How embryo retention affects assisted reproductive technology outcomes: a systematic review and meta-analysis

Abstract

Background

Retransferring retained embryos during the embryo transfer (ET) procedure has raised concerns about its adverse effects on assisted reproductive technology (ART) outcomes. Technical challenges associated with embryo retention (ER) may compromise implantation success and lead to increased complications.

Objective

This systematic review and meta-analysis aimed to evaluate the impact of retransferring retained embryos on key ART outcomes, including clinical pregnancy rate (CPR), biochemical pregnancy rate (BPR), ectopic pregnancy rate (EP), miscarriage rate (MR), and live birth rate (LBR).

Methods

We conducted a systematic search in PubMed, Scopus, and Cochrane databases from inception to April 11, 2025. ART outcomes were extracted and pooled Mantel–Haenszel odds ratios (OR) with 95% confidence intervals (CI) were calculated under both fixed- and random-effects models. Subgroup analyses were performed based on study design (matched versus non-matched retrospective cohorts) and ET technique (afterload versus direct). Sensitivity analyses were conducted by excluding studies with high or very high risk of bias, as determined by the ROBINS-E tool.

Results

The overall analysis demonstrated that retransferring retained embryos was associated with a significant reduction in CPR (OR ≈ 0.75, 95% CI: 0.64–0.89, p < 0.001) and LBR, while substantially increasing the risk of EP (OR ≈ 2.36, p = 0.036). Subgroup analysis showed that studies with matched designs and those using the afterload ET technique exhibited more pronounced negative outcomes. Sensitivity analyses confirmed the robustness of the primary findings.

Conclusion

ER negatively impacts ART success, lowering clinical pregnancy and live birth rates and elevating the risk of ectopic pregnancy. These findings highlight the critical need to refine ET protocols and further investigate the biological mechanisms underlying ER. Future well-designed, prospective studies with standardized methodologies are warranted to optimize ER management and improve clinical outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00404-025-08185-2.

Introduction

Embryo transfer (ET) is the final step in an in-vitro fertilization (IVF) cycle, playing a pivotal role in the success of assisted reproductive technology (ART). During the ET procedure, the embryo is carefully delivered into the uterine cavity, a process that demands a well-coordinated collaboration between the gynecologist and embryologist. In rare cases, the embryo may remain within the ET catheter or adhere to its exterior surface. This uncommon phenomenon, known as embryo retention (ER), represents a challenging event that can pose significant anxiety for both the patient and the medical team.

With recent advancements in ET procedure, like ultrasound-guided transfers, standardized ET catheters, improved transfer techniques, rigorous quality control programs, and enhanced operator training, the risk of ER has markedly decreased. Nevertheless, even in ideal conditions, ER can still occur. Potential risk factors for ER include contamination of the transfer catheter with mucus or blood, difficulty during transfer, intrauterine pressure, ET techniques (direct or afterload), the experience of the ET operator, and the volume of the culture medium used to load the catheter [1–6].

Several original studies have examined the impact of retransferring retained embryos on ET outcomes. Whereas many investigations have reported that retransferring embryos retained in the catheter does not adversely affect IVF outcomes [7, 8], other studies have found a diminished implantation and pregnancy rates following ER [6, 9].

Evaluating the effect of ER on ART outcomes is crucial due to its potential impact on embryo implantation and pregnancy success. Prolonged transfer procedures and environmental factors, such as temperature fluctuations or suboptimal handling, can further compromise embryo viability [10–12], with their effects becoming more pronounced in cases of ER. Additionally, coexisting conditions, such as difficult transfers, mucus and blood contamination, may exacerbate the risks by inducing endometrial trauma or triggering uterine contractions, both of which could jeopardize precise embryo implantation [2, 3, 6]. Furthermore, the immediate retransferring of retained embryos poses a concern regarding excessive uterine fluid accumulation, which may reach a threshold capable of "flushing out" previously transferred embryos through the cervix or into the fallopian tubes. Systematically assessing these interconnected risks will help refine ET techniques, enhance clinical protocols, and improve ART outcomes, ultimately ensuring higher success rates and better patient experiences. Therefore, this systematic review and meta-analysis aims to evaluate the effect of retransferring retained embryos on ET outcomes for the first time.

Methods

This systematic review and meta-analysis aimed to evaluate the effect of retransferring retained embryos on ET outcomes in comparison to successful ETs. Our methodology adhered to the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Supplementary File 1).

Study PECOs

P (population)

Patients undergoing ET as part of ART, including fresh and frozen embryo transfer (FET) cycles.

E (exposure)

ER, which refers to the failure of the embryo to be released into the uterine cavity during ET. Instead of successfully depositing in the uterus, the embryo remains within the transfer catheter, adhering to either the inner lumen or outer sheath due to blood, mucus, suboptimal fluid dynamics, intrauterine pressure, or any other reasons. This phenomenon can negatively impact implantation potential, requiring a retransfer procedure to complete the ET process.

Early identification of ER is critical to preventing procedural delays and minimizing the risks associated with retransfer. The diagnosis typically involves the following steps, requiring close examination by an embryologist under a microscope. Immediately after the first ET attempt, the embryologist examines the catheter under a microscope to confirm whether the embryo was successfully deposited in the uterine cavity. The embryologist inspects the inner lumen of the catheter under high magnification to check for the presence of an embryo. If blood or mucus contamination is present, additional difficulty in visualization may occur, requiring careful microscopic assessment. In dual-lumen catheters, the embryologist also inspects the outer sheath for potential embryo adhesion. If the embryo is still in the catheter, the transfer is deemed unsuccessful, necessitating an immediate retransfer attempt. In cases where visualization is inconclusive, an embryologist may verify by flushing the catheter contents into a culture dish for further microscopic assessment.

