Abstract
Purpose
Polycystic ovary syndrome (PCOS) is a common endocrine disorder associated with adverse reproductive and obstetric outcomes, particularly in pregnancies achieved through assisted reproductive technologies. This study aimed to compare obstetric outcomes among programmed cycle-frozen embryo transfer (PC-FET), modified natural cycle-frozen embryo transfer (mNC-FET), and natural cycle-frozen embryo transfer (NC-FET) in women with PCOS using a nationwide Korean health claims database.
Methods
This population-based cohort study utilized data from the National Health Insurance Service–National Health Information Database of South Korea. A retrospective analysis was conducted on 3,873 PCOS patients with singleton pregnancies following FET between October 2017 and December 2021. Patients were categorized into PC-FET, mNC-FET, and NC-FET groups based on prescription records. Obstetric outcomes were analyzed using logistic regression models adjusted for age and medical comorbidities.
Results
mN-FET was associated with significantly lower rates of miscarriage (23.9%) compared to PC-FET (31.3%) and NC-FET (32.0%), as well as a higher term birth rate (69.6% vs. 61.7% and 60.2%, respectively). mNC-FET also showed significantly lower rates of hypertensive disorders of pregnancy, gestational hypertension, cesarean section, and emergency cesarean section compared to PC-FET. No significant differences were observed between mNC-FET and NC-FET in most obstetric complications.
Conclusion
mNC-FET was associated with favorable obstetric outcomes compared to PC-FET and NC-FET in women with PCOS. These findings suggest that mNC-FET may be an effective and safer endometrial preparation strategy in this population. Further prospective studies are needed to validate these results and optimize FET protocols based on individual patient characteristics.
Supplementary Information
The online version contains supplementary material available at 10.1007/s10815-025-03523-4.
Keywords: Polycystic ovarian syndrome, Frozen embryo transfer, Programmed cycle, Modified natural cycle, Natural cycle, Obstetric outcomes, National cohort
Background
Polycystic ovary syndrome (PCOS) is an endocrine and reproductive disorder affecting 7 to 15% of women of reproductive age [1]. It is a leading cause of hyperandrogenism and oligo-anovulation, often contributing to infertility [2]. Moreover, PCOS is frequently associated with various clinical and metabolic disorders, further complicating reproductive health [3, 4].
Women with PCOS are also at an increased risk of obstetric complications, particularly in pregnancies achieved through assisted reproductive technologies (ART) [5, 6]. Research have demonstrated that PCOS patients face higher incidences of pregnancy-induced hypertension (PIH), preeclampsia (PE), gestational diabetes mellitus (GDM), spontaneous preterm birth (PTB), and cesarean sections (CS). These complications are largely driven by common PCOS-associated factors, such as insulin resistance, obesity, and hyperandrogenism. In addition to maternal complications, these pregnancies are at greater risk of adverse neonatal outcomes, including prematurity and fetal growth restrictions, which often result in higher rates of neonatal intensive care unit admissions. These risks underscore the need for meticulous management and monitoring during pregnancy, especially in women undergoing ART [8].
Given the increased risk of ovarian hyperstimulation syndrome (OHSS) during ovarian stimulation in women with PCOS, the freeze-all method is frequently employed to mitigate this risk [9]. Additionally, since PCOS is characterized by anovulation, embryo transfer typically requires endometrial preparation, with programmed cycles (PC) or modified natural cycles (mNC) being the most commonly used methods. PC offers several advantages, such as easier scheduling and fewer hospital visits; however, national cohort studies have shown an association with obstetric complications, including hypertensive disorders of pregnancy (HDP), placenta previa, and placenta accreta, in the general population [10, 11]. Conversely, SC has been reported to result in fewer of these risks in the general population, while also lowering pregnancy loss rates and improving term birth rates [14].
Although women with PCOS are known to have higher risks of adverse obstetric outcomes compared to non-PCOS patients, there is a lack of studies investigating the obstetric outcomes of ART, particularly frozen embryo transfer (FET), in PCOS patients. Therefore, this study aims to utilize a national cohort database to compare the obstetric outcomes of PC, mNC and NC in PCOS patients undergoing FET, with the goal of determining the optimal method for FET in this population.
Methods
Data source
We conducted this nationwide cohort study using data from the National Health Insurance Service–National Health Information Database of South Korea, an organization operating under the Ministry for Health, Welfare, and Family Affairs in the Republic of Korea. The NHIS serves as the primary provider of medical insurance, covering approximately 97% of the Korean population [15]. To ensure accurate reimbursement processes, healthcare utilization data from all medical institutions are systematically submitted and archived in the NHIS database. This comprehensive database encompasses both inpatient and outpatient records, classified according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), as well as claims related to healthcare services and pharmaceutical prescriptions, coded by the Health Insurance Review and Assessment Service of Korea.
The dataset for this study was sourced from the NHIS research database (NHIS-2023–1616). Since October 2017, infertility treatments have been included under Korea’s national health insurance coverage. Accordingly, this study utilized NHIS claims records spanning from October 2017 to December 2021.
Ethical considerations
Individual identifiers in the Korean NHIS database were anonymized. The study was approved by the official assessment panel of the Korean Government and the Institutional Review Board of Pusan National University Hospital (IRB 2304–015–126) in accordance with their guidelines. Because the data were de-identified, the Institutional Review Board waived the need for informed consent.
