Incidence of Retained Products of Conception Following Hormone Replacement Versus Natural Cycles in Frozen Embryo Transfer.

This study was approved by the Ethics Committee of the Gunma University Hospital (approval number: HS2024‐115). Written informed consent was not required owing to the retrospective nature of the study. Instead, an opt‐out consent process was used, as approved by the ethics committee, and relevant information was made publicly available in accordance with the institutional policy.

The authors have nothing to report.

A total of 425 pregnancies were analyzed (HRC, n  = 204; NC, n  = 221). Maternal and baseline characteristics of the HRC and NC groups are summarized in Table  1 . The mean maternal age was 34.8 ± 4.53 years in the HRC group and 34.7 ± 3.74 years in the NC group. A history of miscarriage/abortion was recorded in 83/204 (40.7%) cases in the HRC group and 84/221 (38.0%) in the NC group, and prior miscarriage/abortion surgery was documented in 56/204 (27.5%) and 55/220 (25.0%) cases, respectively. Regarding gynecologic surgical history, uterine surgery (e.g., myomectomy) was noted in 4/204 (2.0%) and 14/221 (6.3%), ovarian surgery in 10/204 (4.9%) and 10/221 (4.5%), and endometriosis surgery in 1/204 (0.5%) and 6/221 (2.7%) cases in the HRC and NC groups, respectively. Cesarean section history was not recorded in either group in this dataset. Detailed distributions are provided in Table  1 . Maternal baseline characteristics of pregnancies conceived by hormone replacement cycle (HRC) and natural ovulation cycle (NC) frozen–thawed embryo transfer. Note: Values are presented as mean ± standard deviation or number (%). For variables with missing data, values are shown as n / N (%), where N indicates the number of available cases. Abbreviations: HRC, hormone replacement cycle; NC, natural ovulation cycle. The primary outcome of this study was the occurrence of RPOC. Overall, RPOC occurred in 14/425 (3.3%) pregnancies. By protocol, the incidence was 11/204 (5.4%) in the HRC group and 3/221 (1.4%) in the NC group (Table  2 ). Regarding pregnancy outcomes, vaginal delivery occurred in 99/204 (48.5%) of the HRC group and 145/221 (65.6%) of the NC group, while cesarean section was performed in 48/204 (23.5%) and 34/221 (15.4%), respectively. Term deliveries (≥ 37 weeks) accounted for 129/204 (63.2%) in the HRC group and 165/221 (74.7%) in the NC group, whereas mid‐trimester loss (≥ 12 to < 22 weeks) was infrequent in both groups (4/204 [2.0%] vs. 4/221 [1.8%]). Pregnancy complications (multiple responses permitted) included hypertensive disorders of pregnancy (12/204 [5.9%] in HRC; 2/221 [0.9%] in NC) and gestational diabetes mellitus (3/204 [1.5%] in HRC; 7/221 [3.2%] in NC). Placental malposition was rarely documented (HRC: 4/203 [2.0%]; NC: 2/221 [0.9%]), with a substantial proportion recorded as unidentified. Detailed distributions are provided in Table  2 . Clinical outcomes and management of RPOC cases, including conservative management (spontaneous expulsion or spontaneous resolution), surgical treatment, and blood transfusion or embolization, are summarized in Table  3 . Among the 11 RPOC cases in the HRC group, 4 (36.4%) were managed conservatively and 7 (63.6%) underwent uterine evacuation. In the NC group, 1 case (33.3%) was managed conservatively, 1 (33.3%) required uterine evacuation, and 1 (33.3%) underwent uterine artery embolization. No blood transfusion or hysterectomy was required in either group. Histopathological confirmation of RPOC was obtained in 6/11 (54.5%) of the HRC cases and 1/3 (33.3%) of the NC cases, for a total of 7/14 (50.0%). Pregnancy outcomes and occurrence of retained products of conception (RPOC) after hormone replacement cycle (HRC) and natural ovulation cycle (NC) frozen–thawed embryo transfer. Note: Values are presented as numbers (%). For variables with missing data, values are shown as n / N (%), where N indicates the number of available cases. Abbreviations: HRC, hormone replacement cycle; NC, natural ovulation cycle; RPOC, retained products of conception. Clinical outcomes and management of pregnancies complicated by retained products of conception (RPOC) in hormone replacement cycle (HRC) and natural ovulation cycle (NC) groups. Note: Values are presented as numbers (%). Conservative management includes spontaneous expulsion or spontaneous resolution. No cases required blood transfusion or hysterectomy. Abbreviations: HRC, hormone replacement cycle; NC, natural ovulation