基于外周血检验指标构建新生儿窒息并发感染性肺炎列线图预测模型

Construction of nomogram prediction model for concurrent infectious pneumonia in neonates with asphyxia based on peripheral blood test indicators

  • 摘要: 目的 基于新生儿临床资料和外周血检验指标,筛选新生儿窒息并发感染性肺炎的预测因素,并构建列线图预测模型。方法 回顾性选取2022年3月-2023年10月在赣州市妇幼保健院生产后出现窒息情况的新生儿334例,根据患儿有无并发感染性肺炎分为肺炎组84例和无肺炎组250例,收集新生儿娩出后24 h外周血检验指标和产妇分娩资料,单因素及多因素logistic回归分析新生儿窒息并发感染性肺炎的预测因素,构建列线图预测模型,通过受试者工作特征(ROC)曲线、校准曲线评估模型的预测性能。结果 外周血白细胞介素-6(IL-6)、前白蛋白(PA)、红细胞沉降率(ESR)、白细胞计数(WBC)、血小板计数与淋巴细胞计数比值(PLR)及出生胎龄、出生体质量、辅助气管插管均是新生儿窒息并发感染性肺炎的预测因素(P<0.05); 构建的列线图预测模型其曲线下面积(AUC)为0.900(95%CI:0.863~0.938),分值临界值为0.219,敏感度为0.881,特异度为0.796,模型具有较好的区分能力; 校准曲线的平均绝对误差为0.020,与45°的理想线部分重合,系统性误差较少。结论 新生儿窒息并发感染性肺炎可通过外周血检验指标及出生胎龄、气管插管等临床因素进行预测,基于各临床因素构建的风险预测模型具有较好的预测效能,可帮助临床上筛选并发感染性肺炎发生风险较高的新生儿,以便及时采取干预措施。

     

    Abstract: OBJECTIVE To screen out the predictive factors for concurrent infectious pneumonia in the neonates with asphyxia based on clinical data of neonatal delivery and peripheral blood test indicators and establish the nomogram prediction model. METHODS A total of 334 neonates who had asphyxia after being given birth in Ganzhou Maternal and Child Health Hospital from Mar. 2022 to Oct. 2023 were retrospectively enrolled in the study and were divided into the pneumonia group with 84 cases and the no pneumonia group with 250 cases according to the status of concurrent infectious pneumonia. The peripheral blood test indicators and clinical data of puerpera were collected from the neonates 24 hours after the birth. Univariate analysis and multivariate logistic regression analysis were performed for the predictive factors for the infectious pneumonia in the neonates with asphyxia, the nomogram prediction model was established, and the predictive efficiency of the model was assessed by receiver operating characteristic (ROC) curves and calibration curves. RESULTS The predictive factors for the concurrent infectious pneumonia in the neonates with asphyxia included the peripheral blood interleukin-6 (IL-6), prealbumin (PA), erythrocyte sedimentation rate (ESR), white blood cell (WBC) counts, platelets to lymphocytes ratio (PLR), gestational age at birth, birth weight, and assisted endotracheal intubation(P<0.05). The area under the curve (AUC) of the nomogram prediction model was 0.900(95%CI:0.863 to 0.938), with the cutoff value 0.219, the sensitivity 0.881, the specificity 0.796; the model had incredible discriminating capability; the mean absolute error (MAE) of the calibration curves was 0.020, they partially overlapped with the ideal 45° line, indicating minimal systematic error. CONCLUSIONS The peripheral blood test indicators and clinical factors such as gestational age at birth and endotracheal intubation can be used for prediction of the concurrent infectious pneumonia in the neonates with asphyxia. The risk prediction model that is established based on the clinical factors has high predictive efficiency, which may facilitate the clinical screening of the neonates at high risk of concurrent infectious pneumonia so as to take interventional measures in a timely manner.

     

/

返回文章
返回