Chapter 167. Kawasaki Disease

Kawasaki disease (KD), the leading cause of acquired heart disease in children in developed nations, is a multisystem inflammatory illness that particularly affects blood vessels, especially the coronary arteries. About 25% of untreated children develop coronary artery abnormalities, including dilatation and aneurysms that can lead to myocardial infarction and sudden death.1,2 The etiology is unknown, but clinical and epidemiologic data support an infectious cause. In KD, an intense inflammatory cell response develops in a wide array of organs and tissues3; in medium-sized arteries such as the coronary arteries, this response can damage collagen and elastin fibers in the vessel walls and lead to loss of their normal structural integrity with resultant ballooning or aneurysm formation. Despite a limited understanding of KD pathogenesis, a very effective therapy exists in the form of IVIG and aspirin; when given in the first 10 days of fever, this therapy reduces the prevalence of coronary artery abnormalities from 25% in untreated patients to 5% in those who receive the therapy.4 Because the etiology is unknown, no diagnostic test exists, and the diagnosis is made clinically. Classic KD is diagnosed in a patient with prolonged fever and four of five other clinical findings (Box 167-1). However, incomplete forms of illness are well-recognized, in which a child manifests prolonged fever with fewer than four other clinical features of the illness and subsequently develops coronary artery abnormalities. The existence of these incomplete forms of illness results in a major diagnostic dilemma for physicians in establishing the diagnosis accurately in children with prolonged fever of uncertain cause.

KD is named for Dr. Tomisaku Kawasaki, a Japanese pediatrician who first recognized the clinical features of the illness. He described 50 cases of a new illness in 1967 in the Japanese-language literature that he termed mucocutaneous lymph node syndrome.5 It was not until later that some of the children with this newly described illness experienced sudden death; autopsy revealed myocardial infarction from thrombosis of coronary artery aneurysms. Prior to Dr. Kawasaki's description of the clinical features, KD was recognized only by pathologists at autopsy, who called the disease “infantile periarteritis nodosa.”6 Dr. Kawasaki described the illness in the English-language literature in 1974; this report was closely followed by a description of the same illness, observed independently in the early 1970s in Hawaii by Dr. Marian Melish and colleagues.7,8 Since that time, it has become clear that although the attack rate of KD is highest in Asian, particularly Japanese, Korean, and Chinese children, all racial and ethnic groups around the world are affected by the illness.

KD is predominantly an illness of young children, with 80% of cases occurring in children ages 6 months to 5 years.9,10 However, infants less than 6 months of age can be affected, often manifest incomplete forms of the illness, and can have particularly severe KD.11,12 Similarly, KD can occur in older children and teenagers, in whom the diagnosis is often delayed, and who may also have more severe KD with a higher prevalence of coronary artery abnormalities.13–15 Therefore, the diagnosis must be considered in all pediatric age groups.

Boys are more commonly affected than girls at a ratio of 3:2. The incidence of KD is approximately tenfold higher in Japanese than in Caucasian children.9,10 About 1 in 100 Japanese children develop KD by the age of 5 years; the peak age of illness is 9–11 months of age in both Japan and the United States.9,16 The higher attack rate in Japanese children persists in those who adopt a Western diet and lifestyle, and is likely related to a genetic predisposition to KD among Asian children.17 The risk of KD in siblings is tenfold higher than in the general population, and the incidence of KD in children born to parents who had KD is twice as high as in the general population.18,19 Recurrence is rare, occurring in about 3% of cases in Japan.16

Many epidemiologic features of KD suggest an infectious agent is the cause. Among these are the well-described epidemics of illness17,20–22 and the geographic wave-like spread of illness during an epidemic, compatible with spread of an infectious agent.22 Cases in the United States are more common in the winter and spring.9

The etiology of KD remains unknown. The hypothesis that best fits the available clinical, immunologic, and epidemiologic data is that KD results from infection with a ubiquitous etiologic agent that usually results in asymptomatic infection, but causes KD in a small subset of genetically predisposed individuals. Genetic predisposition is a common theme in susceptibility and host response to infectious diseases. It is likely that KD is polygenic. Intensive investigation of genetic factors influencing susceptibility to KD is ongoing by several international collaborative groups. A functional polymorphism in the ITPKC gene, which is a negative regulator of T-cell activation, has been associated with KD susceptibility and risk of developing coronary artery abnormalities23; it is likely that other susceptibility genes will be discovered in the near future.

