Shedding light on the ‘black box’: the luteal phase from end to end

Published 23 May 2022

The first Campus meeting to be held with ‘live’ participants for two years covered the luteal phase in ART, from progesterone measurement to luteal support to implantation and how to improve it, and finally to pregnancy outcome. As ever, there were many questions raised, but rarely answers supported by strong evidence.

A marathon Campus meeting held in Lubeck in early May and organised by the SIG Reproductive Endocrinology was the first to be staged as a live (plus virtual) event after lockdowns and attracted a large attendance – 115 registrants on-site and 179 online. The agenda covered every stage of the luteal phase, from progesterone levels in the late follicular phase to perinatal and maternal outcomes of pregnancy. Many of the interventions discussed (and debated) were designed to improve endometrial receptivity, particularly in women with repeated implantation failure; many of these procedures are now designated as add-ons by any authorities. And as has been repeatedly stressed in these pages, many were lacking in strong evidence of efficacy or safety.

In his introduction to the meeting local organiser Georg Griesinger said that there had been huge advances in research on the follicular phase and on the embryo but comparatively little on the luteal phase. ‘So we are still transferring our embryos into a black box,’ he said, hoping that this Campus might ‘fill the gaps’. Those gaps in the first sessions focused on progesterone levels (and support) in the early follicular and luteal phases.

First, Samuel Santos-Ribeiro from IVI Lisbon asked if elevated serum progesterone levels in the late follicular phase were clinically relevant or simply a mirror of what might be seen in the mid-luteal phase. He confirmed the great variability in progesterone levels at this stage, noting that elevated levels would regulate the timing of ovulation, follicular development and initiation of the secretory phase of the endometrium. However, he added that the potential effect of a premature progesterone rise on oocyte and embryo competence, whilst possible, ‘is still very debatable’. Strategies to avoid premature progesterone production were presently focused on lowering the gonadotrophin dosage, using a step-down protocol and pre-hCG measurements – ‘but more solutions may be on the way’.

In considering the mid-luteal phase Peter Humaidan from Aarhus, Denmark, stressed the huge disparity in LH levels depending on the trigger – 1.5 IU/l with an agonist trigger, 0.2 IU/l with hCG, and 6.0 IU/l in a natural cycle. LH activity, he explained, was ‘totally responsible’ for the steroidogenic activity of the corpus luteum and the upregulation of growth factors involved in implantation. Similarly, progesterone in the luteal phase initiates the secretory activity of the endometrium, and regulates the window of implantation. The question is, however, how much progesterone is needed for optimal implantation – with proposals now emerging for a difference in cut-off levels between fresh and frozen embryo transfers. Haumaidan’s conclusion was that 50 ng/ml progesterone in the mid-luteal phase might be a reasonable cut-off in a fresh transfer cycle, and 10 ng/ml in a frozen – and he concluded that monitoring mid-luteal progesterone levels in fresh and FET cycles was ‘the future way to increase reproductive outcomes’.

It was, however, revealing to hear the clinical take on these recommendations when Ying Cheong from the University of Southampton and co-ordinator elect of the SIG RE asked if it was possible to pinpoint the window of implantation and so define when best to transfer the embryo. Referring to the systematic review of Craciunas et al on markers of endometrial receptivity, Cheong reaffirmed that various degrees of receptivity exist within the window of implantation, but many of the markers used to assess their predictability of live birth were inaccurate and unreliable.(1) Currently, said Cheong, our tools for monitoring the uterus - pelvic ultrasound, hysteroscopy, endometrial biopsy, uterine fluid sampling and indeed blood tests for progesterone or LH – ‘are blunt’. However, given the shorter luteal length of the stimulated cycle and the strength of the evidence, she underlined luteal phase support for preparing the endometrium for pregnancy.(2) She further cited the recent ESHRE guidelines on ovarian stimulation to note that luteal support should start in a period between the evening before oocyte retrieval and three days post-retrieval – and continue until the pregnancy test at least. This regime, however, would change in FET cycles and would depend on the type of endometrial preparation used – HRT in a programmed cycle or a ‘natural’ or modified cycle transfer. ‘Clinicians need to be aware of “best days” to transfer will depend on medicated, natural and modified natural cycle timings,’ she said.

One ‘strategy’ in Cheong’s menu to fine tune implantation was adoption of a freeze-all approach, which was separately subject to an entertaining debate later in the programme. This debate, freeze-all ‘for all’, largely followed the evidence of recent RCTs, with potential for an improvement in LBR in predicted high responders and the likely reduction in OHSS risk following the ‘segmentation’ principle. However, Stratis Kolibianakis, a former co-ordinator of the SIG RE, after qualifying that ‘for all’ would actually mean ‘for most’, insisted that the freeze-all debate was long over – and the only difference between a freeze-all and a routine cycle was that the former simply skipped the first fresh transfer.

