Haste makes waste: don’t rush for a fresh embryo transfer in high responders

In this issue of Human Reproduction, Santos-Ribeiro et al. (2020) report a 9% rate of moderate–severe ovarian hyperstimulation syndrome (OHSS) when high-responder women underwent a fresh embryo transfer following a single 1500 IU dose of hCG given by injection 35 h after a GnRH analogue (GnRHa) trigger in the context of a randomized controlled trial (RCT). There was no moderate–severe OHSS in the elective freeze-all group, who received the GnRHa trigger without any hCG. Women in the freeze-all arm enjoyed statistically similar, albeit slightly higher, live birth rates following the first embryo transfer than the participants in the fresh transfer arm (41.6% vs 39.4%, respectively). The trial included a total of 212 women, who had ≥18 follicles measuring ≥11 mm on the day of GnRHa trigger. The median number of oocytes (17 vs 18) and median serum estradiol levels on the day of trigger (2720 vs 2714 pg/ml) were similar in the freeze-all and fresh transfer groups, respectively. More than 90% of the transfers were at the blastocyst stage and more than 85% of the participants received a single embryo. The trial is methodologically sound and reflects the current ART practice of elective single blastocyst transfer and embryo cryopreservation with vitrification. Thus, the results are generalizable to similar women with high ovarian response. OHSS is a potentially life-threatening iatrogenic complication faced by young, otherwise healthy women undergoing ovarian stimulation. Both exogenous hCG used to trigger oocyte maturation and, when present, endogenous hCG of pregnancy exert a sustained stimulus to luteinized granulosa cells of the corpora lutea to produce vascular endothelial growth factor (VEGF) (Pellicer et al., 1999). In the presence of an adequately large pool of granulosa cells, i.e. multi-follicular growth, excessive VEGF production leads to increased vascular permeability, the hallmark of OHSS. The syndrome very rarely occurs in the absence of hCG. Thus, withholding exogenous hCG and avoiding pregnancy until the demise of corpora lutea are two of the several available measures to prevent OHSS (Dahan et al., 2018). In the past, withholding the hCG trigger required canceling the cycle without oocyte collection, while avoiding pregnancy would require cryopreservation of any available embryos for transfer in a subsequent unstimulated cycle with lower success rates. In the case of pending OHSS, both ART professionals and patients needed to make a choice between a fresh and a frozen embryo transfer in order to balance the risk of OHSS with the benefit of a fresh transfer. However, several advancements have rendered this an easier dilemma and today there is hardly any reason to rush for a fresh embryo transfer for such women. GnRH antagonists have widely replaced GnRHa for pituitary suppression. More recently, progestins are also being used as an alternative (Ata et al., 2020). These enable replacing the usual 5000–10 000 IU hCG trigger with a single bolus of a GnRHa, since the gonadotrophs remain responsive to GnRH in an antagonist or progestin suppressed cycle (Fauser et al., 2002). A GnRHa bolus triggers an endogenous LH surge that lasts long enough to induce oocyte maturation, yet is too short to maintain viability of the corpus luteum, which is solely dependent on LH in the absence of the long-lived hCG (Fauser et al., 2002). However, endogenous LH is suppressed due to high levels of estradiol and progesterone produced by multiple corpora lutea, depriving them of the ubiquitous LH stimulus. Accelerated atresia of corpora lutea following a GnRHa trigger almost always prevents sufficient VEGF production to cause OHSS (Fatemi et al., 2014). However, the severely deficient luteal phase leads to unacceptably low live birth rates following a fresh embryo transfer after a GnRHa trigger alone (Kolibianakis et al., 2005). The addition of a single low-dose hCG injection 35 h after a GnRHa trigger was introduced to rescue the luteal phase and allow a fresh transfer while preventing OHSS (Humaidan, 2009). Initial studies, which did not exclusively recruit women at high risk of OHSS, suggested complete elimination of the syndrome while maintaining live birth rates (Humaidan et al., 2010). However, early onset moderate–severe OHSS, which develops due to exogenous hCG, was reported to occur with this method (Seyhan et al., 2013). Santos-Ribeiro et al. (2020) reaffirm these findings in the