Association between urine and serum estradiol levels in in vitro fertilization cycles

To study the correlation between urine and serum estradiol (E2) controlled ovarian hyperstimulation (COH). This is a cross-sectional analytical study that was conducted in a tertiary care hospital. Seventy-seven urine and blood samplings were collected from infertile women who were treated with COH. An electrochemiluminescent immunoassay was performed to evaluate E2 levels between urine and serum samples on the 6th day and the day of ovarian trigger. In addition, the correlations were evaluated between urine E2 level and number of follicles, retrieved, metaphase II oocytes, and fertilization rate. A sub-analysis was performed for age, responding status and BMI. Seventy-seven infertile women were recruited. The medians of serum and urine E2 level levels on the day 6th of ovarian stimulation were 833.20 pg/ml (IQR; 516.90–1371.00) and 3.67 (IQR; 2.84–4.81), respectively. On the day of ovarian trigger, the median of serum E2 level was 2113.00 pg/ml (IQR; 1382.00–3885.00) and urine E2 level (E2/creatinine) was 6.84 (IQR; 5.34–8.70). The correlation between serum and urine E2 level on day 6th was 0.53 and the day of ovarian trigger was 0.59, p < 0.001. Moreover, the correlations of urine E2 level on the day of ovarian trigger to number of follicles, number of oocytes retrieved, metaphase II oocytes and fertilization rate were 0.57, 0.58, 0.61, and 0.64 (p < 0.001). The urine E2 level was moderately correlated to serum E2, number of follicles growth, oocytes retrieved and fertilization rate.


Material and methods
Study participant and study design. A cross-sectional study was conducted. The serum and urine samples were obtained in the infertility clinic between April 2020 and November 2020 from infertile women treated with controlled ovarian hyperstimulation for IVF or ICSI at Srinagarind Hospital, Khon Kaen University. The inclusion criteria were women aged > 20 years-old without contraindication for hormonal treatment. After receiving written informed consents of the participants, urine specimens were collected at the same time of routine blood collections in the morning on the 6th day and the last day of ovarian stimulation. Other information such as the cause of infertility, body mass index (BMI), type of hormone, and follicular sizes were recorded, and the follicular sizes were measured by 2 dimensional (2D) transvaginal ultrasonography. This research was approved by the Khon Kaen University Ethics Committee on Human Research (HE621557).
Controlled ovarian hyperstimulation protocol. Transvaginal ultrasound was performed for all participants on the second or third day of menstruation in order to identify the confounding factors such as ovarian cyst that may be affect on estradiol level. In the woman with normal transvaginal ultrasound finding, daily gonadotropin injection treatment, the HMG (Menopur, Ferring AG, Germany) or rFSH (Gonal-F, Merk, France) was started on the 2nd or 3rd day of menstruation. The dosage of gonadotropin stimulation using depended on antral follicle count measurement under ESHRE guideline 13 . According to type of gonadotropin using, the protocol depended on physician decision. Long-acting rFSH treatment at a subcutaneous dose (100-150 mcg) of corifollitropin alpha (Elonva, MSD, USA) was injected on the second or third day of menstruation. The treatment continued from day six of stimulation with rFSH or HMG and the dosage of gonadotropin was adjusted following the size of follicle growth under EHSRE guideline 13 . The follicle size was more than 10 mm. the gonadotropin was continue in the same dosage. In contrast, if dominant follicles size was less than 10 mm, we would add 75 unit of gonadotropin. The GnRH antagonist (Orgalutran, MSD, USA) was used daily until triggering occurred when the leading follicles reached a mean diameter of 14 mm. The ovulation was triggered by human chorionic gonadotropin (HCG) or GnRH agonist alone or dual trigger (combined GnRH agonist with HCG) in case of > 2 dominant follicles of > 17 mm.
Estradiol analysis. Both urine and serum E2 levels were analyzed by an Electrochemiluminescent immunoassay (ECLIA). The ECLIA at the KKU center is Elecsys Estradiol III Cobase 801 that employs a competitive test principle using two monoclonal antibodies specifically directed against 17-β E2 and standardized with ID-GC/MS Calibration monthly.The coefficient of variation was 3.6%. The duration of the automated assay was 18 min, calibrated with an internal quality control daily and external quality control monthly. For control the confounding factor, urine E2 level (pg/ml) was adjusted its level with urine creatinine (mg/dl) as the urine E2-creatinine ratio (E2/creatinine) to correct the glomerular infiltration rate.
Serum and urine collection: Ten mL of blood samples were collected in plain tubes that included coagulation activator followed by followed by centrifugation at 1720g for 10 min; serum was collected and stored at 4 °C until the assay was performed within 48 h after sample collection. Fifty mL of urine specimens were collected in the morning on the same day as blood collection. Urine specimens were subjected to centrifugation at 430g for 10 min; supernatants were collected and stored at 4 °C until the assay was performed, which was within 48 h after the sample collection. Urine and serum E2 levels were measured by Electrochemiluminescent immunoassay (ECLIA) 15 . Our study was the first study that using ECLIA to evaluate urine E2. Therefore, we developed pilot study for testing accuracy of this method. We applied 10 serum and urine E2 sampling from other women who received COH regimen on day 6th of ovarian stimulation and day of ovarian trigger. The correlation (Rho) between urine and serum E2 level on day 6th of stimulation and day of ovarian trigger was 0.67 and 0.81, respectively. Data analysis. Based on the result of Maskarinec et al. 10 , with a 10% dropout rate, 77 participants were needed. The general appearances of participants were analyzed by SPSS V.23 in terms of mean and standard deviation (SD). The correlations between urine E2 to serum E2, ultrasonographic follicles, retrieved oocytes, metaphase II oocytes and fertilization rate were analyzed by Spearman's correlation. A p value < 0.05 indicated statistical significance.