cycle; RPOC, retained products of conception; UAE, uterine artery embolization. Embryo transfer–related parameters are summarized in Table  4 . Use of a high‐concentration hyaluronic acid culture medium was documented in 42/204 (20.6%) cycles in the HRC group and 29/221 (13.1%) in the NC group (total 71/425, 16.7%). Two‐step embryo transfer was performed in 37/204 (18.1%) and 17/221 (7.7%) cycles, respectively (total 54/425, 12.7%). Endometrial scratching was recorded in 55/204 (27.0%) in HRC and 4/221 (1.8%) in NC (total 59/425, 13.9%). Regarding the number of embryos transferred, single‐embryo transfer accounted for 136/204 (66.7%) in HRC and 191/221 (86.4%) in NC (total 327/425, 76.9%), whereas double‐embryo transfer occurred in 68/204 (33.3%) and 30/221 (13.6%), respectively (total 98/425, 23.1%). These data are presented descriptively without inferential testing. Details of estradiol and progesterone supplementation in both HRC and NC protocols, including limited supplementation in a subset of NC cases, are summarized in Table  S1 . Embryo transfer procedures in hormone replacement cycle (HRC) and natural ovulation cycle (NC) groups. Note: Values are presented as a number (%). For variables with missing data, values are shown as n / N (%), where N indicates the number of available cases. Abbreviations: CHM, concentrated hyaluronic acid medium; HRC, hormone replacement cycle; NC, natural ovulation cycle; SET, single embryo transfer. Table  5 presents a gestational‐age–stratified summary of RPOC by transfer protocol (HRC vs. NC) and pooled estimates across strata. Overall, RPOC occurred in 11/204 (5.4%) cases in the HRC group and 3/221 (1.4%) in the NC group (unstratified RR, 3.97). After accounting for gestational‐age strata, the Cochran–Mantel–Haenszel common odds ratio indicated higher odds in the HRC group (common OR 4.47; exact common OR 4.26). A minimally adjusted Firth's penalized logistic model (adjusted for gestational‐age category) yielded an adjusted OR of 3.915. In the 12 ≤ 22 and 22 ≤ 37 week strata, no RPOC events were observed in the NC group (zero cells), as shown in Table  5 . Occurrence of retained products of conception (RPOC) stratified by gestational age at termination of pregnancy in hormone replacement cycle (HRC) and natural ovulation cycle (NC) groups. Note: Exact (conditional logistic): conditional maximum‐likelihood estimate of a common odds ratio using the noncentral hypergeometric distribution, stratified by gestational‐age categories (< 12, 12 ≤ 22, 22 ≤ 37, 37 ≤ weeks); two‐sided exact p ‐value reported. Firth's aOR: adjusted odds ratio from Firth's penalized logistic regression, adjusting for gestational‐age category. Stratum‐specific RR/OR were computed from 2 × 2 tables; when a zero‐cell occurred, the Haldane–Anscombe 0.5 continuity correction was applied (this correction is not used for the exact pooled estimate). CMH (asymptotic)=Cochran–Mantel–Haenszel common odds ratio; Breslow–Day P–test homogeneity across strata. Abbreviations: CI, confidence interval; CMH, Cochran–Mantel–Haenszel; HRC, hormone replacement cycle; NC, natural ovulation cycle; OR, odds ratio; RD, risk difference (HRC‐NC, percentage points); RR, risk ratio. Table  6 dichotomizes pregnancies by gestational age at the end of pregnancy, ≥ 22 weeks (deliveries) and < 22 weeks (miscarriages), and summarizes RPOC risks by protocol. For pregnancies ending at ≥ 22 weeks (deliveries), RPOC occurred in 8/144 (5.56%) cases in the HRC group and 3/177 (1.69%) in the NC group (unstratified RR, 3.28 [95% CI 0.89–12.13]; RD, 3.86 percentage points; Fisher's exact p  = 0.070). For pregnancies ending before 22 weeks (miscarriages), RPOC occurred in 3/60 (5.00%) in the HRC group and 0/40 (0.00%) in the NC group. Owing to a zero cell in the NC group, the RR was not directly estimable; with continuity correction, the RR was 4.70 (95% CI 0.25–88.7; Fisher's exact p  = 0.273). Notably, no RPOC events were observed in the NC group in this < 22‐week stratum. Occurrence of retained products of conception (RPOC) with a 22‐week cutoff: Deliveries (≥ 22 weeks) and miscarriages (< 22 weeks) in hormone replacement cycle (HRC) and natural ovulation cycle (NC) groups. Note: Panel A: Deliveries (≥ 22 weeks). Panel B: Miscarriages (< 22 weeks). Values are presented as numbers (%). Relative risks (RR) and odds ratios (OR) were estimated with 95% confidence intervals using standard large‐sample approximations. When a zero cell occurred, the Haldane–Anscombe 0.5 continuity correction was applied. p ‐values are from Fisher's exact test (two‐sided). Abbreviations: CI, confidence interval; HRC, hormone replacement cycle; NC, natural ovulation cycle; OR, odds ratio; pp, percentage points; RPOC, retained products of conception; RR, relative risk.