Study participants
Figure 1 illustrates a flowchart of the study participants. This study focused on women with singleton pregnancies achieved via FET, using data from the National Health Insurance Service (NHIS). Patients were excluded from the study if they lacked childbirth or abortion codes within 360 days (loss to follow-up), had undergone artificial insemination within the previous 90 days (washout period), or had missing age information.
Fig. 1.
Flow diagram showing the distribution of the study population. Abbreviations: FET, frozen embryo transfer; PCOS, polycystic ovarian syndrome; PC-FET, programmed cycle-frozen embryo transfer; mNC-FET, modified natural cycle-frozen embryo transfer; NC-FET, natural cycle-frozen embryo transfer
Patients diagnosed with PCOS were identified based on the presence of the diagnostic code E28.2 within one year of confirmed pregnancy. These PCOS patients were further stratified into three groups based on their prescription records as follows:
PC-FET group: Patients prescribed estradiol (estradiol valerate) within 30 days prior to embryo transfer (ET).
mNC-FET group: Patients prescribed one or more of the following within 30 days prior to ET: human chorionic gonadotropin, clomiphene, letrozole, follicle-stimulating hormone, or human menopausal gonadotropin.
NC-FET group: Patients who did not receive any prescriptions within 30 days prior to ET.
Outcome definitions and confounder variables
This study evaluated the prevalence of pregnancy, obstetric outcomes utilizing ICD-10 diagnostic and procedure codes. Pregnancy outcomes were classified as abortion (O00-06), PTB (Z3700–3701, Z3790–3791, O6010–6011, O6030–6031, O4200-4201, O4210-4211, O4220-4221, O4290-4291), term birth (Z3702, Z3792, and delivery procedure codes), and stillbirth (Z371). Obstetric and fetal complications were analyzed among singleton pregnant patients who delivered, excluding those who experienced abortion. Obstetric complications encompassed threatened abortion (O200), hypertensive disorders of pregnancy (O11, O13–O15), gestational hypertension (O13), PE (O11, O14–O15), placental insufficiency (O438–O439), placental abruption (O45), placenta previa (O44), placenta accreta (O432), postpartum hemorrhage (PPH) (O72), GDM (O244), preterm labor (O602), preterm premature rupture of membranes (PPROM) (O42), cesarean delivery (CS) (O82), and emergency CS (O821).
The confounding variables measured in this study included patient age at the index date, obstetric history, and medical history within one year prior to the initial pregnancy confirmation. Medical history was assessed using ICD-10 codes for hypertension (I10–I15), diabetes (E10–E14), endometriosis (N80), dyslipidemia (E78), recurrent pregnancy loss (RPL) (N96, O26.2), autoimmune diseases (M05–M06, M32–M35, M45, M79, D56, D89, K50–K51, L40), asthma (J45–J46), hypothyroidism (E02–E03), hyperthyroidism (E05), depressive disorders (F32–F33), anxiety disorders (F40–F41), bipolar disorder (F30–F31), and gravida ≥ 1.
Statistical analysis
Categorical data were compared between groups using the chi-square test, and odds ratios (OR) with 95% confidence intervals (CIs) were calculated through binary logistic regression, adjusted for age and medical history. All statistical analyses were performed using SAS Enterprise Guide version 7.15 (2017, SAS Institute Inc., Cary, NC, USA) and R version 4.3.1 (2023, R Core Team, R Foundation for Statistical Computing, Vienna, Austria). A post hoc power analysis was performed for GDM, to assess the statistical power of detecting group differences given the observed sample sizes and event rates.
Results
This study included 18,322 women who achieved pregnancy following frozen embryo transfer (FET) between October 2017 and December 2021. Among them, 3,873 women were diagnosed with polycystic ovary syndrome (PCOS), of whom 2,227 underwent PC-NET, 790 underwent mNC-FET, and 856 underwent NC-FET. The overall maternal mean age was 33.99 years, with no statistically significant differences among the three groups (P = 0.085). The prevalence of diabetes was highest in the PC-FET group (6.2%), followed by the mNC-FET (3.9%) and NC-FET (3.7%) groups. Post hoc analysis revealed that the prevalence in the PC-FET group was significantly higher than in both the mNC-FET and NC-FET groups (P = 0.005). The prevalence of recurrent pregnancy loss was highest in the NC-FET group (7.8%), followed by the PC-FET (7.0%) and mNC-FET (4.8%) groups, with a statistically significant difference among the groups (P = 0.039) (Table 1).
Table 1.
