Freeze-All vs Fresh Embryo Transfer in PCOS: What the Trials Show

Freeze-All vs Fresh Embryo Transfer in PCOS: What Do the Trials Actually Show?

The largest randomized trial ever run on this question in PCOS specifically found a higher live birth rate with freeze-all than with fresh transfer — 49.3% versus 42.0% — in 1,508 women with PCOS undergoing their first IVF cycle. The same trial found ovarian hyperstimulation syndrome (OHSS) more than five times less common with freeze-all, and pregnancy loss lower too. It also found preeclampsia more than three times more common in the freeze-all group — a result the freeze-all story often leaves out. This page reports what that trial, a second large randomized trial outside PCOS, and two independent meta-analyses found for live birth, OHSS, pregnancy loss and obstetric outcomes. It does not tell you which strategy to choose: that is an IVF protocol decision made between you and your clinic, based on your specific stimulation response, embryo number and endometrial findings, not something population-level trial data can settle for an individual case — what a fertility specialist actually does with that information is a separate, useful question from the outcome data itself.

“Fresh transfer” means an embryo is placed into the uterus three to five days after egg retrieval, in the same hormonally stimulated cycle. “Freeze-all” (also called elective cryopreservation, or a segmented cycle) means every embryo from that retrieval is frozen instead, and transfer happens in a separate, later cycle once stimulation hormones have returned to baseline. Both approaches use the same egg-retrieval and lab process; they differ only in what happens after the embryo exists.

You may see PCOS written as polyendocrine metabolic ovarian syndrome (PMOS), after a 2026 global consensus of more than 50 organisations renamed it. Every figure on this page applies under either name — only the label changed.

Why Does This Question Matter More in PCOS Than in Other IVF Cycles?

A prospective cohort of 290 women under 37 having their first IVF cycle found severe OHSS requiring hospitalisation in 15.4% of women with PCOS, against 2.7% in women with normal ovaries — one of the largest documented risk gaps in IVF, driven by a higher resting pool of small antral follicles that makes more follicles respond at once to stimulation. A wider comparison of PCOS egg counts and IVF outcomes covers that stimulation-response picture beyond OHSS alone. The 2023 international guideline states plainly that women with PCOS having IVF or ICSI should be counselled before treatment starts about this elevated OHSS risk, and offered options to reduce it — and it names a GnRH-antagonist protocol combined with an agonist trigger and freezing all embryos as one accepted risk-reduction option, not a required one. That guideline framing — an option to discuss, not a default — is the same report-only distinction this page keeps to throughout.

Most severe OHSS cases are driven by the hormone surge of an early pregnancy landing on top of an already-overstimulated ovary. Delaying transfer to a later, unstimulated cycle removes that overlap, which is the actual mechanism behind freeze-all’s OHSS advantage, not a general safety upgrade over fresh transfer in every situation.

The Head-to-Head Numbers in PCOS: Table 1

The 2016 trial randomly assigned 1,508 women with PCOS having their first IVF cycle to either fresh or frozen transfer of up to two embryos, cultured for three days, and followed every major outcome through delivery.

Every one of those numbers moved in the same trial, in the same women, which is what makes this the strongest single piece of evidence available for PCOS specifically: freeze-all did better on live birth, pregnancy loss and OHSS, and worse on preeclampsia. The five-versus-zero neonatal-death gap did not reach statistical significance, but the trial’s own authors reported it rather than omitting it, and this page does the same rather than rounding it to “no difference.”

Does This Same Pattern Show Up Outside PCOS-Specific Trials?

A 2019 randomized trial of 1,650 ovulatory women with a good IVF prognosis — not selected for PCOS — found an almost identical preeclampsia signal in a completely different population: 3.1% with frozen transfer versus 1.0% with fresh, a rate ratio of 3.13, essentially identical to the PCOS trial’s 3.12. That trial, run across 21 centres in China and transferring a single blastocyst-stage embryo, is the strongest evidence this page can offer that the trade-off is real rather than a one-trial fluke. Frozen transfer in that population also produced a higher singleton live-birth rate (50% versus 40%).

Two independent meta-analyses, pooling mixed and general infertility populations rather than PCOS specifically, found the same direction of effect for hypertensive disorders and a broader, genuinely mixed obstetric picture: some risks lower with frozen transfer, not only higher.

