Understanding the importance of cerebrovascular involvement in Kawasaki disease

Article information

Korean J Pediatr. 2019;62(9):334-339

Publication date (electronic) : 2019 May 16

doi : https://doi.org/10.3345/kjp.2019.00143

Jung Sook Yeom, MD, PhD, Jae Young Cho, MD, Hyang-Ok Woo, MD, PhD

Department of Pediatrics, Gyeongsang National University Hospital, Gyeongsang Institute of Health Science, Gyeongsang National University College of Medicine, Jinju, Korea

Corresponding author: Hyang-Ok Woo, MD, PhD Department of Pediatrics, Gyeongsang National University Hospital, Gyenogsang Institute of Health Science, Gyeongsang National University College of Medicine, 15, Jinju-daero 816 beon-gil, Jinju 52727, Korea Tel: +82-55-750-8157 Fax: +82-55-752-9339 E-mail: howoo@gnu.ac.kr

Received 2019 February 13; Revised 2019 May 2; Accepted 2019 May 15.

Abstract

Why do we focus on cerebral vascular involvement in KD?

Kawasaki disease (KD) is a self-limiting systemic vasculitis that occurs in infants and young children that preferentially involves medium-sized muscular arteries, particularly the coronary arteries [1]. Coronary artery aneurysm, the most severe complication of KD, develops in 15%–25% of untreated patients and may lead to lethal manifestations including myocardial ischemia, infarction, and sudden death [2]. Since the introduction of intravenous immunoglobulin (IVIG) therapy, the risk of coronary artery aneurysm has remarkably decreased, but it still occurs in 4%–6% of patients [3,4]. KD is among the leading causes of acquired heart disease in developed countries [4]. However, research focuses on the coronary artery because coronary arterial lesions are directly associated with fatality. However, KD is a systemic vascular disease. Aneurysms have been reported in other extracardiac muscular arteries such as the celiac, mesenteric, femoral, iliac, renal, and branchial [5]. Extracardiac artery aneurysms reportedly occur in younger infants and those with severe acute vasculitis. The fate of extracardiac artery aneurysms resembles that of coronary artery aneurysms. Larger extracardiac artery aneurysms seem to cause stenotic lesions in the late period [5]. Arterial strokes are very rarely reported in KD [6]. Based on these findings, some authors suggested that noninvasive imaging modalities to confirm extracardiac artery aneurysms should be performed in severe cases of KD with coronary aneurysms [5]. Attention to cerebral vascular involvement in KD patients is also needed considering its fatal consequences. Autopsy cases showed endoarteritis and periarteritis in the parenchymal and pial arteries despite such lesions being less extensive than those seen in other areas such as the coronary and iliac arteries [7]. However, until now, there has been little interest in cerebral vasculitis, which may lead to stroke in KD patients. Since KD was first recognized in the late-1960s, patients who have recovered will now be middle-aged or younger. Therefore, the follow-up of these patients has been insufficient to reveal the disease’s natural history [8]. This report aims to review brain vascular involvement in KD focusing on cerebral vasculitis. Here we summarize case reports of arterial strokes, discuss the evidence of cerebral vasculitis, and suggest future research directions.

Review of arterial stroke in KD patients

1. Ischemic stroke

To our knowledge, only 16 cases of cerebral ischemic stroke have been reported in pediatric patients with KD. In the literature, cerebral infarction onset associated with KD ranges from 3 days to 4 months [6]. Most cases developed during the acute to subacute phase of KD [6]. Among them, we summarize only those studies published in English (Table 1) [6,9-19]. However, too few cases were reported to identify the clinical predictors of stroke in KD patients. Nevertheless, a prolonged fever or severe KD presentation seems to be noteworthy in connection with ischemic stroke. All patients but one suffered from a fever lasting more than a week (median 10.5 days). In some patients, a devastating course developed. This analysis shows that we must monitor for cerebrovascular involvement as well as cardiovascular complications in KD patients with prolonged fever or severe progression.

Table 1.

