Introduction

Establishment of pregnancy in human IVF requires transfer of morphologically adequate embryos into the uterus. This is achieved by the retrieval of multiple cumulus-oocyte complexes (COC) after ovarian stimulation. Ovarian stimulation is necessary for multiple follicular developments and is accompanied by supraphysiological serum estradiol levels. Assessment of the role of estradiol levels for IVF outcome has been the focus of interest for many years.

At present, the importance of supraphysiological estradiol levels on the day of hCG administration for the probability of pregnancy in IVF remains unclear [1]. The role of estradiol in the luteal phase is also not clear and it is suggested that it only plays a permissive role [2, 3].

Significantly lower implantation and pregnancy rates were shown in cycles with high serum estradiol concentrations on the ovulatory (hCG) day [4–6]. The proposed mechanism of possible adverse effects of elevated estradiol levels is the altered endometrial receptivity [5–9]. Valbuena et al. [10] suggested that high E2 levels are deleterious to embryo adhesion in vitro, mainly because they have a direct toxic effect on the embryo that may occur at the cleavage stage. On the contrary, Bianco et al. concluded that elevated E2 levels in donors were not found to affect pregnancy outcome in oocyte donation cycles. This suggests elevated E2 levels do not compromise oocyte quality or embryo development in vitro, but that elevated E2 levels may diminish endometrial receptivity [11].

The impact of elevated E2 levels on the day of hCG administration (and of a high number of retrieved oocytes) in the subgroup of IVF patients known as “high responders” on implantation during assisted reproduction has been the subject of debate. Various studies showed that the different threshold levels above which a patient could be defined as a high responder make the meaningful interpretation confusing. High responders were defined variously in these studies as patients who had a peak E2 level of >3,000 pg/ml [7], >15 retrieved oocytes [5, 8] or >10 retrieved oocytes [9, 12].

Sharara and Mc Clamrock [3] found that the ratio of the day of hCG estradiol to mid-luteal estradiol was highly predictive of successful outcome in IVF-embryo transfer. The ongoing pregnancy and implantation rates (PR and IR, respectively) were significantly impaired if the above ratio was >5. They postulated that the endometrial integrity might become compromised when there was a dramatic decline in estradiol concentrations around the mid-luteal period as reflected by a high ratio.

The purpose of this retrospective study was to examine the effects of the peak and mid-luteal E2 levels, and the degree of decline in E2 levels from its peak level (on the day of hCG injection) to the lower levels of mid-luteal phase (day eight of hCG injection), as it is reflected by the E2 ratio (E2 level on the day of hCG injection/E2 on the day eight of hCG injection), on clinical PRs in an IVF-embryo transfer program.

Material and method

Of the patients who underwent IVF-ET at the outpatient clinic of Zekai Tahir Burak Woman’s Health Education and Research Hospital IVF Unit between June and October 2008, 106 normal and high responders to the first cycle of COH with GnRH-agonist were included in the study. Approval for the study was obtained from the local ethics committee of the hospital. Data were obtained from patient records.

Patients with unexplained infertility, ovulatory dysfunction, male and tubal factors were included in the study. For the patients included in the study, the causes of infertility were male factor 34.9% (n = 37), tubal factor 14.1% (n = 15), ovulatory dysfunction 14.1% (n = 15), and unexplained in 36.7% (n = 39) of cases.

Study inclusion criteria included:

  1. 1.

    A basal FSH hormone level <10 IU/L

  2. 2.

    Age between 19–38

  3. 3.

    Long protocol with GnRH-a and rFSH

  4. 4.

    First cycle of IVF treatment.

Exclusion criteria were poor responders (women who achieved an E2 level <500 pg/ml on the day of hCG administration and/or women in whom <5 oocytes were retrieved), endometriosis patients, frozen-thawed cycles and other stimulation protocols.

The patients were divided into two groups with respect to peak E2 levels (Group 1: <2,500 pg/ml, Group 2: >2,500 pg/ml). Also the patients were compared of the women who got pregnant and did not get pregnant and the patients were divided into three groups with respect to E2 ratios (peak E2 level/mid-luteal E2 level) (Group 1 = 0.6–2.5, Group 2 = 2.5–5, Group 3 = >5). A comparison between the groups was made regarding ovarian stimulation characteristics, fertilization, implantation and pregnancy rates.

