Predictive value of amniotic fluid index for oligohydramnios in patients with prolonged pregnancies.

The objective of this study was to evaluate the predictive values of the amniotic fluid index for measures of perinatal morbidity and for clinical observations consistent with oligohydramnios. We evaluated positive and negative predictive value of the amniotic fluid index for measures of perinatal morbidity and for clinical observations consistent with oligohydramnios at various cutoff values for amniotic fluid index in a cohort of 449 consecutive postdates patients who had a clinician's observation of amniotic fluid quantity and quality recorded at the time of rupture of membranes. Newborn morbidity was a rare event. Clinical observations consistent with oligohydramnios had significant positive and negative predictive values for some measures of newborn morbidity. The last amniotic fluid index performed during antepartum testing had 95% confidence intervals for relative risks for these measures of newborn morbidity that included unity and therefore were not significant. At a cutoff value of 5.0 cm, the positive predictive value of the amniotic fluid index for clinical observations consistent with oligohydramnios was 50%; the negative predictive value was 85%, with a prevalence of clinical observations consistent with oligohydramnios of 19%. The presence of fetal heart rate decelerations did not significantly improve the positive predictive value of the amniotic fluid index. Higher positive predictive values were obtained at cutoff values of 4 cm and 3 cm with minimal loss in negative predictive value. The amniotic fluid index did not possess significant predictive value for measures of newborn morbidity. Clinical observations consistent with oligohydramnios at the time of rupture of membranes did have predictive value for some of these measures and thus probably are a reflection of the actual amount of fluid present inside the uterus prior to rupture of membranes. The amniotic fluid index is only a fair predictor of clinical observations consistent with oligohydramnios. Thus, a positive test correctly predicted these observations 50% of the time, with 50% false-positive results. Undertaking delivery in the 50% of patients without clinical observations consistent with oligohydramnios may lead to a higher cesarean section rate since these patients do not require induction and are subject to the risk of a failed induction of labor. A negative test correctly predicted observations consistent with normal fluid 85% of the time, with a false-negative rate of 15%. Thus, a negative test was no guarantee that observations consistent with oligohydramnios, and thus newborn morbidity, would not subsequently appear. Frequent testing with multiple modalities and induction of labor when the Bishop score is favorable remain sensible options. Induction of labor in postdates patients with a low amniotic fluid index needs to be evaluated in a yet-to-be-performed prospective randomized control trial before a low amniotic fluid index is assumed to be the sole indicator for induction of labor. More stringent cutoff values for amniotic fluid index may be justified.