Diagnosis, Treatment, and Long-Term Management of Kawasaki Disease: A Scientific Statement for Health Professionals From the American Heart Association

Fever is typically high spiking (>39°C to 40°C) and remittent. In the absence of appropriate therapy, fever continues for 1 to 3 weeks. The spontaneous resolution of fever after 7 days should not be regarded as evidence that the diagnosis of KD has been excluded. Fever usually resolves within 36 hours after IVIG infusion has been completed; if not, the patient is considered to have resistance to IVIG, and further therapy is required.

Changes in the extremities are distinctive. Erythema of the palms and soles and firm and sometimes painful induration of the hands or feet often occur in the acute phase. Desquamation of the fingers and toes usually begins in the periungual region within 2 to 3 weeks after the onset of fever and may extend to involve the palms and soles. At 1 to 2 months after fever onset, deep transverse grooves across the nails (Beau’s lines) may be noted.

An erythematous rash usually appears within 5 days of fever onset. Most commonly, this is a diffuse maculopapular eruption. Scarlatiniform erythroderma and erythema multiforme-like rashes are also common. Less commonly, urticarial or fine micropustular eruptions are observed. The rash is usually extensive, primarily involving the trunk and extremities, and accentuation in the groin with early desquamation is a characteristic feature. An unusually severe form of psoriasis with plaques and pustular features can rarely occur during or after the acute KD illness.⁸⁰ Patients may also experience a flare of new-onset atopic dermatitis during the subacute phase. Bullous, vesicular, and petechial rashes are not consistent with KD and should prompt a search for an alternative diagnosis.

Bilateral bulbar nonexudative conjunctival injection usually begins shortly after fever onset and often spares the limbus, an avascular zone around the iris. Anterior uveitis is often observed by slit-lamp examination during the first week of fever.^81,82 Subconjunctival hemorrhage and punctate keratitis are occasionally observed.^82,83

Changes of the lips and oral cavity include (1) erythema, dryness, fissuring, peeling, cracking, and bleeding of the lips; (2) a “strawberry tongue,” with erythema and prominent fungiform papillae; and (3) diffuse erythema of the oropharyngeal mucosa. Oral ulcers and pharyngeal exudates are not consistent with KD.

Cervical lymphadenopathy is the least common of the principal clinical features. Lymph node swelling is usually unilateral, ≥1.5 cm in diameter, and confined to the anterior cervical triangle. In a small subset of patients, lymph node findings may be the most notable and sometimes only initial clinical finding, prompting a clinical diagnosis of bacterial lymphadenitis and significantly delaying KD diagnosis.⁸⁴ In such cases, fever persists, and other typical KD features, such as rash and conjunctival injection, will follow. Imaging studies including ultrasound and computed tomography (CT) can be helpful in differentiating KD lymphadenopathy from bacterial lymphadenitis. In KD, multiple lymph nodes are enlarged, and retropharyngeal edema or phlegmon is common. In contrast, bacterial lymphadenitis is most frequently associated with a single node with a hypoechoic core.⁸⁴ It has been increasingly recognized that cervical lymphadenopathy can be associated with deep neck inflammation leading to parapharyngeal and retropharyngeal edema and nonsuppurative phlegmon.^84,85

Other illnesses with similar clinical features (Table 3) should be considered before the diagnosis of KD is made, because the principal clinical findings that fulfill the diagnostic criteria are not specific. The presence of exudative conjunctivitis, exudative pharyngitis, oral ulcerations, splenomegaly, and vesiculobullous or petechial rashes should prompt consideration of another diagnosis.⁴⁶ Measles shares many clinical features with KD and should be considered in the differential diagnosis in any unimmunized infant or child. KD occurs more commonly in the winter and spring in nontemperate climates, when many respiratory viruses circulate, and a child with KD may have concurrent infection with a respiratory viral pathogen. In a child with clinical findings compatible with classic KD, the detection of respiratory viruses such as respiratory syncytial virus, metapneumovirus, coronaviruses, parainfluenza viruses, or influenza viruses does not exclude the diagnosis of KD.^86–88 The detection of adenovirus in a nasopharyngeal sample from a patient with suspected KD poses a particular challenge, because the illnesses have some similar clinical features.⁸⁹ Adenoviruses (particularly species C) can persist in tonsil or adenoid tissue, potentially confusing diagnosis of a subsequent febrile illness.⁹⁰ In a patient with fever, exudative pharyngitis, exudative conjunctivitis, and a nasopharyngeal sample positive for adenovirus by respiratory polymerase chain reaction assay, KD is extremely unlikely; however, the diagnosis of KD should still be considered if adenovirus is detected in a patient with nonexudative pharyngitis. Other diagnostic features of KD not commonly observed in adenovirus infection include erythema and swelling of the hands and feet, strawberry tongue, and a desquamating groin rash.⁹¹ In children with some clinical features of KD and a positive rapid test or culture for group A streptococcus who do not improve after 24 to 48 hours of effective antibiotic therapy (streptococcal carriers), the diagnosis of KD should be again considered.

Although important long-term sequelae are confined to the arterial tree (in particular, the coronary arteries), multiple other organs and tissues are inflamed during the acute illness and cause clinical symptoms. Common neurological findings include extreme irritability exceeding that observed in other febrile illnesses and aseptic meningitis in those children who undergo lumbar puncture.⁹⁸ Transient unilateral, and rarely bilateral, peripheral facial nerve palsy has been noted in rare case reports.⁹⁹ Profound sensorineural hearing loss is a rare but serious complication.^100,101 Common gastrointestinal findings include hepatitis, diarrhea, vomiting, abdominal pain, and gallbladder hydrops; pancreatitis and jaundice are less common. Genitourinary findings include urethritis, which is common, and hydrocele and phimosis, which are less common. Musculoskeletal findings include arthralgia and arthritis, involving multiple small interphalangeal joints and large weight-bearing joints during the first week of illness and predominantly large weight-bearing joints, especially the knees and ankles, in the second to third week of illness.^102,103 Respiratory findings include peribronchial and interstitial infiltrates on chest radiography; nodular infiltrates occur rarely. Erythema and induration at the site of a previous vaccination with bacille Calmette-Guérin is common in children with KD born in countries where it is used widely.¹⁰⁴ Macrophage activation syndrome occurs rarely and is often associated with IVIG resistance.¹⁰⁵

Laboratory tests, although nonspecific, provide support for a diagnosis of KD in patients with nonclassic but suggestive clinical features. Clinical experience suggests that KD is unlikely if the erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and platelet count are normal after day 7 of illness. In addition, low white blood cell count and lymphocyte predominance suggest an alternative diagnosis.

The evolution of the laboratory findings during and after the acute KD illness was summarized recently.¹⁰⁶ Leukocytosis is typical during the acute stage of illness, with a predominance of immature and mature granulocytes. Leukopenia and lymphocyte predominance suggest an alternative diagnosis. Anemia occurs commonly, is normochromic and normocytic, and resolves with resolution of inflammation. Elevation of acute-phase reactants such as ESR and CRP is nearly universal; the degree of elevation of ESR and CRP may be discrepant. The CRP normalizes more quickly than the ESR during resolution of inflammation. Moreover, the ESR is elevated by IVIG therapy, and therefore, a decreased ESR during follow-up should not be used to assess response to treatment with IVIG. The CRP is more useful as a marker of inflammation after treatment of the acute illness. Finding of a minimally elevated ESR in the setting of severe clinical disease should prompt investigation for disseminated intravascular coagulation.⁵⁵

Thrombocytosis is a characteristic feature of KD but generally does not occur until the second week, peaking in the third week (mean ≈700 000 per mm³) and normalizing by 4 to 6 weeks after onset in most cases. Thrombocytopenia is rare but may occur in the first 1 to 2 weeks of illness. Thrombocytopenia can be a sign of disseminated intravascular coagulation and is a risk factor for the development of coronary artery abnormalities. In patients with arthritis, arthrocentesis typically yields purulent-appearing fluid with a white blood cell count of 125 000 to 300 000 per mm³, a normal glucose level, and negative Gram stain and cultures.

Mild to moderate elevations in serum transaminases or gammaglutamyl transpeptidase occur in 40% to 60% of patients, and mild hyperbilirubinemia occurs in ≈10%.^106,107 Hypoalbuminemia is common and associated with more severe and more prolonged acute disease. Urinalysis may show pyuria in up to 80% of children, although this finding lacks specificity for KD.¹⁰⁸ In children who undergo lumbar puncture, ≈30% demonstrate pleocytosis with a mononuclear cell predominance, normal glucose levels, and generally normal protein levels.⁹⁸

In the absence of a diagnostic test, identification of serum or urine biomarkers of KD is an active area of research, but no biomarkers presently available have been demonstrated to be superior to elevated CRP or ESR. N-terminal moiety of B-type natriuretic peptide (NT-proBNP), likely indicative of myocardial involvement, may be elevated in some patients with KD, but this biomarker may not have sufficient discriminative ability to differentiate KD, and cut-point values for a positive result have not been clearly defined.^109,110

Cardiovascular manifestations can be prominent during the acute KD episode and are the leading cause of long-term morbidity and mortality. The pericardium, myocardium, endocardium including valves, and the coronary arteries all may be inflamed. Clinical findings during the acute illness may include a hyperdynamic precordium and tachycardia. Innocent systolic flow murmurs may be accentuated, and a gallop rhythm suggesting decreased compliance (diastolic dysfunction) of the ventricle secondary to myocardial inflammation and edema may be present. The presence of a pericardial rub, or clinical signs of pericardial tamponade, is very rare, although echocardiographic findings of small pericardial effusions are common. Valvar dysfunction occurs in ≈25% of patients regardless of coronary artery involvement and most often involves the mitral valve.¹¹¹ Children with clinically important mitral regurgitation (MR) may have a pansystolic murmur heard best between the low left sternal border and the apex. A diastolic murmur associated with important aortic regurgitation (AR) is rare.

