IVF Reads / Progesterone Support in IVF and IUI: Who Needs It, How Long
Progesterone Support in IVF and IUI: Who Needs It, How Long

Progesterone for luteal phase support is recommended after fresh IVF/ICSI by the ESHRE 2025 ovarian stimulation guideline (a strong recommendation on very-low-quality evidence), and the Cochrane review of luteal phase support (van der Linden 2015) found higher live birth or ongoing pregnancy with progesterone than placebo (OR 1.77, 95% CI 1.09 to 2.86; 5 RCTs, 642 women). It has not been shown to raise live births after IUI or timed intercourse with ovarian stimulation (Cochrane 2022, RR 1.10, 95% CI 0.81 to 1.48; 7 RCTs, 1,792 women). In the PROMISE trial (836 women with unexplained recurrent miscarriage) live birth was 65.8% with progesterone versus 63.3% with placebo, and in PRISM (4,153 women with early-pregnancy bleeding) 75% versus 72%; neither difference was statistically significant. ESHRE's 2023 recurrent pregnancy loss guideline says vaginal progesterone does not improve live birth in unexplained recurrent loss, but may in women with three or more losses and bleeding in the next pregnancy (both conditional recommendations).
- ESHRE ovarian stimulation guideline, 2025 update (Hum Reprod 2026): 'Progesterone is recommended for luteal phase support after IVF/ICSI' (Strong, very-low-quality evidence); any non-oral route of natural progesterone can be used (good practice point); oral dydrogesterone 'is probably recommended' (Conditional, moderate-quality evidence) with a note that reported associations with congenital malformations cannot be read as causal; progesterone should continue until at least the day of the pregnancy test (good practice point). NICE NG257 (March 2026) refers UK clinicians to this ESHRE section for luteal support.
- Cochrane, van der Linden 2015 (94 RCTs, 26,198 women): progesterone vs placebo, live birth or ongoing pregnancy OR 1.77 (95% CI 1.09 to 2.86; 5 RCTs, 642 women, very-low-quality evidence); 'the route of progesterone administration is not associated with an improvement in outcomes'; hCG as luteal support raised ovarian hyperstimulation syndrome (OR 4.28, 95% CI 1.91 to 9.6; 1 RCT).
- LOTUS I (Tournaye 2017, double-blind, 1,031 women, 38 sites, Abbott-sponsored): pregnancy at 12 weeks 37.6% on oral dydrogesterone vs 33.1% on vaginal progesterone (difference 4.7%, 95% CI -1.2 to 10.6); live birth 34.6% vs 29.8%. LOTUS II (Griesinger 2018, open-label, 1,034 women): 38.7% (191/494) vs 35.0% (171/489); live birth 34.4% vs 32.5%. Both trials were fresh IVF cycles.
- Stopping at the positive test vs continuing (Watters 2020 meta-analysis, 7 RCTs, 1,627 women, fresh cycles): live birth RR 0.94 (95% CI 0.88 to 1.00), miscarriage RR 0.91 (0.69 to 1.20), ongoing pregnancy RR 0.98 (0.91 to 1.05).
- Miscarriage: PROMISE (NEJM 2015) 65.8% (262/398) vs 63.3% (271/428) live birth, RR 1.04 (0.94 to 1.15); PRISM (NEJM 2019) 75% (1,513/2,025) vs 72% (1,459/2,013), RR 1.03 (1.00 to 1.07); Cochrane 2025 (Haas, recurrent miscarriage, 8 trials, 1,276 women) miscarriage RR 0.91 (0.76 to 1.07), moderate certainty. ASRM 2026 luteal phase deficiency opinion: 'there is no evidence that progesterone is beneficial for the treatment of LPD'.
- Frozen transfer (Erden 2026, Hum Reprod Update): vaginal progesterone raised live birth in true natural-cycle FET (RR 1.43, 95% CI 1.16 to 1.78; 923 cycles, 2 RCTs, moderate certainty) but not in hCG-triggered modified natural cycles (RR 1.04, 0.82 to 1.32; 667 cycles). Labarta 2021 (1,150 programmed cycles on vaginal progesterone): serum progesterone under 8.8 ng/mL on transfer day, ongoing pregnancy 36.6% vs 54.4%; observational.
