IVF E2/P Ratio Before Progesterone: Signal, Not a Target

A hormone ratio measured before progesterone starts may predict outcomes in hormone replacement therapy frozen embryo transfer (HRT-FET) cycles. It cannot yet tell an individual patient whether her transfer will work, nor does evidence show that changing the ratio improves live birth.

That distinction is the whole story. In a retrospective study of more than 25,000 HRT-FET cycles, a higher estradiol-to-progesterone (E2/P) ratio immediately before progesterone initiation was independently associated with lower clinical pregnancy and live birth rates. But the ratio's ability to sort successful from unsuccessful individual cycles was weak: its area under the receiver-operating-characteristic curve, or AUC, was 0.559. A coin flip scores 0.50.

For people in an active HRT-FET cycle, this is a reasonable topic to raise with a reproductive endocrinologist. It is not a number to optimize at home, and it is not evidence to delay or alter medication without a clinic-specific plan.

What the large HRT-FET study found

Shuai et al. reviewed more than 25,000 HRT-FET cycles performed at one Chinese tertiary IVF center between 2017 and 2022. Every participant used vaginal micronized progesterone, Utrogestan 200 mg three times daily. The researchers measured estradiol and progesterone immediately before progesterone supplementation began, converted progesterone to pg/mL to align the units, and calculated the E2/P ratio.

They grouped cycles into four ratio categories: below 0.62, 0.62 to below 1.14, 1.14 to below 2.91, and 2.91 or higher. The highest-ratio group had the poorest clinical pregnancy and live birth outcomes. The association remained statistically significant across progressively adjusted logistic-regression models that accounted for age, body mass index, infertility type, embryo stage, number of embryos transferred, and endometrial thickness when progesterone began.

That is useful evidence of an association. It also explains why the finding deserves more attention than a tiny, unadjusted clinic audit. Still, regression adjustment cannot remove every difference between patients or every treatment choice that led to their hormone values. Estradiol dose was chosen at physician discretion, for example. Those choices may track patient characteristics or clinical judgments that the model could not fully capture.

The study's central limitation is its predictive performance. An AUC of 0.559 means the ratio adds a small amount of information across a large population but performs poorly as a standalone test for a single transfer. The authors did not offer a clinically actionable cutoff. A statistically significant result in 25,000 cycles can be real and still be too imprecise for bedside decision-making.

Why the ratio is biologically plausible

Estradiol prepares the endometrium for progesterone by increasing progesterone-receptor expression. Progesterone then drives the secretory transformation that supports implantation. A disproportionately high estradiol environment at progesterone initiation could, in theory, reduce or desensitize progesterone signaling and shift the implantation window.

That is a plausible explanation, not proof of mechanism in these patients. The Shuai study measured hormone levels and pregnancy outcomes; it did not measure endometrial gene expression, receptor function, histology, or the timing of the implantation window. Biology gives the association a coherent story, but it does not turn a retrospective association into a treatment target.

The stronger evidence concerns progesterone after it starts

The pre-progesterone E2/P ratio asks one question: does the hormonal setting appear favorable when progesterone begins? Most clinically developed monitoring research asks another: did the patient achieve adequate progesterone exposure later in the cycle?

A 2021 systematic review and meta-analysis by Melo et al. found that low luteal-phase serum progesterone in HRT-FET was associated with worse outcomes. For vaginal micronized progesterone, the literature has proposed a threshold near 9.2 ng/mL; individual included studies reported adjusted odds ratios of roughly 1.8 to 2.5 for live birth below that threshold. Thresholds can vary by assay and progesterone route, so this is not a universal target for every protocol.

A prospective cohort study by Dashti et al. also found that serum progesterone on transfer day independently predicted clinical pregnancy and live birth. A separate 2026 retrospective cohort study by Alpçetin et al. found that progesterone on the day of the beta-hCG pregnancy test predicted ongoing pregnancy in FET cycles. These are different time points and, in the latter study, a different endpoint. Together, they support the idea that progesterone bioavailability matters throughout the luteal phase.

They do not prove that E2/P before progesterone begins predicts the same risk through the same pathway. Shuai et al. did not measure transfer-day progesterone in the same cycles. The ratio could reflect an independent upstream issue, later progesterone absorption, both, or an unmeasured feature of patient management.

There is also early intervention evidence for the downstream problem. A 2025 study by Yazbeck et al. reported higher live birth rates with a luteal-phase rescue protocol that escalated progesterone supplementation when luteal deficiency was detected. That makes downstream progesterone monitoring more clinically compelling than an upstream ratio, at least for now. Whether correcting E2/P before progesterone adds anything beyond rescue remains untested.

Different preparation protocols complicate the story

The ratio finding applies to a specific HRT protocol, not every frozen transfer. All cycles in the primary study used vaginal micronized progesterone, and different oral, subcutaneous, and intramuscular regimens produce different serum and endometrial progesterone patterns.

Protocol comparisons also suggest that the endometrium can reach similar outcomes under meaningfully different hormonal conditions. In a retrospective comparison of 1,880 FET cycles, Guan et al. found no statistically significant difference in live birth between letrozole-plus-human menopausal gonadotropin preparation and standard HRT after adjustment (odds ratio 1.145, 95% confidence interval 0.932 to 1.407; P = 0.198). Letrozole-plus-HMG uses endogenous follicular estradiol and an ovulation-induction pattern, unlike fixed-dose exogenous estrogen in HRT.

This does not refute Shuai et al. Guan et al. did not test pre-progesterone E2/P as a predictor. It does temper the claim that one ratio should dominate clinical decisions: substantially different preparation strategies can produce comparable outcomes.

Likewise, a 2025 retrospective study by Huang et al. found that GnRH-agonist pretreatment before HRT improved FET outcomes among overweight and obese women. That intervention may alter endometrial receptivity through LH suppression and reduced androgen exposure. It is not evidence that lowering E2/P improves results, but it shows that the hormonal environment before progesterone can matter in selected populations.

What keeps this from changing practice

• No intervention trial: No study has tested whether lowering estradiol, delaying progesterone, or changing preparation in response to a high E2/P ratio increases live birth.
• Embryo ploidy remains unresolved: The primary study excluded preimplantation genetic testing cycles. Without known embryo euploidy, the ratio's independent contribution cannot be separated fully from embryo-level variation, a major determinant of implantation.
• Limited generalizability: Results come from one center, one progesterone formulation and route, and local estrogen-dosing practices that may differ elsewhere.
• The extremes are incomplete: About 60 cycles with estradiol above 1,000 pg/mL were excluded because of the assay dilution method, so the highest end of the ratio distribution remains less certain.
• Effect sizes by quartile are unavailable: The published evidence available here confirms significant adjusted trends but does not provide enough complete per-quartile adjusted estimates to quantify the outcome difference precisely.

What patients can do with this information

If you are preparing for an HRT-FET, ask whether your clinic measures estradiol and progesterone together before progesterone starts, how it interprets those values, and whether it monitors serum progesterone after supplementation begins. Ask how its approach changes for your progesterone route, embryo testing status, and medical history.

There is no validated pre-progesterone E2/P target, no proven correction strategy, and no cutoff that reliably predicts an individual cycle's outcome.

The E2/P ratio is a credible hypothesis-generating marker from a very large observational study. Its weak discrimination and lack of intervention data keep it on the research side of the clinic door. For now, the more established conversation is whether progesterone exposure later in HRT-FET is adequate and whether a clinic has an evidence-based rescue protocol when it is not.