Kawasaki disease is a systemic vasculitis and the leading cause of acquired heart disease in North American and Japanese children. The epidemiology, cause, and clinical characteristics of this disease are reviewed. The diagnostic challenge of Kawasaki disease and its implications for coronary artery outcomes are discussed, as are the recommended treatment, ongoing treatment controversies, concerns associated with treatment resistance, and the importance of ongoing follow up.
Historical perspectives
Kawasaki disease (KD) is a systemic vasculitis of unknown cause and the leading cause of acquired heart disease in North American and Japanese children. This disease was first reported in 1967 by Dr Tomisaku Kawasaki. His original case series of pediatric acute mucocutaneous lymph node syndrome described 50 Japanese patients with the clinical signs and symptoms that we now refer to as KD. In 1965, Dr Noboru Tanaka was the first pathologist to recognize the potentially serious and fatal complications of this disease when he discovered coronary artery thrombosis in a child who died unexpectedly. Dr Takajiro Yamamoto was the first physician to note cardiac complications in nonfatal cases. He described electrocardiogram abnormalities in these patients and published this finding in a report in 1968.
In 1970, the first Japanese nationwide epidemiologic survey of KD was undertaken. This study clearly established the link between KD and coronary vasculitis. Although KD was first described in Japan, similar cases were being seen around the world in the 1960s and 1970s. The first North American description was published in 1976 from Hawaii.
Epidemiology
There have been 20 nationwide epidemiologic surveys of KD in Japan. The average annual incidence in 2005 and 2006 was 184.6 per 100 000 children less than 5 years of age. In 2008, in the latest survey, the incidence increased to 218.6 per 100 000 children less than 5 years of age. The latest incidence is even higher than the rates seen in the epidemic years of 1979, 1982, and 1986. As in previous surveys, the incidence was highest in children aged 6 to 11 months. Patients with affected siblings were seen in 1.4%; and 0.7% of patients had at least 1 parent with a previous history of KD. Recurrence of the disease was seen in 3.5% of patients.
KD has been described in more than 60 countries. The annual incidence of KD is highest in Asian countries. After Japan, Korea (105/100 000) and Taiwan (68/100 000) have the next highest incidences. It is presumed that the incidence is high in other Asian countries but cases may not be as well documented as in Japan.
In the United States, the incidence in the year 2006 was estimated to be 20.8/100,000 in children less than 5 years of age. Race-specific incidence rates showed that the disease was most common among Americans of Asian and Pacific Island descent. In Ontario, the most populous province in Canada, the incidence increased over time from 14.4/100 000 in children less than 5 years of age in 1995 to 1997 to 26.2/100 000 in 2004 to 2006. This is most likely because of better disease recognition over the years, particularly for incomplete cases. In this study, a seasonal pattern was observed, with an increase in cases in the late fall and winter, similar to Japan. The male/female ratio was 1.62:1. Children less than 5 years of age made up 73% of cases seen. Although race was not evaluated, Ontario has a large Asian population and it was suggested that the incidence of KD in Ontario, and possibly of Canada, may be one of the highest outside Asia. Other reported incidences include Ireland (15.2/100 000), England (8.1/100 000), New Zealand (8.0/100 000), and Australia (3.7/100 000).
Epidemiology
There have been 20 nationwide epidemiologic surveys of KD in Japan. The average annual incidence in 2005 and 2006 was 184.6 per 100 000 children less than 5 years of age. In 2008, in the latest survey, the incidence increased to 218.6 per 100 000 children less than 5 years of age. The latest incidence is even higher than the rates seen in the epidemic years of 1979, 1982, and 1986. As in previous surveys, the incidence was highest in children aged 6 to 11 months. Patients with affected siblings were seen in 1.4%; and 0.7% of patients had at least 1 parent with a previous history of KD. Recurrence of the disease was seen in 3.5% of patients.
