Neonatal and Pediatric Medicine
Open Access

Kawasaki disease (KD; also called mucocutaneous lymph node syndrome and Kawasaki syndrome) is an acute autoimmune systemic vasculitis disease that mainly occurs in infancy and younger children between 6 months to 5 years of age. The acute illness is self- limited and characterized by fever>5 days and at least 4 of the followings; bilateral non exudative bulbar conjunctivitis, changes in lips and oral mucosa, polymorphous rash, cervical adenopathy (greater than 1.5 cm), and changes in peripheral extremities, including swollen hands and feet, red and edematous palms and soles, and later subungual peeling. The definitive etiology remains unclear but some studies found some genetic factors and familial predisposition to KD [1].

Respectively while the incidence of KD in Thailand ranged from 2.14 to 3.43 cases per 100,000 children under 5 years and has become the leading cause of acquired heart disease among children in Thailand [2].

The autopsy studies of fatal cases have clearly demonstrated that inflammation occurs in multiple organs and tissues in KD, although the coronary artery lesions (CALs) are the most clinically significant aspect of the illness includes coronary artery dilatation, coronary aneurysm, coronary fistula, myocardial infarction, and thus sudden dead. The respiratory, gastrointestinal, dermatologic, urinary, nervous and lymphoreticular systems can all be involved [3].

The risk of CALs develop in approximately 20%-25%in children with KD who are not early treatment with high- dose intravenous immunoglobulin (IVIG) and high-dose aspirin, the appropriate treatment in patients who respond well to IVIG have the better chance to decrease the risk of CALs to 3%-5% [4].

The early diagnosis and treatment with IVIG within 10 days of fever onset is the crucial management in patients with KD. Based upon prospective, multicenter treatment trials of IVIG optimal therapy is 2 g/kg with high- dose aspirin (80-100 mg/kg/day) as soon as possible after diagnosis during the acute febrile phase of illness, followed by low-dose aspirin (3-5 mg/kg/day) until follow-up echocardiograms indicate a lack of coronary abnormalities are the current standard treatment in KD [5]. Approximately 10%-20% of patients with KD were IVIG- resistant or non- responder and defined as the KD patients with persistent or recurrent fever at least 36 hours after the end of IVIG infusion. Furthermore, many studies have shown that these patients are at increased risk of developing CALs [5].

The previously classic risk factors from three Japanese risk-scoring systems in predicting outcomes of KD such as Kobayashi, Egami and Sano risk scores include young ages, high C-reactive protein, high peripheral neutrophil percentage, high serum alanine transaminase (ALT), high bilirubin level, low blood hemoglobin, low serum albumin and low serum sodium are widely accepted to predict IVIG resistance however the usefulness, effectiveness and accuracy for prediction IVIG resistance among patients with KD remains challenging in Thailand. The main objective of our study was to develop and validate the predictive factors for IVIG-resistant KD in Thailand.

Study population

A single-center-based retrospective study, reviewed 102 consecutive pediatric patients under 15 years of age who were hospitalized and treated in department of pediatrics at Bhumibol Adulyadej Hospital between January 2010 and December 2020 with newly diagnosis of KD according to the American Heart Association (AHA) guidelines. The study protocol was reviewed and approved by the Ethics Committee of Bhumibol Adulyadej Hospital Affiliated, and with its approval. Sample size was calculated by using single group proportion with 0.1 alpha (α) error. There are at least 68 cases a good for statistical significant. The protocol was approved by the ethic review committee (IRB 76/63).

All patients were receiving initially high-dose IVIG (2 g/kg) as a single 8-12 hours infusion according to the AHA recommendation combined with high- dose aspirin (80-100 mg/kg/day) every 6 hours. After 36 hours at the end of the first dose of IVIG infusion, the patients were grouped as IVIG- responder or IVIG- resistant (non-responder) KD. For a refractory case to the initial treatment with the 1st dose of IVIG, repeated dose of IVIG were provided and methylprednisolone pulse therapy or infliximab were provided after second-dose of highdose IVIG in patients with re-fever.

