Kawasaki disease (KD) is a multisystem acute febrile vasculitic syndrome of presumably infectious origin that affects predominantly infants and young children. The diagnosis is based on characteristic clinical features ( Box 77.1 ). Serious complications include coronary arteritis, coronary artery aneurysms and stenoses, coronary thrombosis leading to myocardial infarction, and, very rarely, rupture of a coronary aneurysm. KD has become the leading cause of acquired heart disease in children in most developed countries, including the United States and Japan. It has been reported in children of all racial groups and from all continents. Although the etiology of KD remains unknown and a specific diagnostic test is lacking, establishing a timely diagnosis is very important because administration of intravenous immunoglobulin (IVIG) and aspirin before the 10th day of illness generally has a dramatic effect on the clinical manifestations and markedly reduces the likelihood of development of coronary abnormalities (see the later sections on treatment).
Fever for at least 5 days a
a In the presence of classic features, experienced clinicians may be able to establish the diagnosis before the fifth day of illness.
plus four of the following features:
- •
Bilateral conjunctival injection
- •
Polymorphous exanthem
- •
Changes in the lips and oral cavity (erythema, cracking of lips; oropharyngeal erythema; strawberry tongue)
- •
Peripheral extremity changes (erythema and swelling of hands and feet; later periungual desquamation, Beau lines)
- •
Cervical lymphadenopathy (≥1.5 cm in diameter)
- •
Exclusion of other diseases with similar features
The finding of fever plus three criteria in the presence of coronary abnormalities qualifies.
Synonyms for KD include Kawasaki syndrome and mucocutaneous lymph node syndrome (MCLS, MLNS, or MCLNS). The disease also has been referred to as lymphomucocutaneous syndrome. An earlier term used in autopsy reports was infantile periarteritis nodosa (IPAN), which is pathologically indistinguishable from fatal KD. The International Classification of Diseases (ICD-9) designates the condition as both KD and mucocutaneous lymph node syndrome (acute) (febrile) (infantile) under rubric 446.1 and ICD-10 under M30.3. Before 1983, the National Library of Medicine listed KD publications under various subject headings, particularly “Lymphatic Diseases.” Since 1984, publications have been listed under “Mucocutaneous Lymph Node Syndrome,” a name that has been replaced by KD.
History
The illness now bearing his name was first recognized as a clinical entity in 1961 by Dr. Tomisaku Kawasaki, who subsequently became chairman of the Department of Pediatrics at Tokyo’s Japan Red Cross Medical Center. Between 1961 and 1967, Kawasaki identified 50 infants and young children who manifested a distinctive constellation of signs that included prolonged high fever, unilateral cervical lymphadenopathy, bilateral conjunctival injection, a polymorphous erythematous rash, changes of the mucosa of the lips and oral cavity, and edema and erythema of the extremities, with subsequent desquamation of the finger and toes. Although the syndrome was distinctive, its signs were nonspecific. Laboratory tests ruled out other disorders. A series of the first seven cases was presented by Kawasaki at the 61st Chiba General Meeting of the Japan Pediatric Society in 1962. Convinced that he was observing a distinct clinical syndrome, Kawasaki published a report of his experience with 50 cases of “febrile oculo-oro-cutaneo-acrodesquamatous syndrome with or without acute nonsuppurative cervical lymphadenitis” in 1967. Other Japanese physicians quickly recognized the syndrome after Kawasaki’s report, although considerable discussion ensued about whether it was a distinct entity or an illness related to Stevens-Johnson syndrome.
Cardiac involvement in this illness was suspected first in 1968, when Yamamoto and Kimura reported an infant with KD who had transient tachycardia with a gallop rhythm, cardiomegaly, and minor electrocardiographic (ECG) abnormalities. In 1970, Kawasaki obtained funding to establish the Research Committee of Mucocutaneous Lymph Node Syndrome, sponsored by the Japanese Ministry of Health and Welfare, which was organized with Dr. Fumio Kosaki as chair. In the first national survey of this committee, four autopsied and six nonautopsied cases of children who had died of coronary artery complications after apparent KD were identified. These children were predominantly younger than 2 years of age and had died suddenly within 30 days of onset of disease, with evidence of coronary aneurysms and acute thrombosis. The first biennial Japanese national epidemiologic survey was conducted by this committee in 1970 under the leadership of Dr. I. Shigematsu and was published in 1972. By this time, it had become well established that some patients who recovered apparently uneventfully from this acute illness were at risk for sudden cardiac death, with findings of acute myocardial infarction secondary to thrombosis within coronary arteries damaged by a severe vasculitic process.
In 1971, physicians at the University of Hawaii who were unaware of the Japanese experience began to recognize patients with an unusual Reiter syndrome–like illness. When information about KD was published in the English-language literature, the illness in Hawaii was recognized clearly as KD. Information exchanged between Japanese and U.S. investigators led to the 1974 publication of English-language articles by both groups and triggered worldwide recognition of cases. In the early 1970s, death from myocardial infarction was reported to occur in approximately 2% of cases of KD; more recent data reflect much lower mortality rates of less than 0.02%. Four deaths were recorded in Japan among 26,691 patients in 2011 to 2012 (0.015%). Four deaths occurred among 23,349 KD patients in Taiwan from 2000 to 2010 (0.017%). With the availability of echocardiography in the late 1970s, researchers determined that 20% to 25% of patients developed evidence of coronary artery abnormalities.
In the decades before Kawasaki recognized the clinical features of the illness, many individual reports of fatal coronary arteritis in children (usually labeled IPAN) were published in the non-Japanese pediatric and pathology literature. Clinical details of these cases generally are highly suggestive of KD, and the pathologic features of IPAN are indistinguishable from those of KD, as demonstrated conclusively by Landing and Larson in 1977. Almost 100 years before Kawasaki’s description was published, Samuel Gee of St. Bartholomew’s Hospital in London in 1871 reported the case of a 7-year-old boy who at death (“following scarlatinal dropsy”) had three coronary aneurysms, each filled with a fresh clot; histologic examination of that patient’s cardiac tissue is compatible with inactive KD with extensive coronary luminal myofibroblastic proliferation. Shibuya and colleagues retrospectively identified cases of illnesses compatible with KD that occurred in Japan up to two decades before Kawasaki’s description was published. Likely, patients with KD in previous decades were misdiagnosed as having measles, scarlet fever, rubella, or other once common conditions, and reductions in the numbers of cases of those illnesses helped to facilitate recognition of KD.
Kawasaki’s clinical description of the syndrome has remained the foundation of diagnosis and the basis of the clinical and epidemiologic case definitions in use today (see Box 77.1 ). In 2017, the American Heart Association Committee on Rheumatic Fever, Bacterial Endocarditis, and Kawasaki Disease published updated guidelines for the management of patients with incomplete (or atypical) KD (see later discussion).
Epidemiology
Sources of Epidemiologic Data
In the absence of a confirmatory diagnostic test, the epidemiologic case definitions of KD are relatively strict and exclude from surveillance data other exanthematous conditions that could dilute “true” cases and thus could obscure secular trends. However, the original epidemiologic case definitions were not intended for clinical application, a very important point since effective treatment became available. Thus, less strict application of clinical case criteria is appropriate for management of patients. Clinicians must be aware that children often present with clinical illnesses that do not completely fulfill the diagnostic criteria for KD but who are nonetheless at risk for developing coronary artery sequelae and therefore warrant therapy. These patients generally are considered to have incomplete or atypical KD. Incomplete presentations of KD are particularly common occurrences in young infants in whom clinical signs often are subtle or fleeting but who are at the highest risk for development of coronary artery abnormalities. In the United States, the Centers for Disease Control and Prevention (CDC) case definition usually is used for epidemiologic purposes and the American Heart Association published an algorithm to aid in the diagnosis of incomplete (atypical) KD. The current Japanese diagnostic guidelines were revised in 2014 and also reflect the importance of incomplete cases.
Incidence Rates
The incidence of KD varies throughout the world and reflects primarily the racial composition of various countries. Rates in Japan have climbed steadily, with an annual rate of 243.1 per 100,000 children younger than 5 years of age in 2011 and 264.8 in 2012 (22nd national survey). The highest reported rate in Korea was 134 per 100,000 children younger than 5 years old in 2009–11. Taiwan has the third highest reported rate, 82.8 per 100,000 children less than 5 years old. In countries with predominantly non-Asian populations, the rate is 15 to 20 per 100,000 children younger than 5 years.
