Increased interleukin-6 concentrations in cervical secretions are associated with preterm delivery

Abstract

This study sought to determine whether elevated concentrations of interleukin-6 in the cervical and vaginal secretions of patients between 24 and 36 weeks' gestation predicted subsequent preterm delivery and/or identified those preterm deliveries associated with maternal infectious morbidity. A cohort study was undertaken with cervical and vaginal samples collected from 161 consenting patients seen at 3- to 4-week intervals between 24 and 36 weeks. Levels of interleukin-6 were measured by immunoassay. Demographic, obstetric, neonatal, and laboratory data were analyzed by Fisher's exact test, Student t test, or Wilcoxon rank sum test, linear and multiple logistic regression, and receiver-operator characteristic curve analysis. There were 4.2-fold and 3.4-fold increases in maximal cervical and vaginal interleukin-6 concentrations, respectively, among patients with preterm deliveries versus term deliveries. The receiver-operator characteristics curve analysis indicated that a single cervical interleukin-6 value > 250 pg/ml of sample buffer, present between 24 and 36 weeks' gestation, optimally identified patients with subsequent preterm deliveries versus term deliveries (sensitivity 50.0%, 95% confidence interval 33.2% to 66.8%; specificity 85.0%, 95% confidence interval 78.8% to 91.2%; positive predictive value 47.2%, 95% confidence interval 30.9% to 63.5%; negative predictive value 86.4%, 95% confidence interval 80.4% to 92.4%). The optimal vaginal interleukin-6 cutoff value (> 125 pg/ml) proved less sensitive (45.5%, 95% confidence interval 28.5% to 62.4%) but equally specific (86.6%, 95% confidence interval 80.7% to 92.5%). Multiple logistic regression indicated that a cervical interleukin-6 level > 250 pg/ml was an independent predictor of preterm delivery (adjusted odds ratio 4.8, 95% confidence interval 1.7 to 14.3). Cervical interleukin-6 levels did not correlate with cervical change or gestational age at sampling. Among patients delivered preterm there were no differences in the mean white blood cell count on admission or the prevalence of vaginal pathogens, alkaline vaginal pH, chorioamnionitis, or endometritis among patients with cervical interleukin-6 values > 150 or < or = 250 pg/ml. Cervical interleukin-6 is a relatively insensitive, although fairly specific discriminator of patients with subsequent preterm deliveries. Among patients delivered preterm, elevated cervical interleukin-6 values are not apparently associated with maternal infectious morbidity.