Impact of serum progesterone levels and rescue progesterone supplemental luteal support on pregnancy outcomes in frozen embryo transfer: a controlled trial

Abstract

Background

The role of serum progesterone levels and the benefit of rescue luteal support in frozen embryo transfer (FET) cycles remain debated. While low serum progesterone has been linked to poorer outcomes, evidence on whether supplementation improves pregnancy rates is inconsistent.

Objective

To evaluate the impact of serum progesterone levels and additional progesterone supplementation on pregnancy outcomes in women undergoing FET.

Methods

In this controlled, single-blind trial, 74 women were allocated into three groups: Group I (n = 26, progesterone < 10 ng/mL, standard luteal support), Group II (n = 23, progesterone < 10 ng/mL, standard luteal support + 400 mg progesterone), and Group III (n = 25, progesterone ≥ 10 ng/mL, standard luteal support). Primary outcome was clinical pregnancy; secondary outcomes included biochemical and ongoing pregnancy, and pregnancy loss. Logistic regression identified independent predictors.

Results

Progesterone levels differed significantly between groups both on transfer day and three days later (p < .001). Pregnancy outcomes did not differ significantly: biochemical pregnancy (50.0%, 34.8%, 44.0%), clinical pregnancy (42.3%, 34.8%, 44.0%), ongoing pregnancy (42.3%, 30.4%, 44.0%), and pregnancy loss (7.7%, 4.3%, 0%) for Groups I–III, respectively. Maternal age was the only consistent predictor across models, negatively associated with pregnancy outcomes (OR 0.85–0.87, p < .05).

Conclusion

Supplemental luteal support in women with low progesterone mitigate negative effects, as outcomes were comparable to women with normal progesterone even though it did not significantly improve pregnancy outcomes. Maternal age remained the strongest predictor. Larger trials are needed to clarify benefits in high-risk subgroups and to define optimal progesterone thresholds.

Trial registration

Iranian Registry of Clinical Trials (IRCT), IRCT20200905048630N2, registered on 22 Aguest 2021.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-026-08645-w.

Introduction

Frozen embryo transfer (FET) has evolved from a supplementary technique to a cornerstone of modern assisted reproductive technology (ART). Its application has expanded markedly, with the proportion of FET cycles now rivaling or exceeding that of fresh transfers in many regions [1]. This shift is driven by advances in vitrification techniques yielding superior embryo survival rates, the avoidance of ovarian hyperstimulation syndrome (OHSS), and the flexibility to optimize endometrial receptivity without the deleterious effects of supraphysiologic hormone levels from stimulation [2–4]. Critically, evidence from meta-analyses suggests that FET cycles may be associated with higher implantation, ongoing pregnancy, and live birth rates for certain patient populations compared to fresh transfers, further solidifying its pivotal role in infertility treatment [5].

The success of any embryo transfer, whether fresh or frozen, depends on the delicate synchronization of a viable embryo with a receptive endometrium. In artificial FET cycles, this receptivity is entirely orchestrated through exogenous hormonal administration [6]. Estradiol primes the endometrium for proliferation, but it is progesterone that is indispensable for orchestrating the complex molecular and morphological changes of the secretory transformation, creating the narrow window of implantation [7]. Despite its recognized criticality, the management of progesterone support remains one of the most debated aspects of ART, with significant variations in clinical practice worldwide [8].

A compelling and growing hypothesis posits that a subset of women exhibit suboptimal progesterone absorption or metabolism, leading to unexpectedly low serum levels despite standard progesterone administration [9]. Emerging evidence indicates that these low concentrations on or around the day of transfer are associated with significantly reduced implantation potential, higher rates of early pregnancy loss, and consequently, lower live birth rates [10]. This has prompted investigations into the concept of personalized, or individualized, luteal phase support. The proposed strategy involves identifying at-risk women through serum progesterone monitoring and supplementing them with additional progesterone to potentially “rescue” the cycle [11, 12].

The literature presents conflicting evidence on progesterone supplementation [11]. While some studies report that correcting low levels with supplemental progesterone normalizes pregnancy rates to match those of non-deficient patients [13], others show no significant benefit to clinical pregnancy rates beyond standard protocols [14].

Given these uncertainties, the present study was designed to evaluate the impact of serum progesterone concentration and additional progesterone supplementation on pregnancy outcomes in women undergoing frozen embryo transfer cycles. By comparing biochemical, clinical, and ongoing pregnancy rates, as well as pregnancy loss across three groups defined by progesterone levels and luteal support protocols, this study aimed to clarify the role of progesterone optimization in improving ART success.

