Mid-pregnancy circulating placental growth factor (PlGF) and obstetrical outcomes following nonovulatory and ovulatory frozen embryo transfer cycles

Abstract

Purpose

We compared serum placental growth factor (PlGF) concentration in pregnancies from ovulatory vs. non-ovulatory FET cycles and evaluated associations with obstetrical and neonatal outcomes.

Methods

We conducted a secondary analysis of a prospective single-center screening study at a large urban center in Canada where PlGF was performed between April 2020 and December 2022. Baseline characteristics were compared between FET ovulatory cycle and non-ovulatory cycle groups. We compared PlGF concentration in unadjusted and adjusted analyses using multivariable linear regression. Secondary outcomes were compared using chi square and Fisher exact tests for categorical variables and Wilcoxon rank sum tests for continuous variables.

Results

We identified 340 patients, 246 (72.4%) in the non-ovulatory group and 94 (27.7%) in the ovulatory group. The median gestational age (IQR) at PlGF testing was 27.9 weeks (26.6–28.3) in the non-ovulatory group and 27.7 weeks (25.6–28.1) in the ovulatory group (p = 0.07). Univariable analysis comparing the difference in PlGF showed no association between groups [β − 51.7 (95% CI − 43.7–384.1), p = 0.44]. After adjusting for maternal age, race, parity, pre-pregnancy BMI, gestational diabetes, gestational age of PlGF test, donor egg cycle, and maternal and previous obstetrical comorbidities, the difference in PlGF concentration remained non-significant. Obstetrical and neonatal outcomes were similar between groups.

Conclusion

Our findings indicate no significant difference in either mid-pregnancy circulating PlGF concentration or obstetrical outcomes comparing successful FETs in ovulatory vs. non-ovulatory cycles. Larger-scale prospective research should be prioritized to further elucidate the role of PlGF in predicting the potential risk of adverse pregnancy outcomes.

Introduction

Frozen embryo transfer (FET), after in vitro fertilization (IVF), has become the predominant method of embryo transfer in assisted reproductive technology (ART), comprising approximately 79% of transfers in the USA in 2019 [1]. The widespread adoption of FETs has been driven by advancements in cryopreservation techniques, the increasing use of preimplantation genetic testing, elective “freeze-all” protocols, and the need to mitigate ovarian hyperstimulation syndrome risk. The key to success in FET is synchronizing the endometrium with the developmental stage of the embryo to facilitate implantation [2, 3]. Common FET endometrial preparation protocols include non-ovulatory (also known as hormone-orchestrated or programmed cycle) or ovulatory cycle (also known as a natural cycle). While clinical pregnancy rates appear comparable between these protocols, emerging evidence suggests that non-ovulatory FET cycles may be associated with adverse obstetrical and neonatal outcomes [3, 4]. These complications include a higher risk of early pregnancy loss, hypertensive disorders of pregnancy (HDP), gestational diabetes (GDM), higher birthweights, and more postpartum hemorrhage (PPH) in non-ovulatory FET cycles compared to ovulatory cycles [4–6].

The absence of the corpus luteum in non-ovulatory cycles is postulated to be responsible for these poorer outcomes since the pharmacologic suppression of a dominant follicle destined to ovulate prevents its transformation into a corpus luteum [5]. In ovulatory cycles, the resulting corpus luteum can provide circulating progesterone and additional hormonal and cytokine support, which are considered important for successful embryo implantation and early placental development [7]. Without the corpus luteum and its crucial angiogenic proteins, early placental angiogenesis could be disrupted and manifest later as maternal vascular malperfusion (MVM) disease of the placenta [8].

MVM is the most common placental pathology mediating adverse maternal (preeclampsia) and fetal (growth restriction, stillbirth, abruption) outcomes. The impaired placental development, characterized initially by insufficient extravillous trophoblast invasion of the uterine wall, results in reduced utero-placental vascular remodeling and chronic placental hypoperfusion, causing damage to the syncytiotrophoblast surface of the placental villi that secrete PlGF into maternal blood [9]. MVM-associated obstetric complications are more commonly observed in ART than in spontaneous pregnancies [10–13]. Research has demonstrated that the circulating concentration of angiogenic PlGF is depressed prior to clinical signs of these complications arising, prompting multiple studies to investigate PlGF as a predictive risk marker [14–17].

