The relative myth of elective single embryo transfer

Abstract

The option of single embryo transfer (SET) has recently dominated the pages of this and other medical journals. Opinions, in regards to the utility of such an approach, appear to differ between Europe and the US. While US guidelines promote a more individualized approach, European opinions, at times, even advocate mandated practice patterns. The European approach, however, fails to recognize the rather significant differences in supportive arguments between the historical switch from multiple embryo transfers to 2-embryo transfers and the current discussion, favouring a switch from 2-embryo transfer to elective (e)-SET. In the former, a significant risk of (at times, high-order) multiple pregnancies was reduced without loss of pregnancy potential. In the latter, a comparably relatively low twinning risk is reduced at the expense of declining pregnancy rates, a need for more treatment cycles, a potential delay in treatment success and, potentially, higher treatment costs. These consequences of e-SET, together with the preference of some infertility patients to actually conceive twins, raise serious questions about the wide utilization of e-SET, as has been propagated by many authorities. According to US guidelines, e-SET, therefore, appears to represent an appropriate transfer option for only a small minority of IVF patients. Argument in favour of indiscriminate SET appears unrealistic and should be reconsidered.

Introduction

The treatment of infertility is, in most cases, based on replacement of the, naturally occurring, mono-follicular by a, medically induced, poly-follicular ovarian response. The consequence to the release of multiple oocytes is an increase in multiple pregnancies, which historically has been associated with practically all infertility treatments (American Society for Reproductive Medicine, 2000; Fauser et al., 2005). In vitro fertilization (IVF) is the first treatment which allows control over this risk by limiting the number of embryos transferred into the uterus (Gleicher et al., 2000; Fauser et al., 2005). Recent US data have demonstrated that voluntary restrictions of embryo numbers have already positively impacted on multiple pregnancy rates (Toner, 2002; Center for Disease Control and Prevention, Division of Reproductive Health, American Society for Reproductive Medicine/Society for Assisted Reproductive Technology, 2004; Jain et al., 2004), though a larger impact should become visible after the effects of the most recently introduced American Society for Reproductive Medicine (ASRM) guidelines will become apparent (American Society for Reproductive Medicine, 2004).

The effort to limit the number of transferred embryos has been primarily spearheaded by European investigators. Templeton and Morris probably published the key paper, demonstrating that, in young women, the transfer of more than two embryos did not further increase pregnancy rates (Templeton and Morris, 1998). Their data has since been confirmed by many others, including a national US data set (Center for Disease Control and Prevention, Division of Reproductive Health, American Society for Reproductive Medicine/Society for Assisted Reproductive Technology, 2004). The limitation of maximally two embryos transferred has, therefore, in young US women, become the generally recommended practice (American Society for Reproductive Medicine, 1999; Toner, 2002).

Two (2-)embryo transfers result, however, still in a significant twinning risk. Especially European (Gerris and Van Royen, 2000; Templeton, 2000; ESHRE Campus Course Report, 2001; Barlow, 2005; Bergh, 2005), but also some US academics (Rowland Hogue, 2002) have argued that, under maximal safety guidelines for the performance of IVF, no twinning risk was any longer acceptable. Indeed, none less than the president of the European Society for Human Reproduction and Embryology (ESHRE), at the time, voiced the opinion that ‘the achievement of singleton pregnancies alone should be considered as the leading quality of care parameter for IVF programs’ (Land and Evers, 2004).

Only e-SET can guarantee the non-occurrence of multiple pregnancies (Van Montfoort et al., 2006). Consequently, especially in Europe, e-SET has become a principle issue of professional discourse, at times supported by arguments which in the US might be considered radical (Saldeen and Sunderstrom, 2004). This article is meant to review the utilization of e-SET and discuss its place in modern infertility care.

Historical perspectives

A few years ago, a somewhat controversial paper in this Journal suggested that had the concept of multiple embryo transfer ever been submitted to an institutional review board (IRB) for approval, it would have never passed scrutiny (Lambert, 2002). In reaching this conclusion, the author, of course, failed to consider IVF’s historical developments, as was pointed out to him in a responding article (Gleicher, 2003), since IVF only painstakingly slowly evolved into a practical clinical treatment option, following the Brown birth (Edwards et al., 1980).

