INTRODUCTION

Frozen embryo transfer (FET) as a standard of care technique is being utilised increasingly worldwide in assisted reproductive technology practices.^[1] Safety and efficacy of vitrification protocols, combined with a reduced risk of ovarian hyperstimulation syndrome, expanding use of preimplantation genetic testing, single embryo transfer, elective freeze-all and avoiding embryo-endometrial asynchrony in fresh transfers, have all contributed to this trend.^[2–4]

There is no consensus on the best endometrial preparation protocol for FET cycles for achieving pregnancy.^[5] The options are AC, NC - true (tNC) and modified (mNC) and stimulated cycles. The AC protocol has been the predominant protocol due to its simplicity in monitoring, flexibility and good pregnancy rates. Consistent evidence from observational studies and a recent truncated randomised controlled trial (RCT) emphasises the higher obstetrical and perinatal risks, specifically hypertensive disorders of pregnancy (HDP), preeclampsia and placental bed disorders, in pregnancies conceived by AC-FET as compared to NC-FET.^[6–8] Some studies have also shown a greater incidence of early pregnancy loss and first-trimester bleeding in AC-FET compared to NC-FET pregnancies.^[9,10] A recent RCT showed comparable live birth rates (LBRs) in ovulatory patients with tNC, mNC and AC-FET cycles.^[11] Certain limitations in this RCT have been highlighted, which may restrict the application of the results to general practice.^[12]

The suboptimal endometrial preparation and absence of ovulation/circulating corpus luteal vasoactive peptides, mainly relaxin, in the AC-FET protocol have been implicated in the aetiology of adverse obstetrical and perinatal outcomes, specifically HDP in AC-FET pregnancies.^[13,14]

For these reasons, in ovulatory patients, the expanded use of ovulatory FET protocols is recommended.^[15,16]

Some consistently reported issues with the NC-FET protocol are pinpointing the time of ovulation and, as an extrapolation, the start of progesterone production so as to time the embryo transfer. The aim is to synchronise the embryo age with the duration of exposure of the endometrium to progesterone. The detection of ovulation is through surrogate markers in the tNC-serum/urinary luteinising hormone (LH) surge, maybe combined with serum oestradiol, progesterone levels and ultrasound of the growing dominant follicle.^[17,18] In the mNC protocol, human chorionic gonadotropin (hCG) trigger is given to reduce the endocrinological monitoring for detecting ovulation and to increase the flexibility for scheduling the embryo transfers. A hCG trigger mimics the natural LH surge, bypassing endogenous hormonal signalling, allowing greater control over the timing and predictability of ovulation.

To optimise the NC protocols: tNC and mNC—there is no consensus in the literature on the use of LPS. We present a review to date regarding various aspects of progesterone as LPS in tNC-FET and mNC-FET protocols in relation to the endocrinology and clinical evidence. Is progesterone supplementation required as LPS? Can it improve pregnancy outcomes? What is the significance of serum progesterone levels in the luteal phase? Can we individualise LPS with progesterone? Is there an optimum time to start progesterone in the luteal phase?

Progesterone from the CL is essential for the development of a receptive endometrium, optimal embryo implantation and continuation of early pregnancy till the luteal placental shift takes over.^[19] It also leads to uterine quiescence and exerts a regulatory role on the maternal immune response, both systemic and intrinsic to the uterus.^[20,21]

DOES THE CL SECRETE ADEQUATE PROGESTERONE IN ALL PATIENTS WITH REGULAR MENSTRUAL CYCLES?

