The impact of estradiol supplementation on endometrial thickness and intrauterine insemination outcomes.

A total of 2,281 completed IUI cycles from March 2017 through March 2023 were included in the study. The overall median age for the entire study was 37 years old; the median age for the E2 cohort was 38 years compared to 37 years for the no E2 cohort (p=0.001). Both cohorts had similar BMIs, race/ethnicity, total motile sperm count, and number of IUI cycles completed ( Table 1 ). The two most common types of IUI cycles for the entire study population were letrozole (39.9%) and clomiphene (38.8%), and the E2 cohort had a higher proportion of letrozole IUI (50.2% vs. 38.3% for no E2) and lower proportion of natural cycle IUIs (3.2% vs 15.4% for no E2) (p<0.0001). With regards to infertility diagnosis, the two cohorts overall had a similar proportion of diagnoses, except that the E2 cohort had a statistically significant lower prevalence of “male factor”, higher prevalence of “single female”, higher prevalence of “recurrent pregnancy loss”, and lower prevalence of “female sexual dysfunction” diagnoses ( Table 1 ). The reference cohort that received no E2 was further categorized into cycles with thin pre-ovulatory lining (EMT <7mm) and normal pre-ovulatory lining (EMT ≥7mm) for analysis of outcomes. At baseline, the E2 cohort had a statistically significantly lower median EMT of 4.2mm compared to 4.6mm for the thin-lining reference (p=0.001) and 4.8mm for normal-lining reference (p<0.0001) ( Table 2 ). At ovulation trigger, the E2 cohort had a median EMT of 6.6mm that was greater than the thin-lining reference cohort’s median EMT of 6.4mm (p<0.0001), but significantly less than the normal-lining reference cohort’s median EMT of 8.4mm (p<0.0001). With regards to changes in EMT from baseline to ovulation trigger, the E2 cohort had a statistically significant greater change in EMT from baseline to ovulation trigger compared to the thin-lining reference cohort (2.4cm vs 1.9cm, P=<0.0001). However, when compared to the normal-lining reference cohort, the E2 cohort had a statistically significant smaller change in EMT from baseline to trigger (2.4cm vs 3.8cm, P=<0.0001). In subanalysis stratifying IUI type by letrozole and clomiphene, the observed EMT patterns persisted for letrozole cycles, but not clomiphene cycles ( Supplemental Table 1 ). The median change in EMT for the E2 cohort was 3.0mm for the letrozole IUI cycles and 1.7mm for the clomiphene IUI cycles. For clomiphene cycles, the E2 cohort had a lower change in EMT from baseline to trigger compared to both the thin and normal-lining reference cohorts; median change in EMT for E2 was 1.7mm compared to 1.9mm for thin-lining reference which was not statistically significant (p=0.88) and compared to 3.7 for normal-lining reference which was statistically significant (p<0.0001) With regards to pregnancy outcomes, there were no significant differences in the rates of positive serum βhCG, clinical pregnancy, and live birth rates between the E2 and no E2 reference groups ( Table 3a ). The clinical pregnancy rate for E2 cohort was 14.4% compared to the thin-lining reference cohort’s 13.8% (p=0.41) and to the normal-lining reference cohort’s 15.3% (p=0.35). The live birth rate for E2 cohort was 7.8% compared to the thin-lining reference cohort’s 8.6% (p=0.68) and to the normal-lining reference cohort’s 10.7% (p=0.15). There were also similar rates of biochemical and ectopic pregnancy. There was a trend towards higher rate of miscarriage for the E2 cohort with a rate of 40.1%, compared to 30.7% for the thin-lining reference cohort (p=0.26) and to the normal-lining cohort’s rate of 26.2% (p=0.05, Table 3a ). Furthermore, when the GEE models were adjusted for potential confounders including age, BMI, race/ethnicity, infertility diagnosis, total motile sperm count, and EMT at trigger, the E2 cohort had a statistically significant increased odds of miscarriage compared to the reference no E2 cohort (adjusted OR 2.46, 95% CI [1.18, 5.14], p=0.02, Table 3b ). In subanalysis stratifying IUI type, the letrozole group exhibited similar rates of clinical pregnancy, miscarriage, and live birth between the E2 and no E2 reference cohorts. The clomiphene E2 cohort had a greater magnitude of difference in its higher miscarriage rate of 44.0% compared to the normal-lining cohort’s rate of 26.4% (p=0.09), and comparison with the thin-lining reference cohort’s miscarriage rate of 32.0% was also not statistically significant (p=0.38) ( Supplemental table 1 ).