Data collection

Data referring to demographic characteristics, clinical manifestations, laboratory examination and echocardiographic results were obtained including age in month at the time of diagnosis, gender, number of days with fever prior to initial treatment with IVIG, presence and degree of CALs, and ineluctable related clinical manifestations.

Two- dimensional echocardiographic data was obtained at the time of diagnosis prior to initial treatment and at 6 weeks and 6 months after treatment. All dimensions of coronary arteries were obtained by a standardized protocol and performed by a single pediatric cardiologist.

Presence of CALs was defined as coronary artery diameter more than 2.5 mm. in patients aged less than 3 years old, more than 3. 0 mm. in patients aged 3 and 9 years old and more than 3.5 mm in patients older than 9 years old. As for the degree of CALs (D- CALs), localized dilatation with internal diameter (ID) less than 4 mm the dilatation with ID between 4 mm and 8 mm and the dilatation with internal diameter more than 8 mm were defined as slight CALs, moderate CALs and severe CALs, respectively.

Laboratory data were obtained, including hemoglobin (Hb), Hematocrit (Hct), white blood cell (WBC) count and morphology, percentage of lymphocyte (LYM) count, percentage of neutrophil (PMN) count, percentage of monocyte and platelet (PLT) counts, serum albumin, serum alanine aminotransferase (ALT), serum aspartate aminotransferase (AST), serum total and direct bilirubin, serum blood urea nitrogen (BUN) and creatinine, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), serum sodium, serum potassium and serum chloride. All those laboratory variables were routinely obtained in our clinical practice and recorded at the time of diagnosis.

For using the previous models, we kept the predictors in those models and re-estimate the coefficients with the current data set to build the models. Predictive ability of the new model was compared with the previous models including [6-8] scoring systems from Japan and Yang prediction model from China.

Statistical analysis

SPSS statistical software package, version 27. OJ (SPSS, Chicago, USA) was used for statistical analyses. Continuous variables were summarized using mean and standard deviation (mean ± SD) or using frequencies (%) of patients to describe categorical variables. Student’s t-test and Mann-Whitney U tests for continuous data and qui square test for categorical data were used to compare variables between the 2 groups. For all analyses a two- sided probability value below 0.05 (p-value< 0.05) was considered to indicate statistical significance. Results and significance values are summarized in the relevant tables.

Variables selected by the univariate analysis with p-value< 0.05 were applied for multivariate logistic regression to explore independent risk factors for IVIG resistance. To construct the scoring system, the score of independent risk factors were determined by the odd ratios (OR). The cut-off point was selected by the receiver-operator characteristic (ROC) curves and the sensitivity and specificity of the scoring system were analyzed.

Demographic and clinical manifestations

A total 102 consecutive pediatric patients diagnosed with Kawasaki disease was excluded due to lack of data for 5 patients. A total of 97 patients which were 60 male patients and 37 female patients (maleto- female ratio 1.62) met the inclusion criteria according to the AHA guidelines and were enrolled into our study (Figure 1). The age ranged was 1 to 144 months with a median of 27.3 months. Eighty (80) patients occurred in IVIG-responder group (80/97, 71.4%) included 46 male patients (46/80, 57.5%) and 34 female patients (34/80, 42.5%) with a median of 28 months. Thirty-four (34) patients were diagnosed with incomplete KD (35%). Seventeen (17) patients occurred in IVIGresistant group (17/97, 28.5%) included 14 male patients (14/17, 82.4%) and 3 female patients (3/17, 17.6%) with a median of 21 months. No case was recurrent KD after at least 3 years follow-up including no history of KD in siblings (Table 1).

Table 1: Baseline demographic characteristics among responder (n=80) and resistant (n=17) to intravascular immunoglobulin (IVIG).

Figure 1: Shown 102 patients was diagnosed Kawasaki disease, 5cases with missing data was excluded from our study, Remained 97 patient with IVIG resistant consist 17 patients and IVIG responders consist 80 patients.

