This article has been updated with comments from two commercial NIPT testing labs and a clinical expert.
NEW YORK (GenomeWeb) – Results from an undercover performance test of five commercial noninvasive prenatal testing laboratories suggest a need for better quality standards, as three of the labs reported normal results in samples from women who were not even pregnant.
The assessment, which involved just two samples and was conducted by Tamara Takoudes and Benjamin Hamar, maternal-fetal medicine specialists based in Massachusetts, was reported in a Letter to the Editor in Ultrasound in Obstetrics & Gynecology this week. One of the commercial testing labs, Ariosa Diagnostics, provided financial support for the study but its laboratory was not informed about the samples it received.
Critics have called the study flawed because it used samples from women who were not pregnant, subjects they said the tests are not designed to analyze.
All NIPT relies on the presence of cell-free fetal DNA in the mother's blood, though companies differ in how they measure and analyze the DNA to find chromosomal abnormalities, such as trisomies or microdeletions.
The relative amount of free fetal DNA compared to maternal DNA, called fetal fraction, influences the accuracy of the test results and needs to be determined, according to some laboratories, while others maintain that reliable results can be obtained without knowing the fetal fraction. According to the authors, the amount of fetal DNA in the blood is influenced by a number of factors, including pregnancy stage, maternal weight, sample collection methods, and shipping conditions.
To see how different NIPTs deal with samples that have not just very little but no fetal DNA, the doctors sent blood samples from two non-pregnant women separately to five commercial testing laboratories in the US, declaring them as samples from women 12 weeks into their pregnancies.
Three laboratories – Sequenom, Illumina, and LabCorp's Integrated Genetics – returned normal, or negative, test results for a female fetus, according to Takoudes, even though the samples had no fetal DNA in them. The other two laboratories – Ariosa and Natera –provided no test result, citing insufficient fetal DNA.
Ariosa, Natera, and Sequenom all measure fetal fraction as part of their tests, but Sequenom incorrectly determined the fetal fraction to be around 4 percent in both samples. Illumina and LabCorp's Integrated Genetics, which offers a NIPT based on Illumina's Verifi test, don't measure fetal fraction.
The results "raise concerns about the need for quality standards in NIPT," according to the authors. Similar to other prenatal assays, like karyotyping and fluorescence in situ hybridization, which require a minimum number of fetal cells, "it seems reasonable that for NIPT, an analogous control measure should be applied," they wrote.
While NIPT has shown promising accuracy, "we urge professional medical and laboratory societies to set and enforce appropriate quality-control guidelines for NIPT that are consistent with standard laboratory practice as in other commercially available tests," they said.
Takoudes told GenomeWeb that she was concerned by the fact that NIPT providers increasingly market their tests to the general OB/GYN community, while no standardized methodologies exist for the assays or how their results are reported. Labs also seemed to differ in how frequently they were unable to return results for technical reasons, prompting her to compare them side by side.
Takoudes is part of Boston Maternal Fetal Medicine, a physician group with a number of locations in Massachusetts. Adam Wolfberg, one of her colleagues at Boston MFM, serves as the associate director of medical affairs for Ariosa.
Natera, which distinguishes between fetal and maternal DNA for its test, agreed with the need for better quality controls. Commenting on the study via e-mail, Solomon Moshkevich, Natera's vice president of marketing and business development, said that "from now on doctors and actual pregnant patients must think twice when receiving a negative result with female sex from one of those three labs [that provided a test result in the study], because they won't know if it reflects the fetal or maternal DNA."
Illumina did not agree with the design of the comparison, however, in particular the fact that it used samples from non-pregnant women, which the tests are not designed to analyze.
"This assessment of NIPT assay quality is simply not valid," Tristan Orpin, senior vice president and general manager of Illumina's reproductive health and genetics business, told GenomeWeb in an e-mail. "It is akin to submitting a female sample for a prostate-specific antigen test and receiving a 'normal' result." Illumina's Verifi test, he said, is not a pregnancy detection test but is designed to detect fetal aneuploidy from maternal serum.
He also pointed out that Illumina's reported result – 'no aneuploidy detected' – was correct, since there was no aneuploidy to detect, and that the presence of two X chromosomes was also a correct result, which did not imply there was a female fetus.
"Peer-reviewed publications have shown Verifi to have industry-best performance and the lowest false negative rates without the requirement to measure fetal fraction or exclude low fetal fraction samples," Orpin said.
Anthony Gregg, a maternal-fetal medicine specialist and professor at the University of Florida, agreed that Illumina and Integrated Genetics provided correct results in reporting no aneuploidy and two X chromosomes. Only Sequenom's result was incorrect because it did not measure the fetal fraction correctly, he said. Overall, "the authors fail to recognize differences in platforms, methods for determining fetal fraction and the important point that the test is useful only in established pregnancies," he told GenomeWeb via e-mail.
It is unknown how many pregnant women with extremely little or no fetal DNA in their blood may have received a false negative result from labs that do not take fetal fraction into account or measure it incorrectly. Because the prevalence of Down syndrome in women receiving NIPT screening is relatively low, the likelihood of receiving a correct negative result is quite high, according to Takoudes. "I think it's a very good test, but it should not be used as a diagnostic test, and the patient needs to receive a lot of counseling," she said.
Takoudes said she sees patients at different locations, which limits her choice of tests for various reasons, but she will continue to use NIPTs from different providers in her practice, currently mostly from Ariosa, Sequenom, and Integrated Genetics. "In the end, I still think that all of the labs are good companies, but I think there needs to be more quality control," which should ideally come from the companies themselves, she said.
Sequenom and Ariosa did not respond to requests for comment for this article.