This study is a retrospective chart review of IUI cycles completed at Stanford Fertility and Reproductive Medicine Center from March 2017 through March 2023. All monitored IUI cycles completed by patients aged 18 to 45 years old with a normal uterine cavity were included for a total of 2,281 cycles: 309 cycles had E2 supplementation and 1,972 reference cycles had no E2. The reference cohort was further categorized into cycles with thin-lining prior to ovulation (EMT <7mm, n=536) and normal-lining (EMT ≥7mm, n=1,436). Uterine cavity evaluations included hysterosalpingogram, saline infusion sonohysterography, hysteroscopy, or transvaginal pelvic ultrasound. Demographic and clinical characteristics, fertility treatments, and pregnancy outcomes were collected from medical records. Baseline characteristics included maternal age, race/ethnicity, body mass index (BMI; kg/m2), infertility diagnosis, type of IUI, and total motile sperm count. EMT at baseline, mid-cycle, and ovulation trigger ultrasound scans were collected. Serum levels of beta human chorionic gonadotropin (βhCG) were evaluated around 2 weeks after an IUI procedure. The Stanford University Institutional Review Board approved the study protocol. The primary outcomes studied were the differences in EMT measurements as well as EMT growth between IUI cycles with and without E2 supplementation. The secondary outcomes studied were the differences in rates of positive serum βhCG (>5 mIU/mL), clinical pregnancy, clinical miscarriage, and live birth. Clinical pregnancy was defined as a viable intrauterine pregnancy with fetal cardiac activity, biochemical pregnancy was defined as rise and fall of serum βhCG without a gestational sac ever visualized, ectopic pregnancy was defined as a pregnancy that occurred outside of the uterine cavity, clinical miscarriage was defined as pregnancy loss prior to 20 weeks of gestation, and live birth was defined as live infant born on and after 24 weeks of gestation. The standard protocol for intrauterine insemination started with a baseline ultrasound evaluation on cycle days 2 to 5. Thereafter, patients monitored ovulation with home detection kits and/or clinic ultrasounds. Patients undergoing medicated cycles used either letrozole (2.5–7.5mg daily starting on cycle days 2–3 for 5 days), clomiphene citrate (50–100mg/day starting on cycle day 2–3 for 5 days), or gonadotropins (150–225 IU starting on cycle day 2–3 until trigger). Around 7–9 days after the initiation of ovulation induction medications, ultrasound monitoring was performed until a follicle achieved at least 18mm in size, after which ovulation was triggered with either hCG (5,000 IU) or recombinant hCG (250 mcg, Ovidrel, EMD Serono, Darmstadt, Germany). IUI was performed around 36 hours after ovulation. Patients in the estradiol supplementation cohort started vaginal micronized estradiol (Estrace, 2mg tablets twice a day per vagina, Teva Pharmaceuticals, North Wales, Pennsylvania, USA) in the follicular/proliferative phase if the lining thickness was 15mm; this was continued until they were confirmed to either have a negative pregnancy test or viable intrauterine pregnancy (presence of fetal cardiac activity). REDcap was utilized to manage the study data [ 12 ], and analyses were performed using SAS Version 9.4 statistical software (SAS Institute Inc., Cary, North Carolina, USA). Continuous variables were presented as median and interquartile range and compared using Wilcoxon rank-sum test. Categorical variables were presented as frequencies and percentages and compared using chi-squared or Fisher’s exact test. Overall pregnancy rate (positive serum pregnancy test) was calculated per total number of completed IUI cycles. Clinical pregnancy was calculated per total number of completed IUI cycles. Clinical and biochemical miscarriage rates were calculated per total number of pregnancies. Live birth was calculated per total number of completed IUI cycles. Patients undergoing multiple cycles were accounted for utilizing generalized estimating equations (GEE) to assess associations between E2 supplementation and IUI outcomes (clinical pregnancy, clinical miscarriage, and live birth). Odds ratios (ORs) and 95% confidence intervals (CIs) were reported, and models were adjusted for age, BMI, race/ethnicity, infertility diagnosis, and EMT at trigger. All statistical tests were two-tailed and p<0.05 indicated statistical significance.

