Receiver operating characteristic curves of ultrasonographic estimates of fetal weight for prediction of fetal growth restriction in prolonged pregnancies

Citation
Cp. O'Reilly-green et My. Divon, Receiver operating characteristic curves of ultrasonographic estimates of fetal weight for prediction of fetal growth restriction in prolonged pregnancies, AM J OBST G, 181(5), 1999, pp. 1133-1138
Citations number
25
Categorie Soggetti
Reproductive Medicine","da verificare
Journal title
AMERICAN JOURNAL OF OBSTETRICS AND GYNECOLOGY
ISSN journal
00029378 → ACNP
Volume
181
Issue
5
Year of publication
1999
Part
1
Pages
1133 - 1138
Database
ISI
SICI code
0002-9378(199911)181:5<1133:ROCCOU>2.0.ZU;2-T
Abstract
OBJECTIVE: Recent studies have documented increased perinatal morbidity and mortality rates in the growth-restricted postterm fetus. Our purpose was t o evaluate the receiver operating characteristic curve of ultrasonographica lly estimated fetal weight as a predictor of fetal growth restriction in pr olonged pregnancies. STUDY DESIGN: Fetal weight was estimated ultrasonographically within 9 days of delivery (mode 1 day) in members of a cohort of 410 patients with prolo nged pregnancies (>41 weeks). Estimated fetal weights were compared with bi rth weights in receiver operating characteristic curve analysis. RESULTS: The areas under the receiver operating characteristic curves for p redicting birth weights <10th percentile (3125 g in this population) and <5 th percentile (2930 g in this population) were 0.89 and 0.96, respectively. Both areas were significantly different from an area indicating a useless test. The estimated fetal weight values corresponding to the inflection poi nts for the receiver operating characteristic curves predicting birth weigh ts <10th percentile and <5th percentile were 3370 and 3200 g, respectively. With estimated fetal weight at less than these test cutoff values, the rel ative risks for a fetus to have a birth weight <10th percentile or <5th per centile were 14.6 (95% confidence interval, 6.25-33.8) and 89.8 (95% confid ence interval, 12.1-665), respectively. Analysis of the receiver operating characteristic curves resulted in improved test characteristics relative to using the actual 10th and 5th birth weight percentiles as cutoff values fo r estimated fetal weight (relative risk of 14.6 vs 9.5 and 89.8 vs 26.0, re spectively). CONCLUSIONS: Ultrasonographic estimation of fetal weight is a useful test f or predicting fetal growth restriction in prolonged pregnancies. Future stu dies should evaluate whether intervention on the basis of this identificati on results in improved perinatal outcome.