Characteristics of singleton pregnancies depending on endometrial preparation methods in PCOS patients
| Variables | Overall | PC-FET | mNC-FET | NC-FET | P value |
|---|---|---|---|---|---|
| N = 3873 | N = 2,227 | N = 790 | N = 856 | ||
| No. of cycles in the index year | < 0.001 | ||||
| 2017 | 19 (0.5) | 17 (0.8) | 1 (0.1) | 1 (0.1) | |
| 2018 | 893 (23.1) | 465 (20.9) | 212 (26.8) | 216 (25.2) | |
| 2019 | 1099 (28.4) | 642 (28.8) | 235 (29.7) | 222 (25.9) | |
| 2020 | 1164 (30.1) | 732 (32.9) | 228 (28.9) | 204 (23.8) | |
| 2021 | 698 (18.0) | 371 (16.7) | 114 (14.4) | 213 (24.9) | |
| Maternal age (y) | 0.817 | ||||
| < 30 | 366 (9.5) | 220 (9.9) | 70 (8.9) | 76 (8.9) | |
| 30–34 | 1878 (48.5) | 1084 (48.7) | 383 (48.5) | 411 (48.0) | |
| 35–39 | 1349 (34.8) | 773 (34.7) | 276 (34.9) | 300 (35.0) | |
| ≥ 40 | 280 (7.2) | 150 (6.7) | 61 (7.7) | 69 (8.1) | |
| Maternal mean age | 33.99 (3.66) | 33.88 (3.63) | 34.09 (3.64) | 34.19 (3.74) | 0.085 |
| Medical comorbidities | |||||
| Hypertension | 107 (2.8) | 58 (2.6) | 22 (2.8) | 27 (3.2) | 0.706 |
| Diabetes | 200 (5.2) | 137 (6.2) | 31 (3.9)† | 32 (3.7)† | 0.005 |
| Endometriosis | 245 (6.3) | 123 (5.5) | 58 (7.3) | 64 (7.5) | 0.058 |
| Recurrent pregnancy loss | 260 (6.7) | 155 (7.0) | 38 (4.8)† | 67 (7.8)‡ | 0.039 |
| Dyslipidemia | 453 (11.7) | 268 (12.0) | 74 (9.4) | 111 (13.0) | 0.057 |
| Autoimmune disease | 992 (25.6) | 571 (25.6) | 202 (25.6) | 219 (25.6) | 0.999 |
| Asthma | 264 (6.8) | 143 (6.4) | 62 (7.8) | 59 (6.9) | 0.391 |
| Hypothyroidism | 607 (15.7) | 339 (15.2) | 141 (17.8) | 127 (14.8) | 0.163 |
| Hyperthyroidism | 97 (2.5) | 50 (2.2) | 21 (2.7) | 26 (3.0) | 0.431 |
| Obesity | 10 (0.3) | 6 (0.3) | 1 (0.1) | 3 (0.4) | 0.662 |
| Depression disorders | 75 (1.9) | 36 (1.6) | 18 (2.3) | 21 (2.5) | 0.236 |
| Anxiety disorders | 158 (4.1) | 106 (4.8) | 21 (2.7) † | 31 (3.6) | 0.028 |
| Bipolar disorders | 16 (0.4) | 10 (0.4) | 3 (0.4) | 3 (0.4) | 0.917 |
| Gravida ≥ 1 | 257 (6.6) | 147 (6.6) | 54 (6.8) | 56 (6.5) | 0.967 |
Data are presented as number of patients (%) or the mean (SD)
A P value < 0.05 was considered statistically significant. Post hoc pairwise comparisons were conducted using Bonferroni correction for variables with P < 0.05
† Indicates a significant difference compared with the PC-FET group (P < 0.05)
‡ Indicates a significant difference between the mNC-FET and NC-FET groups (P < 0.05)
Abbreviations: PCOS polycystic ovarian syndrome, PC-FET programmed cycle-frozen embryo transfer, mNC-FET modified natural cycle-frozen embryo transfer, NC-FET natural cycle-frozen embryo tran
Pregnancy outcomes
Using mNC-FET as the reference, both PC-FET and NC-FET were associated with significantly higher risks of abortion (adjusted OR 1.50, 95% CI 1.25–1.80 for PC-FET; adjusted OR 1.62, 95% CI 1.31–2.01 for NC-FET). Similarly, the risk of miscarriage was significantly increased in the PC-FET (adjusted OR 1.49, 95% CI 1.23–1.79) and NC-FET (adjusted OR 1.52, 95% CI 1.22–1.89) groups. Term birth was significantly less likely in both PC-FET (adjusted OR 0.69, 95% CI 0.57–0.82) and NC-FET (adjusted OR 0.65, 95% CI 0.53–0.80) compared with mNC-FET (Table 2).
Table 2.
Pregnancy outcomes by endometrial preparation method in PCOS patients who were pregnant (mNC-FET as Reference)
| Outcomes | mNC-FET | PC-FET | Adjusted OR | NC-FET | Adjusted OR |
|---|---|---|---|---|---|
| N = 790 | N = 2,227 | (PC vs mNC) | N = 856 | (NC vs mNC) | |
| Abortion | 207 (26.3) | 761 (34.2) |
1.50* (1.25–1.80) |
310 (36.2) |
1.62* (1.31–2.01) |
| Ectopic pregnancy | 24 (3.0) | 80 (3.6) |
1.23 (0.77–1.96) |
42 (4.9) |
1.66* (0.99–2.78) |
| Miscarriage | 189 (23.9) | 698 (31.3) |
1.49* (1.23–1.79) |
274 (32.0) |
1.52 (1.22–1.89)* |
| Artificial abortion | 14 (1.8) | 70 (3.1) |
1.75 (0.97–3.13) |
22 (2.6) |
1.45 (0.73–2.86) |
| Preterm birth | 31 (3.9) | 89 (4.0) |
1.03 (0.68–1.57) |
31 (3.6) |
0.93 (0.56–1.55) |
| < 34 weeks | 16 (2.0) | 48 (2.2) |
1.09 (0.61–1.93) |
16 (1.9) |
0.93 (0.46–1.87) |
| < 37 weeks | 15 (1.9) | 41 (1.8) |
0.96 (0.53–1.75) |
15 (1.8) |
0.94 (0.45–1.93) |
| Term birth | 550 (69.6) | 1374 (61.7) |
0.69* (0.57–0.82) |
515 (60.2) |
0.65* (0.53–0.80) |
| Still birth | 2 (0.3) | 3 (0.1) |
0.53 (0.09–3.20) |
0 (0.0) |
< 0.001 (< 0.001– > 999.99) |
Data are presented as number of patients (%) or odds ratio (95% CI)
A P value < 0.05 was considered statistically significant and is indicated by an asterisk (*)
Adjusted OR were adjusted for age and medical comorbidities (hypertension, diabetes, endometriosis, recurrent pregnancy loss, dyslipidemia, autoimmune disease, asthma, hypothyroidism, hyperthyroidism, depression disorders, anxiety disorders, bipolar disorders and gravida ≥ 1)
Abbreviations: PCOS polycystic ovarian syndrome, mNC-FET modified natural cycle-frozen embryo transfer, PC-FET programmed cycle-frozen embryo transfer, NC-FET natural cycle-frozen embryo transfer, PC programmed cycle, mNC modified natural cycle, NC natural cycle, OR odds ratio, CI confidence interval
Obstetric complications
The PC-FET group had a higher rate of obstetric complications compared to the mNC-FET group, including threatened abortion (adjusted OR 1.40, 95% CI 1.13–1.73), HDP (adjusted OR 1.40, 95% CI 1.13–1.73), particularly gestational hypertension (adjusted OR 1.87, 95% CI 1.11–3.15), CS (adjusted OR 1.35, 95% CI 1.08–1.70), and emergency CS (adjusted OR 1.67, 95% CI 1.12–2.49). In contrast, when comparing the NC-FET group with the mNC-FET group no significant differences were observed among the groups in the rates of preeclampsia, placenta previa, placenta accreta, PPH, GDM, preterm labor, PPROM (Table 3).
