Ovarian stimulation is the engine of modern in vitro fertilization. For decades, the standard approach has been to start injectable gonadotrophins early in the menstrual cycle, typically on day two or three, and to co-administer a GnRH antagonist partway through the stimulation to prevent a premature surge of luteinizing hormone (LH). Such a surge would otherwise trigger ovulation before the eggs could be retrieved, ruining the cycle. The antagonist acts as a chemical clamp on the pituitary gland, blocking the receptors through which GnRH from the hypothalamus would normally instruct the release of LH and follicle-stimulating hormone. It is an effective safeguard, but it adds injections, cost, and monitoring complexity to an already demanding process.

Late follicular phase stimulation emerged as an alternative concept in the context of freeze-all strategies, in which no fresh embryo transfer takes place. Instead, all embryos are cryopreserved and transferred in a later, unstimulated or artificially prepared cycle. Because the transfer is decoupled from the stimulation, clinicians have more flexibility in how the stimulation itself is conducted. In a late stimulation protocol, gonadotrophins are not started at the beginning of the cycle but later, after a dominant follicle has already been selected by the body’s own hormonal signals. The idea is to recruit the cohort of antral follicles that is naturally available in that cycle while working with, rather than against, the endocrine rhythm of the ovary.

Until now, however, even late stimulation protocols have relied on exogenous pituitary modulators, usually a GnRH antagonist added once stimulation begins, to guard against a premature LH surge. Whether the antagonist could be dropped altogether in late stimulation was an open question, and robust randomized evidence was lacking. The new trial, led by Xiuxian Zhu, Jingwen Lang, and colleagues at Shanghai First Maternity and Infant Hospital affiliated with Tongji University School of Medicine, was designed to answer precisely that question using a non-inferiority framework, the appropriate statistical design when the goal is to show that a simpler treatment is not meaningfully worse than the standard one.

The researchers enrolled 300 infertile women with normal ovulatory function and randomized them one-to-one between February 2022 and October 2023. Half received the conventional GnRH antagonist protocol, with stimulation initiated in the early follicular phase and antagonist co-administration. The other half received the antagonist-free late stimulation protocol, in which gonadotrophin injections began later in the follicular phase without any pituitary-blocking drug. All participants were undergoing IVF or intracytoplasmic sperm injection (ICSI) with a freeze-all strategy, meaning every viable embryo was frozen for later transfer. The trial was registered with the Chinese Clinical Trial Registry (ChiCTR2200055972) and approved by the hospital’s ethics committee, with all participants providing written informed consent.

The primary outcome was the number of usable embryos, a measure that captures the clinically meaningful yield of a stimulation cycle better than raw egg counts alone. In the late stimulation group, the median number of usable embryos was 3 (interquartile range 1 to 5), while in the antagonist group it was also 3 (interquartile range 1 to 4). The median difference between groups was 0, with a 95 percent confidence interval of minus 1 to 0. The prespecified non-inferiority margin was set at the lower bound of that interval, and because the lower bound of minus 1 met the margin, the investigators concluded that the antagonist-free late protocol was non-inferior to the antagonist regimen. In practical terms, women who skipped the antagonist did not end up with fewer embryos fit for freezing and later transfer.

Several secondary outcomes added nuance to this headline result. The total number of oocytes retrieved did not differ significantly between the two groups, but the late stimulation group produced significantly more mature oocytes, with a median of 7 (interquartile range 4 to 10.25) compared with 5.5 (interquartile range 3 to 8) in the antagonist group, a difference that reached statistical significance at P equals 0.018. Maturity matters because only mature oocytes, those that have completed the first meiotic division, can be fertilized. A protocol that yields a higher proportion of mature eggs may therefore translate the same number of retrieved eggs into more embryos, which is consistent with the comparable usable embryo counts observed despite any differences in retrieval totals.

Pregnancy outcomes from the subsequent frozen embryo transfers leaned in favor of the late stimulation protocol, though not all differences were conclusive. The implantation rate, the proportion of transferred embryos that successfully implanted, was significantly higher in the late stimulation group at 68.15 percent versus 54.86 percent in the antagonist group (P equals 0.013). The clinical pregnancy rate was also numerically higher in the late stimulation group, 66.01 percent versus 55.36 percent, but this difference fell just short of statistical significance at P equals 0.051. Early miscarriage rates were comparable between the groups, at 14.85 percent and 12.9 percent respectively (P equals 0.695), providing reassurance that the simplified protocol did not compromise early pregnancy outcomes.

The biological rationale for these findings is worth unpacking. In a natural cycle, rising estradiol from the developing dominant follicle feeds back on the hypothalamus and pituitary, and the late follicular phase is characterized by a delicate interplay in which follicle-stimulating hormone is restrained while LH remains controlled until the mid-cycle surge. Starting gonadotrophins late, when a dominant follicle has already been selected, may allow the remaining cohort of smaller follicles to be recruited under conditions that more closely mimic natural physiology. The higher maturity rate of the retrieved oocytes in the late stimulation group suggests that the follicles developed in a more synchronized hormonal environment, possibly because the endogenous hormonal milieu was not flattened by antagonist administration from the earliest days of the cycle.

The authors caution that their study was a single-center trial conducted in normo-ovulatory women, a population whose regular cycles make them suitable candidates for a protocol that depends on the timing of natural follicular development. Women with irregular ovulation, such as those with polycystic ovary syndrome, may not be appropriate candidates, since the protocol relies on the body’s own selection of a dominant follicle. The investigators themselves state that larger multicenter trials are warranted to confirm the findings before the protocol can be considered for broad clinical adoption. Still, the trial, funded by the hospital, the National Natural Science Foundation of China, and the Science and Technology Commission of Shanghai Municipality, represents the first robust randomized evidence that late follicular phase stimulation can succeed without any exogenous pituitary modulator at all.

For patients, the implications are tangible. Removing the GnRH antagonist from a stimulation cycle means fewer daily injections, lower medication costs, and a simpler monitoring schedule, all without a measurable loss of usable embryos in this study population. For the freeze-all paradigm, which has grown rapidly because it avoids the ovarian hyperstimulation syndrome risk and endometrial asynchrony associated with fresh transfers, the late stimulation protocol offers a physiologically elegant complement: stimulate later, work with the cycle rather than suppressing it, freeze everything, and transfer later. As assisted reproduction continues to move toward personalized, less burdensome protocols, this Shanghai trial provides a data point suggesting that sometimes the best way to control the ovary is to trust it a little more.

Subject of Research: GnRH antagonist-free late follicular phase ovarian stimulation for freeze-all IVF/ICSI in normo-ovulatory women

Article Title: GnRH antagonist‑free late follicular phase ovarian stimulation versus an early follicular GnRH antagonist regimen in normo-ovulatory women undergoing freeze-all IVF/ICSI: a randomized controlled trial

Article References: GnRH antagonist‑free late follicular phase ovarian stimulation versus an early follicular GnRH antagonist regimen in normo-ovulatory women undergoing freeze-all IVF/ICSI: a randomized controlled trial. (n.d.). https://doi.org/10.1186/s13048-026-02285-6

Image Credits: AI Generated

DOI: 10.1186/s13048-026-02285-6

Keywords: IVF, ICSI, ovarian stimulation, GnRH antagonist, late follicular phase, freeze-all strategy, randomized controlled trial, fertility treatment, oocyte maturity, implantation rate, reproductive medicine, frozen embryo transfer