Salivary estriol as risk assessment for preterm labor: a prospective trial.

Abstract

Increased understanding of human parturition allows for novel approaches to (1) identification of women at increased risk for preterm birth and (2) development and controlled testing of etiology-based strategies to prevent preterm birth. Five hundred forty-two women were enrolled at five study sites in a prospective evaluation of salivary estriol in samples obtained weekly beginning at 22 weeks' gestation (Salest, Biex, Inc., Boulder, Colo.). Estriol concentrations were determined with a well-characterized enzyme-linked immunoassay. Women adjudged at either high risk or low risk for prematurity were evaluated through to delivery. A total of 267 women submitted serial samples that were analyzed; 241 women with singleton pregnancies submitted sufficient samples. Twenty-three women with singleton fetuses went into idiopathic preterm labor (without prior rupture of membranes) and were delivered preterm (mean 35 weeks' gestation); 182 were delivered at term (> or = 37 weeks' gestation). Mean (geometric) estriol concentrations were higher from 24 to 34 weeks in women with singleton pregnancies delivering preterm (p < 0.05). A surge in estriol concentrations occurred approximately 3 weeks before the onset of labor in both women delivering at term and those delivering preterm. This increase occurred approximately 4 weeks earlier in women delivered preterm versus term. Receiver-operator curve analyses showed that exceeding a 2.3 ng/ml saliva estriol level was associated with occurrence of preterm labor (71% sensitivity, 77% specificity, 23% false-positive rate). Detection of an early estriol surge or increased level (> or = 2.3 ng/ml) may be clinically helpful in identifying women at elevated risk for preterm labor and birth, allowing for evaluation of biologically based interventions in controlled trials.