Gender
In virtually all population-based studies in many countries, the ratio of male to female patients with KD approximates 1.5 : 1. In addition, serious and fatal complications also are significantly more common findings among male patients with KD compared with female patients. Examination of fatal Japanese cases indicated almost three times as many Kawasaki-related fatalities among males compared with females, with a higher ratio in infancy. The basis for the preponderance of KD in males and for the even greater predominance of serious coronary artery disease in males with KD remains unclear. Of interest is that a male predominance is observed in many infectious diseases.
Race or Ethnic Background
The first cases of KD were recognized in Japanese children and in Hawaiian children of predominantly Japanese ethnicity, and subsequent data consistently support higher rates in those of Asian background. Annual incidence rates in Japan have climbed steadily to 264.8 cases per 100,000 children younger than 5 years of age in 2012, from 102.6 in 1995 and 184.6 in 2005 to 2006, and they are the highest in the world, approached only by rates for Japanese-American children in Hawaii (210.5 per 100,000 of children younger than 5 years). More than 13,917 cases of KD were reported in Japan in 2012, with local clusters rather than nationwide epidemics, as were seen in 1979, 1982, and 1985–86. In epidemic years in Japan, the annual age-specific incidence rates reached or exceeded 200 per 100,000 children younger than 5 years. Incidence rates in white children in many communities are much lower, most often approximating 10 to 20 per 100,000 children younger than 5 years. Surveys in countries with almost exclusively white populations often yield rates of 5 to 10 cases per 100,000 children younger than 5 years. In Washington state, ethnic group–specific incidence rates per 100,000 children younger than 5 years were estimated to be 33.3 for Asian Americans, 23.4 for blacks, and 12.7 for whites. In Hawaii, with its complex racial-ethnic makeup, the overall annual incidence is about 45 per 100,000 children younger than 5 years. The yearly incidence for Japanese children in Hawaii approaches 200 cases per 100,000 children, and for whites it is 35 per 100,000, with intermediate rates for those of native Hawaiian and Chinese, Filipino, and other Asian ancestry. Prospective Kawasaki disease surveillance in Germany in 2011–12 corrected for underreporting yielded an incidence of 7.2 per 100,000 children younger than 5 years. Extrapolation of data from surveys of U.S. hospitals with large children’s services led to estimates of approximately 2500 cases per year in the United States from 1984 to 1993, with more recent estimates as high as 5000 cases annually. A study of KD in the United States identified approximately 4200 hospitalizations in the year 2000, with highest rates among Asian and Pacific Islanders, lowest rates in whites, and intermediate rates in blacks and Hispanics.
Age
KD occurs almost exclusively in children. In the United States and Japan, adult cases are quite rare, although some reports of adults diagnosed by accepted diagnostic criteria have been published. Many of these adult patients have been infected with human immunodeficiency virus (HIV). Most of the early reported adult cases, however, probably actually represented toxic shock syndrome or drug hypersensitivity reactions. Perhaps the best-documented adult case (confirmed by Dr. Kawasaki) was that of a 31-year-old Japanese man who later developed bilateral coronary artery aneurysms. Because the signs and symptoms are nonspecific, adults suspected to have KD should be evaluated carefully for infectious, toxic, and other possible causes of illness.
The distribution of KD by age in childhood is highly characteristic. The disease occurs most frequently in young children: 50% are younger than 2 years, 80% are younger than 5 years, and cases seldom occur in those older than 12 years. In 2009–10, 0.7% of Japanese patients with KD were 10 years old or older. Infants in the first 3 to 5 months of life have a relatively low incidence of KD, but the incidence rises rapidly from that point. In the United States, the peak age is about 10 months, whereas more recent data from Japan identified 25% as younger than 1 year, 88% as younger than 5 years, and the peak age incidence as 9 to 11 months of age. In Hawaii, 29% of cases are in infants younger than 1 year, and 85% of patients are younger than 5 years. The age-incidence curve may be helpful in elucidating risk factors for developing KD. Such a pattern is compatible with highly transmissible infectious agents, particularly respiratory agents, and suggests possible transplacental immunity. The features of 28 patients with KD who were aged 8 years and older at the time of diagnosis at Children’s Memorial Hospital in Chicago have been reported. Delays in establishing the diagnosis and in providing treatment were common occurrences and were at least partially related to the prominence of arthritic and gastrointestinal symptoms in this population. We recently cared for a 30-year-old man with classic KD.
Earlier Japanese mortality data suggested that fatality rates are approximately three times higher in children younger than 1 year at the time of onset of disease, compared with older children, and that fatalities occur predominantly in the first several months after onset of KD. Male patients account for a disproportionate number of deaths in infants and older children. Long-term mortality rates in a large Japanese KD cohort study were increased over background rates, particularly during and shortly after the acute stage of illness.
A Taiwanese series of 25 infants who developed KD at 3 months or younger included 19 males and 6 females. Only 6 infants met KD diagnostic criteria, with the other 19 classified as incomplete. Coronary artery involvement was noted in 80%, although only one infant had persistent abnormalities at 1-year follow-up.
Recurrent Kawasaki Disease
A recurrence of KD is defined generally as a new episode of illness meeting clinical criteria for KD that begins at least 3 months after the initial episode and after inflammatory markers such as the erythrocyte sedimentation rate (ESR) or serum C-reactive protein (CRP) level have normalized. The frequency of recurrences after first cases of KD in Japan was estimated to be approximately 1.9% during a 3-year follow-up period, with approximately 0.07% of patients experiencing a third episode. This corresponds to a rate of 5.21 per 1000 person-years for one or more recurrences. With longer follow-up, recurrence rates in Japan may approach 3%. Data from the United States suggest a recurrence rate of 1% to 2%. Recurrences occur most frequently within the first 2 years after the initial episode, especially in male patients and in children who have their initial episode before reaching their second birthday. The frequency of recurrences in Chicago appears to be approximately 1%, occurring more often in Asian children, and recurrences were documented in about 2% of Hawaiian cases from 1996 to 2000, a finding likely reflecting a racial difference. In the CDC’s passive KD surveillance data, 97 of 5557 (1.7%) patients from 1984 to 2008 had recurrent KD, including 3.5% among Asian/Pacific Islander patients. The true recurrence rate will be determined only when a specific diagnostic test becomes available for KD and thus minimizes recognition bias.
Family Cases
Simultaneous or sequential cases of KD in siblings, twins, or other family contacts also have been reported, particularly during outbreaks in Japan. Japanese epidemiologists have documented secondary sibling cases in approximately 1% of cases, a rate that is approximately 10 times greater than that in the general child population. However, such figures are difficult to interpret because they may be influenced by recognition and reporting biases. In the 2011–12 Japanese survey data, 1.5% of KD cases had one or more infected siblings, and 0.89% had at least one parent with history of KD. Sibling cases are reported to occur more frequently in twins than in nontwins. Only three sibling pairs were recognized in more than 2100 Chicago patients, and none was reported in 400 children in Los Angeles. A report from two US medical centers noted 18 families with multiple affected KD patients, including nine families with two affected siblings and nine with KD in two generations or in multiple affected members.
Epidemics and Outbreaks
Japanese investigators noted large nationwide outbreaks of KD in 1979, 1982, and 1985 to 1986, with wavelike spread occurring from one prefecture to the next, suggesting an infectious origin. The 1982 Japanese epidemic started simultaneously in four areas and spread outward from each region, similar to the way that epidemic influenza spreads in Europe and America. In the 1985–86 Japanese epidemic, investigators identified epidemic “waves” that spread outward from an initial focus in the Tokyo metropolitan area and extended simultaneously northward and southward to involve most of the country within approximately 4 months. A similar but less distinctive pattern of interprefectural progression in waves was noted in the 1982 Japanese epidemic. Within the northern Tohoku District, for example, KD spread from prefecture to prefecture over a period of approximately 7 months. Korean epidemics were detected 7 and 15 months after the Japanese epidemics of 1979 and 1985–86, respectively. More localized outbreaks also have been observed. In the United States and elsewhere, community-wide outbreaks were documented beginning in 1977. Investigation of outbreaks provides opportunities to study potential risk and etiologic factors. Clustering of cases within families, schools, or neighborhoods is quite unusual, even during large-scale epidemics. Japanese investigators have associated epidemics with a significantly increased likelihood of second cases occurring in families, fatalities, and recurrent cases, but the implications of these findings are uncertain. It is striking that, since 1986, no further nationwide Japanese outbreaks have been identified, suggesting that the epidemiology of KD may have changed. However, it is important to recognize that the current endemic rate of KD in Japan exceeds that of the largest previous endemic.
