2.1 Baseline characteristics
During the observation period, 353 children were enrolled in this study. Of these, 90 (25.6%) children had CAA, including 15 with medium-sized and 5 with large-sized CAAs. The mean age of the medium-large CAA group was 37 months (range 3–158 months). The male-to-female ratio of children in the medium-large CAA group was 3:1 (15 boys and 5 girls), with a ratio of incomplete KD of 0.54:1 (7/13); there were 6 cases (30%) of patients with IVIG resistance and 9 cases (45%) who had thrombosis. The mean age of the small-sized CAA group was 23 months (range 2–152 months). The male-to-female ratio of children in the small-sized CAA group was 2.9:1 (52 boys and 18 girls), with a ratio of incomplete KD of 0.49:1 (23/47); there were 14 cases (20%) of patients with IVIG resistance and 2 cases (2.9%) who had thrombosis. No deaths occurred in either group.
2.2 Comparison between the medium-large and small-sized CAA groups to analyze the risk factors for KD with medium-large CAA
2.2.1 Comparison Of Clinical Characteristics
The sum of the Harada risk scores and the incidence of thrombosis was higher in the medium-large CAA group than those in the small-sized CAA group, with a significant difference in both groups (P < 0.05). However, there were no statistically significant differences between the groups in terms of age, gender, height, weight, BMI, pSOFA score, fever duration before admission, days of illness at primary treatment, incidence of incomplete KD, and IVIG resistance or steroid therapy (all, P > 0.05), as shown in Table 1.
Table 1
Comparison of the clinical characteristics between medium-large and small-sized CAA groups
|
Total
(n=90)
|
medium-large CAA group
(n=20)
|
small-sized CAA group
(n=70)
|
P
|
Age
[month, P50(P25, P75)]
|
16.50
(9.00, 31.25)
|
16.50
(8.00, 57.50)
|
16.50
(9.75, 29.00)
|
0.563
|
<12 months[n(%)]
|
12(13.3)
|
3(15.0)
|
9(12.9)
|
1.000
|
Male[n(%)]
|
67(74.4)
|
15(75.0)
|
52(74.3)
|
1.000
|
Height
[m, P50(P25, P75)]
|
0.80
(0.71, 0.92)
|
0.82
(0.71, 1.05)
|
0.79
(0.72, 0.91)
|
0.446
|
Weight
[kg, P50(P25, P75)]
|
10.00
(8.50, 13.50)
|
10.50
(8.35, 15.88)
|
10.00
(8.50, 12.93)
|
0.387
|
BMI
[kg/m2, P50(P25,P75)]
|
15.89±1.73
|
16.06±1.86
|
15.85±1.71
|
0.631
|
pSOFA score
(point, mean ± SD)
|
0.67±1.57
|
0.80±2.46
|
0.63±1.23
|
0.670
|
Harada score
(point, mean ± SD)
|
4.10±1.59
|
4.85±1.53
|
3.89±1.56
|
0.016
|
Fever duration before admission
(day, mean ± SD)
|
8.83±5.22
|
9.30±6.17
|
8.70±4.96
|
0.653
|
Days of illness at primary treatment
[day, mean ± SD]
|
10.27±5.44
|
10.45±5.77
|
10.22±5.38
|
0.871
|
Incomplete KD[n(%)]
|
30(33.3)
|
7(35.0)
|
23(32.9)
|
0.858
|
IVIG resistance[n(%)]
|
20(22.2)
|
6(30.0)
|
14(20.0)
|
0.520
|
Steroid therapy[n(%)]
|
9(10.0)
|
3(15.0)
|
6(8.6)
|
0.673
|
Thrombosis[n(%)]
|
11(12.2)
|
9(45.0)
|
2(2.9)
|
<0.001
|
CAA, coronary artery aneurysm; IVIG, intravenous immunoglobulin; KD, Kawasaki disease; BMI, body mass index; pSOFA, pediatric sequential organ failure assessment. |
2.2.2 Comparison Of Laboratory Characteristics
The laboratory indices analysis results of the two groups of patients are shown in Table 2. The medium-large CAA group showed higher values in the serum globulin concentration, proportions of CRP > 40 mg/L, serum albumin level < 35 g/L, and lower A/G ratio values than the small-sized CAA group, and the differences were statistically significant (P < 0.05).
