Kawasaki disease (KD) is a systemic vasculitis involving multiple organs, and it can cause coronary artery lesions (CALs), carditis, hepatitis, arthritis, and central nervous system diseases [1]. CALs are one of the most serious complications associated with KD. Characteristic clinical symptoms of KD are prolonged fever, bilateral conjunctival injection, cervical lymphadenopathy, erythematous induration of palms and soles, mucocutaneous changes in oropharynx and lips, and polymorphous skin rashes [1,2].

There have been many reports showing that patients with KD had leukocytosis and elevated erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels, which were associated with CALs [3,4]. However, there have been a few reports on the significance of hyponatremia in patients with KD [5-9]. The previous studies were also conducted in a small number of patients, rarely analyzed the correlations between serum sodium levels and other inflammatory parameters of KD, or were not analyzed by multivariate analysis [5-9]. The aims of the present study were to investigate the risk factors for the development of CALs and determine whether hyponatremia could be associated with CALs in Korean children with KD.

We retrospectively analyzed the data of 105 children (37 girls and 68 boys; mean age, 2.35 ± 2.1 years) with KD admitted to National Health Insurance System (NHIS) Ilsan Hospital between January 2000 and July 2011. The children were divided into the following groups: KD children with and without hyponatremia, and KD children with and without CALs.

Medical charts were reviewed for clinical characteristics, including patient age, sex, duration of fever, and presence of CALs during the clinical course. Laboratory data on admission included complete blood cell count (CBC), ESR, CRP, serum sodium, potassium, chloride, total carbon dioxide (tCO2), blood urea nitrogen (BUN), creatinine, total protein, albumin, cholesterol, uric acid, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), total bilirubin, calcium, phosphorus, creatine kinase (CK), CK-MB, and urinalysis.

Serum sodium levels were measured with the Ion selective electrode (ISE) method using a Hitachi 747 chemistry analyzer (Hitachi Inc., Tokyo, Japan) and Unicel DxC 800 (Beacman Coulter Inc., Brea, CA, USA). Complete blood counts including platelet counts were immediately analyzed within 2 hour by Sysmex XE-2100 (Sysmex Corp., Kobe, Japan). CRP levels were measured with the latex-enhanced turbidimetric assay method using a Hitachi 747 chemistry analyzer and Unicel DxC 800. ESRs were measured using Sysmex VES-Matic cube. Strict quality control procedures were adopted. Echocardiography was performed in all the children with KD to detect CALs.

KD was diagnosed if children had a fever (temperature, >38°C) for at least 5 days in addition to at least four of the following: (1) changes in the mucous membranes of the upper respiratory tract, including injected pharynx, injected or fissured lips and strawberry tongue, (2) bilateral conjunctival injections, (3) cervical lymphadenopathy, (4) polymorphous rash, and (5) changes in the extremities, including peripheral edema or erythema, and periungual desquamation [1,2].

Hyponatremia was defined as a serum sodium concentration (Na+) ≤135 mEq/L. CALs were defined as either an internal diameter of the coronary artery lumen >3 mm in a child <5 years of age or >4 mm in a child ≥5 years of age, the internal diameter of a segment being at least 1.5 times larger than that of an adjacent segment, or the presence of a clearly irregular lumen [10,11]. Intravenous immunoglobulin (IVIG) was defined when additional rescue therapies were required owing to persistent or recrudescent fever (≥38.0 or 100.4℉) at least 48 hours after the end of initial IVIG infusion [12].

All data were analyzed using Statistical Package for the Social Sciences (SPSS) for Windows, version 20.0 (SPSS Inc., Chicago, IL, USA). The data were expressed as mean ± standard deviation (SD). Univariate analysis was performed using the independent t-test and chi-square test. Correlation analysis was performed to determine the relationship between two variables using Pearson correlation. A comparison between two paired groups was performed using the paired t-test. Multiple logistic regression analyses were performed to identify the independent risk factors for CALs in children with KD. Statistical significance was set at a P-value of <0.05.

