Single or double embryo transfer – A medical dilemma?

There has been considerable controversy over the number of embryos to transfer during assisted reproduction procedures. Couples often prefer double embryo transfer (DET) to optimize their chances of conceiving, while physicians advocate the use of single embryo transfer (SET) when possible, in a bid to reduce the risk for adverse maternal and neonatal outcomes associated with increasing order of pregnancy.
Countries such as Belgium, Sweden, and Finland have adopted a policy of elective SET but physician concerns surrounding the large-scale adoption of such a policy include lower pregnancy rates per embryo transfer, increases in the cost of treatment, and a prolongation in the time needed to achieve a pregnancy.
Zev Rosenwaks, a leading figure in the field of assisted reproduction, and team advocate a more individualized approach to decision-making. They are in the process of finalizing a mathematical model for predicting the order of pregnancy, which they say will simplify decision-making on the number of embryos to transfer.
Speaking with MedWire reporter Ingrid Grasmo, Rosenwaks discussed the challenges that doctors face in recommending SET or DET to patients, and how the mathematical model he has developed with his team could impact the process of decision-making by both patients and doctors in the future.

Zev Rosenwaks is Professor of Obstetrics and Gynecology and Reproductive Medicine at Weill-Cornell Medical College in New York. He is also director and physician-in-chief of the Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine at the Weill-Cornell Medical College and the New York Presbyterian Hospital. His work in the field of assisted reproduction is internationally renowned.

Is SET the ultimate solution?
The general consensus among physicians is that the goal of IVF should be a successful healthy singleton pregnancy.

Compared with singletons, twins have a five-fold higher rate of neonatal mortality, and a four-fold and eight-fold increased risk for cerebral palsy and low birth weight, respectively ^1-3.

Furthermore, twin gestations are associated with more maternal and perinatal complications, including 3.7-, 2.0-, and 2.8-fold higher risks for pre-eclampsia, placental abruption, and post-partum hemorrhage, respectively, compared with singleton gestations ².

Study findings agree that DET results in significantly increased multiple birth rates, with one meta-analysis showing a 63-fold higher multiple pregnancy rate compared with SET⁴.

Rosenwaks acknowledged that there are certain situations when having twins may be a desirable outcome, such as in older patients who may only have one chance of achieving a single pregnancy with IVF, in couples for whom severe male factor exists or when they require testicular sperm aspiration, or in patients who undergo preimplantation diagnosis for severe genetic disease and have only two unaffected embryos for transfer.

Indeed, a significant proportion of patients (20–90%) favor twins over singletons^5-7, mainly because many cannot afford more than one treatment cycle and wish to maximize their chance of pregnancy with the transfer of more than one embryo⁵.

“ Also, I will say that for the most part, with good obstetrical care, the absolute risk for twins versus singleton in the context of IVF is not that much greater as far as complications go. There is no question that the risk of having a premature baby is much greater with twins versus singletons. But with good obstetrical care, the outcome for the babies is quite good,” Rosenwaks said.

The addition of cryopreservation to SET has improved cumulative success rates in women younger than 36 years of age to levels comparable to those seen with DET, but the same level of success has not been achieved in women closer to 40 years of age or those with a poor prognosis.

Rosenwaks commented that if one looks at success rates, DET results in higher pregnancy rates at any age compared with SET and this has been confirmed in various prospective studies.

However, he added: “Of course, the argument is that when you deal with young women in clinics that have good or excellent success rates and where cryopreservation is highly successful, one can achieve excellent cumulative pregnancy rates per retrieval.”

Cryopreservation as an adjunct to SET would therefore seem sensible, but Rosenwaks believes that this is not the optimum approach for all women and that care should be individualized, taking into account age and the prognosis for pregnancy in individual couples.

He stressed that age is the single most important variable, as fertility in women approaching the age of 40 years declines to a point where SET may not be a viable option. In addition, clinicians need to account for whether it is the patient’s first IVF cycle, whether there have been previous multiple failures, and, of course, embryo quality.

“I think those are all important factors that need to be looked at in terms of assessing the likelihood of implantation in any given individual,” Rosenwaks stressed.

SET legislation – a push in the wrong direction?
The drive to reduce multiple gestations associated with DET has led to the implementation of an elective SET policy in some countries, a move that does not support the need for a more individualized approach to embryo transfer.

“It is reasonable to replace a single embryo in young women, in women that have anatomical problems with the uterus, and in women that are prone to premature delivery. It is certainly more desirable to have a singleton pregnancy than a multiple pregnancy, as it reduces the complication rate for the baby and for the mother,” said Rosenwaks.