Do I need progesterone after IVF, IUI or a frozen transfer?
After a fresh IVF or ICSI cycle, yes: the ESHRE ovarian stimulation guideline (2025 update) makes progesterone for luteal phase support a strong recommendation, and NICE's fertility guideline NG257 (March 2026) points UK clinicians to that same section. After IUI or timed intercourse with ovarian stimulation, the trials have not shown it raises live births. After a frozen transfer, it depends on how the cycle was prepared.
Fresh IVF is different because of the stimulation itself. The Cochrane review of luteal phase support (van der Linden 2015) explains that after ovarian stimulation and egg collection the body's own progesterone and hCG signals run low, so the luteal phase is propped up from outside. Across five placebo-controlled trials in 642 women, progesterone gave higher live birth or ongoing pregnancy than placebo (OR 1.77, 95% CI 1.09 to 2.86). That evidence is graded very low quality: the placebo trials are old and small, vaginal progesterone 'is routinely used in most clinics' (the LOTUS I investigators' words), and a new placebo trial is unlikely to be run now.
After IUI the picture is different. The Cochrane review of luteal support for IUI and timed intercourse (Salang 2022, 25 RCTs, 5,111 women) found vaginal progesterone did not clearly change live birth or ongoing pregnancy against placebo or no treatment (RR 1.10, 95% CI 0.81 to 1.48; 7 RCTs, 1,792 women, low certainty). Clinical pregnancy rates were somewhat higher with progesterone, but the outcome that matters did not follow. If your clinic prescribes it after IUI, that is a reasonable thing to ask about, not a reason to doubt the clinic. The stimulation drugs themselves are covered in ovulation induction: process and success rates.
Frozen transfers split three ways. In a programmed (hormone-replacement) cycle there is no corpus luteum at all, so the progesterone you take is the whole luteal phase, not a supplement to it. In a true natural cycle, where your own ovulation is tracked, a 2026 systematic review in Human Reproduction Update (Erden) found vaginal progesterone raised live birth (RR 1.43, 95% CI 1.16 to 1.78; 923 cycles, two RCTs, moderate certainty). In a modified natural cycle, where an hCG injection triggers ovulation, the same review found no benefit (RR 1.04, 0.82 to 1.32; 667 cycles, two RCTs, moderate certainty), a finding a 2026 meta-analysis in Reproductive BioMedicine Online (Engesgaard, 3,896 cycles) repeated.
Which form: vaginal, injection or oral dydrogesterone?
No route has been shown to produce more babies than another. The Cochrane review's summary line is that 'the route of progesterone administration is not associated with an improvement in outcomes'. Its largest route comparison, intramuscular versus vaginal or rectal, pooled seven trials in 2,309 women and leaned slightly towards injections (OR 1.24, 95% CI 1.03 to 1.5), but the trials disagreed with each other badly and the evidence was graded very low quality. Subcutaneous versus vaginal gel (two RCTs, 1,465 women) and vaginal versus oral micronised progesterone (four RCTs, 857 women) showed no difference.
ESHRE 2025 therefore says any of the non-oral routes for natural progesterone can be used (a good practice point). Vaginal preparations are the usual default; the LOTUS investigators list their downsides as vaginal irritation, discharge and poor adherence. In the one IUI trial that reported side effects by route, the women randomised to intramuscular injections had far more of them than the vaginal arm (RR 9.25, 95% CI 2.21 to 38.78; 225 women, very low certainty).
Oral dydrogesterone is the exception to 'non-oral'. ESHRE 2025 says it 'is probably recommended' for luteal phase support, a conditional recommendation on moderate-quality evidence, and adds that reports linking dydrogesterone to congenital malformations 'cannot be translated into a conclusion on causality' and are treated as potential associations. The evidence is two manufacturer-sponsored trials. LOTUS I (Tournaye 2017) was double-blind across 38 sites and randomised 1,031 women having fresh IVF to oral dydrogesterone or micronised vaginal progesterone: pregnancy with a fetal heartbeat at 12 weeks was 37.6% versus 33.1% (difference 4.7%, 95% CI -1.2 to 10.6), live birth 34.6% versus 29.8%, a non-inferiority result. LOTUS II (Griesinger 2018) was open-label at 37 centres in 10 countries including India, 1,034 women: 38.7% (191/494) versus 35.0% (171/489) pregnant at 12 weeks, live birth 34.4% versus 32.5%. Both were fresh cycles; the LOTUS II authors say programmed frozen cycles, where there is no corpus luteum, need their own trials.
