产前胎儿体重评估模型的建立与分析

目的利用临床常规指标建立胎儿体重预测模型。方法选取长垣县人民医院893例临产妇为研究对象，观察其体重指数（BMI）、孕期增重、孕妇腹围、宫高，分娩前1周内B超测量胎儿双顶径（BPD）、头围（HC）、腹围（Ac）、股骨长度（FL），分娩后0．5h内称重新生儿体重，分别进行Pearson相关分析、线性回归分析。结果①新生儿体重与孕期增重（R=0．534）、宫高（R=0．589）、孕妇腹围（0．413）、胎儿BPD（R=0．459）、AC（R=0．643）和FL（R=0．602）呈正相关（p〈0．（11）。②多元线性回归分析得出新生儿体重预测多指标模型：新生儿体重（g）=61．94PL＋41．832孕期增重＋48．589胎儿AC＋43．435宫高-5124．661。③独立变量线性回归得单一指标的新生儿体重预测模型：胎儿FL法：新生儿体重（g）=164．667FL（mm）-8828．406。孕期增重法：新生儿体重（g）=101．839孕期增重（kg）＋1555．537。胎儿Ac法：新生儿体重（g）=168．825AC（cIn）-2457．808。宫高法：新生儿体重（g）=169．402宫高（cm）-2605．743。，④多指标法预测体重与新生儿实际体重的相关系数（R=0．865）大于胎儿FL法（R=0．829）、孕期增重法（0．731）、胎儿Ac法（R=0．812）和宫高法（R=0．746），多指标法预测体重的平均绝对误差（241．32g）和相对误差（7．52％）小于胎儿FL法（259．34g，8．25％）、孕期增重法（262．32g，8．39％）、胎儿Ac法（254．96g，8．07％）和宫高法（271．58g，8．62％）。结论以孕期增重、宫高、胎儿Ac、FL综合指标建立的模型较单一指标更能准确预测新生儿体重。

The establishment and evaluation of prenatal predicting model on fetal weight

Objective ] To develop a predicting model on fetal weight based on clinical conventional indexes. Methods ] A total of 893 parturient women were included in the study.The body mass index（BMI）,gestational weight gain（GWG）,uterine fundus height（UFH）, abdmoinal perimeter（AP） of parturient women were observed,and the fetal biparietal diameters（BPD）, head circumferences（HC）,abdmoinal circumferences（AC）, femur lengths（FL） were measured by uhralsound within 1 week of delivery,and neonate were weighted 0.5h after delivery.Pearson correlative anslysis and linear regression method were applied. Resluts ] ① Neonatal weight were positively correlated with GWG（R=O.534）,UFH（R=O.589）,AP（O.413）,BPD（R=0.459）, AC（R=0.643） and FL（R=0.602）. ② Multiple linear regression analysis develop a predicting model was neonatal weight（g）= 61.94FL＋41.832GWG＋48.589AC ＋43.435UFH-5124.661. ③ Single-parameter linear regression analysis develop a predicting model was following：FL methods：neonatal weight（g）=164.667FL（mm）-8828.406, GWG methods：neonatal wei ght（g）=101.839GWG（kg）＋1555.537,AC methods：neonatal weight（g）=168.825AC（em）-2457.808, FUH methods：neonatal weight（g）=169.402UFH（em）-2605.743.④ The correlation coefficient of the neonatal actual weight and the estimated fetal weight by multiple indexes methods（R=0.865） was greater than that of FL methods（R=0.829）,GWG methods（0.731） ,AC methods（R=0.812） and FUH methods（R=0.746）.The mean absolute error（241.32g） and the mean relative error（7.52%） by multiple indexes methods was lower than that of FL methods（259.34g, 8.25%）,GWG methods（262.32g, 8.39%） ,AC methods（254.96g, 8.07%） and FUH methods（271.58g, 8.62%）. Conclusion ] The estimated fetal weight by multiple indexes predicting model based on GWG, FUH ,AC and FL was better than that of single index.