Geography
KD has been diagnosed throughout the United States and Japan and in virtually all developed and many developing countries on all continents, including temperate and tropical zones. No striking rural–urban differences have been noted. Elevation, longitude, and latitude have not been implicated. Travel histories of patients are unremarkable, although an anecdotal report of a 7-month-old infant in Australia documented KD onset 17 days after leaving Japan during that country’s 1982 epidemic, and we diagnosed KD in infant in Chicago about 2 weeks after moving from Japan.
Seasonality
In Japan, KD occurs year round but is most prevalent in the winter, with peaks usually occurring in December or January, with a lower peak in June, and the lowest number of cases in October. In Korea, peaks were observed in summer and winter ; in Taiwan, peaks occur in summer, with a nadir from November to January ; in Beijing and Shanghai, peaks occur in spring and summer , and in spring in Sichuan and Hong Kong. In the United States and other temperate areas, the number of cases peak in the winter and early spring and is lowest in late summer; nonetheless, cases occur throughout the year. No clear seasonality has been seen in Hawaii. Winter predominance has been observed in at least some Southern Hemisphere countries. That sporadic cases are recognized year round is somewhat different from the pattern usually observed with highly transmissible respiratory viral diseases, incidences of which peak sharply in the winter and spring (e.g., measles, rubella, influenza), and virtually disappear in summer.
Communicability
Although little direct evidence exists that KD is transmissible from person to person, considerable circumstantial evidence supports an infectious origin. Secondary or co-primary cases in families occur but are not common. Several outbreak investigations found a higher rate of antecedent respiratory tract illness in patients with KD compared with matched controls. This finding is of particular interest in view of the finding that immunoglobulin A (IgA) plasma cells infiltrate the proximal respiratory tract and that viral-like inclusion bodies are observed in bronchial epithelial cells in fatal cases of KD, and it strongly supports a respiratory mode of spread of an inciting agent (see later discussion). Japanese family data suggest that sibling cases cluster either on the same day as the index case or 7 days later. Because these results are based on questionnaire data, however, the possibility of ascertainment bias is great; thus determining the degree to which Japanese familial cases represent co-primary or secondary cases is difficult.
Other Risk Factors
In addition to demographic risk factors for KD, specific exposures that could be related to an etiologic agent have been linked to the disorder by older epidemiologic investigations. History of more frequent recent antecedent respiratory illnesses in cases compared with controls was documented in the early 1980s. Because many viral agents are prevalent in the winter and spring when KD is most prevalent, some “background” isolation of various viruses from patients and from matched controls is expected.
Other past reported KD associations include exposure to recent carpet cleaning or shampooing, exposure to house dust mites, and residence near bodies of water; very few recent reports support these associations. The association of KD with exposure to recently cleaned or shampooed carpets was observed in some reports but not in others. The significance of a possible association with shampoo is unclear because it is absent entirely in many well-studied outbreaks and because carpets are relatively uncommon in Japan. In a small number of outbreaks, investigators from the CDC found that patients with KD lived closer to bodies of water compared with control patients. However, other studies, including one in Washington state, did not find living in proximity to water to be a risk factor.
Etiology
The origin of KD remains unknown. However, clinical and epidemiologic features strongly suggest that the disease has an infectious cause. A self-limited, usually nonrecurring illness manifested by fever, rash, enanthem, conjunctival injection, swollen and red hands and feet, and cervical adenitis fits well with an infectious cause. The epidemiologic features noted earlier, including the striking age distribution, winter-spring seasonality in most areas, occurrence of community outbreaks with wavelike geographic spread, and apparent epidemic cycles, resemble those of a transmissible disease of childhood. The laboratory features, including leukocytosis with a “left shift,” elevated acute-phase reactants, mildly elevated liver transaminases, and pyuria also suggest infection. The medium-size muscular arteritis of KD may result from infection of arterial walls by the causative agent(s) with a subsequent immune response or an immune attack on presently unknown host molecules as a consequence of infection with the agent(s).
A very attractive hypothesis is that KD is caused by a ubiquitous infectious agent that produces clinically apparent disease only in selected, genetically predisposed individuals, with particular predilection for Asians. Its rarity in the first few months of life and in older children and adults suggests an agent to which virtually all adults are immune and from which very young infants are protected to some degree by passive maternal antibody. Consistent with this hypothesis is the paucity of evidence of person-to-person spread of KD, because most infections are likely asymptomatic, and only very few of those infected likely develop clinical features of KD. This would be analogous to acute poliomyelitis, the relatively uncommon complication of poliovirus infection that occurs in only approximately 1 in 200 of those infected. However, efforts to identify an infectious agent of KD using conventional bacterial and viral culture and serologic methods, as well as inoculation of primates, mice, and guinea pigs, have failed to yield an infectious cause, and a long list of known infectious agents proposed as the cause has been disproved. This includes Leptospira spp., retrovirus, TSST-1–producing Staphylococcus aureus, and coronavirus NL-63. An oligoclonal, antigen-driven immune response has been well-documented ; whether a superantigen response also occurs in some children with KD is unclear and will likely require identification of the causative agent to be determined.
The immune transcriptional profile of coronary arteritis in KD has features of an antiviral immune response such as activated cytotoxic T lymphocyte and type I interferon-induced gene upregulation. Ultrastructural, immunofluorescence, and RNA evidence support the hypothesis of a presently unidentified “new” virus as the cause that enters through the respiratory tract and is carried by macrophages to the coronary arteries. Another hypothesis is that the agent of KD, possibly a fungal toxin, may be carried via the wind as an aerosol.
While some investigators suggest that KD can be triggered by any of a large number of diverse infectious agents in a genetically predisposed host, recurrence of KD is quite uncommon, and KD children appear to experience subsequent infections without developing KD again. This theory also fails to explain epidemics and the wavelike geographic spread of illness during outbreaks.
Identification of the etiologic agent(s) of KD is an extremely important research goal because diagnostic test development, development of improved and specific therapies, and prevention are critical to improved patient care and outcomes but will likely only be achieved after the etiology is determined.
Genetic Susceptibility
The higher rates of KD in those of Japanese and, to a lesser degree, Korean and Taiwanese backgrounds indicate a genetic basis rather than environmental factors, as supported by increased rates among third- and fourth-generation immigrants from Japan to Hawaii. The increased incidence of KD among siblings and parents of patients also supports a genetic predisposition. The genetic basis is complex, and no single human leukocyte antigen (HLA) or MHC class II gene is common to most patients with KD. Immunoglobulin allotypic markers were studied as possible genetic markers for KD, and differences between patients and race-matched controls were found in Japanese and whites, supporting a complex genetic basis for susceptibility to KD. Transmission disequilibrium studies involving children with KD and their parents demonstrated evidence of several functional single nucleotide polymorphisms in the calcineurin-NFAT pathway, the transforming growth factor-β (TGF-β) signaling pathway, and genes encoding Fcγ receptors. These studies show that single nucleotide polymorphisms (SNPs) within the inositol triphosphate kinase-3 (ITPKC) gene on chromosome 19 and in CASP3 on chromosome 4 confer increased risk in Japanese, Taiwanese, and white children.
Genome-wide association studies (GWAS) to define KD susceptibility risk loci have accelerated identification of additional loci on chromosome 8 and other chromosomes. Davila and colleagues, also in a GWAS, identified a polymorphism in FCGR2A on chromosome 1 related to immune activation as a KD risk marker. Other analyses have identified SNPs in ITPKC and CASP3 that are associated with increased risk of IVIG unresponsive and coronary aneurysms. Several additional loci have been identified in the Taiwanese population. The picture is emerging that there are more than 6 to 10 genetic susceptibility markers for KD, each of which is associated with some increased risk for KD compared with controls. Onouchi et al. recently reported an SNP in the ORA/1gene in the Ca 2+ /NFAT pathway that is associated with KD in Japanese and that is present with 20 times higher frequency in Japanese compared to European populations. This is an area of intense research activity.
Pathology and Pathogenesis
Relationship With Infantile Periarteritis Nodosa
When the first Japanese survey of KD in 1970 yielded reports of 10 deaths, the pathologic diagnosis was thought to be IPAN. Japanese and Western investigators quickly recognized the pathologic similarities between fatal KD and IPAN, long known to have an unusual predilection for the coronary arteries. In the most definitive study, Landing and Larson showed clearly that IPAN and fatal KD were indistinguishable pathologically. Published case reports of autopsies of children with IPAN frequently contain clinical histories that include many or all features of acute KD and the development of severe vasculitis affecting primarily the coronary arteries. The failure of some reported patients with IPAN to meet KD case criteria applied retrospectively is likely related to documentation bias and to the young age of many of the patients. Young infants with KD are less likely to have a classic presentation than are older children (see later discussion). When pathologic and clinical criteria are combined, the two diseases appear indistinguishable. Adult-type periarteritis nodosa was described first in 1866 by Kussmaul and Maier, and it differs from IPAN primarily by the presence of hypertension and involvement of small and medium muscular arteries, especially in the lung, kidney, and intestines. An early recorded case of possible IPAN dates from 1899. In 1959, Munro-Faure very clearly delineated a syndrome of infantile necrotizing arteritis with coronary artery involvement, fever, rash, conjunctival and pharyngeal infection, and cervical adenitis, distinct from classic periarteritis nodosa. Roberts and Fetterman and others expanded on these observations to define a distinct clinicopathologic syndrome of IPAN shortly before Kawasaki recognized the clinical syndrome.
