Abstract
Coronary artery lesions (CALs) are the most common and serious complication of Kawasaki disease (KD), and the pathogenesis is unknown. Exploring KD-specific biomarkers and related risk factors is significant for clinical diagnosis and treatment. This study aimed to explore the feasibility of combining clinical indicators with S100A12/TLR2-associated signaling molecules for the predictive modeling of CALs in KD. A total of 346 patients (224 males and 122 females) with KD who visited the rheumatology department of Wuhan Children's Hospital between April 2022 and March 2025 were enrolled and divided into two groups according to the presence or absence of CALS (292 patients had CALs and 54 patients did not). Forty-one variables were collected from the two groups, including demographic characteristics, clinical manifestations, and laboratory data. Single nucleated cells from each patient were extracted, and the expression of the S100A12/TLR2 signal transduction-related molecules S100A12, TLR2, MYD88, and NF-κB were detected by real-time fluorescent quantitative polymerase chain reaction. Statistically significant variables were subjected to logistic regression analysis to determine the independent risk factors for KD with CALs, and a new risk score model was established to assess the predictive efficacy based on receiver operating characteristic curves. Sixteen variables significantly differed between the no-CALs and CALs groups: gender, fever duration, white blood cells (WBC), hemoglobin (HGB), Ce reactive protein (CRP), procalcitonin, serum ferritin (SF), erythrocyte sedimentation rate (ESR), fibrinogen (FIB), aspartate aminotransferase-to-alanine aminotransferase ratio (AST/ALT), serum albumin (ALB), sodium (Na), Interleukin (IL-10), tumor necrosis factor (TNF-α), S100 calcium binding protein A12 (S100A12), and Myeloid Differentiation Factor 88 (MYD88) (p < 0.05). After performing a univariate analysis, 12 variables (gender, fever duration, WBC, HGB, CRP, SF, ESR, FIB, AST/ALT, ALB, Na, and S100A12) were included in the multifactorial binary logistic regression, which showed that fever duration ≥ 6.5 days, ESR ≥ 46.5 mm/h, AST/ALT ≤ 1.51, and S100A12 ≥ 10.02 were independent risk factors for KD with CALs and were assigned scores of 3, 2, 1, and 2, respectively, according to the odds ratio (OR). The total score of each patient was counted, and a new prediction model for KD combined with CALs was established, where < 3.5 was considered low risk and ≥ 3.5 was regarded as high risk; the sensitivity, specificity, Jorden index, and area under the curve of this scoring system were 0.667, 0.836, 0.502, and 0.838, respectively. This new scoring model has good efficacy for predicting the occurrence of KD with CALs. The expression of S100A12 was significantly increased in the CALs group and was an independent risk factor for the occurrence of CALs, and has the potential as a biomarker for predicting KD with CALs.