Abstract
PURPOSE: To examine the interaction between serum progesterone concentration on the trigger day and choice of freeze-all and fresh transfer strategies on live birth in an unselected population as well as in patients over 35 years old. METHODS: We performed a retrospective cohort study of 26,661 patients commencing their first IVF cycle in a large fertility centre between 2015 and 2019, including 4687 patients over 35 years old. We performed a multivariable fractional polynomial interaction analysis within a logistic regression model to investigate the interaction between serum progesterone concentration and the choice of freeze-all or fresh transfer strategy following the first transfer. RESULTS: 15,539 patients underwent a fresh embryo transfer and 11,122 underwent a freeze-all strategy in their first IVF cycle. The freeze-all group had a higher live birth rate compared to the fresh group (43.9% vs 40.3%). After adjusting for confounding factors, there was a positive interaction between serum progesterone concentrations and the choice of a freeze-all versus fresh embryo transfer on live birth (p for interaction 0.0001), with a larger magnitude of effect when progesterone concentration was higher. Such an interaction was also observed in patients over 35 years old (p for interaction 0.01), but the treatment effect curve over progesterone concentrations was almost flat. CONCLUSIONS: In an unselected population, frozen transfer is associated with greater chances of live birth, especially in patients with higher serum progesterone concentration. In patients over 35 years old, the benefit of a freeze-all policy appears small across all serum progesterone concentrations.