Ultrasound evaluation of endothelial dysfunction in immunoglobulin-resistant children with acute Kawasaki disease

Given the evidence that brachial artery flow-mediated dilation (FMD) is declined in children later after the onset of Kawasaki disease (KD), we hypothesized that indicators that detect the situation of the endothelium are useful parameters that can accurately reflect subclinical dysfunction in resistant patients and assist in differentiating patients with KD at a higher risk of IVIG resistance, which may be valuable in better understanding how to protect patients from endothelial and thrombotic complications. Fifty IVIG-resistant KD children, 120 IVIG-responsive KD children, 35 febrile children with acute upper respiratory infection, and 50 healthy controls were recruited, and indicators reflecting endothelial inflammation, including flow-mediated dilation (FMD), were measured. Receiver operating characteristic (ROC) curve analysis was utilized to determine the threshold values of these indicators of IVIG resistance. Multiple logistic regression analysis was performed to determine whether FMD was an independent predictor of IVIG-resistant patients. In comparison with the lab data, PCT, Na + , and FMD exhibited AUCs of 0.727, 0.653, and 0.698 (P < 0.05), respectively, in predicting IVIG resistance in KD through ROC analysis. PCT > 1.69 ng/ml, Na +  < 133.2 mmol/l, and FMD < 5.79% were independent predictors of IVIG resistance in KD (OR 4.257, 3.516, 3.563, 95% CI 1.549 ~ 11.700, 1.277 ~ 9.680, 1.299 ~ 9.772, P < 0.05). More severe endothelial dysfunction, especially lower FMD, was present in IVIG-resistant patients than in IVIG-responsive patients. It is a helpful diagnostic tool that provides supportive criteria to detect KD patients at a higher risk of IVIG resistance when FMD < 5.79% in children. Key Points • IVIG-resistant KD patients have more severe endothelial dysfunction than IVIG-sensitive patients. • FMD < 5.79% may indicate an increased risk of IVIG resistance in children with Kawasaki disease. Key Points • IVIG-resistant KD patients have more severe endothelial dysfunction than IVIG-sensitive patients. • FMD < 5.79% may indicate an increased risk of IVIG resistance in children with Kawasaki disease.


Introduction
Kawasaki disease (KD) is a self-limited systemic vasculitis. Intravenous immunoglobulin (IVIG) in combination with aspirin has been the standard regimen for years [1], approximately 10-20% of children with KD still exhibit persistent fever or the recurrence of the typical symptoms of KD within 36 h after the completion of the first IVIG injection, and this phenomenon is known as IVIG resistance [2]. The early detection of such resistant patients may greatly help clinicians achieve optimal treatment and prognosis. The physician could quickly communicate with the patients' guardians about doubts regarding the effectiveness of IVIG before IVIG resistance appears, which would efficiently reduce the occurrence of medical disputes; early detection may also enable the choice of personalized therapy that is administered based on the classification of IVIG resistance [3,4]. Several inflammatory laboratory parameters and cytokine markers such as interleukins, TNF-α, and interferon γ have been reported to predict IVIG-resistant KD because they participate in the activation of monocytes/macrophages and granulocytes at different levels in the innate immune response.
Nevertheless, these predictors may not be popular in clinical practice because screening for them is limited by technology and cost; moreover, collecting them maybe invasive and some of them lack reliability especially in low-and middle-income countries with few resources [5][6][7][8][9]. In fact, since they are noninvasive ultrasound indicators, f low-mediated dilation (FMD), carotid intima-media thickness (CIMT), and artery stiffness index (SI) have been widely used in recent years in the prognostic follow-up of patients with cardiovascular risk associated with chronic diseases such as diabetes, hypertension, and metabolic syndrome [10,11]; an abnormal decrease in FMD and increases in CIMT and SI suggest a suboptimal state of the endothelium. A number of studies have confirmed that children with a history of KD, especially those with coronary artery lesions, maintain these abnormalities from the beginning to the follow-up years of the disease [12,13]. The pathological basis of KD lies in the systemic inflammation of small and medium-sized blood vessels, which indicates that the immunological mechanism of IVIG resistance may also be inextricably linked to endothelial dysfunction. Therefore, it is of great clinical significance to explore the correlation between endothelial function and drug resistance and their roles in the mechanisms underlying the onset and progression of KD. However, the clinical value of these endothelial indicators in the acute phase of KD has not been fully studied, and their relationship with IVIG resistance is especially understudied. Additionally, the parameters of vascular ultrasonography have not previously been included in the evaluation system to predict IVIG resistance. Therefore, in the present study, we measured the endothelial function of IVIG-resistant KD patients. Given our findings, we further evaluated the relationship between endothelial function and the degree of systemic serum internal environmental inflammation in IVIG-resistant children with KD, and clarified the correlation between endothelial function indicators and IVIG resistance, as well as the clinical significance of this correlation.

