For frozen embryo transfer, a natural cycle matched hormones, with fewer complications

More than half of the embryos created by IVF are now frozen and transferred in a later cycle, and clinics have to decide how to ready the womb lining first. Two approaches dominate. In a programmed cycle, the lining is built entirely with oestrogen and progesterone tablets or patches, which makes scheduling easy but suppresses the woman's own ovulation, so no corpus luteum forms. In a natural-ovulation cycle, clinicians track the woman's own follicle and hormones and time the transfer to her ovulation. A large randomised trial has now tested whether that choice affects the baby and the mother [s1].

What the trial tested

The trial enrolled 4,376 ovulatory women aged 20 to 40 across 24 academic fertility centres in China, all planning to have a single thawed blastocyst transferred [s1]. They were randomly assigned 1:1 to a natural-ovulation regimen, with transfer timed by monitoring follicle growth and serum luteinising hormone, oestradiol and progesterone, or to a programmed regimen using sequential oestrogen and progesterone [s1]. It was a multicentre, assessor-blinded design, registered as ChiCTR2200057990 [s1].

Crucially, the trial set two co-primary outcomes rather than one: a healthy live birth, and pre-eclampsia or eclampsia [s1]. That pairing matters, because a regimen could raise the birth rate while quietly raising risk to the mother, and only measuring both reveals the trade-off.

What it found

On the headline fertility outcome, the two regimens were indistinguishable. A healthy live birth occurred in 910 of 2,185 women (41.6%) in the natural group and 890 of 2,191 (40.6%) in the programmed group, a relative ratio of 1.03 (95% confidence interval 0.96 to 1.10; P=0.49) [s1]. Prespecified per-protocol and subgroup analyses agreed [s1]. For a woman weighing the two options, the chance of taking home a baby was effectively the same.

The difference showed up in the mother's safety. Among women who reached a clinical pregnancy, pre-eclampsia struck 2.9% (38 of 1,302) in the natural group against 4.6% (61 of 1,326) in the programmed group, a relative ratio of 0.63 (0.43 to 0.94; P=0.02) [s1]. Early pregnancy loss was lower too, 12.1% versus 15.2% (0.80, 0.66 to 0.97) [s1]. So was placenta accreta spectrum, a dangerous abnormal attachment of the placenta, at 1.8% versus 3.6% (0.51, 0.31 to 0.83) [s1]. Caesarean section ran at 69.5% versus 75.6% (0.92, 0.87 to 0.97), and postpartum haemorrhage at 2.0% versus 6.1% (0.32, 0.20 to 0.52) [s1]. Birth weight and neonatal complications did not differ between the groups [s1].

None of this came entirely free. The natural regimen was cancelled more often, in 16.2% of cycles (354 of 2,185) against 11.5% (251 of 2,191) in the programmed group (P<0.001), because a natural cycle depends on the body ovulating on cue and sometimes it does not [s1].

Why the corpus luteum matters

The biological thread running through these results is the corpus luteum, the small hormone-producing structure left behind after ovulation. Natural cycles have one; fully programmed cycles do not. The corpus luteum secretes vasoactive hormones thought to help the placenta embed and the mother's blood vessels adapt to pregnancy, which is the leading explanation for why removing it, as a programmed cycle does, tracks with more pre-eclampsia and bleeding. This trial cannot prove that mechanism, but its pattern of harms fits it closely [s1].

The wider evidence base has been circling the same question. A 2025 Cochrane review, updating work first published in 2008 and 2017, set out to compare natural, hormone-therapy and ovulation-induction regimens for endometrial preparation on both effectiveness and safety [s2]. That the field needed a fresh randomised trial this large is a sign of how unsettled the comparison has been [s2].

How to read it

The clearest message is that, for women who ovulate regularly, choosing a natural-cycle transfer does not cost them a live birth and appears to lower several serious pregnancy complications [s1]. That is an unusual result: a less medicated option that is no worse on the outcome patients care about most and better on maternal safety.

The caveats are real. The trial was run in ovulatory women, so it says nothing about those with irregular or absent cycles, who often need a programmed approach [s1]. Natural cycles also carry a higher cancellation rate, which means more monitoring visits and, for some women, a delayed transfer [s1]. And the trial was conducted entirely in China, so practice patterns such as the high baseline caesarean rate may not transfer everywhere [s1].

Still, for a clinic and patient who can manage the extra monitoring, the balance of evidence now leans toward letting the body's own cycle prepare the womb where possible.

This article is informational and is not medical advice. Decisions about frozen embryo transfer should be made with a fertility specialist.

Sources

• [s1] Natural ovulation versus programmed regimens before frozen embryo transfer in ovulatory women: multicentre, randomised clinical trial. BMJ, 21 January 2026; 392:e087045. Chinese Clinical Trial Registry ChiCTR2200057990. https://doi.org/10.1136/bmj-2025-087045
• [s2] Cycle regimens for endometrial preparation prior to frozen embryo transfer. Cochrane Database of Systematic Reviews, 3 June 2025. https://doi.org/10.1002/14651858.CD003414.pub4