dfwcpml1www

i

og

S

s(t

c3ittdttGbe

tdewc(t

aasa

s

D

dpel

a(am2C

Catrbi

t

gwifi

uosltss

a(i

Poster Session II Diabetes, Labor, Ultrasound-Imaging www.AJOG.org

multiplied by the median fetal weight for the gestational age at deliv-ery, resulting in the GAP predicted birth weight. The GAP predictedweight and the term US weight were compared to the actual birthweight. Absolute and percent birth weight errors were compared us-ing Student’s t-test. Each patient served as her own control.RESULTS: 237 subjects were included. The mean (�SD) number of

ays between US and delivery was 22�8d for the GAP US and 3�1dor the term US. The mean absolute error and absolute percent errorere 256�184g and 7.7�5.6%, respectively, for the GAP method

ompared to 236�169g and 7.1�5.1% for the term US. When theredicted weights were further classified as either over or under-esti-ated, the percent error in the over-estimated weights was statistically

ower with the GAP method (6.1�5.7% vs. 7.8�5.6%, p�0.01). Tabledemonstrates the percentage of correct birth weight predictionsithin a specified error range. The sensitivity of predicting birtheight �4000g was 22% for the GAP method and 28% for term US,ith 97% specificity for both methods.

CONCLUSIONS: The GAP method is as accurate as term US in predict-ng birth weight in patients with GDM.

Table 1. The percentage of correct birth weightpredictions within a specified error range

Birth weightwithin 5%

Birth weightwithin 10%

Birth weightwithin 15%

GAP 39.4% 69.5% 90.1%..........................................................................................................................................................................................

Term 39.4% 73.5% 91.2%..........................................................................................................................................................................................

p 1 0.2 0.5..........................................................................................................................................................................................

275 Does gestational age affect the accuracyf birth weight prediction when using theestation-adjusted projection method?

Rachelle Schwartz1, Barak Rosenn1,ophia Havraniak1, Lisa Simmonds1

1St. Luke’s-Roosevelt Hospital, New York, NYOBJECTIVE: The gestation adjusted projection (GAP) method has beenhown to be a reliable predictor of birth weight in gestational diabetesGDM). The purpose of this study was to determine the ideal gesta-ional age at which to apply this method.

STUDY DESIGN: Women with GDM, a well defined estimated date ofonfinement, an ultrasound estimated fetal weight (EFW) between4-36 6/7 weeks, and delivery at least 7 days after ultrasound werencluded. The birth weight was predicted by calculating a ratio be-ween the EFW and the median fetal weight for the gestational age athe time of the sonogram. This ratio was then multiplied by the me-ian fetal weight for the gestational age at delivery, which resulted inhe GAP derived predicted birth weight. The gestational age at whichhe ultrasound was performed was divided into three groups. TheAP predicted weights in each group were compared to the actualirth weight. The absolute birth weight errors and absolute percentrrors in each group were compared using Student’s t-test.

RESULTS: 485 subjects met inclusion criteria. The mean (�SD) gesta-ional age at ultrasound was 35.2�0.8 weeks and the mean number ofays between ultrasound and delivery was 22�8d. The mean absoluterrors and mean absolute percent errors are shown in the Table. Thereas no difference between groups with respect to absolute error. Per-

ent error was significantly smaller in group 3 compared to group 2p�.04) and group 1 (p�.02). There was no statistical difference be-ween groups with respect to prediction of macrosomia.

CONCLUSIONS: Although using the GAP method at 36 weeks to predictctual birth weight resulted in the lowest absolute percent error, thectual difference in grams between the three groups is not clinicallyignificant. The GAP method can reliably predict birth weight as early

s 34 weeks gestation. a

S116 American Journal of Obstetrics & Gynecology Supplement to JANUARY 2

Table 1. The mean absolute errors and meanabsolute percent errors for the three groups.

Group 134-34 6/7

n�198

Group 235-35 6/7

n�199

Group 336-36 6/7

n�88

Mean Absolute Error 260 � 198g 249 � 178g 214 � 203g..........................................................................................................................................................................................

Mean Absolute Percent Error 8.1 � 6.6% 7.7 � 5.7% 6.2 � 5.6%..........................................................................................................................................................................................

276 Male gender and the risk of oxidativetress related birth defects

April Adams1, Ray O. Bahado-Singh1,evika Maulik2, Michael Kruger1

1Wayne State University–School of Medicine, Detroit, MI, 2UniversityOf Missouri-Kansas City, School Of Medicine, Kansas City, MOOBJECTIVE: Oxidative stress (OS) is an important mechanism for the

evelopment of congenital anomalies (CA). Recent data from multi-le sources suggest an increased susceptibility to OS in fetal males. Wevaluated whether there was an increased risk of CA’s putativelyinked to OS, in male newborns.

STUDY DESIGN: CDC – National Center for Health Statistics birth datafor 31 US states were reviewed between 2004-2006. We excluded chro-mosome anomalies, genetic syndromes and diabetic pregnancies. Ad-justed Odds Ratio (aOR) for major CA linked to OS were calculatedfor male newborns. Primary CA considered were heart and renalanomalies, cleft lip (CL) and adactyly (ADAC). Female newbornswere used as the reference group. Stepwise logistic regression was usedto control for the following confounders: maternal age, race, Hispanicethnicity and parity. P� 0.05 was considered significant.RESULTS: There were a total of 8,026,304 eligible cases after exclusionsbove and elimination of cases with missing data. There were 5,6350.07%) cases with ADAC, 5,487 (0.07%) with CL, 866 (0.01%) renalnd 9,235 (0.12%) heart defect cases. The aOR (95% CI) for CA inales were 1.09 (1.04, 1.13), p�0.001 for heart defects, 1.81 (1.58,

.08), p�0.001 for renal anomalies, 1.24 (1.17, 1.131), p�0.001 forL and 1.32 (1.25, 1.39), p�0.001 for ADAC.

CONCLUSIONS: Male gender appears to significantly increase the risk ofA’s that are etiologically linked to OS, especially in the case of renalnomalies and ADAC. Fetal gender may thus play an important role inhe mechanism of development of common birth defects. Given theecognized importance of OS in the development of a wide range ofirth defects, this information could be important in clinical counsel-

ng and merits further scientific investigation.

277 Are there third trimester ultrasound findingshat could identify late onset gestational diabetes?

Rekuc Emily1, Chaya Ray1, James Airoldi1

1St. Luke’s Hospital, Bethlehem, PAOBJECTIVE: The risk of gestational diabetes mellitus (GDM) increases asestational age advances yet routine screening is recommended at 26-28eeks. Some women are not diagnosed until repeat third trimester testing

s performed. The objective is to assess if there are simple ultrasoundndings that could identify late onset gestational diabetes.

STUDY DESIGN: This is a prospective cohort study. Late pregnancytrasound findings that increase the suspicion of late onset GDM werebtained. The following findings were used to justify repeat diabetescreening: overall large fetus, large abdominal circumference (AC) ateast 2 weeks ahead of gestational age, asymmetry with the AC greaterhan the head circumference by 25 mm, and polyhydramnios. Chiquare and Fisher exact test were used for categorical variables andtudent T-test were used for continuous variables.

RESULTS: Thirty seven women with suspicious ultrasound findings asbove and complete repeat GDM testing were identified. Nine women24%) were diagnosed with late pregnancy GDM. There were no signif-cant differences between the GDM and no GDM group with respect to

symmetry (p�0.64), AC 2-3 weeks ahead (p�1.00), AC 3-4 weeks

011