Influence of the interval between time points on individual fetal growth curve standards derived from Rossavik models and two ultrasound scans before 26 weeks, menstrual age

We studied the influence of the interval between the two scans used before 26 weeks' menstrual age to generate individual fetal growth curve standards utilizing the Rossavik growth model: P = c(t) kappa + s(t) (model specification functions previously reported). Intervals of 3 weeks to 12 weeks were suitable for predicting the growth of the abdominal and head circumferences and femur diaphysis length in individual fetuses. However, large systematic and random errors were found with intervals less than 5 weeks for three-dimensional parameters such as the head and abdominal cubes and estimated fetal weight. In addition, the data suggest that the systematic errors for these latter parameters may increase with intervals of 10 weeks or more. Overall, optimal individual fetal growth curve standards were best generated from two scans before 26 weeks' menstrual age separated by 5 weeks to 9 weeks.