|Home | About | Journals | Submit | Contact Us | Français|
Kawasaki disease (KD) is an acute systemic inflammatory illness of young children that can result in coronary artery aneurysms, myocardial infarction and sudden death in previously healthy children. Clinical and epidemiologic features support an infectious cause, but the etiology remains unknown four decades after KD was first identified by Tomisaku Kawasaki. Finding the cause of KD is a pediatric research priority. We review the unique immunopathology of KD and describe the current treatment. New research has led to identification of viral-like cytoplasmic inclusion bodies in acute KD tissues; this finding could lead to identification of the elusive etiologic agent and result in significant advances in KD diagnosis and treatment. Current management of acute KD is based upon prospective, multicenter treatment trials of intravenous immunoglobulin (IVIG) with high-dose aspirin. Optimal therapy is 2 g/kg IVIG with high-dose aspirin as soon as possible after diagnosis during the acute febrile phase of illness, followed by low-dose aspirin until follow-up echocardiograms indicate a lack of coronary abnormalities. The addition of one dose of intravenous pulse steroid has not been shown to be beneficial. For the 10–15% of patients with refractory KD, few controlled data are available. Options include repeat IVIG (our preference), a 3-day course of intravenous pulse methylprednisolone, or infliximab (Remicade®). Patients with mild-to-moderate coronary abnormalities should receive an antiplatelet agent such as low-dose aspirin (3–5 mg/kg/day) or clopidogrel (1 mg/kg/day up to 75 mg), and those with giant (~8 mm diameter) or multiple coronary aneurysms should receive an antiplatelet agent with an anticoagulant such as warfarin or low-molecular-weight heparin. Acute coronary obstruction requires acute thrombolytic therapy with a surgical or percutaneous interventional procedure.
Kawasaki disease (KD) is the most prevalent cause of acquired heart disease in children in industrialized nations. It usually affects previously healthy children, who may develop coronary artery aneurysms, myocardial infarction and sudden death as a result of the illness. The etiology of KD remains a major pediatric enigma, despite efforts to identify the cause over the last four decades. The scope of this article is to review the unique immunopathology of KD and to describe current treatment, which is primarily aimed at blocking or reducing inflammation in the coronary arteries. We detail the evidence in support of an infectious cause of KD, and we describe exciting new research that has led to discovery of viral-like cytoplasmic inclusion bodies in acute KD tissues; it is hoped that this finding will lead to the identification of the causative agent and to improved diagnostic and treatment modalities. In addition, we update treatment recommendations for patients with KD and its complications.
Kawasaki disease is a systemic inflammatory process that predominately affects children 6 months to 5 years of age, although younger infants and older children can also develop the illness. Autopsy studies of fatal cases have clearly demonstrated that inflammation occurs in multiple organs and tissues in KD , although inflammation of the coronary arteries is the most clinically significant aspect of the illness. The cardiovascular, respiratory, gastrointestinal, dermatologic, urinary, nervous and lymphoreticular systems can all be involved . Some theories of KD pathogenesis focus upon endothelial cell antigens as the exclusive targets of the disease process, but such models do not explain many of the pathologic findings of KD such as bronchitis, prostatitis and pancreatic ductitis. Since the clinical and epidemiologic features of KD support an infectious cause, one speculation is that the infectious agent travels from its portal of entry through the bloodstream and infects many organs and tissues, with the immune response targeting these sites of infection. The theory of KD etiology that best fits the available data is that a ubiquitous infectious agent results in asymptomatic infection in most individuals but causes KD in a subset of genetically predisposed individuals .
