At age 42, Amy Klein had already suffered three miscarriages and gone through several rounds of IVF. She wasn’t done trying to have a baby. But she worried that her age likely meant that her eggs had chromosomal abnormalities that kept her from getting pregnant.
So starting in 2012, the health reporter opted for a controversial addition to the fertility toolkit: She went through four rounds of additional egg retrieval, and had those embryos frozen and genetically analyzed for abnormalities.
The basic in vitro fertilization (IVF) process kick-starts embryo formation by fertilizing an egg, or many, with sperm in a petri dish. Klein’s plan was to add an optional and costly method called preimplantation genetic testing (PGT), to look for the most viable embryos in the bunch. Once the embryo reaches a stage called a blastocyst, technicians take a handful of cells—six or seven, about—to test them for genetic abnormalities that could result in disease or miscarriage.
For parents-in-waiting like Klein, the process holds great appeal. As adults age, the eggs and sperm they produce tend to be less healthy; eggs in particular tend to have more chromosomal abnormalities over time that may lead to miscarriage. Selecting embryos through PGT has the potential to help conceive more healthy children. It can look for common age-related genetic mix-ups like extra chromosomes, in addition to hereditary genes that can cause specific disease. If you’re already invested enough in having children—emotionally and financially—that you’re trying IVF, you’re likely to find some allure in increasing your genetic odds of success.
But while IVF by now has been accepted as a standard part of fertility care, PGT—which most clinics in the US offer for an additional $4,000 to $10,000, usually paid out of pocket—remains controversial.
🎧 For more intel on IVF technologies, listen to the Quartz Obsession podcast episode on egg freezing. Or subscribe via: Apple Podcasts | Spotify | Google | Stitcher.
“It’s never succeeded in doing what it promised, which was to improve [live births],” says Norbert Gleicher, an OB/GYN who treats infertility at the Center for Human Reproduction in New York.
Some clinicians think that PGT is indicative of a larger trend of IVF add-on procedures that are expensive and ineffective at best. Currently, just 4% to 6% of people who undergo IVF elect to try PGT. But as more people use assisted reproduction, more will face the choice of getting an extra genetic assist. And the decision may come down to more than mere success rates.
An evolutionary leap
Couples who choose PGT can look for a few different things.
Klein was interested in PGT-A, a test that looks at whether an embryo has an abnormal number of chromosomes, known as aneuploidy. Each embryo should have 46 chromosomes when it starts dividing; 23 from the egg and 23 from the sperm. But about 20% of the time, even in people who are perfectly healthy and younger than 35, an embryo will have too many or too few chromosomes, which can lead to miscarriage.
These rates are likely to be higher when the eggs used for a pregnancy come from someone who is older than 38. Klein’s doctors told her that because she was 41 when she started trying to get pregnant, aneuploidy was the likely reason she couldn’t stay pregnant.
Not all aneuploidies result in miscarriage, however. Some are so mild, people don’t even realize they have them, like those with three X chromosomes, or one X and two Ys. Others result in clinical presentations, like Down syndrome, which results from three copies of chromosome 21. Down syndrome is the most common version of aneuploidy, and can involve intellectual disabilities, heart disease, and gut problems, although the severity of the condition varies.
In PGT-A, embryologists can manually count the number of chromosomes and look to see if bits of the DNA on these chromosomes has shifted around, which can also result in miscarriage. Then embryologists can tell which embryos will have the best chance of growing into a healthy baby, and discard the rest if the couple chooses.
Another type of test, called PGT-M, screens for monogenic disorders. This kind of screening is useful if potential parents know they have a genetic trait that they don’t want to pass down, like cystic fibrosis. For the most part, these traits are recessive, meaning that a person needs two copies of the unhealthy gene to actually have the condition. People with only one copy of recessive traits are not sick, but they can still pass the gene on to their kids; if they’re trying to have a baby with someone who has the same recessive gene, there’s a 25% chance their child could wind up with two copies and develop the disease.