context of an RCT and provide a more precise estimate of the risk in a well-defined population. Deferring the embryo transfer to a subsequent cycle following high ovarian response is not only the safer choice but also the rational choice from an effectiveness standpoint. Above the threshold of about 15 oocytes, an inverse relationship between ovarian response and live birth rate following fresh embryo transfer has been reported (Sunkara et al., 2011; Polyzos et al., 2018). Widespread use of vitrification has changed the cryopreservation landscape (Rienzi et al., 2017). Currently, vitrified warmed embryos fare as well as fresh embryos in general. However, in the case of hyper-response to stimulation, a fresh embryo transfer has an even lower chance of live birth as compared with a vitrified warmed embryo transfer in a subsequent natural or artificial cycle (Roque et al., 2019; Wei et al., 2019). This is especially evident in the RCT by Wei et al. (2019), which most closely resembles the trial by Santos-Ribeiro et al. (2020), i.e. using vitrification and elective single blastocyst transfer. Despite the exclusion of women deemed to be at risk of OHSS by Wei et al. (2019), the difference between live birth rates was significantly in favor of elective vitrification (53% vs 41% for fresh transfer). Significantly lower live birth rates attained with a fresh embryo transfer in high-responders possibly render discussions regarding the optimal luteal support protocol following a GnRHa trigger, i.e. hCG supplementation, high doses of estradiol and progesterone without hCG or sequential small doses of hCG, obsolete for these women (Humaidan et al., 2015). The lack of a cost-effectiveness analysis (CEA) could be regarded as a limitation of the trial by Santos-Ribeiro et al. (2020). The only study on the relative cost effectiveness of elective freeze-all strategy compared with a first fresh embryo transfer concluded that the former was unlikely to be cost effective from a patient perspective (Le et al., 2018). Although the analyses were based on cumulative live birth rate until delivery or until exhaustion of all available embryos from the stimulation cycle, and took into account the costs of treatment complications including OHSS, the original RCT that informed the assumptions of the CEA recruited women with regular cycles and without polycystic ovary syndrome, i.e. not necessarily high responders (Vuong et al., 2018). Moreover, women considered to be at risk of OHSS were excluded. As would be expected, the average number of oocytes collected was 12, cumulative live birth rates were statistically similar with a small absolute difference of 1.3% favoring freeze all, and the incidence of moderate–severe OHSS was 0.8% and 1% in the freeze all and fresh transfer groups, respectively (Vuong et al., 2018). Clearly, these assumptions do not apply to the trial by Santos-Ribeiro et al. (2020), in which the participants had on average 19 oocytes, there was a wider margin of 2.2% between live birth rates favoring freeze-all, and most importantly a 9% versus 0% OHSS rate, again in favor of the freeze-all strategy. These differences can be expected to tip the balance against a fresh transfer. Furthermore, a CEA is reliably applicable only to the specific setting informing the assumptions and the willingness-to-pay value, which determines the decision for cost effectiveness and could vary across populations. Regardless, it is ethically challenging to attribute a monetary value to a complication with risk of morbidity and mortality, which can be simply prevented by withholding hCG and when similar, if not higher, success can be achieved with a deferred frozen embryo transfer as early as 1 month later, as shown by Santos-Ribeiro et al. Preventative measures for OHSS are not mutually exclusive and different measures can be taken, alone or in combination, for patients at different levels of risk (Dahan et al., 2018). While a fresh embryo transfer following a GnRHa trigger and low dose hCG rescue can be appropriate and cost effective for women at low risk, an elective freeze-all policy without any hCG at all is absolutely a safer and probably a more (cost) effective choice for women at high risk of OHSS, as clearly shown by Santos-Ribeiro et al. (2020). I don’t think equipoise still exists to justify any further trials to answer this question, ever. Drafting the manuscript. No funding was received for this commentary. The author has no conflict of interests associated with this commentary.