Methodology

Study design and setting

This study was conducted as a controlled, single-blind trial at the Infertility Center of Hormozgan University of Medical Sciences between November 2022 and December 2024. It was conducted and is reported in accordance with the CONSORT guidelines for clinical trials. The objective was to evaluate the effect of different serum progesterone levels and additional progesterone supplementation on pregnancy outcomes in women undergoing frozen embryo transfer (FET). This study was approved by the Medical Ethics Committee of Hormozgan University of Medical Sciences with ethical code of IR.HUMS.REC.1402.394.

Participants

Eligible participants were infertile women aged 18–37 years who were candidates for frozen embryo transfer of day-3 embryos, according to ASRM guidelines [15]. Inclusion criteria were: Artificial cycle endometrial preparation with estradiol and progesterone, availability of 2–3 good-quality embryos and, normal uterine cavity confirmed by ultrasound or hysteroscopy. Exclusion criteria included: uterine anomalies (e.g., fibroids, polyps, intrauterine adhesions), recurrent implantation failure (RIF) or recurrent miscarriage, advanced endometriosis (grade 3–4), poor-quality embryos, body mass index (BMI) > 30 kg/m², age > 37 years, and endometrial thickness < 7 mm after hormone therapy. All participants who met al.l exclusion and inclusion criteria were selected for this survey.

Sample size and grouping

Sample size was determined based on previous studies evaluating the impact of serum progesterone levels and supplemental luteal support in frozen embryo transfer (FET) cycles [13, 16, 17]. A total of 74 women were enrolled, which was considered sufficient to detect large differences in pregnancy outcomes between groups, but not powered to identify small or modest effects. Group I (n = 26): Progesterone < 10 ng/mL, routine luteal support only, Group II (n = 23): routine luteal support plus additional vaginal progesterone (400 mg daily) and, Group III (n = 25): Progesterone ≥ 10 ng/mL, routine luteal support only (Control Group) [18].

Endometrial Preparation and luteal support

All women received oral estradiol valerate (6 mg/day, starting on cycle day 2) for endometrial preparation. When endometrial thickness reached ≥ 7 mm, intramuscular progesterone 50 mg/day was initiated. Embryo transfer was performed 4 days after the start of progesterone administration. Luteal support was continued until the pregnancy test and, if positive, up to 10 weeks of gestation [19].

Hormonal measurements

Serum progesterone was measured 1h before embryo transfer and again 3 days after embryo transfer, using a standardized chemiluminescent immunoassay. Serum estradiol (E2) levels were also recorded prior to embryo transfer, consistent with established monitoring practices [10, 20].

Embryo transfer procedure

Two to three good-quality day-3 embryos were transferred under ultrasound guidance. Embryo quality was classified based on morphology and fragmentation, following established grading systems: Grade I and II (good or fair quality) defined as 8 cell embryos with evenly sized blastomeres and ≤ 5% fragmentation; or 7–10 with minor defects or 10–15% fragmentation; and Grade III (poor quality) as < 7 or > 10 cells, ≥ 10% fragmentation, or multiple defects [21].

Outcome measures

The primary outcome was the clinical pregnancy rate, defined as the presence of a gestational sac with fetal cardiac activity on ultrasound performed four weeks after embryo transfer. Secondary outcomes included: biochemical pregnancy (positive serum β-hCG 14 days after transfer), ongoing pregnancy (a viable intrauterine pregnancy beyond 12 weeks of gestation), and pregnancy loss (miscarriage occurring before 12 weeks of gestation) [22].

Statistical analysis

Data were analyzed using SPSS software (version 26). The distribution of continuous variables was assessed using the Shapiro-Wilk test. Variables that were not normally distributed were compared across the three groups using the Kruskal-Wallis H test, with pairwise comparisons performed using the Mann-Whitney U test where appropriate. For normally distributed data, such as maternal age, the One-way ANOVA test was used. Categorical variables were compared using the chi-square or Fisher’s exact test.

To identify independent predictors for each pregnancy outcome (biochemical, clinical, ongoing, and abortion), multivariable binary logistic regression analyses were performed. A single model was constructed for each outcome. The independent predictor variables included in each model were: treatment group (with Group III as the reference), maternal age, progesterone levels on transfer day, progesterone 3 days post-transfer, estradiol levels, and BMI. Results are presented as adjusted odds ratios (OR) with 95% confidence intervals (CI).

