Natural vs. Programmed FET: What the Latest Lancet Data Reveals About Your Best Option

The Bottom Line

Live birth rates are close to identical between natural and programmed frozen embryo transfer (FET). The difference that is clearly established is pregnancy loss: endometrial preparation that involves ovulation carries less of it, shown with high-certainty evidence across 17 randomised trials. The widely repeated claim that natural cycles prevent hypertensive disorders of pregnancy has not been established.

This article previously stated that the 2024 Lancet trial showed natural FET to carry a lower risk of preeclampsia, placenta previa and preterm birth. That trial does not say that, and we are correcting it. The detail is below.

This doesn't mean natural FET is universally better — but it does add an important data point to the conversation with your fertility specialist.

Two Approaches to FET: What's the Difference?

Frozen embryo transfer involves thawing a previously frozen embryo and placing it into the uterus. The key decision: how to prepare the uterine lining.

Natural Cycle FET

Uses your body's own ovulation cycle to prepare the endometrium. Best suited for those with regular menstrual cycles.

Pros: Fewer medications, lower cost, more physiological
Cons: Requires precise ovulation monitoring, less scheduling flexibility

Programmed (Hormone Replacement) Cycle FET

Uses exogenous estrogen and progesterone to artificially prepare the endometrium. Works regardless of cycle regularity.

Pros: Predictable scheduling, works for irregular cycles, lower cancellation rates
Cons: More medications, potential side effects (nausea, headaches)

What the 2024 Lancet Trial Actually Found

Ho VNA, Vuong LN, et al. Lancet. 2024 (PMID: 38944045). An open-label randomised trial at a single hospital in Ho Chi Minh City, Vietnam, which allocated 1,428 ovulatory women aged 18 to 45 (1:1:1) to natural, modified natural, or artificial (programmed) endometrial preparation.

Live birth after one FET: natural 37% (174/476), modified natural 33% (159/476), artificial 34% (162/476). Relative risk 1.07 (95% CI 0.87 to 1.33) for natural versus artificial, and 0.98 (0.79 to 1.22) for modified natural versus artificial. No difference was demonstrated.
Maternal and neonatal complications: the paper's own wording is that these "did not differ significantly between groups, as the power to detect small differences was low." Its conclusion: "no definitive conclusions can be made regarding the comparative safety of the three approaches." It does not report lower preeclampsia, placenta previa or preterm birth in natural cycles.
Cycle cancellation: 99 of 476 first FET cycles were cancelled in the natural group and 99 of 476 in the modified natural group, versus none in the artificial cycle group. Roughly 21% against 0% — a clear, practical advantage for programmed cycles.

Level of evidence

The trial was open-label (unavoidable given the interventions, but it can influence subjective assessments) and was run at one hospital in Vietnam, so it does not automatically transfer to other countries, populations or health systems. Funding is listed as "None." On complications the authors themselves acknowledge being underpowered, so the correct reading is "we could not tell," not "there was no difference."

So where does the choice of cycle genuinely make a difference?

Pregnancy loss is lower when the cycle involves ovulation (high certainty)

A meta-analysis of 17 randomised controlled trials with 10,611 participants (Martins WP & Nastri CO, Ultrasound Obstet Gynecol. 2026. PMID: 42640814) gives the strongest answer available.

Pregnancy loss: risk ratio 0.75 (95% CI 0.65 to 0.86), high-certainty evidence — less loss with endometrial preparation that maintains a corpus luteum (natural, modified natural or stimulated cycles).
Clinical pregnancy rates were equivalent (RR 1.02, 95% CI 0.98 to 1.06; high certainty).
Live birth was comparable overall (RR 1.07, 95% CI 0.95 to 1.22; moderate certainty). But excluding studies at high risk of bias and one outlier trial with a very large between-group difference in cancellation rate, live birth was higher with ovulation (RR 1.15, 95% CI 1.01 to 1.31).
The authors' conclusion: clinicians should prioritise endometrial preparation with ovulation, while aiming for a low cycle-cancellation rate.

This article previously said early miscarriage was slightly higher in natural cycles. The high-certainty evidence points the other way, and we are correcting it.

For hypertensive disorders, there is a trend and nothing firmer

In the same meta-analysis, hypertensive disorders of pregnancy came out at risk ratio 0.77 (95% CI 0.55 to 1.07), moderate certainty. The interval crosses 1, so "a protective trend" is as far as it goes.

The large figures usually quoted come from a meta-analysis of observational studies (Busnelli A, et al. Hum Reprod. 2022. PMID: 35553678). Compared with natural cycles, programmed cycles showed hypertensive disorders at odds ratio 1.90 (95% CI 1.64 to 2.20; 12 studies; very low quality), preeclampsia 2.11 (1.87 to 2.39; 8 studies; low quality), postpartum haemorrhage 2.53 (2.19 to 2.93; low quality) and caesarean section 1.62 (very low quality).

Two points there cannot be skipped.

