SUMMER REPLAY: How often might a surrogate baby actually be the surrogate's own?

In 2016, a surrogate's single embryo transfer resulted in a twin birth — and one of the babies, it turned out, was the genetic child of the surrogate and her husband. The story made headlines. But research suggests that natural conceptions during IVF happen more often than you think.

I first heard about the case of Jessica Allen in early autumn 2017. Allen, an American surrogate for a couple in China, had had only one embryo transferred to her uterus, but twins were eventually detected on ultrasound. She was told the embryo had split and that the twins were identical. After the birth, however, genetic testing eventually revealed that one of the babies was her own, naturally-conceived, genetically-related child.

I did not at first believe that the second child — clearly not identical — was their own child. A more likely explanation, I thought, was clinic error. But after seeing DNA tests and other documentary evidence, I understood that somehow their own child had been conceived alongside the IVF child.

The consensus seemed to be that such a thing was extremely rare. No one I asked had seen anything like it, they told me, and certainly not in a surrogate. But a few months after my documentary about their case aired on CBC radio, I came across a very interesting scientific paper in the journal Fertility and Sterility entitled, "Not all twins are monozygotic after elective single embryo transfer: analysis of 32,600 elective single embryo transfer cycles as reported to the Society for Assisted Reproductive Technology."

Below I describe that research, some research that preceded it, the various findings, and the implications for surrogate pregnancies.

6 minute read

And then there were two

It's an obvious question: if you put only one embryo into a uterus, how do you end up with two babies? About two per cent of all single embryo transfers (SETs) result in multiple babies.

The most common explanation is that the single embryo has split. Various research has suggested that splitting may be more common when intracytoplasmic sperm injection (ICSI) or assisted hatching are used, when the embryo is transferred into the uterus on the fifth day after fertilization, or when the embryo is thawed from frozen; the jury is still out.

But if a single embryo splits, the two babies born will necessarily be monozygotic or "identical" twins. According to a paper published in the journal Human Reproduction back in 2014, however, as many as one in five of those twin pairs are not identical.

The researchers, from Monash University in Australia, looked back at records from 4701 patients who had had a single embryo transferred. Of those, 137 twin pairs were detected on ultrasound by heartbeat and 109 were delivered live.

Twelve of the 109 were gender discordant, meaning that the pair consisted of a boy and a girl. Clearly, these children did not result from the same embryo. Rather, they were dizygotic, or "fraternal." Since it's just as likely to have same-sex fraternal twins — two girls who are fraternal, for instance, or two boys — as to have opposite-sex fraternal twins, the researchers estimated that approximately 24 of the twin sets were actually fraternal. Since only one embryo was put in, the presumption is that the second child had been conceived naturally.

A larger analysis of 32,600 single embryo transfer (SET) cycles was published in January 2018 in Fertility and Sterility. These researchers, from Albert Einstein College of Medicine in the Bronx, looked at all the SETs done in the US between 2010 and 2013 in women under age 35. Almost all the babies — 98.3 per cent — were singletons. But in 255 cases — 1.7 per cent — there were multiples.

For 249 of the infants, sex data was available. In 23 instances, the multiples contained both sexes. (There were 5 sets of triplets and 18 sets of twins.) They calculated the rate of sex discordancy at 9 per cent.

Like the Monash researchers, they reasoned that if 9 per cent were opposite-sex fraternal multiples, another 9 per cent or so were likely same-sex fraternal multiples. So they estimated that approximately 18 per cent of all twins born following SET include a child naturally conceived alongside the IVF-conceived individual.

Since that paper, an even larger study has come out. In October, Human Reproduction published a paper which looked at almost a million SETs carried out in Japan. The purpose of their study was to identify the risks behind embryo splitting, not the odds of natural conception, and their methods were a bit different. Still, the results were broadly similar: the total multiple pregnancy rate after SET was 1.60 per cent, and dizygotic twins were estimated to account for some 27 percent of those.

Surrogate own child

This finding is interesting in itself. A woman setting out to have only one child by IVF might conceive another, naturally, on the side. But the finding is of even greater interest for surrogates. They will relinquish all children at or shortly after birth.

It's important to note that these numbers don't transfer perfectly to surrogates. First and foremost, surrogates are told not to have unprotected sex for an extended period around the transfer; people seeking IVF for infertility are only sometimes told that. Obviously, this could have a huge impact on the number of unintended natural conceptions. (Jessica Allen and Wardell Jasper, the couple who made headlines, are clear that they followed doctor's orders on this, only having sex after detection of a fetal heartbeat.)

Women who are surrogates are also typically prescribed a drug to suppress ovulation. But patients undergoing frozen embryo transfers for themselves are also often prescribed this kind of drug. In fact, one-third of the discordant twin pairs in the 2018 study (2 of 6) were born to women taking such drugs. (This was not prescribed to Allen.)

Also, women who are surrogates are more likely to be fertile, whereas all the couples in these studies had fertility issues. In the 2018 study, 16.7 percent of the couples had "unexplained" infertility, and those couples were four times more likely to have sex discordant multiples after single embryo transfer.

All that to say, it is not easy to calculate how many unintended natural conceptions occur among surrogates. But as a thought experiment, let's try to estimate, using the year 2015.

According to the American Society for Reproductive Medicine (ASRM), 2,256 gestational surrogates gave birth that year to at least one baby following embryo transfer.

If the numbers from the 2018 paper were to hold true for surrogates, about 1.7 percent would have produced an "extra" baby, surplus to the number of embryos transferred. So around 38 babies could have been an unintended twin — either one created by a split embryo or one created through natural conception. (Personal communication with the 2014 researchers confirms that double embryo transfer or higher would be expected to confer the same risk of unexpected natural conception.)

About 18 percent of those unintended twins, if that 2018 research were to hold true for surrogates, could be the result of one twin having been conceived naturally on the side. That could be as many as 6 or 7 babies.    

For all the reasons above, there were almost certainly not that many accidental conceptions. It was probably many fewer. But there is good reason to suspect it might not always be zero.

In the Allen-Jasper case, the error was only caught when the Chinese embassy required DNA evidence before issuing passports. Testing showed that the intended parents were not the biological parents of one of the infants. The naturally-conceived baby was also a different race than his IVF twin — he had a white mother and a black father, whereas the surrogate baby had two Asian parents — so it would eventually have come to light even without testing.

Perhaps it's time for all babies born to gestational surrogates to be genetically tested, and for all surrogacy contracts to contemplate what will happen in cases where own babies are accidentally conceived.

This item was originally published in HeyReprotech on December 11, 2018.