Once ER is confirmed, the retransfer process is initiated. The retained embryo is carefully repositioned into fresh culture medium to preserve viability. If necessary, a new catheter is used to ensure optimal embryo delivery, particularly if mucus or blood contamination was observed. The embryo is reloaded into the catheter and transferred into the uterine cavity. Following retransfer, the catheter is re-examined under a microscope to ensure that the embryo has been successfully deposited. If retention persists, clinicians may reassess ET conditions, considering patient-specific factors such as uterine contractility or anatomical challenges.

C (comparator)

Successful ET without retention, where the embryo was deposited in the uterine cavity in the first attempt without requiring retransferring.

O (outcome)

Key ART outcomes, including clinical pregnancy rate (CPR), as the primary outcome, and biochemical pregnancy rate (BPR), live birth rate (LBR), miscarriage rate (MR), and ectopic pregnancy rate (EP), as the secondary outcomes.

Eligibility criteria

Inclusion criteria

We included all types of retrospective and prospective observational studies, including case–control and cohort studies encompassing women recruited from clinical, hospital-based, and population-based settings. Just full-text articles providing sufficient data for statistical analyses were eligible for inclusion. No restrictions were applied on date, language, or geographical region.

Exclusion criteria

We excluded case reports, case series, conference abstracts, pre-prints, reviews, editorials, animal studies, in-vitro studies, and studies that lacked sufficient methodological or result details. Also we excluded studies that compared outcomes of ER with control groups other than successful ETs like difficult transfer.

Search strategy

We conducted a systematic search in PubMed, Scopus, and Cochrane databases from inception to April 11, 2025. Additional studies were identified through hand-searching references and citation lists. Search terms included combinations of “embryo retention”, “retained embryo”, “embryo transfer catheter”, and “embryo transfer complications" (for detailed search strategy, see Supplementary File S2).

Study selection

Relevant studies were organized into a structured Excel database, with duplicates removed. Two independent reviewers (NE and R.K.) screened titles and abstracts against eligibility criteria. Full texts of potentially relevant studies were retrieved for comprehensive evaluation. Discrepancies were resolved through discussion with a third reviewer (F.A.). A full list of studies that were excluded by abstract and full-text screening with reasons for exclusion is provided in Supplementary File S4.

Data extraction

Data from each qualifying study were extracted and verified by two reviewers (R.K. and N.E.) using a standardized electronic data capture form. Extracted data included the following.

Baseline characteristics

Methodological aspects (publication year, location, overall incidence of ER, study duration, and design), technical and clinical considerations (including cryopreservation status [fresh vs frozen], number of transferred embryos, stage of transferred embryos, ultrasound visualization, ET technique [afterload vs direct], catheter type [hard vs soft, and single vs double lumen], distance of catheter’s tip from uterine fundus, medium volume loaded in catheter, loading technique, mucus removal before catheter insertion, catheter excision technique, and endometrial preparation protocol).

Primary outcome measure:

CPR: This is the number of intrauterine clinical pregnancies diagnosed by ultrasonographic observation of fetal heartbeats, divided by the total number of ET cycles.

Secondary outcome measures:

• BPR: This refers to the number of chemical pregnancies diagnosed by the detection of beta human chorionic gonadotropin (β-hCG) in blood or urine, relative to the total number of ET cycles.

• MR: This represents any type of pregnancy loss occurring before 20 weeks of gestation, relative to the total number of pregnancies. We consider all forms of miscarriage, biochemical, clinical, spontaneous, induced, early, or late, as part of this rate. Ideally, the total number of positive BhCG tests is used as the number of pregnancies; however, in some studies that only reported clinical miscarriages or did not provide sufficient data for biochemical pregnancies, the number of clinical pregnancies was used for calculating the MR.

• LBR: This is the proportion of all ET cycles that resulted in the birth of at least one live baby, counted as the birth of one or more live babies from a single pregnancy, regardless of whether this involves singletons or multiples.

• EP Rate: Defined as the number of fertilized oocytes or gestational sacs implanted outside the uterine cavity relative to the total number of pregnancies. EP is most commonly diagnosed in the fallopian tube, although implantation may also occur at other extrauterine sites, such as the ovary, cervix, or abdominal cavity.

Quality appraisal

We employed the ROBINS-E (Risk of Bias in Non-randomized Studies of Exposures) tool to assess the risk of bias in included studies. Two independent reviewers (A.M-H. and R.K.) judged about each study’s quality and any discrepancies were resolved through discussion with a third reviewer (F.A.).

Statistical analysis

ART outcomes were extracted, and the Mantel–Haenszel odds ratio (MHOR) with 95% confidence intervals (CI) was calculated for each categorical endpoint. The pooled MHOR with 95% CI was obtained using both fixed and random-effects models. Heterogeneity was assessed using forest plots, Chi-square-based Q statistic, and I² value, with significance P value or I² > 40%. Statistical analyses were performed using Comprehensive Meta-Analysis (CMA) software (Version 3.0, Biostat Inc., USA).

Subgroup analyses

Before data extraction, we planned to perform subgroup analyses based on a wide range of variables, including contamination with blood or mucus, transfer difficulty, cryopreservation status, catheter type, distance from the catheter’s tip to the uterine fundus, the volume of medium loaded in the catheter, loading technique, mucus removal, catheter excision technique, endometrial preparation, the number of transferred embryos, and the stage of transferred embryos.

