过敏性紫癜患儿发生消化道出血的危险因素分析及列线图预测模型构建  

Analysis of risk factors of gastrointestinal bleeding in children with henoch-schonlein purpura and construction of Nomogram prediction model

作　　者：丁晏 徐德万 詹姗姗 DING Yan;XU Dewan;ZHAN Shanshan(Department of Pediatrics,Shangrao Municipal Hospital,Jiangxi Province,Shangrao 334000,China;Department of Pediatrics,the First Hospital of Jiaxing,Zhejiang Province,Jiaxing 314000,China)

机构地区：[1]江西省上饶市立医院儿科,江西上饶334000 [2]浙江省嘉兴市第一医院儿科,浙江嘉兴314000

出　　处：《中国当代医药》2024年第34期28-32,共5页China Modern Medicine

基　　金：江西省卫生健康委科技计划项目(202410980)。

摘　　要：目的探讨过敏性紫癜(HSP)患儿发生消化道出血的危险因素,并构建列线图(Nomogram)预测模型。方法回顾性分析2021年11月至2023年11月上饶市立医院和嘉兴市第一医院收治的120例HSP患儿的临床资料,按照是否发生消化道出血分为发生组(53例)与未发生组(67例)。比较两组患儿的临床资料,通过受试者工作特征(ROC)曲线分析相关参数对HSP患儿发生消化道出血的预测价值;通过logistic回归分析影响HSP患儿发生消化道出血的独立危险因素;通过R语言软件4.0“rms”包构建HSP患儿发生消化道出血的Nomogram风险预测模型,并进行验证。结果单因素分析结果显示,两组患儿的皮疹分布≥3个比例、呕吐比例、中性粒细胞/淋巴细胞比值(NLR)、C反应蛋白(CRP)、白细胞(WBC)、血小板计数(PLT)、中性粒细胞计数(NEU)水平比较,差异有统计学意义(P<0.05)。ROC分析结果显示,NLR、CRP、WBC、PLT、NEU曲线下面积(AUC)分别为0.737、0.829、0.719、0.830、0.876,敏感度分别为62.30%、66.00%、41.50%、79.20%、71.70%,特异度分别为77.60%、89.60%、91.00%、74.60%、94.00%,最佳截断值分别为>4.1、>12.46 mg/L、>13.9×10^(9)/L、>249×10^(9)/L、>8×10^(9)/L。多因素分析结果显示,皮疹分布≥3个(β=0.704,OR=2.022,95%CI:1.457~2.963)、呕吐(β=0.374,OR=1.453,95%CI:1.225~1.856)、NLR>4.1(β=0.586,OR=1.796,95%CI:1.021~2.356)、CRP>12.46 mg/L(β=0.758,OR=2.134,95%CI:1.563~2.796)均是HSP患儿发生消化道出血的独立危险因素(P<0.05)。Nomogram风险预测模型预测HSP患儿发生消化道出血的AUC为0.853(95%CI:0.814~0.892),校正曲线趋于理想曲线;预测阈值>0.18,Nomogram风险预测模型提供临床净收益,且临床净收益均高于独立预测因子。结论基于皮疹分布、呕吐情况、NLR、CRP水平可以构建HSP患儿发生消化道出血的Nomogram风险预测模型,该预测模型具有重要的临床价值。Objective To explore the risk factors of gastrointestinal bleeding in children with henoch-schonlein purpura(HSP)and to construct a Nomogram prediction model.Methods The clinical data of 120 children with HSP admitted to Shangrao Municipal Hospital and the First Hospital of Jiaxing from November 2021 to November 2023 were retrospectively analyzed.They were divided into occurrence group(53 cases)and non-occurrence group(67 cases)according to whether gastrointestinal bleeding occurred.The clinical data of the two groups were compared,and the receiver operating characteristic(ROC)curve was used to analyze the predictive value of relevant parameters for gastrointestinal bleeding in children with HSP.Logistic regression was used to analyze the independent risk factors for gastrointestinal bleeding in children with HSP.A Nomogram risk prediction model for gastrointestinal bleeding in children with HSP was constructed and verified by R language software 4.0"rms"package.Results The results of single factor analysis showed that there were statistically significant differences between the occurrence group and the non-occurrence group in the proportion of distribution of≥3 rashes,the proportion of vomiting,and the levels of neutrophil/lymphocyte ratio(NLR),C-reactive protein(CRP),white blood cells(WBC),platelet(PLT)and neutrophil(NEU)(P<0.05).The areas under the curve(AUC)of NLR,CRP,WBC,PLT and NEU were 0.737,0.829,0.719,0.830,0.876 respectively, the sensitivities were 62.30%, 66.00%, 41.50%, 79.20% and 71.70% respectively, and the specificities were respectively 77.60%, 89.60%, 91.00%, 74.60%, 94.00%, and the optimal cut-off values were >4.1, >12.46 mg/L, >13.9×10^(9)/L, >249×10^(9)/L, and >8×10^(9)/L, respectively. Multivariate analysis showed that rash ≥3 (β=0.704, OR=2.022, 95%CI: 1.457-2.963), vomiting (β=0.374, OR=1.453, 95%CI: 1.225-1.856), NLR>4.1 (β=0.586, OR= 1.796, 95%CI: 1.021-2.356), CRP>12.46 mg/L (β=0.758, OR=2.134, 95%CI: 1.563-2.796) were independent risk factors for gastrointestinal bleeding in

关 键 词：过敏性紫癜 消化道出血 危险因素 列线图 

分 类 号：R593.1[医药卫生—内科学] R725.9[医药卫生—临床医学]