儿童发生缺铁性贫血的影响因素及其风险预测模型的建立

Influencing factors for iron deficiency anemia in children and establishment of risk prediction model

目的 分析儿童发生缺铁性贫血(IDA)的影响因素,并构建列线图风险预测模型。方法 182例体检儿童中,发生IDA 43例(IDA组),未发生IDA 139例(非IDA组)。采用多因素logistic回归分析筛选儿童发生IDA的独立危险因素。构建列线图风险预测模型,并使用C指数、ROC曲线、校准曲线及决策曲线评估该模型对IDA的预测效能。结果 与非IDA组相比,IDA组母亲孕期IDA、早产、出生体重<2.5kg、母乳喂养及挑食比例较高(P<0.05)。多因素logistic回归分析显示,母亲孕期IDA、早产、出生体重<2.5kg、母乳喂养和挑食均是儿童发生IDA的独立危险因素(P<0.05)。基于独立危险因素构建列线图模型,C指数为0.863。ROC曲线分析显示,列线图模型预测儿童发生IDA的AUC为0.826[95%CI(0.796~0.856),P<0.05],其对应的灵敏度和特异度分别为82.62%和89.41%。校准曲线与理想曲线基本一致,表明该列线图模型的拟合效果较好。决策曲线显示阈值概率为25%~95%时,列线图模型预测儿童发生IDA的净获益值较高。结论 母亲孕期IDA、早产、出生低体重、母乳喂养和挑食均是儿童发生IDA的独立危险因素,基于此构建的列线图模型具有较好的预测效能。

Objective To analyze the influencing factors for iron deficiency anemia(IDA),and construct a nomogram prediction model in children.Methods A total of 182 children underwent health examination were divided into two groups of A(with IDA,43 cases)and B(without IDA,139 cases).Multivariate logistic regression analysis was used to analyze the influencing factors for IDA in children,and the nomogram prediction model was constructed.The C-index,ROC curve,calibration curve and decision curve were used to evaluate the predictive efficiency.Results Compared with group B,the proportions of the children with maternal IDA during pregnancy,premature birth,birth weight<2.5 kg,breast feeding and picky eating were higher in group A(P<0.05).The results of multivariate logistic regression analysis showed that the maternal IDA during pregnancy,premature birth,birth weight<2.5 kg,breast feeding and picky eating were the independent risk factors for IDA in children(P<0.05).The C-index of the nomogram model was 0.863.The AUC value of predicting IDA in children by the nomogram model was 0.826[95%CI(0.796~0.856),P<0.05],with the sensitivity of 82.62%and specificity of 89.41%.The calibration curve was basically consistent with the ideal curve,which indicated the nomogram model fitted well.When the decision curve showed a threshold probability of 25%to 95%,the nomogram model had a higher net benefit value for predicting IDA in children.Conclusion The maternal IDA during pregnancy,premature birth,low birth weight,breast feeding and picky eating are the independent risk factors for IDA in children.The constructed nomogram model has a good predictive value for IDA in children.