In this single‐center, retrospective exploratory cohort, we observed preliminary evidence that HRC was associated with a higher occurrence of RPOC than NC. Overall, RPOC occurred in 14/425 pregnancies (3.3%): by protocol, 11/204 (5.4%) in HRC and 3/221 (1.4%) in NC. After accounting for gestational age at the end of pregnancy, the Cochran–Mantel–Haenszel common odds ratio indicated higher odds in HRC (common OR 4.47, 95% CI 1.29–16.39; exact common OR 4.26, 95% CI 1.11–24.87), and a minimally adjusted Firth's penalized logistic regression yielded an adjusted OR of 3.915 (95% CI 1.18–12.96). Given the small number of events and zero cells in some NC strata, these estimates are imprecise and should be regarded as hypothesis‐generating rather than confirmatory. Beyond incidence, the clinical course of RPOC cases also differed by protocol: in the HRC, 4/11 cases were managed conservatively and 7/11 required uterine evacuation, whereas in the NC, 1/3 of the cases were managed conservatively, 1/3 underwent uterine evacuation, and 1/3 required uterine artery embolization. No blood transfusion or hysterectomy was needed in either group, and histopathological confirmation of RPOC was obtained in 6/11 HRC cases and 1/3 NC cases (50% overall) (Table  3 ). These observations are directionally consistent with prior reports indicating an increased risk of RPOC in ART pregnancies—particularly in FET cycles employing exogenous hormonal support [ 20 ]. Related literature linking hormone‐replacement FET to abnormalities of placental attachment (e.g., non‐previa PAS) provides a plausible context [ 11 , 15 ], although those studies did not primarily evaluate RPOC incidence. Prior work varies in diagnostic criteria, imaging thresholds, and follow‐up intensity, and few studies have directly contrasted HRC with NC. With this heterogeneous background, our cohort contributes protocol‐specific, gestational‐age–aware estimates using exact and Firth‐adjusted analyses, while still warranting cautious interpretation due to sparse data. Several biological pathways may plausibly underlie the observed pattern. In HRC, the absence of a functional corpus luteum means that endometrial receptivity and early placentation are driven by exogenous estradiol and progesterone rather than a full corpus luteum–derived endocrine milieu. Corpus luteum factors—including progesterone, estrogen, and vasoactive peptides such as relaxin—have been implicated in decidualization, immune tolerance, and vascular remodeling at the maternal–fetal interface; their deficiency may impair orderly placentation and subsequent placental separation [ 21 , 22 , 23 , 24 ]. Although not directly evaluating RPOC, a recent large cohort reported higher risks of hypertensive disorders, low birth weight, and preterm birth in HRC compared with NC, suggesting that the absence of corpus luteum–derived endocrine complexity may contribute to adverse pregnancy outcomes [ 25 ]. Exogenous‐hormone programming may also affect endometrial gene expression (e.g., adhesion molecules) and spiral‐artery remodeling, consistent with reports of abnormal placental attachment in hormone‐replacement FET [ 11 , 15 ]. Prolonged exogenous support might delay endometrial involution and dampen myometrial contractility, contributing to incomplete placental expulsion in some cases [ 20 ]. Finally, gestational age at the end of pregnancy itself influences RPOC risk—earlier losses carry higher risk—which may interact with protocol‐related biology [ 18 ]. These mechanistic considerations remain speculative and cannot be confirmed by our dataset, which lacked direct biological measurements; included limited histopathologic verification and imaging standardization; and had few events (including only half of the RPOC cases with histopathological confirmation, as shown in Table  3 ) in a retrospective single‐center design with partial hormonal supplementation in some NC cases (Table  S1 ). In our descriptive analysis, no consistent pattern emerged linking embryo‐transfer–related parameters to RPOC risk. Add‐on procedures at transfer—such as high‐concentration hyaluronic acid medium, two‐step transfer, and endometrial scratching—showed no obvious differences that would