Do semiquantitative amniotic fluid indexes reflect actual volume?

Objective: Our objective was to determine how well semiquantitative ultrasonographic measures of amniotic fluid, i.e., maximal amniotic fluid vertical pocket and amniotic fluid index, reflect actual amniotic fluid volumes in 50 near-term patients whose amniotic fluid volume estimates were normal by visual inspection.
Study Design: Before amniocentesis for fetal lung maturity, each patient had visual amniotic fluid volume estimates, maximal amniotic fluid vertical pocket, and amniotic fluid index performed by the same examiner, and then each received intraamniotic injection of a 10% paraaminohippurate solution. Amniotic fluid volume was quantitated by spectrophotometric assay of paraaminohippurate concentration. Oligohydramnios and polyhydramnios were defined as < 300 and > 2000 ml, respectively.
Results: Quantitative amniotic fluid volume was positively related to both amniotic fluid index and maximal amniotic fluid vertical pocket (r = 0.75 and 0.60, respectively). True-positive rates for oligohydramnios (amniotic fluid index < 5 cm or maximal amniotic fluid vertical pocket < 2 cm) were 100% and 0%, respectively; false-positive rates with either method were 0%. True-positive rates for polyhydramnios (amniotic fluid index > 20 cm and maximal amniotic fluid vertical pocket > 8 cm) were 0%; false-positive rates were 16% and 24%, respectively.
Conclusions: Amniotic fluid index appears to be slightly better than maximal amniotic fluid vertical pocket for reflecting actual amniotic fluid volume. Both indirect methods tend to overestimate actual amniotic fluid volume at the upper end of its extremes.