KD is often referred to as an autoimmune illness, but there is no compelling reason to classify it as such. The spontaneous resolution of the febrile phase of illness, the rarity of recurrent KD, and the fact that the highest incidence occurs in male infants, a group who rarely develop autoimmune diseases, all argue strongly against this possibility. Immune complexes, although detectable in serum in the subacute phase, are not deposited in tissues and do not appear to play a prominent role in pathogenesis. Although inflammation in KD results in tissue damage, the damage is likely to be primarily the result of an immune response targeting antigen(s) of an infectious agent in the tissue.

Many cytokines are upregulated in KD, which has led some investigators to propose a superantigen as the cause of the illness. However, cytokines are upregulated in many infectious diseases, and the cardinal feature of a superantigen-mediated illness, the paralysis of adaptive immunity,24 is not observed in acute KD. Oligoclonal, antigen-driven CD8 T lymphocyte,25 IgA plasma cell,26 and IgM B lymphocyte27 adaptive immune responses are detected in acute KD.

In KD, an intense inflammatory cell infiltrate is observed in many organs and tissues, particularly in the medium-sized arteries.3 These inflammatory cells secrete matrix metalloproteinases and other enzymes that can disrupt collagen and elastin fibers in the coronary arteries, impairing the integrity of the arterial wall and resulting in ballooning or aneurysm formation.28 The initial insult that causes this immune response is unknown, but the nature of the inflammatory cell infiltrate provides clues. Although neutrophils contribute to the very early inflammatory-cell infiltrate in KD tissues, they are quickly replaced by large mononuclear cells and lymphocytes.2 By two weeks after the onset of fever, the inflammatory infiltrate in the arterial walls largely consists of T lymphocytes (predominately CD8 T lymphocytes), macrophages, plasma cells (predominately of the IgA isotype), and eosinophils.29,30 IgA plasma cells are increased in the respiratory tract, particularly in a peribronchial distribution.31 The presence of IgA plasma cells and CD8 T cells as primary components of the inflammatory infiltrate in acute KD suggests an immune response to an intracellular pathogen with a respiratory portal of entry.29 Sequencing of IgA genes in the arterial walls in acute KD shows an oligoclonal, or antigen-driven, response.26 Synthetic versions of these oligoclonal IgA antibodies identify antigen in acute KD tissues.32 Light and electron microscopic studies show that the antigen detected by KD synthetic antibodies resides in cytoplasmic inclusion bodies that are consistent with aggregates of viral protein and RNA.33,34 Identification of the specific proteins and RNA in the inclusion bodies should lead to identification of the etiologic agent of KD. These studies are hampered by the lack of unfixed tissue specimens from fatal cases, and by the likelihood that the etiologic agent is an RNA virus without significant homology to known viruses, thereby making its RNA sequence difficult to identify.34 It is hoped that advances in high throughput sequencing and bioinformatics analysis will yield the solution to the mystery of the etiology of KD in the near future.

Classic Diagnostic Criteria for KD

A diagnosis of classic KD is made in the presence of prolonged fever ≥5 days, with four of the following five clinical features in the absence of another explanation for the illness: (1) nonpurulent bulbar conjunctival injection (Fig. 167-1); (2) red, swollen, dry lips, which may crack and bleed (Fig. 167-1); (3) redness and swelling of the hands and feet; (4) rash; and (5) cervical lymphadenopathy, ≥1.5 cm in diameter. Experienced physicians can make a diagnosis of KD before the fifth day of fever in children with classic features. A patient with prolonged fever and fewer than four of the other features of the illness can be diagnosed with KD if coronary artery abnormalities develop. Incomplete (or atypical) KD refers to children with prolonged fever and fewer than four of the other features of illness who have a laboratory profile compatible with KD. Such patients should undergo echocardiography and be considered for treatment with IVIG because KD patients, particularly infants, do not always manifest classic diagnostic criteria during the acute febrile phase of illness, yet can develop coronary artery abnormalities (Box 167-1).

History

KD should be considered in the differential diagnosis of any child with prolonged fever without other explanation. In KD, all clinical features may not be present simultaneously. Therefore, it is important to query parents and physicians who saw the patient during the course of a prolonged febrile illness as to the presence of the other five clinical features of the illness: (1) conjunctival injection, (2) oral mucosal changes, (3) changes of the hands and feet, (4) rash, and (5) cervical adenopathy. Children with KD often have significant enough swelling and discomfort of the hands and feet that they will refuse to pick up objects or to walk. This is not commonly observed in children with most other illnesses in the differential diagnosis of KD and can provide an important clue to the diagnosis. Similarly, extreme irritability is common in KD and not as common in most other illnesses in the differential diagnosis. Without specific therapy, fever in KD is daily, high spiking, intermittent, and lasts for one to two weeks. The illness is often divided into three stages: (1) the acute febrile phase, (2) the subacute phase (that begins when fever resolves and continues until all clinical features have normalized), and (3) the convalescent phase (that follows the subacute phase and continues until the erythrocyte sedimentation rate (ESR) normalizes, usually at 6–8 weeks after the onset of fever).

