The Sixth International Kawasaki Disease Symposium brought together clinicians and researchers from around the world to report and discuss the state of the art related to Kawasaki disease, in the magnificent setting of the Hilton Waikoloa Village Resort, Waikoloa, Hawaii. Registration and participation exceeded previous symposia, with over 120 oral and poster presentations, and several invited and named lectures from leading authorities. The Symposium has alternated its location between Japan and Hawaii.

EPIDEMIOLOGY AND CLINICAL CHARACTERISTICS

Jane W.

Newburger

With heightened awareness and better treatment for Kawasaki disease, its epidemiology continues to evolve. Yanagawa et al. (Tochigi, Japan [Abstract 1]) reported that the mortality from Kawasaki disease in Japan has decreased from 1% in 1974 to < 0.1% in 1995, and that the incidence in 1996 was 108 per 100,000 children less than age five years. This incidence is much higher than that reported by Han et al. in Ontario, Canada [Abstract 2] (11.3 per 100,000 children less than age 5 y). Although the differences in incidence may be related to increased susceptibility among Japanese children, delays in diagnosis and treatment were found to be common in the Canadian study, especially among those in age ranges of five to nine years (20%) and ten to 14 y (38%). Treatment delays were also reported by Kuijpers et al. (Amsterdam [Abstract 48]) to be common in the Netherlands, attributed by the authors to late diagnoses by caregivers at primary care facilities. In the United States, the Centers for Disease Control (CDC) maintain only a passive surveillance system, and no nationwide increases were noted by Belay et al. (Atlanta, USA [Abstract 4]) in the occurrence of Kawasaki disease in 1994 through 1997. However, cases may be underreported to the CDC. Indeed, Dille et al. (San Diego, USA [Abstract 7]) found that, of the 21 per 100,000 children less than age 5 y in San Diego County with Kawasaki disease in 1996 and 1997, only half were reported to state or national health authorities. Epidemiologic studies also explored risk factors for Kawasaki disease and for development of coronary artery aneurysms. Using cases and controls matched for age, sex, race, and area of residence to investigate a 1997 outbreak in Colorado, Treadwell et al. (Atlanta, USA [Abstract P1]) at the Centers for Disease Control found that Kawasaki disease was associated with a recent respiratory illness and with the use of a humidifier in the affected child's room. The Harada Index has been used in Japan to identify patients at high risk for coronary artery abnormalities, and to guide selective treatment of patients. Honkanen et al. (Toronto, Canada [Abstract 40]) found that low serum albumin and infant age, but not other risk factors in the Harada index, were significant independent risk factors for aneurysms.

Clinical features of Kawasaki disease continue to be a subject of investigation. Ostrow et al. (Los Angeles, USA [Abstract 8]) reviewed patients with coronary abnormalities encountered between 1984 and 1998, to explore which features could assist in recognition of atypical Kawasaki disease by the tenth day of illness. Approximately 40% of patients who ultimately developed coronary artery abnormalities had atypical presentations. Occurring most commonly among children younger than age one or older than age five years, cases with atypical Kawasaki disease were more likely to lack criteria of cervical adenopathy, early extremity changes, and erythema of the mucous membranes. Fever, hypoalbuminemia, and a left shift in the white blood cell differential were common features in both typical and atypical presentations. Characteristics and outcomes of Kawasaki disease in older children, ages 8 to 15 y, were reviewed by Stockheim et al. (Chicago, USA [Abstract 9]). These investigators found high frequencies of delay in diagnosis and of coronary abnormalities. Gastrointestinal disturbances, meningeal irritation, and joint involvement occurred often in this age group, and these symptoms seemed to delay diagnosis despite fulfillment of criteria. Bagga et al. (London, UK [Abstract 10]) reported the occurrence of vasospastic reversible peripheral arterial disease after Kawasaki disease in infancy, suggesting that Raynaud's syndrome can be produced by Kawasaki disease and persists for a number of years. Psoriasis was reported by Eberhard et al. (Boston, USA [Abstract 34]) in the subacute and convalescent phases of Kawasaki disease, with a prevalence higher than that expected among North American children. Takahashi and Mason (Los Angeles, USA [Abstract 33]) reported three infants with pervasive developmental disorder resembling autism after Kawasaki disease, and Inage et al. (Tokyo, Japan [Abstract 35]) reported three neonates with Kawasaki disease, all with atypical presentations.