During the early days of IVF, rates per embryo were in the low single digits. Indeed, clinical pregnancy rates were so low, that only the transfer of, what today would be considered very large embryo numbers, resulted in minimally acceptable clinical pregnancy rates. It is also important to note that, at least in the US, IVF was developed without any federal funding. Its clinical success was, therefore, a precondition for its economic survival and also funded the ongoing research efforts in the field (Gleicher, 2003).

As Toner recently noted (Toner, 2002), the US can be proud about the progress IVF outcomes have made over time. Not only have clinical pregnancy rates continuously improved over the years, and have exceeded outcomes elsewhere (Gleicher et al., submitted for publication), but multiple pregnancy rates have, as noted above, started to decline. Indeed, the multiple pregnancy curves, observed over the years, followed a logical and expected trend: As the implantation rates per embryo improved, the transfer of large embryo numbers would increase multiple pregnancy rates. Once recognized by a responsible profession, this observation would, however, immediately result in appropriate countermeasures. The introduction of transfer guidelines, therefore, represented a logical and ethical next step (American Society for Reproductive Medicine, 1999). A key event in this process was the publication of a study by Templeton and Morris who demonstrated that, in selected and usually younger patients, the benefit from the transfer of multiple embryos on pregnancy rate peaks at two (Templeton and Morris, 1998). This paper, rightly, exerted considerable influence on the worldwide practice pattern of IVF (Meldrum and Gardner, 1998). In the US, it led to the recommendation that in younger women, with normal ovarian function, in principle, only two embryos be transferred. This recommendation was initially restricted to women below age 35 (American Society for Reproductive Medicine, 1999) but in a more recent recommendation was expanded to women above age 35 with exceptionally good ovarian function, as witnessed by large numbers of high quality oocytes and embryos (Gleicher, 2004).

The most recent US guidelines, however, also, for the first time, address e-SET and suggest the transfer of only one embryo in carefully selected and mostly younger patients (American Society for Reproductive Medicine, 2004), where e-SET has been shown to result in excellent pregnancy rates (Bergh, 2005; Mantikainen et al., 2001; De Neubourg et al., 2002; Gerris et al., 2002; Thurin et al., 2004; Lukassen et al., 2005). Some authors have, further, suggested that the continuous culture of embryos to blastocyst stage (day 5 or 6 after fertilization) allows for improved embryo selection and, therefore, enhances pregnancy chances after e-SET (Milki et al., 2004). Others have argued that day 3- or blastocyst stage-transfers do not differ in outcomes if day-3 embryo selection is done well (American Society for Reproductive Medicine, 2001). Whether embryos should be transferred at cleavage or blastocyst stages has, therefore, to be considered as unresolved. As with so many other issues, the correct answer is probably dependent on the patient’s age, her ovarian function and other clinical parameters. Whatever the embryo transfer timing, current US guidelines call for the consideration of e-SET in only carefully chosen patients who, with the transfer of more than one embryo, would run a significant risk of twinning (American Society for Reproductive Medicine, 2004). US guidelines (Table I), however, are voluntary. This means that the ultimate decision remains with patients and physicians.

Such an approach, of course, allows for the individualization of patient care, and the consideration of patient preferences. It thus follows closely the ethical and legal concepts of self-determination that patients have become accustomed and privy to within the US health care system. It, however, also means that the decision, how many embryos should be transferred, may be subject to undue biases such as economic pressures and insurance circumstances (Gleicher, 1998; Jain et al., 2001), or limited patient knowledge about risk factors, associated with the various possible outcomes (Ryan and Van Voorhis, 2004).

While some European colleagues have been sympathetic to such a flexible approach towards embryo transfer (Hernandez, 2001; Bocket and Tan, 2004; Barlow, 2005), many others consider such a voluntary approach as too soft. Indeed, some European countries, in almost unprecedented ways, have withheld choices from patients and their physicians as to embryo transfer numbers and restricted available options through legislation (German Act for the Protection of Embryos, 1990; Bernat and Vranes, 1993; Jones and Cohen, 2004).

