Correlation Between Luteal Phase Serum Progesterone Levels and Pregnancy Outcome in Frozen Embryo Transfer Cycles: A Prospective Cohort Study

INTRODUCTION

The progesterone (P₄) secretion by the corpus luteum following ovulation plays a vital role in orchestrating the necessary changes in the endometrium for successful implantation and a positive pregnancy outcome. P₄ has a regulatory effect on key ultrastructural features of the secretory phase endometrium, such as giant mitochondria, glycogen accumulations in subnuclear regions, pinopodes and the nucleolar channel system, all of which facilitate the process of implantation.^[1] Additionally, it is believed to support implantation by stimulating the immune system to generate non-inflammatory T-helper-2 cytokines.^[2,3]

In the presence of optimum levels of P₄, CD56 cells are known to produce a substance known as P₄-induced blocking factor (PIBF), which is believed to possess anti-abortive properties.^[4] Furthermore, P₄ plays an important role in enhancing nitric oxide production, which, in turn, promotes improved blood circulation and oxygen supply to the endometrium.^[5]

Luteal phase defect (LPD) is commonly observed in most of the assisted reproductive technology (ART) cycles, which often necessitates the supplementation of additional P₄. The causes of LPD have been associated with various factors, such as the removal of granulosa cells during oocyte retrieval, extended pituitary suppression due to agonist administration in gonadotropin-releasing hormone (GnRH) agonist cycles or elevated levels of serum estradiol and P₄ that inhibit endogenous luteinising hormone release via negative feedback mechanisms at the hypothalamic-pituitary-ovarian axis.^[6–8] As a result, there is a consensus within the medical community regarding the requirement for P₄ supplementation during the luteal phase of all stimulated ART cycles.^[9,10]

The significance of P₄ in the context of implantation and early pregnancy is undeniable, and it’s reasonable to consider that LPD could potentially contribute to unsuccessful implantation and early pregnancy loss (EPL) during ART cycles.^[11,12] Research indicates that in fresh in vitro fertilisation-embryo transfer (IVF-ET) cycles, where conceptions occur, there is a more rapid increase in P₄ levels and higher luteal phase P₄ levels compared to cycles without conception.^[13]

However, achieving optimal pregnancy rates in frozen embryo transfer (FET) cycles hinges on adequate luteal phase support (LPS), primarily through exogenous P₄ administration.^[14] Literature is scarce regarding luteal phase serum P₄ levels in FET cycles, in which there is no endogenous production of P₄ due to the absence of ovulation and, in turn, the corpus luteum.^[15]

Endogenous production of P₄ during pregnancy may begin around week 6, coinciding with the uteroplacental shift.^[16] Therefore, in artificial cycles, serum P₄ levels during the mid-to-late luteal phase transition might solely rely on the dosage of exogenous P₄ administered to the patient.^[17]

A specific threshold for luteal phase serum P₄ that may affect the outcomes of hormone replacement therapy (HRT)-FET cycles has also not been established till now. Thus, we conducted a study to find a correlation between luteal phase serum P₄ levels and pregnancy outcome in FET cycles and the cut-off level of luteal phase serum P₄ which favours successful pregnancy outcome.

MATERIAL AND METHODS

RESULTS

A total of 100 patients undergoing HRT-FET were included in our study. The mean age of our patients was 33.58 years, and the mean BMI was 23.91 kg/m². Patients were divided into four groups on the basis of raw data of the luteal phase serum P₄ levels available after completion of the study: Group 1 (<15 ng/mL), Group 2 (15.1-25 ng/mL), Group 3 (25.1-35 ng/mL) and Group 4 (>35 ng/mL). The baseline characteristics of all four groups were similar [Table 1].

