Thirty-one children diagnosed with KD at CHA University Bundang Medical Center, between November 2007 and November 2008, were enrolled in this study. All enrolled patients met the following criteria for complete KD: fever and at least 4 of the following 5 clinical symptoms—rash, conjunctival injection, cervical lymphadenopathy, changes in the oral mucosa, and changes in the extremities; or fever and 3 of the above clinical symptoms plus coronary artery abnormalities (dilatation or aneurysm) documented by echocardiography⁴⁾. Children who presented with incomplete KD were excluded from our study. To exclude other febrile illnesses resembling KD, we determined the serum titers of antistreptolysin O, anti-Epstein-Barr viral, antimumps viral, and antimycoplasma antibodies in all samples from patients with KD. Furthermore, we performed multiplex polymerase chain reaction on nasopharyngeal aspirates taken from patients with KD, for common respiratory viruses (adenovirus, respiratory syncytial virus, parainfluenza virus, influenza virus, metapneumovirus, coronavirus, and rhinovirus). We also performed neck ultrasonography on these patients to rule out suppurative lymphadenitis. All patients received 2 g/kg IVIG in a single dose at the time of diagnosis and were treated with high doses of oral aspirin (80 mg/kg/day), until they became afebrile for 3 to 4 days, as specified in the 2004 American Heart Association guidelines⁴⁾. Afterwards, they were given low doses of aspirin (5 mg/kg/day). Serial echocardiography was performed on all patients with KD at the time of diagnosis and 1 month later, to study CAL development in the acute stage of KD. Echocardiography was conducted on all patients by a single cardiologist to evaluate the presence of structural heart diseases as well as to obtain M-mode echocardiographic parameters, including left ventricular end-diastolic dimensions (LVEDD), left ventricular end-systolic dimensions (LVESD), and fractional shortening (FS). We measured the internal diameter of the proximal right coronary artery (RCA) and proximal left anterior descending coronary artery (LAD) to classify patients with CAL development (CAL (+)) and those who without CAL development (CAL (−)). We did not include the measurements of the left main coronary arteries (LMCAs), as normal anatomic variations in LMCAs have been reported to be frequent, and the probability of isolated dilatation of LMCA without accompanying dilatation of LAD has been reported to be low²²⁾. The z scores of coronary arteries were obtained using RCA and LAD measurements as well as the nonlinear regression equations based on the body surface area (BSA)²²⁾. To calculate the z scores of RCA and LAD measurements, we derived the predicted values of RCA and LAD measurements for a patient with a given BSA using the following regression equations, reported by McCrindle et al.²²⁾:

Predicted value of LAD = 0.26108 × (BSA^0.37893) − 0.02852

Standard deviation of LAD = 0.01465 + (0.01996 × BSA)

Predicted value of RCA = 0.26117 × (BSA^0.39992) − 0.02756

Standard deviation of RCA = 0.02407 + (0.01597 × BSA)

The z scores were calculated by dividing the differences between the actual measurements of RCA and LAD and the corresponding predicted values by the corresponding standard deviations, as described by McCrindle et al.²²⁾.

CAL (+) patients had dilated (2.5 ≤z score<4.0) or aneurysmal (focal or diffuse dilatation of a coronary artery segment with z score ≥4.0) coronary arteries, based on the maximal internal diameters of the RCA and LAD one month after the initial diagnosis. CAL (−) patients had normal (z score <2.5) coronary arteries one month after the initial diagnosis⁴⁾.

Patients were defined as IVIG nonresponders if they had a persistent or recrudescent fever after ≥48 hours of completion of initial IVIG infusion⁵⁾.

Twenty-one children with no history of KD, autoimmune, or allergic diseases, who were admitted for treatment of common febrile illness, such as pneumonia, gastroenteritis, or urinary tract infection, were included as febrile controls (FCs).

Demographic variables followed were age, sex, duration of fever before and after the initial treatment, prevalence of CAL (+), and resistance to IVIG.

Venous blood samples were taken from FCs on admission and from patients with KD before initial IVIG treatment (i.e., on days 4 to 8 of the illness) and 7 days after the IVIG treatment. The day of onset of fever was determined as the day one of illness.

White blood cell counts and serum levels of hemoglobin, albumin, aspartate aminotransferase, alanine aminotransferase, and C-reactive protein were determined before IVIG treatment in patients with KD and FCs. Serum samples were stored at −80℃ until the analyses.

Whole blood samples were taken from FCs on admission and from patients with KD before initial IVIG treatment (i.e., on days 4 to 8 of the illness) and 7 days after IVIG treatment. Fresh 50-µL aliquots of whole blood samples in Falcon polystyrene tubes (Becton Dickinson, Lincoln Park, NJ, USA) were incubated with APC-conjugated anti-human CD14 and FITC-conjugated anti-human TLR2 monoclonal antibodies (eBioscience, San Diego, CA, USA). Isotype matched monoclonal antibodies (eBioscience) were used as controls. After the incubation, red blood cells were lysed, washed with cold phosphate-buffered saline, and fixed with 1% paraformaldehyde. Total 50,000 cells were counted from each sample, and after appropriate gating, monocyte population was identified as CD14⁺ cells in the scatter diagrams. FTLR2% was analyzed by BD FACScan (Immunocytometry Systems, San Jose, CA, USA), using CellQuest software.