Sample size.
There are not any studies comparing between urine and serum estradiol in women treated with assisted technology. The study of Maskarinec et al. 10 was cited to calculate the sample size. Maskarinec G's study was the largest study, included Asian and young population, studied about correlation between urine and serum estradiol of premenopausal women that was 0.35, p < 0.001. We provided type I error as 0.05 and 80% of power. A total of 77 patients were recruited in this study included 10% dropouts. This study used this formular of sample size calculation.

Ethical consideration. The Human Research Ethics Committee of Khon Kaen University reviewed and
approved the study per the Helsinki Declaration and the Good Clinical Practice Guidelines (HE621557).

Consent for publication.
All of the authors consent to publishing and hereby grant the Publisher exclusive license of the full copyright.
The sub-analysis presented that on the day 6th of ovarian stimulation, women who had an age ≥ 35 years the urine E2 level 0.72 (p < 0.001) was significantly correlated when compared to women with age < 35 years 0.35 (p = 0.18). While on the day of ovarian trigger, the urine E2 levels and serum E2 levels were significantly correlated in both age groups: 0.52 in age < 35 years and 0.60 in age ≥ 35 years. Moreover, the urine E2 level was moderately correlated in a group of women with normal BMI (0.50, p < 0.001) and high BMI (0.54, p < 0.001) while low the BMI group was lowly correlated (0.13, p = 0.39) on day 6. Furthermore, the urine E2 levels in the normal responder group were significantly correlated with serum E2 levels on day 6 of ovarian stimulation (0.82, p < 0.001) compared to the poor and high responder groups. In contrast, on the day of ovarian trigger, the poor responding group had a statistically significant difference in the correlation between the urine and serum E2 levels (0.42, p < 0.05) when compared to the other groups (Table 4).   The correlations between serum E2 at 0.72 and sonographic follicles on the 6th day and the last day of ovarian stimulation were 0.78 (p < 0.001), which showed a higher correlation than urine E2 and sonographic follicles.
There were not any COH complications such as OHSS or oocyte retrieval such as bleeding or infection. The specimens were collected and analyzed without any mistakes or complications.