However, the available studies did not provide sufficient or consistent data for many of these variables. For example, only two studies reported separate data for cases with mucus or blood contamination [6], [24], while one study exclusively included cases without such contamination [25]; most studies presented mixed data, precluding a focused analysis in this area. Similarly, no studies reported on transfer difficulty, and although cryopreservation status was mentioned, most did not differentiate between fresh and FET cycles. Additionally, parameters such as the distance from the catheter’s tip to the uterine fundus, medium volume loaded, loading technique, mucus removal, catheter excision, endometrial preparation, number of embryos transferred, and the stage of embryos were either reported heterogeneously or not reported in enough detail to allow for meaningful subgroup comparisons. The ultrasound visualization status during the transfer also was used in most of studies, so this parameter was not considered for subgroup analysis.

Given these limitations, subgroup analyses were ultimately performed based on three factors: study design (matched versus non-matched retrospective cohorts) and ET technique employed (afterload versus direct).

In the direct approach, the catheter is preloaded with the embryo, and the transfer is executed immediately without any prior trial insertion. This method can be performed using either a single-lumen catheter or a double-lumen catheter, where the inner, softer lumen carries the embryo as the outer rigid lumen facilitates placement. In contrast, the afterload technique separates the outer sheath from the inner catheter. The outer sheath is first advanced through the internal cervical os until it reaches the upper cervical canal, where it is left in place. The inner catheter, preloaded with the embryo, is then threaded through the outer sheath into the endometrial cavity for embryo release.

Sensitivity analyses

To ensure the robustness of our findings, we conducted a sensitivity analysis by excluding studies with a high and very high risk of bias from the meta-analysis.

Publication bias

Publication bias for each outcome was assessed using funnel plots. The trim-and-fill analysis was also conducted to adjust for any 'missing' studies due to publication bias.

Results

Summary of the literature search

Our initial search across electronic databases retrieved a total of 324 records from databases. This included 106 from PubMed and 218 from Scopus. Also an additional 13 records found through manual and citation searching. After removing 98 duplicate records, 239 records were screened. Of these, 200 records were excluded based on the screening criteria.

39 reports were sought for retrieval, and except for one study, all were successfully retrieved. These reports were assessed for eligibility, resulting in the exclusion of 28 reports for the following reasons: reporting technical issues other than ER (n = 19), not eligible publication type (n = 7), lack of a favorable control group (n = 1), and lack of sufficient data for meta-analysis (n = 1).

Ultimately, 10 studies were included in this systematic review with a total of 46,842 ET cycles. This included 45,965 successful ETs (control) and 877 cycles complicated by ER, indicating a pooled ER rate of 1.9%.

The study flow of our literature search and study selection is depicted in the PRISMA 2020 flow diagram [13] (Fig. 1). For transparency, we have also provided a complete list of the screened studies, along with their details and reasons for exclusion at each screening phase, in Supplementary File 4.

Fig. 1.

PRISMA (preferred reporting items for systematic reviews and meta-analyses) 2020 flow diagram

Study characteristics

Methodological characteristics

Table 1 summarizes the key methodological and technical characteristics of all included studies. These studies spanned various geographic locations, study designs, and clinical protocols.

Table 1.

Methodological characteristics and details of technical considerations in reviewed studies

Study ID	Location	Design	Cryopreservation status	Overall incidence of embryo retention	Ultrasound visualization	ET technique	Catheter type	Distance from catheter’s tip to uterine fundus	Medium volume loaded in catheter	Loading technique	Mucus removal	Catheter excision technique	Endometrial preparation	Study outcomes	Number of transferred embryos	Stage of transferred embryos
He et al., 2025	China	Matched retrospective cohort	FET	0.47% (185/39,118 ETs) in a 6 years period	Yes	Afterload	Dual-lumen	1–1.5 cm	20 mL	Small air bubbles were strategically placed at the tip	Wiping	A short delay and rotation	Various	BPR, CPR, MR, EP, LBR, CS, GA at birth, BW	1–2	Cleavage and Blastocyst
Jung et al. 2003	Korea	Retrospective	Fresh	NM	No	Direct	Firm single-lumen catheter in easy transfers and dual-lumen in difficult cases	0.5–1 cm	10–20 mL	5cc of air, 10 to 20cc of culture medium and embryo, and 5cc of air	Wiping	A short delay without rotation	NA	BPR, FR, IR	4–5	Cleavage
Kadour-Peero et al., 2022	Canada	Matched retrospective cohort	FET and Fresh	0.47% (213/15,321 ETs) in a 10 years period	Yes	Afterload	Dual-lumen	1–1.5 cm	20 mL	A 5-mL air bubble at the tip of the inner catheter	Wiping	A short delay and rotation	Various	IR, CPR, LBR, EP, MR	1–3	Cleavage and blastocyst
Lee et al., 2004	USA	Retrospective	FET and Fresh	3.9% (71/1809) in a 5 years period	only in half of study period	Some cases direct and some cases afterload	Soft single-lumen catheter in easy transfers and malleable stylets in difficult cases	NM	20–35 μL	10–15 μL of fluid is drawn into the catheter first, the embryo(s) are then aspirated, Another 10–15 μL of fluid	Some cases wiping and some cases aspiration	10 to 60 s delay without rotation	NM	BPR, IR, CPR, LBR (No sufficient data was reported for CPR and LBR)	NM	Cleavage and blastocyst
Nabi et al., 1997	UK	Retrospective	Fresh	5.7% (69/1204) in a 4 years period	No	Direct	Soft single-lumen and dual-lumen	0.5–1 cm	20 μL	Air bubbles, culture medium containing embryos and air bubbles again	Wiping	No detail provided	NA	CPR	Maximum of 3	Cleavage
Oraif et al., 2014	UK	Retrospective	FET and Fresh	7.5% (49/652) in a 3 years period	Yes	Direct	Dual-lumen	1.2 cm	NM	NM	Wiping	2 min delay without rotation	NM	CPR	NM	Cleavage
Vicdan et al., 2007	Turkey	Retrospective	FET and Fresh	2.8% (41/1454) in a 2 years period	Yes	Direct	Dual-lumen	1 cm	20–30 μL	No air was loaded in catheter, just embryo containing media was loaded	Wiping and Aspiration	No detail provided	NM	BPR, CPR, IR	NM	NM
Xu et al., 2020	China	Matched retrospective cohort	FET and Fresh	1.59% (97/6,089) in a 5 year period	Yes	Afterload	Dual-lumen	1–1.5 cm	20 μL	A 5-μL air bubble at the tip of the inner catheter	Wiping	A short delay and rotation	NM	CPR, IR, EP, MR, LBR, GA at birth, BW	NM	Cleavage and blastocyst
Yi et al., 2016	Korea	Retrospective	FET and Fresh	2.8% (32/1,131) in a 10 year period	Yes	Afterload	Dual-lumen	NM	20 μL	A 5-μL air bubble at the tip of the inner catheter	Wiping	No detail provided	NM	BPR, IR, CPR, LBR, EP	NM	Blastocyst
Zhang et al., 2023	China	Matched retrospective cohort	FET and Fresh	0.33% (95/29,160) in a 5 year period	Yes	Afterload	Dual-lumen	1–1.5 cm	20 μL	A small air bubble at the tip of the inner catheter	Wiping	1 min delay and no rotation	various	BPR, CPR, MR, EP, LBR	NM	Cleavage and Blastocyst