Cutaneous Features

Rash is commonly observed in KD, is most pronounced on the trunk and extremities, and generally takes one of three forms: (1) an erythematous maculopapular exanthema (Fig. 167-2), (2) erythema multiforme with typical target lesions, or (3) a scarlatiniform exanthema. Bullae, vesicles, and ulcerative lesions are not observed, but a fine micropustular rash, especially on the extensor surfaces, occasionally occurs. The rash may be pruritic. In the acute febrile phase of illness, groin erythema and desquamation are commonly observed (Fig. 167-3), and can be mistaken for candidal diaper dermatitis or even a staphylococcal scalded skin syndrome. The skin changes in the groin can be seen both in children in diapers and toilet-trained children. Classic periungual desquamation of the fingers and toes does not begin until the second to third week after fever begins, and can progress to involve the entire hand and foot (Fig. 167-4); treatment should be administered well before its appearance. In the third to sixth week after illness, transverse lines across the fingernails (Beau lines) are often apparent. These grow out with the nail.

In countries where BCG vaccine is routinely administered, such as in Japan, a common finding in children with KD is erythema and swelling at the site of BCG vaccine; the mechanism is unknown but the process resolves with treatment of KD. Although KD results in a vasculitis affecting all arteries and veins in the body, blood vessels in the skin are not prominently affected, and skin biopsy is not useful for diagnosis because pathologic findings are nonspecific.

Related Physical Findings

Conjunctival injection in KD is bilateral and nonexudative. There may be limbal sparing (Fig. 167-5). Photophobia is a common accompanying feature. Oral findings include red, swollen, dry, cracked lips that may bleed, a “strawberry” tongue, and erythema of the mouth and throat. Oral ulcers are not a feature of KD. Palmar and plantar erythema is a common feature, and there can be an abrupt transition from marked erythema to normal skin at the wrists and ankles. The hands and feet can be edematous and painful. Cervical adenopathy is the least commonly observed clinical feature, occurring in only about 75% of children with classic KD, but can be the most prominent feature in a subset of patients, who are often treated with multiple different courses of antibiotic therapy without improvement before the correct diagnosis is made.35 Cervical adenopathy is usually unilateral, may or may not be associated with superficial erythema and/or tenderness to palpation, and is nonfluctuant. Because KD is a multisystem inflammatory process, many organs and tissues are involved in the inflammatory process, leading to a variety of associated clinical features (Box 167-2). In particular, arthritis can occur during the acute febrile phase, involving the small interphalangeal joints and larger joints, or may occur during the subacute phase of illness, usually involving the larger joints such as the knees and ankles. Aseptic meningitis is a common finding in patients who undergo lumbar puncture.

In the absence of knowledge of the etiology of KD, a diagnostic test is not available. Laboratory findings in acute KD are nonspecific but quite characteristic. A complete blood count reveals either a normal or elevated white blood cell count with a neutrophil predominance. A low white blood cell count with lymphocyte predominance would be distinctly unusual in KD. A normochromic, normocytic anemia can be present, and resolves spontaneously with resolution of KD. The platelet count is normal in the first week of illness, although thrombocytopenia has been reported to be associated with a more severe outcome. Thrombocytosis, with platelet counts sometimes exceeding 1,000,000/mm3, is characteristic of the subacute phase of KD, peaking in the second to third week after the onset of fever. This feature, like periungual desquamation, is not useful in making a diagnosis of KD in the first week of fever. Patients with anemia and low albumin levels may be at higher risk of developing coronary artery disease. A mild elevation of the liver transaminases is commonly observed in acute KD. Occasionally, obstructive jaundice occurs. Gallbladder hydrops, with accompanying right upper quadrant abdominal pain, resolves spontaneously and does not require surgical intervention. Sterile pyuria is also commonly observed. Acute-phase reactants such as the C-reactive protein (CRP) and the ESR are characteristically elevated in patients with acute KD, and the CRP is sometimes used to follow clinical response in patients refractory to IVIG therapy. Once IVIG is given, the ESR cannot be used to follow clinical response, because IVIG itself transiently increases the ESR.