Luteal phase deficiency (LPD) is a condition in which progesterone secretion from the CL is not sufficient or of enough duration (less than 9–11 days) to allow embryo implantation and growth. In a study, it was observed that 13% of ovulatory menstrual cycles are <10 days in duration.^[22] No minimum serum progesterone concentration defines normal luteal function.^[23] Mid-luteal serum progesterone <10 ng/ml has been established as a more liberal definition of LPD.^[24,25] In ovulatory cycles, luteal progesterone <5 ng/ml in 8.4% and <10 ng/ml in 24.4% of women was observed.^[22] LPD prevalence of 31.4% among women with regular ovulatory cycles has been reported.^[26] A higher incidence of LPD has been reported in normoovulatory women and infertility.^[27]

The luteal phase can be affected by many medical conditions, and if undetected, could give the false impression that a woman has normal ovulatory cycles.^[28] Conditions that have been associated with LPD include extremes of reproductive age, obesity, stress, excessive exercise, thyroid/prolactin disorders, and unexplained infertility. These patients may have normal menstrual cycles but with altered corpus luteal function.^[23]

Further, corpus luteal function varies from cycle to cycle in normal ovulatory women. This has recently been corroborated in a prospective study; 26 healthy women with regular menstrual cycles were studied for three consecutive cycles regarding their hormonal profiles. Only 10 out of 26 women had progesterone levels greater than 30 nmol/l (9.4 ng/ml) on cycle days 20 or 25 in all three cycles. In total, only 45/75 of the cycles had serum progesterone ≥30 nmol/l (9.4 ng/ml).^[29]

These observations emphasise that in a tNC-FET protocol where CL is the only source of progesterone, it may not always be adequate.

CAN hCG TRIGGER IN A mNC FET CYCLE ENSURE ADEQUATE PROGESTERONE SECRETION BY THE CL?

Both hCG and LH share a common receptor, LH/CG, but elicit a unique cascade of events following receptor binding. hCG has a longer half-life and a more potent steroidogenic effect. In a mNC-FET cycle, hCG trigger may have an inherent additional LPS.^[20] Certain studies have reported a negative feedback of hCG on hypothalamic GnRh secretion, decreasing LH levels by a short-loop negative feedback, while others have not found such an association.^[30,31] hCG trigger also may alter the pattern of progesterone rise postovulation, being unphysiologically high, a faster and an earlier peak compared to the natural progesterone secretion by the CL. This all can collectively alter embryo-endometrial synchrony.^[32]

In a study after the hCG trigger in the mNC-FET protocol, serum progesterone was tested on day 3 post-hCG; a level of >3 ng/ml was considered day 1 post-ovulation. In one-third of patients, serum progesterone >3 ng/ml was seen on one or more days later than on day 3, as otherwise expected. (Cédrin-Durnerin I unpublished data).

On the whole, the hCG trigger in a mNC-FET cycle does have some inherent luteotropic effect, but the intensity and pattern of progesterone secretion vary from the normal progesterone secretion by the CL. This may have a negative effect on embryo implantation.

DOES CLINICAL EVIDENCE SUPPORT THE USE OF PROGESTERONE AS LPS IN tNC-FET AND mNC-FET CYCLES?

A meta-analysis by Seol A et al. (2020) including one RCT (n = 435) and three retrospective studies (n = 3033) found no significant difference between clinical pregnancy rate (CPR) in NC-FET cycles with progesterone support compared to no LPS (odds ratio [OR], 0.96; 95% confidence interval [CI], 0.60–1.55). A sub-analysis of two retrospective cohort studies that reported the live birth rate (LBR) showed a significantly higher LBR in the vaginal progesterone group (OR, 1.72; 95% CI, 1.21–2.46).^[33]

A systematic review and meta-analysis by Mizrachi et al. (2021) on whether women following NC-FET should receive luteal support included a total of eight studies. Two studies (n = 858) used hCG, and six studies (n = 1507) used progesterone for LPS. Four studies were RCTs, and four were historical cohort studies. hCG administration for LPS did not increase the CPR (OR, 0.85; 95% CI, 0.64–1.14). Progesterone LPS was associated with a higher CPR (OR, 1.48; 95% CI, 1.14–1.94) and a higher LBR (three studies OR, 1.67; 95% CI, 1.19–2.36). The association between progesterone LPS and the LBR remained significant after excluding non-randomised studies. Subgroup analysis showed that progesterone supplementation was associated with higher LBR and CPR in tNC-FET cycles. In addition, a subgroup analysis of studies in which hCG was used as a trigger also showed higher odds of clinical pregnancy in patients receiving progesterone for LPS (four studies, OR, 1.56; 95% CI, 1.10–2.20, I² = 0%).^[34]