Univariate analysis

According to the univariate analysis (Table 2), 4 variables were selected and weighted proportionally to their respective odds ratio for IVIG-resistant. Two variables were significantly lower in IVIG-resistant group than IVIG-responder group, including serum albumin (3.01 ± 0.51 vs. 3.48 ± 0.49 g/dL, p-value=0.01) and serum sodium (133.24 ± 3.43 vs. 135.79 ± 3.19 m Eq/L, p-value=0.01) meanwhile age at time of diagnosis less than 6 months (29.4% vs. 8.75%, p-value=0.02) and serum ALT (78.29 ± 74.19 vs. 71.01 ± 80.19 IU/L, p-value=0.02) were significantly higher in IVIG-resistant group than IVIG-responder group.

Table 2: Baseline characteristics of participants in our study.

Mean age of IVIG-responder group was higher than IVIG-resistant group but no statistical significance (21 ± 26.16 vs. 28.56 ± 24.98 months, p-value=0.30) as shown in Table 1.

There was no difference of gender among responder and resistant groups. The remaining predictive values among both groups were comparable as shown in Table 1.

Multivariate analysis

We explored multivariate predictive factors for IVIG resistance in our data (Table 3). All study subjects were analyzed, the results indicated that independent risk factors for IVIG resistance include only 3 variables; high total peripheral WBC count>18,000 mm³ (OR=2.34, p-value=0.117 and c-statistic 1.35), high serum ALT>45 IU/L (OR=3.97, p-value=0.025 and c-statistic 1.76) and low serum albumin

Table 3: Risk factors for predicting IVIG-resistant by multivariate analysis.

Scoring system for IVIG resistant prediction

To construct the scoring system (Tables 4-6), both of total peripheral WBC count (>18,000 mm³) and serum albumin (45 IU/L) was given 1.5 points. The total scores were calculated for each KD patient. ROC analysis indicated that the area under the curve (AUC) was 0.79 with 95% CI, 0.69-0.89; p-value

Table 4: Risk scoring for predict IVIG resistant group.

KD is an acute autoimmune systemic vasculitis disease that commonly affects the small and medium-size arteries, especially coronary arteries. CALs are the most serious complications of KD, which occur in approximately 20-25% in children who are not early treatment with IVIG [1]. In our study, the frequency of developing CALs was 29% in IVIG resistant group.

Some previous studies shown CALs at onset of acute phase developed more often in IVIG resistant group [2]. All our patients received IVIG administration as initial treatment the relationship between IVIG resistant and CALs in our study at prior treatment has significant association but no association at 2nd(6 weeks) and 3rd times (6 months) of echocardiogram follow-up after initial treatment, this might be considering that nonresponse may reflect only severity of ongoing inflammation in acute phase whereas development of CAAs might be affected by both ongoing inflammation and hemodynamics after initiation of treatment [9].

Over 80% of occurs in infancy and younger children between 6 months to 4 years of age, rarely reported in adolescents and adults [10]. In our study shown the mean age of 27.3 months and KD patients with age less than 6 months occur only 13.4% (13/97) suggests that most infants with KD are protected by transplacental maternal immunity.7 In addition, we also found KD patients with IVIG resistant group was younger than IVIG responder group and the frequency of KD patients with age younger than 6 months were higher in IVIG resistant group than IVIG responder group significantly [11].

Many previous studies shown evidence of low serum sodium, low serum albumin and high serum ALT all strongly predicted IVIG unresponsiveness while in our study showed only low serum albumin and high serum ALT.

Nakamura study noted that hyponatremia might be the most useful to predict giant coronary aneurysms (diameter 8 mm and higher) caused by KD. The relationship between hyponatremia and severity of KD might involve permeability of the endothelium, dehydration, duration of fever and severity of illness but pathogenesis of hyponatremia in KD remains clearly understood, possible mechanism including the syndrome of inappropriate antidiuretic syndrome (SIADH), hyponatremic dehydration and hypo-osmolar fluid intake [12].

