When she learned she was pregnant, pediatric nurse Erin Acosta, already a mom of three, wasn’t worried. But at 37 years old, she knew she should be screened for Down syndrome, a genetic disorder more common in babies born to older mothers. “It was important for me to know ahead of the time of delivery so that I could prepare myself, my children and family,” she says.
Acosta decided to take advantage of non-invasive pregnancy testing, or NIPT, a relatively new prenatal screening method. NIPT analyzes the mother’s blood to assess the genetic makeup of a fetus and is recommended by medical professional organizations for high-risk pregnancies, including expectant mothers who are over 35. It can be performed earlier in the pregnancy than other types of screening, making it a popular option for anxious mothers.
NIPT makers promise peace of mind. “Clear answers,” advertises Ariosa. “Simple, clear results,” offers Sequenom.
“It seemed like a good idea,” Acosta says now.
But for this nurse with 14 years experience, NIPT proved to be a nerve-wracking experience, leaving her more confused and distraught than assured. The experiences of women like Acosta, coupled with the extensive marketing and commercial success of NIPTs, has experts worried. They say the companies have yet to prove with large-scale real-world studies that the technology really works as promised for some applications. They worry that it’s being introduced in an ad hoc way, and that women aren’t getting sufficient genetic counseling to interpret the results.
“There are significant benefits of NIPT, but the way it was introduced into current practice and use has been chaotic and outside how other technologies get introduced,” says Ronald Wapner, MD, director of reproductive genetics at Columbia University Medical Center. “NIPT was basically offered to the world by salespeople.”
When they first launched in 2011, basic NIPT screens primarily provided information on Down syndrome. NIPT, also known as cell-free DNA screens, is not considered diagnostic testing, but is often covered by insurance for high-risk and some regular pregnancies. Otherwise, it can range in price from $800 to $3,000.
NIPTs were quickly successful, with sales in the hundreds of thousands each year (more than 603,000 were sold in 2015 by just three of the six main US providers, based on publicly reported sales figures.) “If what you’re screening for is patients with Down syndrome, [NIPT] is fantastic,” Wapner says. “Nothing’s ever been better.” (Some of Wapner’s research is supported by industry funding.)
But starting in 2012, NIPT companies also began offering to detect chromosomal conditions trisomy 13 and trisomy 18, both of which cause severe birth defects and developmental disabilities. Then, in 2013 and 2014 they added the ability to detect “sex chromosome aneuploidies”—conditions like Turner or Klinefelter syndrome that are caused by missing or duplicated sex chromosomes.
NIPT also added genetic disorders called microdeletions, such as DiGeorge or Wolf-Hirschhorn syndrome, which previously couldn’t be detected at all without an invasive test. Because these disorders are rare, NIPT is not as accurate for these conditions as it is for the more familiar Down syndrome. But that’s not explained clearly in the brochures and websites that describe the screens.
A few weeks after sending in her Sequenom’s MaterniT21 Plus test, Acosta’s results came back. She learnt that she was carrying a girl. She also discovered the fetus tested positive for Turner syndrome, which can cause mild to moderate intellectual disability and other medical problems.
Acosta was shocked. She hadn’t even realized that she was being screened for Turner, she says. She was now terrified she’d miscarry, a frequent outcome in Turner syndrome pregnancies. “I cried more than I have ever cried in my life,” she recalls. “I worried constantly that any pain or cramp was because I was going to lose her.”
As with traditional prenatal screens, medical guidelines advise that a positive NIPT result be confirmed with a definitive test like amniocentesis, which analyzes cells from the womb. But because amnio has a very small risk of miscarriage, Acosta decided against taking the test. It wasn’t until after she gave birth last May that she found out her baby girl’s chromosomes were normal. The test was a false positive.
It turns out false positives are common with these kinds of screens. And yet companies are gearing up to offer even more of the tests, by expanding them to the millions of women in the US with low-risk pregnancies, a move they’ve announced in their annual reports and at investor meetings. The shift is supported by insurers. Recently, Anthem Blue Cross Blue Shield said it will begin covering NIPT for low-risk pregnancies.
Few large studies to date have measured how well microdeletion screens work in real-world conditions—the standard benchmark for a screening test. Some microdeletions are so rare that such studies are actually difficult to do. So while the data are still sparse, most published research finds the microdeletion and or sex chromosome aneuploidies (SCA) screens to be far less reliable than the Down syndrome screen.
But you won’t find that in the patient brochures of the four biggest NIPT manufacturers. All warn that NIPT is imperfect, but none provide comprehensive statistics for consumers.