The SCOPE (Screening for Pregnancy Endpoints) study assessed the role of PlGF-1 levels measured at 14–16 weeks’ gestation in predicting preterm preeclampsia in low-risk nulliparous, singleton pregnancies. This study found that serum PlGF-1 measured at 14–16 weeks gestational age improved the identification of those at risk. This was demonstrated by an improvement in the area under a receiver operating characteristic curve (AUC) of 0.76 (95% CI 0.67–0.84) using previously reported clinical risk variables, to 0.84 (95% CI 0.77–0.91) after the addition of PlGF [18]. Additionally, serum PlGF at 14–16 weeks gestational age was significantly lower in concentration in those who delivered preterm with preeclampsia compared to normal pregnancies (preeclampsia: median 642 ng/mL (IQR 347–1252) vs. no preeclampsia: 977 ng/mL (IQR 545–1649), p < 0.001) [19]. The AUC for PlGF alone was 0.61 (95% CI 0.58–0.65). Furthermore, prior research has demonstrated altered angiogenic marker profiles in ART pregnancies compared to spontaneous conceptions [7, 20, 21]. These included soluble fms-like tyrosine kinase (sFLT-1) and PlGF, with ART pregnancies showing an anti-angiogenic profile with elevated sFLT-1 and decreased PlGF at multiple timepoints throughout gestation compared to spontaneous pregnancies. However, no study has directly compared PlGF concentrations between ovulatory and non-ovulatory FETs or their association with perinatal outcomes.

Given prior literature linking non-ovulatory FET cycles to abnormal placentation, we aimed to compare second-trimester serum PlGF concentrations between individuals who underwent successful ovulatory and non-ovulatory FET cycles [4–6]. Our objective was to assess differences in PlGF concentrations and their association with adverse obstetrical and neonatal outcomes, potentially mediated by underlying MVM of the placenta. We hypothesized that non-ovulatory FET cycles would be associated with lower PlGF concentrations and a higher incidence of adverse perinatal outcomes compared to ovulatory cycles, due to the absence of a functional corpus luteum during implantation and early placental development.

Material and methods

Study design

We conducted a single-center, nested cohort study at a tertiary maternal referral center in Toronto, Canada, including patients from April 2020 to December 2022. Institutional research ethics board approval was obtained for this study (REB #20–0086-E-E). Participants and data were identified from a published prospective cohort study that included pregnant people aged 18 years or older with viable, singleton pregnancies under the care of obstetricians, family physicians, or midwives at Mount Sinai Hospital [11]. All patients receiving care at this institute who fulfilled eligibility criteria during this timeframe were included as part of an opt-out protocol, with the addition of PlGF testing (Elecsys, Roche Diagnostics; assay reference 05144671–190) added to routine bloodwork assessing for gestational diabetes at 24–28 weeks gestational age.

Participants were those who conceived through FET at our institution-affiliated fertility clinic. Multi-fetal pregnancies were excluded. Patients were categorized into two groups based on their FET endometrial preparation protocol: ovulatory FET, where cycles were monitored for endogenous ovulation, and non-ovulatory FET, where hormone-orchestrated cycles involved the administration of exogenous estrogen and progesterone.

Procedure protocols

Patients undergoing a non-ovulatory FET protocol were given Estrace 4 mg twice daily starting Day 2. A transvaginal ultrasound on cycle Days 13–15 was performed to assess endometrial thickness and to confirm that no dominant follicle had developed. The optimal endometrial thickness was defined as ≥ 8 mm. Additional transdermal estradiol was used as needed, and patients were brought back every 5–7 days for assessment. Patients were started on luteal support consisting of either Prometrium (Organon), Endometrin (Ferring), or intramuscular progesterone (Hikma), or a combination of these once optimal endometrial thickness was achieved. Choice of luteal support was based on individual patient characteristics and physician discretion. A frozen embryo transfer was performed on the 6th day of progesterone exposure under ultrasound guidance. A pregnancy test was performed 9 days after blastocyst embryo transfer, and an ultrasound was performed to confirm intrauterine pregnancy at 6.5–7.5 weeks gestation. Progesterone and estrogen supplementation was continued until 10 weeks gestational age.

