This Scientific Short is part of a series on Academy Guidance Documents. View the full document click.
Preterm birth (PTB), the delivery of a baby prior to 37 weeks gestation, is the second most common cause of infant death in the United States and surviving infants are at increased risk of lifelong complications (1). The cost of medical care from delivery through adulthood, as well as diminished workplace productivity in individuals with PTB-associated developmental disabilities is estimated at $26 billion annually (2,3). Clearly, this represents a significant public health issue yet despite substantial effort, the preterm birth rate remains essentially unchanged at 10% (3,4).
Identifying women who will deliver prematurely is difficult as the clinical signs and symptoms are very nonspecific. Complicating matters is the fact that very few women with preterm labor symptoms deliver within seven days. At U.S. medical centers, the typical rate of such deliveries is 2-3%. In other words, 97-98% of women with symptoms of preterm labor will not deliver within one week (5,6).
Given the negative health consequences of PTB, there is a clear need for diagnostic tools that distinguish the small number of women who will deliver preterm from the much larger pool of women who will not. A diagnostic test that identifies the “needle in the haystack” – the 2-3% of women who will deliver, with a high degree of certainty must have a high positive predictive value (PPV: people with a positive test result who have the disease or condition).
Unfortunately, currently available tests for the evaluation of preterm labor have a low PPV and a high negative predictive value (NPV: people with a negative test result who do not have the disease or condition) and they provide limited additional information. Remember that without doing any clinical laboratory testing, prevalence alone provides a NPV of 97-98%. Fetal fibronectin (fFN), the most widely used clinical laboratory test for the evaluation of PTB, consistently demonstrates a NPV of 99% for PTB within seven days. A negative result provides reassurance that delivery within the next week is very unlikely but this doesn’t substantially improve patient management as the same reassurance is provided by prevalence alone. Furthermore, the PPV of fFN for PTB within one week is consistently 15-20% across different patient populations (7–9). As <20% of women with a positive test deliver within one week, this doesn’t substantially change the pre-test probability but may increase stress and prompt unnecessary further evaluation (10). An alternative PTB test, placental alpha microglobulin-1 (PAMG-1), consistently exhibits a higher PPV than fFN but it still falls short of the threshold needed to add value in a population with a low prevalence (11). Limiting testing to those at high risk (and thereby increasing prevalence in the tested population) through ultrasound screening of cervical length increases PPV but this is not routinely performed at most U.S. medical centers (12).
A clinical laboratory test that identifies the minority of women who will deliver prematurely with a high degree of certainty would have immense value. Ultimately, the search continues, as none of the available markers meet the “high PPV” requirement.