Marketing materials usually point to detection or “accuracy” rates: If a problem is there, NIPT does well at picking it up, often upwards of 98%. But the gold standard for screening tests is actually a metric called positive predictive value (PPV). It answers the question: If you get a positive result, how likely is it to be correct? And by that measure, NIPT is far less successful.
For example, for the extremely rare chromosomal condition trisomy-13, or Patau syndrome, the average accuracy of NIPT in a 35-year-old woman is 91%, meaning that the test is fairly successful at picking up a possible problem. But a calculation based on the pooled results of 37 published studies of NIPT indicate that the average PPV is only 21%. That means NIPT catches most pregnancies with trisomy-13, but it makes a lot of false predictions too. A positive result is 79% likely to be wrong.
Acosta’s experience is all too common. From posts she read online, and from a genetic counselor she contacted after her NIPT, Acosta found out that for an expecting mother her age, a positive Turner result is more likely to be wrong than right. Nobody told her that beforehand.
“If I had to do it all over again, I would absolutely not choose to have any NIPTs done at all,” she says, at least not until the companies do more research and perfect the technology.
She’s not the only one who feels burned. One post about a false positive result on Babycenter.com elicited more than 1550 comments, mostly from women angry that they weren’t warned about high NIPT false-positive rates.
Women “are not saying, I don’t want the test,” says Mayo Clinic bioethicist Megan Allyse, PhD, who is currently leading a study interviewing women who’ve had positive test results with NIPT. “They’re saying: I wish I’d known” that the false positive rate was so high—“I wish somebody had explained this to me.”
Screens for microdeletions—especially rarer ones—seem to be even more error-prone. For the most common microdeletion, DiGeorge syndrome (also called 22q11.2 deletion), recent studies calculate PPVs as low as 5.3% and as high as 90%. Research by a team including Dutch and Israeli prenatal specialists and the director of medical affairs at Ariosa, an NIPT manufacturer that does not offer microdeletion screens, estimated PPVs for most microdeletions to be around 10%—meaning 9 false-positives for every 1 true positive. (A Sequenom-supported study last year calculated higher PPVs for its screens, but for several microdeletions there was only one case detected, making accurate statistics a challenge.)
Given the lack of solid data, the American College of Obstetricians and Gynecologists and the Society for Maternal Fetal Medicine recently recommended against routine use of NIPT for microdeletions in otherwise normal pregnancies.
Even when the tests are correct, it’s not always clear what a positive result means. The variability and possible consequences of Down syndrome are well understood, and most women have heard of it. But they probably haven’t heard of microdeletions like 22q11.2 or Cri-du-chat—some of which not even ob/gyns are familiar with, say genetic counselors. Both are rare, with unpredictable effects. Children with 22q11.2 (1 in 4,000 births) have heart problems, immune system problems and cognitive delays, but some have quite mild symptoms; Cri-du-chat (1 in 20,000 to 50,000 births) causes delayed development and intellectual disability.
Last July, NIPT-manufacturer Sequenom introduced an NIPT that goes one step further, looking not just for known problems but for any large missing or extra chunks of DNA—even ones unknown to medicine.
So it’s possible that a couple expecting a baby could hear something like this: Your fetus seems to have a genetic abnormality, but we’re not sure because we don’t really know how accurate this NIPT is, and in any case we can’t tell you what it means for the baby. “There’s an emotional cost to consider with that,” says Heidi Lindh, MS, CGC, executive director of the Genetic Support Foundation, an independent educational nonprofit.
Sequenom didn’t respond to a request from Quartz for comment on test performance statistics and recommendations against microdeletion screening. Ariosa general manager Dave Mullarkey, when asked about PPVs, emphasized that the company had genetic counselors on hand to help healthcare providers interpret results when needed.
Indeed, manufacturers often say it’s up to doctors to inform women. But doctors may not understand these stats either. In a recent study quizzing MDs and medical students on the mathematics of screening tests, three-quarters of respondents answered incorrectly.
Medical associations have asked manufacturers to provide information on PPV, but so far, only a few do. Independent groups are trying to fill the gap. The Perinatal Quality Foundation and the National Society of Genetic Counselors created a predictive value calculator that even math-challenged doctors can use to explain NIPT. Lindh’s organization is also putting together information about NIPT’s strengths and weaknesses.
“If there’s one thing I’d say, it’s that we need to educate patients and give them the autonomy to decide what they want to do,” says Wapner. Only then will the technology deliver on its promise: Clear information that really can ease an anxious expectant parent’s mind.