW or VLBW compared to the luteal agonist protocol; for those without endometriosis there was no effect of the protocol. CONCLUSIONS: The association between BW and the type of protocol used does not appear to be clinically relevant, although a modest increase in VLBW among antagonist cycles suggest suboptimal placentation in some patients. Data from hormonal monitoring during the cycle -not available for analysis in the SART database- might be useful in understanding differences in obstetric outcomes of singleton births after fresh autologous cycles.