Social and economic issues also attracted some investigation. Using the Child Health Questionnaire, Baker et al. (Boston, USA [Abstract 49]) examined the late impact of Kawasaki disease on physical and psychosocial health in children. The parents of Kawasaki disease patients without coronary artery aneurysms reported long-term physical and psychosocial health in their children similar to those reported by parents of the general pediatric population. However, patients with giant coronary aneurysms exhibited lower scores in multiple health dimensions. Scott and Ettedgui (Pittsburgh, USA [Abstract 50]) examined the utility of the one-year echocardiogram currently recommended in American Heart Association guidelines for patients with no coronary abnormalities noted 2 to 6 wk after illness onset. Among children whose echocardiogram at the 4 to 6 wk assessment was normal, the late echocardiogram never demonstrated coronary artery abnormalities. If confirmed in larger studies, these data would suggest that the one-year echocardiogram could be eliminated among patients with no coronary artery abnormalities in the early phases of the illness. Erickson et al. (Boston, USA [Abstract 51]) analyzed the determinants of hospital length-of-stay among patients with Kawasaki disease hospitalized in New York State between 1992 and 1997 using administrative New York State datasets. Lower case volume, nonwhite race/ethnicity, and Medicaid insurance type were all associated with longer length of stay, implying that decreased clinical experience and lower patient socioeconomic status impact on health service utilization and resource requirements.

ETIOLOGY AND PATHOGENESIS

Stanford T.

Shulman

The etiology and pathogenesis of Kawasaki disease were the topics of many interesting investigations. One important question is whether Kawasaki disease is triggered by a superantigen (or superantigens) or is a response to a specific antigen or antigens (presumably of infectious origin). Several relevant abstracts were presented. Rowley et al. (Chicago, USA [Abstract 13]) presented data showing that IgA-producing plasma cells infiltrate the vascular tissues in fatal acute Kawasaki disease cases from the United States and Japan. She also found that sequencing the α and kappa chain genes indicates oligoclonality and somatic hypermutation, features that strongly suggest an antigen-driven IgA immune process. Abe et al. (Tokyo, Japan [Abstract 12]) studied the isolation rates of S. aureus from Kawasaki disease children and controls and the presence of staphylococcal and TSST-1 and enterotoxin genes among those isolates. No difference in rates of S. aureus recovery or presence of toxin genes was found, although younger control children had higher rates of S. aureus and of TSST-1 and SEI genes compared with older children. The significance of this is unclear.

