Increasing endometrial thickness beyond 8mm does not alter clinical pregnancy rate after single euploid embryo transfer

Abstract

The aim of this study was to investigate if variation in endometrial thickness affects clinical pregnancy and live birth rates among patients undergoing single euploid embryo transfer (SET).

A retrospective review of IVF cycles performed at a single private fertility institution between 2015–2020 was performed. Patients with normal uterine anatomy undergoing their first SET of a euploid embryo undergoing their first cycle at the center were included, for a total of 796 cycles. Endometrial thickness was measured by transvaginal ultrasound following 10–14 days of estradiol exposure.

Specific infertility diagnoses did not significantly impact endometrial lining thickness with means across diagnoses ranging from 9.3–11.0mm. Endometrial thickness was grouped into five categories: <8mm, 8–10mm, 10–13mm, 13–15mm, and ≥15mm. Using 8–10mm as the reference group, the odds ratio of live birth was 0.5, 1.22, 1.05, and 1.05 for <8mm, 10–13mm, 13–15mm, and ≥15mm groups, respectively. Risk of first trimester miscarriage was equivalent across groups. There was a trend toward an increased rate of biochemical pregnancies in patients with a <8mm and ≥15mm endometrium; however, this was not statistically significant. The clinical pregnancy and live birth rate were lowest in patients with <8 mm endometrial thickness.

For single euploid embryo transfers, an endometrial lining greater than or equal to 8mm confers optimal live birth rates following a medicated FET cycle. These data confirm the findings of prior studies in fresh embryo transfers without the confounders of supraphysiologic ovarian hormone concentrations and genetically untested embryos.

Introduction

Optimization of endometrial receptivity to improve in vitro fertilization (IVF) outcomes has been widely studied. The implantation environment has been an increasing area of interest as the complexities of the endometrial microenvironment have become elucidated. Endometrial thickness has been implicated as a prognostic factor in IVF success (1), however, this has largely been studied in the context of minimum endometrial thickness requirements. Whether there is a threshold at which a thicker endometrium becomes detrimental to implantation and pregnancy outcome has been unclear (2) and a recent meta-analysis was inconclusive (3).

Thin endometrium has been associated with lower pregnancy rates following IVF (4, 5), yet no clear optimal range of endometrial thickness has been established. The impact of an atypically thick endometrium is less studied compared to thin endometrium, and results have been conflicting (2, 6). A retrospective analysis of a study designed to compare progesterone regimens showed no difference in pregnancy rates across a range of endometrial thicknesses up to 18mm, although there was wide variation in pregnancy rates at the thin and thick ends of the spectrum (7). A single-center retrospective study showed decreased pregnancy rates with an endometrium >14mm prior to fresh transfer of day 2 cleavage stage embryos, however, in addition to the clear confounders related to day 2 embryo transfer and potential for early progesterone rise that can be associated with fresh cycles (8, 9), the authors also note older age in patients in the group with thicker endometrium, however, these patients may have had more aneuploidy related to age that impacted their outcomes (2). A study of IVF cycles with fresh embryo transfers used 11mm as a cut-point and demonstrated improved live birth rates (LBR) above this threshold (10). Yet another study which concluded that endometrial thickness did not have an impact on pregnancy rates drew this conclusion based on averages (11), which is unlikely to be useful since the vast majority of patients will have an endometrium >8mm (1), thus the average is above what would typically be considered thin.

In contrast to previous studies, the current study provides a unique, updated perspective on the relationship between endometrial thickness and pregnancy outcomes that is relevant to current clinical practices. By including only blastocyst stage, euploid embryos that were transferred as an SET within a standardized, medicated frozen embryo transfer protocol, numerous confounders present within prior studies have been eliminated. Specifically, this study reduces the risk of aneuploidy potentially affecting study results by solely including embryos deemed euploid by preimplantation genetic testing for aneuploidy (PGT-A). Furthermore, the analysis of transfer outcomes following vitrification and warming of blastocyst stage embryos is highly relevant since this approach is reflective of current practice in many fertility centers today.