This study design and the use of patients’ information stored in the hospital database were approved by the Institutional Review Board (IRB) at NHIS Ilsan Hospital. We were given exemption from getting informed consents by the IRB because the present study was a retrospective study and personal identifiers were completely removed and the data were analyzed anonymously. Our study was conducted according to the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Clinical characteristics were not significantly different between KD children with CALs and those without (Table 1). However, IVIG-resistance was significantly higher (32.4% vs. 8.8%, P = 0.003) in KD children with CALs than those without. ESR levels were higher (50.3 ± 39.4 mm/hr vs. 24.8 ± 18.8 mm/hr, P = 0.013), sodium levels were lower (134 ± 2.8 mmol/L vs 135 ± 3.0 mmol/L, P = 0.027), tCO2 levels were lower (18.2 ± 3.5 mmol/L vs. 19.4 ± 2.7 mmol/L, P = 0.041) and total bilirubin levels were higher (1.5 ± 1.4 mg/dL vs. 0.9 ± 1.1 mg/dL, P = 0.017) in KD children with CALs than those without (Table 2).

Clinical characteristics were not significantly different between KD children with hyponatremia and those without (Table 3). However, CALs were significantly higher (47.8% vs 28.4%, P = 0.034) in KD children with hyponatremia than those without. In KD children with hyponatremia, white blood cell (WBC) counts were higher (15,355 ± 6,159/μL vs. 12,335 ± 4,794/μL, P = 0.006), neutrophil counts were higher (72.6 ± 13.8% vs. 63.2 ± 16.8%, P = 0.003), lymphocyte counts were lower (20.2 ± 10.5% vs. 27.7 ± 14.8%, P = 0.005), neutrophil-lymphocyte ratio were higher (5.1 ± 4.1 vs. 3.6 ± 3.2, P = 0.048), chloride levels were lower (100 ± 3.8 mmol/L vs. 103 ± 3.6 mmol/L, P = 0.001), tCO2 levels were lower (18.1 ± 3.1 mmol/L vs. 19.6 ± 2.9 mmol/L, P = 0.013), albumin levels were lower (3.3 ± 0.5 g/dL vs. 3.6 ± 0.6 g/dL, P = 0.009), uric acid levels were higher (4.3 ± 1.8 mg/dL vs. 3.7 ± 1.3 mg/dL, P = 0.037), and phosphorus levels were lower (3.6 ± 0.9 mg/dL vs. 4.3 ± 1.0 mg/dL, P = 0.001) than those without hyponatremia (Table 4).

On multiple logistic regression analysis, hyponatremia (Odd ration (OR) = 3.148, 95% confidence interval (CI) 1.161-8.536, P = 0.024) and IVIG–resistance (Odd ratio = 4.816, 95% CI 1.234-18.797, P = 0.024) were independent risk factors for CALs in children with KD. Also, low serum chloride levels (Odd ratio = 1.246, 95% CI 1.049-1.479, P = 0.012), low serum tCO2 levels (Odd ratio = 1.311, 95% CI 1.040-1.653, P = 0.022) and low serum phosphorus levels (Odd ratio = 2.670, 95% CI 1.241-5.747, P = 0.012) were independent risk factors for the development of hyponatremia in children with KD (Table 5).

Furthermore, serum sodium levels were correlated negatively with WBC counts (r = -0.276, P = 0.004), neutrophil counts (r = -0.414, P < 0.001), lymphocyte counts (r = 0.398, P <0.001), neutrophil-lymphocyte ratio (r = -0.348, P <0.001), total bilirubin levels (r = -0.280, P = 0.005) and positively with albumin levels (r = 0.255, P= 0.009) (Table 6). However, IVIG–resistance was not correlated with WBC counts (P = 0.287), ESR levels (P = 0.160), CRP levels (P = 0.268), total bilirubin levels (P = 0.999) and albumin levels (P = 0.230).