PGT-M looks at specific snippets of the cells’ DNA, and replicates them millions of times so it’s easier to see the specific pattern of nucleic acids within them. This way, embryologists can identify—and discard—any embryos that could result in disease.
Both PGT-M and PGT-A cost a couple of thousand dollars up front. Sometimes, clinics will give a discount if people want to test multiple embryos at once. Theoretically, it could save money and time because potential parents will only implant embryos thought to be more viable. “The time to pregnancy is shorter, there are fewer failed cycles and fewer multiple pregnancies, and financially it’s cheaper,” says Alan Copperman, a reproductive endocrinologist and medical director of the Reproductive Medicine Associates of New York.
But in practice, it’s much less straightforward.
Before PGT, embryologists used to guess which embryos were the most likely to take in the uterus, based on their size and shape (paywall). Usually this act of biological divination happened just three days after fertilization, when the embryo was only 16 cells big. It was a highly subjective process that even embryologists didn’t always trust. So they often implanted multiple embryos, hoping that at least one would stick.
But multiple implantations can result in multiple pregnancies, which are dangerous for both the pregnant parent and the unborn babies. There’s a higher likelihood that a person will deliver prematurely, develop gestational diabetes or anemia, or even miscarry. “What gave [IVF] a bad name is that we yielded to the need of making patients pregnant at some point,” says Antonio Gargiulo, an infertility specialist at Brigham and Women’s Hospital in Boston. “We’ve been lax in the number of embryos that we’ve allowed them to receive at once.”
In theory, PGT helps cut down the number of embryos that clinicians choose to implant. But it hasn’t totally eliminated the need to make tough decisions.
In the case of testing for a single disease, even if an embryo looks healthy, most clinics recommend that the results are confirmed by another form of genetic testing on the placenta during pregnancy, called chorionic villus sampling—which all people who are pregnant can elect to do. Although the risks are low, this kind of testing can also lead to miscarriage. And, in a way it negates the purpose of PGT, which is that it allows for testing before pregnancy.
In the case of testing for aneuploidy, looking at the DNA from more embryonic cells means there are more opportunities for a technician to miscount the number of chromosomes present. In some embryos, some cells are normal while others have aneuploidy—a phenomenon known as mosaicism. There’s no great data to say where mosaic embryos will grow to be healthy babies or not. People who get these kind of results face the difficult decision of trying to transfer some of these embryos, knowing that they may result in miscarriage, or discarding them, knowing they could have been perfectly healthy babies.
And because most people who elect PGT-A are usually already older and at a higher risk of aneuploidy that results in miscarriage, they may not be in a position to waste any embryos.
This was the dilemma presented to Klein, who, after her experience with PGT and IVF, wrote the forthcoming book The Trying Game about navigating the world of fertility care. Her doctors told her that none of her embryos came back normal. She and her husband opted not to transfer any of them.
Klein says that at the time, her doctors didn’t explain what “abnormal” meant. They didn’t unpack the concept of mosaicism, or that the tests could have been wrong. Instead, they told her that her eggs were causing her repeat miscarriages—which is why she made the difficult decision to use donor eggs instead.
But then, she miscarried a fourth time. Yet another specialist, a reproductive immunologist, told her that her body’s immune system was likely causing her miscarriages instead. At this point, Klein was 43. She knew that even if her eggs had been healthy when she started seeking fertility treatment, they likely weren’t now. She tried again to have a child, continuing to use donor eggs, which resulted in the birth of her daughter, when Klein was 44.
Now she knows more about the inconsistencies of testing, Klein is still unsettled about her choice. After her daughter was born, she went back to the clinic that told her her eggs were abnormal, to see if maybe she could implant them—this time under the watch of her last doctor. But they had all been discarded. “To this day,” she writes, “I wonder if any of those embryos could have resulted in healthy pregnancies.”