For patients undergoing an ovulatory FET protocol, this included either a natural cycle with or without use of letrozole. If letrozole was used, 5 mg from cycle Days 3 to 7 was administered. The patient was followed up on cycle Day 12 with bloodwork and ultrasound. When the endometrial thickness was ≥ 7 mm with presence of a dominant follicle > 18 mm, hCG was administered (Ovidrel 250 mcg IM) and progesterone was started the following day. A frozen embryo transfer was performed as described above on the 6th day of progesterone exposure, which was continued until 8 weeks gestational age.

Eligible participants underwent PlGF serum testing at the time of gestational diabetes screening, generally at 24 to 28 weeks’ gestation, although participants tested within 2 weeks of this range (22 to 30 weeks) were also included if they did not have a PlGF value drawn from 24 to 28 weeks. Daily PlGF testing occurred in the hospital central laboratories (Elecsys, Roche Diagnostics). Results were posted to the hospital electronic medical record (EMR) within 2 h of sampling.

Definition of outcomes

The primary outcome of the study was second-trimester serum PlGF concentration, measured between 24–28 weeks gestation. For patients with multiple PlGF tests in pregnancy, the earliest available PlGF test between 24 and 28 weeks gestation was used for analysis. Secondary outcomes included obstetrical and neonatal outcomes. Spontaneous preterm birth was defined as birth at either less than 37 weeks or less than 34 weeks with labor and/or spontaneous rupture of membranes. Hypertensive disorders of pregnancy (HDP) included gestational hypertension, preeclampsia, or severe preeclampsia as defined by the Society of Obstetricians and Gynaecologists of Canada criteria [22]. The diagnosis of gestational diabetes adhered to the Diabetes Canada Clinical Practice Guideline and included hemoglobin A1c (HbA1c) greater than 5.8% (to convert to proportion of total hemoglobin, multiply by 0.01), which temporarily replaced glucose challenge testing at our institution during the COVID-19 pandemic [23]. Stillbirth was defined as the loss of a clinical pregnancy > 20 weeks. Neonatal death was defined as the death of a live born infant, regardless of gestational age at birth, within the first 28 completed days of life.

Statistical analysis

Maternal demographic characteristics, obstetrical history, and pregnancy outcomes were extracted from electronic medical records. Continuous variables were reported as mean with standard deviation or median with interquartile range (IQR). Group comparisons were conducted using t-tests or Kruskal–Wallis tests, depending on data distribution. Categorical variables were summarized as percentages, with comparisons made using chi-squared tests. The primary outcome of PlGF concentration at 24–28 weeks was compared with t-tests (unadjusted) and using multivariable regression to adjust for potential confounders identified through univariable analysis and clinical relevance. All analyses were performed using STATA software (version 18.0).

Results

A total of 340 patients met inclusion criteria, with 246 (72.4%) undergoing non-ovulatory FET and 94 (27.7%) undergoing ovulatory FET. Within the ovulatory FET group, 12 patients (12.8%) underwent a cycle with Letrozole and 82 (87.2%) without. Baseline characteristics, including maternal age at egg retrieval, BMI, antral follicle count, and AMH levels, were comparable between groups (Table 1). There were no significant differences in the distribution of race, parity, or infertility diagnoses. There were significant differences in donor cycles between groups, with a higher proportion of donor cycles in the non-ovulatory group.

Table 1.