Although neutrophils are the predominant inflammatory cell type in the peripheral blood in acute KD, they are not the predominant cell type in the inflammatory infiltrate in KD tissues . Inflamed tissues in acute KD show large mononuclear cells, lymphocytes, macrophages and plasma cells, with a smaller number of neutrophils [1,3–5]. CD8 T lymphocytes have been found to predominate over CD4 T lymphocytes in the inflamed arterial wall , which is consistent with an immune response to an intracellular pathogen. Activated myeloid dendritic cells are present and colocalize with CD3 T lymphocytes . We showed that although IgG and IgM plasma cells are evident in the arterial wall, IgA plasma cells predominate [4,5]. The presence of a predominant IgA plasma cell response in a non-lymphoid, non-mucosal tissue such as arterial tissue is distinctly unusual. The presence of IgA deposits or IgA immune complexes in the skin or kidney is characteristic of Henoch–Schönlein purpura and IgA nephropathy, but IgA plasma cells are not observed in these tissues in these diseases. Such a prevalent IgA plasma cell response as seen in the acute stage of KD suggested stimulation of the immune system at a mucosal site, with switch of B lymphocytes to IgA. Our further studies indicated a marked IgA plasma cell response in the respiratory tract in acute KD, with IgA plasma cell infiltration of additional inflamed KD tissues such as the kidney and pancreas . The pattern of an IgA B lymphocyte and a CD8 T-lymphocyte response is consistent with an adaptive immune response to an intracellular pathogen such as a virus entering at a mucosal site .
As would be expected in any illness associated with prolonged fever and multiorgan system inflammation, many cytokines are upregulated in the peripheral blood in acute KD [7–9]. In addition, macrophages, T lymphocytes and myofibroblasts secrete VEGF, matrix metalloproteinases, and other cytokines and enzymes within the arterial wall [10,11]. Secreted enzymes destroy collagen and elastin fibers, and the structural support of the arterial wall weakens, resulting in dilatation or aneurysm formation. Angiogenesis ensues, with neovascularization, the development of small blood vessels within the damaged arterial wall . These new blood vessels secrete E-selectin and VCAM 1, and may serve to potentiate inflammatory cell infiltration by promoting endothelial cell interactions with circulating inflammatory cells . Damage to the arterial wall is patchy, with normal areas adjacent to severely inflamed areas even along the course of the same artery, which is more consistent with an inflammatory reaction to localized areas of infection than with widespread targeting of a human endothelial cell antigen.
The presence of a marked IgA plasma cell response in acute KD tissues suggested a specific immune response, potentially directed at antigens involved in disease pathogenesis, such as an inciting etiologic agent. Because antigen-driven B-lymphocyte responses are oligoclonal, restricted to specific epitopes of the antigenic target, we performed studies to determine the clonality of the IgA response in acute KD. Analysis of gene sequences in the CDR3 region of the IgA plasma cells in the acute KD arterial wall revealed a strikingly oligoclonal response, with additional features of an antigen-driven process, such as somatic mutation, apparent among the sequences . The presence of an oligoclonal IgA immune response provided us with the opportunity to prepare synthetic versions of the most prevalent IgA antibodies in the KD arterial wall. The goal of these studies was to identify antigenic targets of the KD IgA antibodies prevalent in acute KD tissues. Synthetic KD antibodies were used in immunohistochemistry experiments on acute KD and control tissues. These studies revealed the presence of a specific antigen in acute KD bronchial epithelium and in macrophages within inflamed KD tissues . We detected the antigen in bronchial epithelium in ten out of 13 patients with acute KD but in none of nine control infants, and in coronary artery aneurysms in nine out of 12 acute KD and none out of ten control coronary arteries . A panel of monoclonal synthetic KD antibodies was developed from KD IgA antibodies, and those that bound to KD tissues identified the same antigen in all cases, despite the difference in primary sequence of the antibodies, confirming the presence of an antigen-driven IgA immune response in the acute stage of KD. We found that synthetic antibodies made from IgA sequences that were more prevalent in KD arterial tissues showed stronger binding to acute KD tissues than those made from less prevalent sequences .
Using light and electron microscopy, we further studied the antigen identified by the synthetic KD antibodies. These studies showed that the antigen in acute KD bronchial epithelium is present in cytoplasmic inclusion bodies that are consistent with aggregates of viral proteins and nucleic acids . Additional studies indicate that KD cytoplasmic inclusion bodies contain RNA but not DNA, and that they can be identified in 85% of acute and late-stage KD fatalities and 25% of adult controls, suggesting that the KD agent is a previously unidentified, ubiquitous RNA virus that can result in persistent infection .
Thus, all available data indicate that IgA antibodies in KD are targeting a microbial antigen, not an autoantigen.
