[The use of Rossavik's mathematical model in determining individual intrauterine growth curves. Our experience].

Objective: To evaluate the performance of Rossavik growth models, derived from II trimester ultrasound measurements, to predict growth in normally growing fetuses.

Design: Comparison between observed measurements after 25 weeks and those predicted by Rossavik growth models determined from the data collected in 2 ultrasound examinations at approximately 16 weeks and 24 weeks.

Setting: Teaching hospital obstetric unit, in Florence.

Subjects: Thirty women who delivered normal term fetuses in our unit, between January 1991 and December 1992.

Main Outcome Measures: Determination of the expected growth curve after 25 weeks from the appropriate growth models, for these fetal parameters: biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC) and femur length (FL). The differences between sonographic measurement of each parameter, at various time points, and predicted values were expressed as a percentage of the predicted values. Head circumference and weight measured in each infant were also compared with the predicted values at term.

Results: Mean percent deviation values were comprised between -0.02% (+/- 2 SD: 3.9%) for BPD and +0.75% (+/- 2 SD: 4.6%) for FL. Pearson's correlation coefficients between predicted and observed fetal parameters, ranged from 0.89 for AC and 0.94 for BPD. For HC and estimated weight at birth, the percent deviations were 0.84% (+/- 2 SD: 5.6%) and 1.23% (+/- 2 SD: 10.5%), respectively. The mean percent deviations for all parameters, were not significantly different from zero.

Conclusions: Our results confirm the accuracy of Rossavik growth model in predicting growth after 25 weeks in normally growing fetuses.