Demographics of patients in non-ovulatory vs ovulatory FET groups (N = 340)

Characteristic	Non-ovulatory FET (%)	Ovulatory FET (%)	p-valuec
N	246 (72.4)	94 (27.7)
Age at egg retrieval  ≤ 30 31–37 38–44	31 (12.6) 158 (64.2) 57 (23.2)	5 (5.3) 61 (64.9) 28 (29.8)	0.10
Age, median (IQR)	35 (33–37)	35.5 (32–38)	0.44
Gravida 1 2 3 4  ≥ 5	108 (43.9) 63 (25.6) 40 (16.3) 22 (8.9) 13 (5.3)	39 (41.4) 27 (28.7) 10 (10.6) 10 (10.6) 8 (8.5)	0.53
Parity 0  ≥ 1	164 (66.7) 82 (33.3)	68 (72.3) 26 (27.7)	0.32
Maternal race White Black Asian or South Asian Other Unknown	43 (17.5) 3 (1.2) 15 (6.1) 127 (51.6) 58 (23.6)	15 (16.0) 2 (2.1) 9 (9.6) 37 (39.4) 31 (33.0)	0.20
Diagnosis Unexplained Male factor Tubal factor DOR RPL Uterine factor Anovulation/PCOS Mixed Other	50 (20.3) 37 (15.0) 15 (6.1) 23 (9.4) 10 (4.1) 2 (0.8) 12 (4.9) 80 (32.5) 17 (6.9)	22 (23.4) 14 (14.9) 8 (8.5) 8 (8.5) 4 (4.3) 0 5 (5.3) 26 (27.7) 7 (7.5)	0.97
BMI, median (IQR)a	23.6 (21.1–26.6)	23.8 (21.1–26.6)	0.89
AFC, median (IQR)	18 (12.0–25.5)	18.5 (13.0–26.0)	0.54
AMH (pmol/L), median (IQR)a	21 (12.8–34.0)	23 (14.0–32.0)	0.58
Regular cycles q21-35d Yes No	195 (79.3) 51 (20.7)	79 (84.0) 15 (16.0)	0.32
Donor cycle Yes No	21 (8.5) 225 (91.5)	2 (2.1) 92 (97.9)	0.04
# embryos transferred 1 2	243 (98.8) 3 (1.2)	92 (97.9) 2 (2.1)	0.53
Day of embryo cryopreservation Day 5 Day 6 Day 7	174 (70.7) 71 (28.9) 1 (0.4)	56 (59.6) 38 (40.4) 0	0.11
Preimplantation genetic testing Untested Euploid Mosaic	128 (52.0) 116 (47.2) 2 (0.8)	41 (43.6) 52 (55.3) 1 (1.1)	0.38
Luteal support medication PV PV + PIO PIO	187 (76.0) 56 (22.8) 3 (1.2)	78 (83.0) 13 (13.8) 3 (3.2)	0.10
Previous medical co-morbidities POI PCOS Stroke VTE Essential HTN Diabetes	7 (2.9) 30 (12.2) 1 (0.4) 0 5 (2.0) 3 (1.2)	0 9 (9.6) 1 (1.1) 1 (1.1) 0 2 (2.1)	0.10 0.50 0.48 0.11 0.16 0.53
Previous obstetrical co-morbidities HDP GDM Placental disorderb Preterm birth	n = 138 6 (4.3) 4 (2.9) 10 (7.2) 7 (5.1)	n = 55 1 (1.8) 2 (3.6) 0 2 (3.6)	0.53 0.60 0.09 1.00

FET frozen embryo transfer, DOR decreased ovarian reserve, RPL recurrent pregnancy loss, PCOS polycystic ovarian syndrome, AFC antral follicle count, AMH anti-mullerian hormone, BMI body mass index, PV per vagina, PIO progesterone in oil, POI premature ovarian insufficiency, VTE venous thromboembolism, HTN hypertension, HDP hypertensive disorder of pregnancy, GDM gestational diabetes mellitus

^aBMI: 2 missing values, AMH: 35 missing values

^bIncludes abruption, placenta accreta, placenta previa, retained placenta or stillbirth