During the acute illness, electrocardiography may show arrhythmia, including sinus node and atrioventricular node functional abnormalities, with prolonged PR interval and nonspecific ST and T-wave changes or low voltage if there is myocardial or pericardial involvement.¹¹² Increased QT dispersion, abnormalities of ventricular repolarization, and electrocardiographic signs suggestive of left ventricular (LV) dilation have been reported.^113,114 Rarely, malignant ventricular arrhythmias may be seen in the setting of myocarditis or myocardial ischemia.^115,116

Approximately 5% of children with KD in the continental United States present with cardiovascular collapse and hypotension requiring the initiation of volume expanders, the infusion of vasoactive agents, or transfer to the intensive care unit. The presence of thrombocytopenia and coagulopathy in such cases is notable, and a diagnosis of bacterial sepsis is frequently suspected at the outset. In such cases, when bacterial cultures are negative and fever persists, the diagnosis of KD should be considered. Children with shock presentation appear to be at higher risk of IVIG resistance, coronary artery abnormalities, MR, and prolonged myocardial dysfunction.^117–119

Myocarditis occurs frequently in acute KD. Reports of myocardial biopsies performed early in the disease course suggested a nearly universal incidence.¹²⁰ More recent data indicate that myocardial inflammation can be documented in 50% to 70% of patients using gallium citrate Ga 67 scans and technetium Tc 99m–labeled white blood cell scans.¹²¹ Recently, it has been demonstrated that myocardial inflammatory changes in KD occur before coronary artery abnormalities and that without concurrent ischemic damage, there is myocardial edema but little associated permanent cellular disruption or cell loss.¹²² Thus, most often, myocarditis in KD develops early, and acute LV dysfunction is generally transient and responds readily to anti-inflammatory treatment.¹¹¹ The rapid improvement in LV function differs from that observed in other causes of myocarditis. Myocarditis in KD likely improves rapidly as the inflammatory process subsides because it results from interstitial edema and inflammation and only rarely from myocardial cell necrosis.^73,122 Infrequently, acute myocardial inflammation is associated with overt ventricular ectopy, although recent information indicates more common repolarization impact than may be clinically apparent (see Long-Term Management, Arrhythmias). The exception to the more typical short-term impact of mild myocarditis in KD is the KD shock syndrome.

Early studies in KD found wide variability in the incidence of MR depending on techniques of diagnosis and variability of inclusion and exclusion criteria.^123,124 However, other clinical studies, including a contemporary multicenter US study,¹¹¹ have demonstrated a more consistent incidence of MR of 23% to 27% acutely. When detected early, the preponderance of MR as assessed with echocardiography is in the mild to moderate range of severity and does not appear to persist on follow-up. MR has been correlated with other laboratory markers of inflammation early in the course of KD, and it has been postulated to result from a pancarditis, or a “shared inflammatory mechanism” with other KD changes during the acute illness.

AR is much less common at presentation (1% of patients).¹¹¹ AR in KD is usually associated with aortic root dilation and becomes apparent early in the course of the disease. It is associated with coronary artery dilation as well.^111,125 Aortic root dilation (as indicated by an increased ascending aortic Z-score measurement) has been reported in ≈10% of patients during the acute illness.¹¹¹

The pathophysiology and pathology of coronary artery abnormalities are described in previous sections. Clinically, coronary artery abnormalities have been detected and defined based on luminal dimensions, as assessed with echocardiography or angiography. The presence of coronary artery abnormalities is considered a specific criterion supportive of the diagnosis of KD, particularly for those patients who do not meet the full clinical criteria for a diagnosis of complete KD. The coronary artery abnormalities associated with KD can be differentiated from lesser degrees of dilation that may be rarely present with other febrile illnesses.⁹⁵ The prevalence of coronary artery abnormalities in a clinical trial of initial treatment was 23% at 4 weeks after enrollment, reduced to 8% with 4 infusions of low-dose IVIG.¹²⁶ In a subsequent trial of single high-dose IVIG, this was further reduced to ≈4%.¹²⁷ These trials used absolute luminal dimensions and Japanese Ministry of Health cut points to define abnormalities and did not exclude patients with abnormalities at baseline.

Coronary artery abnormalities during the acute illness range from dilation only to aneurysms of various numbers, sizes, and characteristics, with the involvement occurring first in proximal segments and then extending distally. It is very rare to have distal involvement without some abnormalities being evident in proximal segments. In up to 80% of those patients who have significant dilation or aneurysms as noted on later echocardiograms, some abnormality is evident on the initial baseline echocardiogram obtained in the first 10 days of illness.¹²⁸ The largest proportion of patients with coronary artery abnormalities will have dilation only, characterized by luminal measurements outside the normal range but with a maximal Z score of <2.5. Dilation resolves within 4 to 8 weeks in the majority. Some patients will have coronary artery dimensions always within the normal range but with serial measurements will demonstrate reductions in luminal dimensions suggestive of dilation, using the patient as his or her own control.^129,130 The prevalence of these patients may range from 32% to 50%, which may indicate that coronary artery dilation may be more common than previously thought. However, it is unclear whether such reductions in dimensions represent resolution of inflammatory changes in the arterial walls or hemodynamic or functional factors related to fever and circulating inflammatory mediators.^95,96

Patients with severe coronary artery involvement (extensive or large/giant aneurysms) do not have cardiac symptoms unless myocardial ischemia develops secondary to severe coronary artery flow disturbances or thromboses. Symptoms and signs of myocardial ischemia/infarction may be atypical and nonspecific, particularly in infants. There have been rare case reports of rupture of a coronary artery aneurysm with subsequent myocardial ischemia and pericardial tamponade. This usually occurs during the acute illness, when aneurysms may be rapidly enlarging.

Patients with severe coronary artery involvement may also develop aneurysms of other medium-sized arteries, with rare occurrences of thromboses or rupture at these sites.^73,131 Common sites include the axillary, subclavian, brachial, femoral, iliac, splanchnic, and mesenteric arteries, usually near or at branching points. These may present clinically as pulsatile masses and bruits. The pathology is probably similar to that of coronary artery involvement, with a similar natural history that can lead to thromboses and stenoses, although often not associated with clinical symptoms, signs, or sequelae during childhood, because collateralization is common. Another rare but important complication is peripheral gangrene, often with resulting loss of digits.^132,133

Echocardiography is the primary imaging modality for cardiac assessment because it is noninvasive and has a high sensitivity and specificity for the detection of abnormalities of the proximal coronary artery segments.¹³⁴ The initial echocardiogram should be performed as soon as the diagnosis is suspected, but initiation of treatment should not be delayed by the timing of the study. Because detailed echocardiographic imaging is compromised if a child is uncooperative, sedation is frequently needed for those <3 years of age and may also be required in older, irritable children.¹³⁵ If a poor-quality initial echocardiogram is obtained because sedation was not administered, a sedated study should be repeated as soon as possible within the 48 hours after diagnosis and initial treatment. This initial study establishes a baseline for longitudinal follow-up monitoring of coronary artery morphology, LV wall motion, valvular regurgitation, and pericardial effusion. An initial echocardiogram in the first week of illness is typically normal and does not rule out the diagnosis.

Transesophageal echocardiography, invasive angiography, CMRI, and CTA can be of value in the assessment of selected patients but are not routinely indicated for diagnosis and management of the acute illness. Invasive angiography is rarely performed during the acute illness. Transesophageal echocardiography, CTA, and CMRI can be useful for the evaluation of older children and adolescents in whom visualization of the coronary arteries with transthoracic echocardiography is inadequate.^153,154 Evaluation of potential aneurysmal involvement in other arterial beds can be assessed with CMRI, CTA, and, rarely, invasive angiography, but such assessment is best performed after recovery from the acute illness, and usually for patients with severe coronary artery involvement or symptoms or signs, such as the presence of a pulsatile axillary mass.^155,156

The efficacy of IVIG administered in the acute phase of KD is well established to reduce the prevalence of coronary artery abnormalities.^126,158,159 Meta-analyses of IVIG compared with placebo have shown a conclusive decrease in new coronary artery abnormalities among IVIG-treated patients.^157,160 The mechanism of action of IVIG in treatment of KD is unknown. IVIG appears to have a generalized anti-inflammatory effect. Possible mechanisms of action include modulation of cytokine production, neutralization of toxins or other pathogenic agents, augmentation of regulatory T-cell activity, suppression of antibody synthesis, and provision of anti-idiotypic antibodies.¹⁶¹

Patients should be treated with IVIG 2 g/kg as a single infusion, usually given over 10 to 12 hours, together with ASA.¹⁶⁰ A variety of dose regimens have been used in Japan and the United States in the past. Two meta-analyses have demonstrated a dose-response effect, with higher doses given in a single infusion having the greatest efficacy.^{126,158,159,162} Furthermore, peak adjusted serum immunoglobulin G levels are lower among patients who subsequently develop coronary artery abnormalities and are inversely related to fever duration and laboratory indices of acute inflammation.^126,163 The association of lower peak immunoglobulin G levels with worse outcomes lends further support to the concept of a relationship between serum immunoglobulin G concentration and therapeutic effectiveness.

IVIG is a biological product made from pooled donor plasma, and potentially important product manufacturing differences exist. Perhaps for this reason, adverse effects appear to vary considerably among products.^164–166 Recently, Coombs-positive hemolytic anemia complicating IVIG administration has been reported, especially in individuals with AB blood type.^167–169 Aseptic meningitis can occur as a result of IVIG infusion, but it resolves quickly without neurological sequelae.¹⁷⁰ Clinical studies comparing the efficacy of different immune globulin products have been conflicting.^171–173

Measles, mumps, and varicella immunizations should be deferred for 11 months after receiving high-dose IVIG.¹⁷⁴ However, children in whom risk of exposure to measles is high may receive vaccination earlier and then be re-immunized at least 11 months after IVIG administration if they have an inadequate serological response.

Even when treated with high-dose IVIG regimens within the first 10 days of illness, 20% of children will develop transient coronary artery dilation in the proximal LAD or proximal RCA by Z-score criteria, 5% will develop coronary artery aneurysms (Z >2.5), and 1% will develop giant aneurysms according to the Japanese Ministry of Health Criteria.^158,159,175 An even greater percentage of patients (30%) will be classified as having coronary artery dilation when a Z-score cut point of 2.0 is used to define dilation.^111,140,176

Additional therapies of potential benefit are discussed below, but optimal treatment awaits delineation of the specific agent(s)/trigger(s) and pathogenesis of KD.

