Holistic Screening for Preeclampsia in First Trimester Bests Clinical Factors Alone

Screening for preeclampsia in the first trimester with the Fetal Medicine Foundation (FMF)’s model—which considers not only clinical characteristics but also additional factors like biomarkers and ultrasound results—is an effective early means of predicting who will develop the condition, according to the prospective PREDICTION study.

Developed in the United Kingdom, the FMF’s first-trimester screening test for preeclampsia detected nearly two-thirds of preterm cases, with half the false-positive rate seen with the American College of Obstetricians and Gynecologists (ACOG) criteria, which are based on risk factors like comorbidities, family history, body mass index, and age, as well as maternal medical history.

However, the role of the FMF algorithm continues to be debated, said senior author Emmanuel Bujold, MD (CHU De Québec-Université Laval Research Center, Canada). To help overcome some of the wariness towards tool, and to see if it might be a good fit for everyday practice in North America, the PREDICTION researchers “decided just to evaluate whether it was working or not.”

Their trial, published earlier this month in Hypertension, enrolled exclusively women who’d not yet given birth and thus lacked a maternal medicine history to inform screening—a key group of clinical interest.

Speaking with TCTMD, Bujold said that their new findings aren’t surprising, given what was known from prior studies. Still, “we were extremely happy, because now we know that it’s working, so women will have the opportunity to have the test,” he said. “Now we’re looking more at: how we can implement it? How can we make it accessible to all women?”

A necessary quality of a screening tool is that its results inspire patients’ trust, something that’s especially important in pregnancy, a time when many women don’t want to take extra medications that might add risk, Bujold pointed out. With other approaches, “the problem is the very high false-positive rate,” he said. Recent research done in Ontario has hinted that “since women do not trust the screening, because the false-positive rate is way too high, most do not take the aspirin.”

Laura A. Magee, MD (Addison House, Guy’s Campus, London, England), whose own work has also explored the FMF algorithm and other approaches to preeclampsia screening, commented on the PREDICTION results for TCTMD. She, too, said the findings are “consistent with the literature.”

“What is fabulous about it is that this has been done by a group not involved in [the FMF algorithm’s development], so this is a true external validation of the test in real-world settings across multiple sites. All of those things are important in building our confidence in the fact that the test should work for us,” she noted.

Magee said she hopes that this helps to calm debates over whether the FMF model is something worth doing. “We already screen women for preeclampsia in practice. It’s not like: should we start screening? What we really need to be asking ourselves is, why are we sticking with an inferior form of screening [and] why aren’t we using a better form of screening?” Implementation research, though more difficult and perhaps less exciting, is “critically important,” she emphasized. “If it doesn’t come off the shelf and if you don’t make it work for your setting, for your patients, it in many ways doesn’t matter that [the earlier-stage studies] happened.”

What we really need to be asking ourselves is, why are we sticking with an inferior form of screening [and] why aren’t we using a better form of screening? Laura A. Magee

Lead author Paul Guerby, MD, PhD (CHU De Québec-Université Laval Research Center), and colleagues analyzed data for 7,325 nulliparous women who were enrolled across five hospitals between 11 and 14 weeks of gestation and followed until delivery.

At their first recruitment visit, the researchers collected information on maternal age, ethnicity, smoking status, method of conception, and chronic diseases (including hypertension, diabetes, and antiphospholipid syndrome) as well as body mass index and mean arterial blood pressure. The women also gave blood samples and underwent fetal ultrasound. Then, between 3 and 6 months after the expected delivery date, the women’s medical files were examined to obtain gestational age at delivery, birth weight, Apgar score, and details of adverse perinatal outcomes such as gestational hypertension and preeclampsia.

Among them, 65 (0.9%) developed the primary endpoint of preterm preeclampsia with delivery before 37 weeks of gestation and 22 (0.3%) developed early-onset preeclampsia, where the women delivered before 34 weeks.

With the FMF algorithm, the ideal cutoff was ≥ 1 in 110 for preterm preeclampsia, which had a sensitivity of 63.1% and specificity of 15.8%. By comparison, with the ACOG model, sensitivity was 61.5% and specificity was 34.3%. Area under the receiver operating characteristics curve (AUC) for the FMF and ACOG approaches were 0.79 versus 0.64, respectively (P < 0.001).

Using that same cutoff, the FMF algorithm had 77.3% sensitivity and 16.0% specificity. With the ACOG model, those values were 59.1% and 34.5%. The AUC again was greater with the FMF versus ACOG screening (0.89 vs 0.62; P < 0.001).

“The impact of our study on practice suggests that first-trimester screening based on a combination of biophysical, biochemical, and ultrasound variables should be favored over current screening based on clinical risk factors,” the investigators conclude. “As aspirin for the prevention of preterm preeclampsia is most effective when initiated before the 16th week of pregnancy, and as nuchal translucency measurement for the screening of aneuploidies is routinely performed in many settings, it would, therefore, be possible to progressively offer this screening in many North American settings.”

For decades, there’s been interest in the potential for aspirin to prevent preterm preeclampsia, Guerby et al note. “Several national societies now recommend initiating aspirin before the 16th week of gestation in pregnant women identified as high risk for preeclampsia based on the presence of specific single or multiple maternal demographic or medical historical risk factors as part of published checklists. However, this checklist approach is associated with a limited sensitivity or specificity.”

The more-holistic FMF algorithm could help aid prevention efforts, the researchers suggest. Using the FMF model, they estimated the number needed to treat (NNT) with aspirin to prevent one case of preterm preeclampsia is 46. By comparison, the ACOG model has an NNT of 101.

Bujold and Magee both noted that the FMF algorithm is already gaining traction in some areas of the world. For instance, the tool is embedded in the astraia software and available on the FMF website. “There is more and more interest in integrating the approach into clinical practice, and that’s really followed, in particular, the ASPRE trial,” said Magee.

The 2017 ASPRE trial, which showed the benefits of low-dose aspirin in women at high risk for preterm preeclampsia, incorporated the FMF tool as part of its design. The study was done at centers in the United Kingdom, Spain, Italy, Belgium, Greece, and Israel. “From what I’ve heard, many of those centers now keep using it,” said Bujold.

Following their own experience with the 7,300-patient PREDICTION study, he said, the researchers are “trying to repeat a similarly sized study where [after FMF screening] we are giving the aspirin where women are at high risk. We hope to see a decrease of preeclampsia [compared with] a historical cohort.” A goal going forward is to ensure the screening is accessible for all women, even those in lower-income or geographically isolated areas, Bujold added.

Among the barriers to uptake, said Magee, are that the FMF model requires a few extra minutes to do ultrasound and measurement of placental growth factor has an up-front testing cost. “But preeclampsia care costs a lot of money,” she stressed, adding that if the FMF approach to screening were implemented on a wide scale, it would ultimately save money.