P–693 Gonadotropin stimulation reduces the implantation and live birth but not the miscarriage rate – a study based on the comparison of stimulated and unstimulated IVF

Abstract

Abstract Study question Does gonadotropin stimulation in conventional IVF (cIVF) affect the implantation, miscarriage and live birth rates? Summary answer Gonadotropin stimulation negatively affects the implantation and live birth but not the miscarriage rate in IVF treatments. What is known already Literature hypothesizes that embryos derived from unstimulated, natural cycle IVF (NC-IVF) have a higher implantation potential compared to embryos from cIVF. In NC-IVF, recruitment of the leading follicle is based on natural selection. Hormonal stimulation might not only affect the embryo but also endometrial function. It’s possible to compare outcomes of NC-IVF and cIVF if cIVF is performed without embryo selection, in other words, if only those zygotes, which will be transferred 1–2 days later, are left in culture and all other zygotes are cryopreserved. To test this hypothesis, we compared success rates in NC-IVF and in cIVF. Study design, size, duration We performed a cohort study from 2011–2016 including data on IVF cycles with transfer of fresh embryos on day 2–3 at a University based infertility center. Our sample consisted of 640 women with 1482 embryos transferred in 996 cycles. We defined implantation rate as the number of sonograhically detected amniotic sacs per transferred embryos. Data originated from the Swiss ART registry “FIVNAT” and the Bern IVF Cohort and was completed using medical and delivery records. Participants/materials, setting, methods We defined NC-IVF as IVF without stimulation of follicular growth and cIVF as IVF with gonadotropin stimulation ≥75 IE/d and >3 retrieved oocytes. We performed zygote, but not embryo selection and transferred embryos on day 2–3. We calculated implantation and live birth per transferred embryo as binary outcomes using bi- and multivariable multilevel logistic regression models accounting for two clusters; the women and the cycle; and adjusting for maternal and infertility characteristics using STATA. Main results and the role of chance Age of women (p = 0.531), parity (p = 0.194) and type of infertility (primary vs secondary) (p = 0.463) did not differ between women undergoing NC-IVF or cIVF. In NC-IVF, 468 (31.6%) embryos were transferred, 450 as single, 18 as double transfers. In cIVF, 1014 (68.4%) embryos were transferred, 91 as single, 830 as double and 93 as triple transfers. Implantation rate was higher in NC-IVF. In NC-IVF 80 (17.1%) and in cIVF 132 (13.0%) embryos developed into an amniotic sac (OR 1.58; 95% CI 1.01–2.46; p = 0.042). After adjustment for maternal age (p < 0.001), parity (p < 0.001), type of infertility (p = 0.037), duration of subfertility and indication for IVF, aOR for implantation per transferred embryo increased to 1.87 (95% CI 1.21–2.91; p = 0.005). Miscarriage rate was similar. In NC-IVF and cIVF 25% (n = 20; n = 33) miscarried and 75% (n = 60; n = 99) ended in a live birth, respectively (OR 0.91; 95% CI 0.32–2.60; p = 0.855; aOR 1.0; 95% CI 0.42–2.36; p = 1.000). Live birth rate per transferred embryo was increased in NC-IVF; 60 of 468 (12.8%) embryos in NC-IVF compared to 99 of 1041 (9.8%) embryos in cIVF resulted in a live birth (OR 1.51; 95% CI 0.92–2.49; p = 0.106); and became significantly higher after adjustment (aOR 1.85; 95% CI 1.16–2.95; p = 0.010). Limitations, reasons for caution This study analyses observational data from a clinic offering NC-IVF and cIVF treatment as equivalent options. NC-IVF is a model for natural fertility and allows us to study the impact of gonadotropins. However, it is not a randomised study and therefore prone to selection bias. Wider implications of the findings: The study suggests that gonadotropin stimulation might reduce the implantation potential and subsequently live birth rates, by possibly affecting embryo and endometrium quality. Clinicians should consider lower gonadotropin doses for stimulation. Trial registration number Not applicable