Among the included studies, five employed a matched retrospective cohort design, while six were retrospective studies without matched controls.

Technical characteristics

Cryopreservation status varied across studies, with some focusing exclusively on FET, others including both fresh and frozen cycles, and a few solely analyzing fresh transfers.

While most studies employed trans-abdominal ultrasound guidance during ET, some either lacked ultrasound assistance entirely or implemented it inconsistently throughout the study period.

The ET technique used in the studies was either the afterload or direct approach. Studies using the afterload method predominantly employed dual-lumen catheters, whereas those using the direct approach varied between firm single-lumen, soft single-lumen, or dual-lumen catheters, depending on transfer difficulty.

Significant procedural differences were observed regarding catheter positioning, medium volume loaded, embryo loading techniques, mucus removal protocols, and catheter excision strategies. Distances from the catheter’s tip to the uterine fundus ranged from 0.5 to 1.5 cm, with medium volumes varying between 10 and 35 μL. Loading techniques often involved small air bubbles strategically placed at the catheter tip, though some studies did not load air at all. Mucus removal was generally achieved only via wiping, though a few studies employed aspiration. Catheter excision techniques varied, with some studies incorporating rotation and short delays, while others omitted rotation altogether or implemented extended delays.

Endometrial preparation protocols were heterogeneous, with some studies specifying methods while others provided limited details.

The number of transferred embryos varied among studies, ranging from one to five embryos per cycle. Likewise, embryo stage differed, with transfers involving cleavage-stage embryos, blastocysts, or a combination of both.

Reported ER rates

The reported incidence of ER varied widely across studies, ranging from 0.33 to 7.5% of total ET cycles, and there appears to be a trend when considering the publication dates. For example, more recent studies, such as He et al. (2025, China), Kadour-Peero et al. (2022, Canada), and Zhang et al. (2023, China), reported very low ER rates (ranging from 0.33 to 0.47%). In contrast, earlier studies like Nabi et al. (1997, UK) and Lee et al. (2004, USA) reported considerably higher ER rates at 5.7% and 3.9%, respectively. Additionally, the study by Oraif et al. (2014, UK) reported an ER rate of 7.5%, which might partly reflect differences in technique and reporting rather than a linear time trend.

Notably, the adoption of different ET techniques over time also coincides with these differences. Studies employing the afterload technique (a method more frequently reported in recent publications) generally exhibited lower ER rates, while those using direct transfer, which was more common in older cohorts, tended to report higher ER. Unfortunately, inclusion of heterogeneous procedural variables or incomplete reporting on these variables, including cryopreservation status (fresh vs FET), developmental stage of embryos (cleavage vs blastocyst), media volume, catheter type, and operator skill limited our ability to provide more comparisons about risk factors of ER. Thus, while publication date appears related to observed ER rates, these differences are also likely influenced by concurrent improvements in technique and reporting practices over time.

Quality and risk-of-bias assessment

The summary of the risk-of-bias assessment is shown in Fig. 2. A total of ten studies were evaluated for risk of bias using the ROBINS-E tool. The final scoring was as follows: three studies were rated as having a low risk of bias, three studies had some concerns, three studies were rated as high risk, and one study were rated as having a very high risk of bias.

Fig. 2.

Risk of bias according to ROBINS-E Tool

Domain 1 (risk of bias due to confounding): Studies that did not match groups or adjust for key confounding variables were rated as having "some concerns." This issue was notable in several studies where the lack of group matching might have introduced bias in the comparison of outcomes.

Domain 2 (risk of bias arising from measurement of the exposure): Most studies demonstrated a low risk in this domain, reflecting standardized procedures for assessing the exposure (ER) across the research.

Domain 3 (risk of bias in the selection of participants into the study or analysis): Retrospective study designs inherently raised concerns in this domain. Many studies were rated as having "some concerns," particularly those that were retrospective with unmatched control groups. In a few instances, the risk was deemed "high" when these issues were pronounced, further compounded by factors, such as not using ultrasound guidance consistently, undefined inclusion or exclusion criteria, a mixed approach combining both fresh and frozen cycles, or varying operators during ET.