explain the observed RPOC occurrence (Table  4 ); however, this study was not designed or powered to evaluate these exposures, no formal hypothesis testing or adjusted modeling was performed for these parameters, and causal inference is not possible. Confounding by indication is plausible (e.g., add‐ons selected for perceived low receptivity), and the number of embryos transferred differed between protocols (single vs. two embryos), further limiting interpretability. With only 14 outcome events, including add‐ons in multivariable models would violate events‐per‐variable considerations even with Firth's correction, and calendar‐time effects could not be addressed in this retrospective design. These observations should therefore be regarded as hypothesis‐generating. Our diagnostic pathway relied primarily on transvaginal ultrasound with grayscale assessment of lesion size and color Doppler evaluation of vascularity, with serum hCG measured when feasible and histopathological confirmation performed when tissue was obtained. When ultrasonographic impressions were uncertain, additional imaging—most commonly contrast‐enhanced MRI—was considered, recognizing that universal MRI was not feasible in our primary‐care setting. We referenced a published power Doppler–based algorithm to contextualize decision‐making [ 7 ] without implementing it verbatim. In practice, lesions with low vascularity (operationally, ≤ 50% of the lesion area showing flow) and clinically stable/minimal bleeding were managed expectantly, whereas progressive or symptomatic cases were considered for procedures such as hysteroscopic removal or, when indicated, uterine artery embolization; actual interventions are reported in the Results. Notably, our NC protocol was not strictly “natural”: when measured on day 5 after ovulation, estradiol ≤ 100 pg/mL or progesterone ≤ 10 ng/mL prompted supplemental support. This real‐world feature, summarized in Table  S1 , highlights that a subset of NC cases received limited hormonal supplementation; however, despite this, the incidence of RPOC in the NC remained substantially lower than in the HRC. From a clinical standpoint, these findings are descriptive and should be regarded as hypothesis‐generating. They may inform neutral counseling by indicating that RPOC occurred more frequently after HRC than after NC in this cohort, but they neither mandate protocol changes nor imply that HRC should be avoided. Determining whether HRC intrinsically increases RPOC risk—beyond confounding and case‐mix differences—will require adequately powered prospective studies (preferably multicenter) with standardized diagnostic criteria and prespecified strategies to control confounding. The strengths of this study include a protocol‐specific comparison within a single center with consistent clinical pathways; clearly defined inclusion and exclusion criteria summarized in a flow diagram; and an analysis plan that prespecified descriptive summaries for background characteristics and focused inference on the primary exposure (HRC vs. NC). This analytic strategy incorporated gestational‐age–aware evaluations (stratification and a 22‐week dichotomy) and small‐sample–appropriate methods (exact inference and Firth's penalized logistic regression). The diagnostic framework was explicitly stated, and real‐world features (e.g., targeted hormonal supplementation in some NC cases) were documented to enhance interpretability. The study's limitations include the retrospective single‐center design; the small number of outcome events with zero cells in several strata yielding imprecise estimates; infeasibility of multivariable adjustment beyond a minimal Firth model without violating events‐per‐variable constraints; absence of hypothesis testing for descriptive background tables; potential confounding by indication, unmeasured confounders, and calendar‐time effects; heterogeneous access to advanced imaging and limited histopathologic verification; and external validity restricted to FET pregnancies in a Japanese single‐institution setting (fresh transfers and other ART modalities not included). These constraints should be considered when interpreting the magnitude and generalizability of the observed associations. In conclusion, in this single‐center retrospective exploratory cohort, we observed a higher occurrence of RPOC after HRC than after NC; these preliminary findings are hypothesis‐generating and warrant confirmation in adequately powered prospective multicenter studies.

