To evaluate the presence of fetal fibronectin in the cervix and vagina as a screening test for spontaneous preterm birth.
Two thousand nine hundred twenty-nine women at ten centers were routinely screened every 2 weeks from 22–24 to 30 weeks for cervical and vaginal fetal fibronectin. A positive test was defined as a value equal to or greater than 50 ng/mL. The relation between a positive test at four gestational ages and spotaneous pretern birth at various intervals after the test was determined.
In each testing period 3–4% of the fetal fibronectin tests were positive. The correlation between cervical and vaginal fetal firbonectin at the same visit was always approximately 0.7 (P < .001), and that between cervical or vaginal fetal fibronectin in consecutive visits was between 0.17 and 0.25 (P < .001). The sensitivity of fetal fibronectin at 22–24 weeks to predict spontaneous preterm birth at less than 28 weeks was 0.63, and the relative risk for a positive versus negative test was 59. The specificity was always 96–98%, whereas the positive predictive value rose from 13% to 36% as the upper limit of the definition of preterm birth was increased from less than 28 to less than 37 weeks. The relative risk for spontaneous preterm birth after a positive fetal fibronectin test compared with a negative fetal fibronectin test varied substantially by testing period and by the definition of spontaneous preterm birth, but always remained greater than 4 and statistically significant.
A positive cervical or vaginal fetal fibronectin test at 22–24 weeks predicted more than half of the spontaneous preterm births at less than 28 weeks (sensitivity 0.63). As the definition of spontaneous preterm birth was extended to include later gestational ages or when the fetal fibronectin test was performed later in pregnancy, the level of association between a positive fetal fibronectin test and spontaneous preterm birth, while remaining highly significant, tended to decrease. Although fetal fibronectin is an excellent test for predicting spontaneous preterm birth, we present no evidence that the use of this test will result in a reduction in spontaneous preterm birth.