Literature review of studies of ischemic stroke associated with Kawasaki disease (English publications only)

Several potential mechanisms may explain the occurrence of stroke in KD. In most cases, cerebral infarction occurred in the middle cerebral artery (MCA) territory in KD patients. Most cases of MCA territory infarction have been attributed to cardiogenic embolism, internal carotid artery occlusion, or internal carotid artery dissection [20]. Embolism induced by myocardial infarct likely provokes the cerebral infarction in the course of KD [21]. IVIG-associated thrombosis also has been suggested as a cause of stroke in KD [13]. Complete occlusion of the internal carotid artery was reported in an 18-month-old girl 10 days after the onset of fever and the day after she received IVIG [6]. Thrombotic complication is one of the well-known complications of immunoglobulin therapy [22]. IVIG-associated thrombosis may be related with an increased blood viscosity, procoagulant activity of immunoglobulin itself, and immunoglobulin-induced arterial vasospasm [6]. The authors of this case suggested evaluating hydration status before IVIG treatment [6]. A 30-month-old girl with hemiplegia showed ongoing inflammatory processes of the cerebral vessels as the cause of her stroke and no abnormal findings of magnetic resonance angiography (MRA) [16]. However, in this patient, stroke occurred shortly after aspirin therapy was stopped [16]. Therefore, in addition to vascular inflammation, acquired thrombophilia related to inflammatory syndrome with hyperthrombocytosis might have been a mechanism of stroke in this patient [16].

An arteritic complication was also suggested as a possible mechanism of stroke in a male infant with cerebral infarct because complete occlusion of the MCA and systemic aneurysms (including coronary aneurysm) were noticed simultaneously [23]. In this patient, stroke occurred 45 days after KD symptom onset and he was not treated with IVIG [23]. This case suggested that cerebral infarction involving large-sized blood vessels might be associated with arterial occlusion resulting from a nonspecific inflammation burst and prothrombotic inflammatory state of the large vessels [19]. Cerebral vasculitis affecting medium- and small-sized vessels has been suggested as a cause of stroke in KD [19]. A 4-year-old female patient presented with rapidly catastrophic KD in which tetraplegia developed 15 days after KD onset [19]. Multiple supra- and infratentorial microhemorrhages were observed, but neuroimaging revealed normal arteries of the circle of Willis [19]. Based on these findings, the authors suggested diffuse brain vasculitis as a cause of ischemic damage in this patient [19].

2. Hemorrhagic stroke

Only 5 cases have been reported to date on hemorrhagic stroke as a possible complication of KD [24-28]. Among them, we summarize 3 cases were published in English (Table 2) [24-26]. All 3 cases were caused by aneurysm rupture that occurred from 7 months to a decade after KD onset [24-26]. Aneurysm rupture appears a more long-term complication of KD compared with ischemic stroke. Among the 3 cases, a saccular aneurysm at the bifurcation of the major arteries in 2 cases and a stalk-like aneurysm located at the trunk of the major artery in 1 case [24-26]. Histopathologic findings of a 12-year-old male patient seemed to mimic the acute inflammatory changes observed in the coronary arteries in KD because of hypertrophy and invasion of the inflammatory cells resulting in endothelial wall thickness [24]. This patient had KD at 3 years of age and had a stroke 9 years later [24]. On the other hand, there was only mild invasion of inflammatory cells but complete absence of the elastic lamina in a 20-year-old man [26]. The authors suggested that all the vasculitis changes during infancy completely destroy elastic lamina and tunica intima which caused the aneurysmal formation [26]. The pathological findings of the 2 cases suggested that inflammation of cerebral arteries might be associated with aneurysm formation in KD [24,26]. Although it is difficult to draw any conclusions from the 2 cases, these findings show why attention must be paid to inflammation of the cerebral vessels in KD.

Table 2.

Literature review of studies of hemorrhagic stroke associated with Kawasaki disease (English publications only)

Asymptomatic cerebral vasculitis may be more common than expected

To discuss this issue, we must review a prospective systematic review of KD patients with coronary artery lesions [15]. Muneuchi et al. [15] prospectively evaluated brain lesions using magnetic resonance imaging (MRI) and MRA in 24 patients with coronary artery lesions at 0.1–21.2 years after the onset of KD. The median age at KD onset was 1.2 years, and no patients had significant neurological symptoms or signs [15]. Of 24 patients, approximately half (11 of 24) had a giant aneurysm in the coronary artery, while 17 (70%) were treated with aspirin and IVIG. An ischemic lesion was observed in one of 24 patients (4%) on brain images.15) MRI on the 21st day of the illness revealed an ischemic lesion on the cerebellum with severe stenosis of the right posterior inferior cerebellar artery in a 4-year-old boy [15]. He was unresponsive to 2 trials of IVIG and had a giant aneurysm. Interestingly, he had no obvious neurological symptoms or signs throughout the KD course [15]. The authors assumed the need to consider the possibility of brain vascular lesions in severe cases of KD with or without neurological symptoms [15]. Considering the long-term consequences of cerebrovascular lesions, the incidence of 4% seen in this study is not low.

