Research On Factors Influencing The Accuracy Of Sonographic Estimated Fetal Weight

Objectives:Accurate estimation of fatal weight prenatally is essential for obstetrician to make delivery managements.At present,two-dimensional ultrasonography is the most widely used method for estimating fetal weight.However,sonographic estimated fetal weigh(SEFW)lacks accuracy.Apart from the intrinsic reasons of the method,various clinical factors and ultrasonic factors influence the accuracy of SEFW.Therefore,the aim of the study was to evaluate the accuracy of SEFW and to determine the relevant clinical and ultrasonic factors influencing the accuracy of SEFW in a large retrospective cohort study on Chinese population.Methods:1.This retrospective study included 7,016 singleton pregnancies in the department of obstetrics at the First Affiliated Hospital of Soochow University,China,between February 2014 and July 2017.Inclusion criteria were a live-birth singleton pregnancy,gestational age(GA)≥ 28 weeks,ultrasound examination with complete biometric parameters – biparietal diameter(BPD),head circumference(HC),abdominal circumference(AC),femur length(FL)and amniotic fluid index(AFI)– within 7 days prior to delivery.Fetus with structural malformations and intrauterine deaths were excluded.2.We selected 10 widely used formulas,containing 1 to 4 parameters of BPD,HC,AC and FL,to calculate the SEFW.The accuracy of SEFW was assessed by calculating:(a)correlation with actual birth weight(BW)(using Pearson’s correlation coefficient);(b)systematic error(also known as mean percentage error,MPE)(mean of(EFW-BW)/BW*100),which reflects the systematic deviation of a model from the actual birth weight;(c)random error(SD of the systematic error×100)-a measure of precision;(d)Mean absolute percentage error(MAPE),which reflects both systematic error and random error;(e)accuracy rate: if SEFW is within 10% of BW,it is defined as accuracy,if not,inaccuracy.By comparing the accuracy of SEFW of the 10 formulae,we chose the most appropriate formula in the following analysis.3.Then we classified the study population into accuracy group and inaccuracy group according to the most appropriate formula above.We compared the maternal-fetal and ultrasonic data between the two groups.In addition,we conducted a single factor analysis on the factors that may influence the accuracy of SEFW,including maternal body mass index(BMI),neonatal birth weight,gestational age,fetal sex,fetal presentation,amniotic fluid volume,the time interval between ultrasonic assessment and delivery(TI)and the seniority of sonographers.4.We then conducted the multiple factors binary logistic regression analysis of the factors above to determine whether those factors influence the accuracy of SEFW.5.Data analysis was performed by Statistical Package for the Social Sciences(SPSS),version 19.0.The one-sample t-test,Student’s t-test and the analysis of variance(ANOVA)were used in comparison of the continuous variables.The chi-square test was used for categorical data.Binary logistic regression analysis was used to determine whether the potential factors influence the SEFW.Random errors were compared using the Levene’s test.A Bonferroni adjustment was carried out if the comparison was conducted in 3 or more groups.Difference was considered significant when p<0.05.Results:All models had high correlation coefficient between SEFW and BW(r=0.819 to 0.843).Model 3 showed the lowest systematic error(0.36%).Except for model 3 and model 4,the other models all had negative systematic error,which meant a tendency of underestimation of fetal weight.Model 10 showed the lowest random error(8.23%);Model 6 had the lowest MAPE(6.59%).Only the MAPE of model 3(6.69%)showed no difference compared to model 6;Model 6 had the highest accuracy rate(78.18%),followed by model 3(78.12%).In general,combination systematic error,MAPE and accuracy rate,model 3(Hadlock et al,lg BW=1.1134+0.05845(AC)-0.000604(AC)2-0.007365(BPD)2+0.000595(BPD)(AC)+0.1694(BPD))was the most appropriate formula for our population.Therefore,model 3 was chosen for the following analysis of the potential factors influencing the accuracy of SEFW.2.The single variant analysis of potential factors influencing SEFW2.1 The maternal-fetal data and ultrasonic data of the study populationAccording to model 3,the study population was classified into accuracy group and inaccuracy group and the maternal-fetal data and ultrasonic outcomes were compared between them.We can see that women in accuracy group had higher vaginal birth rate,larger neonatal birth weight and gestational age,and the distribution of SGA/AGA/LGA and sonographers seniority showed a significant difference(p<0.05).Characteristics such as maternal age,maternal height,maternal weight,maternal BMI,maternal parity,the rate of maternal DM,fetal sex,distribution of amniotic fluid volume and TI all showed no significant difference between the two groups(p>0.05).2.2 The single variant analysis of potential factors influencing SEFWThen we made a stratified analysis to the 9 potential factors.According