标题：重型颅脑损伤机械通气输血患者发生急性肺损伤的影响因素及其列线图模型的构建

      作者：王云鹏，王跃玲，吕旭方，张新蔚，马宝亮，徐丽    河南大学第一附属医院医学检验科，河南 开封 475001

      卷次： 2024年35卷14期

      【摘要】 目的 探究重型颅脑损伤(STBI)机械通气输血患者急性肺损伤(ALI)发生影响因素，并构建列线图模型。方法 收集2020年5月至2023年5月河南大学第一附属医院收治的258例STBI机械通气输血患者的临床资料进行回顾性研究，根据有无发生ALI分为ALI组(n=67)和N-ALI组(n=191)，比较两组患者的一般资料和早期预警标志物，采用Logistic回归方程筛选ALI的影响因素，采用R软件构建列线图预测模型，绘制受试者工作特征(ROC)曲线及曲线下面积(AUC)、校准曲线、决策曲线(DCA)分析列线图预测模型区分度、一致性、临床有效性，同时基于列线图模型建立危险分层系统。结果 ALI组患者的输血量、Murray评分、机械通气时间及血清C-X-C型趋化因子受体4 (CXCR4)、含热蛋白结构域6的NOD样受体(NLRP6)、含热蛋白结构域3的NOD样受体(NLRP3)水平分别为(600.24±180.18) mL、(0.30±0.12)分、(7.30±1.34) d、(20.24±6.07) μg/L、(53.96±4.48) μg/L、(42.41±5.54) μg/L，明显高于N-ALI组的(418.18±125.12) mL、(0.10±0.03)分、(5.52±1.00) d、(14.14±4.18) μg/L、(45.56±5.29) μg/L、(36.61±4.48) μg/L，低氧血症发生率为56.72% (38/67)，也明显高于N-ALI组的34.56% (66/191)，差异均有统计学意义(P<0.05)；Logistic回归方程分析结果显示，输血量、Murray评分、机械通气时间、低氧血症及血清CXCR4、NLRP6、NLRP3均是ALI发生的影响因素(P<0.05)；基于Logistic回归方程结果构建列线图预测模型，AUC为0.955，且预测概率与实际概率一致性较高，在阈值0~100%概率内，该模型临床获益率高；在整个研究人群中，低风险、中等风险、高风险患者ALI发生率分别为8.75%、24.00%、40.78%。结论 含血清CXCR4、NLRP6、NLRP3的列线图模型预测STBI机械通气输血患者ALI效能较高，可帮助临床医生识别高风险人群，采取合理治疗措施，降低ALI发生风险。
      【关键词】 重型颅脑损伤；机械通气；输血；急性肺损伤；列线图模型；预警标志物
      【中图分类号】 R651.1+5 【文献标识码】 A 【文章编号】 1003—6350（2024）14—2001—06

Influencing factors of acute lung injury in patients with severe craniocerebral injury undergoing mechanicalventilation and blood transfusion and the construction of its nomogram model.
WANG Yun-peng, WANG Yue-ling,LYU Xu-fang, ZHANG Xin-wei, MA Bao-liang, XU Li. Department of Clinical Laboratory, the First Affiliated Hospital ofHenan University, Kaifeng 475001, Henan, CHINA
【Abstract】 Objective To explore the influencing factors of acute lung injury (ALI) in patients with severecraniobrain injury (STBI) undergoing mechanical ventilation and blood transfusion, and to construct a nomogram mod-el. Methods The clinical data of 258 patients with STBI undergoing mechanical ventilation and blood transfusion ad-mitted to the First Affiliated Hospital of Henan University from May 2020 to May 2023 were collected for retrospectivestudy. All patients were divided into ALI group (n=67, with ALI) and N-ALI group (n=191, without ALI) according tothe occurrence of ALI. The general data and early warning markers of the two groups were compared. Logistic regres-sion equation was used to screen the influencing factors of ALI. R software was used to construct a nomogram predic-tion model, and the receiver operating characteristic (ROC) curve, area under ROC curve (AUC), calibration curve, de-cision curve analysis (DCA) were used to analyze the nomogram prediction model for discrimination, consistency, andclinical effectiveness. At the same time, a risk stratification system was established based on the nomogram model.Results The blood transfusion volume, Murray score, mechanical ventilation time, and serum levels of C-X-C che-mokine receptor 4 (CXCR4), NOD-like receptor with pyrin domain 6 (NLRP6), and NOD-like receptor with pyrin do-main 3 (NLRP3) in the ALI group were (600.24±180.18) mL, (0.30±0.12) points, (7.30±1.34) d, (20.24±6.07) μg/L,(53.96±4.48) μg/L, and (42.41±5.54) μg/L, respectively, which were significantly higher than (418.18±125.12) mL,(0.10±0.03) points, (5.52±1.00) d, (14.14±4.18) μg/L, (45.56±5.29) μg/L, and (36.61±4.48) μg/L in the N-ALI group,and the incidence of hypoxemia was 56.72% (38/67), significantly higher than 34.56% (66/191) in the N-ALI group (P<0.05). Logistic regression equation analysis showed that blood transfusion volume, Murray score, mechanical ventilationtime, hypoxemia, and serum CXCR4, NLRP6, and NLRP3 were all factors affecting the occurrence of ALI (P<0.05).Based on the logistic regression equation results, a nomogram prediction model was constructed with an AUC of 0.955,and the consistency between the predicted probability and the actual probability was high. Within the threshold of 0 to100% probability, the model had a high clinical benefit rate. Among the entire study population, the incidence of ALI inlow-risk, intermediate-risk, and high-risk patients was 8.75%, 24.00%, and 40.78%, respectively. Conclusion The no-mographic model containing serum CXCR4, NLRP6, and NLRP3 has high efficacy in predicting ALI in patients withSTBI mechanical ventilation transfusion, which can help clinicians identify high-risk groups and take reasonable treat-ment measures to reduce the risk of ALI.
      【Key words】 Severe craniocerebral injury; Mechanical ventilation; Blood transfusion; Acute lung injury; Nomo-gram model; Warning marker