| Objective: A coronary artery lesion(CAL)is more likely to occur in patients who are resistant to intravenous immunoglobulin(IVIG)therapy for Kawasaki disease(KD).The aim of this retrospective study was to examine the clinical cases and laboratory results of KD patients who were hospitalized over the past five years.The study analyzed the data to identify the risk factors for IVIG resistance and CAL in KD and then developed a prediction scoring model to predict IVIG resistance and CAL.The data was divided into a modeling group and a validation group in a 7:3 ratio,and crucial indicators were analyzed through statistical methods.Methods: A retrospective analysis was conducted on the medical records and ultrasound results of patients with KD admitted to Anhui Provincial Hospital in the past 6 years,and the data was divided into a modeling data set and a validation data set according to the ratio of 7:3.According to the response of KD children to IVIG treatment and the inner diameter of the coronary artery,the children were divided into an IVIG sensitive group and an IVIG resistant group,a normal CAL group(NCAL group),and a coronary artery lesion group(CAL group).Variables with differences in univariate analysis were entered into multivariate logistic regression to determine the risk factors for IVIG resistance and CAL in KD,and the ROC curve was used to evaluate the prediction effect of the model.The area under the curve(AUC),sensitivity,and specificity were calculated and validated in the validation group.In all the data,meaningful continuous variables were converted into binary variables according to the best prediction cutoff value determined by the receiver operator characteristic curve(ROC)and assigned according to the β value to establish IVIG resistance and the CAL predictive scoring system.Results: 660 children with KD were included in the study,including 459 cases in the modeling data set and 201 cases in the verification data set.The multivariate analysis result in the modeling data showed that Hb,ESR,ALB,and PLR were independent risk factors for IVIG resistance in children with KD;the area under the ROC curve was 0.753;the sensitivity was82.50%;and the specificity was 61.10%.In the validation data,the AUC was 0.746,the sensitivity was 78.60%,and the specificity was 72.70%.A simple scoring model was constructed using the above four variables in all the data.The area under the ROC of the scoring model was 0.784,and the area under the ROC of the logistic regression model was0.786.The critical value of the scoring model is above 4 points,the cutoff value of the logistic regression model is above 0.10,the sensitivities of the two models are 73.50% and 72.00%,the specificities are 75.20% and 74.10%,respectively,and the Kappa consistency test value is 0.75.Male,days of fever,IVIG resistance,and ALB were found to be independent risk factors for coronary artery damage in a multivariate logistic regression analysis of the CAL modeling group.The AUC under the ROC was 0.773,the sensitivity was 58.50%,and the specificity was 82.50%.The AUC in the verification group was 0.816,with a sensitivity of 73.70% and a specificity of 73.60%.The cut-off values of ALB and fever time were determined as 38.05 g/L and 8 days,respectively,by the same method,and a scoring system was established.The logistic regression model of all the data showed that the AUC was 0.765.For a cutoff value above 0.14,the sensitivity and specificity were 59.70% and 82.10%,respectively.The AUC of the simple scoring model was 0.745 when the cutoff value was 4 or more,the sensitivity and specificity were 56.40 and 76.90%,and the Kappa consistency test value was 0.65.Conclusion:(1)ESR,ALB,Hb and PLR are independent risk factors of IVIG resistance in KD;male,fever days,IVIG resistance and ALB were independent risk factors for CAL.(2)The prediction system for IVIG resistance and CAL based on the results of multivariate logistic regression analysis has high accuracy and a certain guiding significance. |
PDF Full Text Request