Is freeze-all better than fresh embryo transfer when IVF prognosis is low? A randomized trial and a separate observational study found results favoring a fresh-transfer strategy in the populations studied. The findings challenge routine freeze-all use, but do not override reasons to defer transfer or establish the best choice for every patient.
Freeze-all versus fresh transfer in low-prognosis IVF: weighing live birth evidence and reasons to defer transfer
- Re-evaluating the ‘freeze-all’ dogma in women with low oocyte yields or diminished ovarian reserve
- Multicenter RCT of 838 low-prognosis patients: Live birth rate was 32% for freeze-all vs. 40% for fresh transfer; 1-year cumulative rate reached 44% vs. 51%
- Freezing is not a universally superior strategy: Protocol selection must be individualized based on ovarian response, endometrial environment, and OHSS risk
“We only got a couple of eggs—do we really need to freeze all the embryos?”
This is a recurring question among women undergoing In Vitro Fertilization (IVF). Online patient communities frequently reinforce the belief that “frozen embryo transfer (FET) always yields higher success rates than a fresh transfer” and that “it is always better to let the uterus rest for a month or two after egg retrieval.”
In clinical practice, the “freeze-all” strategy—cryopreserving the entire cohort of embryos during the retrieval cycle and postponing transfer to a subsequent unmedicated or programmed cycle—has indeed become widespread.
Yet, does applying a universal “freeze-all” policy to a woman who retrieved only two or three eggs offer the same clinical advantages as it does for a woman who retrieved twenty? Emerging research indicates that for “low-prognosis” patients with low ovarian reserve or few retrieved eggs, the clinical equation changes dramatically.
Freeze-all versus fresh transfer: two different study designs
A study published online on September 3, 2026, in Reproductive BioMedicine Online compared freeze-all with a fresh-transfer strategy using target trial emulation. It retrospectively analyzed 2,102 women undergoing a first or second IVF/ICSI cycle: 912 in the freeze-all group and 1,190 in the fresh group. Weighting balanced 20 measured baseline variables, but did not actually randomize treatment. Weighted cumulative live birth per initiated cycle was 40.0% versus 46.7%, an association subject to residual confounding.
The premise of the study is straightforward: questioning whether the documented benefits of the freeze-all strategy apply equally across all patient phenotypes. In particular, the authors sought to determine whether freezing the entire embryo cohort and delaying transfer offers any tangible advantage to women who have very few embryos available from the start.
One rationale for freeze-all is to avoid transferring during a stimulated cycle when hormone levels may affect endometrial receptivity. It is also an important OHSS risk-reduction strategy in high responders. A GnRH agonist trigger is given before egg retrieval; embryos can subsequently be frozen instead of transferred fresh. ASRM guidance supports these measures to reduce moderate-to-severe OHSS, without guaranteeing that all OHSS is eliminated.
Low-prognosis patients may have fewer oocytes or reduced ovarian reserve, but their hormone levels and endometrial findings still vary. Whether routinely deferring transfer improves live birth in this group must be tested clinically, rather than assumed from a low egg count.
Randomized trial: live birth after fresh and frozen transfer
Robust randomized evidence addressing this question came from a large-scale, multicenter randomized controlled trial conducted across nine reproductive medicine centers in China, published in The BMJ. The investigators randomized 838 women with a low prognosis for IVF into either a freeze-all group or a fresh embryo transfer group. Study eligibility included women with 9 or fewer retrieved oocytes, fewer than 5 antral follicles, or low baseline AMH levels.
The trial assigned 419 women to a freeze-all strategy and 419 to a fresh-transfer strategy. Outcomes were analyzed by randomized group, including participants who crossed over or did not proceed to transfer within the follow-up period.
The trial’s findings diverged sharply from the popular assumption that “frozen is always superior”:
- Live Birth Rate: 32% in the freeze-all group vs. 40% in the fresh transfer group
- Clinical Pregnancy Rate: 39% in the freeze-all group vs. 47% in the fresh transfer group
The reported 1-year cumulative live birth rate was 51% in the fresh group versus 44% in the freeze-all group. It counted outcomes from transfers occurring within one year after randomization, rather than requiring delivery within that year. A table footnote states that 11 ongoing pregnancies in the fresh group and 15 in the frozen group were counted as live births; these were not all confirmed deliveries at reporting.
For low-prognosis women, the rule that “freezing everything and letting the uterus rest automatically improves outcomes” did not hold up.
