Neonatal Outcomes after Maternal Biomarker-Guided Preterm Birth Intervention: The AVERT PRETERM Trial

Matthew Hoffman, C. Kitto, Zugui Zhang, Jing Shi, Michael Walker, Babak Shahbaba, Kelly Ruhstaller

This paper presents the results of the AVERT PRETERM trial, a study investigating whether a proteomic blood test can identify "low-risk" pregnancies headed for preterm birth and if targeted intervention can improve the eventual health of the newborns.

TL;DR

The AVERT PRETERM trial demonstrates that a proteomic biomarker test can identify high-risk pregnancies in women traditionally considered "low-risk." By targeting these individuals with a combination of aspirin, progesterone, and care management, researchers achieved a 21% reduction in neonatal hospital stays and an 18% reduction in severe neonatal morbidity.

What Problem Does This Paper Address?

Preterm birth (PTB) remains the leading cause of perinatal mortality and long-term developmental delays. While medical interventions like vaginal progesterone and low-dose aspirin exist, the clinical challenge lies in identification: who actually needs them?

Currently, doctors rely on a patient's history of previous preterm births or ultrasound measurements of cervical length. However, the majority of preterm births occur in women with no prior history and "normal" cervical lengths. This leaves a massive "unidentified" at-risk population. The authors sought to determine if a molecular-level screening tool could find these hidden risks and if early intervention based on those results would actually lead to healthier babies.

Methods

The study utilized a "test-and-treat" strategy within a prospective trial arm and compared outcomes against a large historical cohort.

  • The Screening Tool: Researchers used the PreTRM test, a validated serum proteomic biomarker that measures the ratio of two proteins—insulin-like growth factor binding protein 4 (IGFBP4) and sex hormone-binding globulin (SHBG)—between 19 and 20 weeks of gestation.
  • The Intervention: Participants identified as "higher-risk" (defined as a $\ge$16.0% predicted risk of spontaneous PTB) were offered a multimodal intervention:
  • Low-dose aspirin (81 mg daily).
  • Vaginal progesterone (200 mg daily).
  • Care management (twice-weekly nursing contacts to monitor symptoms and compliance).
  • Statistical Approach: To account for the fact that most births are healthy and full-term, the researchers focused their primary analysis on the "extremes"—the highest 8.5% quantile of neonatal stays and morbidity—to ensure the signal of the intervention wasn't diluted by normal births.

Datasets

  • Prospective Arm: 1,463 pregnant women screened at ChristianaCare (Newark, DE) between 2018 and 2020.
  • Historical Arm: 10,000 consecutive pregnancies from the same institution immediately prior to the study that met the same eligibility criteria (singleton, no prior PTB, no known anomalies).
  • Diversity: The study site served a racially diverse population; notably, Black women represented 26.5% of both study arms, which is roughly double the U.S. national average.

Key Insights and Ideas

  • Molecular vs. Clinical Risk: The study confirms that proteomic biomarkers can "see" risks that traditional clinical history and ultrasounds miss. 34.7% of the "low-risk" population actually screened as high-risk.
  • The Power of Multimodal Care: The intervention wasn't just a pill; the inclusion of "care management" (nurse calls) likely played a critical role in patient compliance and early symptom detection.
  • Impact on the "Sickest" Neonates: The most significant impact was seen in very early preterm births. While the overall average gestational age didn't shift dramatically, there was a significant reduction in births occurring before 32 weeks. For these babies, the intervention extended gestation by an average of 2.5 weeks.

Key Results

The study found statistically significant improvements in neonatal outcomes for the prospective arm compared to the historical controls:

  • Neonatal Hospital Length of Stay (NNLOS): Decreased by 21% on average (Hazard Ratio 1.35, $P=0.01$).
  • Neonatal Morbidity Index (NMI): Severe morbidity (NMI $\ge$ 3) was reduced by 18% (Odds Ratio 0.81, $P=0.031$).
  • Very Preterm Births: For pregnancies delivering before 32 weeks, the risk was nearly halved (Hazard Ratio 0.52, $P=0.009$).
  • NICU Stays: Among babies born before 32 weeks, the mean hospital stay dropped from 97.2 days (historical) to 68.5 days (prospective).

Limitations and Future Work

The authors acknowledge several limitations:

  • Historical Control: This was not a randomized controlled trial (RCT). Comparing a prospective group to a historical one introduces potential "secular" bias (changes in general medical care over time).
  • COVID-19 Impact: The trial was halted early due to the pandemic. The researchers had to adjust their statistical analysis plan (while blinded) to exclude subjects impacted by the local spread of SARS-CoV-2 to avoid confounding data.
  • Demographic Shifts: The prospective arm was older and had higher rates of obesity and hypertension than the historical arm. While the authors adjusted for these in their models, these factors typically increase PTB risk, suggesting the intervention's success might be even more robust than the numbers show.

Future Work: Prospective, randomized trials where results are blinded to some participants are already underway to provide more definitive evidence for universal proteomic screening.

Summary updated

Abstract

The AVERT PRETERM trial (NCT03151330) evaluated whether screening clinically low-risk pregnancies with a validated maternal blood biomarker test for spontaneous preterm birth (sPTB) risk, followed by preventive treatments for those screening positive, would improve neonatal outcomes compared to a clinically low-risk historical population that had received the usual care. Prospective arm participants with singleton non-anomalous pregnancies and no PTB history were tested for sPTB risk at 191/7–206/7 weeks’ gestation and followed up with after neonatal discharge. Screen-positive individuals (≥16% sPTB risk) were offered vaginal progesterone (200 mg) and aspirin (81 mg) daily, with twice-weekly nurse phone calls. Co-primary outcomes were neonatal morbidity and mortality, measured using a validated composite index (NMI), and neonatal hospital length of stay (NNLOS). Endpoints were assessed using survival analysis and logistic regression in a modified intent-to-treat population comprising screen-negative individuals and screen-positive individuals accepting treatment. Of 1460 eligible participants, 34.7% screened positive; of these, 56.4% accepted interventions and 43.6% declined. Compared to historical controls, prospective arm neonates comprising mothers accepting treatment had lower NMI scores (odds ratio 0.81, 95% CI, 0.67–0.98, p = 0.03) and an 18% reduction in severe morbidity. NNLOS was shorter (hazard ratio 0.73, 95% CI, 0.58–0.92, p = 0.01), with a 21% mean stay decrease among neonates having the longest stays. Sensitivity analyses in the entire intent-to-treat population supported these findings. These results suggest that biomarker sPTB risk stratification and preventive interventions can ameliorate PTB complications in singleton, often nulliparous, pregnancies historically deemed low risk.
BibTeX
@inproceedings{hoffman2024neonatal,
  author = {Matthew Hoffman and C. Kitto and Zugui Zhang and Jing Shi and Michael Walker and Babak Shahbaba and Kelly Ruhstaller},
  title = {{Neonatal Outcomes after Maternal Biomarker-Guided Preterm Birth Intervention: The AVERT PRETERM Trial}},
  booktitle = {Diagnostics},
  year = {2024},
  doi = {10.3390/diagnostics14141462},
}

Loading related content...

Figure