We are describing such legislative interference into the patients’ ability to self-determine their course of treatment as practically unprecedented because, except for legislative controls over induced abortions and sterilizations, largely abandoned over the last few decades, we are unaware of any other, comparable, medical circumstance where legislative intervention, in similar ways, has interrupted the physician–patient relationship. Yet, surprisingly, European colleagues have, at times, been supportive of such an approach and, indeed, have also argued in favour of legislative intervention in a recent drive to mandate e-SET (Saldeen and Sunderstrom, 2004; Braat et al., 2005).

Europe and the US thus appear to approach the question, as to how many embryos to transfer, from different philosophical viewpoints. The US, without doubt, transfers more embryos (Gleicher et al., submitted for publication) and this difference in practice pattern also appears to have carried over to the discussion of e-SET.

European investigators recently reported that cumulative pregnancy rates were similar, and multiple pregnancy rates were radically reduced if, in place of a multiple embryo transfer, consecutive e-SET cycles were performed (Thurin et al., 2004; Lukassen et al., 2005). On a superficial level, this observation, indeed, appears to lend credence to the argument that e‐SET should be the transfer method of choice. A more recent European study, however, quite conclusively demonstrated that in unselected patients e-SET significantly reduces pregnancy rates in comparison to 2-embryo transfers (Van Montfoort et al., 2006). As this article will demonstrate, e-SET, therefore, represents an appropriate transfer choice in only a small minority of IVF cycles.

The (relative) risks of multiple pregnancies

The principle motivation for e-SET is the prevention of multiple pregnancies. There is absolute consensus in the medical literature that the risk to mother and offspring increases with increasing order of pregnancy (Ventura et al., 1999). We frequently forget, however, that this risk association is not linear: While twin pregnancies undoubtedly represent a higher obstetrical risk than singletons, the risk differential between these two is, assuming modern obstetrical care, relatively minor in absolute terms, though, in relative terms, can increase risks to mother and child by up to approximately 50% (Kogan et al., 2000).

Triplet pregnancies, despite improved outcomes with modern neonatal care, represent significant additional risks in comparison to twins (Lipiz et al., 1989). The profession is, therefore, rather unanimous in agreeing that they are to be avoided, a sentiment reflected in the wide utilization of 2-embryo transfers in recent years. Beyond triplets, the perinatal risk becomes truly unacceptable, mostly based on the very high prevalence of premature and very premature births (Gleicher et al., 2000).

Risk is a universal presence in medical care. Every medical intervention creates some risks for adverse outcomes. Much of competent medical care, in all medical specialties, is, indeed, based on the ability to correctly assess risk versus benefit of medical interventions. The risk of twin pregnancies should, therefore, not be seen as an absolute one, but needs to be considered relative, together with what patients may consider as their own, specific potential benefits in delivering twins.

Such an evaluation traditionally starts with a risk assessment, defined by the prevalence of complicating circumstances for mother and offspring. It, then, continues, however, with potential medical, as well as social, benefits to the patient. That infertility patients may derive such social and medical benefits from the birth of a multiple pregnancy is often overlooked. Potential social benefits are best characterized by the repeatedly reported desire of infertile patients for low order multiple births (Gleicher et al., 1995; Karla et al., 2003; Ryan et al., 2004). One can understand such a desire especially well when it also correlates with length of infertility and advancing female age (Gleicher et al., 1995), even if obvious excessive parenting stress is considered that comes with the birth of a twin gestation (Glazebrook et al., 2004).