All the patients received oral contraceptive pills starting 3 weeks previously to ensure ovarian quiescence. Subcutaneous 0.5 mg/dl Leuprolid acetate (Lucrin daily flacon, 1 mg Abott, Cedex, İstanbul) was used for the inhibition and it was started from mid-luteal phase of previous menstrual cycle. The ovarian stimulation was initiated when serum estradiol levels were lower than 50 pg/ml and follicles smaller than 10 mm. rFSH (Gonal F, Serono laboratories, Boulogne, France) was used for ovarian stimulation based 225 IU initiation dose on the 3rd day of vaginal bleeding. After the initiation of gonadotrophin stimulation, the dosage of lucrin was reduced to 0.25 mg/dl and continued until the day of oocyte retrieval. Dose alterations were performed on 4th day of stimulation and continuing days according to sonographic findings and E2 levels. Once three follicles, at least 18 mm in diameter, were observed, ovulation was induced by intramuscular injection of 10,000 IU hCG (Pregnyl-Organon). Oocyte pick up (OPU) was performed 34–36 h after hCG injection. Embryos were transferred from 72 to 78 h after OPU by Wallace catheter (Edwards-Wallace Catheter; Marlow Technologies, Willoughby, OH) under ultrasonography guidance. Luteal phase support was achieved with a vaginal progesterone gel (Crinone gel 8%, Ares-Serono SA, Geneva, Switzerland) beginning from the OPU day and continued at least until pregnancy was ruled out by a negative serum hCG measurement.

The biochemical pregnancy rate was detected with the confirmation of positive serum hCG 2 weeks after ET. The implantation rate was the proportion of embryo transferred resulting in an intrauterine gestational sac. Clinical pregnancies were detected with the confirmation of positive fetal cardiac activities by transvaginal sonography on sixth gestational week. Blood samples were drawn on the hCG day and on the eighth day of embryo transfer (ET) for E2 (peak and mid-luteal E2, respectively) and progesterone levels. E2 ratio was defined as peak E2 level/mid-luteal E2 level [3]. Miscarriage was defined as a loss of a clinical pregnancy before the 13th week of gestation.

Laboratory analysis

The serum levels of E2, FSH, and LH that were studied with Electrochemiluminescence Immunoassay “ECLIA” (Roche) were intended to use on Elecsys and cobas e 601 immunoassay analyzers. The analysis sensitivity of the assay was 5 pg/ml and linear interval of test was 500–4,300 pg/ml for estrogen. E2 levels were assayed with intra- and interassay coefficients of variation of <3.3 and <4.9%, respectively. FSH sensitivity assay 0.13 mIU/ml, range assay 0.13–200 mIU/ml, sensitivity assay 5% was measured. For serum FSH measurement,the intra-and interassay coefficients of variation were <2.8 and <4.5%, respectively.

Statistical analysis

Continuous data were expressed as mean ± SD and were analyzed with 2-sample t test and Mann–Whitney rank sum test. Categorical data were analyzed with Chi-square test or Fisher’s exact test where appropriate. All statistical analysis was performed by SPSS 15.0 package (SPSS, Chicago, IL, USA).

A probability value of <0.05 represented statistical significance.

Power analysis

We estimated 48 patients per group would be needed to show ±1 difference in E2 ratio was defined as peak E2 level/mid-luteal E2 level with in groups, assuming a statistical power of 95% at an alpha level of 0.05. Power analysis was performed using NCSS-PASS package program.

Results

The mean (± SD) patient age was 30.5 ± 5.9 years (range 19–38 years). The mean body mass index was 24.9 ± 3.2 kg (range 18.6–37.2 kg). The mean basal FSH level was 7.8 ± 2.2 IU/L (range 0.2–10 IU/L). The mean number of oocytes retrieved was 10.1 ± 4.4 (range 5–35). The mean peak E2 level was 2,697.4 ± 1,453 pg/ml (range 684–4,983 pg/mL).

Comparing E2 <2,500 pg/ml and E2 >2,500 pg/ml, there were no statistically significant differences in age, serum basal FSH level, fertilization rate, number of replaced embryos and PRs. The number of retrieved oocytes, luteal E2, peak E2 and E2 ratio were significantly higher in E2 >2,500 pg/ml group, however, the implantation rate was significantly lower in this group (Table 1).

Table 1 Outcome of IVF in 106 women grouped according to their peak E2 response as high responders or normal responders
Full size table

Figure 1 shows a sharp decline in the mid-luteal E2, in group >2,500 pg/ml. There were statistically significant differences in peak E2 levels, luteal E2, retrieved oocytes, E2 ratio; of the women who got pregnant and did not get pregnant, all the above parameters were significantly higher in non-pregnant group (Table 2).

Fig. 1
figure 1

Decline in the mid-luteal E2, in group >2,500 pg/ml and <2,500 pg/ml groups

Full size image
Table 2 Demographic and cycle characteristics of patients with or without established pregnancy
Full size table

When the patients were divided into three groups with respect to E2 ratios (Group 1 = 0.6–2.5, Group 2 = 2.5–5, Group 3 = >5). It was seen that a great majority of the patients (92.4%) were in groups 1 and 2. The clinical pregnancy rate was highest in group 1 and significantly lowest in group 3. The mean mid-luteal progesterone (hCG day-8) levels of these three groups were statistically significantly different, and a statistical relationship to group 1 with a p value of 0.001 was found (Table 3).

