OBJECTIVE: The purpose of this study was to evaluate the association of vaginal fetal fibronectin expression to risk of preterm delivery and delivery of very-low-birth-weight infants. STUDY DESIGN: Vaginal secretions were obtained from women between 22 and 35 weeks' pregnant with minimal cervical dilatation (≤2 cm) and threatened preterm delivery. The secretions were analyzed for the presence of fetal fibronectin. Other clinical information including cervical dimensions, uterine activity, serum C-reactive protein concentration, vaginal pH, evidence of vaginal or systemic infection, and vaginal bleeding were also obtained. RESULTS: Of the 112 patients recruited, 40 (35.7%) were delivered prematurely (<37 weeks). For prediction of preterm delivery, the fetal fibronectin test result had a sensitivity, specificity, and positive and negative predictive values of 67.5, 90.3, 79.4, and 83.3%, respectively (odds ratio 19.3, p < 0.0001). Women with a positive fetal fibronectin test had a nearly 13-fold increased probability of being delivered of an infant weighing <1500 gm than did women with a negative fetal fibronectin test (32.4% vs 2.5%, p < 0.0001). Categoric analysis and multiple logistic regression demonstrated that fetal fibronectin was an independent risk factor for prediction of preterm delivery and birth weight <1500 gm. CONCLUSION: Vaginal fetal fibronectin expression is an independent risk factor for preterm delivery and birth of very-low-birth-weight infants in symptomatic women with intact amniotic membranes and minimal cervical dilatation. (AM J OBSTET GYNECOL 1996;174:971-4.)