When only confounder-adjusted odds ratios were pooled, the associations that remained significant were hypertensive disorders, preeclampsia, postpartum haemorrhage and caesarean section — placenta previa (1.27) and preterm birth (1.19) did not survive adjustment. That is a second reason the earlier version of this article was wrong to name placenta previa and preterm birth.
These are all observational studies, with low to very low quality evidence. The authors state plainly that "because of the methodological weaknesses, recommendations for clinical practice cannot be made." Funding: None.

So the honest position is this: observational data keep showing a strong association; randomised trials stop at a trend. Association and causation have to be read separately here.

Why this is thought to happen (the corpus luteum hypothesis)

The suspicion about programmed cycles centres on the absence of the corpus luteum — the structure that forms in the ovary after ovulation.

In an ovulatory cycle the corpus luteum secretes not only progesterone and estradiol but also relaxin and vasoactive factors. The hypothesis is that these are involved not just in preparing the endometrium but in placental formation and uterine blood flow. In a programmed cycle that pathway is missing.

But this is a mechanistic hypothesis, not demonstrated causation. As above, randomised trials have not reached significance for hypertensive disorders. "A corpus luteum protects against preeclampsia" is not something that can be stated as fact; the lower rate of pregnancy loss is the finding that currently stands on firm ground.

Which One Is Right for You?

Preparation with ovulation has a high-certainty advantage on pregnancy loss. Even so, there is no one-size-fits-all answer — waiting for ovulation comes at the cost of roughly one in five first transfers being cancelled (99 of 476 in the Lancet trial, versus none in the programmed group), and the meta-analysis itself attaches the condition "while keeping cancellation low." Consider discussing these factors with your doctor:

Cycle regularity: Natural FET works best with regular cycles
Medical history: Previous pregnancy complications are worth weighing — though there is not yet evidence to say "you had preeclampsia, so choose a natural cycle"
Tolerance for cancellation: how much a cancelled first transfer would cost you in scheduling terms. Keep the one-in-five figure in mind
Scheduling needs: Programmed cycles offer more date flexibility
Medication preferences: Natural cycles require fewer drugs
Your clinic's expertise: Each clinic has its own protocols and experience

The most important step is an informed conversation with your fertility specialist, weighing these findings against your individual situation.

References

Martins WP, Nastri CO. "Endometrial preparation with vs without ovulation for frozen-embryo transfer: systematic review and meta-analysis of randomized controlled trials." Ultrasound Obstet Gynecol. 2026. PMID: 42640814
Busnelli A, et al. "Obstetric and perinatal outcomes following programmed compared to natural frozen-thawed embryo transfer cycles: a systematic review and meta-analysis." Hum Reprod. 2022. PMID: 35553678
Ho VNA, et al. "Livebirth rate after one frozen embryo transfer in ovulatory women starting with natural, modified natural, or artificial endometrial preparation in Viet Nam: an open-label randomised controlled trial." Lancet. 2024. PMID: 38944045

About the evidence in this article

The Lancet trial and the observational meta-analysis both record funding as "None"; no hormone manufacturer or industry involvement is disclosed. The authors of the randomised-trial meta-analysis are affiliated with a private reproductive medicine centre (SEMEAR Fertilidade) — please check the original for affiliations and conflict of interest disclosures.
Level of evidence: pregnancy loss and clinical pregnancy are high certainty. Live birth and hypertensive disorders are moderate. The large complication figures all come from observational studies rated low to very low, and some associations (placenta previa, preterm birth) did not survive adjustment for confounders. The Lancet trial was open-label and run at a single Vietnamese centre, and the authors acknowledge it was underpowered for complications.

FAQ

Q: Is natural FET always better than programmed FET? No. Live birth rates are similar, and clinical pregnancy rates are equivalent with high certainty. What ovulatory preparation does reduce, with high certainty, is pregnancy loss (RR 0.75). Programmed FET remains a well-established option, especially for irregular cycles, and it cancels far fewer cycles.

Q: Can I choose which type of FET to have? Ultimately, yes — but the decision should be made with your doctor based on your cycle regularity, medical history, and treatment goals.

Q: Does this mean I should avoid hormone replacement FET? Not at all. Programmed FET is well established and effective, and in the Lancet trial it cancelled none of 476 first transfers while natural cycles cancelled 99. The evidence favours ovulatory preparation on pregnancy loss; it does not condemn programmed cycles.

Q: I had preeclampsia before. Should I choose a natural cycle? There is not yet evidence to support that. Observational studies repeatedly show more hypertensive disorders with programmed cycles, but at low to very low quality, and the authors of that meta-analysis state that clinical recommendations cannot be made from it. Randomised trials reach only a trend (RR 0.77, 95% CI 0.55 to 1.07). If you have had preeclampsia, a plan for blood pressure monitoring during the pregnancy is firmer ground than the choice of cycle.

Written by the same author — a general guide to preconception care and fertility planning: A Doctor's Guide to Women's Health & Preconception (Amazon)