A thin uterine lining postpones more frozen transfers than any other scan result. The evidence is narrower than the worry it causes: below about 7 mm live-birth odds do fall, but modestly, and how far depends on your preparation protocol — no study supports one millimetre cut-off that every clinic should apply.
Every number in this guide comes from peer-reviewed research indexed in PubMed, compiled by the ProIVF editorial team and reviewed by our medical advisory board.
| Finding | What the data show | Source |
|---|---|---|
| Median lining across 30,676 euploid frozen transfers | 8.9 mm overall — 9.0 in the USA, 8.7 in Spain, 8.0 in the UAE | Hum Reprod 2025, PMID 40639807 |
| Lining below 7 mm, programmed cycle | Live-birth odds down 22% (aOR 0.78, 95% CI 0.70–0.87) | Hum Reprod 2025, PMID 40639807 |
| Lining below 7 mm, fully natural cycle | No significant reduction (aOR 0.85, 95% CI 0.58–1.25, P = 0.41) | Hum Reprod 2025, PMID 40639807 |
| Lining ≤7 mm versus the 8–9 mm band | Live birth aOR 0.428 (0.353–0.519); preterm birth aOR 1.919; gestational age −0.505 weeks | Reprod Biol Endocrinol 2026, PMID 42271370 |
| Live birth by 2 mm band, pooled across 67 studies | 17% at 4–<6 mm, 26% at 6–<8 mm, 35% at 10–<12 mm, 43% at 12–<14 mm | Ultrasound Obstet Gynecol 2025, PMID 40757788 |
| Pattern versus thickness in 237 euploid transfers | Triple-line pattern predicted outcome (p = 0.004); thickness did not (AUC 0.505 for pregnancy) | Front Med 2026, PMID 42661947 |
What is a normal endometrial thickness before embryo transfer?
Around 8–9 mm is what high-volume programmes typically record before a frozen transfer. Across 30,676 single euploid (chromosomally normal) frozen transfers at 25 centres in three countries, the median was 8.9 mm: 9.0 in the USA, 8.7 in Spain, 8.0 in the United Arab Emirates PMID 40639807.
Three reference points from the literature:
- Below 7 mm is where the large datasets first detect a decline in live birth, and is the definition of “thin endometrium” used in most treatment trials.
- 8–9 mm was the comparison group in a 16,566-cycle frozen transfer study — the practical standard band PMID 42271370.
- Above 10 mm was associated with higher odds of live birth (aOR 1.254, 95% CI 1.108–1.420) with no increase in adverse perinatal outcomes in that same 16,566-cycle frozen cohort — the widespread fear of a “too thick” lining was not borne out PMID 42271370.
Where does the 7 mm rule come from?
From clinical convention, not from a trial. The idea that a lining must exceed roughly 6–8 mm is described in the literature as “widely believed” rather than established, and the 2025 multicentre study above tests it hardest: across the 25 centres, live birth fell below 7 mm by amounts that varied between clinics and between cycle types PMID 40639807.
The between-centre spread is not a rounding error: a lining under 7 mm appeared in 2.6% of programmed cycles in the USA, 5.2% in Spain and 12% in the UAE — one biological measurement, three very different populations and scanning practices.
Does a thin lining reduce the chance of live birth?
Yes, but the effect size depends on how the cycle was prepared. Using the same 30,676-cycle dataset:
- Programmed (artificial) cycles — estrogen then progesterone given as medication, no ovulation: a lining under 7 mm carried 22% lower adjusted odds of live birth (aOR 0.78, 95% CI 0.70–0.87, P ≤ 0.001).
- Modified natural cycles — you ovulate on your own but progesterone is added: 41% lower adjusted odds (aOR 0.59, 95% CI 0.49–0.72, P < 0.001).
- Fully natural cycles — monitored and timed on your own hormones alone: no effect at any threshold (aOR 0.85, 95% CI 0.58–1.25, P = 0.41), though only 759 such transfers (2.5% of the 30,676) were available to test it.
In a natural cycle the lining grows under your own follicle and corpus luteum, so a low number means something different from the same number in a cycle built entirely on tablets and patches; the frozen embryo transfer guide explains the three protocols.
A separate cohort of 16,566 frozen transfers reported a steeper drop — live birth aOR 0.428 at ≤7 mm against the 8–9 mm band, plus higher miscarriage (aOR 1.660) and preterm birth (aOR 1.919) and 0.505 fewer weeks of gestation PMID 42271370. The two cohorts disagree on how much a thin lining costs — 22% versus 57% lower adjusted odds — and that gap is the honest state of this evidence.