Domain 4 (risk of bias due to post-exposure interventions): This domain was generally rated as low risk, although a few studies exhibited "some concerns" where subsequent interventions might have influenced outcomes.

Domain 5 (risk of bias due to missing data): Missing data represented a challenge in several studies, leading to a rating of "some concerns" in this domain.

Domain 6 (risk of bias arising from measurement of the outcome): Concerns here stemmed from the inadequate reporting of outcomes, specifically, the absence of precise or absolute values in some studies. Consequently, this domain was frequently rated as "some concerns."

Domain 7 (risk of bias in the selection of the reported result): The majority of studies received a low-risk rating in this domain, suggesting that selective reporting was not a widespread problem.

The overall risk-of-bias ratings varied among the studies. For instance, He et al. (2025), Xu et al. (2020), and Zhang et al. (2023) were classified as having an overall low risk of bias. In contrast, studies, such as Lee et al. (2004), Nabi et al. (1997), and Vicdan et al. (2007), were judged to be at high risk, largely driven by significant issues in participant selection. Oraif et al. (2014) were rated as having a very high risk of bias due to multiple concerns across several domains. This heterogeneity in bias assessments underscores the importance of cautious interpretation when synthesizing the effects of ER on ART outcomes.

Results of meta-analysis

CPR

A total of eight studies were included in the analysis of the CPR. Under the fixed-effects model, the pooled MH OR was 0.75 (95% CI: 0.64–0.88; Z = –3.41, p < 0.001). The random-effects model produced a similar estimate (MHOR = 0.76, 95% CI: 0.63–0.91; p = 0.003). Heterogeneity was low, as indicated by an I² value of 16% (Q = 8.34, df = 7, p = 0.30) (Fig. 3).

Fig. 3.

Comparison of clinical pregnancy rate in embryo transfer cycles complicated by embryo retention versus successful embryo transfers. Forest plot detailed Mantel–Haenszel odd ratio and 95% confidence intervals for biochemical pregnancy rate

These results indicate that retransferring retained embryos is associated with a statistically significant reduction in the odds of achieving a clinical pregnancy, and the low heterogeneity supports the consistency of this finding across studies.

Subgroup analysis for CPR by study design

Matched versus non-matched retrospective cohorts

Matched retrospective cohorts (n = 4): The fixed-effects analysis revealed a pooled OR of 0.68 (95% CI: 0.56–0.83; p < 0.0001) with negligible heterogeneity (I² = 0%, Q = 2.25, df = 3, p = 0.522). In studies where groups were adequately matched (or confounders were adjusted for), retransferring retained embryos was significantly associated with lower CPR.

Non-matched retrospective cohorts (n = 4): The pooled OR was 0.97 (95% CI: 0.71–1.32; p = 0.862) with no detectable heterogeneity (I² = 0%, Q = 2.60, df = 3, p = 0.457). In studies without adequate group matching, no significant effect on CPR was observed, suggesting that confounding may mask the true impact of retransferring retained embryos (Supplementary Fig. 1, Supplementary File S3).

Afterload versus direct approach

Afterload technique (n = 5): The fixed-effects analysis yielded a pooled OR of 0.692 (95% CI: 0.57–0.83; p < 0.0001), with no significant heterogeneity (I² = 0%, Q = 2.34, df = 4, p = 0.674), indicating that studies employing the afterload ET technique demonstrated a significant reduction in CPR.

Direct technique (n = 3): The fixed-effects model produced a pooled OR of 1.030 (95% CI: 0.72–1.45; p = 0.865) suggesting that the direct ET approach did not significantly affect CPR. It seems this technique may mitigate the adverse effects observed with retransferring retained embryos. However, this finding could be caused by the lower number of included studies and a smaller size in this subgroup (Supplementary Fig. S2, Supplementary File S3).

Sensitivity analysis for CPR

Excluding studies with a high or very high risk of bias (n = 4) [14–17] resulted in a fixed-effects OR of 0.69 (95% CI: 0.57–0.83; p < 0.0001). The consistency of these results with the overall analysis confirms the robustness of the observed effect of retransferring retained embryos on reducing clinical pregnancy rates.

Publication bias assessment for CPR

The funnel plot revealed two missing studies (Supplementary Fig. 3, Supplementary File 3), but the adjusted estimate according trim-and-fill test, found similar results to original analyses (adjusted MHOR: 0.70; 95% CI: 0.60–0.81) supporting the robustness of the findings.

BPR

Five studies were analyzed for the BPR. Under the fixed-effects model, the pooled OR was 0.792 (95% CI: 0.644–0.974; p = 0.027), with moderate heterogeneity (I² = 49.4%, Q = 7.899, df = 4, p = 0.095). The random-effects model, however, produced a pooled OR of 0.871 (95% CI: 0.6311–1.2041; p = 0.405) (Supplementary Fig. 4, Supplementary File 3). Although the fixed-effects model suggests a significant reduction in BPR, the random-effects result is not statistically significant. This discrepancy, coupled with moderate heterogeneity, indicates variability among studies that limits definitive conclusions regarding BPR. Considering the small number of eligible studies reporting BPR, subgroup and sensitivity analyses were not performed.

Publication bias assessment for BPR

The funnel plot revealed two missing studies (Supplementary Fig. 5, Supplementary File 3), but the adjusted estimate according trim-and-fill test, found similar results to original analyses (adjusted MHOR: 0.68; 95% CI: 0.56–0.82) supporting the robustness of the findings.

EP rate

Five studies contributed data on EP. The fixed-effects model provided a pooled OR of 2.35 (95% CI: 1.05–5.25; p = 0.036), and the random-effects model yielded nearly identical estimates. Heterogeneity was minimal (I² = 0%, Q = 1.04, df = 4, p = 0.902). These findings suggest that retransferring retained embryos is associated with a significantly increased risk of EP, and the consistency across models and lack of heterogeneity reinforce the reliability of this association (Supplementary Fig. 6, Supplementary File 3).