Table 3.
Obstetric complications by endometrial preparation method in PCOS patients who delivered (mNC-FET as Reference)
| Outcomes | mNC-FET | PC-FET | Adjusted OR | NC-FET | Adjusted OR |
|---|---|---|---|---|---|
| N = 583 | N = 1466 | (PC vs mNC) | N = 546 | (NC vs mNC) | |
| Threatened abortion | 156 (26.8) | 499 (34.0) |
1.40* (1.13–1.73) |
118 (21.6) |
0.76* (0.57–1.00) |
| HDP | 29 (5.0) | 116 (7.9) |
1.64* (1.08–2.50) |
23 (4.2) |
0.84 (0.48–1.47) |
| Gestational HT | 20 (3.4) | 88 (6.0) |
1.87* (1.11–3.15) |
19 (3.5) |
1.01 (0.52–1.97) |
| Preeclampsia | 11 (1.9) | 42 (2.9) |
1.52 (0.77–3.01) |
8 (1.5) |
0.78 (0.31–1.97) |
| Placenta previa | 29 (5.0) | 87 (5.9) |
1.25 (0.81–1.94) |
25 (4.6) |
0.89 (0.51–1.55) |
| Placenta accreta | 6 (1.0) | 13 (0.9) |
0.83 (0.31–2.24) |
2 (0.4) |
0.34 (0.07–1.69) |
| PPH | 44 (7.5) | 127 (8.7) |
1.18 (0.82–1.69) |
33 (6.0) |
0.79 (0.49–1.26) |
| GDM | 99 (17.0) | 210 (14.3) |
0.85 (0.65–1.10) |
82 (15.0) |
0.85 (0.62–1.18) |
| Preterm labor | 169 (29.0) | 425 (29.0) |
0.99 (0.80–1.23) |
170 (31.1) |
1.14 (0.88–1.47) |
| PPROM | 100 (17.2) | 261 (17.8) |
1.03 (0.80–1.34) |
102 (18.7) |
1.12 (0.82–1.52) |
| PPROM < 34 weeks | 14 (2.4) | 47 (3.2) |
1.33 (0.72–2.45) |
14 (2.6) |
1.11 (0.52–2.35) |
| PPROM < 37 weeks | 10 (1.7) | 30 (2.0) |
1.18 (0.57–2.45) |
11 (2.0) |
1.21 (0.51–2.88) |
| CS | 140 (24.0) | 432 (29.5) |
1.35 (1.08–1.70)* |
114 (20.9) |
0.85 (0.64–1.13) |
| Emergency CS | 32 (5.5) | 135 (9.2) |
1.67 (1.12–2.49)* |
39 (7.1) |
1.33 (0.82–2.16) |
Data are presented as number of patients (%) or odds ratio (95% CI)
A P value < 0.05 was considered statistically significant and is indicated by an asterisk (*)
Adjusted OR were adjusted for age and medical comorbidities (hypertension, diabetes, endometriosis, recurrent pregnancy loss, dyslipidemia, autoimmune disease, asthma, hypothyroidism, hyperthyroidism, depression disorders, anxiety disorders, bipolar disorders and gravida ≥ 1)
Abbreviations: PCOS polycystic ovarian syndrome, mNC-FET modified natural cycle-frozen embryo transfer, PC-FET programmed cycle-frozen embryo transfer, NC-FET natural cycle-frozen embryo transfer, PC programmed cycle, mNC modified natural cycle, NC natural cycle, OR odds ratio, HDP hypertensive disorders of pregnancy, HT hypertension, PPH postpartum hemorrhage, GDM gestational diabetes mellitus, PPROM preterm premature rupture of membranes, CS cesarean section, CI confidence interval
Discussion
In this study, mNC-FET was associated with significantly lower rates of miscarriage and higher term birth rate compared to both PC-FET and NC-FET. Additionally, mNC-FET was linked to lower rates of threatened abortion, HDP, gestational hypertension, CS, and emergency CS when compared to PC-FET. Compared to NC-FET, mNC-FET demonstrated comparable obstetric complication rates (Fig. 2).