A quite obvious potential medical benefit from twinning has been completely overlooked by the literature. The distinct possibility that the delay required by any second conception may result in additional loss of fertility potential has remained unreported. Especially for the older patient this is a very relevant concern. She faces two theoretical options: the first is a 2‐embryo ET with considerable chance of twinning; the alternative is an e-SET which, if not successful, can be immediately followed by a frozen/thawed cycle. The literature suggests that both of these options create an approximately equal chance of conception, with greatly diminished twinning risk (Thurin et al., 2004; Lukassen et al., 2005; Pandian et al., 2005). What is overlooked, however, in favouring the second option is that the occurrence of a twin pregnancy is not, necessarily, an undesired option since it produces two children, at once, for the infertile couple. Especially for the older patient, this may represent a significant benefit since, following a first, successful singleton pregnancy, she may, for many reasons, later find herself deprived of the opportunity to conceive and deliver a second child. Especially in older patients, and in women with premature aging ovaries, even one year of delay can make a significant difference in conception chance. In other words, not every twin pregnancy can necessarily be split into two, successful, singleton pregnancies. Avoiding a twin pregnancy by e-SET may, therefore, in such patients result in a net loss of one child.

In their own, private cost–benefit calculation women with older ovaries may, therefore, conclude that, under such circumstances, even considering the reported excessive maternal and fetal risks of twin, over singleton, pregnancies, a delayed, second pregnancy attempt may create more overall risk that she will never have a second child, while a twin pregnancy, of course, would immediately offer such an outcome. This observation, alone, strongly suggests that SET should never be considered a universal goal of IVF and/or be mandated for all patients.

The medical literature is permeated by the allegation that the desire of many infertility patients for twins is the consequence of factual ignorance about risks and long-term consequences (Karla et al., 2003; Glazebrook et al., 2004; Ryan and Van Voorhis, 2004; Ryan et al., 2004). Yet, infertility patients are, as is well known, of generally higher socio-economic standing, and they are usually more educated than patients in most other medical fields (Gleicher et al., 1995). And, indeed, when their knowledge of the subject matter was tested, it uniformly was found to be remarkably high (Gleicher et al., 1995; Karla et al., 2003). The literature, therefore, does not support the denial of patient choices, based on an argument of factual ignorance. Indeed, if such an argument were to be upheld in the generally highly educated infertility population, it would bring the informed consent process in the rest of medicine, where patients are usually much less involved in their treatment choices, to a screeching halt.

A patient’s right to self-determination also allows her to make choices against medical advice. Repeated legal precedent in US courts has established embryos as the personal property of the ‘parents’. They, therefore, at least legally, would appear to have an absolute right to determine when, and under what circumstances, they wish those embryos to be implanted, even if choosing a higher-risk over a lower-risk outcome.

Physicians, of course, also have rights, and may refuse treatments they consider inappropriate. The medical community has, however, not always been very consistent in its approach towards twinning risks. For example, when performing selective multifetal reductions, the risk of twin pregnancies is generally considered acceptable, as evidenced by the widely practised routine of reducing high order multiples to twins, and not to singletons (Macones et al., 1993; Smith-Levitan et al., 1996).

The conclusion of all of this is, of course, that while singleton pregnancies do represent the most desirable outcome for most IVF cycles, they do not represent the most desirable outcome for all patients and under all clinical and social circumstances. Consequently, the achievement of a singleton pregnancy by e‐SET may represent the most desirable clinical approach for many, though, most certainly, not for all patients.

Costs and cost-effectiveness

The economic costs of multiple pregnancies are, indeed, multiple in nature (Callahan et al., 1994). They do not only include the service charges to establish pregnancy, see it through delivery and any eventual neonatal services, but should also address life-long societal benefits from the birth of a healthy child and/or life-long societal costs from the birth of a child with disabilities. It should, therefore, not surprise that the evaluation of cost-effectiveness has remained controversial. As noted before, proponents of e-SET have suggested that this practice is cost-effective (Lukassen et al., 2004; Bergh, 2005; Lukassen et al., 2005; Thurin Kjellberg et al., 2006), though a complete cost-effectiveness analysis, involving prospective randomization and long-term costs, has never been performed. When Lukassen and associates compared charges (up to 6 weeks of pregnancy) between e-SET and 2-embryo control cycles, the mean number of IVF cycles, per live birth, with e-SET was 4.3, but with 2-embryo transfer only 2.8, and direct medical costs were practically identical (Lukassen et al., 2005). Thurin Kjellberg et al. who reported e-SET to be cost effective, included in their cost-effectiveness analysis costs up to only six months after delivery and, of course, did not consider the economic value of live births, or costs of disability, over a lifetime (Thurin Kjellberg et al., 2006).