This study seeks to determine if variation in endometrial thickness impacts clinical pregnancy and live birth rates. By investigating these outcomes within several categories of endometrial thickness, rather than by simply analyzing outcomes above or below a single thickness cutoff, this study delineates the effect of slight fluctuations in endometrial thickness. It has been well documented that a thin endometrial lining is associated with adverse pregnancy outcomes, but this study seeks to clarify the effect of progressively increasing endometrial thickness on pregnancy outcomes. This study’s focus on outcomes across various thicker endometrial lining measurements differentiates it from prior publications and provides a novel contribution to the literature. A primary goal of this publication is to further define the optimal lining thickness at which to perform a frozen embryo transfer.

Materials and Methods

Patients

This retrospective study took place at a single fertility center with procedures performed by six reproductive endocrinology and infertility specialists. It was determined that institutional review board approval was not required as the data that was analyzed was deidentified, retrospective, and there was no experimental intervention. Routine comprehensive informed consent was obtained prior to IVF treatment and preimplantation genetic testing for aneuploidy (PGT-A) testing of embryos.

Patients were included if they underwent a frozen embryo transfer of a single euploid embryo during a medicated cycle of an autologous or donor oocyte derived embryo between January 2015 and December 2020 at the study center. Only the first cycles at the study center were included. Gestational carriers and natural cycle frozen embryo transfers were excluded.

To evaluate the uterus, all patients underwent pelvic ultrasound with 3D evaluation of uterus and office hysteroscopy to ensure a normal uterine cavity. Ovarian stimulation protocols were developed by the primary physician based on age, ovarian reserve testing, and diagnosis. Oocyte retrieval, embryo culture, and PGT-A was performed as in previously established protocols (12).

Endometrial preparation and embryo transfer

The endometrium was prepared with transdermal estradiol patches at a dose of 0.05 mg/day, with patches changed every other day, increasing at the discretion of the primary physician up to four patches per day, with supplementation of estradiol in 2 mg dose increments orally or vaginally in cases where endometrial development was inadequate as described in detail elsewhere (13). Endometrial thickness was measured prior to starting progesterone using General Electric (GE) Voluson S8 ultrasound machines. An exact endometrial thickness measurement was obtained in the sagittal plane at the point of maximum thickness of the endometrial lining. The measurement was made in a perpendicular line to the axis of the central hyperechoic line generated by the apposing surfaces of the anterior and posterior endometrial walls. Patients were treated with five or six days of intramuscular progesterone prior to embryo transfer, as described (13). Specifically, patients were given 50 mg progesterone in oil every other day in addition to vaginal progesterone 100 mg three times daily. The dose of progesterone in oil was increased to daily, up to 75 mg daily, if serum progesterone levels were found to be less than 20 ng/mL two days prior to embryo transfer.

Embryo transfer of a single euploid embryo was performed under ultrasound guidance. Serum β-hCG was tested 9 days after embryo transfer. For patients with a positive serum β-hCG, clinical pregnancy was confirmed approximately two weeks after appropriate increase in serial β-hCG levels. Progesterone supplementation was continued through 10 weeks gestation. Estradiol supplementation was tapered after the pregnancy confirmation ultrasound. Patients were discharged from fertility care to an obstetrician at approximately 7 weeks’ gestation and managed at the discretion of the obstetrician throughout the pregnancy. They were then contacted via phone or electronic correspondence by the study center following their expected date of delivery in order to ascertain pregnancy outcomes and delivery details.