Discussion
In this study, it was found that the correlations of the relative levels of serum and urine E2 were valid on both on the 6th day of ovarian stimulation and the day of ovarian trigger in infertile women who had undergone the COH cycle. There was a moderate correlation between serum and urine E2 concentration on the 6th day of ovarian stimulation (r = 0.53) and on the day of ovarian trigger (r = 0.59). The positive linear correlation in urine E2 levels on the day of ovarian trigger and number of follicle sizes (size ≥ 10 mm r = 0.57), number of retrieved oocytes, r = 0.58, number of metaphase II oocytes, r = 0.61 and rate of fertilization, r = 0.64, were relatively significant. Moreover, subgroup analysis was done for evaluation between serum and urine E2 level in factors of IVF success rate (age, BMI, and responding status) and the results showed that there was a significant correlation in women age ≥ 35 year and normal to high BMI, while this correlation was uncertain in the responding factors.
This study results are consistent with previous studies. A 1981 reported by Miyakawa et al. 14 reported that the urinary estradiol-17 beta-glucuronide (E2-17G) measured by direct radioimmunoassay (RIA) correlation was significantly correlated with serum estrogens in women treated with gonadotropin on the day of ovarian trigger. Frenken et al. 15 reported that total the excretion per 24-h or the concentrations per liter in the 24-h urine collection of estrogen and estrone-3-glucorunide on the day of prior ovarian trigger were in good correlation with serum E2 levels in women treated with HMG. In addition, a 1992 study by Rapi et al. used a chemiluminescence immunoassay (LIA) method in early morning urine (EMU) samples to evaluate the correlation between serum E2 level and urine E1-3G level on the day of ovarian trigger. They revealed that the urine E1-3G excretions were consistent with serum E2 levels in women treated with COH 16 . In the clinical setting, RIA is a complex method that is performed using a 24-h collection. For simplicity, the ECLIA method was used to evaluate spot urine which is convenient for analyses urine samples. Furthermore, the ECLIA showed good sensitivity in normal and high E2 level women and is used worldwide 17,18 .
The present study is the first study that evaluates the correlation between urine E2 levels with the number of follicle growth, number of retrieved oocytes, number of metaphase II oocytes and rate of fertilization with which these data were associated with the IVF outcomes. From these results, we assumed that urine E2 level on the 6th day of stimulation might be a predictor of the mean number of oocytes growth. In addition, urine E2 levels on the day of ovarian trigger might also predict, the mean numbers of retrieval follicles, numbers of MII oocytes and fertilization rate.
We performed sub-analysis to investigate the factors that might affect the outcomes especially, age, BMI and responding status. We found that some factors were slightly impact on the outcomes such as age and BMI. In current literature reviews, there are not information to explain the effect of the age or BMI on the correlation between serum and urine E2 level. Therefore, we need larger sample size and randomized controlled studies to confirm these effects.
To the best of the authors knowledge, this is the first study that used ECLIA which a good sensitivity method to evaluate the urine E2 level. Using urine samples to monitor E2 level is more convenient and simplifies collection and analysis when compared to serum. In addition, the fact that the serum and urine E levels were collected on the same day reduced possible confounding factors from the fluctuations of hormone levels, moreover, urine E2 levels from collected urine E2 to creatinine ratio reduced the effect of the glomerular infiltration on urine E2 excretion. The limitation of this study, other hormones such follicle stimulating hormone (FSH), luteinizing hormone (LH) and progesterone that might affect to IVF outcomes were not evaluated. Furthermore, some hormones especially LH were unstable and had short half-life in the plasma. www.nature.com/scientificreports/ Based on the results of this study, the urine E2 level was correlated with serum E2 level in COH stimulation on the 6th day of ovarian stimulation and the day of ovarian trigger. It may be an advantage to implement this collection method during COH especially in the cases where it is inconvenient to collect serum samples. Further study is needed to confirm the efficacy of this method in women with the risk factor of OHSS.
In conclusion, the urine E2 level was moderately correlated to serum E2, follicle growth, retrieved oocytes and also fertilization rate. We assumed that the urine E2 might be the alternative test for ovarian stimulation monitoring in women who cannot collect blood specimen such as obesity women and anxiety women.