Analysis of risk factors for steroid resistance in children with primary nephritic syndrome and discussion on the prediction model of steroid resistance

Objective To seek the risk factors of steroid resistance in children with primary nephritic syndrome, and construct the predicting model of steroid resistance. Methods The clinical indicators of 185 patients with primary nephrotic syndrome (PNS) were collected, including clinical data, laboratory and imaging examination. The risk factors of steroid resistance were found using single factor analysis, receiver operator characteristic (ROC) and logistic regression test. The predicting model of steroid resistance was constructed based on integral method model. Results The results of single factor analysis, receiver operating characteristic curve (ROC) and logistic regression analysis showed that the age more than 6.5 years old, having microscopic hematuria and the 24 h urine protein content (24 hUP) more than 177.49 mg·kg-1·d-1 were the significant risk factors of steroid resistant nephrotic syndrome (SRNS). Logistic regression prediction model was Y=6.761-2.947X1-3.336X2-2.669X3. The result of receiver operator characteristic showed that when the score was 0.95, the sensitivity and specificity was 56.56%, 96.62% respectively and the area under ROC was 0.86, P <0.05). Conclusions The age more than 6.5 years old, having microscopic hematuria and 24 h urine protein content more than 177.49 mg·kg-1·d-1 are the significant risk factors of SRNS. Key words: Nephrotic syndrome, primary; Drugs resistance, steroid; Children; Predict