Coronary artery aneurysms were described as early as the early 20th century, with a male to female ratio of roughly 3 : 1, including a male preponderance in childhood cases. Childhood death from multiple coronary artery aneurysms was known to occur at least as early as 1871, as reported by Samuel Gee. The cardiac specimen from Gee’s case, formalin fixed for more than 120 years in the pathology museum at St. Bartholomew’s Hospital, London, recently was sectioned and examined histologically. The coronary arteries showed the characteristic histologic findings of inactive KD and IPAN. Whether these early cases with childhood coronary artery aneurysms truly represented early examples of KD is not certain; details of the clinical histories usually were scant. What is clear, however, is that most of these cases greatly resemble (and very likely do, in fact, represent) KD. From the features of many such early case reports, a reasonably accurate picture of KD as recognized today emerges. It is highly likely that in the decades before the introduction of the measles vaccine in the 1960s, cases of KD were misdiagnosed as measles, scarlet fever, drug hypersensitivity, or other common conditions. In fact, an autopsy study of pediatric vasculitis cases in Japan indicates that many pediatric vasculitis fatalities prior to 1977 had clinical and pathologic features that in retrospect were consistent with KD. This study also demonstrated a marked decrease in all pediatric vasculitis fatalities in Japan following the advent of IVIG therapy in the late 1980s, suggesting that the majority of pediatric vasculitis fatalities in Japan prior to this time were actually the result of Kawasaki disease
Pathologic Features of Kawasaki Disease
Cardiac death in KD generally occurs in the subacute or convalescent stages of illness but also can occur earlier. KD results in medium-sized muscular arteritis, with a marked predilection for the coronary arteries. Small arterioles, larger arteries, capillaries, and veins also are affected to a lesser extent. In more than 80% of fatal cases in the acute stage, the immediate cause of death is acute thrombosis of inflamed coronary arteries, with resultant myocardial infarction. Less commonly, death can occur from acute coronary rupture, usually within a few weeks of the onset of illness, and rarely early deaths can result from pancarditis with inflammation of the atrioventricular conduction system and fatal arrhythmia or intractable congestive heart failure. Deaths that occur months to years after the acute episode of KD often are secondary to coronary stenosis, from luminal myofibroblastic proliferation (see below) and/or thromboses, resulting in myocardial ischemia. In some patients with coronary artery abnormalities (virtually always in those who also have quite severe coronary disease), aneurysms of other major medium-sized arteries, including the brachial, renal, and iliac arteries, also may be present. Although phlebitis may be found, vascular inflammation more typically and more severely affects medium-sized muscular arteries in their extraparenchymal portions. In the acute stage of KD, systemic inflammatory changes also are evident in many other organs, including myocardium, pericardium, cardiac valves, meninges, lung, lymph nodes, pancreas, spleen, joints, and liver.
An extensive study of arterial tissues from 32 patients who died and 8 who received heart transplants shed new light on the pathologic features of severely affected KD patients. Three characteristic vasculopathic processes were identified by light and transmission electron microscopy: (1) acute self-limited necrotizing arteritis (NA), (2) subacute/chronic (SA/C) vasculitis, and (3) luminal myofibroblastic proliferation (LMP). NA is a neutrophilic process that starts at the endothelial lining of vessels only in the first 2 weeks of illness, progressively destroying the vessel wall and causing development of saccular aneurysms, which can thrombose or rupture. In its most severe form, NA can result in necrosis of arteries to such an extent that only a thin layer of adventitia remains. These “giant aneurysms” can rupture but more commonly acquire multiple layers of thrombi which organize and can become recanalized and calcify. SA/C vasculitis, consisting of lymphocytes, eosinophils, and plasma cells, with fewer macrophages, begins within the first 2 weeks of illness in the adventitia/perivascular tissue and damages the vessel wall during its progression toward the lumen. SA/C can lead to fusiform and saccular aneurysms that can thrombose, and it is associated with transition of medial and adventitial smooth muscle cells into smooth muscle cell–derived myofibroblasts which, with their matrix products, result in stenosing luminal lesions (SA/C-LMP). LMP can also occur in some areas without apparent local SA/C vasculitis. NA is the only self-limiting of these three processes, is responsible for the earliest morbidity and mortality, and is consistent with acute viral infection. SA/C vasculitis can begin early but can persist for months or even years, whereas LMP can cause progressive arterial stenosis. Stenosis of the coronary arteries months to years after onset can result from LMP and/or thrombi. The mechanism of persistent SA/C arteritis and LMP in a subset of KD children is unknown. Ongoing expression of vascular endothelial growth factor (VEGF) within smooth muscle cells, TGF-β in myofibroblasts, and matrix metalloproteinases in inflammatory and other cells have been proposed to be involved.
Significant atrioventricular conduction system lesions have been identified in 5 of 10 autopsy specimens, with a strong correlation found between ECG findings, especially PQ-segment prolongation, and acute inflammation of the atrioventricular conduction system. Severe acute changes were most pronounced at 21 to 31 days after onset. In a unique study of right ventricular endomyocardial biopsy specimens in 201 patients with uncomplicated acute KD, some degree of myocarditis and cellular infiltration ranging from very mild (subclinical) to quite severe myocardial inflammation was noted in all of the patients studied.
Neutrophils infiltrate the arterial wall at sites of NA; NA appears limited to the first 2 weeks after fever onset. In SA/C arteritis, which begins in the first 2 weeks after fever onset but can persist for months to years in a subset of KD children, lymphocytes (particularly CD8 + lymphocytes ), plasma cells (particularly IgA plasma cells ), and eosinophils predominate, while fewer macrophages are also present. The transcriptome of KD SA/C arteritis demonstrates upregulation of genes associated with cytotoxic T-lymphocyte activation, antigen presentation, immunoglobulin production, and type I interferon response, features suggesting an antiviral immune response. Both CD8 + T lymphocytes and IgA plasma cells have been demonstrated to be oligoclonal (i.e., they appear to be responding in an antigen-driven process).
Although arteritis is the most important pathologic feature of KD and results in the only known long-term sequelae of KD, the illness is associated with multisystem inflammation, and multiple organs and tissues are involved in the acute inflammatory process, including the digestive system (e.g., enteritis, hepatitis, cholangitis, pancreatitis), respiratory system (e.g., bronchitis, pulmonary nodules, segmental interstitial pneumonia), urinary system (focal interstitial nephritis, cystitis, prostatitis), nervous system (e.g., choriomeningitis), and hematopoietic system (e.g., lymphadenitis, splenitis), as well as the cardiovascular system (e.g., myocarditis, pericarditis). IgA plasma cells have been observed in peribronchial, pancreatic, and renal tissues and macrophages in pancreatic acini and islets. Lymph node findings are nondiagnostic and can show severe lymphadenitis with necrosis.
At present, it is not known what factors predispose children with KD to develop coronary artery stenosis from LMP. This complication can occur in children with persisting aneurysms, but can also result in critical stenosis in patients with aneurysms that appear to have “resolved” by echocardiography or angiography. In this latter circumstance, arterial luminal diameter initially returns to a more normal range as LMP fills the lumen, but the lumen subsequently narrows with progressive LMP, and coronary artery blood flow is impaired. Close follow-up of all children with KD with persisting or “resolved” aneurysms is strongly encouraged, according to the recommendations of the American Heart Association Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease.
Clinical Manifestations
Clinical Phases of Illness
The clinical course of KD can be divided into acute, subacute, and late or convalescent phases. The acute febrile phase begins with fever, rash, conjunctival injection, “strawberry” tongue, red swollen lips, edema and erythema of the hands and feet, cervical lymphadenitis, and sometimes aseptic meningitis and mild hepatic dysfunction. Young children often are quite irritable. Evidence of myocarditis, rarely including congestive heart failure or arrhythmias, may develop during this time. Pericardial effusion, mitral regurgitation, and/or depressed myocardial function may be detected by echocardiogram. Without IVIG and aspirin treatment, the acute phase generally lasts 8 to 30 days (mean, 11 days).