How long do I keep taking it?
The guideline floor is clear and the ceiling is not. ESHRE 2025 says to start between the evening of egg collection and the third day after it, and to continue 'until at least the day of the pregnancy test'; both are good practice points, not evidence-graded recommendations. Beyond that, practice ranges from stopping at the positive test to continuing to 12 weeks, which is what the LOTUS trials did.
The trials that compared the two found no difference. A 2020 meta-analysis in Reproductive BioMedicine Online (Watters) pooled seven RCTs in 1,627 women after fresh transfer who were randomised to stop progesterone at the positive test or continue: live birth RR 0.94 (95% CI 0.88 to 1.00, three trials), miscarriage RR 0.91 (0.69 to 1.20), ongoing pregnancy RR 0.98 (0.91 to 1.05). The authors conclude prolonged support 'may be unnecessary' after fresh transfer and ask for a larger trial. Two limits: the lower bound for live birth touches 1.00, so a small loss is not excluded; and every trial was a fresh cycle, where a corpus luteum exists to take over. None of it applies to a programmed frozen cycle, where there is no corpus luteum and progesterone continues until the placenta produces its own.
Whatever the schedule, do not stop on your own because of spotting or because symptoms faded. Bleeding on progesterone is common and has its own page: bleeding after a positive test. And progesterone's own effects (breast tenderness, bloating, tiredness) are the reason symptoms after transfer tell you nothing about whether it worked.
Want your own results explained?
IVY reads your protocol, your transfer type and your clinic's progesterone plan against the guidelines quoted here, and tells you which questions are worth taking back to your doctor.
Does progesterone prevent miscarriage?
For most women, the two largest trials ever run say no. PROMISE (NEJM 2015) randomised 836 women with unexplained recurrent miscarriage who had conceived naturally to vaginal progesterone or placebo from a positive test to 12 weeks: live birth was 65.8% (262/398) with progesterone and 63.3% (271/428) with placebo, RR 1.04 (95% CI 0.94 to 1.15). PRISM (NEJM 2019) randomised 4,153 women with bleeding in early pregnancy at 48 UK hospitals: live birth after 34 weeks was 75% (1,513/2,025) versus 72% (1,459/2,013), RR 1.03 (1.00 to 1.07), P = 0.08.
The Cochrane network meta-analysis that followed (Devall 2021, seven trials, 5,682 women) rated the threatened-miscarriage result high certainty: vaginal progesterone 'probably makes little or no difference' to live birth (RR 1.03, 1.00 to 1.07). The 2025 Cochrane review of recurrent miscarriage of unclear cause (Haas, eight trials, 1,276 women) found miscarriage RR 0.91 (0.76 to 1.07) and live birth RR 1.04 (0.96 to 1.12), both moderate certainty, and notes that two trials in its earlier versions have since been retracted. It excluded IVF pregnancies, so it says nothing about luteal support.
There is one group where the signal is real. When the Cochrane 2021 analysis split the bleeding trials by history, women with previous miscarriages did better on progesterone and women with none did not. ESHRE's 2023 recurrent pregnancy loss guideline reads the pooled PROMISE and PRISM data as a live-birth RR of 1.28 (95% CI 1.08 to 1.51) in women with three or more previous losses who bleed in the next pregnancy, and makes two conditional recommendations on moderate-quality evidence: vaginal progesterone 'does not improve live birth rates in women with unexplained RPL', and it 'may improve live birth rate in women with three or more pregnancy losses and vaginal blood loss in a subsequent pregnancy'. It also finds insufficient evidence for progesterone given for 'luteal phase insufficiency' in recurrent loss, and recommends against testing for it. If you are reading this after a loss, fertility after miscarriage covers what the next cycle looks like.
Will progesterone fix a 'luteal phase defect' if I am trying naturally?
