Prediction of clinical infection in women with preterm labour with intact membranes: a score based on ultrasonographic, clinical and biological markers.

OBJECTIVE: To predict maternal and neonatal clinical infection at admission in women hospitalized for preterm labour (PTL) with intact membranes.

STUDY DESIGN: Prospective study of 371 women hospitalized for preterm labour with intact membranes. The primary outcome was clinical infection, defined by clinical chorioamnionitis at delivery or early-onset neonatal infection.

RESULTS: Clinical infection was identified in 21 cases (5.7%) and was associated with earlier gestational age at admission for PTL, elevated maternal C-reactive protein (CRP) and white blood cell count (WBC), shorter cervical length, and a cervical funnelling on ultrasound. We used ROC curves to determine the cut-off values that minimized the number of false positives and false negatives. The cut-off points chosen were 30 weeks for gestational age at admission, 25 mm for cervical length, 8 mg/l for CRP and 12,000 c/mm(3) for WBC. Each of these variables was assigned a weight on the basis of the adjusted odds ratios in a clinical infection risk score (CIRS). We set a threshold corresponding to a specificity close to 90%, and calculated the positive and negative predictive values and likelihood ratios of each marker and of the CIRS. The CIRS had a sensitivity of 61.9%, while the sensitivity of the other markers ranged from 19.0% to 42.9%. Internal cross-validation was used to estimate the performance of the CIRS in new subjects. The diagnostic values found remained close to the initial values.

CONCLUSION: A clinical infection risk score built from data known at admission for preterm labour helps to identify women and newborns at high risk of clinical infection.