To identify independent predictors of clinical pregnancy, a multivariable logistic regression analysis was performed. The dependent variable was clinical pregnancy outcome (pregnant vs. not pregnant). Predictor variables included treatment group (three categories), maternal age (years), body mass index (BMI, kg/m²), and embryo quality (categorized as high-quality or poor-quality embryos). A p-value < 0.05 was considered statistically significant.

Results

Baseline characteristics

A total of 74 women were included and allocated into three groups (Group I: n = 26, Group II: n = 23, Group III: n = 25). The mean age, body mass index (BMI), infertility duration, endometrial thickness, estradiol (E2) levels before transfer, number of embryos transferred, and embryo quality were comparable among the three groups, with no statistically significant differences (all p > .05), except for progesterone levels which differed significantly (p < .001) (Table 1).

Table 1.

Comparison of patients’ characteristics

Variables		Group I (n = 26)a	Group II (n = 23)b	Group III (n = 25)c	P Value
Age(years)*		35.23 ± 7.25	35.39 ± 6.07	32.32 ± 5.78	0.17
BMI(kg/m2)**		26.1 ± 2.9	25.9 ± 1.8	25.5 ± 2.8	0.23
Infertility duration (years)**		5.52 ± 4.29	6.83 ± 4.38	6.54 ± 4.10	0.39
Endometrial thickness(mm)**		7.97 ± 1.11	8.20 ± 1.32	8.13 ± 0.78	0.31
Progesterone level on embryo transfer day(ng/mL)**		6.11 ± 2.30	8.73 ± 1.70	16.08 ± 2.49	< 0.001 a, b < 0.001 a, c < 0.001 b, c < 0.001
Progesterone level 3 days after embryo transfer(ng/mL)**		14.02 ± 5.19	7.72 ± 2.20	24.20 ± 4.77	< 0.001 a, b < 0.001 a, c < 0.001 b, c < 0.001
Estradiol (E2) levels before transfer(ng/mL)**		176.50 ± 50.45	172.38 ± 53.82	177.92 ± 75.58	0.51
Number of embryos transferred**		2.85 ± 0.46	2.91 ± 0.42	3.04 ± 0.20	0.17
Embryo quality mean(n%)**	High	2.42(85.6%)	2.43(84.8%)	2.76(91.3%)	0.16
	Low	0.38(14.4%)	0.48(15.2%)	0.20(8.7%)	0.32

Values are presented as mean ± standard deviation or median. Bold indicated p < .05 was regarded as a significant value

*The One-way ANOVA was used. The data distribution was normal based on Shapiro–Wilk test

**The Kruskal–Wallis H was used to determine if there are statistically significant differences between all groups Mann–Whitney test was used to compare between two groups. The data distribution was not normal based on Shapiro–Wilk test

Hormonal levels

Serum progesterone levels differed significantly between groups. On the day of embryo transfer, mean progesterone concentrations were 6.11 ± 2.30 ng/mL in Group I, 8.73 ± 1.70 ng/mL in Group II and, 16.08 ± 2.49 ng/mL in Group III (p < .001). Similar differences were observed three days after embryo transfer (14.02 ± 5.19 ng/mL, 7.72 ± 2.20 ng/mL and, 24.20 ± 4.77 ng/mL, respectively; p < .001).

Pregnancy outcomes

No significant differences were observed in pregnancy outcomes among the three groups.

• Biochemical pregnancy occurred in 50.0% of women in Group I, 34.8% in Group II and, 44.0% in Group III (p = .56).

• Clinical pregnancy was achieved in 42.3%, 34.8%, and 44.0% of women in Groups I, II, and III, respectively (p = .79).

• Ongoing pregnancy rates were 42.3%, 30.4%, and 44.0% across the three groups (p = .58).

• Pregnancy loss occurred in 7.7% of Group I, 4.3% of Group II, and none of Group III (p = .38).

Although numerical differences were noted, particularly with slightly higher pregnancy rates in Groups I and II compared with Group III, these did not reach statistical significance (Table 2).

Table 2.

Pregnancy outcomes in different study groups

variables		Group Ia (n = 26)	Group IIb (n = 23)	Group IIIc (n = 25)	P Value
Biochemical pregnancy	Positive	13(50.0%)	8(34.8%)	11(44.0%)	0.56 a, b:0.28 a, c:0.67 b, c:0.51
	Negative	13(50.0%)	15(65.2%)	14(56.0%)
Clinical pregnancy	Positive	11(42.3%)	8(34.8%)	11(44.0%)	0.79 a, b:0.59 a, c:0.90 b, c:0.51
	Negative	15(57.7%)	15(65.2%)	14(56.0%)
Ongoing pregnancy	Positive	11(42.3%)	7(30.4%)	11(44.0%)	0.58 a, b:0.39 a, c:0.90 b, c:0.33
	Negative	15(57.7%)	16(69.6%)	14(56.0%)
Pregnancy loss	Positive	2(7.7%)	1(4.3%)	0(0.0%)	0.38 a, b:0.63 a, c:0.16 b, c:0.29
	Negative	24(92.3%)	22(95.7%)	25(100.0%)