Table 3 Pregnancy and implantation rates in 106 patients grouped according to their E2 ratios (Group 1 = 0.6–2.5, Group 2 = 2.5–5, Group 3 = >5)
Full size table

A total number of 60 clinical pregnancies were noted with 3 missed abortions and 4 extra uterine pregnancies. In 4 women with a peak E2 level ≥2,500 pg/mL, moderate degree ovarian hyper stimulation syndrome (OHSS) was observed (7.8%). They recovered fully with expectant management, but they did not conceive.

Discussion

High serum E2 concentrations may be detrimental to the implantation and pregnancy rates in IVF treatment. The effect of such supraphysiologic E2 levels on the IVF-ET outcome has been the subject of many researches in the literature [13–17]. In a retrospective study by Pellicer et al. [9], where gonadotrophins without clomiphene citrate were used, a negative effect was observed based on numerous oocytes retrieved from 97 patients, constituting the first study in the GnRH agonist era. The patients in this study were divided into three groups based on the number of oocytes retrieved: 1–5, 6–10 and >10. When compared to women who had 1–5 oocytes retrieved, IRs were significantly reduced in the 6–10 oocytes retrieval group. Both fertilization rates and IRs were significantly lower in women with >10 oocytes [9]. Among women who had >10 retrieved oocytes, peak E2 levels were significantly higher as well as E2 levels on days 1 and 2 after hCG administration. The researchers suggested that impaired embryo quality may be the cause of the lower IRs and PRs [9]. Makkar et al. [18] demonstrated that a high serum E2 level caused a decrease interleukin-11 and interleukin-6 expression in the peri-implantation endometrium and the lower IRs and PRs in the high responders may due to this reduction.

High responders were defined as women who had >15 retrieved oocytes by the same group of researchers [5, 8]. Simon et al. [5] studied 59 high responders, and 105 normo-responders (≤15 oocytes) and concluded that although fertilization rates did not change, there was a distinctive decrease of IRs and PRs in high responders. Then, they established peak estradiol threshold levels as, 1,770 pg/ml for high responders and 2,200 pg/ml for normo-responders, above which IRs and PRs were significantly decreased in both groups [5]. When peak E2 levels exceeded 2,500 pg/ml among all cycles, IRs and PRs decreased significantly, independent of the number of retrieved oocytes [5].

Similarly, in this study, when the patients were divided into two groups according to their peak E2 levels, there were no significant differences in PRs, but implantation rate was significantly lower in >2,500 pg/ml group. These results are similar to the results of the study by Sharara et al. [13].

We also evaluated the ratio of peak estradiol/mid-luteal estradiol considering earlier studies [2, 3, 6, 14] investigating the effects of mid-luteal estradiol levels in pre-implantation period on IVF outcome. It was seen that a great majority of the patients (92.4%) were in groups 1 and 2. The clinical pregnancy rate was highest in group 1 and significantly lowest in group 3. A sharp decline in the mid-luteal E2, defined in our study as peak estradiol to mid-luteal estradiol ratio higher than 5, resulted in lower pregnancy rate, which was also confirmed by other studies [3, 6, 14].

Also the patients were evaluated of the women who got pregnant and did not get pregnant. There were significant difference between the peak E2 levels, the luteal E2 levels, E2 ratios of the pregnant patients and the patients in whom pregnancy was not achieved. All the above parameters were significantly higher in non-pregnant group.

While initially the success of IVF outcome used to be associated with E2 level, it was soon resolved that E2/oocyte ratio may be another important parameter in a successful COH cycle.

But, there is no accepted optimal E2/oocyte ratio yet [1, 4, 19]. In a recent study, Var et al. [20] concluded that, IVF success may be increased with a strategy to keeping the E2/oocyte ratio above 100 pg/ml in IVF cycles.

There may be a threshold for peak E2 level and E2 ratio above which IRs and PRs are negatively affected, and this threshold level is likely to be much higher than 2,500–3,000 pg/ml and E2 ratio higher than 5 resulted in lower pregnancy rate, which was also confirmed by other studies [1, 3, 6, 14]. Some researchers believe that IVF outcome is not reduced until peak E2 is >5,000 pg/ml, as previously reported [4, 13, 21, 22].

Friedler et al. concluded that only marked E2 decline (>98%) might influence the miscarriage rate, especially in high responder group therefore neither the significant decline of mid-luteal E2 nor the absolute serum concentration of E2 were detrimental to IVF-ET outcome and their accepted threshold for peak E2 level was 2,500 pg/ml [23, 24].

On the contrary, Joo et al. concluded that the serum E2 levels during COH influenced the IVF outcome in a concentration-dependent manner. They suggested that the optimal range of E2 levels were 3,000–4,000 pg/ml for women <38 years and 2,000–3,000 pg/ml for women ≥38 years [25].

Conclusion

This study has shown that the high E2 level and mid-luteal decline of E2 which were defined as peak E2 level/mid-luteal E2 level were predictive for implantation rate in IVF cycles. Larger randomized controlled clinical trials are still needed to discover the effect of the E2 level on ART.