KD has been described in more than 60 countries. The annual incidence of KD is highest in Asian countries. After Japan, Korea (105/100 000) and Taiwan (68/100 000) have the next highest incidences. It is presumed that the incidence is high in other Asian countries but cases may not be as well documented as in Japan.
In the United States, the incidence in the year 2006 was estimated to be 20.8/100,000 in children less than 5 years of age. Race-specific incidence rates showed that the disease was most common among Americans of Asian and Pacific Island descent. In Ontario, the most populous province in Canada, the incidence increased over time from 14.4/100 000 in children less than 5 years of age in 1995 to 1997 to 26.2/100 000 in 2004 to 2006. This is most likely because of better disease recognition over the years, particularly for incomplete cases. In this study, a seasonal pattern was observed, with an increase in cases in the late fall and winter, similar to Japan. The male/female ratio was 1.62:1. Children less than 5 years of age made up 73% of cases seen. Although race was not evaluated, Ontario has a large Asian population and it was suggested that the incidence of KD in Ontario, and possibly of Canada, may be one of the highest outside Asia. Other reported incidences include Ireland (15.2/100 000), England (8.1/100 000), New Zealand (8.0/100 000), and Australia (3.7/100 000).
Cause
The cause of KD remains unknown. It is suspected that there is activation of the immune system by an infectious trigger in a genetically susceptible host. An infectious cause is suspected for several reasons. First, the clinical characteristics of KD resemble an infection and the illness is self-limited. Second, the epidemiologic features, such as age of affected children, seasonality of cases, and occurrence of community outbreaks and epidemics, are all consistent with an infectious trigger. However, no known infectious agent has been consistently found. Genetics may explain why certain ethnicities are at increased risk. This risk persists despite a move from 1 country to another. As well, the risk of disease in a child born to a parent with a history of KD, or in siblings of affected children, is higher than in the general population.
There is still significant controversy about the mechanism of immune system activation in patients with KD. Some investigators believe that a bacterial superantigen leads to massive stimulation of T lymphocytes. Others suggest that an oligoclonal IgA immune response is occurring rather than a polyclonal one. This theory is supported by the discovery of IgA plasma cells infiltrating the coronary artery aneurysms in patients who died in the acute phase of KD, as well as the detection of virallike cytoplasmic inclusion bodies in ciliated bronchial epithelial cells of patients with KD.
Recent data have suggested that T-cell activation is important in determining the susceptibility and severity of KD. A genetic association study identified a polymorphism in the inositol 1,4,5-triphosphate 3-kinase C (ITPKC) gene on chromosome 19q13.2, which acts as a negative regulator of T-cell activation and may contribute to immune hyperreactivity in KD. This polymorphism was significantly associated with KD susceptibility, and with an increased risk of coronary artery abnormalities (CAA), in both Japanese and American children. An animal mouse model of KD has also identified regulation of T-cell activation as a critical determinant of coronary disease.
Multiple studies are under way to identify genetic markers that influence disease susceptibility, disease severity, and treatment resistance in KD. Genetic studies to date have made significant contributions to this field, and it is hoped that with further studies, the cause of KD will also be elucidated.
Clinical, laboratory, and cardiac features
Diagnostic Criteria
The diagnosis of KD is made clinically. There is currently no diagnostic test for this disease. Clinical criteria were established to help clinicians make the diagnosis of KD ( Box 1 ). The diagnostic criteria can be applied once other diseases with similar findings have been excluded ( Box 2 ). However, it is not uncommon to document a concomitant infection in patients with KD ; in 1 study, 33% of children diagnosed with typical KD had a confirmed concomitant infection. Therefore, suspicion for KD needs to remain high, especially if the documented infection cannot explain all the clinical features.
