ImportanceImplantation failure remains a critical barrier to in vitro fertilization. Prednisone, as an immune-regulatory agent, is widely used to improve the probability of implantation and pregnancy, although the evidence for efficacy is inadequate.ObjectiveTo determine the efficacy of 10 mg of prednisone compared with placebo on live birth among women with recurrent implantation failure.Design, Setting, and ParticipantsA double-blind, placebo-controlled, randomized clinical trial conducted at 8 fertility centers in China. Eligible women who had a history of 2 or more unsuccessful embryo transfer cycles, were younger than 38 years when oocytes were retrieved, and were planning to undergo frozen-thawed embryo transfer with the availability of good-quality embryos were enrolled from November 2018 to August 2020 (final follow-up August 2021).InterventionsParticipants were randomized (1:1) to receive oral pills containing either 10 mg of prednisone (n = 357) or matching placebo (n = 358) once daily, from the day at which they started endometrial preparation for frozen-thawed embryo transfer through early pregnancy.Main Outcomes and MeasuresThe primary outcome was live birth, defined as the delivery of any number of neonates born at 28 or more weeks’ gestation with signs of life.ResultsAmong 715 women randomized (mean age, 32 years), 714 (99.9%) had data available on live birth outcomes and were included in the primary analysis. Live birth occurred among 37.8% of women (135 of 357) in the prednisone group vs 38.8% of women (139 of 358) in the placebo group (absolute difference, −1.0% [95% CI, −8.1% to 6.1%]; relative ratio [RR], 0.97 [95% CI, 0.81 to 1.17]; P = .78). The rates of biochemical pregnancy loss were 17.3% in the prednisone group and 9.9% in the placebo group (absolute difference, 7.5% [95% CI, 0.6% to 14.3%]; RR, 1.75 [95% CI, 1.03 to 2.99]; P = .04). Of those in the prednisone group, preterm delivery occurred among 11.8% and of those in the placebo group, 5.5% of pregnancies (absolute difference, 6.3% [95% CI, 0.2% to 12.4%]; RR, 2.14 [95% CI, 1.00 to 4.58]; P = .04). There were no statistically significant between-group differences in the rates of biochemical pregnancy, clinical pregnancy, implantation, neonatal complications, congenital anomalies, other adverse events, or mean birthweights.Conclusions and RelevanceAmong patients with recurrent implantation failure, treatment with prednisone did not improve live birth rate compared with placebo. Data suggested that the use of prednisone may increase the risk of preterm delivery and biochemical pregnancy loss. Our results challenge the value of prednisone use in clinical practice for the treatment of recurrent implantation failure.Trial RegistrationChinese Clinical Trial Registry Identifier: ChiCTR1800018783