Fig. 2.
Overview of obstetric outcomes comparing PC-FET and mNC-FET in PCOS patients. A PC-FET versus mNC-FET. B NC-FET versus mNC-FET. Presented as odds ratios with 95% confidence intervals, adjusted for age and medical history. Adjusted OR were adjusted for age and medical comorbidities (hypertension, diabetes, endometriosis, recurrent pregnancy loss, dyslipidemia, autoimmune disease, asthma, hypothyroidism, hyperthyroidism, depression disorders, anxiety disorders, bipolar disorders and gravida ≥ 1). Abbreviations: PC-FET, programmed cycle-frozen embryo transfer; mNC-FET, modified natural cycle-frozen embryo transfer; NC-FET, natural cycle-frozen embryo transfer; PCOS, polycystic ovarian syndrome; AOR, adjusted odds ratio; CI, confidence interval; HDP, hypertensive disorders of pregnancy; HT, hypertension; PPH, postpartum hemorrhage; GDM, gestational diabetes mellitus; PPROM, preterm premature rupture of membranes; CS, cesarean section
When examining previous studies on miscarriage rates, meta-analysis by Zeng M et al. included four studies and reported no significant difference in miscarriage rates between mNC-FET and PC-FET (RR = 0.66; 95% CI = 0.37–1.19) [16]. This finding contrasts with the results of our study, which is likely attributable to variations in the stimulation methods used in mNC-FET. Zeng M et al.’s meta-analysis included studies using human menopausal gonadotropin and letrozole, and when analyzing the two studies that utilized letrozole, the miscarriage rate was significantly lower in mNC-FET (RR = 0.53; 95% CI = 0.40–0.70). Two studies published after this meta-analysis corroborate these findings. Wang X et al. reported a reduced miscarriage rate in mNC-FET cycles using letrozole, and Guan L et al. found similar results in a study focusing on overweight/obese women with PCOS [17, 18]. In contrast, Simon V et al. reported no difference in pregnancy loss rates [19]. However, Simon V’s study involved 255 FET cycles conducted between 2011 and 2017 using gonadotropins. While our study did not specifically investigate the mNC-FET stimulation protocols, we observed a lower miscarriage rate in mNC-FET. When considered alongside previous studies, this suggests the possibility that mNC-FET, particularly in cases using letrozole, may be associated with lower miscarriage rates. However, further large-scale research is needed to confirm this association.
In NC-FET, ovulatory PCOS patients are typically considered suitable candidates. However, in our study, ovulatory status could not be clearly distinguished, and it is possible that some anovulatory patients were inadvertently included in the NC-FET group. This may have resulted in inadequate hormonal support, leading to reduced endometrial receptivity or luteal phase insufficiency, ultimately increasing the risk of miscarriage. Furthermore, luteal phase defects have been reported in PCOS patients and are considered part of the common pathophysiological mechanisms contributing to infertility [20]. Therefore, adequate hormonal support may play a crucial role in maintaining pregnancy in this population.
Beyond the endometrial preparation method, it is plausible that PCOS itself influences the miscarriage rate in FET [21]. In prior studies conducted on all FET patients, including those with PCOS, PC-FET was associated with a higher miscarriage rate compared to mNC-FET. However, the magnitude of this difference was found to be lower than that observed in our study focusing specifically on PCOS patients (OR 1.25; 95% CI, 1.14–1.37 vs. OR 1.49; 95% CI, 1.23–1.79) [14]. This suggests that PCOS itself increases the miscarriage rate during FET in ART compared to women without PCOS.
Regarding HDP, the only study analyzing this issue besides the present study was conducted by Wang X et al., which similarly reported a reduction in HDP in stimulated cycles (OR 0.63; 95% CI, 0.42–0.95) [17]. However, that study did not include a subgroup analysis of the various HDP categories. In contrast, our study observed a significant difference in gestational hypertension, whereas no difference was found in PE. The findings of previous literature regarding miscarriage rates and HDP are generally consistent with those of our study. One potential explanation for these differences between mNC-FET and PC-FET is the presence or absence of the corpus luteum and the associated alterations in serum hormone levels in mNC cycles [17, 22, 23]. In PC-FET, the corpus luteum does not form, and while the luteal phase is supported by exogenous progesterone, NC-FET typically involves the formation of a single corpus luteum. In mNC-FET cycles, one corpus luteum is formed when hCG is used, whereas stimulation with letrozole, clomiphene citrate, or gonadotropins may result in the development of multiple corpus luteum. The corpus luteum is known to secrete various substances such as relaxin and vascular endothelial growh factor beyond progesterone that contribute to implantation and pregnancy maintenance [24, 25]. Previous studies demonstrated that pregnancies without a corpus luteum—such as those achieved via programmed FET cycles—are characterized by the lower levels of circulating relaxin, plasma renin activity, aldosterone, and angiogenic factors compared to those with one or more [27, 28]. Furthermore, another study reported that gestational increases in aortic arterial compliance were attenuated in women who conceived via in vitro fertilization technology in the absence of a corpus luteum [25]. These findings suggest that the absence of the corpus luteum impairs maternal cardiovascular adaptation, potentially contributing to the heightened risk of hypertensive disorders of pregnancy [23, 24, 26]. Additionally, differences in serum hormone levels may play a role. Wang et al. reported that estradiol levels were significantly lower, whereas progesterone levels were higher in mNC-FET compared to PC-FET [17]. These hormonal variations could contribute to differences in pregnancy outcomes between the two groups. Although further research is needed to clarify the exact mechanisms, these findings suggest that the absence of the corpus luteum and changes in serum hormone levels may influence clinical outcomes.