IL-10 serum concentrations were determined using enzyme-linked immunosorbent assay kits (R&D Systems Co., Minneapolis, MN, USA). Briefly, we used the quantitative sandwich enzyme immunoassay technique. To a microplate precoated with human IL-0-specific monoclonal antibodies, a buffered protein base, standard, control, and samples were added per well. After incubation and washing with wash buffer, human IL-10 conjugate was added and incubated. After another washing, substrate solution was added, and when the color in the wells turned from blue to yellow, the optical density of the wells was determined using a microplate reader (R&D Systems Co.), according to the manufacturer's instructions.

All procedures involving human participants performed in this study were in accordance with the ethical standards of the Institutional Review Board of CHA University Bundang Medical Center as well as the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The parents or guardians of all children participating in this study provided written informed consent, which was approved by the Ethics Committee of the CHA University Bundang Medical Center.

Values are expressed as mean±standard deviation, or where appropriate, as percentage. Comparisons between variables were made using the Mann-Whitney U test for continuous variables and the chi-square test for categorical variables. Correlation analysis was performed using the Spearman coefficient. To determine the cutoff value of FTLR2% in predicting CAL development and IVIG resistance, receiver operating characteristic (ROC) curve analysis was performed. All statistical analyses were done using IBM SPSS Statistics ver. 20.0 (IBM Co., Armonk, NY, USA). A P value of <0.05 was considered statistically significant.

We studied the clinical characteristics of 31 patients with KD and 21 age-matched FCs (Tables 1, 2). In patients with KD, the day of onset of fever was designated as the day one of illness.

Eleven patients with KD developed CALs 1 month after the initial diagnosis and 10 patients were IVIG nonresponders. Patients with CALs at the time of initial diagnosis, but whose coronary z scores normalized after 1 month, were considered to have transient lesions and were categorized under the CAL (−) group, according to the method previously proposed by Yoshimura et al.²⁾ in their study correlating N-terminal pro-brain natriuretic peptide, CALs, and IVIG resistance in the acute stage of KD. The percentage of IVIG nonresponders did not significantly differ between the CAL (+) and CAL (−) groups (3 out of 11 patients, 27.3%, versus 7 out of 20 patients, 35%, respectively). The percentage of CAL (+) patients was not significantly different between IVIG nonresponders and IVIG responders (3 out of 10, 30%, versus 8 out of 21, 38.1%, respectively).

Echocardiography of all patients with KD revealed no significant structural heart defects, except for a small, hemodynamically insignificant patent foramen ovale in one patient. Small amount of mitral regurgitation was noticed in four patients with KD (3 in the CAL (+) group and 1 in the CAL (−) group), but they were considered hemodynamically insignificant.

No significant difference in LVEDD, LVESD, and FS was found between the CAL (+) and CAL (−) groups, and between the IVIG nonresponders and IVIG responders (Table 3).

Of the parameters previously determined to be correlated with the increased risk of developing CALs (i.e., age, duration of fever, white blood cell count, serum albumin, and C-reactive protein)⁶⁾, none were statistically different between the CAL (+) and CAL (−) groups.

As predicted, duration of fever after IVIG treatment was significantly longer in the IVIG nonresponders compared to that in the IVIG responders. Among the parameters that have previously been found to correlate with the risk of IVIG resistance^5,7,8), the levels of serum hemoglobin and alanine aminotransferase were significantly higher in the IVIG nonresponders compared with those in the IVIG responders.

TLR2 expression in representative cases of CAL (+), CAL (−), and FC groups is shown in Fig. 1. Prior to the initial treatment, FTLR2% was significantly increased in the CAL (+) group compared to that in the CAL (−) group (93.27%±8.34% vs. 79.69%± 16.04%, respectively; P<0.05) (Table 4).

FTLR2% before the initial treatment was usually increased in IVIG nonresponders compared to that in the IVIG responders, but there was no statistically significant difference between the two groups (91.88%±5.81% vs. 80.30%±17.11%, respectively; P>0.05) (Table 4).

As shown in Table 4, there were no significant differences in sIL-10 before the initial treatment, between patients with KD and the FC group or between the CAL (+) and CAL (−) groups of patients with KD. sIL-10 before the initial treatment was usually higher in IVIG nonresponders compared with that in the IVIG responders, but the results were not statistically significant. Additionally, FTLR2% in patients with KD did not correlate significantly with sIL-10.

FTLR2% and sIL-10 in patients with KD did not correlate significantly with previously determined risk factors of CAL development^6,23) or with the known predictors of IVIG resistance^7,8). Additionally, FTLR2% and sIL-10 in these patients did not correlate significantly with echocardiographic variables such as LVEDD, LVESD, or FS.

The cutoff values of FTLR2% and sIL-10 for predicting CAL (+) were obtained from the ROC curve, and the areas under the curve were 0.779 and 0.558, respectively. An FTLR2% cutoff value of 92.62% predicted CAL (+) with an 80% sensitivity and 68.4% specificity, whereas a sIL-10 cutoff value of 15.52 pg/mL predicted CAL (+) with a 40% sensitivity and 73.7% specificity.

The areas under the ROC curves, for identifying IVIG nonresponders using FTLR2% and sIL-10, were 0.746 and 0.841, respectively. An FTLR2% cutoff value of 94.61% predicted IVIG resistance with a 66.7% sensitivity and 71.4% specificity, whereas a sIL-10 cutoff value of 15.52 pg/mL predicted IVIG resistance with a 66.7% sensitivity and 76.2% specificity.