Diagnosis and outcome definitions

Diminished ovarian reserve was defined as an anti-müllerian hormone (AMH) level <1.10 ng/mL or an antral follicle count (AFC) of <5. Recurrent pregnancy loss (RPL) was defined as two or more previous clinical pregnancy losses, in line with criteria outlined by the American Society for Reproductive Medicine (ASRM). Uterine factor infertility was defined as the past or current presence of intrauterine adhesions diagnosed during hysteroscopy, the presence of one or more leiomyomas documented on ultrasound > 2 cm in size, or the presence of a müllerian anomaly. Patients with a history of uterine septum which had undergone excision were included in the uterine factor category. Some patients had more than one diagnosis (i.e. diminished ovarian reserve and endometriosis). Patients listed as “donor oocyte” had transfer of an embryo that was developed from an oocyte donated by a donor who met Food and Drug Administration criteria for eligibility. Donor sperm was used as a diagnosis for patients using sperm from an eligible donor who was not an intimate partner (14). Polycystic ovary syndrome (PCOS) was defined using Rotterdam criteria (15). Patients with a diagnosis of endometriosis were diagnosed by laparoscopy. Male factor infertility was defined as one or more abnormal features of a semen analysis to include strict morphology, count, and motility. Advanced maternal age (AMA) was defined as age >35 at time of embryo creation and/or embryo transfer. Tubal factor infertility was defined as tubal occlusion either unilaterally or bilaterally diagnosed by hysterosalpingogram due to an acquired condition. Hypothalamic infertility was defined by ovulatory dysfunction with follicle stimulating hormone and luteinizing hormone levels below the reference range. Other ovulatory disorders included patients that did not receive a diagnosis of PCOS or hypothalamic dysfunction but had persistent oligo-ovulation. Unexplained infertility was diagnosed in patients with a normal infertility evaluation that excluded any of the above conditions or other identifiable causes.

Biochemical pregnancy was defined as the presence of a positive serum βhCG (>5 mIU/mL) followed subsequently by a decline in βhCG level prior to visualization of a gestational sac on ultrasound. First trimester loss was defined as fetal demise before 14 weeks, but after ultrasound evidence of an intrauterine pregnancy with an embryo. Clinical pregnancy was defined as an intrauterine pregnancy with an embryo and fetal cardiac activity on ultrasound. Live birth was defined as the delivery of a liveborn infant.

Statistical Methods

Outcomes of interest included negative pregnancy, biochemical pregnancy, clinical pregnancy, and live birth. These outcomes and additional demographics and characteristics were summarized using descriptive statistics for all patients. Endometrial thickness was treated as a continuous variable as well as grouped into five categories: < 8mm, 8–10mm, 10–13mm, 13mm-15mm and ≥15mm. Endometrial thickness was also summarized by infertility diagnosis. To investigate the effect of endometrial thickness on pregnancy outcomes, unadjusted logistic regression models were used. Odds ratios and 95% confidence intervals (CI) were obtained. All reported p-values are two-sided and a significance level of 0.05 was used. Statistical analyses were performed using R (version 4.1.2, R Core Team).

Results

Patient characteristics

A summary of demographic information and diagnoses for the 796 patients meeting inclusion and exclusion criteria is in Table 1. The majority of patients were nulliparous (72.5%). Mean age was 35.9 (SD 4.3), reflecting many patients being of advanced maternal age (AMA, age greater than 35 years). Mean BMI was 26.3 (SD 5.4), and 49.3% of patients were normal weight, with 26.3% and 22.8% or patients being overweight or obese, respectively.

Table 1.

Summary of characteristics for all patients.

Variable	All patients (N=796)
Age
Mean (SD)	35.9 (4.3)
Gravidity
Mean (SD)	1.1 (1.5)
Parity
Mean (SD)	0.3 (0.6)
Parity, n (%)
Nulliparous	577 (72.5%)
Multiparous	219 (27.5%)
BMI1
Mean (SD)	26.3 (5.4)
BMI1, n (%)
Underweight	13 (1.6%)
Normal weight	392 (49.3%)
Overweight	209 (26.3%)
Obese	181 (22.8%)
Diagnosis2: n (%)
Unexplained infertility	115 (14.4%)
PCOS	115 (14.4%)
AMA	485 (60.9%)
Endometriosis	61 (7.7%)
Male factor	154 (19.3%)
Tubal factor	55 (6.9%)
Diminished ovarian reserve	150 (18.8%)
Translocation carrier/genetic, desires PGT-A or -M	35 (4.4%)
Recurrent pregnancy loss	81 (10.2%)
Donor sperm	21 (2.6%)
Hypothalamic or ovulatory infertility	28 (3.5%)
Donor oocyte	63 (7.9%)
Uterine factor	17 (2.1%)
Endometrial lining thickness
Mean (SD)	10.3 (1.9)
Endometrial lining thickness, n (%)
<8 mm	33 (4.1%)
8–10 mm	362 (45.5%)
10–13 mm	320 (40.2%)
≥13 mm	81 (10.2%)

¹BMI was missing for 1 patient.