“In clinics that have high implantation rates per embryo, meaning that the per embryo implantation rate is higher than 30%, or with blastocyst transfers when in some instances you can achieve a 50–60% implantation rate per embryo, it is not unreasonable to transfer a single embryo.”

However, when it comes to applying a blanket policy for elective SET, Rosenwaks disagrees.

“So I think one needs to individualize, and this is why we will leave in a strategy of modifying and tailoring the number of embryos based on patient characteristics. ”

“In general, young women tend to have a high implantation rates and high pregnancy rates, particularly in the first or second IVF cycle. This is especially true in women who show good ovarian response, and those who have high quality embryos.”

A potential concern arising from the adoption of a non-flexible legislative approach to the number of embryos transferred is a rise in the number of patients seeking reproductive care abroad, in countries where the transfer of two or more embryos is common.

“There are patients who fail in countries where there exists a limitation on the number of transferred embryos, and these women do seek help in countries where this limitation does not exist. This is especially true after they have failed several cycles,” Rosenwaks explained.

Physicians may be faced with increasing numbers of women returning to their home country for obstetrical care, which would incur high costs on health services as a result of complications associated with high-order pregnancies.

One way to prevent this would be to reduce the number of fetuses before the pregnancy progresses to a later stage, but Rosenwaks does not recommend fetal reduction as a routine strategy, and stated that, “clearly, that should serve as a last resort.”

Rosenwaks believes that fetal reduction “can be utilized in cases where it’s unavoidable, where women have failed IVF so many times, and where one needs to in order to overcome the low implantation in some rare instances where you have a high-order multiple, but not as a routine.”

The individualized approach – a predictor of pregnancy for every woman

Given the lack of consensus on the number of embryos that should be transferred, Rosenwaks and team developed a mathematical model to calculate the likelihood of a pregnancy and of a multiple pregnancy in individual women.

Mathematical models for the prediction of pregnancy have previously been developed, but each has reported limitations, and parameters that have well-known influences on IVF outcomes have not been accounted for.

For their model, Rosenwaks and colleagues combined three existing mathematical models, namely the binomial, barrier, and collaborative models, into a single equation while accounting for various factors such as age, implantation rate, multiple IVF failures, and whether the woman is a first-time IVF patient.

“So you can look at it with every single parameter that we know impacts on IVF success,” commented Rosenwaks.

“We’re able to show that the actual pregnancy rate and the multiple pregnancy rate can be predicted within 2–3% of the actual observation.”

The binomial model calculates pregnancy rates based on the implantation rate of individual embryos, while the barrier model calculates pregnancy rates based on factors that may impede attachment of an embryo.

The collaborative model calculation is based on the implantation rate of initial and subsequent embryos, given that it becomes progressively easier with each successful implantation. Indeed, the collaborative model is accurate in predicting both the pregnancy and multiple pregnancy rate, with the implantation probability of each subsequent embryo increased by 22% per previous implantation⁸.

“We used our retrospective data [close to 20,000 cycles with over 60,000 embryos] to see if we could predict pregnancy rates and multiple pregnancy incidents by using these formulae and by giving weight to the barrier and collaborative methods we reported in ESHRE as 50/50. It may be that to be the best predictive model it will be weighted slightly differently… We are working on it, but it’s very close to 50/50,” said Rosenwaks.

He added that retrospective analysis of the team’s own data and prospective data from the literature using their mathematical model showed that it was possible to predict the multiple pregnancy rate and the pregnancy rate based on a woman’s implantation rate with great accuracy.

“We’re able to show that the actual pregnancy rate and the multiple pregnancy rate can be predicted within 2–3% of the actual observation,”Rosenwaks pointed out.

He explained: “If the implantation rate was, say, 24%, then we would be able to tell you that the twin rate would be 8% when you replace two embryos in a 34- to 36-year-old woman.

“It’s still being refined but it works – at least in our hands – very, very well.”

He hopes that the mathematical model, used in the form of a simple calculator running on a computer, could become a routine part of IVF practice. But its use would have to be demonstrated by more than one clinic.

In addition, implementation of the tool would be limited in clinics dealing with very few IVF cases, as the sample size would be too small to give a statistically significant implantation rate.

“You would need a relatively large clinic with a record of consistent implantation rates and pregnancy rates to be able to apply it with confidence. But that’s true for any statistical tool I believe. You need certain numbers to prove it workable,” Rosenwaks said.

“So I think one needs to individualize, and this is why we will leave in a strategy of modifying and tailoring the number of embryos based on patient characteristics.”