[Comparison of sonographic weight estimation to actual birth weight: a retrospective analysis]

This retrospective study aims at determining the accuracy of sonographic estimation of birth weight based on ultrasound examinations performed at the department of gynaecology of Philipps University at Marburg, Germany.

630 children were born at the university department of gynaecology during the first six months of 1998. 519 babies had been examined sonographically and their birth weight estimated within ten days prior to delivery. 176 (33,91 %) of these examinations were carried out by experienced sonographers complying to the level II standard of performance set out by DEGUM. 343 (66.09 %) of examinations were performed by less experienced junior doctors at the time of the mothers' admission to the delivery room. Two reference tables published by Hansmann und Ferrero were used to estimate birth weight.

The lower and upper quartile of deviation between estimated weight and actual weight came to -200 g and + 200 g using the Hansmann method and -180 g and + 220 g respectively based on the method of Ferrero. It has to be noted that even greater differences occurred: the 10 % least exact estimates deviated by 500 g and more. The difference between estimated and actual weight increased with the length of pregnancy. The fully trained physician (DEGUM II) tended to estimate the birth weight more accurately. The average birth weight estimated sonographically was lower in "DEGUM II-babies" than that of children whose weight was estimated at the point of admission to the delivery room (Mann-Whitney p < 0,0001). Junior doctors showed a tendency to over-estimate the birth weight of babies actually being underweight. The opposite happened with babies who presented a relatively higher birth weight: their projected birth weight was underestimated. On the other hand, the regression line of the sonographic estimations of birth weight performed by the expert (level II DEGUM) fit to the expected bisector of the angle (Passing-Bablock p > 0,05).

When clinical decisions are based on estimated values of birth weight, the possible deviation of this value from the actual weight has to be taken into account. In our study this difference came up to 500 g independent of the level of the examiner's ultrasonographic training.