Discussion
Approximately 15%–25% of patients with KD do not respond to the standard regimen of oral aspirin and IVIG.12 Predicting the therapeutic response of patients with KD to the standard treatment regimen in the early phase of the disease is critical in preventing and treating coronary injury and has gained momentum in recent years. Kobayashi et al
13 devised a scoring system in 2006 by using sodium, days of illness at initial treatment, aspartate aminotransferase, neutrophil, CRP, age and PLT. Another study4 reported that erythrocyte sedimentation rate, glutamic-pyruvic transamine, WBC, serum albumin, the time from onset to IVIG treatment and IVIG dosage were risk factors for refractory KD. Uehara et al
3 considered that the initial IVIG before the fifth day of illness, boys, low PLT, elevated alanine aminotransferase and CRP were significantly related to IVIG non-response. The above parameters could be classified as some common clinical profiles and laboratory findings. Our previous research14 found the Kobayashi system and some other models established by these indicators were deficiently precise to be clinically applied to our population. Therefore, we speculate that some other factors might have not been found in regard to the IVIG therapeutic response of children with KD during the acute phase. Cytokines, also have been reported as early predictor for IVIG non-responders.15 16 Nevertheless, some patients with other diseases can also be observed with a high level of cytokines before treatment, such as bacterial infections, pneumonia and sepsis, sometimes even ‘cytokine storms’ occurring after infection. Therefore, cytokines are not stable and specific biomarkers for assessing the severity of KD. In view of these reasons, we aimed to find a simple novel biomarker to predict the risk of IVIG resistance.
We have previously reported that plasma levels of TSP-2 and TSP-1 were significantly increased in children with KD presenting with coronary artery dilatation.17 Here, we show that in the acute phase of KD, the plasma TSP-2 and TSP-1 concentration was significantly elevated relative to healthy control and patients with an unrelated fever. The TSP-2 concentration in IVIG non-responders was significantly higher than in IVIG responders. To examine the possible predictive value of plasma levels of TSP-2 for the responsiveness of IVIG therapy, we used ROC analysis and chose the cut-off value as 31.50 ng/mL, so that the Youden index attained its maximum. The results showed that plasma levels of TSP-2 had both higher sensitivity (82.35%) and specificity (64.81%) in predicting the non-response of IVIG for treating KD. It revealed that when the plasma level of TSP-2 was >31.50 ng/mL, >80% suspected IVIG-resistant children would be recognised. Therefore, the plasma level of TSP-2 can be taken as one of the reference indices in choosing medication for children with KD. Thus, patients at high risk resistant to IVIG would be identified in time and they might benefit from a more appropriate individualised initial treatment. A scientific statement for KD from the American Heart Association in 201718 states that adjunctive therapies, such as the addition of corticosteroid or infliximab therapy to IVIG and ASA in the primary therapy of KD, lower the prevalence of coronary artery abnormalities at high risk for IVIG resistance. Therefore, it is of important clinical implications to predict high-risk IVIG non-response patients before initial treatment.
The mechanism responsible for the increased TSP-2 level in patients with KD of non-responders is unknown. As the member of extracellular matrix protein family, TSPs have an important regulatory role in cardiomyopathy, coronary heart disease, myocardial infarction and other cardiovascular diseases.19–21 Genetic polymorphisms in TSP-1, TSP-2 and TSP-4 have been linked to coronary heart disease and myocardial infarction,22 23 and elevated TSP-2 was detected in fibrotic and narrowed aortic valves.24 Agah et al
25 reported in a wound healing assay that the release of TSP-2 from VSMCs was not detected until 3 days following injury, and peaked as far out as 10 days postinjury, whereas TSP-1 was detected early after injury, peaking at 3 days and subsiding by 7 days. This finding suggests TSP-1 may function more as an acute phase reactant, but TSP-2 may be more important during the subsequent inflammation and remoulding phase. So, it is possible that the elevated TSP-2 level observed in patients with KD of IVIG non-responders related to the excessive inflammatory reaction.
IVIG is a polyclonal IgG preparation, and it remains unclear why it is beneficial in KD. One possibility is that IVIG can reduce tissue inflammation and excessive immune activation by binding to blood mononuclear cells, endothelial cells or platelet surface Fc receptor, and by doing so prevent immune-mediated injury of the intimal surface.26 IVIG has also been shown to reduce the levels of matrix metalloproteinase-2 (MMP-2), MMP-3 and MMP-9, which are significantly elevated in patients with KD.27 28 Reducing the MMP expression level played a protective role in the subendothelial basement membrane and intimal elastin. TSP plays a role in inhibiting angiogenesis by blocking binding of vascular endothelial growth factor that is assembled in the extracellular matrix to its receptor by binding to MMPs. It has been reported that TSP-1 regulates MMP activity by inhibiting conversion of pro-MMP-2 and pro-MMP-9 to the active form, and can also directly inhibit the activity of MMP-9 via MMP-3 and trypsin.21 Calabro et al have also reported an interaction between TSP-2 and MMP-2 that affects angiogenesis by reducing intercellular adhesion.29 Our present findings suggested that elevated plasma TSP-2 level might be connected with the immune response and MMP activation.
The present study has some limitations. First, the number of patients might be relatively small and all specimens were only collected from one hospital. However, considering the occurrence rate2–4 of IVIG resistance and the feature of this retrospective study, it was hardly to collect affluent samples in a short term. We hope that this finding will lead to more attention and inspiration, so that more investigations and clinical trials from multicentre will be conducted to validate these results. Second, the levels of TSP-2 were not correlated with other known parameters, such as clinical manifestations and laboratory indicators, and if these elements were added, the predicting effect might be improved. Third, the specificity of TSP-2 was lower, which may lead to misdiagnosis of some patients.
In conclusion, the plasma concentration of TSP-2 might serve as a useful indicator for predicting the risk of IVIG non-response in children with KD before starting treatment and help paediatricians planning individualised initial therapy regimens.