Another debate – much in line with a central theme of this meeting and also featuring Kolibianakis - raised the everyday applicability of endometrial receptor analysis (ERA) and thereby, according to Diana Valbuena Perilla from Igenomix, the developers of the ERA assay, a greater move towards personalised medicine in ART. The debate, both for and against, rested largely on the results of a five-year RCT published in 2020 and which for Kolibianakis was too full of methodological anomalies (including a 50% drop-out rate) to be considered strong evidence.(3)

ERA was listed in a table of more than 20 add-ons cited by Bulent Urman from Istanbul, who asked if it was possible to make the endometrium more receptive in women with implantation failures. Although there was much evidence that clinics do provide both preconceptional and treatment options in such cases, he too cited the 2020 Cochrane conclusion that ‘none have been evaluated in high quality studies that they actually help people to get pregnant and have a baby’. Moreover, by referring to the 2021 study of Pirtea et al, he proposed that endometrial causes of RIF are very rare and thereby ‘throw doubts into ERA and other add-on interventions to improve endometrial receptivity’. None of these add-ons, he added, have shown merit in RCTs.(4) Among the more common approaches, however, he mentioned endometrial scratch, ERA, intrauterine hCG, intrauterine peripheral blood mononuclear cells (PBMC), and intrauterine platelet-rich plasma – none of which again had as yet been proven as ‘beneficial beyond doubt’. Concluding his review, he noted that implantation involves a ‘complex cascade of events making it extremely difficult to hit the desired target with a single bullet’.

However, one of these approaches – endometrial scratch – may yet merit a rethink following the presentation of Nienke van Hoogenhuijze from UMC Utrecht. While acknowledging the doubts consistently emerging from the RCTs, she noted the heterogeneity of these studies, particularly with respect to the timing of the scratch, which her own analysis of individual patient data (which included data from 12 RCTs, including her own of 2020 but not the most recent of Metwally et al) had sought to resolve. Thus, this IPD analysis had included data on more than 4000 subjects and found an odds ratio for live birth after a first fresh or frozen transfer of 1.28 after adjustments – though this was just non-significant. She thus concluded that endometrial scratching is ‘likely’ to improve live birth rates, though still there are too many unknowns to ‘actively’ offer it as a routine procedure.

At the end of this journey through the luteal phase Swedish gynecologist Christina Bergh, drawing mainly on registry data compiled by the Nordic groups, highlighted the perinatal and maternal complications of pregnancies following FET, and showed that rates of preterm birth, low birth weight and small-for-gestational age were all higher than found in spontaneous pregnancies. Such trends, she argued, continue to gain in significance with the ongoing rise in rates of FET, as evident in all registries. The CoNARTaS study, like others, had also found maternal complications in a higher risk of hypertensive disorders of pregnancy from frozen than from fresh transfers, and most recently in programmed cycles (without a corpus luteum).(5)

Indeed, the corpus luteum, or its absence, was a recurring theme in this meeting. In the spontaneous pregnancy, explained Frauke von Versen-Höynck from Hanover, the corpus luteum developing from the leading follicle provides estradiol, progesterone, relaxin and other vasoactive products for the pregnancy’s greater nutritional needs, until the placental hormones begin to take over at around 20 weeks gestation. Following ovarian stimulation, several corpus lutea form, but retain their physiological effect after a single embryo transfer. In a programmed artificial cycle, however, the spontaneously developed corpus luteum is absent and, as many studies now affirm, the risk of pre-eclampsia is raised. This, explained Versen-Höynck, results from vascular impairment in endothelial function, arterial stiffness, and pulse wave velocity - which may be explained by low levels or relaxin.(6) Thus, the substitution of corpus luteum hormones by relaxin, for example, may be one preventive possibility. However, questions from the audience seemed to favour adaption of the transfer to a modified natural cycle with low dose stimulation - though it should also be said that some comments even questioned the wisdom of a programmed cycle at all.

1. Craciunas L, Gallos I, Chu J, et al. Conventional and modern markers of endometrial receptivity: a systematic review and meta-analysis. Hum Reprod Update 2019; 25: 202-223.
2. Van der Linden M, Buckingham K, Farquhar C, et al. Luteal phase support for assisted reproduction cycles. hrane Database Syst Rev 2015; 2015(7): CD009154.
3. Simon C, Gómez C, Cabanillas S, et al.
A 5-year multicentre randomized controlled trial comparing personalized, frozen and fresh blastocyst transfer in IVF. Reprod Biomed Online 2020; 41: 402-415. doi:10.1016/j.rbmo.2020.06.002.
4. Pirtea P, Cicinelli E, De Nola R, et al. Endometrial causes of recurrent pregnancy losses: endometriosis, adenomyosis, and chronic endometritis. Fertil Steril 2021; 115: 546-560. doi:10.1016/j.fertnstert.2020.12.010.
5. Ginström Ernstad E, Wennerholm UB, Khatibi A, et al. Neonatal and maternal outcome after frozen embryo transfer: Increased risks in programmed cycles Am J Obstet Gynecol 2019; 221: 126.e1-126.e18.
6. Van Versen-Höynck F, Schaub AM, Chi Y-Y, et al. Increased preeclampsia risk and reduced aortic compliance with in vitro fertilization cycles in the absence of a corpus luteum. Hypertension 2019; 73: 640–649.
doi: 10.1161/HYPERTENSIONAHA.118.12043

Get notified of new articles with our ESHRE newsletter.

Sign up and never miss an update