Do increased levels of progesterone and progesterone/estradiol ratio on the day of human chorionic gonadotropin affects pregnancy outcome in long agonist protocol in fresh in vitro fertilization/intracytoplasmic sperm injection cycles?

BACKGROUND: The effect of elevated levels of serum progesterone (P4) and estradiol (E2) on the day of human chorionic gonadotropin and their cut-off value on in vitro fertilization (IVF) outcomes is still not clear. AIMS: The aim was to evaluate the association between serum P4, E2 and progesterone/estradiol ratio (P4/E2) on pregnancy outcome in IVF/intracytoplasmic sperm injection (ICSI) cycles with long agonist protocol. SETTING AND
DESIGN: Retrospective, single center, cohort study.
MATERIALS AND METHODS: A review of complete data of 544 women undergoing fresh IVF/ICSI cycles (539 cycles) with long agonist protocol from January 2012 to February 2014 was done. Data were stratified into Three groups according to the number of oocytes retrieved: low (≤4 oocytes obtained), intermediate (5-19 oocytes obtained), and high ovarian response (≥20 oocytes obtained). STATISTICAL ANALYSIS: Fishers exact test/Chi-square was carried for comparing categorical data. Receiver operating characteristics analysis was performed to determine the cut-off value for P4 and P4/E2 detrimental for pregnancy.
RESULTS: A negative association was observed between pregnancy rate (PR) and serum P4 and P4/E2 levels with no effect on fertilization and cleavage rate. The overall cut-off value of serum P4 and P4/E2 ratio detrimental for pregnancy was found to be 1.075 and ≥0.35, respectively. Different P4 threshold according to the ovarian responders were calculated, 1.075 for intermediate and 1.275 for high responders. Serum E2 levels were not found to be significantly associated with PR.
CONCLUSION: Serum P4 levels and P4/E2 ratio are a significant predictor for pregnancy outcome without affecting cleavage and fertilization rate while serum estradiol levels do not seem to affect PR.

INTRODUCTION

Pregnancy outcomes during in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) cycles may be influenced by supra-physiological concentrations of estradiol (E2) and progesterone (P4).[1] The current use of both Gonadotropin-releasing hormone (GnRH) agonist and antagonist, highly effective to prevent luteinizing hormone (LH) surge, has limited the need for determination of serum P levels. Still, studies have shown a rise in P4 levels in 35% (5–35%) of GnRH agonists cycles and 38% (20–38%) with GnRH antagonists[123] but the effect on IVF outcome is still unclear. The mechanisms that account for this P4 rise are not clearly understood. Recently, this elevated P4 has been linked with the number of mature follicles and secretion of P4 in late follicular phase.[456] Several questions have been raised regarding this P4 elevation at the time of human chorionic gonadotropin (hCG) administration. The underlying mechanism of its effect on IVF success rate is not yet clear. Several studies-deny any association between P4 and pregnancy rate (PR)[789] while others show a negative association.[101112] In several studies, the increasing P4 levels have shown an adverse effect on oocyte maturation, fertilization, or early cleavage[13] while other studies have denied the concept of poor embryo quality[141516] and found it to be associated with impaired endometrial receptivity as a consequence of disturbed endometrial development and maturity[617] and gene expression.[18] Different detrimental cut-off values of P4 and P4/E2 has also been determined, but a consensus could not be reached.

This study aims to investigate the relationship between P4 and E2 levels on the day of hCG administration with PR, oocyte number, fertilization and cleavage rate in GnRH long agonist protocol in nondonor fresh IVF/ICSI cycles.

MATERIALS AND METHODS

This is a retrospective, single-center cohort study of patients undergoing IVF/ICSI treatment. Complete data of 544 women who underwent nondonor fresh IVF/ICSI cycles between January 2012 and February 2014 with long agonist protocol was reviewed. The primary or combined indications for fertility treatment were tubal pathology (45.1%), male subfertility (22.7%), endometriosis (10.7%), polycystic ovarian syndrome (5.9%), unexplained infertility (14.4%). To eliminate the confounding factors that might affect the outcome, following inclusion criteria was kept: Age < 40 years, follicle stimulating hormone (FSH) <10, antimullerian hormone (AMH) >1, fresh cycles (frozen cycles excluded). The study involves no violation of animal or human rights.