Classic/Complete KD:
- •
Fever for ≥5 days and ≥4 of the 5 following clinical criteria:
- ○
Bilateral, nonexudative, bulbar conjunctivitis
- ○
Oropharyngeal changes with any of the following:
- ▪
Strawberry tongue
- ▪
Diffuse erythema of the oropharyngeal mucosa
- ▪
Erythema or cracking of the lips
- ▪
- ○
Cervical lymphadenopathy (>1.5 cm diameter, usually unilateral)
- ○
Polymorphous rash
- ○
Peripheral extremity changes with any of the following:
- ▪
Erythema or edema of the palms or soles
- ▪
Periungal desquamation in the subacute phase
- ▪
- ○
Incomplete KD:
- •
Fever ≥5 days with 2 or 3 of the above clinical criteria
Atypical KD:
- •
This term is used for patients who fulfill criteria for KD but who have a clinical feature that is not usually seen with KD (eg, a patient with KD with renal impairment)
Infectious:
Viruses : measles, adenovirus, Epstein-Barr virus, enterovirus, influenza, roseola infantum
Bacterial illnesses : cervical adenitis, scarlet fever, staphylococcal scalded skin syndrome, toxic shock syndrome, leptospirosis
Rickettsial illness : Rocky Mountain spotted fever
Immune-mediated:
Stevens-Johnson syndrome, serum sickness, rheumatic fever, systemic onset juvenile idiopathic arthritis, other systemic vasculitis (eg, polyarteritis nodosa), connective tissue diseases (eg, systemic lupus erythematosus)
Hereditary autoinflammatory syndromes:
Tumor necrosis factor (TNF) receptor associated periodic syndrome, hyper-IgD syndrome, cryopyrin-associated periodic syndrome
Poisoning:
Mercury
The fever of KD is usually high and unresponsive to antibiotics, and often to antipyretics as well. The clinical features may not all be present at 1 time and therefore, it is important to inquire about these features on history. At times, the diagnosis may become clear as one notes the evolution of symptoms over time. Therefore, it is important to reevaluate a child with a persistent fever or a fever without a focus or a child whose fever is not responding to antibiotics, because the diagnosis may become clear on reevaluation. If untreated, the febrile illness typically lasts 10 to 14 days.
Conjunctivitis, oropharyngeal changes, and rash are the most common clinical features. The conjunctivitis is nonexudative and primarily involves the bulbar rather than the palpebral conjunctivae ( Fig. 1 A). Oropharyngeal changes may include erythema, cracking, fissuring, or bleeding of the lips (see Fig. 1 B), strawberry tongue (see Fig. 1 C), and diffuse erythema of the oropharynx without exudate. The rash is polymorphous. The most common rash is a nonspecific, diffuse maculopapular eruption. Many different rashes have been described with KD, except for bullous or vesicular eruptions. Accentuation of the rash in the perineal region with desquamation occurs frequently. Unlike the periungual desquamation that occurs in the subacute phase (see later discussion), the perineal desquamation occurs during the acute phase.
Peripheral extremity changes can include erythema or edema of the palms or soles in the acute phase ( Fig. 2 A) and periungual desquamation in the subacute phase (see Fig. 2 B). In 1 North American series, 68% had desquamation on follow-up and this correlated with having had peripheral extremity changes in the acute phase. It is important to evaluate the periungual region even in the acute phase, because periungual lifting, or detachment of the skin beneath the nail plate (hyponychium) without frank desquamation, can occur within the febrile period and may help when evaluating a child with prolonged fever. Within 1 to 2 months after the onset of fever, a variety of nail abnormalities have been described in these patients of which Beau’s lines (deep transverse grooves across the nails) are the most characteristic (see Fig. 2 C).
Cervical adenopathy is the least common clinical manifestation of KD. It is usually unilateral, with a diameter greater than 1.5 cm, and the node is usually firm but nonfluctuant ( Fig. 3 ). It can be confused with bacterial cervical adenitis and the diagnosis of KD needs to be considered in children with this diagnosis who are not responding to antibiotics. In 1 study, approximately 12% of patients had fever and cervical adenopathy as the initial presenting features of KD.
