Comparing frozen embryo transfer/FET protocols

The post discusses various frozen embryo transfer protocols, which includes a medicated FET, natural cycle FET (true natural/modified natural), and mild ovarian stimulated FET, along with their scheduling. In addition, it examines whether one FET protocol is superior to another, the relevance of the corpus luteum in medicated vs natural cycle FETs, and optimal progesterone levels.

Embryo transfers come in many different flavors. Fresh transfers can be done as part of a stimulated IVF cycle, however many people choose to do a frozen embryo transfer for its convenience or if there’s extra embryos available after a fresh transfer.

There’s different types of FET protocols, including:

• A medicated FET, also called a programmed FET, or hormone replacement therapy FET, or an artificial FET
• A natural cycle FET – which can include either a true natural cycle FET or a modified natural cycle FET.
• A mild ovarian stimulated FET (which is essentially a natural cycle FET with ovarian stimulation).
• A natural proliferative phase (NPP) FET, where the endometrium develops naturally, but progesterone is started before ovulation.

As we’ll see, a medicated FET doesn’t include the corpus luteum because ovulation doesn’t occur, whereas ovulation occurs in a natural cycle FET and a mild stimulated FET so the corpus luteum is present.

In this post I’ll go through the different types of FET protocols.

---

Overview of the menstrual cycle

Before getting into the different FET protocols, it is very important to go over the menstrual cycle so we can make sense of it all! You may be familiar with this graph that shows hormone levels during the menstrual cycle:

Cycle day 1 (CD1) to about day 5 is menstruation. During this period, GnRH acts on the pituitary gland to secrete LH/FSH. LH/FSH causes follicles in the ovaries to develop. These developing follicles begin to produce estrogen, or more specifically estradiol (E2), which causes the endometrium to enter the proliferative phase (CD6 to CD14). During this phase, the cells of the endometrium proliferate and thicken. Menstruation and the proliferative phase both occur while the follicle is developing, so CD1-CD14 is can be called the follicular phase.

E2 eventually peaks and this causes an LH surge to trigger ovulation. The leading follicle ovulates and the egg is released into a fallopian tube.

From CD15 to CD28 the endometrium is now in the secretory phase (aka the luteal phase). The empty follicle is now called the corpus luteum and begins secreting a variety of hormones such as progesterone (P4), so P4 levels begin to rise. This maintains the thickened endometrium and transforms its cells and architecture so it can accept an embryo for implantation. The embryo implants during this phase and begins secreting hCG that signals to the corpus luteum to keep producing progesterone (this continues for about 10 weeks). If there’s no implantation, then there’s no hCG produced and the corpus luteum eventually degrades. This causes progesterone levels to drop and triggers menstruation at which point the endometrium is shed and the cycle starts over.

In frozen embryo transfers, estrogen and progesterone may need to be administered to prepare the endometrium artificially. This is called “endometrial preparation”. Sometimes you might hear of “luteal phase support”, and this refers to administering progesterone to mimic the luteal phase as described above.

---

Medicated FET cycle protocol

The below diagram is showing cycle days 1-21 (which can vary) and a fresh ET as a reference (which isn’t showing progesterone administration for simplicity). I’m also including the natural cycle hormone levels to show how a medicated FET protocol mimics a natural cycle.

In a medicated FET, complete hormone replacement is needed because there is no follicle here to produce estradiol, and no corpus luteum to produce progesterone. These need to be administered exogenously (from outside the body)! Both estradiol and progesterone are taken until about 10 weeks of pregnancy, at which point the placenta starts to produce progesterone itself (taking over the job of the corpus luteum) in the so-called luteo-placenta shift.

Birth control is often taken before starting estradiol administration to better manage the timing of the medicated FET cycle.

Estradiol

In a medicated FET protocol, the endometrium is first prepared using estradiol (E2) which causes the cells of the endometrium to proliferate. Estradiol also prevents further follicles from growing because it blocks GnRH and prevents FSH/LH from being produced. In the past, a GnRH agonist would be used to prevent the LH surge, but estradiol alone is able to suppress this.

E2 can be administered either as a fixed dose (ie. 6 mg daily) or can be increased over time (ie. 2 mg/day for days 1-7, then 4 mg/day for days 8-12, then 6 mg/day from day 13 to transfer). Madero et al. (2016) compared these two methods and found no difference in live birth rates.

E2 can be administered by oral, vaginal, or transdermal routes, with oral E2 being the most common. There doesn’t seem to be any difference in reproductive outcomes (Glujovsky et al. 2010).