Stimulation protocol

Patients underwent controlled ovarian hyperstimulation (COH) with use of a GnRH agonist long protocol. The pituitary down-regulation was achieved by subcutaneous injection of 1 mg of leuprolide acetate daily from the midluteal phase of the preceding cycle. Ovarian stimulation was done with 150–300 IU recombinant FSH follitropin alpha (Gonal-F, Merck Serono) SC (subcutaneously) and dose adjusted according to the response. Recombinant Chorionic Gonadotropin alpha (250 mg; Ovitrelle) was given to trigger ovulation when at least two-three leading follicles reached a mean diameter of 18 mm. Serum P4 and E2 levels were measured on the day of hCG administration by the chemiluminescent immunoassay using Access 2 Immunoassay system (Beckman Coulter) in the same laboratory. Oocyte pickup was done 36 h after hCG administration. Oocytes were cultured in G-IVF plus media (VITROLIFE) containing 10% of human serum albumin with gentamicin as an antibacterial agent and inseminated with motile sperm prepared by the two-layer percoll gradient method. Fertilization was defined as oocytes with two pronuclei 16–20 h after insemination. Embryos were transferred to G-IVF plus media and were classified by blastomere equalization and cytoplasmic fragment. Day 3 or 5 embryo transfer was done depending upon the number of embryos, and excess good-quality embryos were cryopreserved for subsequent frozen embryo transfer cycles with different grades of embryos (1–3). An ongoing pregnancy was defined as the pregnancy test done after 14 days of embryo transfer with a positive heartbeat by ultrasound at 6 weeks of gestation.

Grouping of high, intermediate and poor ovarian responders

Patients were categorized into 3 groups according to the number of oocytes retrieved: Low (≤4 oocytes), intermediate (5–19 oocytes) or high ovarian response (≥20 oocytes).

Six groups according to P4 levels

Group A <1.0, group B: 1.0–1.25, group C 1.26–1.50, group D 1.51–1.75, group E 1.76–2.0 and group F > 2.0.

Five groups according to E2 levels

Group A (<1000 pg/ml), group B (1000–2000 pg/ml), group C (2000–3000 pg/ml), group D (3000–4000 pg/ml), and group E (>4000 pg/ml).

Statistical analysis

All the statistical analyses were carried out using Statistical package for the social sciences (SPSS) IBM version 19.0 (Chicago, Illinois). Data were expressed as mean, standard deviation (SD) or frequencies and percentages. For comparing categorical data, Chi-square/Fishers exact test was carried out as appropriate. Receiver operating characteristics (ROC) analysis was performed to determine the cut-off value for P4 and P4/E2 at an approximately equivalent sensitivity and specificity, which may discriminate between pregnancy and nonpregnancy. A P < 0.05 was considered to be statistically significant.

RESULTS

A total of 544 ovarian stimulation cycles were included in the study. Of these, 5 cycles need to be cancelled as no oocyte could be retrieved. Of the total 539 cycles with long agonist protocol, 319 (59.1%) were conventional IVF cycles and 220 (40.8%) were ICSI cycles. Of all, 146 (27.08%) clinical pregnancies were noted. Baseline characteristics such as age, body mass index, serum AMH, FSH, LH (follicular phase), total gonadotropin dose, endometrial thickness on day of hCG, E2 and P4 on the day of hCG, number of oocytes retrieved, fertilization rate, cleavage rate and PR for all the subjects among the different ovarian responders are shown in Table 1.

Table 1

Baseline characteristics of subjects in three groups according to ovarian response

Overall, the mean age of the patients in our study was 31.6 ± 3.6 years (21–40). Average (SD) values of P4 among different ovarian responders (≤4 oocytes, 5–19 oocytes and ≥20 oocytes) were 1.06 (0.7), 1.3 (0.8) and 1.4 (0.8), respectively, with an increasing value among high responders. Serum E2 levels and numbers of mature oocytes retrieved also showed an increasing trend in high responders with a statistically significant difference between the three groups (P < 0.01).

We analyzed the correlation between the serum P4 levels on fertilization rate, cleavage rate and PR among the different groups [Table 2]. It was found that PR was affected significantly with the change in serum P4 levels. The total number of oocytes retrieved also varied significantly between the groups showing increasing trend with P4 values.

Table 2

Comparison of fertilization rate, cleavage rate, pregnancy rate and number of oocytes retrieved in different groups according to serum P4 levels

The trend of PR according to P4 levels is as shown in Figure 1. PR showed a significantly decreasing trend (Chi-square trend in proportion = 8.25; P = 0.004) with increasing level of P4 with maximum PR in the range of 1.01–1.25, but after a serum P level of 2 ng/ml, relatively stable effect was seen.

Figure 1

Trend of pregnancy rate with 95% confidence limits according to P4 levels

To assess an optimum P4 level for an equivalent sensitivity and specificity for pregnancy status, ROC analysis was carried out. The area under curve (AUC) 0.58 (95% confidence interval [CI]: 0.53–0.63) was significant (P = 0.006). Overall, P4 level was found to be 1.075 for an equivalent sensitivity and specificity value of 55% [Figure 2] with positive and negative predictive value of 31.3% and 77%, respectively. For intermediate and high responders, the cut-off value was found to be 1.075 for an equivalent sensitivity and specificity value of 56% and 1.275 for the sensitivity and specificity value of 62%, respectively, while it could not be predicted for low responders as AUC was not statistically significant.