Costs may also greatly vary based on geography. Within the US, medical charges differ greatly between regions and costs in Europe are known to be significantly lower than in the US.

The cost to provide services can also vary with volume. For example, the cost to provide IVF services does not correlate in linear fashion to volume. Since a considerable portion of overall costs are fixed, unit costs will decrease with increasing cycle volume until a next incremental step is reached, where capacity needs once again to be expanded.

And, finally, costs also depend on insurance coverage (Gleicher, 1998). In Europe, fertility services are frequently an insured benefit (Jones and Cohen, 2004). In the US, they often are not (Gleicher, 1998; Jones and Cohen, 2004). Where governments sponsor insurance coverage, they often also exert the right to deny such services. Governments in Europe are, therefore, often involved in establishing practice patterns for IVF.

Cost-effectiveness also involves the very individual values a patient assigns to achieving her goal of pregnancy, and to doing so in timely fashion. Europe’s utilization of IVF at approximately twice the US rate is, therefore, significant (Gleicher et al., submitted for publication), and suggests, together with lower per cycle pregnancy rates, that European women will have to wait longer for pregnancy success than their US counterparts. Equally important, of course, is what percentage of European and US patients, respectively, will conceive at all, and will reach their goal of successful delivery; and how many will fail. Since, as noted earlier, available cost data are still limited, it is quite difficult to establish valid cost-effectiveness assessments. To use the cost-effectiveness argument in support of e-SET, therefore, appears, at least as of this point, premature.

The patient’s right to free choice

Under the Helsinki Declaration a patient’s free choice was established as a basic human right (World Medical Association Declarations of Helsinki, 2005). Free choice is, of course, predicated on access to relevant information before reaching a decision. Providing patients with balanced and accurate information is, therefore, an absolute prerequisite for a properly executed informed consent. It should, amongst others, address maternal risks during pregnancy, fetal and neonatal risks, including life-long disabilities from prematurity, as well as the social and economic pressures that arise from multiple births (Bocket and Tan, 2004; Glazebrook et al., 2004). Such risks should not be minimized because, even with modern obstetrical care, twin pregnancies carry significantly increased risks over singletons.

As already noted earlier, infertile patients, if given a choice, actually prefer low order multiples (Gleicher et al., 1995; Karla et al., 2003; Ryan et al., 2004). Assuming that patients have reached their decision after receiving proper and complete informed consent, one cannot deny that they have an absolute right to make such a choice.

Since the choice to have, under such circumstances, a potential multiple birth should be considered as ‘elective’, it would seem appropriate to treat this choice akin to the choice of other elective procedures. Payers may, therefore, be entitled to refuse financial coverage.

The clinical argument

The learning curve in IVF has been steep. We are only a few years removed from the groundbreaking study of Templeton and Morris (Templeton and Morris, 1998). Their work has, since, been repeatedly proven correct (Devreker et al., 1999; Center for Disease Control and Prevention, Division of Reproductive Health, American Society for Reproductive Medicine/Society for Assisted Reproductive Technology, 2004). The logical conclusion to these findings, represented by voluntary US guidelines (American Society for Reproductive Medicine, 1999; American Society for Reproductive Medicine, 2004), was that young women should, with few exceptions, never be transferred with more than two embryos (Table I).