After defervescence, the physical findings rapidly disappear, but, during this subacute phase, the child may remain irritable and anorectic, with decreased activity. Some conjunctival injection may persist. Arthritis or arthralgia, mainly of larger joints, may develop in the subacute phase. Desquamation of fingers and toes, typically beginning in the periungual region, and thrombocytosis are very common manifestations during this period. The subacute phase persists until the child has returned to his or her normal state of health at 6 weeks after the onset of fever, with normalization of inflammatory markers. The time of greatest risk for sudden death occurring from acute coronary artery thrombosis in patients with coronary lesions is during the subacute phase and the early convalescent phase. The convalescent phase begins when all clinical signs and symptoms have disappeared and inflammatory markers are normal, usually at 6 to 8 weeks after onset.
In its complete form, KD is a distinctive clinical entity with a fairly predictable course. The principal clinical diagnostic criteria are presented in Box 77.1 . A diagnosis of typical or classic KD, according to these accepted clinical criteria, is made in patients with fever and at least four of the five classic clinical criteria and with exclusion of other illnesses that mimic KD. Each of the five clinical features is present in 80% to 90% of patients with typical cases, except cervical lymphadenopathy, which is present in approximately 50% of patients. However, children who do not fulfill the classic criteria for diagnosis of KD, in fact, may have the illness and are at risk for developing complications, particularly coronary artery disease (see later discussion section on incomplete or atypical KD). KD should be considered in the differential diagnosis of infants and children with fever for at least 5 days associated with two or more of the five classic features: (1) generalized polymorphous erythematous rash, (2) conjunctival injection, (3) characteristic changes of the lips and mouth, (4) bilateral redness and swelling of the hands and feet, and (5) unilateral nonfluctuant cervical lymph node enlargement greater than 1.5 cm. All features are not necessarily present at the same time. The most commonly encountered diseases to be excluded are febrile exanthems, presumably viral, including measles; acute streptococcal and staphylococcal infections; and drug hypersensitivity reactions. Japanese diagnostic guidelines for KD are similar to the U.S. guidelines with rare exception: they consider fever and the other five major features to be six equal criteria and require at least five of those six criteria for diagnosis. Thus a rare child could have no fever but all five other criteria and be considered a typical case by the Japanese criteria. The Japanese criteria also accept a course of fever lasting less than 5 days if it is shortened by early IVIG treatment.
Several scoring systems using clinical and laboratory features have been developed, primarily in Japan. The initial goal of these systems was to identify those patients with KD at highest risk for development of coronary abnormalities and who, therefore, would benefit most from receiving IVIG therapy, as well as those at low risk for coronary changes who could be spared IVIG treatment. This practice has decreased in Japan because no such scoring system appears sufficiently sensitive and specific to enable selective therapy (i.e., to allow nontreatment of patients predicted to be at low risk for development of coronary abnormalities). More recent scoring systems have focused on identifying patients at high risk for IVIG treatment failure and coronary abnormalities who may benefit from more aggressive primary therapy. Patients who are at greatest risk are as follows: those younger than 1 year; male patients; those with prolonged or recurrent fever; and those with anemia, hypoalbuminemia, hyponatremia, and thrombocytopenia. The Kobayashi and other scoring systems are not effective in predicting treatment failure in multiracial populations as seen in the United States. All patients diagnosed with KD within the first 10 days after onset of disease and those diagnosed later who are still manifesting significant inflammation should be treated with IVIG and aspirin.
In KD, fever typically is high spiking and remittent, with peak temperatures generally exceeding 39°C (102.2°F) and in many cases exceeding 40°C (104°F). Unless treatment is initiated with IVIG and aspirin, fever persists for a mean of 11 days, but it may continue for 3 to 4 weeks and rarely longer. In patients whose disease is treated with 80 to 100 mg/kg per day of aspirin and a single 2 g/kg dose of IVIG, fever generally resolves within 1 to 2 days.
Bilateral painless vascular injection of the bulbar conjunctivae, clearly more severe than injection of the palpebral conjunctivae and often sparing the limbic region around the cornea, generally is seen in the first week of illness, usually beginning shortly after onset of fever. It generally is not associated with exudate, conjunctival edema, or corneal ulceration, thus distinguishing the eye findings of KD from purulent conjunctivitis and from Stevens-Johnson syndrome. Mild acute iridocyclitis or anterior uveitis, which may be noted by slit lamp examination, resolves rapidly and only rarely is associated with photophobia or eye pain. Less common ocular findings include superficial punctate keratitis, vitreous opacities, vitreous and chorioretinal inflammation, lateral rectus palsy, and papilledema and other optic disk changes.
Changes of the mouth and lips consist of the following: (1) erythema, dryness, fissuring, peeling, cracking, and bleeding of the lips; (2) a strawberry tongue indistinguishable from that associated with streptococcal scarlet fever, with erythema and prominent papillae; and (3) diffuse erythema of the oropharyngeal mucosae. Oral ulcerations, pharyngeal exudates, and Koplik spots are rarely if ever found in KD and when present help to exclude the diagnosis.
Changes in the extremities are among the most distinctive features of KD. The hands and feet become indurated and swollen with stretched, shiny skin, sometimes with painful induration. The palms and soles become erythematous, often with an abrupt change to normal skin at the wrist and ankle. Infants and young children frequently refuse to hold objects or to bear weight. In the subacute phase, a distinctive pattern of periungual desquamation of fingers and toes may occur from 2 weeks to 2 months after onset of KD in 50% to 70% of affected patients. Beau lines, which are transverse grooves across the nails, may appear at the nail base 1 to 2 months after a case of acute KD and grow out over several months. Peripheral gangrene is a rare complication (see later discussion).
The erythematous rash associated with KD may take many forms. Most common is a nonspecific, diffuse, maculopapular, primarily truncal erythematous rash. Occasionally diffuse scarlatiniform erythroderma, urticaria, or an erythema multiforme–like rash with target lesions develops. Vesicles and bullae are not seen, although very fine pustules occur rarely. Perineal erythema and then desquamation are quite common manifestations in diapered as well as toilet-trained children in the acute stage of illness. Although rashes in KD tend to be most prominent on the trunk, with perineal accentuation they frequently also involve the face and extremities.
Cervical lymphadenopathy is the least common of the five principal diagnostic criteria, but it sometimes is the dominant clinical feature along with fever. It usually is unilateral and confined to the anterior cervical triangle. To fulfill diagnostic criteria, the enlarged node or mass of nodes exceeds 1.5 cm, is not fluctuant, usually is not associated with erythema of the overlying skin, and is not tender or only moderately tender. Lymphadenopathy generally is benign and transient. Clinicians should be aware that children with suspected acute bacterial cervical adenitis that is unresponsive to antibiotic therapy may have KD. Because other features of KD often are present but overlooked, KD should be considered in such febrile children without an alternate diagnosis. The ultrasound appearance of lymph nodes in KD has been described to be similar to that of acute Epstein-Barr virus infection and distinct from bacterial adenitis. The absence of changes suggesting suppuration on imaging (e.g., ultrasonography or computed tomography [CT]) of such lesions should strengthen the suspicion of possible KD in this setting. Impressive cervical adenopathy in patients with KD generally resolves remarkably promptly after administration of appropriate therapy.
The associated features of KD reflect its multisystemic nature ( Box 77.2 ). Sterile pyuria as a manifestation of urethritis, occasionally with meatitis, is found in approximately half of patients. Arthritis appearing during the first week of illness can be polyarticular or oligoarticular, including the small interphalangeal joints as well as large weight-bearing joints, with a reported prevalence of approximately 7.5%. Arthrocentesis during this early phase yields purulent-appearing fluid, with a mean white blood cell count of 125,000 to 300,000/mm 3 , normal glucose levels, and negative Gram stain and bacterial cultures. Arthritis developing after the 10th day of illness has a predilection for large weight-bearing joints, especially the knees and ankles, with a somewhat lower synovial fluid white blood cell count. Gastrointestinal complaints occur in approximately one-third of patients, especially in older patients; these complaints may be severe, even leading to laparotomy, and they include nausea, abdominal pain, and some diarrhea. These findings may be related to a hydropic gallbladder, pancreatitis, or appendicular vasculitis. Obstructive jaundice and acute hydrops of the gallbladder are not uncommon findings, whereas mild to moderate elevations of serum aminotransferases occur in almost half of patients. Central nervous system involvement including aseptic meningitis occurs in almost half of patients. Transient unilateral lower motor neuron facial nerve palsy occurs rarely, as does sensorineural hearing loss. Characteristic marked irritability is very common, especially in young infants. Reactivation of inflammation at the site of a previous bacillus Calmette-Guérin (BCG) vaccination coincident with acute KD is a common finding in Japan, where BCG is used widely, and Terp and associates observed a child in whom both a BCG site and a purified protein derivative (PPD) test site reactivated with acute KD. Some patients experience a flare of atopic dermatitis or psoriasis during or after experiencing acute KD. Various pulmonary manifestations of KD, including isolated pulmonary nodules, pleural effusions, acute respiratory distress syndrome, and pulmonary infiltrates, all of which are interesting in view of the increased numbers of IgA plasma cells found in respiratory tract tissue, have been observed.
