Objective To evaluate the predictive values of fetal fibronectin, bacterial vaginosis, endotoxin and cervical length for preterm birth (< 35 and < 37 weeks) and neonatal morbidity in twin pregnancies.
Participants One–hundred and twenty–one women with twin pregnancies recruited into a prospective longitudinal study at three antenatal clinics in the southwest of Sweden.
Methods Cervical or vaginal fluid was sampled and determined for fetal fibronectin ( 0.05 μg/mL was used as cutoff), endotoxin ( 100 pg/mL) and bacterial vaginosis (presence of clue cells) at two week intervals from 24 to 34 weeks of gestation. The cervical length was measured with transvaginal sonography at the same time intervals.
Main outcome measures Occurrence of preterm birth (< 35 and < 37 weeks of gestation) and neonatal morbidity.
Results All positive fetal fibronectin samples obtained at screening between 24 and 34 weeks predicted birth < 35 weeks (RR 18.0; 95% CI2.2–145.9). A positive fetal fibronectin at 28 weeks of gestation predicted delivery < 35 weeks (RR 6.3; 95% CI 2.6–15.1) with a sensitivity, specificity, positive and negative predictive value of 50.0, 92.0, 62.5 and 87.3%, respectively. An independent association between fetal fibronectin at 28 weeks and preterm birth (< 35 weeks) was verified with logistic regression (P= 0.03). A positive fetal fibronectin at 28 weeks of gestation predicted neonatal morbidity (RR 5.1; 95% CI 2.4–11.0) and a longer period of care at the neonatal intensive care unit. The predictive power of cervical sonography was generally low but cervical length (cutoff 33 mm) measured at 28 weeks of gestation was significantly associated with birth < 37 weeks (RR 2.2; 95% CI 1.1–4.2). The presence of endotoxin correlated to bacterial vaginosis, but these tests were not significantly related to preterm birth or neonatal morbidity.
Conclusions Fetal fibronectin predicted preterm birth and neonatal morbidity in twin pregnancies. The predictive value of cervical length determinations was low. Endotoxin and bacterial vaginosis had no predictive power for preterm delivery in this study.