The wisdom of these conclusions lies in their flexibility and in the resultant adaptability of IVF cycle protocols to, not only the chronological, but also the ovarian age of the patient. Approximately 10% of young women suffer from prematurely aging ovaries (Nikolaou and Templeton, 2004). They are, of course, disproportionately represented amongst infertile patients (Nikolaou and Templeton, 2003). Indeed, in a recent review of our own practice experience, they represented more than 60% of patients under the age of 37, when sole diagnostic criteria were either elevated baseline FSH levels and/or severe ovarian resistance to stimulation with gonadotropins (Gleicher N, unpublished data). The percentage of patients with prematurely aging ovaries above age 37 may, therefore, be even higher. Even proponents of large scale e-SET would not consider such women good candidates since normal ovarian function has been uniformly cited as a selection criteria for e-SET (Gerris and Van Royen, 2000; Templeton, 2000; Standell et al., 2000; ESHRE Campus Course Report, 2001; Mantikainen et al., 2001; De Neubourg et al., 2002; Gerris et al., 2002; Rowland Hogue, 2002; Land and Evers, 2004; Thurin et al., 2004; Barlow, 2005; Bergh, 2005; Lukassen et al., 2005). Any logical embryo transfer policy, even when considering e-SET, has, therefore, to be flexible enough to adjust embryo transfer numbers to ovarian age. Current US guidelines provide for such flexibility into all directions.

This is important to note because ovarian function may deviate from the age-specific norm in two ways: ovarian age may be enhanced (i.e., older) or be reduced (i.e., younger). As premature aging of the ovaries is characterized by diminished oocyte/embryo numbers and poor oocyte/embryo quality (Nikolaou and Templeton, 2003; Nikolaou and Templeton, 2004), so is the younger ovary characterized by large oocyte/embryo numbers and excellent oocyte/embryo quality (Volpes et al., 2004; Yih et al., 2005). Therefore, current US guidelines also provide for flexibility in reducing embryo numbers if patients demonstrate evidence of ‘younger’ ovaries (American Society for Reproductive Medicine, 2004; Table I).

In practical terms this means that, considering the obvious limitations in our current ability to assess ovarian function, younger patients with excellent ovarian function, indeed, become candidates for SET. However, the medical decision to recommend e-SET is individualized, based on the patient’s ovarian age, and not outright mandated, or even legislated, as has been suggested in Europe (Saldeen and Sunderstrom, 2004; Braat et al., 2005).

Why 2-embryo transfer differs from e-SET

The choice to transfer only two embryos in young women was relatively simple based on the logic of Templeton and Morris’ original article (Templeton and Morris, 1998) and confirming data (Devreker et al., 1999; Center for Disease Control and Prevention, Division of Reproductive Health, American Society for Reproductive Medicine/Society for Assisted Reproductive Technology, 2004) After all, their basic argument was that 2-embryo transfer maintained the same pregnancy rate as 3-, 4- or 5- embryo transfers. This is, however, not the case when it comes to e-SET!

While some have, indeed, claimed that pregnancy rates remained unaffected by e-SET, those who made such claims did so in uncontrolled fashion by either comparing program data to prior years or by carefully selecting patients for e-SET (Mantikainen et al., 2001; De Neubourg et al., 2002; Gerris et al., 2002; Thurin et al., 2004). Nobody who argues in favour of e-SET has ever been able to prove that e-SET can maintain pregnancy rates at the level of 2-embryo transfers. Indeed, a very recent Cochrane review confirmed that e-SET reduces the pregnancy rate and that only after a second, frozen/thawed e‐SET a comparable pregnancy and live birth rate is reached. (Pandian et al., 2005) and an even more recent randomized study also confirmed this point (Van Montfoort et al., 2006). If published studies and editorial opinions are scrutinized, their language usually speaks of similar or acceptably high pregnancy rates in carefully selected patients, but never claims, as Templeton and Moris were able to, when they propagated 2-embryo transfer, equivalency in pregnancy rates. Indeed, the studies that are most frequently quoted in support of e-SET are actually based on the premise that there is no equivalency between 2-embryo transfer and e-SET because they specifically claim similar pregnancy rates between a single IVF cycle with 2-embryo transfer, and two IVF cycles (one fresh and one frozen/thawed) with e-SET (Thurin et al., 2004; Lukassen et al., 2005; Pandian et al., 2005).