ASA has been used in treatment of KD for many years. Although ASA has important anti-inflammatory activity (at high doses) and antiplatelet activity (at low doses), it does not appear to lower the frequency of development of coronary abnormalities.¹⁷⁷ During the acute phase of illness, ASA is administered every 6 hours, with a total daily dose of 80 to 100 mg·kg⁻¹·d⁻¹ in the United States and 30 to 50 mg·kg⁻¹·d⁻¹ in Japan and Western Europe (see Reye Syndrome). There are no data to suggest that either dose of ASA is superior. Practices regarding duration of high-dose ASA administration vary across institutions, with many centers reducing the ASA dose after the child has been afebrile for 48 to 72 hours. Other clinicians continue high-dose ASA until the 14th day of illness and at least 48 to 72 hours after cessation of fever. When high-dose ASA is discontinued, low-dose ASA (3 to 5 mg·kg⁻¹·d⁻¹) is begun and continued until the patient has no evidence of coronary changes by 6 to 8 weeks after onset of illness. For children who develop coronary abnormalities, ASA may be continued indefinitely. Of note, concomitant use of ibuprofen antagonizes the irreversible platelet inhibition induced by ASA¹⁷⁸; thus, ibuprofen generally should be avoided in children with coronary artery aneurysms taking ASA for its antiplatelet effects.

Although corticosteroids are the treatment of choice in other forms of vasculitis, their use has been controversial for children with KD.¹⁸² Corticosteroids were used as the initial therapy for KD long before the first report of IVIG efficacy by Furusho et al in 1984,¹⁵⁹ and studies have shown either no ill effects or possible benefit. In a randomized trial of high-dose intravenous methylprednisolone plus heparin compared with heparin alone, Kijima et al¹⁸³ found that steroid therapy was associated with a greater rate of improvement in coronary artery abnormalities. In a retrospective review, Shinohara et al¹⁸⁴ found that treatment regimens that included prednisolone were associated with significantly shorter fever duration and a lower prevalence of coronary artery aneurysms. In a prospective randomized trial in 178 children treated with IVIG (1 g/kg for 2 consecutive days) and ASA (30 mg·kg⁻¹·d⁻¹) plus intravenous prednisolone (2 mg·kg⁻¹·d⁻¹) followed by an oral taper, Inoue et al¹⁸⁵ reported a lower incidence of coronary artery abnormalities and retreatment, shorter duration of fever, and more rapid decrease in CRP levels in the steroid group.

The National Heart, Lung, and Blood Institute’s Pediatric Heart Network conducted a randomized, double-blinded, placebo-controlled trial to assess the efficacy of a single dose of pulsed intravenous methylprednisolone (30 mg/kg per dose) added to IVIG (2 g/kg for 1 day) and ASA (80–100 mg·kg⁻¹·d⁻¹).¹⁷⁶ The trial showed similar coronary dimensions expressed as Z scores adjusted for BSA, absolute dimensions, and changes in dimensions, although a post hoc subgroup analysis suggested that primary corticosteroid therapy might be beneficial in preventing coronary artery abnormalities in children at highest risk for resistance to initial IVIG.

In Japan, 3 clinical studies were conducted to assess the efficacy of steroid therapy for patients at high risk for nonresponse to IVIG defined by scoring systems. Okada et al¹⁸⁶ reported a multicenter study to assess the effectiveness of pulse methylprednisolone (30 mg·kg⁻¹·d⁻¹ for a single infusion) with IVIG (2 g/kg for a day) and aspirin (30 mg·kg⁻¹·d⁻¹) as a primary treatment for high-risk patients defined by the Sano score. The steroid group had a lower incidence of coronary artery abnormalities and treatment failure compared with a historical control group. Egami et al¹⁸⁷ and Ogata et al¹⁸⁸ reported a single-center randomized trial to assess the efficacy of single-dose methylprednisolone (30 mg/kg per dose) with IVIG (2 g/kg for 1 day) and ASA (30 mg·kg⁻¹·d⁻¹) as a primary treatment for those predicted to be at high risk for resistance to IVIG treatment based on the Egami score. They noted that the steroid group had a lower incidence of coronary artery abnormalities defined by Z scores and treatment resistance. From a total of 248 patients predicted to be resistant to IVIG defined by the Kobayashi score,¹⁸⁹ the RAISE (Randomized Controlled Trial to Assess Immunoglobulin Plus Steroid Efficacy for Kawasaki Disease) Study Group conducted a multicenter, prospective, randomized, open-label, blinded end-points trial to assess the efficacy of IVIG (2 g/kg for 1 day) and ASA (30 mg·kg⁻¹·d⁻¹) plus intravenous prednisolone (2 mg·kg⁻¹·d⁻¹) for 5 days followed by an oral taper over weeks.¹⁹⁰ The steroid group had a lower incidence of coronary artery abnormalities and treatment resistance, lower coronary artery Z scores, and more rapid resolution of fever and decline in CRP levels. A recent meta-analysis that included these trials, using different regimens of steroids and different prediction scores, found that a combination of corticosteroid with standard-dose IVIG as an initial treatment in high-risk patients reduced the rate of coronary artery abnormalities.¹⁹¹ Thus, the addition of corticosteroid therapy to IVIG and ASA in the primary therapy of KD lowers the prevalence of coronary artery abnormalities, duration of fever, and inflammation among Japanese children at highest risk for IVIG resistance. However, the Japanese scoring systems for IVIG resistance and aneurysms have low sensitivity in North American populations.¹⁹² Further research is thus needed to develop predictive instruments or scores for reliable identification of high-risk children outside Japan and to test the efficacy of the RAISE steroid regimen in this population.

Early studies from Japan documented high levels of the proinflammatory cytokine TNF-α in the plasma of patients with acute KD.¹⁹³ Levels were highest in patients who went on to develop coronary artery abnormalities.

Infliximab, a chimeric monoclonal antibody that binds with high affinity to TNF-α, was the first anti-TNF-α monoclonal antibody therapy to be approved for pediatric patients. Numerous case reports and small series described successful use of infliximab to halt inflammation in highly resistant KD.^194,195 In a study of 11 KD patients with IVIG resistance who were treated with infliximab as rescue therapy, there was an apparent clinical response. Proinflammatory cytokine levels fell after infliximab treatment, but markers of vasculitis, including vascular endothelial growth factor and the S100 proteins, remained elevated.¹⁹⁶ This suggested the possibility that infliximab was effective in reducing systemic levels of inflammation but not in suppressing the vasculitis.

A 2-center, randomized, double-blind, placebo-controlled trial of infliximab plus IVIG for intensification of initial treatment enrolled 196 subjects.¹⁹⁷ The study was powered for the primary outcome measure of reducing IVIG resistance from 20% to 5%. Secondary outcome measures included reduction of inflammatory parameters and the change in coronary artery Z scores. Although the number of fever days was shortened and inflammatory parameters normalized more rapidly in the infliximab-treated subjects, the rates of IVIG resistance were identical between the 2 arms. A striking finding was the complete prevention of IVIG infusion reactions in children randomized to the infliximab arm compared with a 13% reaction rate in subjects who received placebo before their IVIG infusion. There was a significant decrease in Z score for the LAD in favor of infliximab. However, there was no difference in the rate of coronary artery aneurysms between the groups, although the study was inadequately powered for this end point. On the basis of current information, addition of infliximab to initial therapy with IVIG is safe but does not prevent recrudescent fever.

A more limited experience with etanercept (soluble TNF receptor) plus IVIG for intensification of initial therapy was reported recently.¹⁹⁸ In this prospective, open-label study of 15 patients, etanercept was administered subcutaneously after IVIG infusion and again at 1 and 2 weeks later. Most patients received 0.8 mg/kg per dose. The pharmacokinetics were similar to those reported in older children, and there were no adverse reactions attributable to etanercept. On the basis of these results, a phase III randomized, placebo-controlled trial was initiated and is still enrolling subjects.¹⁹⁹ Recommendations for primary adjunctive treatment with etanercept await publication of the results of this clinical trial. The potential advantage of etanercept might be the shorter half-life if secondary infections are of concern. However, the soluble receptor only binds to circulating and not cell-bound TNF-α, which could reduce the anti-inflammatory effect.

Approximately 10% to 20% of patients with KD have persistent or recurrent fever after primary therapy with IVIG plus ASA.^204,205 Many studies have shown that patients who are resistant to initial IVIG are at increased risk of developing coronary artery abnormalities.^171,206,207 Thus, scoring systems have been constructed to identify patients likely to be resistant to IVIG and who may benefit from more aggressive initial therapy. In 2006, 3 Japanese groups devised scoring systems to predict resistance to IVIG.^{187,189,208,209} However, currently available risk prediction models for Asian populations are insufficiently accurate to be clinically useful in North American patients in predicting response to initial treatment with IVIG.^192,210 Better predictive models, perhaps incorporating biomarkers or genetic variants, will need to be developed for use outside Japan.

Many experts recommend retreatment with IVIG 2 g/kg. The putative dose-response effect of IVIG forms the theoretical basis for this approach. Retrospective series have suggested efficacy, but IVIG retreatment has never been tested in an adequately powered randomized trial.^171,211

Corticosteroids have also been used to treat patients who have failed to respond to initial therapy for KD.¹⁸² Several small case series and observational studies have described children with KD with recrudescent or persistent fever despite IVIG treatment in whom the administration of steroid therapy was associated with an improvement in symptoms and the absence of an important progression in coronary artery abnormalities or adverse effects.^{219,221–225} In a study of first-line rescue therapy, Furukawa et al²²⁶ reported 63 IVIG-resistant patients, of whom 44 were then given intravenous methyl prednisolone (30 mg·kg⁻¹·d⁻¹ for 3 consecutive days) followed by oral methylprednisolone tapered over 7 days; 19 patients were given a second infusion of IVIG. The incidence of coronary artery abnormalities and treatment failure was similar between the 2 treatment groups, although power was limited by the small sample size. Fever tended to recur more frequently in the group treated with pulse steroids. Ogata et al²²⁷ compared the effectiveness of treatment strategies for 27 IVIG-resistant patients: 14 patients were treated with additional IVIG (2 g/kg for 1 day), and 13 were treated with pulse intravenous methylprednisolone (30 mg/kg for 3 consecutive days). Patients in the steroid group had shorter duration of fever and lower medical costs. Three patients (21%) treated with IVIG and no patients treated with steroids had coronary artery abnormalities. Again, the study was underpowered to show a significant difference between the groups. Teraguchi et al²²⁸ studied 41 IVIG-resistant patients and reported that intravenous methylprednisolone (30 mg·kg⁻¹·d⁻¹ for 3 consecutive days) followed by prednisolone (1 mg·kg⁻¹·d⁻¹) for 7 days did not improve clinical or coronary artery outcomes compared with IVIG retreatment. Kobayashi et al²¹² reported a retrospective study to assess the efficacy of intravenous prednisolone followed by an oral taper (2 mg·kg⁻¹·d⁻¹ tapered over 2 weeks after CRP normalized) using a database of 359 consecutive IVIG-resistant patients. Patients treated with IVIG plus prednisolone had significantly lower rates of persistent or recrudescent fever and coronary artery abnormalities than the group that received IVIG monotherapy. It is hypothesized that the improved outcomes associated with the longer steroid course in the Japanese studies might be attributed to suppression of persistent vascular inflammation, although there have been no clinical trials comparing different steroid regimens for patients who do not respond to an initial or a second dose of IVIG. The optimal steroid regimen is therefore not known, and both pulsed and longer-term steroid therapy remain options.