Several abstracts examined issues related to pathogenesis of Kawasaki disease. The involvement of pro-inflammatory cytokines has been described. Bronstein and colleagues (San Diego, USA [Abstract 11]) studied genotype frequencies at the first intron region of the lymphotoxin α gene and at the promotor region of the TNFα gene to identify alleles associated with increased cytokine production. Results in Caucasian Kawasaki disease patients were compared with published Caucasian and Japanese control genotype frequency data. Compared with Caucasian controls, both Caucasian Kawasaki disease patients and Japanese controls had higher representation of the “G” allele at the TNFα locus. The relationship between this observation and Kawasaki disease susceptibility merits further study. Two abstracts presented data related to the selectins, adhesion molecules expressed on endothelial cells, leukocytes and platelets. Ishii et al. (Kurume, Japan [Abstract 14]) assessed circulating plasma levels of selectins in acute and convalescent Kawasaki disease patients and controls. They found that levels of soluble E-selectin were increased in acute Kawasaki disease, and soluble P-selectin was increased in subacute Kawasaki disease. Levels of both P and E-selectin were more elevated in Kawasaki disease patients who developed coronary abnormalities than in those who did not, particularly during the acute stage. Nakatani (Tokorozawa, Japan [Abstract 15]) used reverse-transcriptase-PCR to detect circulating endothelial cells with detectable E-selectin mRNA. Such cells were identified in 9/18 acute stage, 10/16 subacute stage and 1/13 convalescent stage Kawasaki disease patients, as well as in 1/7 healthy controls, suggesting that this may be a marker of endothelial cell injury in Kawasaki disease. Four abstracts presented data related to vascular endothelial growth factor (VEGF), a multifunctional cytokine that is an important regulator of angiogenesis and vascular permeability and that interacts with endothelial cell tyrosine kinases. Yasukawa et al. (Chiba, Japan [Abstract 16]), Ohno et al. (Fukuoka, Japan [Abstract 17]), and Ichida et al. (Toyama, Japan [Abstract 18]) separately reported the presence of elevated circulating levels of VEGF in acute Kawasaki disease, and correlations with plasma TGFβ but not with TNFα. Ghazizadeh et al. [Abstract 19] also identified increased expression of VEGF and its receptor Flt1 in cardiac tissue of Kawasaki disease patients. Terai et al. (Chiba, Japan [Abstract 21]) studied the role of monocyte chemotactic and activating factor (MCAF or MCP-1) in acute Kawasaki disease. Immunohistochemical staining in fatal Kawasaki disease showed dramatically increased MCP-1 expression in coronary adventitia and other areas of inflammation. Markedly elevated circulating MCP-1 levels were noted in acute Kawasaki disease before treatment with i.v. immune globulin (IVGG), with rapid decrease in levels after IVGG. Similar findings were not seen with IL-8 or MIP-1α. Various lots of IVGG bound to MCP-1 but not to IL-8 and were capable of blocking MCP-1-mediated calcium influx.

Sakata et al. (Kyoto, Japan [Abstract 24]) measured plasma levels of matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs) in acute and subacute Kawasaki disease patients and controls. They found markedly increased MMP-9 and TIMP-1 levels before and after IVGG and at 1 mo follow-up, and they also showed that acute Kawasaki disease sera (pre-IVGG) up-regulated both MMP-9 and TIMP-1 mRNA in human endothelial or smooth muscle cells.

TREATMENT AND RETREATMENT IN THE ACUTE PHASE

Jane C.

Burns

Several abstracts presented in this section reflect an increasing interest in the potential benefits of steroids in the management of acute Kawasaki disease. Sundel et al. (Boston, USA [Abstract 26]) reported on a pilot study of pulse methylprednisolone plus standard therapy (2 gm/kg of i.v. gamma globulin (IVIG) and aspirin 100 mg/kg/day) compared with standard therapy alone for initial treatment of acute Kawasaki disease. Nonaka et al. (Jikei, Japan [Abstract P38]) also reported on combination therapy with steroids for initial treatment. The combination therapy appeared to be safe and well-tolerated, which suggests that larger scale, prospective trials may be warranted.

Management of the patient with Kawasaki disease who fails initial standard therapy with IVIG and aspirin remains controversial. Several abstracts were presented that described the management of such patients with either retreatment with IVIG or treatment with high dose methylprednisolone. In a multicenter survey from North America reported by Burns et al. (San Diego, USA [Abstract 27]), a single infusion of high dose IVIG in conjunction with high dose aspirin resulted in cessation of fever at 48 h post-completion of the infusion in greater than 85% of patients. The number of treatment failures (new coronary artery abnormalities appearing after IVIG treatment) was low and similar to the failure rates (2% to 3%) reported in previous multicenter studies from the U.S. The retreatment of persistently febrile patients with additional doses of IVIG or with steroids was commonly practiced by the North American centers in the survey. Han et al. (Toronto, Canada [Abstract P36]) reported that whereas patients who remained febrile after IVIG and who were treated with adjunctive therapies had a higher rate of initial coronary artery abnormalities, their outcome one year after onset of Kawasaki disease was similar to that of the group who defervesced initially with standard therapy.