Other Clinical Features
Other clinical features may be seen besides those included in the diagnostic criteria. These features may help support the diagnosis of KD. Severe irritability occurs commonly. Aseptic meningitis may also be seen. Mild acute iridocyclitis or anterior uveitis may be noted on slit lamp examination. Obtaining an ophthalmologic evaluation may be warranted in those who do not fulfill criteria. Arthritis and arthralgias can occur, and in 1 study, the reported prevalence of arthritis was 7.5%. Oligoarticular and polyarticular involvement were almost equally observed. In another study, gastrointestinal complaints including diarrhea, vomiting, and abdominal pain were seen in 61% of patients, and respiratory symptoms including rhinorrhea or cough were seen in 35%. Otitis media or tympanitis can also be seen in KD and may represent inflammation rather than infection. Urethritis or meatitis is common and may result in sterile pyuria. Hydrops of the gallbladder occurs in approximately 15% of patients in the acute phase and can be seen on abdominal ultrasonography. In those who do not fulfill diagnostic criteria, abdominal ultrasonography may also be useful in supporting the diagnosis of KD. Facial nerve palsies have been described. Transient sensorineural hearing loss can also occur. In countries where the BCG vaccine is administered, erythema and induration are commonly seen at the previous site of this vaccine. Psoriatic skin eruptions can occur in both the acute and convalescent phases of KD. Two of the potentially rare but devastating features of KD are peripheral gangrene and macrophage activation syndrome.
Laboratory Findings
Results of laboratory investigations in the acute phase of KD are nonspecific; however, in children who do not fulfill diagnostic criteria, these may also help support the diagnosis. Acute phase reactants are almost always increased and include erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP). In a child suspected of KD, if the acute phase reactants are not increased in the early febrile period, repeat testing usually shows this finding. Increased white blood cell count with a predominance of neutrophils is frequently seen. It is also not uncommon to see toxic granulations within the neutrophils. Anemia may be seen especially with prolonged inflammation. A characteristic feature of KD is thrombocytosis; however, this finding usually occurs in the second week of the illness (subacute phase, see later discussion). Hypoalbuminemia is common, as are increased serum transaminase levels. Sterile pyuria can also be seen.
Leukopenia and thrombocytopenia are both rare in KD. When seen, particularly in association with a lymphocytosis, a viral infection should be suspected. Leukopenia and thrombocytopenia can be present in KD in association with macrophage activation syndrome. Index of suspicion for this entity needs to be high given that this is a potentially fatal complication. Although thrombocytopenia is uncommon at presentation in KD, it is considered a risk factor for the development of CAA.
Cardiovascular biomarkers are being evaluated in KD. One such biomarker is N -terminal pro-B-type natriuretic peptide (NT-proBNP). This biomarker is associated with cardiomyocyte stress. It correlates with markers of inflammation, oxidative stress, and echocardiographic measurements, suggesting diastolic dysfunction. Some investigators suggest that subclinical myocardial inflammation occurs to some degree in all patients with KD and therefore, NT-proBNP may be an excellent biomarker to help support the diagnosis of KD. These investigators showed that NT-proBNP was increased in both patients with KD who fulfilled criteria and those who did not, compared with febrile controls. They suggested that NT-proBNP may be a valid adjunctive diagnostic test in patients with KD particularly in those with incomplete diagnostic criteria. Another study showed that NT-proBNP was increased in patients with KD who were developing CAA and therefore suggested that this could be a valuable tool for disease prognosis.
Clinical Phases
KD has 3 clinical phases: acute, subacute, and convalescent. The acute phase begins with the onset of fever and ends with its resolution. It usually lasts for a mean of 11 days (or shorter if therapy is instituted). The subacute phase begins with the resolution of fever and ends when all clinical features resolve. It typically begins 10 days into the illness and lasts for 2 weeks. Thrombocytosis and periungual desquamation of the fingers or toes occur at this stage. The convalescent phase begins at the end of the subacute phase and continues until the sedimentation rate and platelet count return to normal, which is usually 4 to 8 weeks after the onset of illness. This stage is when the nail abnormalities can be seen.