Musculoskeletal System
Arthritis or arthralgia
Central Nervous System
Aseptic meningitis
Facial nerve palsy
Marked irritability
Sensorineural hearing loss
Gastrointestinal System
Hydrops of gallbladder
Abdominal pain, diarrhea
Hepatic dysfunction, obstructive jaundice
Pancreatitis
Genitourinary System
Urethritis, meatitis
Respiratory System
Perihilar infiltrates or pulmonary nodules
Preceding respiratory illness
Other
Erythema and induration of bacille Calmette-Guérin vaccine site
Anterior uveitis (mild)
Desquamating groin rash
Flare of atopic dermatitis or psoriasis
Peripheral gangrene (young infants)
By far the most important associated feature of KD is cardiac involvement. Cardiac manifestations can be prominent in acute KD and are the major cause of long-term morbidity and mortality. In addition to the coronary artery abnormalities that develop in 20% to 25% of untreated children, pericardial effusion and myocarditis with congestive heart failure, tachycardia, gallop rhythm, nonspecific changes on the ECG, or arrhythmia may occur. An imperfect correlation exists between clinically apparent cardiac involvement and echocardiographic evidence of coronary abnormalities, although in the pre-IVIG era, echocardiographic evidence of mitral regurgitation, impaired left ventricular function, or pericardial effusion in the acute stage was shown to be predictive of subsequent coronary abnormalities. Echocardiographic evidence of perivascular brightness of coronary arteries was suggested as a criterion for early diagnosis of KD, but subsequent reports did not confirm the reliability of this finding as a useful sign of incomplete KD. Acute KD may involve pericardium, myocardium, endocardium, coronary arteries, and cardiac valves. Clinical and auscultatory features may include a hyperdynamic precordium, tachycardia out of proportion for the child’s age and temperature, a gallop rhythm, and a flow murmur. Some infants may manifest very low cardiac output, and ECG changes (ST-segment and T-wave changes, prolonged PR interval, and arrhythmias) may be present.
A relatively recently recognized unusual presentation of acute KD is hypotension or shock. Approximately 5% of patients may require admission to an intensive care unit for suspected septic shock or toxic shock but are ultimately recognized to have features of KD. In the few small series of cases, females and older patients have predominated, laboratory studies reveal more dramatically elevated markers of inflammation, and patients were more refractory to therapy compared with control subjects with KD. Hyponatremia was reported recently in a group of these patients, suggesting a possible association with the syndrome of inappropriate secretion of antidiuretic hormone.
Incomplete or Atypical Kawasaki Disease
A substantial subset of children have illnesses that do not completely fulfill diagnostic criteria for KD but that include at least 5 days of fever and one, two, or three features of the disease. Incomplete or atypical KD ( incomplete is the preferred term) is associated with a substantial risk for development of coronary artery aneurysms, but it can be very difficult to diagnose. Incomplete KD occurs most frequently in young infants, who unfortunately are at greatest risk for developing coronary disease with KD, and fatalities have occurred. The laboratory profile of incomplete cases is similar to that of classic cases, and laboratory results can increase or decrease the likelihood of the presence of KD in a particular patient. Echocardiographic findings in the acute stage, including coronary ectasia, decreased myocardial contractility, pericardial effusion, and mild mitral regurgitation, can support the diagnosis of KD. Individual manifestations of KD in young infants tend to be more subtle than those in older children and can be fleeting. KD should be considered in the differential diagnosis of prolonged fever in infants because patients are described in whom such fevers are virtually the sole manifestation of KD.
A committee of the American Heart Association developed a valuable algorithm to assist in the evaluation of patients with suspected incomplete KD ( Fig. 77.1 ), with emphasis placed on clinical assessment and measurement of acute-phase reactants (CRP, ESR) in patients with 5 or more days of fever and two or three features of KD or in infants with 7 days or more of fever without other explanation. In addition to elevated ESR (≥40 mm/h) and/or CRP (≥3.0 mg/dL), a set of six supplementary laboratory criteria that can be useful in this regard include albumin, 3.0 g/dL or less; anemia for age; increased alanine aminotransferase; platelets after day 7 of more than 450,000/mm 3 ; white blood cell count of 15,000/mm 3 or more; and 10 or more white blood cells per high-power field in the urine sample. Hyponatremia is also useful in this regard. Echocardiographic features that support a KD diagnosis are Z score of left anterior descending (LAD) or right coronary artery (RCA) of 2.5 or greater, coronary aneurysm observed or three or more other suggestive features, including decreased LV function, mitral regurgitation, pericardial effusion, or LAD or RCA Z score of 2 to 2.5.
Retrospective diagnosis of KD often is based on finding coronary abnormalities on an echocardiogram, although the clinician’s goal is to identify patients with incomplete KD before coronary changes have occurred. The existence of these patients again emphasizes the need to identify the etiologic agent of KD so that a diagnostic test can be developed. When possible, patients with illnesses suggesting incomplete KD should be referred to physicians with considerable experience in making the diagnosis.
Laboratory Findings
A specific diagnostic test for KD is not available and awaits discovery of the etiologic agent of the illness ( Box 77.3 ). The laboratory features of KD, although quite nonspecific, are nonetheless characteristic of the illness. Leukocytosis, especially with neutrophilia, is typical in the acute stage, with a predominance of immature and mature granulocytes. White blood cell counts in excess of 30,000/mm 3 occur rarely, and those in excess of 15,000/mm 3 occur in approximately 50% of patients. Leukopenia is unusual. Toxic granulations and Döhle bodies occasionally are seen on peripheral blood smear. Anemia may develop, usually with normocytic red blood cell indices, particularly in patients with more prolonged duration of active inflammation. Severe hemolytic anemia requiring transfusions occurs but is unusual and usually is related to IVIG therapy. Curiously Kawasaki’s first patient in 1961 (and only rare subsequent patients) manifested Coombs-positive hemolytic anemia.
Leukocytosis with neutrophilia
Elevated erythrocyte sedimentation rate
Elevated C-reactive protein (and other acute-phase reactants)
Anemia
Thrombocytosis after week 1
Sterile pyuria
Hypoalbuminemia
Hyponatremia
Elevated serum levels of aminotransferases and γ-glutamyltransferase
Plasma lipid abnormalities
Cerebrospinal fluid pleocytosis
Synovial fluid pleocytosis
Elevation of acute-phase reactants such as ESR, CRP, and α 1 -antitrypsin is nearly universal in KD. CRP values rise and fall much more quickly than do ESR values. Additionally, IVIG therapy per se leads to elevation of the ESR (but not CRP) for several weeks, and thus ESR is not useful in assessing the degree of inflammatory activity in IVIG-treated patients; CRP or other acute-phase reactants clearly are superior for this purpose.
A very characteristic feature of the subacute phase of illness is thrombocytosis, with platelet counts ranging from 500,000 to more than 1.5 million/mm 3 . Thrombocytosis rarely is present in the first week of illness, usually appears in the second week, and peaks in the third week, with a gradual return to normal by 4 to 8 weeks after onset in uncomplicated cases. The mean peak platelet count is approximately 700,000/mm 3 . In one study, infants younger than 1 year with fever without a source who had platelet counts greater than 800,000/mm 3 were 17 times more likely to be diagnosed with KD ultimately than were infants with platelet counts lower than 800,000/mm 3 . No difference exists in chromium-65–labeled autologous platelet survival between cases and controls, and little correlation exists between thrombocytosis and increased platelet aggregation. The latter has been detected in patients with KD from a few days until a year after onset. The rare patients with thrombocytopenia in the acute stage of KD, most often young patients, are at increased risk for development of coronary artery disease and myocardial infarction. The mechanism of thrombocytopenia appears to be low-grade consumptive coagulopathy.
