Loading…

Amniotic fluid volume in normal singleton pregnancies: modeling with quantile regression

Purpose To develop uniform and reliable reference ranges for amniotic fluid volume (AFV) across gestation in normal singleton pregnancies using quantile regression (QR). Methods An analysis of true AFVs determined by dye-dilution techniques or by direct measurement at cesarean delivery in normal sin...

Full description

Saved in:
Bibliographic Details
Published in:Archives of gynecology and obstetrics 2014-05, Vol.289 (5), p.967-972
Main Authors: Sandlin, Adam T., Ounpraseuth, Songthip T., Spencer, Horace J., Sick, Courtney L., Lang, Patrick M., Magann, Everett F.
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Purpose To develop uniform and reliable reference ranges for amniotic fluid volume (AFV) across gestation in normal singleton pregnancies using quantile regression (QR). Methods An analysis of true AFVs determined by dye-dilution techniques or by direct measurement at cesarean delivery in normal singleton pregnancies. AFV centiles were established by QR, a flexible semi-parametric approach of estimating rates of change across the entire distribution of AFV rather than just in the mean as is observed with standard linear regression. Results The study evaluated 379 women with normal singleton pregnancies between 16 and 41 weeks gestation. QR was used to determine the association between AFV and gestational age (GA). A second-order quantile regression model indicated a nonlinear relationship between AFV and gestational age at the upper centile range (≥80th percentile). Conclusion This study defines normative centile charts for true AFVs between 16 and 41 weeks gestation in normal singleton pregnancies using QR. This statistical approach more appropriately reflects true AFV across gestation at each centile of interest (e.g. 5th, 50th, 95th, etc.) as compared to standard linear regression.
ISSN:0932-0067
1432-0711
DOI:10.1007/s00404-013-3087-2