^cWilcoxon rank sum test, Pearson χ² test, or Fisher exact test

The median gestational age at PlGF testing was 27.9 weeks (IQR 26.6–28.3) in the non-ovulatory group and 27.7 weeks (IQR 25.6–28.1) in the ovulatory group (p = 0.07). Median PlGF concentrations were slightly higher in the non-ovulatory group (728 pg/mL, IQR 506–1138) compared to the ovulatory group (691 pg/mL, IQR 477–993) (Table 2). The unadjusted mean difference in PlGF concentration was not statistically significant (β − 51.7, 95% CI − 183.6 to 80.1, p = 0.44). After adjusting for maternal age, race, parity, pre-pregnancy BMI, gestational diabetes, gestational age of PlGF test, donor cycle, maternal comorbidities, and previous obstetrical comorbidities, the mean difference in PlGF concentration was higher in the ovulatory group, but this difference was not statistically significant (β 170.6, 95% CI − 43.7 to 384.1, p = 0.12). A separate model including adjustment for PGT-A confirmed no differences in our findings (β 175.0, 95% CI − 36.8 to 386.9, p = 0.11).

Table 2.

Crude and adjusted coefficients for PlGF concentrations drawn between 24 and 28 weeks GA in non-ovulatory vs ovulatory FET patients (N = 340)

Model	FET group	PlGF level, median (IQR), pg/mLa	B coefficient (95% CI)	p-value
PlGF-1 at 24–28 weeks GA (n = 340)
1	Non-ovulatory Ovulatory	728 (506–1138) 691 (477–993)	0  − 51.7 (− 183.6–80.1)	0.44
2	Non-ovulatory Ovulatory		0 170.6 (− 43.7–384.09)	0.12

Model 1 (univariable): unadjusted

Model 2 (multivariable): adjusted for maternal age, race, parity, pre-pregnancy BMI, gestational diabetes, gestational age of PlGF-1 test, donor cycle, maternal comorbidities and previous obstetrical comorbidities

^aIf no values existed within the 24–28 weeks GA range, the value closest to this range (before or after) was selected

Obstetrical and neonatal outcomes were similar between groups (Table 3). Rates of hypertensive disorders of pregnancy, including gestational hypertension and preeclampsia, did not significantly differ between the non-ovulatory and ovulatory FET groups (30.9% vs. 23.4%, p = 0.17). Preterm birth rates were comparable (6.1% vs. 7.5%, p = 0.65) as were median birth weights (3.4 kg vs. 3.4 kg, p = 0.28). Mode of delivery (p = 0.55) and fetal outcomes (p = 0.38) did not differ.

Table 3.

Obstetrical and neonatal outcomes in non-ovulatory vs ovulatory FET (n = 340)

Outcome	Non-ovulatory FET (%)	Ovulatory FET (%)	p-valueb
N	246 (72.4)	94 (27.7)
Maternal age at delivery, years  ≤ 30 31–37 38–44  ≥ 45	9 (3.7) 131 (53.3) 100 (40.7) 6 (2.4)	1 (1.1) 46 (48.9) 47 (50.0) 0	0.18
Preterm birth	15 (6.1)	7 (7.5)	0.65
Any HDP	76 (30.9)	22 (23.4)	0.17
Gestational hypertension	55 (22.4)	19 (20.2)	0.67
Preeclampsia	21 (8.5)	3 (3.2)	0.10
Gestational diabetes	18 (7.4)	7 (7.5)	0.57
Birth mode Spontaneous vaginal Assisted vaginal Cesarean	88 (35.8) 43 (17.5) 115 (46.8)	39 (41.5) 13 (13.8) 42 (44.7)	0.55
Birthweight (kg), median (IQR)a	3.4 (3.1–3.8)	3.4 (3.0–3.7)	0.28
Birthweight category (kg)a  > 4000 2500 to 4000  < 2500	29 (11.8) 205 (83.3) 12 (4.9)	7 (7.5) 83 (88.3) 4 (4.3)	0.53
Fetal outcome Live birth Stillbirth Neonatal death	244 (99.2) 2 (0.8) 0	94 (100) 0 0	0.38