The American Society for Reproductive Medicine's committee opinion on luteal phase deficiency, reissued in 2026, is blunt: LPD 'has not been proven to be an independent entity causing infertility or recurrent pregnancy loss', there is 'no reproducible and clinically practical standard' to diagnose it, 'there are no RCTs investigating progesterone supplementation for women with LPD', and 'there is no evidence that progesterone is beneficial for fertility in natural cycles'. Short luteal phases (under 10 days) turn up in 9 to 13% of ovulatory cycles in the cohorts of normally menstruating women it cites, and a single progesterone reading can swing eight-fold within 90 minutes.
Its advice is to look for an underlying cause (thyroid, prolactin, hypothalamic disorders) and treat that; if none is found, empiric treatment 'generally is not recommended'. What a luteal progesterone blood test can and cannot tell you is covered in low progesterone: what the test can and cannot show.
What if my progesterone level is low on transfer day?
This question only has data for programmed frozen cycles on vaginal progesterone, and the data are observational. Labarta and colleagues in Valencia measured serum progesterone immediately before transfer in 1,150 women (own eggs with or without PGT-A, and donor eggs): the lowest 30% of readings, below 8.8 ng/mL, carried an ongoing pregnancy rate of 36.6% against 54.4% for everyone else, and live birth 35.5% against 52.0%. Heavier body weight was the one independent predictor of a low reading.
What this establishes is that the same vaginal regimen produces very different blood levels in different women. What it does not establish, from anything cited here, is that adding more progesterone when the level is low restores the pregnancy rate; the authors recommend checking and adjusting, but that is an inference from a cohort, not a randomised result. A low reading in a fresh or natural cycle means something different again, because a corpus luteum is present and its output is pulsatile. If your clinic checks levels on transfer day, ask what cut-off they use and what they would change if you fell below it.
What the evidence does not establish
- That one route or product gives more live births than another. Cochrane 2015 found no route advantage; the LOTUS trials showed dydrogesterone was not inferior to vaginal progesterone, not that it was better.
- That progesterone after IUI raises live births. Cochrane 2022: RR 1.10 (0.81 to 1.48), low certainty.
- That progesterone prevents miscarriage in women without a history of loss. PRISM and the Cochrane 2021 network meta-analysis: little or no difference, high certainty.
- That continuing past the positive test adds anything in a fresh cycle. Seven RCTs found no difference, but the live-birth confidence interval touches 1.00 and the reviewers asked for a bigger trial.
- That topping up a low transfer-day level improves outcomes. The Labarta cohorts show association, not the effect of intervening.
- That 'luteal phase defect' is a diagnosis that progesterone treats. ASRM 2026: no RCTs, no validated test.
- That dydrogesterone is safe or effective in programmed frozen cycles. Both LOTUS trials were fresh cycles, and ESHRE 2025 records the malformation reports as unresolved associations.
What happens next
If you are mid-cycle, take the progesterone as prescribed and raise questions at the next visit rather than by changing the schedule yourself: the trials above compared fixed regimens, not improvised ones. Worth asking: which route you are on and why; when the clinic plans to stop and what that date rests on; and, for a frozen cycle, whether it is programmed or natural, because the two have opposite answers on whether progesterone is optional.
If you are offered progesterone outside those settings, to 'support' a natural cycle, after IUI, or to prevent a first miscarriage, the fair summary is that it has not been shown to help, and the attention may be better spent on the evaluation not yet done. The question is the same one to put to the IVF add-ons: what is the trial, and who was in it. It is reasonable to find this frustrating; the drug everyone takes is the one with the least modern evidence behind it, precisely because nobody will now randomise women to go without.
Not sure what your results mean?
Upload your protocol sheet or transfer-day bloods and IVY will map them to the trials and guidelines on this page, with the gaps marked.