Will my transfer be cancelled because my lining is thin?
Not automatically, and the trend in the literature is against cancelling. The 2026 cohort authors concluded that a thin endometrium alone may not warrant cancelling, and recommended individualising protocols by age and BMI PMID 42271370.
The 2025 systematic review of 67 studies put it in prognostic language:
“Endometrial thickness is associated with reproductive outcomes, but demonstrates a gradient of effectiveness as a prognostic indicator, rather than offering a critical threshold below which ET should be avoided.” — Ultrasound in Obstetrics & Gynecology 2025;66:271, PMID 40757788
Clinics still cancel, usually at 6, 7 or 8 mm by local protocol, : postponing costs a clinic little, a low-odds transfer costs a patient a great deal.
Transfer-day thickness is also a weak discriminator. In the 30,676-cycle dataset a model including thickness scored AUC 0.597 against 0.591 without it, not a significant difference (P = 0.052) PMID 40639807. In the 237-cycle series the live-birth AUC was 0.5103, no better than a coin toss PMID 42661947.
Does endometrial pattern matter more than thickness?
Possibly, and it is under-measured in most clinics. In 237 hormone-replacement euploid transfers, the ultrasound pattern grades (Type A triple-line through Type C homogeneous) differed significantly for total pregnancy, biochemical pregnancy, miscarriage and live-birth rates (p = 0.004), with Type A highest — while thickness itself did not differ between pattern groups (p = 0.3819) or between pregnant and non-pregnant cycles (p = 0.5676) PMID 42661947.
Two patients at 7.5 mm can therefore be genuinely different, and the ruler hides it. The same study did find higher live-birth rates above 10.1 mm, so pattern and thickness are two partial views rather than rivals. If your scan report notes a “triple-line” endometrium, ask for that grade to be recorded alongside the millimetres — our guide to preparing before embryo transfer covers what to bring.
How is the lining measured, and how repeatable is the number?
Transvaginal ultrasound in the sagittal plane, measuring the maximum anterior-to-posterior distance of the endometrial echo, excluding the junctional zone. It takes seconds and is operator-dependent in a way blood tests are not.
Three sources of spread:
Where the caliper is placed — only the central maximum thickness is recorded, so intrauterine fluid or a tilted uterus can be measured either way.
When in the day you are scanned — in a programmed cycle the lining responds to estradiol exposure, so clinics that dose by blood level scan at a fixed interval after the dose. Serum estradiol and thickness generally rise together across an estrogen-built cycle; see our estradiol levels during IVF guide.
Premature progesterone rise — elevated progesterone (≥2 ng/mL) on the day of oocyte retrieval was associated with a higher frequency of thin endometrium (<7 mm) in a study of 128 IVF/ICSI cycles (p < 0.007) PMID 40943727. That is one reason a thin lining in a fresh cycle is often a hormone problem rather than a uterine one, covered in our progesterone levels after IVF transfer guide.
Does the same thickness mean the same thing in a fresh cycle?
No — in a fresh cycle the lining has already grown under supraphysiologic hormone levels, so the number is read against the stimulation response rather than against a frozen-cycle reference. Timing matters more than clinics admit: a study of 1,512 frozen-transfer cycles found the thickness measured on the day of oocyte retrieval in the preceding cycle predicted live birth better than the thickness on the actual transfer day (P = 0.03 versus P = 0.261) PMID 29369190.
It also gives the clearest dose-response curve available: clinical pregnancy rose from 50.0% at ≤6 mm to 84.2% above 16 mm, live birth from 33.3% to 63.2%, and each extra millimetre on retrieval day carried OR 1.069 (95% CI 1.011–1.130), with 8.75 mm as the best single ROC cut-off.
The pooled view across 67 studies is a gradient rather than a cliff PMID 40757788:
| Band or cut-off | Pooled clinical-pregnancy rate (bands) or odds ratio (cut-offs) |
|---|---|
| 4–<6 mm | 17% (95% CI 14–20%) |
| 6–<8 mm (reference) | 26% (95% CI 22–30%) |
| 10–<12 mm | 35% (95% CI 28–42%) |
| 12–<14 mm | 43% (95% CI 33–53%) |
| 14–<16 mm | 39% (95% CI 27–51%) |
| Fresh ET, thicker versus thinner at cut-off ≥5 mm | OR 5.66 (95% CI 1.10–28.98) |
| Fresh ET, cut-off ≥15 mm | OR 1.49 (95% CI 1.26–1.77) |
| Frozen-thawed ET, cut-off ≥5 mm | OR 2.65 (95% CI 1.23–5.72) |
| Frozen-thawed ET, cut-off ≥8 mm | OR 1.17 (95% CI 1.10–1.24) |
Clearing a low cut-off looks dramatic (OR 5.66), a high one much less (OR 1.49); effect size fell linearly as the cut-off rose in both the fresh (R² = 61.5%) and frozen (R² = 73.8%) analyses (P < 0.0001). The review graded 70% of its evidence assessments very low for bias and inconsistency — read the table as a shape, not a menu of promises.
