Two screening tests often used to try to predict which pregnant women are likely to deliver prematurely aren’t effective in low–risk women, according to a national collaborative study of more than 10,000 women, led by clinician–researchers at University of Utah Health Sciences and Intermountain Healthcare.

Researchers found that neither transvaginal cervical measurement or fetal fibronectin tests, used separately or together, adequately predicts preterm birth.

The findings were published in March 14 issue of the Journal of the American Medical Association.

It has become common to use these screening tests to try to predict expectant mothers who are at high risk for preterm birth, which is hard to predict until it begins. Efforts to stop it once labor has started are largely ineffective, said the study’s lead author Sean Esplin, MD, Maternal–Fetal Medicine specialist at Intermountain Medical Center in Salt Lake City and professor of Obstetrics/Gynecology at the University of Utah School of Medicine.

A baby born at 24 weeks, for example, weighs just over a pound and has only a 70 percent chance of survival. The risk of having a long–term complication related to being born premature is about 50 percent at that point.

If cervical thinning is spotted soon enough, however, progesterone therapy can be used to intervene and prevent preterm birth. That’s led some experts to measure transvaginal cervical length by doing ultrasounds. A length of less than 25 mm is considered short and thus risky.

Because of the human and financial toll of preterm births, some experts have advocated screening all pregnant women that way.

The other test measures fetal fibronectin, an extra–cellular matrix protein that acts like glue between the membrane and uterus lining. As labor and delivery approach, the quantity of fibronectin that leaks from the cervix rises, so swabbing to test vaginal secretions for the protein is sometimes done in hopes of predicting and preventing preterm birth.

To see the effectiveness of those tests — or combining both of them — in predicting preterm birth, researchers at eight clinical centers including, the University of Utah/Intermountain Healthcare, Columbia University, the University of Indiana, Northwestern University, the University of Pennsylvania, Ohio State University, the University of Pittsburgh and the University of California Irvine tested the women at three points during their pregnancies to see whether the results predicted which of them would turn out to deliver prematurely.

The tests were conducted, on average, at around 12 weeks, 19 weeks, and 28 weeks of gestation. All of the women were “nulliparous,” meaning they hadn’t previously given birth, so there was no history of a preterm birth or identifiable risk factors other than being pregnant. The study didn’t include women who’d miscarried before 20 weeks gestation or who terminated a previous pregnancy.

“What we found is that neither of these tests is very accurate,” said Esplin. They identify a very small portion of women who are going to have a preterm birth. Of those who have a short cervix, only a portion of them go on to have a preterm delivery.”

Now, researchers are looking at other marker combinations to see if they can identify risk factors of adverse pregnancy outcomes, including preterm birth, focusing on protein markers in the blood and social risk factors like age, nutrition and socio-economic factors. They hope to identify early the women at highest risk for preterm birth in order to have time to prevent their preterm births in the future.