Given that the primary pathophysiologic mechanism of PCOS is insulin resistance, the occurrence of GDM in PCOS patients has been a subject of continued interest [29, 30]. The pooled global standardized prevalence of GDM has been reported to be 14.0% [32]. In terms of the occurrence of GDM in PCOS patients, a cohort study conducted in the United States involving 988 participants reported a prevalence of 19%, which was higher compared to the general population [33], however, there is limited information regarding the prevalence of GDM in PCOS patients undergoing ART. Wang et al. reported that the prevalence of GDM was 16.7% in the letrozole-induced mNC-FET group and 20.7% in the PC-FET group, with a statistically significant difference between the two groups (OR 0.71; 95% CI, 0.53–0.93) [17]. In the present study, the prevalence of GDM was 17.0% in the mNC-FET group and 14.3% in the PC-FET group, with no statistically significant difference observed between the two groups. One possible explanation for the discrepancy may lie in the differences in stimulation protocols. While Wang’s study exclusively analyzed letrozole-induced mNC-FET cycles, our mNC-FET group included patients who underwent stimulation with clomiphene citrate, hCG, or gonadotropins in addition to letrozole. These methodological variations may have contributed to the differing outcomes. In addition, to further evaluate whether the lack of statistical significance in our findings may be attributable to insufficient power, we performed a post hoc power analysis. Based on the observed event rates and group sizes, the calculated power was approximately 8.4%, suggesting that the study may have been underpowered to detect small differences in GDM risk. Therefore, the possibility of a clinically meaningful difference cannot be entirely excluded and should be interpreted with caution.
Regarding the lower rates of CS and emergency CS, no prior literature specifically addresses this observation. However, considering the increase in HDP observed in the PC-FET group in this study, the higher CS rate in this group may be related to this factor.
To the best of our knowledge, this is the first nationwide, large-scale cohort study encompassing all women diagnosed with PCOS who underwent FET in South Korea, minimizing selection bias. The extensive dataset provides real-world evidence reflecting clinical practice patterns in a large, unselected population. Moreover, the longitudinal nature of the NHIS database allows for the assessment of long-term reproductive and obstetric outcomes. The robustness of the dataset ensures high external validity, making the findings generalizable to populations with similar healthcare systems and insurance-based registries. Additionally, the diagnoses of PCOS and obstetric complications were based on ICD-10 codes recorded by physicians, rather than self-reported data, enhancing diagnostic accuracy.
However, this study also has some limitations due to the inherent constraints of the NHIS claims database. Potential coding inaccuracies may exist, and the inability to extract data on specific mNC-FET methods limits further analysis. The NHIS system does not allow for the extraction of data on individual medications, preventing a detailed evaluation of ovulation induction and stimulation protocols. Another important limitation of this study is the lack of access to body mass index (BMI) data, which precluded adjustment for obesity—a well-established risk factor for adverse pregnancy outcomes in women with PCOS. Although we attempted to identify obesity using ICD-10 diagnostic codes, the observed prevalence was lower than expected. This likely reflects undercoding within the Korean healthcare system, where a diagnosis of obesity is typically recorded only when active treatment, such as pharmacologic or surgical intervention, is initiated. As such, the true prevalence of obesity may have been underestimated in our cohort, and residual confounding due to unmeasured BMI cannot be excluded. Additionally, key biochemical and clinical parameters of PCOS patients, such as hormonal profiles, metabolic markers, ovulatory status and phenotype classification based on the Rotterdam criteria, were unavailable. Embryo-related variables, including embryo quality and the number of embryos transferred, could not be assessed. Furthermore, the NHIS database only includes reimbursed procedures, which precluded the evaluation of preimplantation genetic testing for aneuploidy and certain luteal support medications, such as different progesterone formulations. In addition, we were unable to exclude patients with underlying uterine factor such as septate uterus, unicornuate uterus, or uterine fibroids, as the NHIS database does not provide sufficient anatomical detail to determine the extent of septation or the size and location of fibroids. Moreover, because maternal and neonatal medical records are not directly linked within the NHIS data system, it was not possible to trace long-term offspring outcomes such as birth defects or childhood metabolic disorders. To address this gap, we are currently exploring methods to enable linkage between maternal and neonatal datasets and are planning future studies to evaluate long-term offspring health outcomes in this context.
Despite these limitations, the nationwide, large-scale design of this study and its use of a comprehensive claims database strengthen its validity and relevance. The inclusion of all PCOS patients undergoing FET in South Korea minimizes selection bias and provides valuable insights into clinical outcomes in a real-world setting. As South Korea is considered an ethnically homogeneous country—with over 96% of the population being ethnic Korean [34]— and has a high level of healthcare accessibility, caution is warranted when generalizing the results of this Korean national cohort study to populations with different ethnic backgrounds, healthcare infrastructures, or diagnostic practices for PCOS.
In summary, this nationwide cohort study demonstrated that mNC-FET was associated with significantly lower miscarriage rates and higher term birth rates compared to both PC-FET and NC-FET in women with PCOS. mNC-FET also showed favorable outcomes in terms of HDP, gestational hypertension, CS, and emergency CS compared to PC-FET, while NC-FET exhibited similar obstetric outcomes to mNC-FET. These findings suggest that mNC-FET, particularly when optimized with appropriate stimulation protocols, may offer the most favorable balance between safety and efficacy in PCOS patients undergoing FET. However, the inability to distinguish ovulatory status or stimulation protocols within the NHIS dataset limits the interpretation of subgroup effects. Further prospective, mechanistic studies are warranted to confirm these observations and guide personalized FET strategies in this population.