The Maheshwari review adds a detail that keeps this from reading as one-sided in either direction: alongside the higher hypertensive risk, singleton babies conceived from frozen embryos had a significantly lower relative risk of preterm delivery (0.90), low birth weight (0.72) and being small for gestational age (0.61) than those from fresh transfers, but a higher risk of large gestational age (1.54) and high birth weight (1.85). Freeze-all is not simply “riskier” or “safer” across every obstetric measure — it shifts the risk profile in more than one direction at once, and the honest picture includes all of it.

The Hypertensive-Disorder Signal: Where Freeze-All Comes Out Worse

Two randomized trials and two meta-analyses, covering PCOS-specific, good-prognosis and general infertility populations, all found the same direction of effect: preeclampsia or a broader hypertensive-disorders category was more common after frozen embryo transfer than after fresh transfer. That consistency across different study designs and different populations is the reason this finding belongs on the page rather than being left out because it complicates a simpler “freeze-all is better” story.

Researchers who have studied this pattern have proposed that the absence of a corpus luteum — the temporary hormone-producing structure an ovary forms after ovulation, which is bypassed in a fully hormone-programmed frozen cycle — may play a role, since the corpus luteum secretes relaxin and other factors thought to support normal blood-vessel adaptation in early pregnancy. That remains a proposed mechanism, not a confirmed one, and it has not been tested specifically within PCOS populations. Whether a given frozen cycle is hormone-programmed or built around a natural ovulation is a protocol detail your clinic decides based on your own cycle, not a variable this page can recommend for you.

So Is Freeze-All Better Than Fresh Transfer in PCOS?

One trial found four outcomes moving in freeze-all’s favour and one moving against it — a trade-off, not a single winner, and this page reports that trade-off rather than resolving it. In the PCOS-specific trial, freeze-all showed a higher live birth rate (49.3% vs 42.0%), a lower pregnancy-loss rate (22.0% vs 32.7%) and a substantially lower OHSS rate (1.3% vs 7.1%); it also showed a higher preeclampsia rate (4.4% vs 1.4%), a finding an independent trial outside PCOS reproduced almost exactly. Freeze-all additionally requires an extra step most fresh cycles do not — a delay of at least one full cycle between egg retrieval and transfer, while stimulation hormones return to baseline — which carries its own time and, depending on a clinic’s fee structure, cost implications worth asking about directly. Which of these factors matters most for a specific IVF cycle — a person’s individual OHSS risk, how many good-quality embryos were created, what the endometrium looks like on ultrasound, and how urgent timing is — is exactly the conversation a fertility specialist has with a patient armed with her own test results, not a conclusion population averages can reach on her behalf.

Who This Evidence Does Not Cover

• Neither major trial broke results down by PCOS phenotype. Whether the lean and insulin-resistant patterns of PCOS show the same trade-off is not something either randomized trial tested; both enrolled PCOS as a single diagnostic category.
• Both trials enrolled women on their first IVF cycle. Neither result has been established for a second or later cycle, or for women with a prior poor response to stimulation.
• The two trials used different embryo-culture timing. The PCOS trial transferred day-3 cleavage-stage embryos; the ovulatory-population trial transferred blastocysts — a genuine difference in practice, not a detail this page can resolve as equivalent.
• Both randomized trials were run in China, in specific clinical protocols and patient populations. That does not invalidate the findings, but it is a real limit on assuming identical numbers would appear in every country’s clinics.
• The corpus-luteum mechanism behind the preeclampsia signal is a hypothesis, not a demonstrated cause, and it has not been tested as a PCOS-specific question.

If IUI versus IVF, rather than which IVF transfer strategy, is the actual decision point right now, the separate IUI-versus-IVF decision for PCOS answers that question directly. Some clinics also offer a lower-stimulation alternative built specifically to avoid OHSS risk rather than manage it after the fact: in-vitro maturation as a lower-stimulation IVF alternative walks through how that option compares. When a transfer itself does not take despite a reasonable embryo, what implantation failure means in an IVF context covers endometrial receptivity and what scratch, intralipid and ERA testing have actually shown in trials. Beyond this one decision point, the full range of PCOS fertility topics this site covers includes testing, ovulation induction and the rest of the IVF pathway.