A phase I multicenter, randomized, open clinical trial of infliximab (5 mg/kg intravenously over 2 hours) versus a second infusion of IVIG (2 g/kg) was performed to determine the safety, tolerability, and pharmacokinetics of infliximab for rescue therapy for patients who had fever at least 36 hours after the end of the initial IVIG infusion.¹⁹⁵ The study enrolled 24 subjects with IVIG-resistant KD and determined that infliximab was well tolerated in infants and children with KD and that the pharmacokinetics were similar to adults, with circulating levels of the monoclonal antibody detected out to 10 weeks. In the Japanese trial, 20 KD patients resistant to 2 consecutive IVIG infusions (2 g/kg each) were treated with infliximab (5 mg/kg), and an apparent clinical response was achieved in 18 (90%).²²⁹ The 2 unresponsive patients were treated with plasma exchange with resolution of their inflammation. The coronary artery abnormalities detected by echocardiogram all subsequently resolved. There were no adverse reactions attributed to infliximab among the study subjects.

A retrospective review of 2 centers that consistently administered either a second dose of IVIG or infliximab to IVIG-resistant patients suggested that patients receiving infliximab had shorter hospitalization and fewer days of fever, but coronary artery outcomes and adverse events were similar.²¹³ On the basis of these retrospective data, infliximab can be considered as an alternative to a second infusion of IVIG for resistant patients.

Highly inflamed patients who fail to respond to either a second infusion of IVIG, steroids, or infliximab require additional therapy to control inflammation.

To date, no randomized clinical trials have evaluated the safety and efficacy of antithrombotic regimens for prophylaxis of coronary thrombosis in KD, in part because the power for such trials is limited by small patient numbers and few thrombotic events. For this reason, recommendations for use of antithrombotic agents have relied on reasoning from first principles, retrospective studies, practices in atherosclerotic coronary artery disease (CAD), and expert consensus.

Antiplatelet agents are considered to be standard of care in the therapeutic armamentarium for patients with coronary artery aneurysms. For patients with small coronary artery aneurysms, monotherapy with low-dose ASA therapy is sufficient for prophylaxis of thrombosis. In patients with moderate but not large or giant aneurysms, ASA therapy may be combined with a thienopyridine (eg, clopidogrel) to antagonize ADP-mediated platelet activation, a practice supported by the superior efficacy of such a regimen, compared with ASA alone, among adults with coronary artery or cerebrovascular disease.^241–244 Finally, ibuprofen and other nonsteroidal anti-inflammatory drugs with known or potential involvement of the cyclooxygenase pathway interfere with the antiplatelet effect of ASA to prevent thrombosis. If nonsteroidal anti-inflammatory drugs are needed for treatment of arthritis in patients with coronary artery aneurysms who are taking ASA, alternative antiplatelet therapies (eg, thienopyridines) should be considered.

Patients with large or giant aneurysms, that is, with an internal luminal diameter Z score ≥10 or absolute dimension ≥8 mm, are at particularly high risk for coronary artery thrombosis. In affected arterial segments, coronary artery thrombosis is promoted by markedly abnormal flow conditions, with low wall sheer stress and stasis, together with activation of platelets, clotting factors, and the endothelium. Over time, stenoses often develop, causing activation of platelets and endothelial dysfunction from turbulence of flow when located proximally and occluding flow and worsening stasis when located distal to an aneurysm. Indeed, most giant aneurysms seen at postmortem examination are lined by chronic thrombus.⁷³ Because both platelets and humoral clotting factors promote thrombus formation within giant aneurysms, patients are treated with a combination of antiplatelet and anticoagulant therapy, most commonly low-dose ASA together with either warfarin, maintaining an international normalized ratio of 2.0. to 3.0, or low-molecular-weight heparin (LMWH).²⁴⁵

Anticoagulation with therapeutic doses of LMWH is generally substituted for warfarin in infants and is also occasionally used in the older child in whom the international normalized ratio cannot be adequately controlled. Transition from LMWH to warfarin may be considered once aneurysms have stopped expanding and the patient is stable. During the acute phase, the anti-inflammatory actions of the LMWH could be an added advantage. Transient low levels of antithrombin occur in more than half of KD patients during the acute illness, are related to increased antithrombin consumption, and can affect the antithrombotic action of LMWH.⁵⁵ If patients fail to achieve the desired activated factor Xa level (0.5–1.0) on an appropriate dose (Table 7), then antithrombin levels should be measured. If deficient, fresh-frozen plasma or antithrombin supplementation may be given.

More aggressive regimens may be used in patients with exceptionally high risk for coronary artery thrombosis. Infants and children who recently required thrombolysis for coronary artery thrombosis may be maintained for a limited time on 3 agents, that is, ASA, a thienopyridine, and an anticoagulant. Because the risk of bleeding is greater with such a regimen, clinicians must consider its risk/benefit ratio on an individual basis. Newer oral direct factor Xa inhibitors or direct thrombin inhibitors are not yet approved for use in pediatrics but could supplant warfarin and LMWH in the future.

Because of its rarity, treatment of acute coronary thrombosis in KD patients has not been tested in randomized controlled trials. Rather, recommendations for therapy are based on guidelines in adults with ACS and small pediatric case series, with goals of reestablishing coronary artery patency and flow, salvaging myocardium, and improving survival.^247–249 Compared with coronary artery thrombosis in the adult with atherosclerotic CAD, thrombus mass in the KD patients is typically much greater. Furthermore, coronary artery thrombosis in the adult with CAD is most often caused by plaque rupture or inflammation, with exposure of lipids and extracellular matrix to the coagulation system, whereas highly abnormal flow characteristics form the basis for coronary artery thrombosis in KD patients. Methods used to reestablish coronary artery perfusion can vary with the size of the patient and expertise of caregivers; the optimal treatment is that which re-establishes blood flow most rapidly. Coronary artery thrombosis with actual or impending occlusion of the arterial lumen should be treated with thrombolytic therapy or, in patients of sufficient size, by mechanical restoration of coronary artery blood flow at cardiac catheterization.

Thrombolytic therapy with tissue-type plasminogen activator (tPA) is the most commonly administered therapeutic regimen for occlusive or near-occlusive coronary artery thrombosis in infants and children (Table 7). A common regimen of tPA is 0.5 mg·kg⁻¹·h⁻¹ over 6 hours.²⁵⁰ Of note, an alternative regimen of tPA, used in adult coronary artery thrombosis, is 0.2 mg/kg intravenously (maximum 15 mg), then 0.75 mg/kg over 30 minutes (maximum 50 mg) followed by 0.5 mg/kg over 60 minutes (maximum 35 mg).¹³⁷ It should be administered together with low-dose ASA and low-dose intravenous heparin (10 U·kg⁻¹·h⁻¹) with careful monitoring of coagulation parameters to prevent bleeding, maintaining the fibrinogen level >100 mg/dL and platelet count >50 000/mm³.²⁵¹ After completion of tPA, heparin dosage is increased as appropriate for age. The coronary artery thrombus should be reassessed with echocardiographic imaging after completion of the tPA infusion.

The large thrombus burden in the KD patient with coronary artery thrombosis, as well as the tendency for rebound thrombosis in such patients, has led some clinicians to use reduced-dose thrombolytic therapy together with a glycoprotein IIb/IIIa inhibitor, such as the monoclonal antibody abciximab (0.25 mg/kg bolus over 30 minutes, followed by an infusion of 0.125 μg·kg⁻¹·min⁻¹ for 12 hours).^252,253 In adults with ACS, inhibition of this receptor has been shown to improve outcomes, both with and without the use of thrombolytic drugs.^254–256 It may be reasonable to treat coronary artery thrombosis with substantial thrombus burden and high risk of occlusion with a reduced-dose thrombolytic therapy and abciximab. When echocardiographic surveillance in the first weeks of the illness reveals a small mural coronary artery thrombus that does not pose an urgent threat of occlusion, it may be reasonable to use abciximab rather than tPA to prevent clot extension.