E2 is administered for around 2 weeks and then an ultrasound is performed to check for endometrial thickness (and to make sure a follicle isn’t developing!). In some cases, this E2 timing can be shorter (5-7 days) or longer (28 or as high as 36 days). Once the endometrium is 7 mm or greater, progesterone can start to be administered in combination with E2.

Progesterone

Since there is no ovulation (and therefore no corpus luteum!) in a medicated FET cycle, there is no progesterone secretion, so it needs to be administered exogenously (from outside the body).

There are a number of ways to do it, but most commonly is by vaginal suppository (ie. endometrin) or progesterone in oil (PIO) intramuscular shots. Some studies have suggested there is no difference, while a recent RCT found that intramuscular injections alone, or in combination with vaginal progesterone, is superior to vaginal progesterone alone (Devine et al. 2021; you can read my summary of that article here). More studies are needed.

For intramuscular P (PIO), 50 mg a day is common. For vaginal P, there are different forms/doses that can be used such as adhesive gels (1 dose x 90 mg daily), micronized tablets (3 x 100 mg), capsules (3 x 200 mg) or suppositories (2 x 400 mg). These are common doses but your specific dose and schedule are at the discretion of your doctor and can vary.

Once the endometrium is >7 mm (about 14 days after estradiol administration), progesterone is started. Once progesterone is administered, it’s concentration starts to build up in the body. It’s generally believed that progesterone needs to rise to a certain threshold before the endometrium transforms and becomes receptive.

How long to administer progesterone for depends on whether you’re having a Day 3 cleavage stage embryo transferred or a Day 5 blastocyst transferred. The timing of the endometrium and embryo should be synchronized. You can see in the graph above that a day 3 embryo is typically transferred on P+3 and a day 5 embryo on P+5.

Natural cycle FET protocol

For natural cycles, there are two types: a true natural cycle and a modified natural cycle. In a true natural cycle, it is very important to have regular menstrual cycles because the embryo is transferred based on when the LH surge occurs. In a modified natural cycle, a trigger shot is given to start ovulation, so there’s more flexibility.

The LH surge leads to ovulation (and the production of the corpus luteum) between 25-56 hours, and about 36-48 hours with hCG (Wan et al. 2020). Big ranges here and everyone is different! Below I’ve indicated ovulation as occurring 24 hours after the LH surge, and 48 hours after hCG for simplicity, but your specific timing will vary based on your cycle. Also note that the below diagram is showing cycle days 1-21 (which can vary) and a fresh ET as a reference (which isn’t showing progesterone administration for simplicity).

True natural cycle FET protocol

For a true natural cycle, an ultrasound is performed on CD2 or CD3 to ensure the corpus luteum from the previous cycle is gone. Cycles can be canceled if P4 is more than 1.5 ng/ml (5 nmol/L) on these days (which is true for fresh transfers as well).

Serum E2, LH and P4 is measured beginning on day 8-10 when the leading follicle is about 15 mm in size. Monitoring can be done daily or alternate daily at this point to predict exactly when ovulation will occur.

It’s not clear whether or not having a high P4 (>1.5 ng/ml or >5 nmol/L) is detrimental before the LH surge. One study found that 28.4% of women had elevated P4 on the day of the LH surge without any impact on ongoing pregnancy rates (Lee et al. 2014). The same study did find that women who had an elevated P4 for 2 days or more before the LH surge were less likely to become pregnant (39.4% vs 20.7%).

For fresh transfers, levels of P4 above 1.5 ng/ml (at time of trigger) are accepted as the threshold for cancellation (Wei et al. 2022). Elevated progesterone levels can prematurely initiate the luteal phase, potentially misaligning endometrial receptivity with the timing of the embryo transfer, which could decrease the chance of successful pregnancy.

The LH surge is usually used to pinpoint ovulation, but the exact definition varies between clinics. Some define it as an 80% rise compared to the last LH value, others use an absolute cut-off like ≥10–17 IU/L, sometimes with a drop in estrogen the next day. This lack of consistency means the timing of transfer can differ. P4 normally starts rising the day after the LH surge, confirming that ovulation has happened, and in some cases ovulation can be confirmed when P4 increases past 1.5 ng/ml (this happens a day after ovulation). Ultrasound signs like follicle collapse can also be used, although ovulation does not always occur at the exact same time after the surge.

Once the LH surge is detected, a Day 3 embryo is transferred on LH+4 and a Day 5 embryo on LH+6 (as shown above).

Modified natural cycle FET protocol

A modified natural cycle FET is similar to a true natural cycle except hCG is used to trigger ovulation once the leading follicle reaches 16-20 mm. The hCG trigger also causes progesterone levels to rise during the luteal phase, so this acts to some degree as luteal phase support. This happens because hCG signals to the corpus luteum to produce progesterone.

