Reproducibility of sonographic assessment of fetal liver length in diabetic pregnancies

Abstract

Assessing fetal liver size might be useful in monitoring the effectiveness of the treatment of diabetes in pregnancy. We aimed to assess the reproducibility of fetal liver-length measurement in pregnant women with diabetes. From 3 April 2006 to 5 December 2006, we assessed intraobserver and interobserver variation of fetal liver-length measurements on 55 sonograms in 44 pregnant women with diabetes, 34 of whom had gestational diabetes. The mean maternal age was 33 years, the mean maternal weight was 92 kg and the mean body mass index (BMI) was 33.7 kg/m(-2). The effect of covariates BMI, gestational age and maternal age on the reproducibility of fetal liver length was assessed by calculating intraobserver SD ratios. We compared liver length with abdominal circumference and gestational age. Nine of 12 sonographers scanned, on average, six women (range, 3-12) as the first sonographer, and all 12 sonographers scanned, on average, four women (range, 1-10) as the second sonographer. The data were analyzed using a hierarchical linear model. Measurement of fetal liver length was reproducible. The intraobserver SD was 3.06 (95% CI, 2.68-3.59) mm; the interobserver SD was 2.17 (95% CI, 0.59-4.83) mm; the intraobserver correlation was 0.77 (95% CI, 0.63-0.87), and the interobserver correlation was 0.84 (95% CI, 0.51-0.99). The covariate effects were minimal, the SD for a 1-unit increase in the covariate was 1.06 for gestational age, 0.98 for BMI, and 0.97 for maternal age. Measurement of fetal liver length in the diabetic pregnancy is reproducible. It is worthy of further investigation as a parameter for monitoring the effectiveness of treatment of the diabetic pregnancy.