Patients with KD were retrospectively recruited at Chengdu Women's and Children's Central Hospital between January 2017 and December 2019. The diagnosis of KD (including incomplete KD) was made according to the 2017 AHA guideline1 and confirmed by two experienced KD specialists.

The inclusion criteria: (1) patients with an initial diagnosis of KD, (2) <18 years old, (3) received standard treatment with IVIG 2 g/kg as a single infusion in the acute phase. The authors excluded patients who met the exclusion criteria: (1) patients with severe infection, allergy, or systemic juvenile idiopathic arthritis, or (2) recurrent KD, or (3) who refused treatment with IVIG or received an initial IVIG dose of less than 2 g/kg, or (4) who had received IVIG treatment in the first 3 months after enrollment, (5) patients with unavailable clinical or laboratory data.

All patients were treated with 2 g/kg IVIG as a single infusion within 12 h of KD diagnosis and received aspirin at a dose of 30–50 mg/kg/day during the acute phase. Aspirin (3–5 mg/kg/day) was administered for 6–8 weeks until all signs of inflammation had resolved or regression of CAL was observed on two-dimensional echocardiography. Additional IVIG (2 g/kg) was administered if patients had a persistent fever for more than 36 h or had recurrent fever associated with KD symptoms after an afebrile phase. Other therapies, including prednisolone, were not used in the initial treatment. IVIG resistance was defined as recurrent or persistent fever for at least 36 h but no longer than 7 days after initial IVIG treatment.1

All data were obtained from medical records, including demographic data, laboratory data, and clinical outcomes. Laboratory data were collected before the initial IVIG treatment. Coronary artery diameter was measured by echocardiography before the first IVIG treatment.

Patients were divided into two groups according to whether they were complicated with CALs in the acute KD phase: the CALs group and the non-CALs group. CALs were defined based on the normalization of body surface area dimensions as Z-scores as follows according to the 2017 AHA guideline1: (1) no involvement (Z-score <2.0); (2) dilation (Z-score ≥2.0 to <2.5), (3) aneurysm (Z-score ≥2.5 to <5 for a small aneurysm; Z-score ≥5 to <10 for a medium aneurysm; Z-score ≥10 for a giant aneurysm) of the coronary arteries, based on the maximum internal diameters of the right coronary artery, left anterior descending artery and left circumflex coronary artery.1,17 All KD patients underwent standardized echocardiograms by an experienced ultrasonographer during the acute phase.

The normality of the distribution of variables was checked using the Shapiro–Wilk test. Continuous variables were expressed as means ± standard deviations or median and IQR (25th, 75th percentiles) if non-normally distributed. Categorical variables were expressed by presenting the frequency and proportion in each category. Chi-square analysis was applied for categorical variables, and the Student-t-test or Mann–Whitney U test was used based on the normality of variables for continuous variables.

The univariable and multivariable logistic regression models were constructed, and results were expressed as odds ratios (OR) with a 95% confidence interval (CI) to identify independent risk factors for CALs. To assess the discriminatory capacity of CRP in predicting CALs, the authors performed the receiver operating characteristic (ROC) curve analysis. The authors performed subgroup analysis by dividing the CALs group into several subgroups based on CRP and hemoglobin (HB) according to Z-score. The value of independent risk factors at different Z-score levels is using ANOVA or the Kruskal–Wallis test based on the normality of distribution. The chi-square test was also applied to compare the ratio of the different Z-score groups in each independent risk factor, and the Bonferroni test was used for further pairwise tests. A p-value <0.05 was considered statistically significant. All statistical tests were performed using SPSS 12.0 for Windows XP (SPSS, Inc, Chicago, USA).

A total of 851 KD patients were included, 60.28% of whom were men (n = 513). Of the 851 cases, 206 had CALs (Z ≥ 2), accounting for 24.21%. The general characteristics and laboratory indicators of patients with and without CALs are shown in Table 1. There were no significant differences in white blood cell (WBC), platelet, neutrophils (%), lymphocyte (%), alanine transaminase (AST), and erythrocyte sedimentation rate (ESR) between these two groups.

The CALs group had a longer hospitalization time and younger age than the non-CALs group (p < 0.05). In addition, children in the CALs group had a higher proportion of male gender, incomplete KD, and IVIG resistance than children in the non-CALs group (all p < 0.05). Of the routine laboratory inflammatory indicators, CRP and ALT levels were significantly higher in the CALs group than in the non-CALs group (p < 0.05).

The univariable and multivariable logistic regression models were constructed to identify independent risk factors for CALs. The result is shown in Table 2. The multivariable logistic regression analysis showed that incomplete KD, male, lower hemoglobin (HB), and CRP were independent risk factors for CALs in KD patients (all p < 0.05). The results of ROC curve analysis for CRP to predict CALs are shown in Figure 2; the area under the ROC curve (AUC) was 0.588 (95% CI = 0.554–0.622, p < 0.001). The sensitivity and specificity in predicting CALs were 47.57% and 69.61%, respectively, with a cut-off level of 105.5 mg/L determined by the Youden index. For HB, the sensitivity and specificity in predicting CALs were 38.83% and 70.08%, respectively, with a cut-off value of 102 g/L.

Figure 2.

The ROC curve analysis for the prediction of CALs. Note: CRP, C-reactive protein; HB, hemoglobin.

(0.14MB).

The authors also performed subgroup analysis by dividing the CALs group into multiple groups by Z score to compare CRP differences among subgroups with different degrees of CALs. As shown in Appendix S1 and Appendix S2, there was no significant difference in CRP or HB between the subgroups with different degrees of CALs.

The authors use the cut-off value determined by Youden's index to compare the incidence of CALs in KD patients with different levels of CRP or HB. There were 556 patients with a CRP level of <105.5 mg/L, 108 of whom developed CALs (32 dilations and 76 CAA). There were 295 patients with a CRP level ≥105.5 mg/L, and 98 of them had CALs (26 dilations and 72 CAA). Meanwhile, there were 80 KD patients with CALs in the lower HB group (<102 g/L) and 126 KD patients with CALs in the higher HB group (≥102 g/L). KD children with a higher CRP level (≥105.5 mg/L) or a lower HB value (<102 g/L) had a higher incidence of CALs than children with a lower CRP (<105.5 mg/L) (33% vs. 19%, p < 0.001) or higher HB (29% vs. 22%, p < 0.05). Further Bonferroni test showed that KD patients with a higher CRP level (≥105.5 mg/L) or a lower HB level (<102 g/L) had a higher incidence of small aneurysms than children with a lower CRP (<105.5 mg/L) or higher HB (≥102 g/L) (p < 0.05) (Appendix S3).

There were 122 KD patients with a higher level of CRP (105.5 mg/L) and lower HB level (<102 g/L), 43 of whom developed CALs. There were 729 patients with a CRP level ≥ of 105.5 mg/L, and 163 had CALs. Similarly, KD patients with a higher level of CRP and lower HB had a higher incidence of CALs than children with a lower CRP and higher HB (35% vs. 22%, p = 0.01) (Appendix S4).

This study has several limitations. First, the retrospective design limits the present research, which inevitably leads to selection and information bias. Second, the authors could not evaluate inflammatory markers other than CRP and WBC. Third, all KD patients were Chinese, which limits the generalizability of the results. Therefore, further prospective studies are needed to confirm the present results.