To advocate a switch from 2-embryo transfer to e-SET is, therefore, very different from Templeton and Morris’ argument in favour of 2-embryo transfer, in place of multiple embryo transfer. Both recommendations, undoubtedly, reduce multiple pregnancy rates; however, the switch to 2-embryo transfer did so without reducing overall pregnancy rates, while any large scale switch to e-SET, equally undoubtedly, will result in such a reduction. Considering the already very low European pregnancy rates, in comparison to US rates, such an approach should be troubling for European patients.

One additional consideration: When Templeton and Morris recommended the 2-embryo transfer, it was meant to replace the transfer of three or more embryos and, with it, the risk for triplet or even higher order, multiple pregnancies. In contrast, the recommendation for e-SET replaces mostly 2-embryo transfers and, with it, at most, a twinning risk. The perinatal risks, associated with twin pregnancies, are, of course, smaller than those from a triplet or higher order multiple pregnancies. Consequently, the switch from multiple embryo transfer to 2-embryo transfer reduced a clinically highly significant high order multiple pregnancy risk, with no loss of pregnancy potential, while the currently widely advocated switch from 2-embryo transfer to e-SET reduces a much smaller twinning risk, but at the expense of lower pregnancy rates. Any recommendations for e-SET, therefore, warrant considerably more scrutiny.

Who is a candidate for e-SET?

All of the above described considerations suggest that the pool of candidates for e-SET may be surprisingly small. As Table II describes, patients may qualify for e-SET based on social and medical considerations. For example, patients who are fully opposed to any twinning risk, whether for medical, social or economic reasons, will, of course, be appropriate candidates. Others may have medical histories which make multiple pregnancies a contraindication. One may consider in this category the woman with a previously ruptured uterus, with a history of premature labour in a singleton or twin pregnancy or the patient with a Müellerian uterine anomaly that may place her at significant risk for premature labour, even with a singleton pregnancy. Both of these patient groups are, of course, willing to accept reduced pregnancy chances that come with e-SET, and are supported in their decisions by their individual risk–benefit analyses. They are, therefore, the clearest choices for e‐SET.

The remaining candidates will be more controversial because most, if not all, will decrease their pregnancy chances by switching from 2-embryo transfer to e-SET. Furthermore, e-SET may result in a need to conduct two cycles in order to achieve a similar pregnancy rate to the one expected from a single 2-embryo transfer (Pandian et al., 2005).

Candidates for e-SET would have to be young (preferably below age 35), lack any evidence of prematurely aging ovaries (i.e., should have normal baseline FSH levels, no evidence of diminished ovarian reserve or ovarian resistance to stimulation), should have a large number of high quality oocytes/embryos (in our program this means four or more high quality embryos on day 3 after fertilization in women under age 35) and no uterine or other medical factors that may reduce IVF pregnancy chances. Whether they also should have blastocyst-stage embryo transfers has so far remained undetermined in the literature.

How many patients, under such guidelines, will qualify for e-SET will, of course, vary depending on what kind of patient population an IVF centre serves. In preliminary calculations, we concluded that in our program less than 10% of patients would qualify for e-SET since our centre serves a disproportionately older patient population. However, even where the population is much younger, only a small minority of patients would appear candidates for e-SET.

Conclusions

While e-SET obviously should be the embryo transfer of choice in qualified candidates, considering the small pool of candidates for e-SET, the argument in favour of large-scale e-SET, based on US guidelines, appears out of place. The US should continue to pursue the promulgation of voluntary clinical guidelines, rather than allow for government interventions (Barbieri, 2005), which, as the experience in some European countries has demonstrated, restrict patient choices, interfere with the physician–patient relationship and, in the end, result in inferior pregnancy outcomes. European countries, which either have already implemented, or are contemplating to implement, strict e-SET guidelines and/or legislation, are advised to reconsider such an approach in favour of flexible embryo transfer guidelines, akin to those issued in the US.

References

Author notes

1Center for Human Reproduction, New York, NY, 2Foundation for Reproductive Medicine, 3Department of Obstetrics and Gynecology, Yale University School of Medicine, Chicago, IL and 4Department of Epidemiology and Social Medicine and Department of Obstetrics and Obstetrics and Gynecology & Women’s Health, Albert Einstein College of Medicine, Bronx, NY, USA