The main benefit of a modified natural cycle is that you don’t need to be so concerned with monitoring to catch the LH surge, since hCG is used to trigger ovulation. Because hCG is used, a Day 3 embryo can be transferred on hCG+5 and a Day 5 embryo on hCG+7.

Is luteal phase support needed during a natural cycle FET protocol?

Progesterone is produced naturally by the corpus luteum during a natural cycle, but is this enough?

In a true natural cycle, Bjuresten et al. (2010) found that administration of 400 mg of vaginal micronized P4 twice a day (starting on the evening of the FET) increased live birth rates (30% vs 20%, p=0.0272). Several other studies, as reported by Mumusoglu et al. (2021), have found that luteal phase support has no benefit for true natural cycle FETs. The same author also listed numerous studies that show no benefit for luteal phase support in a modified natural cycle FET.

However, many will still use progesterone for luteal phase support because there doesn’t seem to be any disadvantage (and if there isn’t enough progesterone produced by the corpus luteum then this can lead to a miscarriage!).

Mild ovarian stimulation FET protocol

The final type of FET protocol we’ll discuss is the mild ovarian stimulation FET. This can be used for patients who are anovulatory/dysovulatory (ie. PCOS) to induce ovulation. Many doctors use letrozole for the induction, so sometimes this is called a letrozole FET.

This type of protocol is essentially a natural cycle with mild ovarian stimulation to induce follicle development and ovulation. Some people will trigger ovulation with hCG, and others will simply let the LH surge do its job, so this varies! Also note that the below diagram is showing cycle days 1-21 (which can vary) and a fresh ET as a reference (which isn’t showing progesterone administration for simplicity).

A mild ovarian stimulation is achieved using clomid, letrozole, and/or exogenous gonadotropins starting on CD2 or CD3. Monitoring is done by ultrasound and/or serum measurements until the leading follicle is >17 mm, endometrial thickness is >7 mm, and serum E2 is >150 pg/ml (this can vary based on your clinic). At this point hCG can be used to trigger ovulation and Day 3 embryos are transferred on hCG+5 (or LH+5 if no hCG was used) and Day 5 embryos on hCG+7 (or LH+7).

Luteal phase support with progesterone is often used to ensure high progesterone levels, despite the corpus luteum being present, although there aren’t any RCTs that evaluate this.

Natural proliferative phase (NPP) FET protocol

This is a newer approach where the endometrium develops naturally during the first half of the cycle, but progesterone is started once the lining reaches an adequate thickness (typically ≥7 mm), without relying on ovulation for timing. Embryo transfer is then timed based on a set number of days of progesterone exposure (eg. P+5 for blastocysts), similar to a medicated FET. Ovulation is not triggered or used to schedule the transfer, which allows for more flexibility compared to a natural cycle (Godinho et al. 2024).

The idea is to combine the more natural hormonal environment of a natural cycle with the scheduling flexibility of a medicated cycle. Ovulation can still occur naturally, but it isn’t used to guide timing like in a natural FET.

Godinho et al. 2024 compared NPP, natural cycles, and medicated cycles, finding that live birth rates were higher with NPP (49.1%) and natural cycles (45.2%) compared to medicated cycles (38.4%). Miscarriage rates were also lower with NPP and natural cycles. A meta-analysis by Erden et al. (2026) of three studies also found a 25% higher relative live birth rate with NPP compared to medicated cycles, while outcomes were similar between NPP and natural cycles.

However, evidence for this protocol is limited and more studies are needed before this approach can be widely adopted.

Is one FET protocol better than another?

A recent 2020 Cochrane review combined the available data from 31 RCTs.

When comparing a medicated FET vs a natural cycle, they found:

• No significant difference in live births (4 RCTs)
• No significant difference in clinical pregnancies (5 RCTs)
• No significant difference in miscarriage rates (3 RCT)
• Significantly lower cancelation rate with natural FETs compared to medicated FETs (1 RCT, 25.6% vs 36.5%, odds ratio [95% CI]: 0.60 [0.44 – 0.82])

When comparing mild ovarian stimulated FET to a medicated FET, they found:

• No difference in live birth rates when stimulated using letrozole (1 RCT)
• No difference in clinical pregnancy rates when stimulated with FSH (1 RCT), or clomid (1 RCT), but there was an increase with letrozole (3 RCTs, 39.0% vs 25.0%, odds ratio [95% CI]: 1.24 – 3.04])
• No difference in miscarriage rates when stimulated with letrozole (2 RCTs), or clomid (1 RCT).

