Effect of the pretest probability of intrauterine growth retardation on the predictiveness of sonographic estimated fetal weight in detecting IUGR: a clinical application of Bayes' theorem.
Four hundred and five women with singleton pregnancies and fetal age determination by crown-rump length were classified on the basis of their prenatal clinical findings into four risk categories for intrauterine growth retardation (IUGR), defined as a neonatal weight below the 10th percentile of age-dependent birth weight distribution curve. The incidence of IUGR in these four groups were 3.5% (very low risk), 20.6% (low risk), 49.6% (intermediate risk), and 88.0% (high risk). Severe growth retardation (birth weight less than 2.5th percentile) increased from 0% to 76.0% as the incidence of IUGR increased throughout the risk groups. The effect of these pretest risks on the prediction of severe IUGR by sonographic estimated fetal weight (EFW) was evaluated. The positive predictive value of the test, as well as the probability of having a growth-retarded infant after a normal EFW was obtained were considerably higher when the pretest probability of IUGR increased. In the very low risk group, the probability of severe IUGR was negligible regardless of the EFW. When the EFW was less than 10th percentile of our age-dependent EFW curve, the probability of severe IUGR in the other risk groups was high enough to warrant fetal well-being surveillance and/or timely interruption of gestation as appropriate. However, when the pretest probability was high, the risk of severe IUGR in spite of an EFW within the 10th percentile to 90th percentile remained sufficient to require fetal well-being surveillance as well. The study shows that placing ultrasound results in the context of the pretest risk of IUGR may improve clinical decision making in pregnancies complicated by fetal growth retardation.