IVF Study Finds Progesterone-to-Oocyte Ratio Adds Little Beyond Its Parts

A large retrospective analysis of more than 7,700 in vitro fertilization and intracytoplasmic sperm injection cycle records has found that a widely discussed laboratory index, the progesterone-to-metaphase II oocyte ratio, was associated with embryo yield and recorded pregnancy outcomes, but that its apparent usefulness largely dissolved once its two components were accounted for. The study, published in Frontiers in Endocrinology, analyzed cycles performed at the Reproductive Medicine Center of Liuzhou Hospital, Guangzhou Women and Children’s Medical Center, in China, between August 2020 and December 2024, and concluded that the ratio does not support a clinical cutoff for trigger-day decisions or freeze-all strategies.

The progesterone-to-metaphase II oocyte ratio, abbreviated P/MII, is calculated by dividing the serum progesterone concentration measured on the day of final oocyte maturation trigger by the number of mature, metaphase II oocytes retrieved. Elevated late-follicular progesterone during ovarian stimulation has long been debated as a possible marker of impaired pregnancy chances after fresh embryo transfer, with proposed mechanisms centered on premature advancement of endometrial maturation and embryo–endometrium asynchrony. Previous studies have divided progesterone by follicle number, retrieved oocyte number, or mature oocyte number to construct indices, but the authors emphasize that such calculations describe progesterone relative to a measure of ovarian response and do not establish a physiological measure of endometrial progesterone exposure or receptivity.

The analytic cohort comprised 7,778 cycles in women aged 20 to 40 years, in which 98,558 oocytes were retrieved, including 85,956 mature MII oocytes, 60,309 recorded two-pronuclei zygotes, and 33,360 blastocysts. The source retrieval-order field classified 6,148 cycles as first retrievals and 1,630 as later retrievals. Using segmented Poisson regression, the researchers identified an exploratory changepoint of 0.0711 ng/ml per MII oocyte, dividing the cohort into 5,408 lower-ratio and 2,370 higher-ratio records. The authors stress that this internally derived value is retained only for analytic reproducibility and does not define a validated clinical threshold.

In unadjusted comparisons, cycles with P/MII above the changepoint showed lower blastocyst yield per recorded 2PN zygote (39.9% versus 58.3%), lower good-quality blastocyst yield (15.7% versus 25.1%), and lower rates of implantation-positive cycles, clinical pregnancy, and live birth among transfer records. Women in the higher-ratio group were older, had lower AMH and antral follicle counts, higher basal FSH, and markedly fewer retrieved, mature, and fertilized oocytes, with median MII counts of five versus twelve, illustrating how strongly the ratio tracks ovarian response.

After multivariable adjustment for age, infertility duration, gonadotropin dose, stimulation protocol, diagnosis, cycle year, oocyte count, trigger-day estradiol, antral follicle count, and AMH, the high-versus-low ratio remained associated with fewer blastocysts per recorded 2PN zygote (adjusted rate ratio 0.796; 95% CI 0.762–0.831) and a modestly lower good-quality Day-3 embryo rate (aRR 0.959; 95% CI 0.924–0.995). Among the 4,273 records with at least one embryo transferred, the higher-ratio category was associated with lower adjusted odds of an implantation-positive cycle (aOR 0.693), a clinical pregnancy flag (aOR 0.694), and a live birth flag (aOR 0.796), with no clear association with miscarriage among clinical pregnancies (aOR 0.758; P = 0.096).

The most consequential findings came from sensitivity analyses using newly linked laboratory source fields. The original 2PN denominator included embryos allocated to cleavage-stage transfer, cryopreservation, and extended culture, mixing embryo development with embryo allocation; in 173 records, blastocyst counts even exceeded 2PN counts. When the analysis was restricted to the 5,821 records with positive recorded extended culture counts and the denominator changed from 2PN zygotes to recorded cultured embryos, the categorical association attenuated to an adjusted rate ratio of 0.974 (95% CI 0.945–1.005; P = 0.101), while a small continuous association persisted (aRR 0.993 per 0.01-unit increase; 95% CI 0.989–0.997). In 4,960 first-retrieval culture records with no thawing evidence, the categorical estimate was similarly uncertain (0.973; P = 0.108).

Pregnancy associations also weakened under scrutiny. In 3,526 first-retrieval, date-consistent records without thawing evidence, the ratio term after simultaneous inclusion of progesterone and MII count was associated with clinical pregnancy flags (aOR 0.952 per 0.01; 95% CI 0.919–0.986; P = 0.006) but was uncertain for live birth flags (aOR 0.968; 95% CI 0.935–1.002; P = 0.067). In the full primary transfer cohort, the ratio estimate for live birth after adjustment for both components was 0.988 per 0.01 (95% CI 0.961–1.015; P = 0.368). Because P/MII is a deterministic function of progesterone and MII count, the authors note that adding it to a component model introduces a particular nonlinear function rather than an independently manipulable hormonal exposure.

Predictive performance comparisons reinforced the restrained interpretation. In fivefold cross-validation that grouped records by source identifier so that no identifier appeared in both training and test sets, live birth areas under the receiver operating characteristic curve were 0.623 for the base model plus both components, 0.623 for the base model plus P/MII, and 0.624 for the combined model. These small descriptive differences lacked paired uncertainty estimates and, in the authors’ assessment, do not establish superiority or clinical usefulness. The study also faced verification limits: outcome flags were preserved rather than clinically re-adjudicated, miscarriage and live birth definitions could not be validated for every flagged record, and 15 positive-transfer records contained thawing evidence leaving embryo origins uncertain.

The authors conclude that recorded associations depended on the culture population, the denominator, and the outcome, and that incremental clinical usefulness was not established. The data do not support a physiological ratio mechanism or a clinical cutoff for trigger-day or freeze-all decisions, and P/MII, calculated only after retrieval, is not a direct measure of endometrial progesterone exposure. Future studies, they suggest, should evaluate comparable treatment populations using embryo-level provenance, clinically adjudicated endpoints, flexible component modeling, and independent external validation.

Subject of Research: Associations between the trigger-day progesterone-to-mature-oocyte ratio and embryo yield and recorded IVF/ICSI transfer outcomes in a retrospective cohort

Article Title: Trigger-day progesterone-to-mature-oocyte ratio, cycle-level blastocyst yield, and recorded embryo-transfer outcomes: a retrospective cohort study

Article References: Wei, Y., Xu, Y., Wei, N., Chen, Y., Zhang, Y., Li, X., & Wei, H. (2026). Trigger-day progesterone-to-mature-oocyte ratio, cycle-level blastocyst yield, and recorded embryo-transfer outcomes: a retrospective cohort study. Frontiers in Endocrinology, 17, Article 1938120. https://doi.org/10.3389/fendo.2026.1938120

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1938120

Keywords: IVF, ICSI, progesterone, P/MII ratio, blastocyst yield, embryo transfer, live birth, retrospective cohort, ovarian stimulation, endometrial receptivity, reproductive medicine, Trigger-day