Fetal tibia length assessment by prenatal ultrasound: Part I. Using absolute residuals for constructing age-related reference percentiles

Background: Although a variety of fetal short limb syndromes involve tibia growth, no chart of fetal tibia length is available in Taiwan. The purpose of this study was to attempt to construct a chart in Taiwan of normal fetal tibia length (TL) at various gestational periods, based on a pure cross-sectional design without the common weaknesses found in previous literature. Materials and Methods: Normal singleton fetuses of between 17 and 40 weeks of gestation were enrolled in this study at the Antenatal Ultrasound Unit of National Cheng-Kung University Hospital. Only one scan for each subject was included. We analyzed the data by polynomial regression analysis and searched for the best-fitting equations in predicting fetal growth. Altman's model, using absolute residuals for age-related reference percentiles, was applied to calculate the standard deviation of each gestational age (GA) before constructing charts of fetal growth percentiles. All fetuses were followed-up to parturition to ensure they were normal. Results: In total, 2,215 fetuses of normal growth were included in this cross-sectional study. The best-fitting equation in predicting fetal TL using GA is TL = 0.30979 x GA - 2.21881 x 10^-3 x GA² -2.57671 (R = 0.94546, p < 0.0001). The best-fitting equation in predicting fetal GA using TL is GA = - 16.11136 x TL + 4.52078 x TL² - 0.30975 x TL³ + 37.66178 (R = 0.95438, p < 0.0001). Tables for a TL growth percentile chart are generated as a reference for prenatal examination. Conclusion: With strict criteria of study design and appropriate statistical modeling, we believe that our fetal TL values are more accurate for assessing the growth of fetal long bones in Taiwan. Their clinical use should enable the detection of fetal short limb syndromes.