Other treatment modalities were presented. Plasma exchange was reported by Imagawa et al. (Yokohama, Japan [Abstract 28]) as a safe and effective alternative treatment for patients who failed to become afebrile after IVIG treatment. The use of the neutrophil elastase inhibitor, ulinastatin, as adjunctive therapy was reported by Ozawa et al. (Tokyo, Japan [Abstract P35]) and M. Nakano (Gifu, Japan [Abstract P42]), and was shown to be effective in reducing inflammatory markers and may prevent vascular damage.

Taken together, these data suggest that whereas standard therapy with high-dose IVIG and aspirin is effective in the majority of patients, other modalities to reduce vascular inflammation should be considered. There appears to be a consensus that patients who fail standard therapy are at greater risk for coronary artery abnormalities, at least in the short-term. Only large-scale, prospective clinical trials with randomized treatment arms can successfully address the question of how to most efficiently arrest the destructive inflammatory process in the arterial wall.

DIAGNOSTIC EVALUATIONS AND IMAGING

Brian W.

McCrindle

The optimal method for imaging the coronary and peripheral arterial vasculature, and assessing its adequacy in terms of providing optimal perfusion remains a topic for important investigation. Several abstracts were presented regarding the utility of imaging with ultrasound, computerized tomography, radioisotopic scans, magnetic resonance scans and positron emission tomography, both in states of rest and stress, and related to the magnitude of coronary artery involvement. The current criterion standard for imaging remains coronary angiography, with nuclear medicine perfusion scanning at rest and stress remaining the preferred method for assessing myocardial perfusion.

Hamaoka et al. (Kyoto, Japan [Abstract 37]) used intracoronary Doppler echocardiography to assess flow dynamics and reserve in patients with coronary artery abnormalities, and showed a good correlation between angiographic severity of stenosis and abnormalities in flow dynamics. Coronary flow reserve was reduced in vessels with intermediate to large-sized aneurysms and stenotic vessels, regardless of the degree of stenosis. Several patients with angiographically normal vessels but reduced perfusion on stress nuclear medicine scans showed abnormal coronary flow reserve. It would appear that intracoronary ultrasound shows promise in providing additional information regarding functional coronary artery abnormalities and the finding of abnormalities in patients with angiographically normal vessels might pose some long-term concerns.

Assessment of ventricular performance under conditions of stress is increasingly being used as an adjunct in the assessment of myocardial perfusion. Hamamichi et al. (Toyama, Japan [Abstract 41]) used low-dose Dobutamine stress radionuclide ventriculography to assess left ventricular performance in 3 groups of patients with varying degrees of coronary artery involvement. Diastolic performance was reduced only in those patients with coronary involvement and areas of abnormal perfusion on stress nuclear medicine perfusion scans. Minette et al. (Portland, USA [Abstract P45]) used dobutamine stress echocardiography in a small group of patients without coronary artery abnormalities, and showed that, whereas systolic ventricular function was normal, they were more likely to have increased end systolic wall stress and reduced left ventricular mass.

Much attention has been focused on magnetic resonance imaging as an innovative noninvasive imaging technique. Duerinckx and Takahashi (Los Angeles, USA [Abstract 38]) used magnetic resonance angiography in a small group of patients and showed that all aneurysms noted on echocardiography were adequately detected, with more distal aneurysms as noted on conventional coronary angiography also being detected. Katayama et al. (Osaka, Japan [Abstract 39]) used three-dimensional magnetic resonance angiography to assess peripheral artery involvement, and showed that the iliac and axillary arteries could be imaged well, with problems imaging the mesenteric and renal arteries. Ichida et al. (Toyama, Japan [Abstract P44]) used dipyridamole stress cine magnetic resonance imaging to assess coronary aneurysms and stenotic lesions, and showed good dynamic visualization of flow patterns within aneurysms, with less sensitivity in detecting stenotic lesions, particularly in the right coronary artery. Improving technology and experience with magnetic resonance techniques will enhance the assessment of Kawasaki disease patients.