The original descriptions of KD did not recognize the presence of coronary artery abnormalities until it was observed that 1% to 2% of patients died suddenly of cardiac complications.²⁵⁹ An angiographic study of 1100 patients showed coronary artery lesions in 24%, with aneurysms in 8% and a number of patients with stenoses and occlusions.²⁶⁰ The early reports of the prevalence of abnormalities vary widely given the lack of uniformity in the timing of angiography and the definition of abnormalities, and they predate echocardiography and treatment with IVIG. The clinical trial of 4-day IVIG treatment with strict entry criteria (classic KD presenting within 10 days of fever onset) and using the 1984 Japanese Ministry of Health criteria (based on absolute luminal dimensions) noted a prevalence of coronary artery abnormalities of 23% in the ASA-only group versus 8% in the IVIG-plus-ASA group at 2 weeks, with a lower prevalence at 7 weeks.¹⁷⁶ A similar subsequent trial of a single high dose of IVIG showed coronary artery abnormalities, again using Japanese Ministry of Health criteria, at 2 weeks in 9.1% of those treated with 4-day IVIG treatment versus 4.6% in those treated with high-dose IVIG (further reduced to 2.4% when those with coronary artery abnormalities at presentation were excluded).¹²⁷ Using a cut point of a Z score of 2.5 together with the Japanese Ministry of Health criteria, a further trial of single-dose methylprednisolone in addition to high-dose IVIG showed a prevalence of coronary artery abnormalities of 30% in both groups at 1 week.¹⁷⁶ The incorporation of dilation into the definition results in a higher prevalence of coronary artery abnormalities. The prevalence of dilation is further increased if one includes those patients whose coronary artery Z scores were below the cut point for abnormality (<2) but decreased significantly during follow-up.^129,130

These studies define the prevalence of coronary artery abnormalities in homogenous populations; however, some patients in clinical populations may have incomplete criteria, present late, have a missed diagnosis, or have not received appropriate treatment, all of which are risk factors for more severe coronary artery involvement. In addition, some patients presenting within 10 days of fever onset may have coronary artery abnormalities at the time of initial echocardiogram and before or at the time that IVIG is given.²⁶¹ Michihata et al²⁶² showed that patient treatment that followed existing guidelines was associated with a 4.9% prevalence of aneurysms versus 9.9% if it did not. A classification system based on Z scores alone in a clinical population of 1356 patients with serial echocardiograms has been proposed and showed overlap in Z scores for classifications based on absolute dimensions only.¹⁴⁰ An analysis of 1082 patients from centers in the United States and Japan used 2 different sources of normal values for calculation of Z scores.¹⁴¹ Using the Z-score equations of Kobayashi et al,¹⁴⁵ 26% of US subjects and 39% of Japanese subjects had a maximal Z score ≥2.5, with 4.1% and 3.1%, respectively, having a Z score ≥5. Using the Z-score equations of Dallaire et al,¹⁴⁶ 30% of US subjects and 44% of Japanese subjects had a maximal Z score ≥2.5, with 5.8% and 6.2%, respectively, having a Z score ≥5. The higher Z scores in Japanese patients remained significant after adjustment for younger age, male sex, late treatment, and failure to respond to initial IVIG. Thus, many factors need to be considered when determining and comparing the prevalence of coronary artery abnormalities, particularly the population and definitions.

Initial definition of the natural history of coronary artery abnormalities was determined from tracking luminal dimensions from serial angiography together with clinical follow-up for events. Kato et al²⁶³ reported outcomes in 598 patients diagnosed from 1973 to 1983 and followed up for up to 10 to 21 years. Aneurysms were diagnosed in 25%, with 49% of these having reduced to a normal luminal dimension 6 to 18 months later, increasing to 55% with ongoing follow-up. All aneurysms that reduced in size to a normal luminal dimension were originally small or moderate in size. Stenoses developed in 28 patients and showed a more constant risk over time. Coronary artery bypass surgery was performed in 7 patients, 2 had interventional catheter procedures, and 16 required thrombolytic treatment. MI occurred in 11 patients (8 with giant aneurysms), all with severe stenotic lesions in 2 or 3 branches. Patients without coronary artery abnormalities had no symptoms or events during follow-up. Akagi et al²⁶⁴ reported similar outcomes in a cohort of 583 patients and showed that normalization of luminal dimensions occurred more frequently and more rapidly in those with smaller aneurysms and did not occur for giant aneurysms, with MI occurring only in those with giant aneurysms. This remained true in a subsequent study of 1356 patients diagnosed from 1990 to 2007 and followed up with serial echocardiograms for up to 15.7 years.¹⁴⁰ Coronary artery events (thrombosis, stenosis, intervention, MI, death) occurred in 1% of those with an aneurysm Z score <10 and an absolute dimension <8 mm, in 29% of those with a Z score ≥10 but an absolute dimension <8 mm, and in 48% of those with both a Z score ≥10 and an absolute dimension ≥8 mm. Longitudinal studies of outcomes based on Z-score classifications alone have yet to be performed.

Late development or increases in size of aneurysms have been reported in case reports.^265–267 Tsuda et al²⁶⁸ in an angiography study of 562 patients noted new dilated or expanding lesions in 15 patients. The new aneurysms occurred at sites where previous aneurysms had diminished in size but were all associated with a localized stenosis, although none were associated with cardiac events.

MI as a result of thrombotic occlusion of an aneurysm or the development of critical stenosis attributable to LMP occurs mainly in those patients with more severe coronary artery abnormalities and can cause sudden death. In young patients, it can be clinically silent or present with atypical symptoms.²⁶³ The gradual progression of LMP and laminar thrombosis to obstructive lesions may be accompanied by the development of collateral vessels, particularly in the presence of segmental stenosis and in younger patients, regardless of the occurrence of MI.²⁶⁹ Prompt and effective management of acute MI can improve outcomes, although the reported experience is limited.²⁷⁰ Myocardial dysfunction present shortly after MI can improve, although some patients will develop adverse ventricular remodeling and ventricular aneurysms if the damage is extensive.²⁷¹ Patients who have had an MI are at further risk of subsequent MI, although the risk is reduced with effective revascularization. Tsuda et al²⁷² reported long-term outcomes up to 33 years after surviving MI in 60 patients. The 30-year survival rate was 63%, and the 25-year ventricular tachycardia–free survival rate was 29%. Low post-MI LV ejection fraction was predictive of poor outcomes.

Long-term changes in coronary artery structure and function precede clinical events and reflect chronic pathological vascular processes in areas that were acutely involved by KD. Arterial wall structure can be imaged noninvasively and has provided insights into the natural history. Intravascular ultrasound (IVUS) has been used to demonstrate symmetrical and asymmetrical wall thickening in aneurysms, particularly in those aneurysmal segments that have progressed toward normal luminal dimensions.^273,274 IVUS has also been used to characterize wall elements, noting areas of fibrofatty changes, necrotic core, and dense calcification.²⁷⁵ More recently, these changes have been assessed with optical coherence tomography (OCT).²⁷⁶ Studies have been more equivocal as to whether or not wall thickening may be evident in coronary artery segments without previous dilation or aneurysm.

These structural findings are associated with functional abnormalities, with impaired dilation in response to nitroglycerin²⁷⁷ or adenosine,²⁷⁸ resulting in impaired coronary artery flow reserve. Iemura et al²⁷⁴ studied KD patients with various degrees of coronary artery involvement who had all progressed to a normal luminal dimension on angiography. They noted that those segments that previously had large aneurysms showed paradoxical vasoconstriction in response to acetylcholine and diminished vasodilation in response to nitroglycerin. On positron emission tomography (PET) imaging, sites of previous aneurysms showed both impaired vasodilation coupled with reduced myocardial blood flow and flow reserve, particularly in segments with stenosis.²⁷⁹ In contrast, a small case-control study of KD patients with no history of coronary artery abnormalities showed reduced myocardial flow reserve on PET.²⁸⁰ PET has also been used to demonstrate persistent inflammation in patients with aneurysms, which may be reduced with statin therapy.²⁸¹ It would appear that long-term structural and functional coronary artery abnormalities are evident and associated with both maximal and current coronary artery status, with the greatest abnormalities in those with a history of large or giant aneurysms that have diminished to normal luminal dimensions. Whether or not long-term abnormalities occur in those patients who had no luminal abnormalities or only transient dilation remains a subject of debate, although the majority of evidence, including the absence of cardiovascular disease (CVD) events and lack of calcification on CT, suggests a normal long-term prognosis.

KD is a systemic vasculitis, and hence, the potential presence of generalized long-term abnormalities of systemic arterial structure and function has been a subject of controversy. Abnormal brachial artery reactivity or abnormal flow-mediated dilation, reflective of endothelial dysfunction, after KD was initially reported by Dhillon et al²⁸² in a case-control study. Since then, there have been many reports comparing noninvasive surrogate vascular markers, some additionally including measures of arterial stiffness and intima-media thickness. A systematic review and meta-analysis of 30 studies showed important heterogeneity between studies.²⁸³ Similar to studies of coronary artery structure and function, evidence for systemic arterial endothelial dysfunction was most commonly reported for KD patients in general but more consistently for those with aneurysms, whereas results for those without a history of coronary artery abnormalities were not consistent. Similar findings were noted for arterial stiffness, whereas increased mean carotid intima-media thickness was not noted for KD patients, although studies of maximal carotid intima-media thickness were conflicting. The authors concluded that surrogate vascular markers were abnormal only in KD patients with aneurysms. Serological evidence of ongoing systemic inflammation has been noted in those patients with persistent aneurysms, with higher serum amyloid A and IL-6²⁸⁴ and CRP,²⁸⁵ and imaging evidence has been suggested on PET scanning.²⁸⁶ Nonetheless, the clinical impact of these abnormalities has not been defined.

When detected early, the preponderance of MR as assessed with echocardiography is in the mild to moderate range of severity and does not appear to persist on follow-up. MR can occur after the acute stage from myocardial ischemia. Late-onset valvulitis of the mitral and aortic valves can occur very rarely and may require valve replacement.^263,287 MR that becomes severe or that persists into late phases of KD appears to occur with lower frequency and can result from persistent ischemia or, rarely, from more resistant inflammatory processes that result in fixed structural abnormalities of the valve apparatus.^287,288 This circumstance indicates more serious valvular pathology and may require surgical intervention.

AR in KD is usually associated with aortic root dilation (as indicated by an increased aortic sinus Z score) and becomes apparent early in the course of the disease. AR is of lesser severity and appears to persist more consistently, and it is less reliably associated with other inflammatory markers. Nevertheless, progressive aortic valve dysfunction associated with severe AR has been reported in a patient with recurrent KD,²⁸⁹ and aortic valve replacement was required.