Klein’s difficult choice was created by an information void. But others face the same choice knowing their child could, or will, be living with certain medical conditions. Some families, out of other embryos and unable to afford donor embryos, must grapple with the choice to implant an embryo that has a mutation known to put their baby at a high risk of developing certain cancers later in life. Others have talked about the difficult decision they’d be faced with knowing that their child may have a condition like albinism or Down syndrome, which are not life-threatening but could come with physical differences that could be a target for bullying or discrimination.
This raises the question: Without certainty in genetic assessments of embryos that should have a binary result, how should hopeful parents go about accepting a range of risks?
The ethics of selection
In the US, IVF success rates have been on the decline. After peaking at 27% in 2013, across all age groups, they’ve gone down to about 23%, according to analysis by obstetrician-gynecologist Norbert Gleicher and colleagues at the Center for Human Reproduction in New York. Their work shows that blastocyst PGT started taking off right around 2010.
These trends may not be related to PGT at all; as the age of first-time parents increases, the likelihood of successful IVF will decrease.
But Gleicher believes that part of the reason that IVF success rates are on the decline is because PGT may unintentionally lead people to discard embryos that would have been healthy. While it may be the right choice for some people—those who know their carrier status or have a history of certain genetic conditions, or those who have had repeat miscarriages but also have plenty of embryos—it shouldn’t be for everyone.
Those doubts about the efficacy of PGT make its next chapter even more of an ethical conundrum. We’re approaching a world where parents can select embryos not just to avoid miscarriages or life-altering disorders, but to reduce the chances of more common, non-life threatening medical conditions. And even though research suggests that genetic predictors for typical height and intelligence aren’t effective, those choices are just a stone’s throw away from choosing desirable traits.
Nathan Treff is the chief scientific officer at Genomic Prediction, a biotech startup based in New Jersey leading this controversial charge. Unlike other PGT testing companies, Genomic Prediction promises to test an embryo’s risk of 13 different conditions based on the contributions of multiple genes. This testing isn’t black and white: Risk for these polygenic conditions, like type 1 and type 2 diabetes, high cholesterol, and certain types of cancers and heart disease, are influenced by mutations across hundreds of locations on the embryo’s genome—along with environmental factors once a child is born.
So what the company provides is an elevated risk score based on those many mutations. Instead of being told that an embryo has two copies of a recessive gene that would give the future child a 100% chance of developing, say cystic fibrosis, they may get a result that says their embryo has a 4% chance of having type-1 diabetes. Although this absolute risk sounds low, the company’s sample reports frame these results in terms of relative risk: The parents would also see that the general population has a 1% risk of developing the disease—so the embryo’s risk is quadrupled.
Two of the traits Genomic Prediction tests for are severe intellectual disability and idiopathic short stature, both of which have resulted in controversial press. In a Radiolab episode that aired last year, Steven Hsu, the company’s founder, said that while their tests couldn’t definitively say if an embryo would be severely cognitively impaired, “we’re just saying the chances the child will have a lot of difficulty in modern society—that probability is elevated.”
The same year, a piece from MIT Tech Review noted that the company is likely to attract couples who aren’t having trouble conceiving, but are interested in and financially able to try anything to eliminate the risk of their future child having any serious health condition. Of course, all parents want their children to be healthy—but what would this level of screening mean for those who can’t afford it?
Treff coauthored some of the work backing up the company’s specific tests, which cost $1,000 as a baseline, with an additional $400 per embryo (cost of egg retrieval and IVF not included). As of yet, no babies that have received this kind of testing have been born, but the company did confirm that there were multiple people pregnant with embryos that had undergone these tests. It’s unclear how successful this kind of testing will be, and whether it leads to more people discarding embryos that would have resulted in mostly healthy babies.
PGT won’t become mainstream until companies like Genomic Prediction provide clear and transparent information about the success rates of their tests, Gleicher believes. This grounded data is going to be particularly key for potential parents, who may be willing to try anything out of desperation, and the medical professionals that advise them.
“The public currently doesn’t understand what is going on,” says Gleicher. “And most of my colleagues don’t know what’s going on, either.”