Sensitivity and specificity of ultrasonography as a screening tool for gestational diabetes mellitus

J Matern Fetal Neonatal Med. 2012 Aug;25(8):1348-53. doi: 10.3109/14767058.2011.634458. Epub 2011 Nov 25.

Abstract

Objective: To evaluate diagnostic accuracy and efficiency of ultrasound markers of gestational diabetes (GDM) and propose an ultrasound based scoring system suitable for screening (UGDS).

Methods: 110 women with singleton pregnancies and established maternal and/or pregnancy related risk factors for GDM were scanned at/or after 24 weeks gestation followed by administration of a 3 hour 100-gram oral glucose tolerance test (oGTT). A number of ultrasound markers were determined/measured, including fetal adipose subcutaneous tissue, asymmetrical macrosomy, cardiac circumference, cardiac width, and interventricular septum thickness, immature appearance of placenta, intensified breathing movements, polyhydramnios and placental thickness. Each ultrasound GDM marker was assigned one point to create the ultrasound gestational diabetes screening score (UGDS).

Results: All ultrasound GDM markers were positively correlated to the disease P < 0.0001. The strongest independent predictor of GDM was an immature appearance of placenta (RR 40.1 95% CI 5.9-271.0, P < 0.0001). Receiver operator characteristics (ROC) showed an area under the curve of 95.7% confirming good ability of UGDS to discriminate between positive and negative oGTT. We propose a UGDS score of 4 providing sensitivity of 90.9%, and specificity 89.6%.

Conclusions: This study suggests that UGDS is a good predictor of GDM.

Publication types

  • Clinical Trial
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Biomarkers / analysis
  • Diabetes, Gestational / diagnostic imaging*
  • Diabetes, Gestational / epidemiology
  • Diabetes, Gestational / etiology
  • Female
  • Glucose Tolerance Test
  • Humans
  • Mass Screening / methods*
  • Pregnancy
  • Prevalence
  • Reproducibility of Results
  • Research Design
  • Risk Factors
  • Sensitivity and Specificity
  • Ultrasonography, Prenatal* / statistics & numerical data

Substances

  • Biomarkers