Publication bias assessment for EP rate

The funnel plot revealed no missing studies (Supplementary Fig. 7, Supplementary File 3), so the adjusted estimate according trim-and-fill test, was similar to original analyses.

Miscarriage rate

Data from four studies showed no significant difference in miscarriage rate. The fixed-effects model yielded an OR of 0.947 (95% CI: 0.599–1.495; p = 0.814), and the random-effects model showed similar results. Heterogeneity was absent (I² = 0%). Therefore, retransferring retained embryos does not appear to influence the miscarriage rate significantly, as evidenced by the lack of effect and consistent findings across studies (Supplementary Fig. 8, Supplementary File 3).

Publication bias assessment for miscarriage rate

The funnel plot revealed no missing studies (Supplementary Fig. 9, Supplementary File 3), so the adjusted estimate according trim-and-fill test was similar to original analyses.

LBR

For LBR, four studies were analyzed. The fixed-effects model indicated a pooled OR of 0.697 (95% CI: 0.560–0.868; p = 0.0012), and the random-effects model provided a similar estimate (OR = 0.7043, 95% CI: 0.5358–0.9259; p = 0.0120). Heterogeneity was low to moderate (I² = 25.2%, Q = 4.013, df = 3, p = 0.2601). These results indicate that retransferring retained embryos is significantly associated with a reduced LBR, a clinically important finding that aligns with the observed reduction in clinical pregnancy outcomes (Supplementary Fig. 10, Supplementary File 3).

Publication bias assessment for LBR

The funnel plot revealed one missing studies (Supplementary Fig. 11, Supplementary File 3), but the adjusted estimate according trim-and-fill test, found similar results to original analyses (adjusted MHOR: 0.65; 95% CI: 0.53–0.81) supporting the robustness of the findings.

Discussion

Summary of study findings

Our meta-analysis showed that ER is associated with adverse clinical outcomes in ART. Overall, the analysis of eight studies revealed that retransferring retained embryos is associated with about 25% reduction in CPR compared to successful ET. The low heterogeneity observed (I² ≈ 16%) reinforces the consistency of this finding across different studies. Subgroup analyses further clarified the impact of study design and the ET technique on outcomes. When stratified by study design, studies that employed matched retrospective cohorts showed a notably lower CPR compared to non-matched retrospective studies. Similarly, the ET technique emerged as an important determinant. The afterload method, used in five studies, was associated with a significant reduction in CPR, whereas the direct transfer approach, used in three studies, showed no significant effect on CPR. This finding should be interpreted with caution, as current clinical practice supports the use of afterload technique to reduce the rate of difficult ET and its associated complications, including RE [5]. On the other hand, subgroup analysis within the direct category was based on fewer studies, which indicate lower statistical power. Therefore, this discrepancy may just be resulted from heterogeneity in study population or improper control of potential confounders. Sensitivity analysis, excluding studies with a high or very high risk of bias, yielded nearly identical results for CPR, confirming the robustness of the overall findings.

Regarding secondary outcomes, the fixed-effects analysis for BPR indicated a modest reduction (≈ 21%), but the random-effects model did not confirm statistical significance, likely reflecting underlying heterogeneity among the studies. Importantly, our findings indicated that retransfer of retained embryos poses 2.36 times higher risk of EP with no detectable heterogeneity. In contrast, miscarriage rates were unaffected by retransferring retained embryos, while live birth rates were significantly reduced (≈ 30%).

Beside the main study outcomes, we also observed some variations in ER rates among studies that reveal important clinical and methodological implications. The trend toward lower ER rates in more recent studies (e.g., 0.33–0.47% in studies published from 2022 to 2025) suggests that advancements in ET techniques, like wider adoption of the afterload approach, may be instrumental in improving embryo expulsion and overall transfer efficiency. Conversely, higher ER rates reported by older studies (such as Nabi et al., 1997 and Lee et al., 2004) could reflect earlier practices using less standardized protocols. Furthermore, the differing ER rates by publication date highlight how evolving technical refinements, such as improved catheter designs (dual-lumen systems) and more precise positioning by implementing ultrasound guidance (with distances of 1–1.5 cm from the uterine fundus), and even improved operators skill over the time, have contributed to decreased ER incidence.

In summary, our findings indicate that retransferring retained embryos is associated with reduced CPR and LBR, and an increased risk of EP; however, it appears to have a limited impact on BPR and miscarriage rates. These results underscore the importance of optimizing ET protocols to minimize ER and its subsequent adverse effects on ART outcomes. In. the other hand, the trends in ER rate by publication date reinforce the narrative that improvements in ET techniques have likely played a key role in reducing the occurrence of ER, which, in turn, has a meaningful impact on ET outcomes. Future investigations should aim to standardize reporting of technical parameters, such as catheter type, loading technique, and medium volume, to further elucidate their contributions to ER and to continue enhancing ET success.

Potential biological mechanisms behind detrimental effects of ER

Existing literature on the risk factors associated with ER and its detrimental effects on ART outcomes is limited, highlighting the need for further research. One of the most critical issues influencing the success of the ET procedure is its difficulty. Difficult ETs are not only linked to the loss of a viable embryo that could implant (approximately one in every four cases) but are also associated with increased blood and mucus contamination, two significant risk factors for ER.

When blood or mucus is present, the embryo can adhere to either the inner catheter or the outer sheath. Cervical tissue secretes mucus, which may accumulate on the outer sheath as it passes through the cervical opening. Consequently, the inner catheter might become coated or obstructed by mucus or blood once it exits the outer sheath, potentially causing the embryo to stick to this contaminant instead of being transferred into the uterus.