Patients after KD have been shown to have functional and anatomic abnormalities of the aorta. Assessment of aortic distensibility has been used in adults in the early screening of atherosclerosis. In a study that compared 40 KD patients without coronary artery aneurysms to 168 healthy children, aortic diameter was measured at both minimum diastolic pressure and maximum systolic pressure by 2D echocardiography; this study found that aortic distensibility varies with age in normal children.²⁹⁰ It was low in infants, increased gradually to peak during the ages 10 to 15 years, and decreased with age thereafter. For KD patients, aortic stiffness was increased during the acute illness and was normal during convalescence. In a more recent study, Oyamada et al²⁹¹ demonstrated that the aortas in 75 patients with a history of KD had altered elastic properties compared with 57 control subjects. Specifically, LV mass index and aortic stiffness were significantly higher, whereas aortic distensibility and strain were lower at 5 to 10 years of follow-up in KD patients. Another study confirmed that in a KD cohort, aortic stiffness and elasticity were increased in 57 patients with KD >1 year after onset of illness, of whom 12 had coronary artery sequelae.²⁹² Both groups had altered elasticity and stiffness of the aorta.

An unusual finding during the acute KD illness is the presence of retrograde holodiastolic flow in the abdominal aorta in the absence of AR. Mori et al²⁹³ described this finding in 15 of 21 children at the time of their acute KD episode, which resolved by 1 month. The abnormal flow pattern may be attributable to an increase in the distensibility of the descending aorta, which is associated with the acute inflammation. Additionally, it was proposed that this pattern may represent diastolic runoff into dilated and dysfunctional peripheral arteries affected by the vasculitis.

An aspect of the vasculitis/inflammation noted during KD may lead to aortic root dilation. Ravekes et al¹²⁵ evaluated 100 children with history of KD from 1993 to 1997 and noted that patients had a greater normalized aortic root dimension that persisted at least to 1-year follow-up. In addition, 4% of these patients also had AR, although this was not evident on auscultation in any patient. A study of 198 children with acute KD noted that aortic root dilatation was present in 8%, remained constant at 1 and 5 weeks after diagnosis, and was associated with larger coronary artery size at diagnosis¹¹¹; however, aortic root size had little association with inflammatory markers. The long-term implications of these alterations in aortic size and properties are unknown.

Although myocarditis is common during the acute illness, complete resolution is expected. Early biopsy studies suggested the presence of myocardial abnormalities, but the relationship to the acute illness and coronary artery abnormalities was unclear.^294–296 In a more recent study of 16 patients with giant aneurysms, initial biopsy samples showed myocyte degeneration, hypertrophy, and inflammatory cell infiltration, whereas follow-up biopsy specimens showed myocyte disarray, interstitial fibrosis, and ongoing inflammatory cell infiltration.²⁹⁷ The pathogenesis of the abnormalities was not clear, but the sites of the aneurysms were not related to the biopsy findings. Subsequent studies of myocardial characterization and function have reported variable findings.^298–300 Long-term myocardial dysfunction, resulting from primary myocardial insult at the time of acute KD and which is independent of long-term coronary artery abnormalities, may very rarely occur, although evidence-based reports are few. Whether the myocarditis that occurs with the acute illness leads to long-term myocardial pathology, such as fibrosis and myocyte dropout, independent of coronary artery abnormalities is not clear. Clinical experience indicates that myocardial function is normal, except among patients with ischemic heart disease from coronary artery stenoses.^153,236

Generally, the development of important rhythm changes in KD has been documented primarily in those patients manifesting more severe forms of myocardial dysfunction, including those patients with overt myocardial ischemia or infarction. Premature ventricular contractions and ventricular tachycardia have been taken as clinical markers of underlying myocardial damage and as potential predictors of long-term consequences, including late sudden death.^272,301 In particular, after MI, the incidence of ventricular tachycardia can be increased. Patients who have sustained severe myocardial injury or infarction may benefit from extended rhythm surveillance (Holter or other long-term electrocardiographic monitoring) to best assess the need for specific antiarrhythmic therapy. However, other recent studies have demonstrated an electrophysiological impact from KD even in the absence of important ventricular dysfunction or of any coronary artery abnormalities. Ghelani et al³⁰² found increased QT interval dispersion, indicating inhomogenous ventricular repolarization, in a group of KD patients from North India. Kuriki et al³⁰³ similarly described elevations in the QT variability index during acute KD that was correlated with serum inflammatory markers and that normalized as the disease regressed.

Although conduction abnormalities have not been characteristic of KD, sinus node and atrioventricular node dysfunction have been demonstrated in patients with moderate to severe coronary artery abnormalities, although only 1 patient among the 40 studied developed evidence of atrioventricular block.¹¹²

Frequency of cardiology assessment (to include history and physical examination, echocardiography, electrocardiography):

Type and frequency of additional cardiology assessment (other cardiology testing):

Cardiovascular risk factor assessment and management:

Medical therapy (β-blockers, angiotensin-converting enzyme inhibitor [ACEI], statin):

Thromboprophylaxis:

Physical activity:

Reproductive counseling:

Frequency of cardiology assessment (to include history and physical examination, echocardiography, electrocardiography):

Type and frequency of additional cardiology assessment (other cardiology testing):

Cardiovascular risk factor assessment and management:

Medical therapy (β-blockers, ACEI, statin):

Thromboprophylaxis:

Physical activity:

Reproductive counseling:

Note: Long-term status is taken to be when the patient is stable after the acute illness and the coronary artery luminal dimensions are not increasing or progressing (usually within 15 to 45 days). Until this point, patients should be managed in accordance with the recommendations in the Acute Treatment section. Failure to follow up closely and to escalate thromboprophylaxis with progressing coronary artery aneurysms is a major contributor to unexpected morbidity and mortality.

Standard assessment includes history and physical examination, electrocardiography, and echocardiography as previously outlined. For patients whose coronary arteries have consistently remained with a maximal Z score <2 and who are therefore defined as having no involvement, discharge from cardiology care is reasonable at between 4 weeks and 12 months, provided that the recommended echocardiograms have been obtained at diagnosis and 1 and 4 to 6 weeks after acute treatment.³⁰⁵ Previous studies have shown that if there are no echocardiographic abnormalities at a 4- to 6-week assessment, further follow-up is not cost-effective, and these patients are not at risk for new onset of abnormalities.^306–308 Likewise, it is reasonable to discharge patients classified as having dilation only with Z score <2.5 if the Z score is documented to have decreased to <2 by the 4- to 6-week assessment. Otherwise, a further follow-up assessment is reasonable after 6 months to 1 year, or until the measurements are normal or an alternative explanation (dominant coronary artery branch) is evident. Although systematic long-term data are not available, evidence suggests that these patients are not at increased risk of late mortality attributable to CVD compared with the general population.^309–313 The association with late events is limited to rare case reports.^314,315

For patients defined as having an aneurysm of any size noted at any assessment, ongoing cardiology follow-up is recommended. The frequency depends on the degree of maximal and current involvement, which can be modified by other characteristics (Table 9). For patients with small or medium aneurysms, the pathological progression toward normal luminal dimensions occurs most quickly during the first year after acute treatment, whereas for patients with large or giant aneurysms, this occurs at a slower and constant rate over a longer period of time, with far fewer patients ever achieving a normal luminal dimension. The frequency of follow-up is informed by the rate of change and risk. It is also noted that the initial degree of involvement influences the frequency of follow-up based on current involvement. For example, a patient with a persistent small aneurysm would be considered at lower risk of myocardial ischemia than a patient with a large or giant aneurysm that evolved to the size of a small aneurysm, and this is reflected in the recommended frequency of follow-up. Ranges of recommended follow-up frequency represent the need to individualize follow-up, taking into account other factors as previously outlined that would increase risk.

For patients with aneurysms whose coronary artery luminal dimensions have reduced to normal or dilation only, it is reasonable to omit imaging of the coronary arteries with 2D echocardiography, although ongoing assessments for inducible myocardial ischemia are valuable.

For patients with coronary artery aneurysms, the pathological progression toward normal luminal dimension (thrombosis, luminal myofibroblastic proliferation) increases the risk of stenoses and obstructions. Hence, periodic surveillance for inducible myocardial ischemia is recommended, with timing of the first assessment and the subsequent testing frequency calibrated to the severity of maximal and current coronary artery abnormalities. In addition, patients with symptoms suggestive of myocardial ischemia should be evaluated for inducible ischemia in a timely manner. If inducible ischemia is present, further imaging is suggested, usually with invasive angiography, to determine the presence of coronary artery stenoses and occlusions. Historically, patients with significant residual coronary artery abnormalities were followed by myocardial perfusion imaging (MPI) and serial coronary angiography; however, angiography is invasive, and both modalities expose the patient to repeated radiation, which is an important issue in children. Although MPI is useful, limitations include modest specificity, lengthy acquisition time, and necessity for sedation in young children, and other modalities have gained preference. The selection of modality for surveillance for inducible myocardial ischemia should take into account the expertise of the institution, with preference for physiological stress with exercise over pharmacological stress, and minimizing the cumulative radiation dose and risks to the patient.

The long-term cardiovascular impact of KD may manifest not only in distortion of coronary artery luminal geometry but also in changes in the structure and function of the arterial endothelium and wall, as well as the myocardium. Advanced imaging methods can be applied to characterize vascular remodeling, flow reserve, endothelial dysfunction, and myocardial fibrosis, any of which can influence the prognosis and risks of selected patients with important coronary artery involvement. In the convalescent KD patient with coronary artery aneurysms, long-term specialized follow-up is recommended.³¹¹

KD patients might have a different pattern or susceptibility to CVD risk factors than the general population. Several reports have documented that high-density lipoprotein (HDL) cholesterol levels are decreased acutely after KD, sometimes together with increases in triglycerides and reduced levels of apolipoproteins AI and AII, with a variable relation to the extent of coronary artery involvement or inflammatory markers.^337–340 Qualitative changes in HDL particles have also been noted and thought to be related to acute levels of serum amyloid A.³⁴¹ Although they are improved after convalescence, lower HDL cholesterol levels can persist, particularly in those patients with severe and ongoing coronary artery aneurysms.^338,342,343 HDL and inflammation are known to be interrelated, and chronic changes may reflect this relationship. Nuclear magnetic resonance lipoprotein particle analysis suggests that pediatric and adult patients with KD, regardless of their aneurysm status, are no more likely than age-similar, healthy control subjects to have lipid patterns associated with increased risk of atherosclerosis.³⁴⁴ Differences in blood pressure have been equivocal, with one study that used ambulatory blood pressure reporting reduced nighttime dipping³⁴⁵ and others reporting no alterations in blood pressure regulation.^{342,346–348} KD patients may be predisposed to adiposity, most likely related to lifestyle factors, particularly reduced levels of physical activity.^349–351 A single study noted elevation of glycosylated hemoglobin relative to normal control subjects.³⁴⁵ It would appear that KD patients do not differ significantly from the general population with regard to CVD risk factors; however, their increased risk based on the presence of CAD merits careful attention and more aggressive management of lifestyle and CVD risk factors.