Study population
From July 2017 to October 2021, 50 IVIG-resistant and 120 IVIG-responsive patients in the acute stage of KD were randomly recruited at the Chengdu Women's and Children's Central Hospital of the School of Medicine at the University of Electronic Science and Technology of China using the numbering method in SPSS 20.0. Seventy cases of children with fever due to acute upper respiratory tract infection (AURI) and 100 normal children completing regular check-ups at our hospital, who were matched for age and sex with patients in the KD group, were also randomly enrolled during the same period using the same approach. The diagnosis of complete KD, incomplete KD, and coronary artery lesions (CALs) met the diagnostic criteria of the American Heart Association in 2017 [1]. CALs were defined as a dilated coronary artery with a z score of ≥ 2.5 in at least one of the right, left anterior descending, and left main coronary arteries. All KD children were treated with IVIG (2 g/kg over 12 h), as well as oral aspirin (50 mg/kg per day divided into 3 doses and given every 8 h; 5 mg/kg per day after the child has been afebrile for 72 h after IVIG treatment) treatment with ibuprofen was avoided. Immunoglobulin resistance was determined as persistent or recurrent fever (under-arm temperature ≥ 37.5 °C) for more than 36 h but not longer than 7 days after the end of IVIG infusion. The exclusion criteria for all subjects were as follows: (1) previous medical history of KD; (2) previous or current medical history of dyslipidemia, diabetes, inherited cardiovascular disease, inherited metabolic disease, cystinuria, chronic kidney disease, blood disease, etc.; and (3) any history of major surgery, blood transfusion, or use of vasoactive agents or NSAIDs except ibuprofen, glucocorticoids, or immunosuppressive drugs within 3 months before our examination.

Laboratory and clinical data
Laboratory and clinical data were collected through medical record review. Clinical data, such as age, sex, body mass index (BMI), time of blood sampling, diagnosis of incomplete KD and CALs, and fever lasting several days before first IVIG, were collected. On the day of admission before IVIG and aspirin treatment (at least 6 h after the application of ibuprofen for those who had used ibuprofen to abatement alleviate fever before hospitalization), laboratory data were collected and include white blood cell (WBC) count percentage of neutrophils (N%), hemoglobin (Hb), platelet (PLT), C-reactive protein (CRP), serum alanine aminotransferase (ALT), serum aspartate aminotransferase (AST), serum albumin (ALB), serum sodium (Na +), procalcitonin (PCT), N-terminal pro-brain natriuretic peptide (NT-proBNP), creatine kinase-MB (CK-MB), d-dimer, and erythrocyte sedimentation rate (ESR).

Assessment of ultrasound data
In this series of tests, data from all participants was gathered 72 h after the initial treatment of IVIG and aspirin as well as before other additional treatments for the KD patients. All subjects were asked to sleep late and lay at rest in the supine position for at least 10 min in a quiet room before the ultrasound assessments. Any uncooperative subjects were sedated with chloral hydrate (10% solution at 0.3-0.5 ml/kg, maximum 10 ml). All ultrasonic indicators were measured under standardized conditions using an ultrasound machine (Philips CX50) with a L12-3 transducer; these measurements were performed by a single experienced sonographer (Wen Y) who was blinded to the diagnoses, and the average of three repeated test records for each indicator was used for statistical analysis. The KD group underwent additional echocardiography tests and was examined before IVIG treatment.
1. Brachial arterial FMD: FMD was identified with reference to the methods described by Celermajer et al. [14] in 1992. The brachial artery was detected above the antecubital fossa of the right arm by the aforementioned ultrasound machine. The distance between the posterior and anterior intima of the vascular wall at the end of diastole (D1) was imaged in the longitudinal axis as a baseline internal diameter. Then, the blood pressure cuff was deflated rapidly after having been inflated to 50 mmHg above the subject's resting systolic blood pressure for 5 min to occlude the brachial artery. The diastolic diameter (D2) was recorded 60 s later. The FMD value was calculated as FMD = [(D2 − D1)/D1] × 100%.
2. CIMT: All measurements were performed according to a standardized scanning protocol for the right common carotid arteries [15]. The transducer footprint was manipulated to be parallel to the near and far walls of the common carotid arteries, and the maximum in the longitudinal plane was supposed to be the lumen diameter. The entire carotid proximal common carotid artery was observed approximately 1.5 cm before the bifurcation. The distance between the leading edges of the luminalintimal interface and the medial-adventitial interface was measured as CIMT. The LVEF was acquired by the biplane disk method (modified Simpson's rule) [17], and the LVFS was obtained according to previously reported guidelines [18].