²More than one diagnosis could be listed for each patient. See Methods for further explanation of diagnoses.

We found that thicker endometrium (≥13) is present in roughly 10% of patients. Patients with a thin endometrium (< 8mm) were uncommon at 4.1%, which likely reflects additional efforts to promote endometrial development and thus few patients undergoing an FET with an endometrium less than 8mm as well as cycle cancellation when these efforts are unsuccessful (Table 1).

Pregnancy outcomes

After transfer of a single euploid embryo, 17.4% of patients had a negative hCG and 10.7% of patients developed a biochemical pregnancy, reflecting factors extrinsic to the ploidy of the embryo that negatively impact pregnancy rates. As expected for euploid embryos, the clinical pregnancy rate was high (71.2%). Due to a percentage of mainly first trimester losses, 63.9% of patients achieved a live birth (Table 2).

Table 2.

Summary of pregnancy outcomes for all patients.

Variable	All patients (N=796)
Biochemical pregnancy, n (%)	85 (10.7%)
Clinical pregnancy, n (%)	567 (71.2%)
Live birth, n (%)	507 (63.9%)
1st trimester loss, n (%)	67 (8.5%)
Negative hCG, n (%)	138 (17.4%)

¹Biochemical pregnancy and live birth were missing for 2 patients. Negative hCG was missing for 3 patients, and 1^st trimester loss was missing for 5 patients. See Methods for further explanation of pregnancy outcome definitions.

Endometrial thickness and infertility diagnosis

Because this study includes several different infertility diagnoses, including those requesting PGT-A for reasons other than infertility, we sought to determine if there were differences in endometrial lining development across diagnoses. The mean endometrial thickness for each diagnosis was between 9.3 and 11mm, demonstrating equivalent endometrial development across diagnoses in medicated FET cycles. The majority of patients had an endometrial lining thickness between 8mm and 13mm. Patients with hypothalamic dysfunction were less likely to achieve an endometrium ≥10mm, potentially related to low levels of endogenous estrogen, however this would not be expected to be notable in a medicated cycle (Table 3).

Table 3.

Endometrial lining thickness by diagnosis for all patients.

Variable	UEI (N=115)	PCOS (N=115)	AMA (N=485)	Endo-metriosis (N=61)	Male factor (N=154)	Tubal factor (N=55)	DOR N=150)	Trans-location/genetic (N=35)	RPL (N=81)	Same sex or unpartnered (N=21)	Hypothalamic (N=28)	Donor egg (N=63)	Uterine factor (N=17)
Mean endometrial thickness (SD)	10.5 (2.0)	10.3 (2.0)	10.3 (2.0)	10.8 (2.0)	10.3 (1.8)	10.5 (2.2)	10.0 (1.7)	9.8 (1.7)	10.3 (2.0)	11.0 (2.1)	9.3 (1.9)	10.0 (1.8)	10.0 (2.0)
Median endometrial thickness (Range)	10.3 (6.6, 17.1)	10.0 (6.7, 14.7)	10.0 (6.0, 19.4)	10.6 (7.7, 17.2)	10.0 (6.0, 16.0)	10.2 (7.0, 16.0)	9.6 (6.7, 14.9)	9.6 (5.4, 13.6)	9.8 (6.0, 19.4)	10.5 (8.0, 15.3)	8.8 (7.5, 16.4)	9.6 (6.8, 13.7)	9.7 (6.7, 14.1)
Endometrial lining thickness, n (%)
<8 mm	4 (3.5%)	5 (4.3%)	24 (4.9%)	1 (1.6%)	6 (3.9%)	1 (1.8%)	10 (6.7%)	2 (5.7%)	1 (1.2%)	0 (0.0%)	1 (3.6%)	7 (11.1%)	2 (11.8%)
8–10 mm	43 (37.4%)	52 (45.2%)	218 (44.9%)	26 (42.6%)	71 (46.1%)	24 (43.6%)	73 (48.7%)	18 (51.4%)	41 (50.6%)	8 (38.1%)	22 (78.6%)	29 (46.0%)	7 (41.2%)
10–13 mm	51 (44.3%)	44 (38.3%)	196 (40.4%)	25 (41.0%)	64 (41.6%)	21 (38.2%)	58 (38.7%)	14 (40.0%)	33 (40.7%)	9 (42.9%)	3 (10.7%)	23 (36.5%)	6 (35.3%)
≥13 mm	17 (14.8%)	14 (12.2%)	47 (9.7%)	9 (14.8%)	13 (8.4%)	9 (16.4%)	9 (6.0%)	1 (2.9%)	6 (7.4%)	4 (19.0%)	2 (7.1%)	4 (6.3%)	2 (11.8%)