What can be done to improve a thin lining?
Adjusting the estrogen is the only routine lever. PRP, scratching and hormonal add-ons all rest on very-low-certainty evidence for live birth.
Adjusting the estrogen — extending the estradiol run, switching route (oral, transdermal, vaginal) or re-running the preparation. Standard practice; it raises measured thickness, but no randomised trial shows it raises live birth.
Intrauterine platelet-rich plasma (PRP) — a 2024 meta-analysis of 8 randomised trials in 678 women with thin endometrium reported thickness increasing by a mean of 1.23 mm (95% CI 0.87–1.59), clinical pregnancy RR 2.04 (1.52–2.76), live birth RR 2.46 (1.57–3.85) and cancellation RR 0.46 (0.23–0.93) PMID 39215227. A 2026 update restricted to seven trials was far more cautious: clinical pregnancy RR 2.32 but 1.02–5.31, thickness gain 0.94 mm at I² = 90.5%, certainty very low, and the pregnancy result lost significance when any single trial was removed PMID 42676517.
The two reviews overlap heavily and point the same way: promising on the ruler, unproven on the baby.
Endometrial injury (“scratch”) — the 2026 Cochrane review of 24 randomised trials in 7,234 women restricted its primary analysis to the eight low-risk-of-bias trials (4,402 women): live birth OR 1.12 (95% CI 0.98–1.28), so a usual 27% chance lands between 27% and 32%; clinical pregnancy OR 1.08; miscarriage OR 1.00; probably mild-to-moderate pain. The reviewers’ own verdict: the effect is unclear PMID 42138348.
Hormonal add-ons for poor responders — a 2025 network meta-analysis of 22 trials in 4,131 women found no significant between-add-on difference in live birth on very-low-quality evidence; growth hormone ranked best for clinical pregnancy, beating estrogen protocols (OR 3.46, 95% CI 1.59–7.53, low evidence) — but that is a stimulation-response outcome, not a lining one PMID 41136866.
If a clinic offers PRP as a paid add-on, do not ask whether it thickens the lining; the pooled answer is roughly a millimetre. Ask instead how much outcome benefit the clinic attributes to it at very low certainty.
How should I read my lining scan?
Read the number beside protocol, pattern and hormone context, never as a grade alone. Four questions for your next appointment:
- “What millimetre do you cancel at, and does that change between natural and programmed cycles?” — because the natural-cycle data show no threshold at all PMID 40639807.
- “What pattern did you see — triple line or homogeneous — and is it on my record?” — pattern separated outcomes at p = 0.004 where thickness did not PMID 42661947.
- “Was there fluid, and was the measurement taken the same hour of the day as last time?” — both move the caliper reading.
- “If we postpone, is the plan more estrogen, a different route, a natural cycle, or an add-on?” — those are four different conversations with four different evidence levels, and “let’s just rescan” is a fifth.
Two linings, two very different cycles
Both accounts below are composites constructed from patterns documented in the literature and in clinic practice. Names, identity details and amounts are illustrative estimates, not identifiable medical records and not price quotes.
Lin, 34, Chengdu → Kuala Lumpur: 6.6 mm and a clinic that proceeded
Lin’s medicated frozen transfer was scanned at 6.6 mm on day 12 of oral estradiol, with a clear triple line. Her previous clinic in Chengdu cancelled at anything under 8 mm; this clinic extended estradiol by four days, measured 7.3 mm, added vaginal progesterone and transferred a day-5 euploid blastocyst.
It implanted, and she delivered a singleton girl at 38 weeks.
She paid about RM 13,500 (roughly USD 2,900) for the medicated cycle including monitoring, storage and transfer. “The first clinic treated my lining like an exam result,” she said. “This one treated it like one input.” Her case is not the average — a programmed cycle under 7 mm still carries about 22% lower adjusted odds — but it is the scenario the “no critical threshold” language refers to.
