基于肺超声评分、血清SDC-1、ANGPTL4构建模型对早产儿NRDS预后的预测价值分析

Predictive value of a model based on lung ultrasound score, serum syndecan-1, and angiopoietin-like protein 4 for the prognosis of neonatal respiratory distress syndrome in premature infants

摘要:

目的 基于肺超声评分、血清多配体聚糖-1（SDC-1）、血管生成素样蛋白4（ANGPTL4）构建早产儿新生儿呼吸窘迫综合征（NRDS）预后的预测模型并进行验证。方法 回顾性分析河北省沧州中西医结合医院2022年7月—2025年6月收治的350例NRDS早产儿的临床资料。采用随机数字表法按照7∶3的比例将患儿分为建模集（n =245）和验证集（n =105），根据建模集患儿预后情况分为良好组（n =131）和不良组（n =114）。比较患儿的一般资料、1 min和5 min Apgar评分、新生儿急性生理学评分围生期补充版Ⅱ（SNAPPE Ⅱ评分）、肺超声评分、白细胞计数、血小板计数、白细胞介素-6（IL-6）、降钙素原（PCT）、SDC-1、ANGPTL4水平。采用多因素逐步Logistic回归分析NRDS早产儿预后不良的影响因素。构建早产儿NRDS预后不良预测模型，采用Bootstrap重复抽样法（n =1 000）对模型进行内部验证；Hosmer-Lemeshow检验评估预测模型的拟合优度，P >0.05表示模型拟合效果良好；ROC曲线、校准曲线、决策曲线评估预测模型的效能。结果 建模集和验证集一般资料比较，差异均无统计学意义（P >0.05）；不良组患儿早期早产占比、SNAPPE Ⅱ评分、通气时间、肺超声评分、SDC-1水平和ANGPTL4水平均高于良好组，出生体重、1 min Apgar评分和5 min Apgar评分均低于良好组（P <0.05）。早期早产（<32周）占比、SNAPPE Ⅱ评分、肺超声评分、SDC-1、ANGPTL4升高是NRDS早产儿预后不良的独立危险因素（P <0.05），1 min Apgar评分升高是独立保护因素（P <0.05）。基于回归分析结果构建NRDS早产儿预后不良的预测模型，建模集和验证集曲线下面积分别为0.815（95% CI：0.761，0.862）、0.800（95% CI：0.711，0.872）。建模集和验证集的H-L拟合优度检验均无统计学意义（P >0.05），建模集和验证集在阈值概率0.05～0.95范围内，模型预测策略较“全干预”及“无干预”策略均具有更高的净获益，其中在0.10～0.80区间内净获益优势尤为显著。结论 基于肺超声评分、血清SDC-1、ANGPTL4构建早产儿NRDS预后的预测模型具有良好的区分度，与实际结果一致性较高，可能为临床治疗决策的制订提供一定参考。

Abstract:

Objective To construct and validate a model for predicting the prognosis of neonatal respiratory distress syndrome (NRDS) in premature infants based on lung ultrasound score, serum syndecan-1 (SDC-1), and angiopoietin-like protein 4 (ANGPTL4).Methods The clinical data of 350 premature infants with NRDS admitted to Cangzhou Hospital of Integrated Traditional Chinese and Western Medicine in Hebei Province from July 2022 to June 2025 were retrospectively analyzed. Using a random number table, the infants were divided into a modeling set (n = 245) and a validation set (n = 105) at a ratio of 7:3. According to outcomes in the modeling set, the infants were further divided into a favorable prognosis group (n = 131) and an unfavorable prognosis group (n = 114). General data, 1-min and 5-min Apgar scores, Score for Neonatal Acute Physiology with Perinatal Extension II (SNAPPE II), lung ultrasound score, white blood cell count, platelet count, interleukin-6 (IL-6), procalcitonin (PCT), SDC-1, and ANGPTL4 levels were compared. Multivariate stepwise logistic regression was used to analyze factors associated with an unfavorable prognosis in premature infants with NRDS. A prediction model was constructed and internally validated using bootstrap resampling (n = 1,000). The Hosmer-Lemeshow test was used to assess model goodness of fit, with P > 0.05 indicating good fit. ROC, calibration, and decision curves were used to evaluate model performance.Results General data did not differ significantly between the modeling and validation sets (P > 0.05). The proportion of early preterm birth, SNAPPE II score, ventilation duration, lung ultrasound score, and SDC-1 and ANGPTL4 levels were higher, whereas birth weight and 1-min and 5-min Apgar scores were lower, in the unfavorable prognosis group than in the favorable prognosis group (P < 0.05). Early preterm birth (< 32 weeks), higher SNAPPE II score, lung ultrasound score, SDC-1 level, and ANGPTL4 level were independent risk factors for an unfavorable prognosis, whereas a higher 1-min Apgar score was an independent protective factor (P < 0.05). Based on the regression analysis, a model for predicting an unfavorable prognosis in premature infants with NRDS was established. The areas under the curve for the modeling and validation sets were 0.815 (95% CI: 0.761, 0.862) and 0.800 (95% CI: 0.711, 0.872), respectively. The Hosmer-Lemeshow goodness-of-fit test was nonsignificant in both sets (P > 0.05). At threshold probabilities of 0.05-0.95, the model-based strategy provided greater net benefit than the treat-all and treat-none strategies in both sets, with a particularly marked advantage at thresholds of 0.10-0.80.Conclusion The model based on lung ultrasound score, serum SDC-1, and ANGPTL4 has good discrimination and high agreement with actual outcomes and may provide a reference for clinical decision-making in premature infants with NRDS.