All authors have reviewed and approved the final version of the manuscript and consent to its publication.

Retained products of conception (RPOC) refer to placental or fetal tissue that remains within the uterine cavity following pregnancy termination, miscarriage, or delivery [ 1 , 2 ]. RPOC diagnosis typically involves a combination of clinical assessment and imaging. Transvaginal ultrasonography, particularly color Doppler, is widely used and demonstrates a moderate to strong correlation with the presence of RPOC [ 2 ]. However, its diagnostic accuracy is limited, particularly in cases with minimal vascularity or equivocal findings. In such situations, hysteroscopy serves as a valuable adjunct, enabling direct visualization of the intrauterine pathology and facilitating simultaneous diagnosis and treatment [ 3 ]. Although dilation and curettage have traditionally been the standard treatments, hysteroscopic removal has emerged as the preferred approach because of its superior safety profile, efficacy, and reduced risk of intrauterine adhesions [ 4 , 5 ]. Nevertheless, histopathological examination remains essential for definitive diagnosis and to distinguish RPOC from conditions such as endometrial polyps or gestational trophoblastic disease [ 2 , 6 ]. RPOC can lead to complications such as prolonged bleeding, infection, and long‐term fertility issues, making a timely and accurate diagnosis critical [ 7 , 8 ]. Recent studies have identified a multifactorial risk profile for RPOC, highlighting the importance of patient‐ and pregnancy‐related factors. A prior history of RPOC is the strongest predictor (odds ratio [OR] 96.82), followed by multiparity (OR 6.1) and conception via assisted reproductive technology (ART), with reported ORs ranging from 2.65 to 6.0 [ 9 , 10 , 11 ]. Pregnancy‐related complications are also strongly associated with RPOC, including early postpartum hemorrhage (OR 15.84), third‐stage placental complications (OR 12.5), and cesarean section in cases of placenta previa (OR 10.70) [ 12 , 13 , 14 ]. Recent evidence further suggests that pregnancies conceived through hormone replacement cycle–frozen embryo transfer (HRC‐FET) may be associated with an increased risk of abnormal placental attachment, including non‐previa placenta accreta spectrum (PAS), possibly owing to altered endometrial receptivity and decidualization [ 15 ]. Specific placental and fetal characteristics, such as a hypervascular appearance of retained tissue (OR up to 12.8), RPOC length ≥ 4 cm (adjusted OR 8.6), and elevated maternal human chorionic gonadotropin (hCG) levels > 2.0 MoM (OR 4.40), further increase this risk [ 10 , 16 , 17 ]. Additionally, early gestational age at the time of miscarriage or termination has been linked to higher RPOC incidence (OR, 3.53) [ 18 ]. Collectively, these findings highlight the need for tailored monitoring and preventive strategies in high‐risk patients. Since April 2022, infertility treatments, including ART, have been covered by Japan's National Health Insurance system, contributing to the growing number of patients undergoing ART and a shift toward advanced maternal age. According to the Japan Society of Obstetrics and Gynecology, the total number of ART cycles was expected to reach a record high of 543,630 in 2022, resulting in over 70,000 live births [ 19 ]. ART‐conceived births now account for approximately 9% of all live births in Japan, highlighting the clinical and societal relevance of evaluating the maternal and neonatal outcomes following ART. Despite this increasing use of ART, data regarding the relationship between ART pregnancies and RPOC remain limited. Few studies have examined the differences in RPOC incidence based on embryo transfer methods or adjunctive implantation procedures. To address this gap, we conducted a single‐center, retrospective, exploratory study of 426 pregnancies achieved through FET cycles. Specifically, we compared the incidence of RPOC between HRC and natural ovulation cycles (NC) to provide preliminary evidence on potential differences in RPOC occurrence. Our findings are intended to generate hypotheses, support future investigations, and inform clinical counseling in infertility treatment.