Moreover, difficult or prolonged ETs can lead to endometrial trauma, causing mechanical irritation or injury to the lining. This trauma may trigger localized inflammation and alter the endometrial environment, reducing its receptivity and disrupting the finely tuned synchronization required for successful implantation.

On the other hand, reloading or retransferring a retained embryo may subject it to additional mechanical and environmental stress. This extra handling could compromise cellular integrity and viability, potentially affecting the embryo’s developmental potential.

Another potential risk factor is intrauterine pressure. One experimental study [4] demonstrated that pressure fluctuations during ET can cause the transferred fluid to flow backward along the catheter, leading to ER. Furthermore, unwanted uterine contractions, induced by the stress and manipulation of a challenging ET or by hormonal stimulation in fresh ET cycles, can increase intrauterine pressure. These contractions may displace the embryo from its optimal deposition site, leading to ER.

The fluid dynamics involved in retransferring a retained embryo also warrant concern. In cases where embryos must be retransferred immediately, the additional volume of fluid combined with the mechanical process of reloading can lead to excessive uterine fluid accumulation. This may generate a "flushing" effect, where embryos previously positioned in the uterine cavity are inadvertently washed away through the cervix or even into the fallopian tubes. Our findings indicate an increased risk of EP in ER cases compared to successful transfers.

Finally, the multiple attempts required for transferring retained embryos increase the risk of contamination with blood and microorganisms. These contaminants can have cytotoxic effects, interfere with the biochemical environment needed for implantation, or provoke an inflammatory response, all of which could impair the embryo's implantation potential.

Each of these mechanisms, whether acting individually or in combination, may contribute to the lower clinical pregnancy and live birth rates, as well as the increased risk of complications such as EP, observed when retained embryos are retransferred. Understanding these biological underpinnings is crucial for refining ET protocols and improving overall ART outcomes.

Robustness of study findings

Several limitations inherent in the included studies should be acknowledged. First, the majority of studies were retrospective, with several employing non-matched designs. This retrospective design increases susceptibility to selection bias and residual confounding, as many studies did not adjust for key variables. Our risk-of-bias assessment using the ROBINS-E tool highlighted concerns in Domain 1 (confounding) and Domain 3 (participant selection), with several studies rated as “some concerns” or even “high risk” due to unmatched control groups, undefined inclusion or exclusion criteria, and the mixing of fresh and frozen cycles without appropriate stratification. A growing body of evidence showed that ER risk can be more pronounced in fresh cycles [6, 8, 9, 18]. This could be attributed to increased cervical mucus production in fresh cycles because of the elevated estradiol levels or increased uterine contractility [19].

Furthermore, critical aspects of the ET procedure, such as loading technique, medium volume, mucus removal, and catheter excision strategy, were reported inconsistently. These issues complicate the interpretation of how procedural nuances affect outcomes, as some studies lacked detailed descriptions, while others employed mixed techniques. Outcome measurement also posed a challenge; for example, in Domain 6, several studies received “some concerns” because outcomes were not reported comprehensively or the absolute values were not provided. Finally, variability with respect to ultrasound guidance and operator proficiency further undermines the uniformity of the evidence. These methodological limitations, alongside the identified risk-of-bias concerns, suggest that the results of individual studies should be interpreted with caution.

Despite rigorous methodology, this systematic review and meta-analysis has its own limitations. First, the review relied exclusively on observational studies, which are inherently more vulnerable to bias compared to randomized-controlled trials. The heterogeneity in study design, ET techniques, and reporting standards made it challenging to perform detailed subgroup analyses for outcomes other than clinical pregnancy rate. In addition, some potentially influential variables, such as the extent of blood or mucus contamination and precise procedural details, were either underreported or inconsistently detailed in the source studies, limiting our ability to adjust fully for these factors. Moreover, while sensitivity analyses were conducted by excluding studies with a high risk of bias, residual confounding cannot be entirely ruled out.

Although these limitations highlight caution in generalizing the findings, the robust analytical methods employed and the consistency of key results, especially for primary outcomes, contribute valuable insights into the impact of retransferring retained embryos on ART outcomes.

Implications of findings in clinical practice

The findings of this review carry several important implications for clinical practice in ART. Given that retransferring retained embryos is associated with unfavorable ART outcomes and an increased risk of ectopic pregnancy, clinicians must focus on refining ET techniques to achieve two key objectives: (1) minimizing the incidence of ER and (2) reducing the detrimental effects associated with ER when it is unavoidable. Based on the data reported in the included studies, we propose the following recommendations:

1. Optimizing catheter design and selection: Consider the use of dual-lumen catheters, particularly those employing the afterload technique. A soft inner catheter within a dual-lumen system facilitates smoother navigation through the cervical canal and more precise embryo deposition, compared with direct transfer methods. Ultrasound guidance to confirm optimal catheter positioning (ideally 1–1.5 cm from the uterine fundus) is also recommended.

2. Minimizing blood and mucus contamination: Careful handling during ET, such as cleaning blood and mucus, can prevent factors that compromise embryo loading and transfer efficacy. Standardizing these preparatory steps also helps reduce variability in outcomes.

3. Enhancing training and standardization: Ongoing operator training, the implementation of standardized ET protocols, and rigorous quality control are essential. These measures ensure that issues such as suboptimal embryo loading and improper catheter handling are minimized.

4. Reducing mechanical irritation and controlling uterine contractility: Strategies that minimize mechanical irritation during transfer, combined with efforts to control uterine contractility, either through targeted pharmacological interventions or refined procedural techniques, may lessen both the incidence of ER and its subsequent negative impact on clinical outcomes.