To determine whether a child with KD develops coronary artery dilatation, an echocardiogram should be performed at diagnosis, at 2–3 weeks after fever onset, and at 6–8 weeks after fever onset.36 A complete blood count and CRP or ESR are also repeated at 2–3 weeks and 6–8 weeks after onset to monitor for resolution of inflammation. The peak time to detect coronary artery dilatation is 2–3 weeks after onset of fever, during the subacute phase of illness. Some centers perform an additional echocardiogram at 1 year after onset, but it is rare to find any evidence of coronary artery abnormalities at 1 year that were not present within the first two months after fever began. Electrocardiogram most often shows a prolonged PR interval and/or nonspecific ST- and T-wave changes.

The KD child with prolonged high fever, marked conjunctival injection, red, swollen hands and feet, erythematous rash, and red, cracked, bleeding lips who has a markedly elevated ESR and/or CRP and an elevated peripheral white blood cell count with a neutrophil predominance has a very distinctive illness and usually does not pose any difficulties in diagnosis. In other patients who have only some of the clinical features of KD, or in whom clinical findings are not as dramatic, other diagnoses must be carefully considered (Box 167-3).

In areas where measles is still prevalent, differentiating the two disorders can be difficult. Classically, patients with measles have rash that begins on the face behind the ears and Koplik spots in the mouth; neither of these features is observed in KD. Later in the course of measles infection, the rash becomes diffuse and Koplik spots are no longer visible. Conjunctival injection and edema of the hands and feet can be observed in both disorders. In uncomplicated measles, the peripheral white blood cell count and the erythrocyte sedimentation rate are generally low. The measles IgM antibody is virtually always positive by the appearance of the measles rash, and is the best single test to differentiate the two conditions.

Group A streptococcal infection should be considered in the differential diagnosis of a KD patient with a scarlatiniform rash, and can be excluded by a negative throat culture. Diagnostic uncertainty could arise in a KD patient who is a group A streptococcal carrier. Administration of antibiotic therapy followed by reevaluation in 24–48 hours generally clarifies the diagnosis; children with group A streptococcal pharyngitis have a rapid response to therapy, while antibiotics are ineffective in KD. Adenovirus infection can mimic KD. The presence of exudative conjunctivitis and exudative pharyngitis suggests adenovirus as the most likely diagnosis. Other viruses can occasionally cause prolonged fever, such as enterovirus; in differentiating KD patients from those with uncomplicated viral infection, laboratory tests such as the ESR, CRP, and urinalysis can be helpful, as pyuria and markedly elevated acute-phase reactants are more typical of KD.

Drug hypersensitivity reactions can mimic KD. Oral and mucosal ulcers observed in Stevens–Johnson syndrome are absent in KD. Edema of the face, particularly around the eyes, is much more suggestive of drug reaction than KD. In general, the ESR and CRP are either normal or only mildly elevated in drug hypersensitivity reactions.

Staphylococcal scalded skin syndrome (SSSS) is easily distinguished by the classic finding of painful skin in SSSS, which is not observed in KD, and by Nikolsky sign, which is present in SSSS and absent in KD.

Although hypotension is unusual in KD, it occurs occasionally,37,38 and toxic shock syndrome is in the differential diagnosis of such patients. Renal involvement and elevated creatinine phosphokinase are more likely to be observed in toxic shock syndrome than in KD. IVIG administered for a possible diagnosis of KD may serve as adjunctive therapy for toxic shock syndrome.

Juvenile rheumatoid arthritis is a diagnosis of exclusion, and an occasional patient with this disorder might initially be diagnosed with and treated for incomplete KD. The correct diagnosis may become apparent when fever recurs following reduction of high-dose aspirin therapy, implying an original clinical response to aspirin rather than to IVIG.

The following are in the differential diagnosis of KD in the appropriate epidemiologic context: (1) leptospirosis, which predominately occurs in tropical climates; (2) Rocky Mountain spotted fever, which tends to be highly regional in incidence; and (3) mercury poisoning, now quite rare.

The American Heart Association Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease has published an algorithm to assist the clinician in diagnosing incomplete KD.36 This algorithm emphasizes the importance of combining clinical features, laboratory findings, and echocardiographic findings in making a diagnosis (see Box 167-4). In difficult cases, consultation with a KD expert should be considered. If the patient is not treated for KD but develops typical periungual desquamation after fever has resolved, and an alternative diagnosis that could explain this clinical finding has not been established (such as scarlet fever), an echocardiogram should be repeated. Infants ≤6 months of age can have mild or subtle clinical findings with KD, but have a high risk of developing coronary artery abnormalities. Therefore, infants with fever for a week or more without other explanation should have laboratory testing performed. If evidence of an inflammatory process is present, an echocardiogram should be ordered and KD considered, even in the absence of other clinical features of the illness.