to BMI,population was classified into 4 groups----normal BMI(18-23.9),overweight(24-27.9),obesity 1(28-30)and obesity 2(>30).The values of systematic error(p<0.001)and MAPE(p=0.013)showed differences among the four groups.Compared to non-DM group,the DM group had higher BW and systematic error and MAPE.As to BW,the group of BW between 2,500 g and 3,999 g had the lowest systematic error(0.74%)and MAPE(6.23%).In addition,the group of macrosomia had a negative systematic error,the group of BW < 2,500 g had a positive systematic error.As to GA,the groups of GA between 37 weeks and 40 weeks and GA > 40 weeks showed no difference in systematic error and MAPE,both the two groups had higher systematic error and lower MAPE than the group of GA < 37 weeks.Considering GA and BW concurrently,the AGA group has the lowest systematic error(0.83%)and MAPE(6.17%),then the LGA group(-6.80% and 8.24% respectively),the SGA group has the highest systematic error(10.06%)and MAPE(11.17%).Female fetus had larger systematic error than male fetus(p=0.002),although the MAPEs of male and female fetus showed no difference(p=0.180).The systematic errors showed no difference of the cephalic presentation group and non-cephalic presentation group,but the former had lower MAPE(6.61% vs.8.04%,p=0.001).Though the systematic errors of the three groups with different AFV showed a difference(p=0.011),the MAPEs of the three groups showed no difference(p=0.199).According to TI,the values of systematic error showed an obvious differences(p<0.001)and the group of TI between 1-3 days had the lowest systematic error.According to seniority,the MAPEs of the 3 groups showed no difference(p=0.162),but the systematic error indicated an obvious difference(p<0.001).Strangely,the results showed that the high seniority group had the highest systematic error and the low seniority group had the lowest systematic error,which is abnormal,therefore we made an analysis of the accuracy of SEFW to the 7 sonographers respectively.We found that in high seniority group,there was a sonographer with the lowest accuracy rate,on the contrary,in low seniority group,one sonographer had the highest accuracy rate.This phenomenon suggested that the accuracy of SEFW was associated to the personal technique of sonographers,not to the seniority of sonographers.3.The results of multiple factors logistic regression analysisThe results showed that maternal BMI,maternal DM,GA,fetal presentation and AFV had no influence on the accuracy of SEFW(p>0.05).Neonatal birth weight,fetal sex,time interval and sonographer showed different influences on the accuracy of SEFW(p<0.05).The group of fetus in 2,500g-3,999 g had the most accurate SEFW.Comparing to fetus in 2,500g-3,999 g group,the accuracy of SEFW of fetus < 2,500 g decreased by 69.10%(p<0.001,OR:0.309),and the accuracy of SEFW of fetus ≥ 4,000 g decreased by 54.10%(p<0.001,OR:0.459).The fetal weight of male fetus would be more accurately estimated than female fetus.The accuracy of SEFW would decrease by 21.70% compared to male fetus(p=0.021,OR: 0.783).SEFW calculated within 3 days prior to delivery was most accurate.Comparing to the SEFW calculating on 1-3 days before delivery,the SEFW performing on the day of delivery showed no significant difference(p=0.149,OR: 0.867),however,the accuracy of SEFW performing on 4-7days before delivery showed a significant decrease(p=0.014,OR: 0.852).The accuracy of SEFW among the three categories of seniority showed an obvious difference(p=0.032).Comparing to low seniority,the accuracy of SEFW in median seniority group showed no difference(p=0.176),but the SEFW of high seniority group indicated an obvious decrease(p=0.009,OR: 0.818),which once again suggested that the accuracy of SEFW is associated to the personal technique level instead of the seniority of sonographers.Conclusion1.Ultrasonic estimated fetal weight is generally accurate.However,there is a tendency of underestimation of fetal weight.2.Model 3(Hadlock et al,lg BW=1.1134+0.05845(AC)-0.000604(AC)2-0.007365(BPD)2+0.000595(BPD)(AC)+0.1694(BPD))was the most appropriate formula for our population.3.Maternal BMI,maternal DM,GA of birth,fetal presentation and AFV had no influence on the accuracy of SEFW.Neonatal BW,fetal sex,time interval and sonographer showed different influences on the accuracy of SEFW.4.BW in 2,500g-3,999 g would be estimated most accurately,then fetus ≥ 4,000 g,SEFW of fetus < 2,500 g were most inaccurate.In addition,macrosomia tend to be underestimated and small fetuses tend to be overestimated,therefore,we suggest making and exerting specialized formulas for small and big fetuses respectively to increase the accuracy.5.Male fetus has higher SEFW than female fetus.6.SEFW calculated within 3 days of delivery was higher than SEFW calculated 4-7days before delivery,hence we suggest fetal weight should be assessed within 3 days prior to delivery.7.The accuracy was influenced by the individual technique level of sonographers.Therefore the technique assessment should be performed regularly to improve the accuracy of the operation.