Guidelines from the AHA have provided recommendations for prevention of infective endocarditis.³⁵² Patients with KD do not require antibiotic prophylaxis, regardless of the degree of past or current coronary artery involvement, coronary artery revascularization including stent placement, or the presence of valvular regurgitation.

Patients with coronary artery aneurysms after KD may merit medical therapy to minimize the risk for and the degree of myocardial ischemia. Myocardial ischemia in patients after KD may result from structural abnormalities, including coronary artery stenoses or obstructions, and extreme sluggishness of flow through capacious aneurysms, particularly those in distal segments.³²⁸ It may also arise from functional abnormalities, including vasospasm, endothelial dysfunction, and impaired myocardial flow reserve. KD patients with CAD may be asymptomatic or may have stable symptoms of ischemic heart disease, such as exertional chest pain or dyspnea. In addition to compromised coronary artery perfusion, some KD patients may also have ventricular dysfunction resulting from prior MI, which increases their potential for myocardial ischemia, symptoms, and progression and can modify the choices of therapy.

The evidence base specific to KD patients is sparse, but data acquired from extensive experience with atherosclerotic disease in adults identifies several effective approaches to myocardial protection in the setting of coronary obstruction. These interventions have been addressed in recent adult guidelines for the management of stable ischemic heart disease.³⁵³ Medical interventions can be considered in 2 roles: first, of avoiding MI and death, and second, of controlling chest pain and other ischemic symptoms.

β-Blockade decreases the risk of MI and death by reducing myocardial oxygen demand. For atherosclerotic disease, β-blockers are a critical part of management, and their effects should extend to the pathophysiology of KD coronary disease as well. Use of β-blockade is a Class I indication for all adult patients who have had MI or ACS for the 3 years after the event, irrespective of LV function (Level of Evidence B), and indefinitely in patients with LV systolic dysfunction (LV ejection fraction <40%) with heart failure or prior MI, unless a contraindication exists (Level of Evidence A). β-Blockers may be considered for KD patients of all ages, particularly those at high risk of myocardial ischemia because of large or giant coronary artery aneurysms. Carvedilol, metoprolol succinate, or bisoprolol are the β-blocking agents that have been shown to reduce risk of death. The consideration of β-blocking agents has been incorporated into the long-term management algorithm for KD patients with large or giant aneurysms that persist.

Renin-angiotensin-aldosterone blocker therapy has also been shown to be protective against MI and death in atherosclerotic CVD, and similar protection may be anticipated in KD patients with reduced ventricular function, although this has not been proven. ACEIs are recommended in all adult patients with atherosclerosis and stable ischemic heart disease who have the incremental risks of hypertension, diabetes mellitus, LV ejection fraction ≤40%, or chronic kidney disease. In patients who are intolerant of ACEIs, angiotensin receptor blockers are recommended.

For relief of symptoms of ischemia, β-blockers should be used as initial therapy (Class I, Level of Evidence B for atherosclerotic patients); if these are inadequate for symptom control, calcium channel blockers or long-acting nitrates should be added or used instead of β-blockers in intolerant patients. Sublingual nitroglycerine or nitroglycerine spray is recommended for immediate control of angina.

Hydroxymethylglutaryl coenzyme-A reductase inhibitors (statins) are a cornerstone of therapy for the primary and secondary prevention of atherosclerotic cardiovascular events in adults.³⁵⁴ In addition to lowering low-density lipoprotein cholesterol, statins have potentially beneficial pleiotropic effects on inflammation, endothelial function, oxidative stress, platelet aggregation, coagulation, and fibrinolysis. Although controversy continues concerning whether the vascular pathology of KD may have features of atherosclerosis, statins could have a role in the long-term management. KD patients have been variably shown to have chronic inflammation and reduced HDL-cholesterol levels. In addition, endothelial dysfunction, increased vascular stiffness, and intima-media thickening have been noted in both affected coronary arteries and in systemic arteries. In the setting of familial hypercholesterolemia, children and adolescents treated with statins showed normalized endothelial function and regression of carotid intima-media thickening.^355,356 To date, these studies have included patients as young as 6 years. Short-term small studies in KD patients with aneurysms treated with statins have shown reductions in high-sensitivity CRP and improved endothelial function.^357–359 A review of empirical statin use in 20 KD patients as young as 8 months with aneurysms who were treated for a median of 2.5 years showed only transient laboratory abnormalities and no effect on growth.³⁶⁰ Given this discussion, empirical treatment with low-dose statin may be considered for KD patients with past or current aneurysms, regardless of age or sex.

ACS includes ST-segment elevation MI (STEMI), non–STEMI, and unstable angina. Patients with KD may present with STEMI in the setting of complete thrombosis of an aneurysm during the acute/subacute phase of the illness, thrombosis of a residual giant aneurysm later in the illness, or rupture of an atherosclerotic plaque that may have formed independently in an adult with a remote history of KD. STEMI is a medical emergency and requires an attempt at prompt restoration of anterograde flow through the vessel.

In young patients in the acute/subacute phase of the illness, the optimal means of achieving restoration of coronary flow is not known. In this setting, there is little experience with mechanical revascularization techniques, either with catheter-based techniques or with coronary artery bypass grafting (CABG) surgery. Systemic thrombolytic or intravenous antiplatelet therapy (ie, abciximab) may be the best option for these patients in potentially reducing thrombus burden and allowing for rapid recanalization of the acutely occluded vessel. Systemic and intracoronary thrombolytic therapy has been used successfully to reduce aneurysmal thrombus burden in patients with more stable presentation.^385,386 CABG should not be considered because of inherent delays in restoring anterograde flow into the occluded vessel. Percutaneous coronary interventions (PCIs) can be considered, although there may be difficulties in successfully passing a coronary guidewire through an acutely occluded aneurysm. If PCI is pursued, consideration should be given to the use of thrombectomy catheters to remove thrombus burden. Balloon angioplasty may not yield a durable result, and it is unlikely that a stent could be deployed in a stable fashion within an acutely occluded aneurysm with thrombus.

The adult with a known remote history of KD presenting with STEMI should be referred for emergency coronary angiography and determination of the best mode of revascularization. Unlike the patient in the acute/subacute phase of KD presenting with STEMI, the adult presenting with STEMI may have typical atherosclerotic disease as the cause of their STEMI, and standard PCI techniques may be appropriate. If the patient is found to have an acutely thrombosed aneurysm, then a judgment decision will need to be made by the interventional cardiologist as to whether PCI should be attempted or a pharmacological strategy should be used.

Management of ACS in adult patients with remote KD can be particularly complicated when the initial diagnosis was missed in childhood or was not followed up after the transition to adult care. Given the high incidence of ACS in the general population, such patients can surprise the adult interventional cardiologist and should be recognized as a clinical challenge unique from conventional atherosclerotic disease, and suspicion of prior KD, particularly in young adults presenting with ACS and in the setting of unanticipated aneurysmal changes, should be maintained. A recent series in a US-based population underscores the particular challenges of acute percutaneous interventions in this population, relating to the presence of coronary calcification and the potential for underestimation of true luminal dimensions and the potential to miss underlying aneurysmal distortion.³⁸⁷ These factors emphasize the importance of IVUS to demonstrate true luminal dimensions, improve stent deployment, and inform potential modifications to postprocedural anticoagulation. Clinical follow-up strategies in the patient after an ACS episode will include the previously listed recommendations for surveillance and management of CAD in adult KD patients.

Urgency for revascularization is less for patients with other forms of ACS (non-STEMI and unstable angina) provided the patient is stable from an ischemic and hemodynamic standpoint. KD patients with non-STEMI/unstable angina may present because of nonocclusive thrombosis of coronary aneurysms with distal embolization or progression of calcified stenoses later in the disease. Coronary CTA or CMRI may be helpful to understand the pathophysiology of the presentation and determine the appropriate next steps. For patients who present because of thrombosis of an aneurysm, consideration may be given to thrombolytic therapy and institution of long-term anticoagulation. Mechanical revascularization is not likely to be of marginal benefit in these patients. For patients presenting with progression of calcified stenoses as the cause of their presentation, cardiac catheterization should be considered, and revascularization can be considered as discussed below.

KD patients with stable angina typically present well after the initial presentation with KD and often will present in early adulthood. They generally have predictable angina with exertion. Symptoms tend to develop gradually, not suddenly as is typically seen with ACS. Stable angina is usually attributable to demand ischemia caused by progressive stenoses located at the inlet or outflow of coronary artery aneurysms that have mostly regressed and stabilized. Unlike typical atherosclerotic plaques seen in CAD, these stenoses tend to be circumferential rather than eccentric and are often heavily calcified.

KD patients with stable angina should undergo revascularization for left main coronary artery involvement, lifestyle-limiting angina despite maximal medical therapy, or high-risk features on noninvasive ischemia assessment. This practice would be in keeping with guidelines for adult patients with typical atherosclerotic CAD.³⁸⁸

Left main CAD, or disease that is equivalent to left main disease (ie, ostial/proximal LAD involvement and left circumflex involvement), represents anatomy that confers a high ischemic burden with activity. According to revascularization guidelines in adults with CAD, a left main stenosis of ≥50% should be considered for revascularization in a patient with symptoms and documentation of ischemia. CABG is the mode of revascularization of choice, although PCI can be considered in selected patients.

There is growing evidence that patients with angina and without high-risk coronary artery anatomy can be safely managed with medical therapy without conferring an increased risk of long-term mortality or MI. Although data are limited, this likely is true for KD patients with stable angina attributable to fixed obstructions within the coronary arteries. Therapies should be aimed at relieving angina, and this can be achieved with β-blockers, calcium channel blockers, and nitrates. However, if the angina cannot be successfully managed to the point of being acceptable for the patient, or the side effects of the antianginal medications cannot be tolerated, consideration should be given to revascularization.