Cardiac Manifestations
Children suspected of having KD should be admitted to hospital and treated. During admission, all children with KD should undergo a two-dimensional echocardiogram and electrocardiogram. In the acute phase, the myocardium, pericardium, endocardium, valves, conduction system, and coronary arteries may all be involved. Myocarditis, when present, occurs during the acute phase. Myocarditis may present with tachycardia, gallop rhythm, electrocardiographic changes, or wall motion abnormalities on echocardiogram. Other cardiovascular abnormalities include pericarditis, valvulitis, and involvement of the conduction system.
KD shock syndrome is another cardiovascular manifestation that, when present, occurs in the acute phase. Patients present with hypotension and a shocklike state. They more often have left ventricular systolic dysfunction, mitral regurgitation, CAA, and resistance to treatment compared with hemodynamically normal patients with KD.
CAA can occur in the acute phase and have been detected as early as 3 days from onset of illness. However, CAA develop more commonly in the subacute phase. They develop in approximately 20% of untreated children and are the most common cause of morbidity and mortality in this disease. Two-dimensional echocardiography has been shown to be sensitive and specific in detecting aneurysms in the proximal portions of the right and left coronary arteries; isolated distal coronary artery aneurysms are uncommon. Children with uncomplicated KD should undergo an echocardiogram at diagnosis, at 2 weeks, and 6 to 8 weeks after the onset of disease. Repeat echocardiography should be considered optional beyond 8 weeks in those patients with previously normal studies. More frequent echocardiograms are required in those with CAA. These patients may eventually also require other investigations such as an angiogram ( Fig. 4 ). Although KD has a predilection for the coronary arteries, other arteries may be involved, including the axillary, iliac, and renal arteries.
Risk Factors for the Development of CAA
Many scoring systems have been developed to try to identify those who are at highest risk for the development of CAA, with the goal of using these to target treatment. The scoring systems are imperfect and it is recommended that all children diagnosed with KD receive treatment. Nonetheless, there are both clinical and laboratory findings that are associated with an increase risk of CAA in KD. The clinical features include male gender, extremes of age, prolonged fever, delay in diagnosis, and persistent fever after treatment. The laboratory features include low hemoglobin, increased white blood cell count, high absolute band count, very increased or persistently increased ESR or CRP, low platelet count, and low albumin.
Incomplete KD
KD is a diagnostic challenge. The signs and symptoms are nonspecific and are often found in viral illnesses. Adding to this challenge are the patients with KD who do not fulfill diagnostic criteria. These patients are considered as having incomplete KD, with at least 5 days of fever but only 2 or 3 of the other clinical criteria (see Box 1 ). The diagnostic criteria serve as a guideline to prevent overdiagnosis; however, they fail to identify children with incomplete forms of the disease. Incomplete KD needs to be considered in children with some of the clinical features and supportive laboratory findings seen in classic KD, especially in children at the extremes of age at which incomplete KD occurs more frequently. Incomplete KD should not be considered as mild KD because the risk of CAA in these patients is comparable with, if not higher than, classic KD.
Children less than 1 year of age, and those older than 9 years, are more likely to present with incomplete KD. Infants less than 6 months of age are more likely to have an incomplete presentation, late intravenous immunoglobulin (IVIG) treatment, CAA, and poor outcome. A high index of suspicion is required for infants younger than 6 months who present with fever for 5 or more days, despite not fulfilling criteria for KD. One study found that although children less than 6 months old and those greater than 9 years old were at an increased risk for CAA, there seemed to be different associated factors. Those less than 6 months of age had the coexistence of 2 associated risk factors: poor laboratory profile and delay in diagnosis. The poor laboratory profile included lower albumin levels, lower hemoglobin levels, and higher platelet counts. Children in the older age group had a delay in diagnosis and a lower rate of receiving appropriate treatment but a more benign laboratory profile. There may be an inherent risk for worse disease in the group less than 6 months of age but it is often coupled with a delay in diagnosis, leading to a worse outcome. This study also found that those more than 9 years of age were more likely to have resistant disease, with longer duration of fever despite treatment. Another study similarly showed that older children more often had a delay in diagnosis, took longer to defervesce, and had a more complicated course compared with younger children. Failing to consider the diagnosis of KD at the extremes of age puts children at risk for CAA.