Statistical analysis
Analyses were performed using SPSS 20.0 software (IBM Co., Armonk, NY, USA). Medians and interquartile ranges were reported for the quantitative ultrasonic and laboratory data. Categorical data are described as counts and percentages. Because some data, such as SI and Na, were not normally distributed, comparisons of continuous variables among groups were evaluated using the Kruskal-Wallis H test, while two groups were compared by the Mann-Whitney U test. Chi-square tests were performed to compare categorical variables. Spearman's partial correlation was used to assess correlations between variables. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the role of different indicators in discriminating IVIG unresponsiveness. The sensitivity and specificity were calculated, and the cut-off value was determined by the Youden index. Multivariate logistic regression was used to identify independent predictors of IVIG resistance in KD. All tests were twosided, and a value of P < 0.05 was considered to indicate statistical significance.

1.
Comparison of clinical, ultrasonic, and laboratory data among groups. As shown in Table 1, there were no significant differences in baseline characteristics (age, sex, and BMI) or SI among the KD, AURI, and normal groups. FMD and CIMT in the KD group were both significantly different from those in the other two groups (P < 0.01). Table 2 shows a comparison of the laboratory data between the subgroups of KD. These 2 tables indicate that FMD and the levels of serum Na + and ALB were significantly lower while CIMT, N%, and the levels of serum PCT and NT-proBNP were significantly higher (P < 0.05) in the IVIG-resistant group than in the IVIG-responsive group. 2. ROC analysis of different indicators for predicting IVIG resistance. The ROC curves of FMD, CIMT, N%, PCT, ALB, Na + , and NT-proBNP for predicting IVIG resistance in children with KD were analyzed ( Fig. 1; Table 3). In the analysis, the sensitivity of FMD was found to be higher than those of other indicators, with a value of 78.0% at its cut-off point of 5.79% (P < 0.001). The analysis of PCT at its cutoff point of 1.69 ng/ml resulted in the largest AUC of 0.727 (P < 0.001).

