163 Electronic fetal monitoring (EFM) patternsin the second stage of labor: association andtest characteristics for fetal acidemiaAlison G. Cahill1, Anthony Odibo1,Kimberly Roehl1, George Macones1

1Washington University in St. Louis, Departmentof Obstetrics and Gynecology, St. Louis, MOOBJECTIVE: To estimate the association and screening efficacy of EFMpatterns prior to delivery and acidemia using NICHD Categories anddefinitions.STUDY DESIGN: We conducted a 5-year retrospective cohort study ofall singleton, non-anomalous gestations delivered at � 37 weeks fromthe second stage of labor. The primary exposure was 30 minutes ofEFM immediately prior to delivery, interpreted by 2 dedicated obstet-ric research nurses, formally trained and re-trained in EFM interpre-tation, blind to clinical and outcome data. EFM patterns were inter-preted using the NICHD nomenclature, including the 3-Categorysystem and feature definitions. EFM was interpreted in 10-minuteepochs, and then overall. The primary outcome was acidemia, definedas an umbilical cord gas pH �7.10. Relative risks and 95% confidenceintervals, and test characteristics for acidemia were calculated.RESULTS: Of 5,388 patients meeting inclusion criteria, 98.9% (5331)had a pH �7.10 and 1.1% (57) were acidemic. In bivariate analyses ofthe EFM patterns 30 minutes prior to delivery, baseline tachycardia(RR 3.3, 1.7 - 6.3) was the only pattern among NICHD-defined EFMCategories or features associated with acidemia; Categories I- III,minimal variability (RR 0.7, 0.3–1.6), marked variability (RR 2.2, 1.0- 4.6), repetitive late (RR 1.3, 0.5–3.4) and prolonged decelerations(RR 3.1, 1.2– 8.2) all demonstrated no association. Test characteristicsfor each EFM Category or features were equally poor (Table).CONCLUSION: The NICHD 3-tiered Category system, as well as specif-ically defined features such as variability and deceleration types, dem-onstrate weak or no association with acidemia and do not demon-strate clinically useful test properties for EFM interpretation in thesecond stage.

164 Glial fibrillary acidic protein as a serum biomarkerfor the early identification of neonatal periventricularwhite matter injury in preterm neonatesAmanda Stewart1, Jacky Jennings2, Aylin Tekes3,Huisman Thierry3, Frances Northington4, WilliamSavage5, Allen Everett6, Ernest Graham1

1Johns Hopkins University School of Medicine, Gynecology andObstetrics, Division of Maternal-Fetal Medicine, Baltimore, MD, 2JohnHopkins University School of Medicine, Pediatrics, Baltimore, MD, 3JohnsHopkins University School of Medicine, Pediatric Neuro-Radiology,Baltimore, MD, 4Johns Hopkins University School of Medicine, Pediatrics,Baltimore, MD, 5Johns Hopkins University School of Medicine,Pathology-Transfusion Medicine, Baltimore, MD, 6Johns HopkinsUniversity School of Medicine, Pediatric Cardiology, Baltimore, MDOBJECTIVE: Periventricular white matter injury(PWMI) in preterminfants results in cerebral palsy in 60-100% of survivors. Glial fibrillaryacidic protein(GFAP), an abundant white matter protein localized toastrocytes, is normally undetectable in serum and only measureableafter CNS injury. We sought to determine if serum GFAP levels can beused to identify neonates developing PWMI.STUDY DESIGN: This is a case control study of all neonates born at asingle tertiary care hospital from 4/09 - 4/11 at 23- 34 weeks gestationdiagnosed with PWMI on a head ultrasound at 6 weeks of life. Eachcase with PWMI was gestational age matched to a subsequent neonatewith a normal head ultrasound. GFAP was measured in cord blood atthe time of birth, at NICU admission, and on days 1-4 of life. Pairedvariables were compared using paired t-tests and McNemar chisquare. Wilcoxon rank sum tests were used to compare GFAP levels.RESULTS: During this 2 year period, 21 cases with PWMI occurred.Cases and controls did not differ in birthweight (1077 �/� 544 grams,963 �/� 492 grams) or incidence of cesarean delivery (61.9%,66.7%). Cases with PWMI had a significantly increased incidence ofintraventricular hemorrhage (90.5%, 23.8%, p�0.001), 5 min Apgar� 7 (66.7%, 23.8%, p�0.01) and seizures (42.9%, 0%, p�0.004).There was no significant difference in umbilical arterial pH (7.22 �/10.19, 7.27�/� 0.06), or base deficit (4.4 �/� 6.8, 2.6 �/� 1.7). GFAPwas not significantly different in cord blood, at NICU admission, orday 1, but was significantly increased at days 2-4 of life (p�0.002,0.0005, 0.04). The 6 cases with cystic PWMI had significantly elevatedGFAP levels compared to other cases (p�0.03). Logistic regression ofGFAP at day 4 of life showed sensitivity 50%, specificity 78.6%, PPV75% and NPV 55% in identifying PWMI.CONCLUSION: Serum GFAP levels were significantly elevated on day oflife 2-4 in neonates diagnosed with PWMI at 6 weeks. Measurement ofGFAP in the perinatal period provides a rapid quantitative way todiagnose white matter injury, and may also correlate with the scale ofinjury and neurocognitive outcomes.

Poster Session I Clinical Obstetrics, Medical-Surgical-Disease, Neonatology, Physiology-Endocrinology www.AJOG.org

S86 American Journal of Obstetrics & Gynecology Supplement to JANUARY 2012