Fetal fibronectin detection as a predictor of preterm birth in actual clinical practice.

OBJECTIVE This study was undertaken to determine whether fetal fibronectin determination is more useful for predicting preterm delivery in clinical practice than it has appeared to be in prospective blinded studies. STUDY DESIGN Charts of 151 patients with fetal fibronectin tests performed during 2 years were reviewed. Patients were included if they had symptoms of preterm labor, a singleton pregnancy at 24 to 35 weeks' gestation, intact membranes, and cervical dilatation < or =3 cm. RESULTS Complete data were available for 85 tests. For delivery within 7 days after specimen collection the sensitivity, specificity, positive predictive value, and negative predictive value were 89%, 84%, 40%, and 98%, respectively. The positive predictive value was greater (P <.002) than those reported in three prospective studies evaluating delivery within 7 days in patients with symptoms. Gestational age at delivery and birth weight were lower for patients with positive results (P <. 0001 and P <.006, respectively). Patients with positive results were also treated more with tocolysis, corticosteroid use, and hospitalization than were patients with negative results. For direct comparison with studies of patients with cervical dilatation <3 cm, only 4 patients with cervical dilatation of 3 cm were enrolled. All 4 had negative results of fetal fibronectin testing, and their outcomes therefore did not affect the positive predictive value. CONCLUSION The positive predictive value of fetal fibronectin measured in actual clinical practice was significantly greater for delivery within 7 days than has been reported in blinded prospective studies.