A Predictive System For Henoch-sch(?)nlein Purpura Nephritis Established By Multivariate Analysis Plus Nomogram Model

Objectives: Henoch-Sch(?)nlein purpura(HSP)is the most commonly systemic vasculitis in childhood,and its long-term prognosis mainly depends on renal involvement.The present study aimed to identify the risk factors of HSP nephritis(HSPN)and establish a new scoring system for predicting renal involvement in HSP patients.Methods: A total of 1655 Chinese children with HSP were recruited from January 2015 to December 2020.The demographic features,clinical manifestations,laboratory findings and patients’ histories associated with HSPN onset were retrospectively assessed.Independent risk factors were confirmed by multivariate analysis and then assembled into a nomogram for predicting HSPN.Results: HSPN developed in 435 patients(26.28%).By multivariate analysis,age >7.5years,relapse,severe abdominal symptoms,intervals from symptoms onset to diagnosis more than 6 days and persistent purpura more than 12 days were confirmed as significant independent risk factors of HSPN,and their odds ratios(OR)were 1.13,6.36,7.80,1.07 and 1.16,respectively.The sensitivity and specificity of this multivariate regression model were 70.40% and 84.20%,with AUC value of 0.858(95% CI0.825-0.890).A nomogram model for predicting HSPN was established based on the above 5 variables.The calibration curves indicated that the nomogram-predicted probabilities corresponded well with the actual probabilities.Conclusions: Interval from symptom onset to diagnosis is a unique predictor of HSPN in the present study.The predicting model for HSPN is established with an excellent discrimination accuracy.Further multicenter-collaborative studies should be advocated to consolidate our preliminary findings prior to widespread application in clinical practice.