Up to 8% of pregnant people develop high blood pressure that can damage their organs or even kill their babies. Often there’s no sign of the condition, known as preeclampsia, until late in pregnancy. Now, by analyzing patterns of gene activity through RNA markers in a pregnant woman’s blood, researchers have come up with a test that could one day predict the complication months before it develops.
“It’s an exciting paper,” says maternal-fetal medicine specialist Andrea Edlow of Massachusetts General Hospital, who was not involved with the work. The approach could lead to new, more personalized treatments for preeclampsia, she says, which would be “a whole different way of doing medicine.”
Preeclampsia is thought to develop in part because blood vessels in the placenta, an organ that helps nourish the growing fetus, don’t develop properly. It can progress to eclampsia, or extremely high blood pressure that can cause seizures, stroke, and death. So when a pregnant woman develops signs of preeclampsia, such as high blood pressure and protein in the urine, their physician may deliver the baby early. But that, too, carries health risks.
Because there are no reliable early biomarkers, doctors now calculate a woman’s risk early in pregnancy using factors such as race (the condition is more common in Black women), age, and body mass index (BMI). Those at higher risk may then be put on low-dose aspirin, the most common preventive measure. But only about 4% of women identified this way actually go on to develop preeclampsia.
In the new study, researchers analyzed blood samples from 1840 pregnant women of several ancestries for RNA, produced when an active gene’s DNA is read to make a protein. (The RNA has leaked into the blood from maternal, fetal, and placental cells.) The team first used an artificial intelligence approach called machine learning to find RNA patterns that fluctuate with normal fetal development over the course of pregnancy. The work firmed up smaller studies suggesting these RNA signatures could predict fetal age as well or better than ultrasound.
Next, to explore whether such patterns could reveal a health problem, the scientists compared RNA patterns for blood samples taken at about 6 months into pregnancy from 72 women who developed preeclampsia with patterns from 452 normal pregnancies. They found a signature based on seven genes—some known to be involved in placental development—that predicted preeclampsia in 75% of women who later developed it. Overall in women who had this RNA pattern, medical records showed 32% went on to develop the condition. That is seven times better than the current prediction method based on factors such as race, age, and BMI, they report today in Nature.
“This work is a step in the right direction,” says clinical chemist Rossa Chiu of the Chinese University of Hong Kong, who helped develop prenatal DNA blood tests for Down syndrome that paved the way for the new work. Edlow, who co-authored a commentary on the paper, adds that there are different types of preeclampsia—some are immune related, others vascular, for example—and distinct RNA signatures for each could help researchers develop tailored drugs.
The test would also be important because it does not depend on race (including race did not improve the predictions), but instead on what is actually happening at a biological level in that individual, the authors say. Race-based risk calculators have been criticized because race is a social construct not based on genetics, and such calculators can result in worse health care for minorities.
One caveat, however, is that such a test would generate many false positives: Among every three women who had the signature RNA pattern in the study, two did not develop preeclampsia. Putting these individuals on low-dose aspirin shouldn’t cause harm, but the test result could create needless anxiety. Study leader Thomas McElrath, a fetal-maternal medicine specialist at Brigham and Women’s Hospital, suggests women who test positive could be sent home with a blood pressure cuff so they could monitor themselves and contact their doctor if they developed high blood pressure.
The startup biotech company Mirvie, which sponsored the study, is continuing to develop the test, which ideally would predict the risk of preeclampsia as early as 3 months into a pregnancy. The company also hopes to use the gene expression maps to predict other complications such as gestational diabetes and premature birth, says McElrath, a paid adviser for Mirvie. “We used preeclampsia as a test case, but there are a whole host of additional things that are possible.”