Another meta-analysis by Wu et al. (2021) looked at 26 RCTs and 113 observational studies (not RCTs) and found:

• From the observational studies, there were lower live birth rates with medicated FETs compared to true natural cycles (29.9% vs 36.5%, odds ratio [95% CI]: 0.81 [0.70 – 0.93]).
• From the observational studies, there were lower live birth rates with medicated FETs compared to modified natural cycles (29.2% vs. 35.7%, odds ratio [95% CI]: 0.85 [0.77 – 0.93]).
• From the RCTs, they found no difference between medicated and natural cycles.

Overall there may be a slight edge with natural cycles compared to medicated cycles. Mild ovarian stimulated FETs also seem to have a benefit over medicated FETs, although more studies need to be done.

Medicated FETs vs natural cycle FETs: does the presence of the corpus luteum matter?

A number of studies are coming out that are showing natural cycle FETs may lead to improved pregnancy and perinatal/obstetric outcomes.

This may be because the corpus luteum, which is produced in natural/mild stimulated FETs but not medicated FETs, produces a number of factors that aren’t replaced in medicated FETs, such as relaxin and VEGF. These additional factors might help implantation, sustain pregnancy, and reduce adverse pregnancy outcomes.

Takeshima et al. (2022) found improved pregnancy rates with natural FETs vs medicated FETs, along with a reduction in:

• Hypertensive disorders of pregnancy
• Placenta accreta
• Cesarean delivery rates
• Preterm delivery rates
• Other complications of pregnancy.

You can read the full summary in my post Worsened perinatal and obstetric outcomes for medicated FETs vs modified natural FETs.

Pape et al. (2022) found reduced miscarriage rates with natural and mild stimulated FETs compared to medicated FETs. They also found medicated FETs had lower live birth rates and increases in bleeding during the 1st trimester. You can check out the full summary in my post Pregnancy outcomes and complications with medicated, natural and mild stimulated FETs.

Godiwala et al. (2022) found reduced ongoing pregnancy and live birth rates with medicated FETs vs natural and letrozole (mild stimulated) FETs. Medicated FETs had higher miscarriage rates. You can check out the full summary in my post Comparing pregnancy outcomes for letrozole, natural and medicated FETs.

Zaat et al. (2023) did a meta-analysis and combined the results of 30 studies, finding that natural FETs led to a reduction in:

• Babies that were large for gestational age
• Macrosomia
• Babies with a low birthweight
• Early pregnancy loss
• Hypertensive disorders of pregnancy
• Preeclampsia
• Postpartum hemorrhage
• Preterm births
• Very preterm births

You can read the full summary in my post Meta-analysis compares medicated and natural FET adverse pregnancy and birth outcomes.

This is mostly based on retrospective studies so the quality of evidence isn’t the best. More RCTs need to be performed.

What level of progesterone is best for reproductive outcomes?

Serum progesterone levels that range between 8.75 to 32.50 ng/ml (measured around transfer day) seem to be optimal, although there is a lot of variation (Mumusoglu et al. 2021).

Having levels of progesterone that are too low can lead to miscarriage. Labarta et al (2021) found that women undergoing a medicated FET with serum progesterone <8.8 ng/ml on the day of embryo transfer had lower live birth rates (35.5% vs 52.0%) and higher miscarriage rates (13.5% vs 23%).

Other studies have found different minimum thresholds, and one study found no association with serum P below 10 ng/ml (Volovsky et al. 2020), however this study seems to be flawed because women with low P were supplemented with more progesterone which could have rescued the cycle.

A number of studies listed by Mumusoglu et al. (2021) show that cycles with low P can be rescued by simply administering more progesterone or by using a different route. However no RCTs have been done to evaluate this.

It’s not clear if having too high a P concentration can cause problems. Numerous studies indicated by Mumusoglu et al. (2021) have found no ceiling to serum progesterone levels, while others have found decreased success rates for levels above 20 ng/ml (Kofinas et al. 2015), 31.1 ng/ml (Yovich et al, 2015), or 32.5 ng/ml (Alyasin et al. 2021). These levels were typically measured on the day of embryo transfer (P+5 or P+6).

 

---

About Embryoman

Embryoman (Sean Lauber) is a former embryologist and the founder of Remembryo, an IVF research and fertility education website. After working in an IVF lab in the US, he returned to Canada and now focuses on making fertility research more accessible. He holds a Master’s in Immunology and launched Remembryo in 2018 to help patients and professionals make sense of IVF research. Sean shares weekly study updates on Facebook, Instagram, and Reddit regularly. He also answers questions on Reddit or in his private Facebook group.

---