Plasma lipids are markedly perturbed in acute KD, with depression of plasma cholesterol, high-density lipoprotein (HDL) cholesterol, and apolipoprotein A-I (apoA-I). Similar changes are observed in other conditions associated with an acute-phase response. Marked appearance of serum amyloid A (SAA) protein in plasma, associated with HDL3-like lipoprotein particles, is seen acutely. Cabana and colleagues also showed that total cholesterol, HDL cholesterol, apoA-I, and triglyceride levels normalize over the course of several weeks and that SAA disappears from plasma. The core composition of HDL normalizes more slowly than do plasma HDL cholesterol and apoA-I levels, a finding suggesting that KD has a profound effect on the lipoprotein profile acutely and a more subtle sustained effect on HDL composition.
Mild to moderate elevations in serum aminotransferase levels are present in as many as 40% of patients, and mild hyperbilirubinemia occurs in approximately 10%. Plasma γ-glutamyl transpeptidase levels are elevated in most patients. Hypoalbuminemia and hyponatremia are associated with more prolonged and more severe disease. Urinalysis reveals intermittent mild to moderate sterile pyuria in approximately one-third of patients, although suprapubic urine generally does not show pyuria, a finding suggesting urethritis. In those children who undergo lumbar puncture, evidence of aseptic meningitis, with a predominance of mononuclear cells, normal glucose, and normal to mildly elevated protein levels, is a common finding.
Laboratory tests, even though nonspecific, can provide diagnostic support in patients with clinical features that are suggestive but not diagnostic, of KD and may aid in prediction of nonresponder patients, at least in Japan. A moderately to markedly elevated CRP (>3.0 mg/dL) or ESR (>40 mm/h), almost universal in KD, is an uncommon finding in most viral exanthems and hypersensitivity reactions. Platelet counts higher than 450,000 mm 3 usually are present in patients with KD after the seventh day of illness. In cases of incomplete KD associated with coronary abnormalities, thrombocytosis and elevated ESR/CRP are very common events in the acute stage. Clinical experience suggests that KD is unlikely if platelet counts and a full panel of acute-phase inflammatory reactants (e.g., ESR, CRP) are normal after the seventh day of illness.
Immunologic Findings
In the first 1 to 2 weeks of KD illness, marked peripheral blood neutrophilia is observed, which coincides with the neutrophilic necrotizing arteritis that can occur in KD children who develop coronary arteritis. Transcriptome studies of the peripheral blood in acute KD reveal neutrophil activation. This prominent neutrophilic innate immune response is accompanied by secretion of many proinflammatory cytokines and chemokines and abates as fever resolves, either spontaneously or in response to IVIG therapy. This response resembles that observed during many different infectious diseases associated with systemic inflammation and therefore has not been useful in establishing a specific KD diagnostic test. Nevertheless the presence of peripheral blood neutrophilia and elevated acute-phase reactants is so characteristic of KD that the absence of these findings argues against the diagnosis.
At 2 to 3 weeks after the onset of fever, elevation of all the immunoglobulins is observed in the peripheral blood, suggesting an acquired B-lymphocyte immune response. Beginning in the first 2 weeks of illness, subacute/chronic arteritis can be observed in medium-sized muscular arteries, especially the coronary arteries, characterized by infiltration of lymphocytes, eosinophils, plasma cells, and macrophages. This process may occur in adjacent or nonadjacent regions of the arteries affected by necrotizing arteritis or in different arteries. This infiltrate is consistent with an acquired T- and B-lymphocytes immune response that appears to be oligoclonal or antigen driven. CD8 T cells are prominent in the coronary artery infiltrate, and plasma cells in this infiltrate and in other inflamed KD tissues produce predominately IgA antibodies in the first month after fever onset, with IgM and IgG plasma cells present in lower numbers. In a subset of children with KD, subacute/chronic arteritis can persist for months to years. It does not appear that autoantibodies are more prevalent in KD sera than in febrile childhood control sera. Circulating immune complexes can be detected in some KD patients but do not appear related to the development of coronary artery disease.
The arterial immune transcriptome of KD subacute/chronic arteritis is notable for activated cytotoxic T lymphocyte and type I interferon-induced gene upregulation, suggestive of an antiviral response, with no apparent dysregulation of the TNF-α gene family.
The stimulus resulting in these immunologic findings is likely to be a presently unidentified infectious agent(s). Whether the causative agent results solely in an antigen-driven immune response or can result in a superantigen response in some KD children is unclear and will likely require identification of the causative agent to be clarified.
The clinical response to IVIG treatment in KD patients generally is dramatic and rapid, and the majority of patients markedly improve during or immediately following completion of the infusion. The mechanism of action of IVIG in KD is unknown. Potential mechanisms include modulation of cytokine production, suppression of endothelial cell activation, or modification of T- and B-lymphocyte response or function. However, the very striking clinical response of KD patients to IVIG is unique and makes the theory of potential neutralization of a KD-specific agent or toxin or other product of the causative agent an appealing one.
In children with KD who do not develop cardiovascular sequelae, immune system function normalizes following resolution of the clinical signs, and inflammatory markers such as the CRP shortly after fever subsides. Although some investigators propose that many infectious agents may result in KD in genetically predisposed children, KD recurrence is actually quite uncommon, and KD children appear to have normal responses to future infections without developing KD recurrences.
Management
Treatment During the Acute Stage
Initial Therapy
Patients diagnosed with KD should be admitted to hospital, undergo a baseline echocardiogram, and receive IVIG 2 g/kg over the course of 10 to 12 hours, with high-dose aspirin at 80 to 100 mg/kg per day in four divided doses as soon as practical ( Box 77.4 ). When administered by the 10th illness day (onset of fever defines first day) this regimen is highly effective in reducing the risk for development of coronary abnormalities. Because few data exist regarding management of patients treated later than the 10th illness day, the goal is to treat by illness day 10 whenever possible. Patients who are diagnosed after the 10th illness day and who are still febrile and/or with elevated inflammatory markers may benefit from therapy, but the ability to prevent coronary changes is less certain. IVIG and aspirin have been shown to prevent the development of giant coronary aneurysms and to have direct benefits on cardiac function. The mechanism of action of IVIG in KD remains unknown. The single high-dose schedule is superior to the earlier regimen of 400 mg/kg/day IVIG for 4 days with high-dose aspirin with respect to rapidity of defervescence and normalization of acute-phase reactants, as well as in preventing development of coronary artery abnormalities. The large single-dose infusion generally is well tolerated, even in patients with decreased myocardial function. Patients should generally remain hospitalized until they have been afebrile for at least 24 hours, to ensure that they are available for retreatment if necessary.
Acute and Subacute Stages
IVIG 2 g/kg infusion over 10–12 hours plus aspirin 80–100 mg/kg per day in four divided doses (until patient is afebrile at least 3–4 days; some recommend until 14th illness day); then 3–5 mg/kg once daily for 6–8 weeks.
IVIG may be repeated if fever persists or recurs together with at least one classic sign of disease and/or elevated C-reactive protein level (see text and Box 77.5 for other alternative “rescue therapies”).
For patients thought to be at particularly high risk for development of coronary complications, an adjunctive course of corticosteroid may be considered, although the optimal dosing and duration are unclear.
Convalescent Stage
No coronary abnormalities: no therapy
Transient coronary abnormalities: aspirin 3–5 mg/kg once daily at least until resolution of coronary abnormalities
Persistent small to medium coronary aneurysms: aspirin 3–5 mg/kg once daily
Giant or multiple small coronary aneurysms: aspirin 3–5 mg/kg once daily, with or without clopidogrel 1 mg/kg/day, with warfarin or low-molecular-weight heparin for most patients
Coronary obstruction: thrombolytic therapy, surgical or interventional procedures
IVIG , Intravenous immunoglobulin.
Whether treatment with IVIG and aspirin earlier than the fifth illness day leads to better or worse outcomes remains unclear. Some reports indicate fewer coronary abnormalities in those treated before the fifth or sixth day of illness. Others report that those treated before day 5 were more likely to require additional therapy and/or have significantly worse coronary outcomes. A more recent study concludes that the observed higher risk for cardiac complications in those treated on illness days 1 to 4 are due to underlying greater initial disease severity enabling earlier diagnosis and that KD patients should be treated as early as possible.
Single infusions of IVIG at doses of less than 2 g/kg have not been demonstrated to be as effective as 2 g/kg. Terai’s comprehensive meta-analysis of all Japanese and North American IVIG treatment trials showed that the coronary artery outcome is correlated directly with the total dose per kilogram of IVIG, with 2 g/kg superior to 1.6 g/kg, which is superior to 1.2 and 1.0 g/kg; the initial dose of aspirin did not appear to influence coronary outcome. Another study confirmed this finding.