Cardiac troponin I, a contractile protein element, may be a biochemical marker for myocardial involvement and a surrogate marker for coronary involvement in acute Kawasaki disease. Kim et al. (Kangwon-do, Korea [Abstract 43]) suggested in a case-control study that troponin I may be acutely elevated but returns to normal after treatment with IVGG. Checchia et al. (Chicago, USA [Abstract 42]) in a similar case-control study showed only subtle increases in troponin I that did not relate to the development of coronary aneurysms. The two studies used different methodologies to assess troponin I.

CORONARY ARTERY INTERVENTIONS

Hirohisa Kato

At the present time, the standard treatment for ischemic heart disease after Kawasaki disease is coronary artery bypass surgery; however, interventional catheterization for coronary artery stenoses has recently been applied. Akagi et al. (Kurume, Japan [Abstract 44]) described an experience with 49 cases from 10 institutions in Japan. Percutaneous transluminal coronary angioplasty (PTCA), one type of catheter intervention, was performed in 34 cases, with success in 74%. The mean interval from the onset of Kawasaki disease to catheter intervention in successful cases was significantly shorter than that in unsuccessful cases. PTCA is therefore effective for patients within the first few years of disease onset; however, neoaneurysm developed in 17% of the patients. Ishii et al. (Kurume, Japan [Abstract 45]) reported the effectiveness of rotational ablation (PTCRA) for severe stenosis with calcification of the coronary artery in long-term patients, with success in 90% (9/10 patients). Tateno et al. (Chiba, Japan [Abstract 47]) reported a novel therapy for myocardial angiogenesis using heparin infusion with exercise, which resulted in the promotion of the development of collateral vessels and improvements in myocardial ischemia.

In the Richard Rowe Memorial Lecture, Dr. Soichiro Kitamura (Osaka, Japan) presented his experience with coronary artery bypass surgery in 100 cases. The mean follow-up period was 6.7 y, with longest interval being 16 y. He concluded that the long-term results of bypass surgery using an intrathoracic artery and other arterial grafts are generally excellent. Preventive revascularization for myocardial ischemia is not recommended, because of the “string phenomenon” of the arterial graft. Suda et al. (Tokyo, Japan [Abstract 46]) reported on 39 cases with bypass surgery. He recommended using the gastroepiploic artery for right coronary artery or circumflex artery bypass, and demonstrated evidence of growth potential and long-term patency. Gotteiner et al. (Chicago, USA [Abstract P58]) reported an experience with bypass surgery in 4 cases. The youngest case was 8 mo-old, and had a successful result.

Catheter intervention is an effective alternative treatment for severe coronary stenosis many years after Kawasaki disease. Interventional catheterization techniques may postpone or substitute for coronary artery bypass surgery, and indeed may evolve as the standard treatment for stenotic lesions in Kawasaki disease. Regardless of the type of treatment, however, long-term results and complications should be further investigated. If coronary artery bypass surgery is performed, use of arterial grafts, e.g. using intrathoracic and gastroepiploic arteries, is preferable. In the treatment or prevention of myocardial ischemia after Kawasaki disease, a combined approach using catheter intervention and bypass surgery might be considered.

LONG-TERM OUTCOMES

Welton M.

Gersony

There is great variation in long-term outcome after Kawasaki disease, depending on the presence and size of coronary artery aneurysms. Giant aneurysms (≥8 mm in diameter) are associated with a higher risk, during the acute phase of the disease, of thrombosis or rupture. Furthermore, arterial segments with giant aneurysms also are far more likely to develop late stenosis, calcification and thrombosis. It has been reported that over 50% of patients with giant aneurysms have developed a late coronary event. Yasui et al. (Yokahama, Japan [Abstract P61]) described the development of obstructive lesions associated with giant coronary artery aneurysms in 17 of 22 right coronary artery lesions and 7 of 15 left coronary artery lesions, at a median interval of 15.3 y after Kawasaki disease. Myocardial infarction occurred in 6 of these patients. None of the giant aneurysms regressed in this study.