Asian children, particularly those of Japanese, Chinese and Korean ethnicity, have the highest incidence of KD, but all racial and ethnic groups are affected. Incidence rates of KD in Japanese children are approximately tenfold higher than those in Caucasian children [19,20], and this higher incidence holds true for Japanese children living a Western lifestyle in the USA , strongly supporting a genetic predisposition for KD. The molecular genetics of KD is an emerging field . Genomewide linkage analysis has provided potential loci of susceptibility to KD ; such studies led to identification of ITPKC, a negative regulator of T-cell activation, as a gene associated with KD susceptibility and an increased risk of coronary artery lesions in both Japanese and American children . Additional candidate genes potentially related to inflammation, apoptosis and cardiovascular pathology have recently been identified . Susceptibility to KD is highly likely to be polygenic, and additional genes conferring susceptibility to KD will be undoubtedly identified over time.
Even though the precise etiology and pathogenesis of KD are incompletely understood, current management of acute KD is based upon prospective, controlled, multicenter treatment trials that clearly demonstrated the efficacy of intravenous immunoglobulin (IVIG) and high-dose aspirin to halt inflammation and to reduce the likelihood of the development of coronary abnormalities when administered by the tenth day of illness [25,26]. Initial studies demonstrated the efficacy of IVIG when given in four-to-five daily doses . Follow-up trials clearly established the superiority of single high-dose IVIG (2 g/kg given over 10–12 h) in combination with high-dose aspirin (80–100 mg/kg/day in four divided doses) to normalize acute-phase reactants and achieve defervescence . It is commonly believed that Asian children experience excessive toxicity when 80–100 mg/kg/day aspirin is used, and 30–50 mg/kg/day dosing is commonly used in Asia. Patients should be treated by the tenth illness day when possible because very few data exist regarding therapy later than the tenth day. While patients treated later than the tenth day often benefit from therapy clinically, the ability of treatment to prevent the development of coronary changes in this situation is much less certain.
It is unclear whether IVIG and aspirin treatment of KD patients instituted earlier than the fifth illness day leads to poorer outcomes, with different patient series reaching different conclusions [27–29]. It is likely that children diagnosed and treated before the fifth day have more severe KD than those who present later for care. A meta-analysis of all Japanese and North American IVIG trials showed that coronary outcome correlates directly with the total dose per kg of IVIG administered and confirmed that total IVIG doses less than 2 g/kg are less effective in KD . Coronary outcome was independent of the initial dose of aspirin administered (30–50 or 80–100 mg/kg/day).
Patients should undergo a baseline echocardiogram early in their hospitalization, and should remain in hospital until they have been afebrile and stable for at least 24 h to ensure that they do not need retreatment .
Adding corticosteroid therapy to IVIG and aspirin for primary therapy has not been clearly shown to be beneficial. Japanese trial findings must be interpreted cautiously because the IVIG and the aspirin regimens that were utilized differ from those used in the USA and because echocardiograms were not interpreted in a blinded fashion [32,33]. The most definitive trial in this regard showed no benefit of the addition of one dose of intravenous methylprednisolone (30 mg/kg) to the standard 2 g/kg IVIG and 80–100 mg/kg/day aspirin regimen used in the USA, with the possible exception of the subgroup of patients who failed standard therapy and needed retreatment . The latter observation will require further study.
Approximately 85–95% of patients treated with IVIG and aspirin respond promptly with defervescence and subsidence of inflammatory manifestations within 48 h of the conclusion of the IVIG infusion . However, a subgroup of approximately 5–15% fails to manifest an excellent clinical response, with persistent fever or only transient improvement with recurrent fever and inflammatory signs. The laboratory correlate is a rising or persistently elevated serum C-reactive protein level. Note that the erythrocyte sedimentation rate cannot be used to follow the degree of inflammation after IVIG has been administered, as it is artifactually elevated following IVIG. In these patients, the diagnosis of KD should be reassessed, as some of these patients may not actually have KD. In those patients who are confirmed to have KD, additional anti-inflammatory therapy is needed.