5. Optimizing the time interval for retransfer: Although all included studies reported retransfer of retained embryos as “immediate,” no studies provided detailed metrics (e.g., precise time intervals or quantifiable operator measures) to assess this parameter further. We suggest that efforts to minimize any delay could reduce the exposure of embryos to unfavorable environmental conditions.

6. Fine-tuning medium volume and embryo loading methods: Precise control of the transfer medium volume is critical, as excessive volume may impede complete embryo expulsion. Some studies have noted that the strategic placement of a small air bubble at the tip of the inner catheter might enhance expulsion efficiency.

In summary, the clinical implications of these findings emphasize the necessity of meticulous and standardized ET techniques, enhanced operator training, and proactive quality control measures. By addressing these areas, clinicians can better optimize ART outcomes and mitigate complications associated with embryo retention.

Recommendations for future research

Our findings underscore the need for further research to pinpoint the precise mechanisms by which embryo retention impacts ART outcomes and to refine ET protocols accordingly. To advance this field, future studies should consider the following recommendations:

1. Adopting prospective, multi-center, and large-scale designs: Many of the included studies were retrospective, which inherently increases the risk of selection bias and residual confounding. Future research should employ prospective cohort studies to allow for more rigorous control of confounding variables. To overcome issues related to sample size and external validity, future studies should be designed as multi-center trials. This would not only enhance the statistical power for subgroup analyses but also ensure that findings are more generalizable across different clinical settings and patient populations.

2. Standard ET protocols and definitions: Variability in ET techniques, catheter types, and procedural parameters (e.g., loading technique, medium volume, and mucus removal) contributed to heterogeneity among studies. Future research should standardize ET protocols and clearly define key procedural elements. Detailed documentation of the techniques used, including the type of catheter, ultrasound guidance, and any modifications to standard protocols, will facilitate more accurate comparisons and meta-analyses.

3. Enhance reporting of confounding variables and outcomes: Several studies did not provide sufficient data on key factors, such as blood/mucus contamination, operator experience, and precise outcome measures. Future studies should ensure comprehensive reporting of these variables, including detailed baseline characteristics and standardized outcome definitions (such as clinical pregnancy rate, live birth rate, and ectopic pregnancy rate). Future research should also aim for more detailed reporting on exact timing of retransfer to better understand how enhanced retransfer techniques might further improve ART outcomes. This would aid in adjusting for potential confounders and in the interpretation of results.

4. Control for procedural variability: Research should address procedural factors known to affect outcomes, such as the timing of catheter sheath withdrawal, timing of retransfer, the degree of mechanical manipulation, and the impact of uterine contractions. Investigators should design studies that either control for or systematically vary these factors to determine their independent effects on embryo retention and subsequent ART outcomes.

5. Exploring novel ET techniques and interventions: Given the potential detrimental effects associated with retransferring retained embryos, future research should evaluate innovative techniques or modifications, such as alternative catheter designs, loading methods, new transfer medias like high-viscosity media enriched with hyaluronic acid (also known as embryo glue), or pharmacological interventions aimed at reducing uterine contractions. These novel approaches could offer more effective strategies to minimize the incidence of embryo retention and improve overall ART outcomes.

By addressing these recommendations, future research will be better positioned to reduce bias, improve methodological rigor, and generate high-quality evidence that can directly inform clinical practice. Ultimately, such efforts are essential for optimizing ET protocols and enhancing patient outcomes in reproductive medicine.

Conclusion

In summary, our systematic review and meta-analysis demonstrate that ER adversely affects ART outcomes. Specifically, the pooled data reveal a significant reduction in clinical pregnancy and live birth rates, along with an increased risk of EP. Subgroup analyses further underscore the importance of study design and ET technique. Studies with well-matched retrospective cohorts and those employing the afterload method reported more consistent and pronounced negative impacts. Furthermore, our review identifies several methodological limitations in the current literature, including retrospective designs, inconsistent reporting, and variable control of confounding factors. Looking forward, future research should address these limitations by employing prospective, randomized study designs with standardized ET protocols. Detailed reporting of procedural variables, comprehensive outcome measures, and multi-center collaboration will be critical to advance our understanding of the mechanisms underlying embryo retention. Ultimately, such studies will foster the development of targeted interventions to reduce ER incidence, thereby improving the overall success of ART.

Registration of systematic review

Our study has been duly registered in the PROSPERO International prospective register of systematic reviews. The approval code for our registration is CRD420251037916.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ART

Assisted reproductive technology

BPR

Biochemical pregnancy rate

CPR

Clinical pregnancy rate

EP

Ectopic pregnancy

ER

Embryo retention

ET

Embryo transfer

LBR

Live birth rate

MR

Miscarriage rate

Author contributions

All authors contributed to the study conception and design and commented on previous versions of the manuscript. All authors read and approved the final manuscript. A detailed contribution statement is provided below: AM-H: data collection (systematic search, and risk of bias assessment), manuscript original draft preparation, statistical analyses, and study graphs and graphical abstract designing; RK: data collection (study selection, data extraction, and risk of bias assessment) and manuscript editing; NE: data collection (study selection and data extraction) and manuscript editing; FA: study conceptualization; supervision, interpretation of results; manuscript critical review and editing.

Funding

The research team received no specific grant from public, commercial, or not-for-profit funding agencies for performing or reporting this study.

Data availability

The full list of screened studies with the reason of exclusion is available in an Excel file (Supplementary File 4). Other study data including data extracted from included studies and data used for meta-analyses are available in another Excel file (Supplementary File 5).

Declarations

Conflict of interest

The authors declare no competing interests.

Ethics approval and consent to participate

NA.

Consent for publication

NA.