甲泼尼龙治疗过敏性紫癜疗效的影响因素分析及Nomogram预测模型构建

Analysis of factors influencing the efficacy of methylprednisolone in the treatment of henoch-sch?nlein purpura and construction of Nomogram prediction model

目的 探讨甲泼尼龙治疗过敏性紫癜(HSP)疗效的影响因素,构建Nomogram预测模型。方法 回顾性选取2021年5月至2023年4月遂宁市中心医院收治的重症HSP患儿资料,根据甲泼尼龙治疗HSP的疗效分为疗效良好组和疗效不佳组。收集患儿的临床资料,通过受试者工作特征(ROC)曲线分析获取各因素的最佳截断值;logistic多元回归模型分析影响甲泼尼龙治疗HSP疗效的危险因素,构建Nomogram预测模型,内部验证及预测效能分别采用校正曲线和决策曲线评估。结果 共纳入227例重症HSP患儿的临床资料,疗效良好组176例,疗效不佳组51例。疗效不佳组年龄、诱发因素为感染比例、既往HSP病史比例、HSP家族史比例,以及中性粒细胞与淋巴细胞比值(NLR)、血小板与淋巴细胞比值(PLR)、系统免疫炎症指数(SII)均高于疗效良好组,饮食干预比例和IgG水平低于疗效良好组,差异均有统计学意义(P<0.05)。ROC结果显示,年龄、IgG、NLR、PLR、SII的最佳截断值分别为7岁、9.34 g/L、3.58、215.45、726.75。logistic多元回归结果显示,年龄、既往HSP病史、HSP家族史、NLR、PLR、SII是影响甲泼尼龙治疗HSP疗效的危险因素,饮食干预是保护因素(均P<0.05)。内部验证结果显示,Nomogram模型的一致性指数为0.925(0.881~0.970),实际值与预测值基本统一;Nomogram模型阈值>0.08,提示该模型提供了显著的临床净收益。结论 基于年龄、既往HSP病史、HSP家族史、NLR、PLR和SII构建的Nomogram预测模型,对甲泼尼龙治疗HSP疗效的预测价值较好,为改善HSP疗效提供依据。

Objective To investigate the factors influencing the efficacy of methylprednisolone in the treatment of henoch-sch?nlein purpura(HSP) and to construct a Nomogram prediction model.Methods The data of children with severe HSP admitted to Suining Central Hospital from May 2021 to April 2023 were retrospectively selected and divided into the group with good efficacy and the group with poor efficacy according to the efficacy of methylprednisolone in treating HSP.Clinical data of the children were collected,and the optimal cut-off values of each factor were obtained through the analysis of subjects' work characteristics(ROC) curves.A logistic multiple regression model was used to analyze the risk factors affecting the efficacy of methylprednisolone in treating HSP.A Nomogram prediction model was constructed,and the internal validation and prediction efficacy were evaluated by the calibration curve and the decision curve respectively.Results A total of 227 patients with severe HSP were included in this study,including 176 patients in the group with good efficacy and 51 patients in the group with poor efficacy.The age of the group with good efficacy,the inducing factors of infection ratio,previous HSP history ratio,HSP family history ratio,neutrophil to lymphocyte ratio(NLR),platelet to lymphocyte ratio(PLR),and systemic immunoinflammatory index(SII) were all higher than those of the group with good efficacy.Dietary intervention ratio and IgG level were lower than those of the group with good efficacy with statistically significant differences(P<0.05).ROC results showed that the optimal cut-off values of age,IgG,NLR,PLR,and SII were 7 years old,9.34 g/L,3.58,215.45,and 726.75,respectively.Logistic multivariate regression showed that age,previous HSP history,family history of HSP,NLR,PLR,and SII were the risk factors affecting the therapeutic effect of methylprednone on HSP,and dietary intervention was the protective factor(P<0.05).Internal verification results showed that the C-index of the Nomogram model was 0.925(0.881 to 0.970),and the actual and predicted values were basically consistent.The Nomogram model threshold was greater than 0.08,indicating that the model provided significant clinical net benefits.Conclusion The Nomogram prediction model constructed based on age,previous history of HSP,family history of HSP,NLR,PLR,and SII has good predictive value for the efficacy of methylprednisolone in the treatment of HSP,which may provide a basis for improving the efficacy of HSP.