In these patients, a second 2 g/kg dose of IVIG is generally effective in controlling disease activity. Up to 70–80% of nonresponders become afebrile, lose inflammatory signs and show decreased serum C-reactive protein values after a second dose of IVIG [35,36]. Some recommend a 3-day course of intravenous pulse methylprednisolone (30 mg/kg/dose) rather than a second dose of IVIG, but we prefer to utilize pulse methylprednisolone in patients who have failed two 2 g/kg doses of IVIG. In a very small number of particularly refractory patients, we have employed a slow oral steroid taper over several months once fever has resolved and the C-reactive protein levels have decreased substantially.
Another treatment modality for IVIG-refractory KD is infliximab (Remicade®, Centocor, PA, USA), a chimeric mouse–human monoclonal antibody directed at soluble and membrane-bound TNF-α. Use of this agent has been reported in an open-label study  and in a small randomized trial of patients with IVIG-refractory KD ; in the latter study there was no significant difference in outcomes between infliximab and those treated with a second dose of IVIG. The package insert for this agent includes a warning that its use may predispose to serious infections such as TB and to lymphoma.
Even more limited published experience exists with respect to other therapies for IVIG-resistant patients; cyclophosphamide with or without methotrexate has been used in a very small number of patients , and the same is true for plasma exchange.
After the acute stage of KD (when fever and inflammatory signs have resolved and C-reactive protein has normalized), patients should receive low-dose aspirin (3–5 mg/kg once daily) until their echocardiograms at 2 weeks and 6–8 weeks after illness onset are normal. To reduce the theoretical (and very remote at this low dose) risk of Reye syndrome in those taking low-dose aspirin, clopidogrel (Plavix®) may be briefly substituted for aspirin in those who develop influenza or varicella. This agent also can be used in the rare child who is allergic to, or intolerant of, aspirin. If a baseline and the 2- and 6–8-week follow-up echocardiograms fail to reveal coronary abnormalities, subsequent echocardiograms are probably unnecessary , although a 6–12-month follow-up echocardiogram is performed at many centers .
The recommended long term management of individual KD patients is best considered by the degree of risk based upon their coronary artery status, as recommended by the American Heart Association .
These patients have no need for aspirin or other antiplatelet medication beyond 2 and 3 months after the onset of KD or for restriction of physical activities. Only routine pediatric care with routine cardiovascular risk assessment is indicated.
These patients with mild coronary changes that resolved by 6–8 weeks should be treated with aspirin until the echocardiogram normalizes. There are no restrictions after 6–8 weeks; angiography is not indicated. Risk assessment and counseling are recommended at 3–5-year intervals.
These patients should receive daily low-dose aspirin (3–5 mg/kg daily) at least until regression is documented on annual echocardiographic follow-up. No restriction of activity is indicated for those under 11 years of age; in older patients, activity should be guided by the results of a biennial stress test or myocardial perfusion assay. Angiography should be performed if stenosis or ischemia is suggested.
These patients should receive long-term antiplatelet therapy (aspirin 3–5 mg/kg/day or clopidogrel 1 mg/kg/day up to adult dose of 75 mg/day) indefinitely. In addition, anticoagulation with warfarin (target INR of 2.0–2.5) or with daily subcutaneous low-molecular-weight heparin is indicated to prevent thrombosis and ischemia. Cardiac evaluation with echocardiogram and ECG should be performed every 6 months, with stress testing approximately once per year. Angiography should be performed 6–12 months after the acute stage of KD and repeated if symptoms or stress tests suggest the presence of myocardial ischemia. Activity should be restricted on the basis of stress test results and level of anticoagulation, with strenuous activity discouraged.
These patients must be evaluated emergently for possible intervention that may include rotablader angioplasty, coronary artery bypass grafting and/or stent placement, arterioplasty and even cardiac transplantation [41–43]. The arterial wall in KD tends to be more fibrotic with more myointimal thickening than observed in adults with atherosclerotic coronary occlusion, which is more often associated with acute thrombosis. Nevertheless, fibrinolytic therapy with tissue plasminogen activator, urokinase or streptokinase is indicated for acute coronary occlusion in KD when coronary thrombosis is suspected, even though the pathophysiology is somewhat different from that of adults in whom these protocols are generally developed . Abciximab, a glycoprotein IIb/IIIa inhibitor, has been used in a handful of KD patients with acute coronary thrombosis . Balloon angioplasty procedures in KD have been associated with high rates of recurrent stenosis or neoaneurysms related to the very high balloon pressures required and should generally be avoided . A recent publication from Japan has documented the 25-year outcomes of coronary artery bypass surgery in 114 KD children and adolescents from 1 to 19 years of age, showing that internal thoracic arterial grafts are optimal and that 20- and 25-year survival rates were 95% .