Keep reading
16 Sources
- ESHRE Guideline Group on Ovarian Stimulation (Ata B, Bosch E, Broer S, et al.). ESHRE guideline: ovarian stimulation for IVF/ICSI: an update in 2025. Hum Reprod 2026;41(4):498-514. doi:10.1093/humrep/deag018. PMID 41732035. Section 'Triggering ovulation and luteal support'. ESHRE / Human Reproduction
- National Institute for Health and Care Excellence. Fertility problems: assessment and treatment. NICE guideline NG257, 31 March 2026. Recommendation 1.45, Triggering ovulation and luteal phase support. NICE
- van der Linden M, Buckingham K, Farquhar C, Kremer JA, Metwally M. Luteal phase support for assisted reproduction cycles. Cochrane Database Syst Rev 2015;(7):CD009154. doi:10.1002/14651858.CD009154.pub3. PMID 26148507. Cochrane
- Salang L, Teixeira DM, Solà I, Sothornwit J, Martins WP, Bofill Rodriguez M, Lumbiganon P. Luteal phase support for women trying to conceive by intrauterine insemination or sexual intercourse. Cochrane Database Syst Rev 2022;8:CD012396. doi:10.1002/14651858.CD012396.pub2. PMID 36000704. Cochrane
- Tournaye H, Sukhikh GT, Kahler E, Griesinger G. A Phase III randomized controlled trial comparing the efficacy, safety and tolerability of oral dydrogesterone versus micronized vaginal progesterone for luteal support in in vitro fertilization (LOTUS I). Hum Reprod 2017;32(5):1019-1027. doi:10.1093/humrep/dex023. PMID 28333318. Erratum Hum Reprod 2017 (PMID 28938733; placebo supplier names only). Human Reproduction
- Griesinger G, Blockeel C, Sukhikh GT, et al. Oral dydrogesterone versus intravaginal micronized progesterone gel for luteal phase support in IVF: a randomized clinical trial (LOTUS II). Hum Reprod 2018;33(12):2212-2221. doi:10.1093/humrep/dey306. PMID 30304457. Human Reproduction
- Watters M, Noble M, Child T, Nelson S. Short versus extended progesterone supplementation for luteal phase support in fresh IVF cycles: a systematic review and meta-analysis. Reprod Biomed Online 2020;40(1):143-150. doi:10.1016/j.rbmo.2019.10.009. PMID 31864902. Reproductive BioMedicine Online
- Coomarasamy A, Williams H, Truchanowicz E, et al. A randomized trial of progesterone in women with recurrent miscarriages (PROMISE). N Engl J Med 2015;373(22):2141-2148. doi:10.1056/NEJMoa1504927. PMID 26605928. New England Journal of Medicine
- Coomarasamy A, Devall AJ, Cheed V, et al. A randomized trial of progesterone in women with bleeding in early pregnancy (PRISM). N Engl J Med 2019;380(19):1815-1824. doi:10.1056/NEJMoa1813730. PMID 31067371. New England Journal of Medicine
- Devall AJ, Papadopoulou A, Podesek M, Haas DM, Price MJ, Coomarasamy A, Gallos ID. Progestogens for preventing miscarriage: a network meta-analysis. Cochrane Database Syst Rev 2021;4:CD013792. doi:10.1002/14651858.CD013792.pub2. PMID 33872382. Cochrane
- Haas DM, Bofill Rodriguez M, Hathaway TJ, Ramsey PS. Progestogen for preventing miscarriage in women with recurrent miscarriage of unclear etiology. Cochrane Database Syst Rev 2025;6:CD003511. doi:10.1002/14651858.CD003511.pub6. PMID 40497447. Cochrane
- ESHRE Guideline Group on RPL (Bender Atik R, Christiansen OB, Elson J, et al.). ESHRE guideline: recurrent pregnancy loss: an update in 2022. Hum Reprod Open 2023;2023(1):hoad002. doi:10.1093/hropen/hoad002. PMID 36873081. Full guideline PDF (January 2023), sections 14.2 and 17.6. ESHRE / Human Reproduction Open
- Practice Committee of the American Society for Reproductive Medicine. Diagnosis and treatment of luteal phase deficiency: a committee opinion. Fertil Steril 2026;126:633-639. Replaces the 2021 document (Fertil Steril 2021;115:1416-23). American Society for Reproductive Medicine
- Labarta E, Mariani G, Paolelli S, et al. Impact of low serum progesterone levels on the day of embryo transfer on pregnancy outcome: a prospective cohort study in artificial cycles with vaginal progesterone. Hum Reprod 2021;36(3):683-692. doi:10.1093/humrep/deaa322. PMID 33340402. Human Reproduction
- Erden M, Mumusoglu S, Uyanik E, Yarali Ozbek I, Esteves SC, Humaidan P, Yarali H. Revisiting natural cycle frozen embryo transfer: a systematic review and meta-analysis. Hum Reprod Update 2026;32(4):439-457. doi:10.1093/humupd/dmag002. PMID 41689846. Human Reproduction Update
- Engesgaard FM, Johansen AS, Saupstad M, Wang NF, Pinborg A, Løssl K. Luteal phase progesterone support in modified natural cycle frozen embryo transfer: a systematic review and meta-analysis. Reprod Biomed Online 2026;53(2):105731. doi:10.1016/j.rbmo.2026.105731. PMID 42160980. Reproductive BioMedicine Online
Frequently asked questions
What people ask once the pessaries or injections start.