Patients diagnosed after the 10th illness day who are still febrile and/or who manifest other signs of active disease may benefit from IVIG and aspirin therapy because this treatment may result in prompt clinical improvement, with subsidence of fever and other signs of inflammation. Little evidence supports that this approach results in lower rates of development of coronary abnormalities, however. Patients beyond the 10th to 12th illness day who have become afebrile and have resolved their clinical features of KD without therapy are unlikely to benefit from IVIG. Such children should be treated instead with low doses of aspirin, 3 to 5 mg/kg once daily, and should be evaluated carefully by serial echocardiograms. In patients who have already developed coronary aneurysms and whose acute manifestations of illness and elevated inflammatory markers have resolved, there is no convincing evidence of a beneficial effect of IVIG.
Various preparations of IVIG differ in manufacturing processes and therefore in composition (e.g., the proportion of IgG monomers and the presence of proteins other than IgG). Although adverse reaction rates differ among products, clinical efficacy does not seem to differ. An exception is a report that showed IVIG prepared with β-propiolactone, which can affect the biologic activity of the Fc portion of IgG, to be less effective in KD. Immunization with live virus vaccines such as for measles and varicella should be deferred for 9 to 11 months after administration of high-dose IVIG because of impaired immune responses.
The aspirin dosage for acute KD that is most thoroughly studied in the United States is 80 to 100 mg/kg per day in four divided doses. We generally recommend that this be maintained long enough to ensure that the patient has been afebrile for at least 3 to 4 days and then reduced to a daily dose of 3 to 5 mg/kg for antiplatelet activity. Others often lower the aspirin dose earlier. The lower dose should be continued until 6 to 8 weeks after onset of illness in those who have not developed coronary abnormalities. Patients who develop coronary abnormalities should continue to take low-dose aspirin. High-dose aspirin is used for its antiinflammatory activity, whereas the much lower dose inhibits platelet aggregation. Japanese clinicians generally use an intermediate antiinflammatory dose of 30 to 50 mg/kg per day because of perceived higher rates of untoward effects in Japan. Impaired absorption and bioavailability and enhanced salicylate clearance is present in acute KD. Serum salicylate levels should be monitored only if symptoms of vomiting, hyperpnea, tinnitus, lethargy, or striking liver function abnormalities develop in children receiving aspirin.
A study that randomized U.S. patients to receive salicylates at 80 to 100 mg/kg per day or at 3 to 5 mg/kg per day for initial therapy (each regimen with 2 g/kg of IVIG) concluded that there was no difference in coronary outcome, but a more prompt clinical antiinflammatory benefit was noted in the high-dose aspirin group. In the absence of IVIG, aspirin therapy does not decrease the frequency of coronary abnormalities. Reye syndrome has been reported rarely in children taking high-dose aspirin for KD, but there are few data to suggest that low-dose aspirin poses this risk.
Adjunctive Primary Therapy
The potential value of adding corticosteroid therapy to IVIG or aspirin for primary therapy has been addressed in several trials. For example, Jibiki and colleagues compared 3 days of intravenous dexamethasone 0.3 mg/kg per day plus heparin and IVIG, with subsequent low-dose aspirin, to IVIG plus higher dose aspirin and found more prompt decreases in fever and inflammatory markers in the dexamethasone-treated group but no difference in rates of coronary abnormalities. Okada and associates showed in high-risk patients that the addition of intravenous prednisolone (one pulse dose of 30 mg/kg) to IVIG and aspirin followed by a long taper of oral prednisolone was associated with shorter duration of fever, lower coronary abnormality rates, fewer treatment failures, and more a quicker fall in CRP and circulating interleukin-2 (IL-2), IL-6, IL-8, and IL-10 levels. Inoue and colleagues compared 88 patients treated with IVIG and aspirin to 90 patients who received in addition intravenous prednisolone 2 mg/kg per day in three divided doses until afebrile and then orally until the CRP value normalized, with a subsequent taper over the course of 15 days, for a median of 23 days of corticosteroids. The results indicated fewer coronary abnormalities, shorter durations of fever, more rapid decline in CRP level, and fewer initial treatment failures in the steroid recipients. Caution must be exercised in the interpretation of the findings of these Japanese trials because the IVIG regimens and aspirin doses usually differ from those used in the United States, and echocardiograms were not interpreted by investigators blinded to the treatment group. A well-designed US trial in this regard showed little if any direct benefit of the addition of a single dose of intravenous methylprednisolone 30 mg/kg to the current IVIG and aspirin regimen, with the possible exception of the subset of patients who failed to respond to standard therapy and who required retreatment of the disease.
Several attempts have been made in Japan to develop scoring systems to predict at the time of initial presentation those patients who are at increased risk for failure to respond to standard therapy and/or to develop coronary lesions after therapy. In a retrospective analysis of 193 patients, Mori and colleagues found that elevated white blood cell and neutrophil counts and CRP levels after IVIG treatment predicted increased risk for subsequent development of coronary lesions. Egami and associates compared 279 patients who responded to initial standard treatment to 41 patients who were treatment resistant. With their scoring system giving one point each for month younger than 6 months, treatment before 4 days of illness, platelet count 300,000/mm 3 or less, and CRP of 8 mg/dL or higher and two points for an alanine aminotransferase level of 80 IU/L or higher, a total score of three points or higher identified the IVIG-resistant group with 78% sensitivity and 76% specificity. A similar predictive scoring system for IVIG unresponsiveness was developed by Kobayashi and colleagues, who gave two points each for initial treatment on illness day 1 to 4, serum sodium level of less than 133 mmol/L, aspartate aminotransferase value of 100 IU/L or higher, and neutrophils of 80% or greater and two points each for CRP of 10 mg/dL or higher, platelet count up to 300,000/mm 3 , and age 12 months or younger. This yielded 86% sensitivity and 67% to 68% specificity for patients with four or more points. The Kobayashi score was used to select patients at high risk for treatment failure in a randomized corticosteroid trial, as discussed earlier.
A more recent Japanese multicenter, randomized, open-label, blinded end points trial (the RAISE study) assessed the benefit of adjunctive primary steroid therapy for KD patients considered at high risk to be refractory to initial therapy. Using the Kobayashi score to identify patients at high risk for nonresponse to IVIG and aspirin, high-risk subjects were randomized to IVIG and aspirin (30 mg/kg per day until afebrile, then 3–5 mg/kg) or to IVIG, aspirin, and intravenous prednisolone 2 mg/kg per day for 5 days, converting to oral dosage once free of fever and then a 15-day corticosteroid taper beginning when the CRP level normalized. Significantly fewer recipients of corticosteroids developed coronary changes (3% vs. 13% at 4 weeks), and coronary z-scores were significantly lower in corticosteroid recipients. In another recent Japanese study, Ogata and colleagues evaluated corticosteroids with IVIG and aspirin 30 to 50 mg/kg per day in those patients with disease predicted to be refractory. Subjects were randomized to IVIG and aspirin alone or with a single 30 mg/kg dose of intravenously administered methylprednisolone and heparin from 2 hours before to 24 hours after the corticosteroid was given. Corticosteroid recipients had more rapid defervescence and decline in CRP levels with somewhat lower mean coronary z-scores (z ≥ 2.5 in 9% of corticosteroid group, 38.5% in the noncorticosteroid group; P = .04), with transient adverse events only in the corticosteroid group.
It can be concluded that the addition of steroid to standard IVIG/ASA regimen in high-risk patients reduces the coronary abnormality rate and IVIG failure.
Unfortunately the Japanese scoring systems to identify high-risk patients are insensitive in ethnically diverse populations as seen in North America, and it is thus difficult to know how to target adjunctive therapy outside Japan.
A well-designed trial evaluated Infliximab (Remicade), a monoclonal anti–TNF-α antibody, as adjunctive primary therapy. The results showed safety but no significant benefit regarding IVIG resistance rates or rates of coronary artery abnormalities. Addition of infliximab for primary therapy with IVIG in not recommended.
Rescue Therapy for IVIG Treatment Failures
Most patients with acute KD respond promptly to treatment with IVIG and aspirin, with defervescence and subsidence of inflammatory manifestations occurring within 36 to 48 hours. A subgroup of 10% to 20%, however, fails to show significant clinical response; these patients remain febrile 36 to 48 hours after receiving IVIG or they manifest only transient improvement, with recurrent fever and clinical evidence of inflammatory signs and elevated CRP levels. These patients need additional antiinflammatory therapy, and specific guidelines or controlled treatment trials do not exist ( Box 77.5 ). Of course, when treating apparently treatment-refractory patients, it is also prudent to reconfirm the initial diagnosis of KD.