Figure 2

Receiver operating characteristic curve for defining optimal detrimental cut-off value for P4 on human chorionic gonadotropin day

Table 3 shows the correlation of E2 on the day of hCG with fertilization, cleavage and PR. Bivariate logistic regression analysis revealed that the variable E2 on day of hCG alone is not a significant predictor of pregnancy status, fertilization rate and cleavage rate although the number of oocytes retrieved increased significantly with an increase in E2 levels.

Table 3

Comparison of fertilization rate, cleavage rate, pregnancy rate and number of oocytes retrieved in different groups according to serum E2 levels

Impact of P4/E2 ratio on the day of human chorionic gonadotropin

To assess threshold level of P4 to E2 ratio (P4 [ng/ml] ×1000/E2[pg/ml]) for PR, ROC analysis was carried out. AUC 0.58 (95% CI: 0.53–0.64) was found to be statistically significant (P = 0.003). For an equivalent sensitivity and specificity value (56%) the corresponding value of P4/E2 ratio was found to be 0.35 [Figure 3]. PR (31.4%) among the patients having ≤ 0.35 P4/E2 ratio was significantly (P = 0.047) higher compared with 23.4% observed among the patients having value >0.35. However, fertilization and cleavage rates were not significantly (P > 0.05) different between the two categories. Similar ROC analysis was carried out according to the ovarian responders. As AUC was not significant, therefore P4/E2 cut-off value could not be predicted.

Figure 3

Receiver operating characteristic curve for defining optimal detrimental cut-off value for P4/E2 ratio on human chorionic gonadotropin day

DISCUSSION

This study has analyzed the correlation between serum P4, E2 and P4/E2 ratio on hCG day with PR. As a secondary outcome, the effect on cleavage and fertilization rate was also studied. The results have shown that serum P4 levels and P4/E2 ratio is a significant predictor for PR while the E2 levels had no significant association. PR showed a significantly decreasing trend with an increasing level of P4 and P4/E2 levels. Different cut-off levels for P4 and P4/E2 were determined among different ovarian responders. Fertilization and cleavage rate were not affected by either P4 or E2 levels.

The potential effect of type of responders on the association between P4 elevation and probability of pregnancy was also explored. Earlier there have been three studies in which data were analyzed according to the type of ovarian response.[1920] Recent study by Xu et al. has concluded that there is a significant decrease in ongoing PR and implantation rate with P elevation in all ovarian responses to COH.[6] Our study has also shown results in agreement with that study and different cut-off values for P4 among intermediate and high responders were predicted.

A positive correlation between number of mature follicles and secretion of P4 in late follicular phase have been reported in earlier studies.[5] In the present study also, an increasing pattern of serum P level was observed with a significant (P < 0.05) difference between the three types of responders thus supporting the concept that increasing P4 levels is a reflection of the number of follicles and not due to premature luteinization.

Till now, there have been numerous studies which have evaluated the association of P4 elevation with PR with conflicting results. Our results were in agreement with a recent meta-analysis published in 2013 which have reported a detrimental effect of P4 elevation on PR in range of 0.8–1.1 ng/ml (odds ratio: 0.79).[21] The present study has used ROC analysis which is a preferred method to identify optimal thresholds to define these detrimental cut-offs and the value was found to be 1.075 for an equivalent (53%) level of sensitivity and specificity.

Serum E2 concentration on the day of hCG alone was not found to be a significant predictor of pregnancy status which is in consensus with previous studies by Kyrou et al. (2012) and Yu Ng et al.[2223]

Not only the P4 levels alone but also P4/E2 has nowadays been considered to be a reasonable option for prediction of pregnancy outcome. A prospective cohort study by Elgindy published in 2012 have found a cut-off levels of P4 >1.5 ng/ml and P4/E2 >0.55 as detrimental to pregnancy.[24] In our study, P4/E2 ratio ≤0.35 was found to be associated with a significantly higher PR compared to those having value >0.35 although no significant association with fertilization and cleavage rates was found. Table 4 are summarizes the most recent studies with different threshold cut-offs of serum P4 and P4/E2 levels.[126252627282930]

Table 4

Recent studies showing the cut-off levels of P4 or P4/E2 levels in women undergoing fresh IVF cycles

Either cryopreservation of pronuclear or cleavage stage embryos or blastocyst transfer has been suggested as a strategy to overcome this problem. A recent study by Corti et al. in 2013 has concluded that fresh blastocyst transfer does not completely overcome the detrimental effects of progesterone rise at hCG on pregnancy outcome.[31]

CONCLUSION

The present study has shown a negative association of increasing P4 levels and P4/E2 with PR. It still remains uncertain whether freezing the embryos and transferring them later in frozen-thawed cycles will be a solution to improve the pregnancy outcome when P4 levels are high. More randomized studies are required to evaluate the effectiveness of frozen cycles in comparison to fresh cycles to overcome the detrimental effect of elevated P4 or P4/E2 ratio.