Supplementary Information
Below is the link to the electronic supplementary material.
(DOCX 19.6 KB)
Acknowledgements
Every statistical analysis performed in this study was professionally discussed and consulted with the Department of Biostatistics, Biomedical Research Institute, Pusan National University Hospital.
Authors contributions
Conceptualization: Yu EH, Joo KJ. Data curation:Yu EH, Formal analysis: Lee HJ, Methodology: Yu EH, Lee HJ. Visualization: Yu EH. Writing – original draft: Yu EH, Joo JK. Writing – review & editing: Joo KJ, Na YJ.
Funding
This research received no external funding.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethical approval
Individual identifiers within the Korean NHIS database were anonymized. The study procedures were approved by the official assessment panel of the Korean Government and Institutional Review Board (IRB) of Pusan National University Hospital (IRB 2304–015-126), adhering strictly to their guidelines. Since the data were deidentified, the IRB waived the necessity for obtaining informed consent.
Conflict of interest statement
The authors declare no financial and non-financial conflicts of interest.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Collée J, Mawet M, Tebache L, Nisolle M, Brichant G. Polycystic ovarian syndrome and infertility: overview and insights of the putative treatments. Gynecol Endocrinol. 2021;37(10):869–74. [DOI] [PubMed] [Google Scholar]
- 2.Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES, Yildiz BO. The prevalence and features of the polycystic ovary syndrome in an unselected population. J Clin Endocrinol Metab. 2004;89(6):2745–9. [DOI] [PubMed] [Google Scholar]
- 3.Wang S, Alvero R. Racial and ethnic differences in physiology and clinical symptoms of polycystic ovary syndrome. Semin Reprod Med. 2013;31(5):365–9. [DOI] [PubMed] [Google Scholar]
- 4.Dennett CC, Simon J. The role of polycystic ovary syndrome in reproductive and metabolic health: overview and approaches for treatment. Diabetes Spectr Publ Am Diabetes Assoc. 2015;28(2):116–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tandulwadkar SR, Lodha PA, Mangeshikar NT. Obstetric complications in women with IVF conceived pregnancies and polycystic ovarian syndrome. J Hum Reprod Sci. 2014;7(1):13–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Epelboin S, Labrosse J, De Mouzon J, Devaux A, Gervoise-Boyer MJ, Hesters L, et al. Higher risk of pre-eclampsia and other vascular disorders with artificial cycle for frozen-thawed embryo transfer compared to ovulatory cycle or to fresh embryo transfer following in vitro fertilization. Front Endocrinol. 2023;14:1182148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Orosz M, Borics F, Rátonyi D, Tibor Krasznai Z, Vida B, Herman T, et al. Endocrine characteristics and obstetric outcomes of PCOS patients with successful IVF and Non-IVF pregnancies. J Clin Med. 2024;13(18):5602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.D’Alterio MN, Sigilli M, Succu AG, Ghisu V, Laganà AS, Sorrentino F, et al. Pregnancy outcomes in women with polycystic ovarian syndrome. Minerva Obstet Gynecol. 2022;74(1):45–59. [DOI] [PubMed] [Google Scholar]
- 9.Ben Rafael Z. Should we still offer elective freezing of all embryos in all IVF cycles? Hum Reprod. 2020;35(10):2179–84. [DOI] [PubMed]
- 10.Litzky JF, Boulet SL, Esfandiari N, Zhang Y, Kissin DM, Theiler RN, et al. Effect of frozen/thawed embryo transfer on birthweight, macrosomia, and low birthweight rates in US singleton infants. Am J Obstet Gynecol. 2018;218(4):433.e1–433.e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Saito K, Kuwahara A, Ishikawa T, Morisaki N, Miyado M, Miyado K, et al. Endometrial preparation methods for frozen-thawed embryo transfer are associated with altered risks of hypertensive disorders of pregnancy, placenta accreta, and gestational diabetes mellitus. Hum Reprod Oxf Engl. 2019;34(8):1567–75. [DOI] [PubMed] [Google Scholar]
- 12.Waschkies F, Kroning L, Schill T, Chandra A, Schippert C, Töpfer D, et al. Pregnancy outcomes after frozen-thawed embryo transfer in the absence of a corpus luteum. Front Med. 2021;8: 727753. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ginström Ernstad E, Wennerholm UB, Khatibi A, Petzold M, Bergh C. Neonatal and maternal outcome after frozen embryo transfer: increased risks in programmed cycles. Am J Obstet Gynecol. 2019;221(2):126.e1–126.e18. [DOI] [PubMed] [Google Scholar]
- 14.Yu EH, Lee HJ, Lee S, Kim J, Kim SC, Joo JK, et al. Obstetric and perinatal outcomes in 44,118 singleton pregnancies: endometrial preparation methods for frozen-thawed embryo transfer. J Korean Med Sci. 2024;39(45):e282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cheol Seong S, Kim YY, Khang YH, Heon Park J, Kang HJ, Lee H, et al. Data resource profile: the national health information database of the national health insurance service in South Korea. Int J Epidemiol. 2017;46(3):799–800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zeng MF, Zhou X, Duan JL. Stimulated cycle versus artificial cycle for frozen embryo transfer in patients with polycystic ovary syndrome: a meta-analysis. Gynecol Endocrinol. 2021;37(4):294–9. [DOI] [PubMed] [Google Scholar]