Another meta-analysis for LPS for NC-FET included nine studies; the LPS group had increased CPR (RR, 1.23; 95% CI, 1.12–1.34; I² = 52%) compared to the group with no LPS. Subgroup analysis according to trigger administration also showed a significant difference between the two groups.^[35]

Further, the results of a recent systematic review and meta-analysis based on RCTs only (four studies, 1116 participants) by Jiang Y et al. (2023) showed that progesterone supplementation was associated with increased LBR (RR, 1.42; 95% CI, 1.15–1.75; I² = 0%, moderate quality evidence) and CPR (RR, 1.30; 95% CI, 1.07–1.57; I² = 0%, moderate quality evidence) in NC-FET cycles. A subgroup analysis for tNC-FET cycles showed increased LBR and CPR with the use of progesterone as LPS. Very low-quality evidence indicated no such association in mNC-FET cycles (only one study).^[36]

Tables 1 and 2 elaborate on the RCTs and non-randomised trials regarding progesterone as LPS in NC-FET cycles, respectively.^[37–45]

Overall, there is modest evidence of progesterone as LPS in NC-FET cycles for improving LBRs and CPRs as compared to cycles without LPS. This association remains for the tNC-FET cycles. For mNC cycles, larger RCTs are required for studying the role of progesterone in these cycles.

CAN WE INDIVIDUALIZE LPS BASED ON SERUM PROGESTERONE LEVELS IN tNC-FET AND mNC-FET CYCLES?

Progesterone production by the CL is pulsatile, secreted in response to LH pulses. This pulsatility is more pronounced in the mid-luteal to late luteal phase. The progesterone levels may fluctuate up to eightfold within 90 minutes, oscillating between 5 and 40 ng/ml, making a single value difficult to interpret.^[46]

Lab variations also matter in determining serum progesterone levels. Different progesterone assays have limited reproducibility.^[47] A prospective study highlights the existence of a diurnal variation in progesterone levels on the day of ET in a mNC-FET cycle.^[48]

The significance of a single value of mid-luteal serum progesterone in relation to adequate CL function has not been defined.^[24] Nor can it relate to the adequacy in duration of corpus luteal function.^[23]

Identifying patients for the need of LPS in tNC and mNC FET protocols by a single random serum progesterone level in the luteal phase is challenging. The pulsatility of serum progesterone levels in the luteal phase, its diurnal variation, association with certain patient characteristics like BMI and significance of the progesterone levels in relation to sustained adequate CL function all underscore the challenges for individualising LPS.

CLINICAL STUDIES DETERMINING THE SERUM PROGESTERONE LEVELS DURING THE LUTEAL PHASE IN tNC AND mNC FET CYCLES

WHEN TO START LUTEAL PROGESTERONE SUPPORT IN tNC-FET AND mNC-FET CYCLES?

In a FET cycle, the aim is to transfer an embryo to a receptive endometrium. Embryo endometrial synchrony is presumed to be optimal when the total duration of the exposure of the endometrium to progesterone is similar to the age of the embryo on the day of transfer.^[55,56]

For this, in a tNC and mNC cycle, we need to determine the exact time of ovulation, as progesterone secretion begins with ovulation. In a tNC serum LH surge is taken as a surrogate marker for identifying the time of ovulation and consequently the time of embryo transfer.