Kobayashi scoring system is now widely used among Japanese populations to predictthe risk of IVIG resistance but unfortunately it did not accurately predict outcomes in non-Japanese populations. Hepatic impairment was frequent found in the patients with KD. This phenomenon caused by activated cytokine-natural killer cells during acute phase of inflammatory process in liver sinusoids caused hepatocytic injury in KD [6] Both Egami and Kobayashi scoring systems; suggest that high level of serum ALT was the predictor of severity of KD. The cut-off points of Egami and Kobayashi risk scoring were more than 80 and 100 IU/L, respectively, while our study, the cutoff point was only equal and higher than 45 IU/L which is less than previous studies.

In our study showed that hypoalbuminemia (serum albumin<3.5 g/dl) isthe predictor for IVIG resistance, this can explained by the pathophysiology of KD in acute phase caused by increased permeability of the vessels due to inflammatory mediators so hypoalbuminemia reflect the more severe leakage of plasma protein and from hepatic impairment reduced the synthesis of albumin because cytokineactivated natural killer cells accumulate at inflammatory lesions in the acute phase of KD, these activated natural killer cells might accumulate in vascular endothelium and along liver sinusoids where they would be likely to participate in hepatocytic injury. IVIG resistance patients tend to have more severe vascular leakage and liver impairment according to lower serum albumin. Nowadays in Japan, low albumin level is used as one of the risk factors in selection of candidates for IVIG treatment of KD [13].

Principal findings of our study to develop and validate new prediction values were that IVIG resistant could be stratified to lower the risk of progress sequel of CAAs (Table 5). Nowadays, there were several prediction models for IVIG resistant, the risk factors including age below 6 months; IVIG treatment within 4 days of illness; abnormal the echocardiographic results prior treatment; high C-reactive protein, high serum ALT, high serum AST, high percentage of neutrophils, low serum sodium and low serum hemoglobin [14-17]. Our predictive model is the scoring system for predicting IVIG-resistant patients in Thai population, the scoring system includes total peripheral WBC count>18,000 cu.mm. (1 point), serum albumin<3.5 g/dl (1 point) and serum ALT>45 IU/L (1.5 points), and a total score equal 2 points and higher yielded a sensitivity and a specificity of 82.35% and 62.50%, respectively for IVIG resistance prediction (Table 5 and Figure 2).

Table 5: Risk scoring differences among the clinical studies.

Figure 2: ROC and AUC of the prediction models for IVIG-resistance. The new predictive model for IVIG resistance show an AUC of 0.79.

Table 6: Predictive validity for resistant to intravascular immunoglobulin (IVIG).

Study limitation

Five cases with missing file data or with a diagnosis other than complete KD were excluded from our study at the beginning.

In conclusion, the predictive values that we note similar to previous studies in Japan, a risk scoring system based on demographic and laboratory characteristics prior to administrated IVIG can accurately prediction the high-risk patients of IVIG resistant, who require more close monitoring which may help physicians make more rational decisions regarding an initial treatment of KD and be the candidates for additional therapies.

The authors would like to thank to (1) Gp.Capt. Napaporn Jiraphongsa M.D., M.Sc., Division of Clinical Epidemiology, Department of Pediatrics, Bhumibol Adulyadej Hospital, Bangkok, Thailand (2) Gp.Capt. Sasawan Chinratanapisit M.D. Ph.D., Division of Allergy & Immunology, Department of Pediatrics, Bhumibol Adulyadej Hospital, Bangkok, Thailand. (3) Ms. Pangpawinee Saybung and (4) Ms. Suwittra Huanraluck for help on statistical analysis of the data. The authors acknowledged the help of Assoc. Prof.Wg. Cdr. Komsun Suwannaruk M.D. and Assoc.Prof. Dr. Kornkarn Bhamarapravatana Ph.D. Department of Preclinical Sciences, Faculty of Medicine, Thammasat University, Pathumthani, Thailand for their assistance in data analysis and manuscript preparation.

The authors declare no conflict of interest.