¹Patients could have more than one diagnosis associated with infertility.

Endometrial thickness ≥8mm is associated with higher pregnancy rates

To determine the optimal endometrial thickness for higher live birth rate during medicated FET cycles, we compared outcomes in each endometrial thickness group (<8mm, 10–13mm, 13–15mm, and ≥15mm) to patients with an endometrium measuring 8mm-10mm, a lining thickness that many patients achieve and would be considered adequate by the study site. Patients with an endometrium <8mm had higher rates of biochemical pregnancy (OR 2.72, 95% CI 1.08, 6.24) and lower clinical pregnancy rates (OR 0.43, 95% CI 0.21, 0.9). Clinical pregnancy rate and live birth rate were otherwise equivalent across groups (10–13mm, 13–15mm, and ≥15mm). Rates of 1^st trimester pregnancy loss once an intrauterine pregnancy was visualized on ultrasound were similar across all groups (Table 4).

Table 4.

Summary of pregnancy outcomes by endometrial lining thickness using descriptive statistics.

Variable	<8mm (N=33)	8–10mm (N=362)	10-13mm (N=320)	13-15mm (N=70)	≥15mm (N=11)	<8mm vs 8–10mm* OR	10–13mm vs 8–10mm* OR	13–15mm vs 8–10mm* OR	≥15mm vs 8–10mm* OR	P-value1
Biochemical pregnancy n (%)	8 (24.2%)	38 (10.5%)	31 (9.7%)	5 (7.1%)	3 (27.3%)	2.72 (1.08, 6.24)	0.91 (0.55, 1.51)	0.66 (0.22, 1.58)	3.19 (0.68, 11.56)	0.073
Clinical pregnancy n (%)	17 (51.5%)	257 (71.0%)	236 (73.8%)	49 (70.0%)	8 (72.7%)	0.43 (0.21, 0.9)	1.15 (0.82, 1.61)	0.95 (0.55, 1.7)	1.09 (0.31, 5.05)	0.149
Live birth n (%)	15 (45.5%)	226 (62.6%)	215 (67.2%)	44 (63.8%)	7 (63.6%)	0.5 (0.24, 1.02)	1.22 (0.89, 1.68)	1.05 (0.62, 1.82)	1.05 (0.31, 4.05)	0.17
1st trimester loss n (%)	3 (9.1%)	36 (10.0%)	22 (6.9%)	5 (7.1%)	1 (9.1%)	0.9 (0.21, 2.69)	0.67 (0.38, 1.15)	0.69 (0.23, 1.68)	0.9 (0.05, 4.9)	0.68
Negative hCG n (%)	8 (24.2%)	63 (17.5%)	52 (16.3%)	15 (21.4%)	0 (0.0%)	1.51 (0.61, 3.37)	0.92 (0.61, 1.37)	1.29 (0.66, 2.37)	N.E.	0.184

¹To investigate the effect of endometrial lining thickness on pregnancy outcomes, unadjusted logistic regression models were used. Odds ratios and 95% confidence intervals were obtained. Bold indicates significant comparison to the reference level of 8–10mm.