Out of the 56% of patients who became pregnant, 52% were clinical pregnancies, and 4% were biochemical pregnancies [Figure 1]. The overall implantation rate among our study population was 36.08%. A significant association was found between luteal phase serum P₄ levels and implantation rate, with the highest implantation rate (45.45%) seen in the 25.1–35 ng/mL group (p = 0.004). The clinical pregnancy rate (CPR) among our study population was 52%, with the highest CPR in the third group (25.1-35 ng/mL; 68.3%). A significant association was seen between CPR and luteal phase serum P₄ levels (p = 0.018). A significant association was also found between the biochemical pregnancy rate (4%) and luteal phase serum P₄ levels (p = 0.013), with all the biochemical pregnancies belonging to the second P₄ group (15.1-25 ng/mL). No significant association was seen between miscarriage rate (10%) and luteal phase serum P₄ levels (p = 0.542). A significant association was found between the ongoing pregnancy rate (42%) and luteal phase serum P₄ levels amongst our study population (p = 0.010), with the highest ongoing pregnancy rate in the third group (25.1-35 ng/mL) followed by the second group (15.1-25 ng/mL) [Figure 2]. The CPR and ongoing pregnancy rate peaked around the 25 ng/mL mark [Figure 3]. Both below and above this value, the pregnancy rates showed a decreasing trend.

Close

Figure 1:

Distribution of study subjects according to pregnancy outcome at 6–7 weeks period of gestation.

Export to PPT

Close

Figure 2:

Distribution of study subjects according to pregnancy outcome at 12 weeks period of gestation.

Export to PPT

Close

Figure 3:

Pregnancy outcome in relation to the luteal phase serum progesterone levels (n = 100).

Export to PPT

The ROC curve defined an optimal cut-off value for serum P₄ of 22.3 ng/mL, with a sensitivity of 86.5% and specificity of 33.3% for predicting clinical pregnancy [Figure 4] and a sensitivity of 88.1% and specificity of 31% for predicting ongoing pregnancy [Figure 5]. The ROC analysis indicated that the luteal phase serum P₄ levels were neither predictive of clinical pregnancy, being the area under the curve (AUC) (95% CI [confidence interval]) = 0.543 (p-value = 0.456), nor predictive of ongoing pregnancy with AUC (95% CI) = 0.497 (p-value = 0.955).

Close

Figure 4:

Receiver Operating Characteristic (ROC) curve for predicting clinical pregnancy (n = 100).

Export to PPT

Close

Figure 5:

Receiver Operating Characteristic (ROC) curve for predicting ongoing pregnancy (n = 100).

Export to PPT

DISCUSSION

Out of the 194 embryos transferred in 100 patients undergoing HRT-FET, 70 embryos were implanted, resulting in an overall implantation rate of 36.08%. A significant association was found between the luteal phase serum P₄ levels and implantation rate, with the highest implantation rate (45.45%) seen in the 25.1 to 35 ng/mL group. Yovich et al. identified that an ideal P₄ range of 70–99 nmol/L (28–39.6 ng/mL; p < 0.005) was linked to optimal outcomes in their study population. P₄ levels below 50 nmol/L (20 ng/mL) or above 99 nmol/L (39.6 ng/mL) were found to correlate with reduced implantation rates.^[17]

Among our study population, the CPR was significantly higher in the third group (25.1-35 ng/mL; 68.3%), followed by the second group (15.1-25 ng/mL; 51.9%; p = 0.018). In the study by Yovich et al., the CPR among 529 HRT-FET cycles was found to be significantly high in the mid-luteal serum P₄ group 70–99 nmol/L (28–39.6 ng/mL).^[17] Similarly, in a study conducted by Varma et al. on HRT-FET cycles with intramuscular P₄ as LPS, the CPR was found to be most significant in the P₄ group 32.2–43.7 ng/mL (p = 0.02).^[18]

The biochemical pregnancy rate was also observed to have a significant association with the luteal phase serum P₄ levels (p = 0.013), with the highest biochemical pregnancy rate found in the second P₄ group (15.1–25 ng/mL; 14.8%). Kaur et al. found a significantly higher biochemical pregnancy rate in the FET group with P₄ levels <15 ng/mL (p = 0.024). This implies that an adequate level of serum P₄ is essential to ensure successful implantation of the embryo and pregnancy maintenance. An embryonic factor may play a role, though all patients in our study were under 40 years old, suggesting a comparable rate of aneuploidy among them.^[15]