Changes in serum sodium levels (Na levels at the convalescent phase – Na levels at the acute phase, ΔNa) correlated positively with changes in lymphocyte counts (Δlymphocyte) (r = 0.236, P = 0.019), chloride (Δchloride) (r = 0.517, P <0.001), tCO2 levels (ΔtCO2) (r = 0.227, P = 0.020) and negatively with neutrophil counts (Δneutrophil) (r = 0.230, P = 0.022), neutrophil-lymphocyte ratio (Δneutrophil-lymphocyte ratio) (r = - 0.263, P = 0.019), AST levels (ΔAST) (r = - 0.271, P = 0.018) (Table 7). Changes in serum sodium levels were negatively correlated with changes in inflammatory markers.

Our study is the first study which showed that hyponatremia was an independent risk factor for the development of CALs in KD using multivariate analyses and it was also closely associated with other well known inflammatory markers of KD such as WBC count, ESR, and serum albumin level. Although WBC count, CRP, low serum albumin levels, or elevated serum ALT were reported to be associated with the development of CALs in KD [13,14], these biomarkers were not statistically significant between children with KD and those without KD in our study. Although previous studies exhibited that increased CRP and total bilirubin on admission were indicated as significant predictors for IVIG resistance and serum albumin levels before IVIG treatment were also suggested to be a useful predictor of IVIG resistance in patients with KD [15,16], we did not find any significant difference between the two groups. These contrasting results may be due to different ethnic groups and disease severities among studies. We also intended to investigate the independent risk factors for the development of hyponatremia in KD and found that low serum chloride, tCO2 levels, and hypophosphatemia were independent risk factors for hyponatremia, which has not been analyzed and reported previously.

Although the pathogenesis of hyponatremia in KD still remains elusive, some plausible explanations have been proposed. Firstly, as we previously hypothesized, hyponatremia can be a result of ADH excess caused by inflammatory cytokines such as interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α [17]. IL-1β and IL-6 have been known to activate the magnocellular ADH-secreting neurons, leading to SIADH [18]. Activation of CD8+ T-cells and macrophages are thought to play a central role in the development of hyponatremia, which results from elevated concentrations of IL-1, IL-6, IL-8, and TNF-α secreted by the CD8+ T-cells1). Secondly, these inflammatory mediators can cause hyponatremia by inhibiting the function of the apical epithelial sodium channel and/or sodium potassium adenosine triphosphatase at the basolateral membrane of renal epithelial cells [19]. Not only renal epithelial cells but also tubular dysfunction occurs in impaired proximal, distal, and collecting tubules due to severe renal interstitial inflammation, which causes hyponatremia. Thirdly, this is also related to renal tubular unresponsiveness to aldosterone in late distal and cortical collecting tubules, leading to secondary pseudohypoaldosteronism [20], because KD can also involve renal tubules [21]. Finally, increased concentrations of atrial and brain natriuretic peptide (ANP; BNP) can cause hyponatremia and natriuresis by inhibiting tubular reabsorption of sodium [22,23].

We speculate that hypophosphatemia which was found to be an independent risk factor for the development of hyponatremia in our study might be caused by renal loss of phosphorus, because urinary phosphorus is mainly reabsorbed by proximal tubules and tubular injury by KD could lead to urinary phosphate wasting. To confirm its pathogenesis, however, assessment of urine output, urinary electrolytes, phosphorous, creatinine, osmolality, serum ADH levels, and plasma rennin and aldosterone concentration should be investigated in the future randomized, prospective, and controlled study.

In our study, low tCO2 levels were more common in KD children with hyponatremia than in those without, presumably indicating hypovolemic hyponatremia, which was consistent with the previous result [6]. This suggests that hypovolemic hyponatremia can frequently occur in KD patients, and it should be treated with isotonic fluids or an acetated Ringer’s solution containing 130 mEq/L Na+ for parenteral fluid therapy instead of 0.25% NaCl saline to prevent aggravation of hyponatremia.

In conclusion, our findings suggest that hyponatremia was a substantial inflammatory marker for the development of CALs in KD, because it was an independent risk factor for the development of CALs in KD and it was also closely associated with other well known inflammatory markers of KD. Because there is no single biomarker for CALs in KD, measurement of serum sodium levels as well as conventional inflammatory markers could improve the prediction for the development of CALs in KD.