AGA appropriate for gestational age, FET frozen embryo transfer, HDP hypertensive disorder of pregnancy, LGA large for gestational age, PlGF placental growth factor, SGA small for gestational age

^aBirthweight: 1 missing value

^bWilcoxon rank-sum test, Pearson χ² test, or Fisher exact test

Discussion

Our study aimed to explore the relationship between PlGF concentration in patients undergoing different FET protocols. Overall, our findings found no significant association between FET cycle type and mid-trimester PlGF concentration.

While our findings did not show any differences in secondary maternal and fetal outcomes between groups, recent meta-analyses have indicated that ovulatory FETs may be associated with a lower risk of complications such as preeclampsia and abnormal placentation compared to non-ovulatory FETs [4]. Several hypotheses may explain these findings, particularly regarding the impact of exogenous estrogen and progesterone on the endometrium and placental development. In early pregnancy, progesterone plays a crucial role in initiating decidualization, supporting extravillous trophoblast invasion, and promoting vascular remodeling [24]. Disruptions in hormone concentrations during non-ovulatory cycles may impair these processes, leading to inadequate trophoblast invasion and defective spiral artery remodeling [25].

Another hypothesis centers on the absence of a corpus luteum in non-ovulatory cycles. The corpus luteum is thought to play a vital role in supporting embryo implantation and early pregnancy development [7]. The impact of the corpus luteum on maternal health has been clinically observed in several studies [7, 26]. In a small prospective cohort study, maternal cardiovascular changes were assessed noninvasively from the follicular phase through pregnancy and postpartum [26]. Women who conceived without a corpus luteum using a non-ovulatory FET showed significantly attenuated or absent first trimester increases in cardiac output, left atrial dimension, early diastolic filling velocity, and arterial compliance. In contrast, pregnancies with one or more corpus lutea exhibited normal cardiovascular responses, suggesting that the corpus luteum plays a critical role in maternal hemodynamic adjustments during early pregnancy.

Studies have specifically analyzed the impact of the absence of the corpus luteum on PlGF concentrations [7, 26]. A prospective study analyzed angiogenic markers in three cohorts: women who conceived spontaneously (one corpus luteum), via non-ovulatory FET (no corpus luteum), and via IVF with a fresh embryo transfer after ovarian stimulation (multiple corpus lutea) [7]. The results demonstrated that at 32 to 35 gestational weeks, a higher sFLT-1/PLGF ratio was observed in women without a corpus luteum from a non-ovulatory FET, suggesting more severe placental pathology; however, no clinical obstetrical outcomes were assessed.

Further supporting the potential link between the corpus luteum and PlGF concentration, another prospective study compared serum PlGF concentration between fresh and ovulatory FETs at 6–7 weeks gestational age [27]. Among 211 pregnancies, 126 were achieved with fresh ET and 85 with an ovulatory FET. There were no significant differences in perinatal outcome, pregnancy complications, and PlGF concentration between the groups. This suggests that when there is a presence of the corpus lutea, regardless of the fresh or frozen nature of the embryo, PlGF concentrations remain similar. Similarly, an additional study showed that while there was no difference in sFlt-1/PlGF ratios at 5 weeks gestational age between fresh and frozen transfers, these ratios were significantly higher in small-for-gestational age infants, suggesting an imbalance in angiogenic markers during placentation [20]. However, the authors failed to report the FET protocol type used, making it difficult to draw conclusions regarding the role of the corpus luteum in this study.

A retrospective study examined PlGF concentrations in single FET pregnancies, comparing those with pre-implantation genetic testing for aneuploidy (PGT-A) to untested embryos [28]. Since PGT-A involves a trophectoderm biopsy, it has been hypothesized that this procedure may be associated with an increased risk of abnormal placental development, which may subsequently lead to lower circulating levels of PlGF. Their results showed that while maternal PlGF concentrations were lower in the PGT-A group, the difference did not reach statistical significance. Most FETs were performed using a non-ovulatory cycle, and sensitivity analysis adjusting for FET protocol showed no substantial impact on the PlGF difference. In our study, use of PGT-A was statistically similar between FET protocol groups, and adjustment for PGT-A confirmed no differences in our findings.

