Hepcidin in Kawasaki disease: upregulation by acute inflammation in patients having resistance to intravenous immunoglobulin therapy

Neutrophil percentage-to-albumin ratio is a potential marker of intravenous immunoglobulin resistance in Kawasaki disease

Intravenous immunoglobulin (IVIG) resistance in Kawasaki disease (KD) was associated with coronary artery lesions. Neutrophil percentage-to-albumin ratio (NPAR) is an index of mortality in several inflammatory diseases. This study focused on the association of NPAR with IVIG- resistance in KD. Clinical and laboratory data of 438 children with KD before IVIG treatment were retrospectively analyzed. Notably, high NPAR was associated with older age, high WBC, NP, ALT, total bilirubin and CRP, as well as with high the incidence of IVIG-resistance, and with low hemoglobin (Hb), PLT, ALB and sodium levels. NPAR (OR: 2.366, 95% CI: 1.46-3.897, p = 0.001) and Hb (OR: 0.967, 95% CI: 0.944-0.989, p = 0.004) were independent risk factors for IVIG-resistance. NPAR showed linear relation with IVIG-resistance (p for nonlinear = 0.711) and the nonlinear correlation was found between IVIG-resistance and Hb (p for nonlinear = 0.002). The predictive performance of NPAR was superior to Beijing model (z = 2.193, p = 0.028), and not inferior to Chongqing model (z = 0.983, p = 0.326) and the combination of NPAR and Hb (z = 1.912, p = 0.056). These findings revealed that NPAR is a reliable predictor of IVIG-resistance.

Evaluation of Formosa score and diagnostic sensitivity and specificity of four Asian risk scores for predicting intravenous immunoglobulin resistance in Kawasaki disease: a bivariate meta-analysis

Background

In 2016, Lin et al. developed a prediction score of non-responsiveness to intravenous immunoglobulin (IVIG) in patients with Kawasaki disease (KD) (Lin et al., 2016). Various studies have attempted to validate the Formosa score, but inconsistent results have given us new opportunities and challenges. The aim of this meta-analysis is to explore the role of the Formosa score as a risk score in detecting IVIG-resistant KD patients and then compare the pooled sensitivity and specificity of four Asian risk scores, Egami, Formosa, Kobayashi, and Sano risk scores.

Methods

A comprehensive search of Cochrane, Embase, and PubMed was conducted through 20 December 2021, using key terms relevant to the research question "What are the sensitivities and specificities of the four Asian predicting scores, Egami, Formosa, Kobayashi, and Sano, in Kawasaki disease patients with IVIG resistance?" The reference lists of the included studies were manually reviewed to identify pertinent references. A random-effects bivariate model was used to estimate the summary of sensitivity and specificity of the tools.

Results

We found 41 relevant studies of the four Asian risk scores that were eligible to analyze for pooled accuracy. Eleven studies involving 5,169 KD patients reported the diagnostic performance of the Formosa score for the risk of IVIG resistance. The overall performance of the Formosa score was as follows: pooled sensitivity, 0.60 [95% confidence interval (CI), 0.48-0.70]; pooled specificity, 0.59 (95% CI, 0.50-0.68); and area under the hierarchical summary receiver operating characteristic curve, 0.62. The Formosa score exhibited the highest sensitivity 0.76 (95% CI, 0.70-0.82) for detecting IVIG-resistant KD patients among the 21,389 children included in the 41 studies. In terms of specificity estimates, Formosa had the lowest specificity of 0.46 (95% CI, 0.41-0.51).

Conclusion

Patients at high risk for IVIG resistance may receive adjunctive treatment to reduce coronary lesions and thus also cardiovascular morbidity. Among all of the included studies, we found Formosa score to have the best sensitivity (0.76) but unsatisfactory specificity (0.46) for predicting IVIG resistance in Kawasaki disease. In the future, network meta-analysis should also incorporate the accuracy of the new scores after they have undergone a certain degree of validation around the world.

Systematic Review Registration

https://www.crd.york.ac.uk/PROSPERO/, PROSPERO CRD42022341410.

A simple scoring model based on machine learning predicts intravenous immunoglobulin resistance in Kawasaki disease

INTRODUCTION

In Kawasaki disease (KD), accurate prediction of intravenous immunoglobulin (IVIG) resistance is crucial to reduce a risk for developing coronary artery lesions.

OBJECTIVE

To establish a simple scoring model predicting IVIG resistance in KD patients based on the machine learning model.

METHODS

A retrospective cohort study of 1002 KD patients diagnosed at 12 facilities for 10 years, in which 22.7% were resistant to initial IVIG treatment. We performed machine learning with diverse models using 30 clinical variables at diagnosis in 801 and 201 cases for training and test datasets, respectively. SHAP was applied to identify the variables that influenced the prediction model. A scoring model was designed using the influential clinical variables based on the Shapley additive explanation results.

RESULTS

Light gradient boosting machine model accurately predicted IVIG resistance (area under the receiver operating characteristic curve (AUC), 0.78; sensitivity, 0.50; specificity, 0.88). Next, using top three influential features (days of illness at initial therapy, serum levels of C-reactive protein, and total cholesterol), we designed a simple scoring system. In spite of its simplicity, it predicted IVIG resistance (AUC, 0.72; sensitivity, 0.49; specificity, 0.82) as accurately as machine learning models. Moreover, accuracy of our scoring system with three clinical features was almost identical to that of Gunma score with seven clinical features (AUC, 0.73; sensitivity, 0.53; specificity, 0.83), a well-known logistic regression scoring model.

CONCLUSION

A simple scoring system based on the findings in machine learning seems to be a useful tool to accurately predict IVIG resistance in KD patients.

Combination of Hemoglobin-for-Age Z-Score and Plasma Hepcidin Identified as a Novel Predictor for Kawasaki Disease

Kawasaki disease (KD) is a febrile coronary vasculitis that affects younger children and includes complications such as coronary artery aneurysm. KD diagnoses are diagnosed based on clinical presentations, a process that still poses a challenge for front-line physicians. In the current study, we developed a novel predictor using the hemoglobin-for-age z-score (HbZ) and plasma hepcidin to differentiate Kawasaki disease (KD) from febrile children (FC). There were 104 FC and 115 KD subjects (89 typical KD; 26 incomplete KD) for this study, and data were collected on the biological parameters of hemoglobin and plasma hepcidin levels. A receiver operating characteristic curve (auROC), multiple logistics regression, and support vector machine analysis were all adopted to develop our prediction condition. We obtained both predictors, HbZ and plasma hepcidin, for distinguishing KD and FC. The auROC of the multivariate logistic regression of both parameters for FC and KD was 0.959 (95% confidence interval = 0.937–0.981), and the sensitivity and specificity were 85.2% and 95.9%, respectively. Furthermore, the auROC for FC and incomplete KD was 0.981, and the sensitivity and specificity were 92.3% and 95.2%, respectively. We further developed a model of support vector machine (SVM) classification with 83.3% sensitivity and 88.0% specificity in the training set, and the blind cohort performed well (78.4% sensitivity and 100% specificity). All data showed that sensitivity and specificity were 81.7% and 91.3%, respectively, by SVM. Overall, our findings demonstrate a novel predictor using a combination of HbZ and plasma hepcidin with a better discriminatory ability for differentiating from WBC and CRP between children with KD and other FC. Using this predictor can assist front-line physicians to recognize and then provide early treatment for KD.