Discussion
There is an increasing amount of recent evidence that the mechanism of IVIG in the treatment of KD mainly involves the immune response, the destruction and remodelling of the blood vessels, and endothelial inflammation [19][20][21]. In this study, the endothelial function of IVIG-resistant KD patients was found to be significantly lower than that of IVIG-sensitive ones, possibly due to vascular remodelling and poor regulation of endothelial inflammation. An endothelial parameter such as FMD lower than a certain value clearly indicates IVIG resistance. Based on these perspectives, it is hypothesized that laboratory markers of inflammation and endothelial function parameters may be reliable predictive biomarkers. There have been numerous studies [22][23][24] about the predictive value of laboratory markers for IVIG resistance in KD, but few reports have examined endothelial function parameters and the correlation between these two main indicator types.
In our investigation, we demonstrated that IVIG-resistant KD patients had both significantly lower FMD and higher CIMT than other groups, despite the differences in SI in these groups not reaching statistical significance. However, we did not find any difference in endothelial parameters between AURI patients and normal children. These findings confirmed that common infection-related febrile factors may not be sufficient to cause either obvious swelling of the vascular medial membrane or serious damage to the structure of the arterial wall and that simple and transient inflammation  in the acute phase may not be sufficient to cause significant premature arteriosclerosis in a short time [25]. In addition, there was lower FMD and higher CIMT in IVIG-resistant patients than in the IVIG-sensitive patients, which indicates that endothelial dysfunction in IVIG-resistant KD patients may be more serious than that in IVIG-sensitive patients and AURI patients. Additionally, there was no significant difference in CIMT observed between the IVIG-sensitive and AURI groups. Therefore, the analysis of FMD may be more sensitive than analyses of CIMT and SI for the assessment of endothelial function in KD children during the acute phase. In further logistic analysis, we found that of the three vascular ultrasonography parameters, only FMD (< 5.79%) was an independent predictor of IVIG resistance, which also confirmed the above hypothesis. FMD was developed as a high-frequency ultrasound technique for a surrogate marker of endothelial vasodilation function by Celermajer et al. [14] in 1992. It works by physically blocking the brachial artery, resulting in transient ischemia and hypoxia, which stimulates the endothelium to release nitric oxide and other associated vasodilators (neurohormones, such as acetylcholine, 5-hydroxytryptamine, and catecholamines). When the physical blockage is removed, vessels with normal endothelial function expand in response to a large number of these cytokines and molecules [26]. Unlike SI and CIMT, which are primarily used to evaluate vascular morphology [27], FMD is mainly used to monitor vasodilation function after artificial mechanical stimulation as a parameter for measuring endothelial function, rather than structure, i.e., physiology rather than anatomy [14,28]. Since endothelial cell activation and dysfunction are earlier observable changes than structure and form in the development of vascular diseases in lesion areas [29] and measurement of FMD includes online monitoring of disturbed blood flow dynamics, FMD can more sensitively reflect the difference between the vascular endothelium of patients with IVIG resistance and other groups under the influence of systemic inflammation. In adults, FMD < 5 indicates impaired endothelial function, but there is no established standard in children [30]. In the present study, the cut-off value predicted by FMD for IVIG resistance was a slightly higher than the normal minimum for adults mentioned above but significantly lower than that of children in the healthy control group. Therefore, it was speculated that most IVIG-resistant KD patients had subclinical endothelial dysfunction, as the correlation between them may be undesirable consequences of inflammatory storms.
In the analysis of laboratory markers of inflammation in KD patients, between the IVIG-sensitive and IVIG-resistant groups, there were statistically significant differences in levels of NT-proBNP, which reflects atrial dilation as well as myocardial ischemia and hypoxia; moreover, there were differences in Na + , which mainly reflects microvascular permeability and interleukin levels [23], and inflammatory parameters, such as N%, PCT levels, and ALB levels,  which also support the trigger and signalling cascades of a severe clinical vasculitis called cytokine storm, which stimulates polyclonal B cell autoantibody production, resulting in acute inflammation and antibody-mediated endothelial damage [31,32] in IVIG-resistant patients. Thus, the level of inflammation was higher and less conducive for remission in these patients, which is consistent with the results of some previous studies [22][23][24]. Among those laboratory data, aggressive inflammatory response (PCT > 1.69 ng/ ml) and hyponatremia (Na + < 133.2 mmol/l) were found to be strong independent predictors of IVIG resistance in KD children, similar to FMD (< 5.79%), through multivariate logistic regression analysis. PCT is an acute-phase reactant whose levels are not affected by the administration of corticosteroids or nonsteroidal anti-inflammatory medication. As a prohormone for calcitonin, it is suppressed or nonresponsive to interferon-γ, moderately responsive to tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6) stimulation, and strongly responsive to interleukin-1β (IL-1β) [24]. Some studies found that the expression of cytokines such as TNF and ILs increased in the serum of IVIG-resistant patients, and that the activation of the pathway was related to IVIG resistance, which may be one of the reasons why the PCT of IR was higher than that of IS [33]. Previous studies reported that increased IL-6 and IL-1β levels may activate antidiuretic hormone (ADH) secretion and aggravate microvascular permeability by activating the ADH receptor 2 (A2R) [34] in KD patients with severe inflammation, leading to syndrome of inappropriate antidiuretic hormone (SIADH) and hyponatremia [23]. In endothelial cells, A2R activation is associated with increased von Willebrand factor (vWF) and nitric oxide (NO) release and increased tissue plasminogen activator (TPA) and p-selectin activity [35]. This process partially has the same inflammatory mediators as PCT, which means that both hyponatremia and elevated PCT levels are closely related to excessive expression of inflammatory factors and the subsequent appearance of permeable microvascular which affects endothelial system function and of course the FMD [36,37]. While only the analysis of PCT levels exhibited a larger AUC than that of FMD, which had the highest sensitivity with a value of 78.0%, indicating that FMD could be a better predictor than most of the laboratory data above.