It is clear that persistent giant aneurysms are associated with an extremely high risk for late coronary obstructive disease and, not unexpectedly, treatment and prevention of this important complication remains an ongoing concern. Takahashi and Mason (Los Angeles, USA [Abstract 60]) indicated that combined warfarin and aspirin therapy may be effective in preventing coronary obstruction in patients with giant coronary artery aneurysms, and they encountered no severe bleeding problems in their patients treated on a long-term basis. Warfarin therapy was discontinued in patients with angiographic evidence of aneurysm regression. In addition, Etheridge et al. (Salt Lake City, USA [Abstract 58]) reported preliminary, encouraging experience with abciximab, a new anti-platelet MAb. Thrombolysis for the acute treatment of thrombotic complications was described by Kowalczyk et al. (Warsaw, Poland [Abstract P64]), with tissue plasminogen activator as the preferred agent. Finally, the innovative use of i.v. heparin as pretreatment before exercise to decrease myocardial ischemia by enhancing coronary collateral function was reported by Tateno et al. (Chiba, Japan [Abstract 47]).

The long-term outcome for patients with smaller aneurysms also is extremely important, given the possible propensity for late coronary disease in this large population of patients with apparent mild initial coronary involvement during the acute illness. The question remains as to whether the many thousands of patients who have had Kawasaki disease are prone to late coronary manifestations, even if aneurysms were small and even if they have regressed. A number of studies have indicated that pathologic changes in the wall of involved segments of coronary arteries persist. Segments of previously involved coronary arteries with regressed lesions and normal appearance by imaging studies, nevertheless, may have abnormal morphology as defined by intravascular ultrasound (IVUS). Functional abnormalities, such as abnormal responses to coronary vasodilator agents, continue to be reported, both at dilation and aneurysm sites with or without regression. Yamamoto et al. (Kyoto, Japan [Abstract P63]) found “localized stenosis” at catheterization 4 to 8 y after angiography had shown normalization of previous aneurysmal segments. Iemura et al. (Kurume, Japan [Abstract 52]) showed significant abnormalities in vascular wall morphology and function in coronary arteries at the sites of regressed aneurysms. This was true whether the original aneurysms had been large or small. They did not find abnormalities in uninvolved segments or in control subjects. Suzuki et al. (Tokyo, Japan [Abstract 55]) presented data indicating that various types of growth factors are found in thickened intimal stenotic sites in Kawasaki disease, whereas in normal children these proteins were expressed only in the media. They suggested that their results indicate that intimal thickening is an active process in the late phases of Kawasaki disease, with a mechanism different from that of atherosclerosis. Liu et al. (Tokyo, Japan [Abstract 57]) reported a high incidence of persistent microaneurysmal changes and thrombosis in the coronary microvascular ultrastructure.

Finally, the question of whether post-Kawasaki disease patients, compared with unaffected individuals, are more prone to adult atherosclerosis continues to be explored. A case-control study by Silva et al. (Toronto, Canada [Abstract 59]) found that Kawasaki disease patients at long-term follow-up had higher blood pressure, adiposity and lipid profile abnormalities than controls, and that these factors were significantly related to each other. No significant differences in endothelial function (a possible early marker of atherosclerosis) were noted as assessed by brachial artery dilatation in response to reactive hyperemia or nitroglycerin administration. Although the aforementioned findings are of concern, there have been very few reports of late unexpected coronary events in patients with persistent small or regressed aneurysms. Thus, although the morphologic and functional abnormalities described are extremely important, epidemiological and clinical data have yet to document their significance in terms of late outcome after Kawasaki disease.