Mei, 37, Guangzhou → Bangkok: three cancellations and one PRP infusion
Mei’s lining measured 5.3 mm, 5.8 mm and 6.1 mm across three consecutive preparation attempts despite 6 mg/day oral estradiol valerate plus vaginal dosing — a genuinely thin-endometrium pattern, not one borderline reading. Her clinic proposed intrauterine platelet-rich plasma; she asked first what it changes, and the honest answer from the meta-analyses above is roughly one millimetre of thickness, with live-birth evidence graded very low.
She had one PRP infusion, reached 8.0 mm, transferred, and had a biochemical pregnancy; the next transfer, from her own frozen batch, gave a twin pregnancy that reduced to a singleton live birth at 35 weeks. Total spend across three failed preparations, one PRP add-on and two transfers was about THB 320,000 (roughly USD 8,900). “It took three cycles and one expensive experiment to stop being scared of the millimetre,” she said. Her case also carries the perinatal trade-off in the 16,566-cycle data: preterm birth odds roughly doubled at ≤7 mm PMID 42271370.
FAQ
Q: How thick should my lining be before embryo transfer?
Aim for the 8–9 mm band: it is where large-programme medians sit (8.9 mm across 30,676 euploid frozen transfers) PMID 40639807 and was the reference group in a 16,566-cycle study PMID 42271370. Between 7 and 8 mm the measured penalty is small and protocol-dependent, not zero.
Q: Is 7 mm too thin for embryo transfer?
Not automatically — below 7 mm, adjusted odds of live birth fell 22% in programmed cycles and 41% in modified-natural cycles, but showed no reduction in fully natural cycles (aOR 0.85, P = 0.41) PMID 40639807.
The 2026 cohort authors themselves concluded that thin lining alone may not justify cancelling PMID 42271370.
Q: Is a 12 mm lining better than an 8 mm lining?
Up to a point, yes — and without the downside often feared. Pooled live-birth proportions rose from 26% at 6–<8 mm to 43% at 12–<14 mm across the 67-study review PMID 40757788.
A lining above 10 mm carried higher live-birth odds (aOR 1.254), with no rise in adverse perinatal outcomes PMID 42271370.
Q: Why was my transfer cancelled for a thin lining?
Because most clinics run a fixed internal cut-off, commonly 6–8 mm, and postponement is cheap for a clinic while a low-odds transfer is expensive for a patient. The evidence describes a gradient, not a threshold below which transfer should be avoided PMID 40757788 — so asking for the alternative plan is legitimate.
Q: Can I thicken my lining naturally?
Movement, diet and supplements have not been shown to change measured endometrial thickness in randomised trials. What does move the number is more or differently-routed estrogen, and intrauterine PRP, which added a mean 1.23 mm in one pooled analysis PMID 39215227 and 0.94 mm at very low certainty in a smaller update PMID 42676517.
Q: Does endometrial scratching help?
The 2026 Cochrane review found live birth OR 1.12 (95% CI 0.98–1.28) in low-risk-of-bias trials — consistent with no effect or a small improvement — plus probable mild-to-moderate pain, across 24 trials and 7,234 women PMID 42138348. It is not a proven procedure.
Q: Does a thin lining mean a higher miscarriage or preterm birth risk?
In one large frozen-transfer cohort, ≤7 mm carried higher miscarriage odds (aOR 1.660), higher preterm birth odds (aOR 1.919) and 0.505 fewer gestational weeks than 8–9 mm PMID 42271370. Absolute risks remained modest.
Q: Does the estrogen level matter more than the thickness?
Estrogen and lining move together, and estrogen has its own numbers: in 412 programmed frozen transfers, a transfer-day (mid-luteal) serum estradiol inside the 292–<409 pg/mL window went with a 59% live birth rate versus 39% below it and 28% above it PMID 38604265. See our estradiol levels during IVF guide.
Ask your clinic for its own numbers
A lining measurement is one number, taken by a human hand and read differently by every clinic. Ask for it alongside protocol type, pattern grade and hormone context, and ask what your clinic does instead of cancelling — that is where practice variation really sits.
Clinic-level results are public in a few markets: US CDC ART success-rate data and the SART registry let you look up per-centre outcomes, and ASRM practice documents set the guidance behind endometrial assessment. Compare programmes in our hospital directory, or send your scan reports to us through the contact page.
This article is health information, not individual medical advice. Where it conflicts with the written instructions from your own clinic, follow your clinic.
Every number above links to its PubMed record or the cited regulatory page; the ProIVF Medical Advisory Board reviewed the final text. About the team: About us.
Last updated: October 3, 2026