There is no consensus on the definition of LH surge. The onset of the LH surge has been defined as absolute levels greater than 15, 17, or 20 mIU/ml, an increase of 180% of the previous levels or an increase of 1.8–6 fold from baseline LH.^[57] All this can be combined with ultrasound documentation of follicular collapse and combinations of oestradiol and progesterone levels to make it more accurate.^[17,18]

LH surges leading to ovulation are highly variable in timing, amplitude and duration.^[58] Further, the duration between the LH surge and ovulation varies between 22 and 56 hours, the mean duration being 33.9 hours.^[57] A retrospective observational study determined the relationship between LH rise and progesterone rise (n = 102) in women undergoing NC-FET cycles by measuring LH, oestradiol and progesterone on three consecutive days till the day of ovulation (progesterone exceeding 1 ng/ml). 20.6% of patients had the LH rise 2 days prior and 9.8% on the same day as the progesterone rise; that is, nearly 30% of patients were outside the expected standard time of ovulation, which is 24 hours after the LH surge.^[59]

Urinary LH testing is also associated with certain limitations: false negative results due to smaller peaks, false positives due to premature peaks and variable time delays due to prolonged urinary clearance of LH, mean time interval to ovulation of 20 ± ٣ hours.^[57,60]

Once ovulation occurs, we are not sure of the adequacy of the CL and consequently the production and duration of adequate progesterone levels.

In a mNC-FET protocol, the hCG trigger has the advantage of greater consistency and control over the ovulation timing. The mean time interval between the trigger and ovulation is 38 hours (34–46 hours).^[32]

CLINICAL STUDIES REGARDING THE TIME OF STARTING LPS WITH PROGESTERONE IN tNC-FET AND mNC-FET CYCLES

Progesterone supplementation as LPS needs to be similar to the pattern of progesterone secreted by CL. There is no consistency in the timing of starting progesterone as LPS in various clinical studies evaluating the role of progesterone supplementation in NC-FET cycles. In the absence of clinical studies comparing the pregnancy outcomes in NC-FET cycles with starting LPS at different time points, it has been suggested to start progesterone from the presumed day of ovulation to avoid the physiological and lab variations seen in LH surge-ovulation/hCG-ovulation intervals.

In a tNC-FET cycle, it has been proposed in one study to start LPS 24 hours after the serum LH surge is detected and after 48 hours of the trigger.^[10] In another review, the time points suggested are 36 hours after either a serum LH surge or an hCG trigger. For urinary LH detection, tests being done twice a day, 24 hours after urinary LH detection.^[55]

In a recently studied new protocol, more commonly called the natural proliferative phase-FET cycle, in a NC with evidence that ovulation has not occurred, progesterone supplementation for the number of days as is the age of the embryo ensured a good pregnancy outcome, emphasising the basis of starting LPS from luteal day 1.^[61–63]

It is proposed to initiate LPS 24–36 hours after the onset of the serum LH surge or 36–48 hours after the hCG trigger or on the day of embryo transfer. Starting LPS too early, that is, before presumed ovulation, which would be on the day of the serum LH surge or up to 24 hours of the hCG trigger, could have a negative impact on pregnancy outcomes due to premature closure of the window of implantation.^{[10,38,50,55,64]}

CONCLUSION

There is considerable heterogeneity in the studies evaluating the role of progesterone as LPS in tNC-FET and mNC-FET protocols regarding the type, dose, and duration of progesterone. The endocrinology and clinical evidence point to improved pregnancy outcomes, both CPR and LBR, with progesterone as LPS in tNC-FET cycles. For the evidence of the role of progesterone as LPS in mNC-FET cycles, as there is inconsistency in its association as reported in different meta-analyses, larger studies are required. Serum progesterone levels around the ET day may not predict successful implantation because of the variability among women and also it is not predictive of sustained corpus luteal function. There is a significant subgroup of patients who may have LPD; till we have parameters to identify this subgroup, progesterone support may be given in tNC-FET cycles. Starting progesterone on luteal day 1 may have an added advantage of overcoming the physiological and lab variations for determining the ovulation time. Comparative studies related to different progesterone types, routes, doses, and durations would make LPS more evidence-based.