- 17.Wang X, Li Y, Tan H, Cai S, Ma S, Peng Y, et al. Letrozole-stimulated endometrial preparation protocol is a superior alternative to hormone replacement treatment for frozen embryo transfer in women with polycystic ovary syndrome, a cohort study. Reprod Biol Endocrinol. 2023;21(1):101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Guan L, Wu H, Wei C, Pang C, Liu D, Yu X, et al. The effect of mildly stimulated cycle versus artificial cycle on pregnancy outcomes in overweight/obese women with PCOS prior to frozen embryo transfer: a retrospective cohort study. BMC Pregnancy Childbirth. 2022;22(1):394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Simon V, Robin G, Decanter C, Dewailly D, Catteau-Jonard S, Barbotin AL, et al. Comparison of two endometrial preparation methods for frozen-thawed embryo transfer in anovulatory PCOS patients: impact on miscarriage rate. J Gynecol Obstet Hum Reprod. 2022;51(6):102399. [DOI] [PubMed] [Google Scholar]
- 20.Boutzios G, Karalaki M, Zapanti E. Common pathophysiological mechanisms involved in luteal phase deficiency and polycystic ovary syndrome: impact on fertility. Endocrine. 2013;43(2):314–7. [DOI] [PubMed] [Google Scholar]
- 21.Sun YF, Zhang J, Xu YM, Cao ZY, Wang YZ, Hao GM, et al. High BMI and insulin resistance are risk factors for spontaneous abortion in patients with polycystic ovary syndrome undergoing assisted reproductive treatment: a systematic review and meta-analysis. Front Endocrinol. 2020;3(11):592495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.von Versen-Hoynck F, Narasimhan P, Selamet Tierney ES, Martinez N, Conrad KP, Baker VL, et al. Absent or excessive corpus luteum number is associated with altered maternal vascular health in early pregnancy. Hypertension. 2019;73(3):680–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.von Versen-Hoynck F, Schaub AM, Chi YY, Chiu KH, Liu J, Lingis M, et al. Increased preeclampsia risk and reduced aortic compliance with in vitro fertilization cycles in the absence of a corpus luteum. Hypertension. 2019;73(3):640–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Singh B, Reschke L, Segars J, Baker VL. Frozen-thawed embryo transfer: the potential importance of the corpus luteum in preventing obstetrical complications. Fertil Steril. 2020;113(2):252–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.von Versen-Hoeynck F, Strauch NK, Fleischmann R, Liu J, Chi Y, Conrad KP, et al. Effect of corpora lutea number on serum electrolytes and creatinine in early pregnancy. Fertil Steril. 2017;108(3):e379–80. [Google Scholar]
- 26.Pereira MM, Mainigi M, Strauss JF. Secretory products of the corpus luteum and preeclampsia. Hum Reprod Update. 2021;27(4):651–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Conrad KP, Graham GM, Chi YY, Zhai X, Li M, Williams RS, et al. Potential influence of the corpus luteum on circulating reproductive and volume regulatory hormones, angiogenic and immunoregulatory factors in pregnant women. Am J Physiol Endocrinol Metab. 2019;317(4):E677–E685. [DOI] [PMC free article] [PubMed]
- 28.von Versen-Höynck F, Strauch NK, Liu J, Chi YY, Keller-Woods M, Conrad KP, et al. Effect of mode of conception on maternal serum relaxin, creatinine, and sodium concentrations in an infertile population. Reprod Sci. 2019;26(3):412–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Bellver J, Rodríguez-Tabernero L, Robles A, Muñoz E, Martínez F, Landeras J, et al. Polycystic ovary syndrome throughout a woman’s life. J Assist Reprod Genet. 2018;35(1):25–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Anagnostis P, Tarlatzis BC, Kauffman RP. Polycystic ovarian syndrome (PCOS): long-term metabolic consequences. Metabolism. 2018;1(86):33–43. [DOI] [PubMed] [Google Scholar]
- 31.Persson S, Elenis E, Turkmen S, Kramer MS, Yong EL, Poromaa IS. Higher risk of type 2 diabetes in women with hyperandrogenic polycystic ovary syndrome. Fertil Steril. 2021;116(3):862–71. [DOI] [PubMed] [Google Scholar]
- 32.Wang H, Li N, Chivese T, Werfalli M, Sun H, Yuen L, et al. IDF Diabetes Atlas: Estimation of Global and Regional Gestational Diabetes Mellitus Prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group’s Criteria. Diabetes Res Clin Pract [Internet]. 2022 Jan 1 [cited 2024 Dec 27];183. Available from: https://www.diabetesresearchclinicalpractice.com/article/S0168-8227%2821%2900409-5/fulltext?utm_source=chatgpt.com. Accessed 2 Jan 2025. [DOI] [PubMed]
- 33.Lo JC, Yang J, Gunderson EP, Hararah MK, Gonzalez JR, Ferrara A. Risk of type 2 diabetes mellitus following gestational diabetes pregnancy in women with polycystic ovary syndrome. J Diabetes Res. 2017;2017(1):5250162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Statistics Korea 2023. Population and Housing Census: Population by nationality [Internet]. Daejeon: Statistics Korea; 2024 [cited 2025 Apr]. Available from: https://kostat.go.kr. Accessed 2 Jan 2025.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 19.6 KB)
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.