Discussion

Our study contributes to the body of literature surrounding endometrial thickness in a novel manner by analyzing outcomes of entirely medicated FET cycles that followed a standardized protocol, culminating in the transfer of single, vitrified-warmed euploid embryos, thus eliminating confounders such as poor ovarian response, supraphysiologic ovarian hormone levels present in fresh embryo transfer cycles, and the uncertain reproductive potential of embryos that are transferred without undergoing PGT-A. The inclusion criteria defined in the current study are in line with modern clinical practice in many fertility centers, making the findings of this study highly relevant for practitioners as they counsel patients and make clinical decisions related to embryo transfer. Many prior studies have evaluated the impact of endometrial thickness in the setting of cleavage stage embryo transfers. While previous studies’ findings are certainly useful, the updated investigation provided in this study regarding the impact of endometrial thickness on pregnancy outcomes in the setting of current practice is crucial. Given differences in patient population and protocols for individual practices, it is possible that a specific measurement (e.g., <8mm) may be less clinically useful than a percentile cutoff (e.g., 95^th percentile and 5^th percentile).

The differences in outcomes between thin (<8mm) and thick (≥15mm) endometrium in our study were not statistically significant, likely due to the small sample size in these groups. The elevated rate of biochemical pregnancies in patients with an endometrium <8mm may suggest that the microarchitecture of a thin endometrium does not support sustained implantation. This could be secondary to conditions such as chronic endometritis or inadequate blood flow to the endometrium. However, our data suggest that the impact of an endometrium <8mm may be less harmful in medicated FET cycles for euploid embryos compared to prior studies. Our ≥15mm cohort was quite small, and there were no statistically significant associations, but there was a trend toward a higher biochemical pregnancy rate in this group as well. One of the unique aspects of the current study is the analysis of pregnancy outcomes over multiple ranges of endometrial thickness, rather than an analysis of outcomes above or below a single thickness threshold. This aspect of study design allows for an interpretation of the effect of slight variations in thickness and holds immense value from a patient counseling perspective. The prevalence of FET cycles is increasing, thus it is important for future studies to assess LBR in this population, as there may be unique qualities about the endometrium developed during fresh embryo transfers compared to FET, which has been suggested in the literature (16, 17). As FETs become more common, a multi-center prospective study confirming our data as well as adverse obstetric outcomes such as morbidly adherent placenta or large for gestational age fetuses would be instructive.

One strength of this study is the uniform requirement for uterine evaluation via 3D ultrasonography within one year of embryo transfer as well as direct visualization with hysteroscopy. Direct visualization confers the ability to identify subtle intrauterine pathology including endometrial polyps that lead to thicker endometrial linings and lower implantation and pregnancy rates (18). Our study also took place in a single center with a limited number of sonographers performing ultrasounds, as well as a limited number of physicians interpreting the ultrasounds, potentially reducing observer bias with endometrial evaluation.

Endometrial histology is not available due to lack of feasibility and other assays to assess endometrial receptivity were not routinely performed. This is a limitation of the current study, and endometrial receptivity testing may be helpful in some patients but is unlikely to be a panacea (19, 20). Whether the microarchitecture or the milieu of soluble and membrane-bound factors (or both) contributes to lower pregnancy rates with decreased or increased endometrial thickness has yet to be determined. A larger study with robust thin (<8mm) and thick (≥15mm) endometrium cohorts would further explore this potential association. Continued advancements in embryo culture and evaluation have improved outcomes for patients, however, a more detailed understanding of the endometrium will ultimately maximize the reproductive potential of each embryo transferred. In euploid SET cycles, our data support that if an endometrial thickness of at least 8mm is achieved, there is not a significant difference in outcomes across a wide variation of measurements.

Availability of data and material (data transparency):

Raw data can be made available upon request.