485: Early prediction of spontaneous preterm birth

Abstract

Jelle M. Schaaf, Anita Ravelli, Ben Mol, Ameen Abu Hanna Academic Medical Center, Medical Informatics, Amsterdam, Netherlands, Academic Medical Center, Medical Informatics, Amsterdam, Netherlands, Academic Medical Center, Department of Obstetrics and Gynaecology, Amsterdam, Netherlands OBJECTIVE: Many separate risk factors for preterm birth have been identified but the pathogenesis of preterm birth and the prognostic value of their combination is not well understood. A prognostic model for assigning risk of preterm birth to individual women is lacking. Assessment of womens risk is essential in reducing the impact of preterm birth as it is the leading cause of perinatal morbidity and mortality. This study aims to develop a prognostic model for predicting preterm birth after a spontaneous onset of labour. STUDY DESIGN: Data were obtained from the Netherlands Perinatal Registry (PRN). We included all singleton pregnancies 22 weeks of gestation that resulted in a spontaneous onset of birth. Preterm birth was defined as birth before 37 completed weeks of gestation. We used logistic regression methods to develop the prognostic model for predicting preterm birth. We used the Hosmer-Lemeshow C-statistic to assess the models goodness of fit, and the AUC to assess models discriminative abilities. RESULTS: We analyzed data of 1,011,047 women who gave birth between January 1999 and December 2007 in The Netherlands. The incidence of spontaneous preterm birth was 5.6% (n 56,698). Independent risk factors of preterm birth were history of preterm birth, maternal age 35 years, Black or South-Asian ethnicity, low socioeconomic status, living in a deprived area, being nulliparous, getting late into obstetric care and being pregnant of a male infant (table 1). The models Hosmer Lemeshow C-test had p-value 0.0001, and the AUC was 0.64. CONCLUSION: The developed prognostic model predicting preterm birth, applicable around 20 weeks of gestation, had fair discrimination ability and poor goodness of fit. It confirms the known risk factors of preterm birth and identifies new ones particularly being late into obstetric care and living in a deprived area. Some risk factors can potentially be influenced by interventions before and during pregnancy in order to reduce the incidence of preterm birth. Our prediction model will be improved with the addition of data on laboratory and ultrasound results. 486 Non-invasive prediction of intra-amniotic infection and/or inflammation in women with preterm labor: various cytokines in cervicovaginal fluid Jeong Woo Park, Kyo Hoon Park, Sung Youn Lee Seoul National University College of Medicine, Deptartment of Obstetrics and Gynecology, Seoul, Korea, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Department of Obstetrics and Gynecology, Seongnam-si, South Korea OBJECTIVE: To determine the value of cervicovaginal interleukin (IL)1 , IL-6, and IL-8 in predicting intra-amniotic infection and/or inflammation (IAI) in women with preterm labor and intact membranes, and to compare the predictability with the amniotic fluid (AF) white blood cell (WBC) count. STUDY DESIGN: Cervicovaginal secretions were collected immediately before amniocentesis in 86 consecutive women admitted for preterm labor (20 0-34 6 weeks) with singleton pregnancies. AF obtained by amniocentesis was cultured and the WBC count was determined, and the following cytokines were measured in cervicovaginal fluid: IL-1 , IL-6, and IL-8. IAI was defined as a positive AF culture for microorganisms and/or the presence of high levels ( 2.6 ng/mL) of IL-6 in AF. RESULTS: The prevalence of a positive AF culture was 9% (8/86) and the prevalence of intra-amniotic inflammation was 23% (20/86). Receiver operating characteristic (ROC) curves demonstrated that cervicovaginal IL-6 and IL-8, but not IL-1 , had abilities to predict IAI. Cervicivaginal IL-6 had a significantly higher area under the curve (AUC) than cervicovaginal IL-8 (0.85 [95% CI, 0.76-0.92]) vs. 0.70 [95% CI, 0.59-0.79], respectively; P 0.01). However, the AUCs for the cervicovaginal IL-6 and AF WBC were not significantly different. Using ROC curve, the best cut-off value for cervicovaginal IL-6 in predicting IAI was 0.216 ng/mL, with a sensitivity of 75% and a specificity of 86%. CONCLUSION: Cervicovaginal IL-6 is the most important non-invasive marker for predicting IAI in women with preterm labor. Overall, the measurement of the cervicovaginal IL-6 level performed as well as the AF WBC counts for predicting IAI. www.AJOG.org Doppler Assessment, Fetus, Prematurity Poster Session III