Table 2
Comparison of laboratory characteristics between medium-large and small-sized CAA groups
|
Total
(n=90)
|
medium-large CAA group
(n=20)
|
small-sized
CAA group
(n=70)
|
P
|
White blood cell count
(×109/L, mean ± SD)
|
15.59±7.76
|
18.27±9.80
|
14.83±6.97
|
0.081
|
>12×109/L[n(%)]
|
57(63.3)
|
14(70.0)
|
43(61.4)
|
0.483
|
Neutrophils count
(×109/L, mean ± SD)
|
9.62±6.95
|
11.93±8.40
|
8.96±6.40
|
0.093
|
≥80% [n(%)]
|
12(13.3)
|
4(20.0)
|
8(11.4)
|
0.534
|
Lymphocytes count
(×109/L, mean ± SD)
|
4.34±2.82
|
4.54±2.85
|
4.29±2.83
|
0.732
|
NLR(mean ± SD)
|
3.45±3.72
|
4.41±4.79
|
3.18±3.35
|
0.195
|
LER(mean ± SD)
|
59.28±123.89
|
56.54±106.39
|
60.07±128.28
|
0.911
|
Hemoglobin (g/L, mean ± SD)
|
103.80±15.73
|
99.91±16.73
|
104.91±15.37
|
0.211
|
≤110g/L[n(%)]
|
56(62.2)
|
14(70.0)
|
42(60.0)
|
0.416
|
Hematocrit(mean ± SD)
|
0.31±0.05
|
0.30±0.05
|
0.32±0.04
|
0.266
|
<0.35[n(%)]
|
71(78.9)
|
16(80.0)
|
55(78.6)
|
1.000
|
Platelet count
[×1012/L, P50(P25, P75)]
|
364.60
(253.83,499.30)
|
408.00
(272.00,637.43)
|
354.50
(245.25,460.75)
|
0.220
|
>350×1012/L[n(%)]
|
46(51.1)
|
11(55.0)
|
35(50.0)
|
0.693
|
PLR(mean ± SD)
|
119.73±83.13
|
124.71±77.29
|
118.31±85.20
|
0.763
|
CRP(mg/L, mean ± SD)
|
76.23±58.81
|
93.69±54.31
|
71.25±59.46
|
0.133
|
>40mg/L[n(%)]
|
58(64.4)
|
17(85.0)
|
41(58.6)
|
0.029
|
Sodium(mmol/L, mean ± SD)
|
136.29±3.12
|
135.43±3.33
|
136.53±3.04
|
0.166
|
≤133mmol/L[n(%)]
|
11(12.2)
|
4(20.0)
|
7(10.0)
|
0.414
|
ALT(U/L, mean ± SD)
|
55.58±54.99
|
67.86±65.78
|
52.08±51.50
|
0.260
|
AST(U/L, mean ± SD)
|
56.64±74.76
|
56.33±67.93
|
56.73±77.06
|
0.983
|
AST/ALT ratio(mean ± SD)
|
1.37±0.96
|
1.26±1.23
|
1.40±0.87
|
0.580
|
Total bilirubin
(umol/L, mean ± SD)
|
8.90±14.43
|
11.14±15.98
|
8.26±14.01
|
0.433
|
Albumin(g/L, mean ± SD)
|
35.33±5.71
|
33.74±5.37
|
36.05±5.74
|
0.111
|
<35g/L[n(%)]
|
43(47.8)
|
14(70.0)
|
29(41.4)
|
0.024
|
Globulin(g/L, mean ± SD)
|
27.46±9.07
|
31.70±10.76
|
26.25±8.22
|
0.017
|
A/G ratio(mean ± SD)
|
1.44±0.56
|
1.17±0.45
|
1.52±0.57
|
0.014
|
CRP/ALB ratio(mean ± SD)
|
2.30±1.91
|
2.89±1.90
|
2.13±1.89
|
0.120
|
CAA, coronary artery aneurysm; NLR, neutrophil count-to-lymphocyte count ratio; LER, lymphocyte count-to-eosinophil count ratio; PLR, platelet count-to-lymphocyte count ratio; CRP, C-reactive protein; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALB, albumin; A/G ratio, albumin-to-globulin ratio. |
2.2.3 Analysis of risk factors for KD with medium-large CAA
Univariate analysis identified five indices that were significantly associated with medium-large CAA development: the Harada risk score, serum globulin level, A/G ratio values, CRP level > 40 mg/L, and serum albumin level < 35 g/L. The Harada risk score was excluded from the multivariate analysis because the variables included in this score, such as CRP > 40 mg/L and serum albumin level < 35 g/L, were included in the multivariable model and all other variables included in this score were not statistically significant on univariable analysis, such as male sex, < 1 year of age, WBC > 12⋅109/L, platelet count > 350⋅1012/L, or hematocrit <0.35. The A/G ratio, calculated as the serum albumin level divided by serum globulin level, was included in the multivariable analysis instead of the two separate indicators. Ultimately, the values of A/G ratio, CRP level > 40 mg/L, and serum albumin level < 35 g/L were entered into the binary logistic regression analysis, which revealed that the A/G ratio was the only significant independent predictor of medium-large CAA in KD, and the risk for the formation of medium-large CAA increased by 2.503 times for every unit of increase (P = 0.039, 95% confidence interval [95% CI] 1.068–11.492).
2.3 Predictive role of A/G ratio in KD with medium-large CAA
Binary logistic regression analysis showed that a low A/G ratio value was significantly associated with medium-large CAA. The area under the ROC curve for the value of A/G ratio was 0.684 (95% CI: 0.558–0.810), and the sensitivity and specificity for predicting medium-large CAA in KD patients were 80% and 59%, respectively, at a cutoff point of 1.35 (Fig. 1). In addition, after 1 month of the disease course, the A/G ratio was negatively correlated with the internal diameter of the left main coronary artery (r = −0.407, P< 0.001). However, there was no correlation between the A/G ratio and the internal diameter of the right coronary artery (r = −0.212, P = 0.053) (Fig. 2A). The same results were found when the A/G ratio was compared with the z-score of coronary artery internal diameter (r = −0.247, P = 0.023 for the left main coronary artery and r = −0.078, P = 0.480 for the right coronary artery) (Fig. 2B). Among the 82 patients whose data for coronary outcome were complete from baseline at weeks 2, 4, and 3 months, coronary artery regression was observed after 1 month of the disease course and the right coronary artery showed early regression (after 2 weeks of onset) in 39 of 82 (47.6%) patients with an A/G ratio of < 1.35 (Fig. 3). The change of the left main coronary artery was significantly larger in patients with an A/G ratio of < 1.35 at month 1 (Fig. 3A and 3C). However, the change in the right coronary artery was not significantly greater at month 1 (Fig. 3B and 3D).