Does progesterone help after IUI with clomiphene or letrozole?
Not that the trials can show. In the Cochrane review of luteal support for IUI and timed intercourse (Salang 2022), women stimulated with oral drugs had a live birth or ongoing pregnancy RR of 0.97 (95% CI 0.58 to 1.64; 2 RCTs, 485 women) with progesterone versus none, low certainty. Clinical pregnancy was higher (RR 1.44, 1.04 to 1.98; 6 RCTs, 1,073 women) but that did not translate into more births in the trials that counted them.
Should oestrogen be added to progesterone for luteal support?
ESHRE 2025 says adding oestradiol to progesterone 'is probably not recommended' (a conditional recommendation on low-quality evidence), citing the Cochrane 2015 comparison of progesterone with and without oestrogen: live birth or ongoing pregnancy OR 1.12 (95% CI 0.91 to 1.38; 9 RCTs, 1,651 women), no difference. Programmed frozen cycles are a separate matter, because there oestrogen is building the lining, not supporting the luteal phase.
Why is hCG not used for luteal support any more?
Because it drives ovarian hyperstimulation syndrome. In Cochrane 2015, hCG luteal support raised OHSS more than four-fold against placebo (OR 4.28, 95% CI 1.91 to 9.6; one RCT) without a clear gain in live birth over progesterone (OR 0.95, 0.65 to 1.38; 5 RCTs, 833 women). ESHRE 2025 makes a strong recommendation against hCG as luteal support in hCG-triggered cycles.
Does the GnRH agonist injection some clinics add in the luteal phase work?
The evidence conflicts. Cochrane 2015 pooled nine RCTs (2,861 women randomised) and found live birth or ongoing pregnancy lower with progesterone alone than with progesterone plus one or more GnRH agonist doses (OR 0.62, 95% CI 0.48 to 0.81), but with high unexplained variation between trials and low-quality evidence. ESHRE 2025, reviewing a 2022 network meta-analysis, says a single agonist bolus added to progesterone 'is probably not recommended' and repeated injections are likewise probably not recommended, both conditional. It is an add-on to ask about, not a standard.
Does progesterone or dydrogesterone cause birth defects?
For vaginal micronised progesterone in threatened miscarriage, the Cochrane 2021 network meta-analysis found congenital abnormalities RR 1.00 (95% CI 0.68 to 1.46) versus placebo, moderate certainty, so probably no difference. For dydrogesterone, ESHRE 2025 notes published reports of an association with congenital malformations and says they 'cannot be translated into a conclusion on causality'; the LOTUS trials reported similar safety profiles for the two drugs but were not sized to detect rare malformations.
I had one earlier miscarriage and am bleeding now. Is progesterone worth asking for?
The Cochrane 2021 network meta-analysis reports that in women with one or more previous miscarriages and early-pregnancy bleeding, vaginal micronised progesterone increased live birth versus placebo (RR 1.08, 95% CI 1.02 to 1.15, high certainty), whereas in women with no previous miscarriage it did not (RR 0.99, 0.95 to 1.04). ESHRE's recurrent pregnancy loss guideline frames its conditional recommendation around three or more losses. Either way, bleeding in early pregnancy needs a scan first; the drug is a conversation to have at that visit.