Values are presented as number (%). P values were only calculated for positive outcomes and were calculated using Chi-square or Fisher’s exact test, as appropriate. p < .05 was regarded as a significant value

Regression analysis

Binary logistic regression analyses were conducted to identify predictors of biochemical pregnancy, clinical pregnancy, ongoing pregnancy, and abortion. Independent variables included treatment group, maternal age, progesterone levels on transfer day, progesterone 3 days after embryo transfer, induction duration, estradiol levels, high-quality embryo, poor-quality embryo, and BMI (Table 3).

Table 3.

Multivariable logistic regression predicting clinical pregnancy

Outcome	Model χ² [23]	Nagelkerke R²	% Correct	Significant Predictor(s)	OR (95% CI) for Age
Biochemical Pregnancy	16.87 (10), p = .077	0.273	71.6%	Age (p = .022)	0.86 (0.76–0.98)
Clinical Pregnancy	16.42 (10), p = .088	0.269	68.9%	Age (p = .033)	0.87 (0.77–0.99)
Ongoing Pregnancy	19.43 (10), p = .035	0.313	71.6%	Age (p = .015)	0.85 (0.74–0.97)
Abortion	25.11 (10), p = .005	1.000*	100%	None reliable (low cases)	—

The table reports odds ratios (OR), P-value and, 95% confidence intervals (CI) for each predictor variable

Biochemical pregnancy

The model was not statistically significant, χ²(10) = 16.87, p = .077, explaining 20.4% (Cox & Snell R²) to 27.3% (Nagelkerke R²) of the variance, with a classification accuracy of 71.6%. Maternal age was a significant negative predictor (B = − 0.148, p = .022, OR = 0.86, 95% CI [0.76–0.98]). Other predictors, including group, embryo quality, estradiol, and BMI, were not significant.

Clinical pregnancy

The model was not significant overall, χ²(10) = 16.42, p = .088, accounting for 19.9–26.9% of the variance, with 68.9% correct classification. Age was again a significant negative predictor (B = − 0.137, p = .033, OR = 0.87, 95% CI [0.77–0.99]). No other variables were significant.

Ongoing pregnancy

The model was statistically significant, χ²(10) = 19.43, p = .035, explaining 23.1–31.3% of the variance, with 71.6% correct classification. Age remained a significant predictor (B = − 0.164, p = .015, OR = 0.85, 95% CI [0.74–0.97]). Estradiol (p = .083) and BMI (p = .107) approached significance but did not reach the threshold.

Abortion

The regression model was significant, χ²(10) = 25.11, p = .005, with variance explained up to 100% (Nagelkerke R²). The model classified all cases correctly (100%). However, due to the very low number of abortion events (n = 3), the regression coefficients were unstable and suggest model overfitting. Therefore, the abortion results should be interpreted with caution.

Discussion

This controlled trial investigated the impact of low serum progesterone levels on the day of embryo transfer and the potential benefit of supplemental luteal support on pregnancy outcomes in women undergoing frozen embryo transfer (FET) cycles. The principal finding of our study is that while serum progesterone levels were significantly lower in Groups I and II compared to the control Group III, the implementation of an intensified luteal support protocol (additional 400 mg progesterone daily) in Group II did not yield a statistically significant improvement in biochemical, clinical, or ongoing pregnancy rates. However, this absence of statistical difference should not be interpreted as a lack of clinical effect. This normalization occurred despite a significant deficiency in Group II’s baseline progesterone levels on transfer day. The significant rise in progesterone levels three days post-transfer in the supplemented group confirms the biological activity of the intervention. While our study may have been underpowered to detect a small but significant benefit above the control group’s rate, it powerfully demonstrates that supplementation can prevent a negative outcome.

Our results align with a segment of the existing literature that questions the universal benefit of progesterone supplementation beyond standard protocols [13, 24, 25]. For instance, Vuong et al. (2021) found no significant difference in live birth rates between patients receiving a combination of micronized progesterone and dydrogesterone versus micronized progesterone alone, suggesting a potential ceiling effect for luteal phase support [14]. Similarly, the lack of significant difference in pregnancy loss, despite a numerical trend, indicates that low progesterone might not be as critical a driver of early miscarriage in all patients as previously hypothesized.