KD can result in myocardial infarction, aneurysm rupture, and sudden death. Myocardial infarction can occur from thrombosis of an aneurysm, which is most common in the first few months after the onset, or from coronary artery stenosis, which occurs months to years following the illness. Aneurysm rupture is less common and usually occurs within the first month after the onset. More than 50% of KD patients have myocarditis during the acute febrile phase, manifested clinically as tachycardia out of proportion to fever, which generally improves rapidly with IVIG therapy. In the acute phase of illness, a pericardial effusion can be present; this resolves spontaneously. In patients with the most severe coronary artery aneurysms, aneurysms of other medium-sized arteries can also be observed, most commonly in the iliac, femoral, and axillary arteries. Rarely, valvulitis significant enough to require valve replacement has been reported. Although KD is a systemic inflammatory disease affecting multiple organs and tissues, there are no known long-term consequences of the disease outside of the heart and blood vessels.

About 85% of KD children treated with IVIG and aspirin within the first 10 days of illness respond with rapid resolution of fever and other clinical signs. Only 5% of KD patients treated within the first 10 days of fever develop coronary artery abnormalities.4,39 Therefore, the vast majority of KD patients who are promptly diagnosed and treated do well, without developing cardiac complications. However, about 15% of KD children treated within the first 10 days of illness continue to have fever following a single infusion of IVIG with aspirin and require additional therapy; these patients have a higher risk of developing coronary artery abnormalities.40,41 Long-term studies show that coronary artery abnormalities regress in 50% of patients by 1–2 years after the onset.42 The larger the aneurysm, the less likely it is to regress. Giant coronary artery aneurysms, classically defined as those ≥8-mm internal diameter, almost never regress.43 Regression of an aneurysm usually is accompanied by intimal proliferation, which can be significant enough in large aneurysms to lead to coronary artery stenosis and an increased risk of myocardial infarction.44 In young infants and children, myocardial infarction usually presents as shock, emesis, and/or abdominal pain.44 Patients with significant coronary artery disease may require catheter intervention procedures,45 coronary artery bypass surgery,46 or rarely, heart transplantation.47 Cardiovascular risk assessment should be performed at intervals dependent upon the severity of coronary artery abnormalities.36

A single infusion of 2 g/kg IVIG with aspirin (80–100 mg/kg/day given every 6 hours orally) should be administered to children with acute KD as soon as possible after diagnosis (Box 167-5).39 This regimen, when administered to children with KD within the first 10 days of fever, reduces the prevalence of coronary artery abnormalities from 25% in untreated patients to 5% in those who receive the therapy. Most children experience rapid resolution of fever and other clinical signs following treatment, as well as improvement in acute-phase reactants. The mechanism of action of IVIG in KD is unknown.

Aspirin is given in high doses during acute KD for anti-inflammatory effect. It is generally continued at 80–100 mg/kg/day until the 14th illness day, or until the patient has been afebrile for at least 2–3 days. Aspirin is then reduced to 3–5 mg/kg/day given in a single daily dose, for its antithrombotic effect. Aspirin is discontinued at 6–8 weeks after onset if all echocardiograms have been normal and acute-phase reactants have normalized. If coronary artery abnormalities develop, low-dose aspirin is continued. Depending upon the severity of coronary artery disease, other therapies such as clopidogrel or warfarin may be indicated36; such decisions should be made in consultation with a pediatric cardiologist. Intravenous methylprednisolone is not indicated for primary therapy of acute KD.48

IVIG and aspirin is also given to KD patients who present after the tenth day of illness if fever and/or clinical and laboratory signs of ongoing inflammation are present, although the efficacy of the therapy in this clinical situation is uncertain. About 15% of acute KD patients do not respond to initial therapy; optimal therapy for these patients with “refractory” KD is unknown. Most of these patients will respond to a second 2 g/kg IVIG infusion.40,41 In those who remain febrile after a second IVIG infusion, intravenous methylprednisolone once daily for 3 days or infliximab have been given with apparent success.49–51 More specific therapy awaits the identification of the cause of KD.

Prevention of KD will not be possible until the etiology of the illness is identified.