The following model of KD pathogenesis incorporates clinical, epidemiologic, immunologic and pathologic features of KD. KD is probably the result of infection with a ubiquitous infectious agent, most probably a single virus or a group of closely related viruses, which results in asymptomatic infection in most individuals but leads to KD only in a very small subset of genetically predisposed individuals. There is no precedent for a febrile systemic illness of childhood with such strikingly consistent clinical features as KD to be the result of infection with multiple disparate infectious agents. In the past, such claims were made for polio, AIDS, roseola and fifth disease prior to identification of their causative agents, all of which were determined to be single infectious agents or a group of very closely related agents.
In a child with KD, the infectious agent probably enters through the respiratory tract, infecting ciliated bronchial epithelium, and both innate and adaptive immune responses ensue, initially in the respiratory tract, with B lymphocytes switching to IgA lymphocytes. The agent enters a small subset of macrophages traveling through the bronchial submucosa and travels within the macrophage through the circulation. As these infected macrophages traverse tissues as part of normal immune surveillance, the agent is able to infect its target tissues such as the coronary artery. Antigen is processed by local lymphoid tissues, and CD8 T lymphocytes and IgA B lymphocytes with a good match for the antigen are expanded. These immune cells leave the lymphoid tissue, traveling back to infected tissues and contributing to the inflammatory response within the tissues. Additional macrophages and some neutrophils also enter inflamed tissues. Under the influence of local cytokines produced by macrophages in the tissues, IgA B lymphocytes terminally differentiate into plasma cells and produce antigen-specific antibodies locally. CD8 T lymphocytes act to eradicate infected cells by cytotoxic mechanisms. Within the arterial wall, the production of matrix metalloproteinases and other enzymes by macrophages results in destruction of collagen and elastin fibers. The wall can lose its structural integrity and dilate or balloon, forming an aneurysm. The immune response is ultimately successful in controlling the pathogen, particularly if IVIG is given, but damage to the coronary arteries may have already occurred.
If KD is the result of infection with a ubiquitous infectious agent as described here, immunoglobulin pooled from adult blood donors should contain antibody to the infectious agent. This suggests a mechanism of action of IVIG in KD. A mechanism of antibody-dependent cellular cytotoxicity, in which immune IgG binds infected cells through its antibody-binding (Fab) end and effector cells through its antibody-tail (Fc) end, leading to destruction of the infected cells, could explain the dramatic efficacy of IVIG in KD. This mechanism appears to account for the similar dramatic clinical response to immune γ-globulin in Argentine hemorrhagic fever, an arenavirus infection of humans [47–49].
Over the next 5 years, we anticipate identification of additional genes influencing susceptibility to KD and to the development of coronary artery aneurysms. Expression profiles of genes involved in KD pathogenesis will be determined in the peripheral blood and tissues of KD patients, which is likely to provide new therapeutic approaches. Use of bioinformatics analysis of DNA and RNA sequences from KD tissues may allow for identification of the elusive etiologic agent of KD. This would truly revolutionize the field and allow for the development of a diagnostic test, improved therapies and prevention of this potentially devastating illness of childhood.
For reprint orders, please contact moc.sweiver-trepxe@stnirper
Financial & competing interests disclosure
Anne H Rowley is supported by National Institutes of Health grant HL63771 and the Kawasaki Disease Fund of the Children’s Memorial Hospital. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Anne H Rowley, Northwestern University Feinberg School of Medicine, Pediatrics, Morton 4-685B, 310 E Superior St, Chicago, IL 60611, USA, Fax: +1 312 503 1181.
Stanford T Shulman, The Division of Infectious Diseases, The Children’s Memorial Hospital, 2300 Children’s Plaza #20, Chicago, IL 60614, USA, Tel.: +1 773 880 4187, Fax: +1 773 880 8226.
Papers of special note have been highlighted as:
• of interest
•• of considerable interest