We then considered the brain vascular involvement in patients without coronary artery lesion. No coronary artery lesion has been reported in some patients with ischemic stroke [6,13]. A study of single-photon emission computed tomography (SPECT) by Ichiyama et al. [29] might answer this question. To investigate brain perfusion in the acute stage of KD, they performed SPECT in 21 children with acute stage KD and performed follow-up SPECT and MRI about 1 month after the first SPECT [29]. SPECT imaging demonstrated localized cerebral hypoperfusion in 6 of 21 children (28.6%), but none had any abnormal neurological symptoms or signs [29]. No coronary lesions were reported in any of these 6 patients.29) The maximum C-reactive protein level or fever duration was not related with SPECT abnormalities [29]. The follow-up SPECT and MRI performed 1 month after the first SPECT revealed no abnormalities [29]. SPECT has been used as a diagnostic tool for cerebral vasculitis [30]. Vasculitic central nervous system (CNS) involvement has been seen with hypoperfusion in SPECT [30]. Based on these findings, Ichiyama et al. [29] suggested that cerebral vasculitis might have been caused by transient localized cerebral hypoperfusion observed on SPECT in patients of this study. The interesting findings of this study are that approximately 30% of patients showed evidence of cerebral vasculitis without coronary lesions and neurological signs [29]. The 2 studies mentioned above [15,29] suggested that asymptomatic involvement of the cerebral vessels or cerebral vasculitis may be more common than expected. Thus, should all KD patients be subjected to further neurological evaluations? There is insufficient data to answer this. However, considering the fact that clinically significant neurological complications are very rare, additional neurological examinations or tests may not be necessary in all KD patients.

Cerebrospinal fluid study evaluating CNS inflammation in KD

From this perspective, we studied the intracranial inflammatory response in patients with acute-phase KD with evaluating the cerebrospinal fluid (CSF) cytokine levels [31]. We hypothesized that the CSF levels of proinflammatory cytokines are elevated if cerebral vasculitis accompanies acute stage KD [31]. Inflammatory cytokines including tumor necrosis factor (TNF)-α and interleukin (IL)-6 might play a critical role in the pathogenesis of KD [32] and the development of inflammation within the cerebral blood vessels [32,33]. These cytokines were secreted by infiltrated immune cells within the vessels [34]. Vascular changes such as endoarteritis, periarteritis, and perivascular cuffing with mild lymphocytic infiltration have been observed in some pial arteries in autopsy cases [7]. Pentraxin-3 (PTX3) is synthesized at the sites of vascular inflammation and serves as a biomarker of vasculitis in many diseases [35]. The whole-blood transcript levels of PTX3-encoding genes and the extent of KD vasculitis have been reported [36]. PTX3 is capable of inducing vascular endothelial dysfunction [37]. Based on these findings, we hypothesized that elevated CSF levels of TNF-α, IL-6, and PTX3 may suggest CNS vascular involvement. To verify our hypothesis, we measured the levels of the aforementioned mediators in patients with KD and compared them to those of the febrile control groups [31]. Contrary to our expectations, PTX3 and TNF-α were rarely detected in KD patients, while the levels of IL-6 did not differ from those of nonspecific viral illness [31]. Our study seems to show that intracranial inflammation including vasculitis might not be insignificant in patients with KD [31]. We found no evidence of intracranial inflammation related with vascular involvement.

However, Korematsu et al. [38] reported that CSF IL-6, soluble TNF-α receptor 1, and the IL-6 CSF/serum ratio were elevated in patients with KD. They revealed that some patients with KD had a higher degree of independent CNS inflammation [38]. Thus, how can we interpret these conflicting results? The major differences between our study31) and that of Korematsu et al. [38] were onset age, disease severity, and the presence of neurological symptoms. The mean patient age in the study of Korematsu et al. [38] was 2.6±2.1 years; all patients had mildly disturbed consciousness, while some had a coronary aneurysm. In contrast, all patients in our study were very young (median, 2.3 months), did not have coronary aneurysm, and underwent lumbar puncture to find the febrile focus [31]. Those findings suggest that intracranial inflammation may be related with coronary artery inflammation severity. To prove this issue, further studies involving larger number of patients with a wide spectrum are needed. The intracranial inflammation status is unknown in most KD patients.

There is growing concern about the predisposition of KD patients to endothelial damage or cardiovascular disease in later life. Some authors advise that individuals with a history of KD receive cardiovascular counseling and make an effort to minimize traditional modifiable cardiovascular risk factors [39]. Stroke might be a complication of KD. A few studies have suggested that asymptomatic cerebral vasculitis might be more common than originally thought [29]. Therefore, pediatricians should be aware of the possibility of vascular involvement beyond the heart in KD, particularly in the cerebral vessels.

Article information Continued