The most consistent and significant predictor of pregnancy outcomes across all our regression models was maternal age, which emerged as a negative correlation. This finding is well-established in ART literature and underscores the paramount influence of ovarian reserve and embryonic aneuploidy rates on success, factors that may outweigh the impact of modifiable endometrial factors like progesterone levels in many cases [4].

An unexpected finding was that the supplemented group (Group II) had the lowest mean progesterone level three days post-transfer (7.72 ng/mL), despite having higher levels than Group I on the day of transfer. High-dose supplementation may cause downregulation of endogenous production or a receptor saturation ceiling effect, limiting further increases in circulating progesterone [26]. Variable absorption of vaginal progesterone and rapid metabolic clearance may also contribute, as individual differences in mucosa permeability or enzyme activity can reduce systemic levels despite higher dosing [27]. Measurement limitations of serum progesterone assays may underestimate bioactive levels at the endometrium, emphasizing that circulating progesterone does not always reflect local activity [28]. These findings suggest that more progesterone does not necessarily increase systemic levels and should be interpreted with caution, warranting further investigation.

However, the interpretation of our findings must be tempered by the study’s limitations, the most critical being the relatively small sample size. The study may be underpowered to detect small but clinically relevant differences, particularly in outcomes like pregnancy loss, which had a very low event rate (n = 3). The numerical trends observed - whereby Groups I and II, despite lower progesterone, achieved pregnancy rates nearly equivalent to the high-progesterone Control Group [26]- are intriguing. It is plausible that the supplemental progesterone in Group II may have indeed “rescued” the cycle for some women, effectively normalizing their outcomes to match the control group. This hypothesis, that supplementation can mitigate the negative effects of low progesterone, is supported by the work of Labarta et al. (2021) [13]. Our study lacked the statistical power to conclusively prove this equalization effect, but the pattern in the data is suggestive.

Furthermore, the significant difference in progesterone levels three days post-transfer between Group II (supplemented) and Group I (non-supplemented) confirms that the intervention was biologically active and successfully elevated serum levels. Yet, this elevation did not translate into a superior pregnancy rate, raising questions about the precise threshold and timing of progesterone adequacy. It is possible that the window for critical progesterone exposure is earlier or later than the times measured, or that endometrial receptivity is determined by factors beyond serum concentration, such as local endometrial tissue response or progesterone receptor density [7, 10].

In conclusion, although the addition of supplemental luteal support did not yield a statistically superior pregnancy rate, the fact that women with initially low progesterone (Group II) achieved pregnancy outcomes comparable to those with normal progesterone levels (Group III) suggests that personalized progesterone supplementation successfully mitigated the detrimental effects of low serum progesterone, effectively normalizing their reproductive prognosis. The most powerful predictor remains maternal age. However, the numerical trends observed do not entirely negate the concept of individualized luteal support. Instead, they highlight the need for larger, sufficiently powered randomized trials focused on specific high-risk subgroups, such as women with a history of previous failed cycles or unexplained implantation failure, who may derive a more pronounced benefit from personalized progesterone adjustment. Future research should also aim to define more precise, evidence-based thresholds for progesterone supplementation and explore the dynamics of progesterone absorption and metabolism throughout the entire luteal phase.

Limitations

The relatively small sample size may have limited the power to detect small but clinically relevant differences in pregnancy outcomes. In addition, as in other studies, heterogeneity in embryo quality and patient characteristics may have contributed to outcome variability despite statistical adjustment. Future larger randomized controlled trials are warranted to further clarify whether there are subtle benefits of intensified luteal support beyond equalization of outcomes.

Authors’ contributions

Maryam Azizi Kutenaee participated in the study design and provided laboratory facilities and equipment through the university.Sonia Falah Khorsand conducted the experimental research and data collection.Farzaneh Fesahat supervised and monitored all stages of the research process.Minoo Vahedi Raad contributed to manuscript writing and performed the statistical analyses.Fatemeh Afshari contributed in sample collection and preparation.Ensieh Salehi participated in the study design and provided overall guidance, laboratory facilities, and equipment support.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data that supports the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Hormozgan University of Medical Sciences (Approval No. IR.HUMS.REC.1402.394). Written informed consent was obtained from all participants prior to enrollment in the study.

Consent for publication

All authors have reviewed the final version of the manuscript and consent to its publication. Written informed consent was obtained from all participants for the publication of their personal and clinical details, including any identifying images, in this study.

Competing interests

The authors declare no competing interests.