Overall, 10% of patients in our study population suffered a first-trimester miscarriage. The highest miscarriage rate was seen in the fourth P₄ group (>35 ng/mL; 18.2%), but it was not statistically significant. Lek et al. found a cut-off level of 35 nmol/L (14 ng/mL) as a predictor of miscarriage, which was not comparable with our study. Thus, we can infer that an optimum level of serum P₄ is required for pregnancy sustenance; both below and above this level, the risk of miscarriage increases.^[19]

Humaidan et al. found an inverse correlation between the mid-luteal serum P₄ concentrations and the rate of biochemical pregnancies. He also observed a significant reduction in the EPL rate, from approximately 80% to 10%, as mid-luteal-phase P₄ levels rose from around 40 nmol/L (12.57 ng/mL) to approximately 80–100 nmol/L (25.15-31.44 ng/mL).^[20]

Alsbjerg et al. investigated the reproductive outcomes in patients undergoing frozen-thawed ET before and after increasing the dosage of vaginal P₄ gel from 90 mg to 180 mg. They reported a notable decrease in the EPL rate (67% versus 44%; p = 0.014) and a significant increase in the delivery rate (9% versus 21%; p = 0.002) with higher P₄ supplementation.^[21]

A significant association was found between the ongoing pregnancy rate (42%) and luteal phase serum P₄ levels in our study (p = 0.010), with the highest ongoing pregnancy rate in the P₄ range 25.1-35 ng/mL (58.6%). Labarta et al. found an overall ongoing pregnancy rate of 47.2% in their study, which is similar to the result of our study (42%).^[22] Most previous studies on mid-luteal serum P₄ levels in FET cycles followed the patients till delivery and hence calculated the live birth rate.^[18] Due to a paucity of time owing to a relatively short study period, we could not follow the patients till childbirth.

The ROC curve defined an optimal cut-off value of 22.3 ng/mL luteal phase serum P₄ level among our study subjects for a favourable pregnancy outcome. Labarta et al. found a cut-off of 9.95 ng/mL in the mid-luteal phase, which is not at all comparable to our cut-off value. This can be explained by the fact that Labarta et al. used vaginal P₄ 400 mg twice daily as the LPS, whereas we used intramuscular P₄ 100 mg/day, which increases the serum P₄ levels much more profoundly as compared to vaginal P₄ which has a more significant first uterine pass effect.^[22]

Ellenbogen et al. concluded in their study that a mid-luteal P₄ level of >15 ng/mL increased the odds of getting pregnant by 3.2 times.^[23] Kaur et al. also observed that in Indian women, mid-luteal serum P₄ levels <15 ng/mL negatively impact pregnancy outcomes in both fresh and frozen-thawed ET cycles.^[15] Mitwally et al. also found a similar cut-off level of 15 ng/mL.^[24]

On the contrary, Aslih et al. conducted a prospective randomised controlled trial, revealing that biochemical pregnancy, clinical pregnancy and live birth rates showed no significant differences between groups, regardless of P₄ levels on day 7 of the luteal phase (P₄ <15 or >15 ng/mL). They found that even increasing P₄ dosage in the low P₄ group did not alter these outcomes, suggesting that enhancing P₄ levels late in the luteal phase may not sufficiently improve the endometrial environment.^[25]

CONCLUSION

Larger studies are needed to establish a threshold level of serum P₄ in the late luteal phase that can differentiate between successful and unsuccessful implantation. However, it is still uncertain whether the issue of unsuccessful implantation can be resolved once it is detected on days 21–23 of the HRT-FET cycle. It is also unclear whether the variation in serum P₄ levels among patients, despite receiving identical P₄ doses for LPS, is a contributing factor or merely a reflection of pregnancy outcomes. We recommend all institutes to identify their personalised ideal range of luteal phase serum P₄ levels according to the HRT-FET protocol followed in their centre, the method of testing used and the timing of the test. Individualisation of the LPS based on that range can then be done to improve pregnancy outcomes.