Possible explanations and limits of the evidence
These findings do not show that freezing damaged the embryos. Frozen embryo transfer remains an established treatment. Survival after warming varies with embryo stage and laboratory practice, so a universal survival rate above 95% should not be assumed from these studies.
Ovarian response and the endometrial environment are relevant considerations, but these studies did not identify one proven mechanism explaining the difference.
Higher steroid levels during stimulation may affect the endometrium, and high responders have greater OHSS risk. Deferring transfer can be clinically useful in selected circumstances; this does not show that every normal or high responder benefits from waiting for an undefined hormonal reset.
The BMJ authors proposed that lower steroid exposure and possible freezing-related effects might help explain their findings, but the mechanism remains uncertain. With only a few embryos, the timing, cost and burden of a deferred transfer matter. The trial did not establish that every low responder has a more receptive endometrium or receives no biological benefit from freezing.
Ultimately, evaluating assisted reproduction as a simplistic contest of “fresh vs. frozen” misses the mark; the real priority is matching the right strategy to the right patient.
The randomized trial was pragmatic: embryo stage, embryo number and transfer preparation were not standardized, and crossover or delayed treatment may affect interpretation. It was not powered to resolve uncommon maternal or neonatal risks. Its participants were suitable for fresh transfer; the result should not be extended to patients for whom fresh transfer was contraindicated.
When fresh transfer may still need to be deferred
These findings should not be oversimplified into a mandate that every low-prognosis patient must undergo a fresh transfer.
Some clinical findings may warrant assessment or deferral before transfer. These are not automatic freeze-all rules, and freezing itself does not treat a uterine or tubal condition:
- Progesterone elevation: A high level on trigger day can inform counseling, but 1.5 ng/mL is not a universal cutoff requiring freeze-all. ESHRE guidance also considers oocyte number, embryo number and embryo quality.
- Endometrial findings: Thickness below 7 mm, cavity fluid, polyps or suspected chronic endometritis require context-specific assessment. A thickness measurement alone does not establish that frozen transfer is preferable or that a fresh transfer cannot succeed.
- Hydrosalpinx: An untreated communicating hydrosalpinx can impair IVF outcomes and requires specialist assessment before transfer. ASRM guidance discusses tubal treatment; simply freezing embryos does not correct it.
- Preimplantation genetic testing: PGT-A or PGT-SR may involve freezing embryos while awaiting results. The transfer schedule depends on biopsy timing and laboratory testing; cryopreservation is not a universal requirement of every PGT workflow.
The central takeaway from recent trials is that the fertility field must move away from assuming that “frozen transfers are universally superior for every patient.”
The optimal embryo transfer protocol must be individualized by weighing baseline ovarian reserve, maternal age, retrieved oocyte numbers, embryo cohort size, endometrial pattern, late follicular hormone levels, and the potential need for genetic screening.
For low-prognosis patients who retrieve only a small cohort of eggs, the clinical question should shift: rather than asking “Is frozen transfer better?”, patients and clinicians should ask “Is there a clear, documented medical reason to freeze all of my embryos in this specific cycle?”
When only one or two embryos are available, the decision should weigh the chance of live birth, treatment safety, time and patient preferences. These results do not establish that one strategy guarantees a healthy baby.
If genetic testing is part of the transfer plan, see our review of PGT-A and cumulative live birth outcomes. Screening and the decision to freeze embryos answer different clinical questions.
※ Sources: Zhang and colleagues, Freeze-all versus fresh embryo transfer strategies in low-prognosis women: an emulated target trial study, Reproductive BioMedicine Online, published online September 3, 2026 (2,102 women); and Wei and colleagues, Frozen versus fresh embryo transfer in women with low prognosis for in vitro fertilisation treatment: pragmatic, multicentre, randomised controlled trial, The BMJ, January 29, 2025 (838 women, nine Chinese fertility centres; registration ChiCTR2100050168). This article does not replace individualized clinical diagnosis or medical care. Treatment decisions should be made with a qualified reproductive specialist.
※ The images associated with this article were generated using generative AI (ChatGPT, OpenAI) as illustrative visual references and do not depict real individuals.
※ Copyright © 2026 THE FERTILITY NOTES. All rights reserved. Unauthorized reproduction, duplication, distribution, scraping, AI training, or data utilization of all content, including articles, photos, and images, is strictly prohibited and may result in legal liability.