Our study demonstrated that IUI cycles with E2 supplementation during the follicular/proliferative phase had a significantly greater change in EMT than unsupplemented cycles with thin pre-ovulatory lining (EMT <7mm). Such difference was not observed when compared to unsupplemented cycles with normal-lining (EMT ≥7mm). However, when examining pregnancy outcomes, there were no significant differences in clinical pregnancy and live birth rates between the cohorts regardless of pre-ovulatory EMT. There was an overall two-fold increased adjusted odds of miscarriage for the E2 supplementation cohort, with a greater magnitude of difference in miscarriage rates observed in clomiphene IUI cycles upon subanalysis stratifying by IUI cycle type. Overall, these results suggest that E2 supplementation may be beneficial in increasing the lining thickness in patients with thinner pre-ovulatory EMTs, but this did not clinically translate into improved IUI and pregnancy outcomes in our study. Endometrial development in the follicular phase of the menstrual cycle is driven by E2 production from the ovaries, which stimulates the growth of epithelial and stromal cells within the endometrium [ 14 ]. As such, follicular growth and EMT are routinely monitored during assisted reproduction cycles as key factors in clinical decision-making. Furthermore, E2 also plays an important role in the transition of the endometrium from a proliferative to a secretory state as it works with progesterone to mediate the endometrial preparation for implantation [ 15 – 19 ]. In fact, prior studies have observed higher pregnancy rates in IUI cycles with higher E2 levels on day of ovulation trigger [ 12 , 20 – 21 ]. Additionally, studies have demonstrated improved pregnancy outcomes with increasing endometrial thickness. Wolff et al.’s cohort study of IUI cycles found pregnancy rates to increase gradually with increasing endometrial thickness through 10 mm, after which the association plateaus [ 12 ]. Similarly, Quaas et al.’s large prospective cohort analysis of a multicenter randomized controlled trial found evidence of decreasing live birth rates with decreasing EMT [ 11 ]. Given that E2 drives the development of the endometrial lining and that EMT is thought to impact pregnancy outcomes, it stands to reason that E2 supplementation may help improve pregnancy outcomes by increasing EMT. In a small RCT comparing the effectiveness of adding E2 in patients undergoing clomiphene IUI, Gerli et al. found that short (5 days) E2 supplementation together with clomiphene administration was associated with increased EMT, decreased miscarriage rate, and increased pregnancy rate, thereby potentially reversing the deleterious effect of clomiphene on endometrial thickness [ 22 ]. However, this study was limited by its small sample size of 64 cycles. Our study of 2,281 IUI cycles (886 of which used clomiphene) did not observe a significant difference in pregnancy rates and actually observed an increased odds of miscarriage in the E2 supplementation cohort. This could be driven by the fact that the E2 cohort was on average older and had thinner endometrial lining, though both maternal age and EMT at trigger were adjusted for in the odds ratio analysis. Thus far, there have been no published studies that have implicated estrogen supplementation with increased miscarriage risk. However, there have been a few studies that have demonstrated an association between thin EMT and increased rate of miscarriage. Luo et al.’s large retrospect cohort study examining factors associated with spontaneous miscarriage in IUI observed significantly increased rates of miscarriage in thinner EMT [ 23 ], while Mahutte et al’s large cohort examining IVF outcomes found increasing rates of pregnancy loss with decreasing EMT [ 24 ]. Upon subanalysis stratifying by letrozole and clomiphene IUI types, the higher incidence of miscarriage seemed to be largely driven by clomiphene IUIs given that the observed differences in miscarriage rates were higher for the clomiphene E2 cohort. This could be due to clomiphene’s anti-estrogenic effect that may be detrimental to the development of the endometrial lining [ 3 – 6 , 25 ]. In fact, the E2 cohort’s higher EMT change from baseline to trigger compared to the thin-lining group was not observed in the subanalysis of clomiphene IUI cycles, suggesting that perhaps estradiol supplementation cannot overcome the anti-estrogenic effect of clomiphene. Nevertheless, E2 supplementation during the follicular phase in IUI cycles has not been well-studied. Thus far, E2 supplementation has largely been studied in IVF/ICSI cycles and in the context of luteal phase support. In multiple older retrospective cohort studies examining IVF and ICSI cycles, E2 supplementation during the luteal phase has been shown to have a higher pregnancy rate [ 26 – 28 ], but it is important to note that these studies utilized oral E2, which is thought to be less effective [ 29 , 30 ] and not as generalizable to contemporary clinical practice utilizing