The authors declare no conflicts of interest. A.I. is the Editor‐in‐Chief of Reproductive Medicine and Biology and is the co‐author of this article. To avoid any potential conflict of interest and minimize bias, Dr. Iwase was not involved in the editorial review or decision‐making process for this manuscript.

This study was approved by the Ethics Committee of Gunma University Hospital (approval number: HS2024‐115; approval date: September 17, 2024). We conducted a single‐center, retrospective cohort study by reviewing the medical records of patients who conceived using ART between January 2017 and December 2021. A total of 455 pregnancies achieved through FET cycles were initially identified. Inclusion criteria were clinical pregnancies confirmed by the presence of an intrauterine gestational sac. The exclusion criteria were: lack of definitive RPOC diagnosis ( n  = 29), loss to follow‐up before delivery outcome was determined ( n  = 10), and incomplete clinical data ( n  = 1). After exclusions, 425 pregnancies were included in the final analysis (Figure  1 ). Data were collected from medical records to investigate factors associated with RPOC occurrence. Our institution is a primary medical facility with inpatient services offering general infertility treatments, ART procedures, and obstetric care. Study flowchart. A total of 455 clinical pregnancies following frozen–thawed embryo transfer (FET) cycles were assessed. Forty pregnancies were excluded: No definitive diagnosis of retained products of conception (RPOC; n  = 29), pregnancy outcome unknown ( n  = 10), or incomplete clinical data ( n  = 1). The final analysis included 425 pregnancies: 204 conceived in hormone replacement cycles (HRC) and 221 in natural ovulation cycles (NC). In the HRC group, RPOC occurred in 11/204 pregnancies (5.4%) and non‐RPOC in 193/204 (94.6%). In the NC group, RPOC occurred in 3/221 pregnancies (1.4%) and non‐RPOC in 218/221 (98.6%). Detailed protocols for endometrial preparation and luteal support in the HRC and NC groups are shown in Figure  2a,b , respectively. In both protocols, cleavage‐stage embryos were transferred on days 2 or 3, and blastocysts on day 5. Pregnancy was confirmed at 4 weeks and 2 days of gestation using a qualitative urine hCG test. For patients who tested positive, an intrauterine gestational sac was confirmed via transvaginal ultrasound at 5 weeks of gestation. In the HRC group, estradiol and progesterone supplementation were continued until 9 weeks of gestation upon confirmation of pregnancy, whereas in the NC group, only progesterone was administered during this period; however, if serum estradiol was ≤ 100 pg/mL or progesterone ≤ 10 ng/mL on day 5 after ovulation, supplemental estradiol or progesterone was provided. We also examined whether additional hormonal supplementation (estradiol or progesterone) was administered in the NC, and the detailed regimens are shown in Table  S1 . The study population included only patients with confirmed intrauterine gestational sacs and clinical pregnancy diagnoses. (a) Hormone replacement protocol for frozen embryo transfer (HRC cycle). This figure illustrates the hormone replacement cycle (HRC) protocol used for endometrial preparation prior to frozen embryo transfer (FET). Transdermal estradiol patches (0.72 mg) were initiated on day 2 of the menstrual cycle and applied every other day (2–3 patches per application). Oral estrogen (0.5 mg) was administered twice daily starting on day 4. Endometrial thickness was assessed using transvaginal ultrasonography on days 11–13. On day 15 (defined as day 0), progesterone supplementation consisting of oral progesterone (5 mg, six tablets per day) and vaginal progesterone was initiated. The patients were given one of four vaginal progesterone formulations: a tablet (100 mg, administered 2–3 times daily), a capsule (200 mg, administered 2–3 times daily), a suppository (400 mg, administered twice daily), or a gel (90 mg, administered once daily). All hormone treatments were continued until 9 weeks of