In summary, a definite association has been demonstrated between preterm labor and genital tract infection. Conclusions regarding the true benefits of antibiotics as adjunctive therapy in treatment of preterm labor are inconsistent. Whereas some of the studies were able to demonstrate significant prolongation of pregnancy, no consistent reduction in either maternal or neonatal morbidity has been demonstrated. However, because the actual incidental morbidity rate is low in the populations studied, the power of this finding is also low. The potential risks for using antimicrobials has yet to be adequately addressed. It has been shown that bacterial resistance can develop when antibiotics are used without specific aim or when a specific bacteria is undertreated. It has been recently shown that prenatal and intrapartum antibiotic use is associated with an increased risk for antibiotic resistant neonatal sepsis if infection occurs. Because of these reasons, we discourage the administration of antibiotic treatment to women in preterm labor for the purpose of pregnancy prolongations. Treatment should be directed towards those with specific indications for treatment (e.g., intrapartum, group B streptococci prophylaxis, urinary tract infection, etc). The primary flaw in these many evaluations of preterm labor is the true incidence of preterm birth. The clinical diagnosis of preterm labor is a difficult one. Approximately one-half of those individuals with preterm contractions will not deliver until term. So, the use of antibiotics for all women in idiopathic preterm labor is destined to treat many women who are unlikely to benefit. If we were able to truly identify those who were in "true" labor, perhaps we could be more selective in determining who may benefit from antibiotics. Biochemical markers such as onco-fetal fibronectin could well-be a helpful marker. Goldberg et al evaluated FFN in vaginal and cervical secretions while attempting to better-predict who would have upper genital tract infection. In this large, multicenter trial, patients were tested for FFN every 2 weeks from 23 to 30 weeks gestation. In those patients who proceeded to deliver before 32 weeks gestation, increased levels of cervical FFN (> 50 ng/ml) were identified in approximately one-quarter. Fetal fibronectin was positive in 4% of their samples and was found to be twice as likely in one with bacterial vaginosis. They showed that the presence of increased FFN was associated with upper genital tract infection (clinical and histologic chorioamnionitis) as a main reason for preterm labor and delivery (increased risk 16-20-fold). Those with increased FFN levels were also shown to have an increased incidence of neonatal sepsis as well. Peaceman et al used FFN to attempt to identify those at risk for preterm delivery among women with contractions between 24 and 34 6/7 weeks gestation. Those with negative FFN were less likely to deliver within 7 days of the test. The negative predictive value was 99.7%, suggesting that this test may be helpful in identifying women who would not benefit from antibiotic treatment. However, if in the absence of prospective clinical trials demonstrating the efficacy of this approach, we discourage the use of FFN screening for this indication.