In the context of immune-mediated injury, the vascular endothelium may trigger a low-level inflammatory response by imitating pathogen-and damage-associated molecular patterns and activating monocyte-derived macrophages, T cells, B cells, dendritic cells, and mast cells by affecting substrate-receptor interactions and the secretion of different mediators [28] of inflammation, coagulation, angiogenesis, and tumor invasion. In our follow-up correlation analysis, the negative correlations between FMD and PCT in both the KD group and subgroups had a larger magnitude in the IR group than in the IS group, which is consistent with the pathological mechanism mentioned above. While hyponatremia was not significantly correlated with FMD possibly because sodium balance is more influenced by endocrine, renal, and neuroregulatory control but much less by endothelial control [23,38]. These findings further confirmed that a severe exaggerated inflammatory response can lead to vascular damage through biochemical and immune damage pathways [28] and is an important mechanism of vascular structure and endothelial function damage in IVIGresistant patients [39,40]. At the same time, there may be considerable overlap between inflammatory responses that lead to elevated PCT and pathophysiological processes (including injury of endothelial cells that release vasodilators and blocking of vasodilator response pathways in the vascular wall that influence vasodilation) in IVIG-resistant patients. This may be related to cytokines such as TNF-α, IL-6, and IL-1β [28]. Even though FMD values of < 5.79% (OR = 3.563) have a good predictive ability for IR, patients with PCT levels of > 1.69 ng/ml had a higher OR (4.257), which indicates that abnormally elevated PCT levels have a stronger correlation with IR than FMD. Therefore, we speculate that changes in the morphology and function of vessel walls lag behind the occurrence of inflammation in KD, which is also believed to be the initial step in many other rheumatic diseases, including systemic lupus erythematosus, ankylosing spondylitis, Bechet's disease, and rheumatoid arthritis [41].
Although further basic experimental proof is needed, the above conclusions might provide a clue for future etiologic studies on IVIG resistance in KD. Additionally, before IVIG treatment, if the FMD, the PCT levels, or the Na levels are found to be abnormally different from their critical values, we recommend the timely strengthening of endothelial protective support, extra anti-inflammatory therapy (such as initial intravenous application of glucocorticoid), and the timely correction of the fluid and sodium imbalance and other measures that can promote the control of inflammation to a certain extent to prevent the occurrence of IVIG resistance [3,21,42,43]. Similarly, in children with IVIG resistance, strengthening the above measures may help to prevent the recurrence of repeated IVIG resistance and refractory KD. While the use of the combination of FMD and PCT levels as a diagnostic tool in routine clinical practice is not yet warranted, further investigation into this combination might increase its predictive performance and prove the utility of this approach in the future.
Our study has several limitations. First, the sample size of this study was small, and all subjects in this study were Chinese. Further large-scale, multicenter prospective studies are needed to confirm our findings. Second, to acquire ultrasonic data, we measured only the right neck for CIMT and SI and only the left brachial artery for FMD, while LVEF and LVFS were measured only in the KD group, which may have led to some selection bias. Third, the laboratory data were limited in our study; we did not collect laboratory data for the AURI and normal groups, and thus, more indicators need to be included prospectively. Fourth, as our hospital is a large children's medical center in southwest China, there may be a certain selection bias, which allows patients with more severe disease to be admitted to our hospital; timely diagnosis and treatment avoid the occurrence of some CALs, and the relatively small number of CAL cases also contributes to the bias.

Conclusion
Changes in the morphology and function of vessel walls may lag behind the occurrence of inflammation in KD patients. IVIG-resistant KD patients have more severe endothelial dysfunction than IVIG-sensitive patients. FMD reflects endothelial function from a physiological perspective by evaluating the vasomotor status of humeral artery in KD patients, and is closely related to the severity of inflammation and the incidence of IVIG resistance in the acute stage of KD. FMD may be a more useful diagnostic parameter than CIMT and SI, providing supporting criteria for identifying KD patients with higher risk of IVIG resistance when FMD is < 5.79% in children. In order to reduce endothelial damage in KD patients, especially patients who are IVIGresistant, more intense or adjuvant therapy should be considered, and the dynamic and intensive monitoring of endothelial function should be enhanced by ultrasound.

Declarations
Ethical approval This study was performed in accordance with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards and was approved by the Ethics Committee of the Chengdu Women's and Children's Central Hospital (approval number: 2017(11)). Clinical informed written consent was obtained from the guardians of each child.