vaginal forms. Furthermore, more recent studies do not support the use of E2 supplementation in the fresh IVF cycle. In an RCT examining the use of luteal phase vaginal E2 supplementation on day of fresh embryo transfer after egg retrieval, Engmann et al. found no significant differences in implantation, clinical pregnancy, and ongoing pregnancy rate [ 31 ]. Gelbaya et al. performed a systematic review and meta-analysis to investigate the effect of E2 supplementation during IVF and similarly found no significant differences in implantation and pregnancy rates [ 30 ]. Thus, the Canadian Fertility and Andrology Society’s 2019 clinical practice guideline for the management of thin endometrium suggested against the use of luteal estradiol given the lack of convincing evidence [ 33 ]. To our knowledge, this is the first study examining the effect of exogenous estradiol supplementation during the follicular (proliferative) phase on both the endometrium and subsequent pregnancy outcomes for women undergoing IUI. It included all completed IUI cycles at our institution within a 6-year period, providing a robust sample size and minimizing selection bias. All patient and cycle information were gathered from medical records to minimize recall bias. The main limitation of our study was its observational and nonrandomized nature. Thus, randomized controlled trials would be valuable to further understand the true causal relationship between estradiol supplementation in the follicular phase of IUI cycles and pregnancy outcomes. Given that our data was from patients within a single academic center, there are also limitations in the generalizability of our data given that our patients overall had normal median BMI and were predominantly of Caucasian or Asian race. Another limitation, which is based on current clinical practices, is the absence of serum estradiol level data in IUI cycles. In efforts to maximize pregnancy outcomes for patients, there has been an ever-expanding menu of adjunctive treatments from which clinicians may “add-on” to standard protocols within assisted reproductive technology. In line with the 2021 systematic review by the Cochrane Gynaecology and Fertility Group summarizing the effectiveness of IVF add-ons [ 34 ], it is important that our field critically analyzes treatments that go beyond the evidence-based protocols. Despite the pathophysiologic plausibility and theoretical value of estradiol supplementation in the follicular phase, our study suggests that although exogenous estradiol may improve the thickness of the endometrial lining – especially for IUI cycles with thin (<7mm) endometrial lining – this does not translate into improved pregnancy and live birth rates. Our hope is that this study provides helpful evidence to guide patient counseling during a critical shared decision-making process.

Ovarian stimulation and ovulation induction have long been used in conjunction with intrauterine insemination (IUI) and is considered first line treatment for patients with unexplained or male-factor infertility [ 1 ]. Options for ovarian stimulation include anti-estrogens such as clomiphene citrate, aromatase inhibitors such as letrozole, or gonadotropins such as human menopausal gonadotropins (HMG), purified urinary follicle stimulating hormone (FSH) or recombinant FSH [ 2 ]. These medications not only stimulate follicular development in the ovary, but also impact the proliferation of the endometrium. In fact, the anti-estrogenic effect of clomiphene has been thought to negatively impact the endometrial lining thickness [ 3 – 6 ]. Adequate endometrial lining development is believed to play a crucial part in embryo implantation, placental development, and subsequent obstetric outcomes including preeclampsia, placental abruption, placenta previa, small for gestational age neonate, and preterm birth [ 7 ]. The clinical significance of endometrial thickness (EMT) on clinical pregnancy and live birth rates in IUI and IVF cycles is still being debated with conflicting results [ 8 – 12 ]. Nevertheless, assessments of the endometrium including serial monitoring of EMT have become a standard part of treatment cycles. Endometrial development in the follicular/proliferative phase is driven by estradiol (E2) production from the ovaries [ 8 ], and thus both E2 and follicular growth are also routinely factored into clinical decision making with regards to hormonal stimulation and ovulation induction. When the endometrium is found to be “thin” during an IUI cycle – thin being defined by various lining thresholds in the literature – physicians are often faced with the question of how they can further optimize the cycle, and potentially even more importantly, the difficult decision of whether to proceed with IUI or cancel the cycle all together. The utility of estrogen supplementation in IUI cycles remains largely unstudied. Thus, to our knowledge, this is the first study to examine the effect of exogenous estradiol supplementation on pregnancy outcomes for women undergoing IUI treatment.