In the 2004 American Heart Association (AHA) scientific statement, an algorithm was proposed to assist in evaluating children with suspected incomplete KD. This algorithm had not been tested prospectively before publication. The goal of this algorithm was to allow for earlier diagnosis and treatment of incomplete KD. As part of the algorithm, the authors suggested that an infant 6 months of age or younger, who is on the seventh day of fever or greater, without a focus, should have laboratory test looking for signs of inflammation. If systemic inflammation is found, with an increased CRP of 3 mg/dL or greater or ESR 40 mm/h or greater, then the authors suggest that the child undergo an echocardiogram even if no clinical criteria are present. If abnormal findings are seen on echocardiogram, treatment of KD is suggested. The definition of an abnormal echocardiogram includes a Z score of left anterior descending (LAD) or right coronary artery (RCA) 2.5 or greater; aneurysm as per the Japanese Ministry of Health Criteria; or at least 3 of the following: perivascular brightness, lack of tapering, decreased left ventricular function, mitral regurgitation, pericardial effusion, or Z score between 2 and 2.5 of LAD or RCA. In children more than 6 months of age with incomplete criteria, if their clinical characteristics are consistent with KD, an ESR and CRP are also advised. If no signs of inflammation are found, these authors suggest that the patient be reevaluated within 48 hours and that the algorithm be repeated at that time. If there is evidence of inflammation, it is suggested that supplemental laboratory testing be carried out to see if there are findings that would support this diagnosis. These findings include low albumin level, anemia, increased alanine aminotransferase level, increased platelet count after the seventh day, increased white blood cell count, and pyuria. If the patient fulfills at least 3 supplemental laboratory criteria, it is suggested that they be treated for KD. If the child has fewer than 3 supplemental laboratory criteria, an echocardiogram is advised. If the echocardiogram is abnormal, treatment is suggested. If negative but fever persists, these authors suggest repeating an echocardiogram and consulting a KD expert. The algorithm helps to use the knowledge of the clinical, laboratory, and echocardiographic features to assist in making the diagnosis of KD in children with incomplete disease.
The 2004 AHA scientific statement also helped to clarify the terms incomplete and atypical KD. Previously, atypical KD referred to a patient who did not fulfill criteria but who had echocardiographic findings of CAA consistent with KD. It has since been suggested that the term incomplete be used for all patients who do not fulfill criteria and that atypical be reserved for patients who fulfill criteria but who have a clinical feature that is not usually seen with KD, such as renal impairment (see Box 1 ).
Treatment
AHA Recommendations
The AHA recommends that all children suspected of having the diagnosis of KD be treated with a single dose of 2 g/kg of IVIG infused over 12 hours plus high-dose (or antiinflammatory dose) aspirin (ASA) at a dose of 80 to 100 mg/kg per day in 4 divided doses ( Box 3 ). Treatment should be started within the first 10 days of illness and ideally within the first 7 days. IVIG should also be given to children who present after the 10th day of illness if fever is still present, or aneurysms are seen with evidence of persistent inflammation. Some centers switch the child to low-dose (or antiplatelet dose) ASA (3–5 mg/kg/d) once the child is afebrile for 48 to 72 hours; others only do so at the 14th day of illness if the child has been afebrile for at least 48 to 72 hours (see Box 3 ). Low-dose ASA is discontinued if there is a normal echocardiogram 6 to 8 weeks after the onset of illness. If the coronary arteries at that time are not normal, then low-dose ASA is continued.