Symptomatic KD patients with high-risk features on noninvasive imaging should also be considered for revascularization. High-risk features include an early positive testing for inducible myocardial ischemia or exercise-induced arrhythmias, or poor exercise tolerance (<3 MET [metabolic equivalent of task] units) because of symptoms (angina and dyspnea). The COURAGE (Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation) study randomized adult patients with stable atherosclerotic coronary disease to receive optimal medical therapy versus revascularization with PCI and optimal medical therapy.³⁸⁹ Patients in the revascularization group fared no better than patients in the optimal medical therapy group in terms of death and MI at follow-up. However, the nuclear substudy suggested a potential mortality benefit in patients who underwent revascularization for CAD that resulted in ≥10% of the myocardial muscle mass becoming ischemic.³⁹⁰ Therefore, in symptomatic KD patients with this threshold of ischemic muscle mass, revascularization may be reasonable.

Prior reports have suggested that symptomatic KD patients with coronary stenoses estimated to be ≥70% on coronary angiography should be considered for revascularization independent of physiological assessment of the lesion. However, revascularization based on lesion severity alone (the “oculostenotic reflex”) in stable patients has not proven to be of benefit. It would be reasonable to consider revascularization in patients if ischemia testing demonstrates ischemia in the myocardial territory subtended by the lesion. In patients referred for cardiac catheterization without prior noninvasive testing, measurement of FFR in the catheterization laboratory should be considered as a risk-stratification tool to determine need for revascularization. Patients with FFR >0.80 can safely be managed medically without an increased risk of death, MI, or delayed need for target-vessel revascularization. However, based on the results of the recent FAME II (Fractional Flow Reserve versus Angiography for Multivessel Evaluation 2) study, patients with FFR ≤0.80 may be at increased risk for urgent readmission for unstable angina and need for target-vessel revascularization.³⁹¹ These patients may be considered for revascularization after FFR assessment.

Patients with silent ischemia (documented ischemia on noninvasive testing or FFR assessment in the absence of symptoms) represent a difficult subset of patients to manage because the optimal management of these patients is not certain. This holds true for patients with KD as well. Patients with silent ischemia are believed to have an altered warning system and do not sense angina in a typical fashion. In stable patients with obstructive coronary lesions, angina relief is the major benefit of coronary revascularization; however, whether revascularization positively affects outcomes of patients with silent ischemia is not well understood at this time. Revascularization should certainly be considered in KD patients with silent ischemia who have left main coronary artery involvement (or left main equivalent involvement) or who have high-risk features on noninvasive assessments for ischemia. It would be reasonable to consider revascularization in KD patients with silent ischemia who have ≥10% of myocardial muscle mass that is ischemic on MPI. It would also be reasonable to consider revascularization in patients with FFR ≤0.80.

Once the decision to proceed with revascularization is made, the decision between CABG and PCI can often be difficult, and the risks and benefits of both procedures have to be weighed carefully before a route is selected.

There are several factors favoring CABG surgery over PCI. Patients with left main coronary involvement or multivessel coronary artery involvement will be better treated with CABG. It is likely that more complete revascularization can be achieved with CABG, particularly if there is the presence of ≥1 chronic total occlusions. Patients with multivessel coronary artery involvement and reduced LV function (because of either prior MI or chronic ischemia) may also benefit more from CABG, again because of the greater likelihood for complete revascularization. A third subset of patients who may fare better with CABG are diabetic patients. The recently published FREEDOM trial (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease) showed a survival benefit for diabetic patients treated with CABG versus multivessel PCI.³⁹² Although these were all patients with typical atherosclerotic CAD, whether these findings would apply in a clinical trial of KD patients with diabetes mellitus is unknown. It stands to reason, however, that the conclusions would be similar. Patients with single-vessel disease requiring revascularization but for whom prior PCI attempts have failed or for whom PCI is deemed unfeasible should also be considered for CABG. Finally, age plays an important role in the decision for CABG versus PCI. CABG is favored in older children and younger adults, although children in the first decade of life have been treated with CABG.

If CABG is deemed the optimal revascularization strategy, every effort should be made to use both mammary arteries for conduits. Unlike saphenous vein grafts, the length and diameter of mammary artery grafts continues to grow as children grow. Furthermore, the patency of mammary artery grafts over the long-term is superior to that of saphenous vein grafts. In a Japanese survey of KD patients who underwent CABG since 1975 (the largest published series to date of KD patients who have undergone surgery), the patency rates for mammary artery grafts at 1, 5, and 15 years were 95%, 91%, and 91%, respectively, when the operation was performed at >12 years of age.³⁹³ In patients for whom the operative age was ≤12 years, the patency rates were less at 93%, 73%, and 65% at the same respective time points. The overall patency of saphenous vein grafts was 65%, 53%, and 48%, respectively, at 5, 10, and 15 years after CABG in all patients. There are no published data regarding the patency of radial artery or gastroepiploic artery grafts in patients with KD.

Although graft failure of mammary artery grafts is uncommon, particularly in older children, it is important to ensure that a mammary artery is used as a bypass conduit only for arteries with physiological stenoses. Mammary artery grafts for angiographically borderline lesions may fail to mature because of significant competitive flow from the native circulation. If prior ischemia testing has failed to show ischemia in the territory supplied by the lesion, strong consideration should be given to FFR assessment in the catheterization laboratory before grafting of the vessel with a mammary artery is considered.

Complications from CABG for KD are no different than for CABG for patients with CAD and include an aggregate risk of 1% to 2% for major complications, including death, MI, major bleeding, stroke, and renal failure. Patients with KD tend to be much younger than patients with CAD and do not have the diffuse atherosclerosis that contributes to major complications from CABG in adults.

The long-term clinical outcome of KD patients treated with CABG appears to be favorable. In the 244 patients who were included in the survey, there were 15 deaths (1 operative death, 12 late deaths, and 2 noncardiac deaths).³⁹³ Fourteen patients required repeat CABG operations, and another 17 patients required PCI for graft stenoses. As would be expected, patients with normal LV function experienced better long-term survival.

The factors that tend to favor PCI as the optimal revascularization strategy include single-vessel disease, multivessel disease with focal and easily treated lesions, normal LV function, and absence of diabetes mellitus. Patients with coronary artery anatomy severe enough to warrant CABG but with significant medical comorbidities that make CABG too high of a risk can also be considered for PCI provided the lesions are technically amenable to PCI and the risk of PCI is acceptable. Finally, patients who refuse CABG can be considered for PCI as well, provided the risks and benefits of PCI compared with CABG are carefully discussed and the patients are aware of the potential long-term consequences of their decisions. Patients may wish to take the less invasive route to minimize the recovery time required, and in most cases, this would not preclude CABG in the future should it become necessary.

Most of the experience with PCI has been accumulated in Japan, and at this time, there are very few large-scale data to evaluate the long-term efficacy of PCI in patients with KD. PCI techniques that can be used to treat stenotic lesions in patients with KD include balloon angioplasty, with or without coronary stenting, and rotational atherectomy (RA) with or without coronary stenting.

Balloon angioplasty is a poor stand-alone technique for the treatment of stenotic lesions in KD. In the years after the acute illness, these lesions become heavily fibrotic and calcified, which renders them extremely difficult to expand with balloon angioplasty alone. If the lesion can be expanded, there is often significant recoil that limits the acute result. Furthermore, the high pressures required to expand these lesions have been associated with the development of neoaneurysms at the site of dilation. There are no data or reports of the use of plaque-incising balloons in calcified stenotic lesions in KD, although their use can be considered.

For patients in whom moderate balloon inflations fail to expand the lesion, or in whom there is clear evidence of heavy calcification at the lesion, consideration should be given to RA to debride the calcium and increase the compliance of the lesion. RA has been used successfully to treat calcified lesions in KD³⁹⁴; however, the short-term and long-term outcomes have not been studied in a systematic fashion. RA also poses unique considerations in KD patients. For instance, it is likely the RA burr will need to traverse an aneurysm to address the stenotic lesion. It is conceivable that high-speed rotation of the burr in an aneurysm that is not completely thrombosed could lead to liberation and embolization of thrombotic material, although this complication has not been reported.

Regardless of whether balloon angioplasty or RA is used for lesion modification, coronary stents should be used. Stenting after balloon angioplasty will reduce the impact of recoil on restenosis, a major limitation in densely fibrotic lesions. The largest routinely used RA burr is 2 mm in diameter. Although a burr this size would be sufficient to favorably alter the compliance of the lesion, the residual lumen of 2 mm may still be too small to allow for relief of ischemia under demand conditions. Coronary artery stenting provides a means of restoring lumen dimensions that are congruent with the native vessel. RA followed by stenting has a success rate of >90% in a published Japanese series.³⁹⁵

When a stent is chosen for PCI, the choice of a bare-metal stent versus a drug-eluting stent (DES) is an important consideration. In patients with atherosclerotic CAD, numerous studies have shown that bare-metal stents are limited by a higher risk of in-stent restenosis than DESs; however, DESs may require longer to achieve complete endothelial coverage, and therefore, the time period of risk for stent thrombosis may be longer. At the present time, guidelines suggest that patients receiving bare-metal stents remain on dual-antiplatelet therapy (ASA and a thienopyridine agent) for a minimum of 30 days, whereas patients receiving DESs should remain on dual-antiplatelet therapy for a minimum of 1 year. The choice of stent will be highly individualized on the basis of the patient’s ability to take multiple antiplatelet/antithrombin agents if they require warfarin for prophylaxis in the setting of giant coronary aneurysm. The bleeding risk of 1 year of “triple therapy” may be excessive for some patients; however, if an anticoagulant is not required, a DES may be the better choice. At the present time, there are no reports regarding the long-term performance of DESs in patients with KD.

Concurrent use of intravascular imaging may be helpful in planning PCI procedures in patients with KD. IVUS is the imaging modality of choice and will provide qualitative information regarding the extent of calcification of the lesion, as well as potentially providing information regarding the composition of any aneurysms. Quantitative information that can be obtained includes reference vessel diameter, which would be helpful in the selection of appropriate stent sizes. There have been a few reports describing the use of OCT in KD patients, although this technology may be limited in patients with coronary artery aneurysms, and there are difficulties in adequately displacing the blood pool during imaging.