Does the formula suggested by Intergrowth-21st project yield the best accuracy in fetal weight estimation?

Issue:

Fetal growth standards are usually created and used on the basis of the local population. The most accurate, best nationaly suitable formula is usually being selected and adjusted into the clinical practise. Nevertheless a widely known and extensively suggested Intergrowth-21^stproject, which has been conducted in 8 world regeons, offers a new formula for sonographic fetal weight estimation as well as refence centiles. It promotes the idea of international, multicultural society and it's requirement for a unified approach in fetal weight and growth estimation. Aim of our study was to check the precision and benefit of by Intergrowth-21^st project suggested formula for fetal weight estimation. Additionaly we had an interest to compare its accuracy with the method proposed in Zurich University Hospital and with the formula of Hadlock.

Methods:

A retrospective data analysis consisted of 2903 singleton pregnancies with no malformations. Sonographic fetal weight estimation has been performed no longer than 7 days prior to delivery. Fetal weight was estimated according formula of Hadlock, Zurich method (Selection of Hadlock or Merz formula, regarding the results of (abdomen circumference x femur length)-regression parameter, published in Ultraschall in Med 2017; 38:60 – 64) and formula which was suggested by the Intergrowth-21^st project. Birth weigt has been merged into 6 groups (< 1500: ≥4000 g). The percentage error (PE) and absolute percentage error (APE) were calculated.

Results:

The lowest PE using Zurich method has been estimated in weight groups starting 3000 g (-0,1%, -0,4%, -3,0%). The lowest APE using Zurich method was registered at the macrosomic group, ≥4000 g. PE while using Hadlock formula was comparable except for bigger newborns (3500 – 3999 g: -3,7%, ≥4000 g: -7,0%). PE was the highest using the formula suggested by Intergrowth-21^st project in all weight groups (15,6%, -6,5%, -2,7%, -4,0%, -6,4%, -10,0%). The highest APE 24,0% was registered in the group < 1500 g and 10,2% when birth weight ≥4000 g.

Conclusion:

Our study concludes that accuracy of fetal weight estimation depends on qualitative fetal biometry and the regression formula used. Formula proposed by Intergrowth-21^st project can not be used for general population in any weight group due to the unacceptable systematic as well as random error. Formula of Hadlock, known since1985 showed much lower systematic as well as random error. However, Zurich method has the most accetable systematic error when birth weight was > 3500 g and shows best accuracy for fetal weight estimation.