Our study is the first to evaluate mid-trimester PlGF between ovulatory and non-ovulatory FET cycles. We were able to reliably identify FET cycle components and included patients who had PlGF concentrations measured routinely as part of an opt-out protocol, potentially minimizing selection effects. It nonetheless has some limitations. This was a small exploratory study, and our sample size may have limited the ability to detect smaller effects or associations. Based on post hoc power analysis, we had 80% power to detect a difference of approximately 189 pg/mL in PlGF, suggesting we were underpowered to identify more modest but potentially meaningful differences. Additionally, our study was not powered to evaluate clinical obstetrical outcomes, such as preterm birth and fetal growth restriction, which would be important in future research in this area.

Another important limitation is the low number of abnormal PlGF values in our cohort. Although we assessed a group that may have been higher risk, especially those undergoing non-ovulatory FET, only two patients had PlGF < 100 pg/mL (both in the non-ovulatory group); this is the threshold below which medium-to-high risk of preterm delivery for preeclampsia is indicated clinically [29, 30]. This restricted variability limits our ability to compare PlGF across groups or link it to outcomes like preeclampsia. Although our findings may reflect a lower-risk ART population in terms of PlGF levels, the overall rate of any HDP in our cohort was 28.8%, including 21.8% with gestational hypertension and 7.1% with preeclampsia. These rates are markedly higher than those reported in the general obstetric population in Ontario between 2017 and 2021, where HDP occurred in 6.9% of pregnancies, gestational hypertension in 4.1%, and preeclampsia in 2.2% [31]. An alternative explanation is that PlGF measured at 24–28 weeks may not be sensitive enough to detect subtle placental dysfunction in this population. Moreover, because our study assessed PlGF at a single time point, we were unable to evaluate its trajectory across gestation, which may have offered more nuanced insight into its role in placental development.

Our study adds to the growing body of literature examining placental health and the role of PlGF in ART pregnancies. Future research with larger cohorts is needed to determine whether PlGF can be utilized as a reliable biomarker to predict obstetrical complications in pregnancies conceived using different FET protocols.

Conclusion

In summary, we present the first comparison of mid-trimester PlGF levels between ovulatory and non-ovulatory FET cycles. While we did not find a significant association between FET protocol type and PlGF concentration, our findings suggest a possible trend toward higher PlGF concentration in ovulatory FET pregnancies after adjusting for confounders. Given the well-documented differences in maternal cardiovascular adaptation and placental development between these protocols, further research is needed to explore the clinical significance of angiogenic markers in ART pregnancies. Larger prospective studies evaluating PlGF trajectories throughout gestation and their relationship to obstetrical outcomes will be essential to determine the potential utility of PlGF as a biomarker for pregnancy complications in FET-conceived pregnancies.

Author contribution

TD: conceptualization, data curation, formal analysis, methodology, project administration, writing—original draft, manuscript review and editing. NSD: data curation, manuscript review and editing. KM: conceptualization, data curation, methodology, manuscript review and editing. JCK: conceptualization, data curation, manuscript review and editing. EMG: conceptualization, data curation, methodology, manuscript review and editing. SD: conceptualization, manuscript review and editing. JWS: conceptualization, data curation, methodology, supervision, writing—review and editing.

Data availability

Data regarding any of the subjects in the study has not been previously published unless specified. Data will be made available to the editors of the journal for review or query upon request. The appropriate checklist for this study design was followed (STROBE Checklist).

Declarations

Ethical approval

Institutional research ethics board approval was obtained for this study (REB #20–0086-E-E).

Conflict of interest

KM, JCK, and JWS have received research grant funding from Roche Diagnostics. TD, NSD, EMG, and SD declare no competing interests.