gestation if pregnancy was achieved. (b) Natural ovulation cycle protocol for frozen embryo transfer (NC cycle). This figure outlines the natural ovulation cycle (NC) protocol used for frozen embryo transfer (FET). Follicular development was monitored via transvaginal ultrasound starting on days 7–10 of the menstrual cycle. In cases of delayed ovulation, an aromatase inhibitor (letrozole, 2.5 mg) was orally administered to induce ovulation. When the dominant follicle reached 18 mm or a positive urinary luteinizing hormone surge was detected, an intramuscular injection of human chorionic gonadotropin (hCG, 5,000 IU) was administered. Ovulation was confirmed 1–2 days later using ultrasonography, and luteal support was initiated with a second hCG injection (5,000 IU) and oral progesterone (5 mg, six tablets per day). Serum estradiol (E2) and progesterone (P) levels were measured on the embryo transfer day. If E2 was < 100 pg/mL, estradiol patches were applied. If P was < 1.0 ng/mL, vaginal progesterone was added using the same options as in the HRC cycle. A third hCG injection (2,000 IU) was administered 5 days after ovulation. After a positive urine hCG test, oral progesterone was continued until 9 weeks of gestation. RPOC was diagnosed via transvaginal ultrasound during routine follow‐up visits after miscarriage or delivery, or when abnormal bleeding was noted. Diagnosis included measurement of the RPOC size and assessment of vascularity using color Doppler. Our diagnostic approach was based primarily on ultrasonographic findings, with reference to the algorithm proposed by Sonehara et al. [ 7 ]. When feasible, serum hCG levels were measured to aid in diagnosis. If the diagnosis was uncertain, additional imaging such as contrast‐enhanced magnetic resonance imaging (MRI) was performed, although this was not feasible in all cases at our primary‐care facility. Cases with low vascularity (≤ 50% of the lesion area on color Doppler) and clinically stable bleeding were generally managed expectantly. In contrast, for progressive or symptomatic cases, our management policy included surgical intervention (hysteroscopic removal) or uterine artery embolization if necessary. In cases where tissue samples were obtained through spontaneous expulsion or uterine evacuation, a histopathological examination was also performed. In addition, information on the clinical outcomes and management of RPOC cases—including conservative management (encompassing spontaneous expulsion or spontaneous resolution), surgical treatment, and the need for blood transfusion or embolization—was collected for analysis. Descriptive statistics were computed to summarize maternal and pregnancy characteristics in the HRC and NC groups. In addition, descriptive data on adjunctive procedures at the time of embryo transfer, such as the use of high‐concentration hyaluronic acid medium, two‐step transfer, and endometrial scratching, were also presented. Continuous variables were expressed as means with standard deviations, and categorical variables as counts and percentages. As this was a retrospective cohort study, no formal hypothesis testing was conducted for descriptive summaries. The incidence of RPOC was calculated for each group, with effect measures presented as absolute risk differences (RDs), relative risks (RRs), and odds ratios (ORs) with 95% CIs. Given the small number of events, exact statistical tests such as Fisher's exact test and the Cochran–Mantel–Haenszel test were applied where appropriate. To further explore the association between embryo transfer protocol and RPOC occurrence, Firth's penalized logistic regression was performed to reduce small‐sample bias. Owing to the limited number of RPOC cases, conventional multivariable logistic regression was not performed to avoid overfitting. Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Two‐sided p ‐values < 0.05 were considered statistically significant.

Table S1: Clinical details of RPOC cases in the NC group: gestational age and additional hormonal supplementation.