Refractory Kawasaki disease

INTRODUCTION — 

Kawasaki disease (KD), also called mucocutaneous lymph node syndrome, is one of the most common vasculitides of childhood [1]. It is typically a self-limited condition with fever and manifestations of acute inflammation lasting for an average of 12 days without therapy. However, cardiovascular complications, particularly coronary artery (CA) aneurysms leading to potential occlusion and cardiac ischemia, occur in a sizeable percentage of children, even those considered low risk per established risk scores [2]. Furthermore, children with baseline CA abnormalities (z scores ≥2.5) and/or age <6 months are at high risk for CA aneurysms despite appropriate treatment [3-5].

The major sequelae of KD are dramatically decreased by therapy with intravenous immune globulin (IVIG). Nonetheless, for decades, prolonged fever has been recognized as the strongest predictor of CA abnormalities. Thus, additional therapy is advised in patients with persistent or recurrent fever after initial IVIG therapy, commonly referred to as IVIG resistance.

The treatment of refractory KD is discussed in this review. This population differs from the high-risk KD population who warrant primary intensification of therapy (ie, adjunctive therapy with the first dose of IVIG). The initial treatment of KD, diagnosis, clinical manifestations, and cardiovascular sequelae are reviewed elsewhere. (See "Kawasaki disease: Initial treatment and prognosis" and "Kawasaki disease: Clinical features and diagnosis" and "Cardiovascular sequelae of Kawasaki disease: Clinical features and evaluation".)

REFRACTORY DISEASE

Definition — Refractory KD results from failure to control the inflammatory process and is defined as any of the following:

●Persistent or recurrent fever 36 hours after completion of initial intravenous immune globulin (IVIG) therapy up to two weeks after the start of treatment [6].

●Other signs of failed initial therapy such as progressive coronary artery dilation or persistent hypotension in the setting of KD shock syndrome.

Monitoring and assessment after initial therapy — Patients should be monitored for persistent or recurrent fever after starting initial therapy. The frequency of repeat echocardiograms depends upon the degree of initial CA involvement as well as the presence of risk factors for CA dilation or IVIG resistance. Follow-up after initial therapy is discussed in greater detail separately. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Follow-up'.)

Incidence — Fever persists or returns within 48 hours in approximately 15 to 20 percent of patients with KD who are initially treated with 2 g/kg IVIG and aspirin [1,7-9]. In regard to temporal trends, the increasing use of primary intensification therapy based on the RAISE and other trials may result in less fever and IVIG resistance after initial therapy, particularly if the primary intensification is with glucocorticoids [10-12].

Risk factors — Just as the mechanism of action of IVIG in KD is not fully understood, the reason that some patients with KD fail initial IVIG therapy is also not clear. Retrospective studies have identified demographic, clinical, and laboratory characteristics at presentation that can predict which patients are more likely to require further therapy for IVIG resistance. Predictive factors that have been identified repeatedly in several studies include male sex [13], age under 12 months and particularly under 6 months [14], Asian race [15], and delayed diagnosis (table 1) [16]. Risk factors for IVIG resistance and for CA aneurysms are reviewed in greater detail separately. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Identification of patients at high risk for IVIG resistance' and "Kawasaki disease: Complications", section on 'Coronary artery abnormalities'.)

Other causes of fever — In general, a child diagnosed with KD who has persistent or recurrent fever plus one or more of the presenting signs of mucocutaneous inflammation 36 hours after the start of the IVIG infusion is likely to have IVIG resistance. However, fever plus the appearance of manifestations of an infection, macrophage activation syndrome (MAS), or other conditions should lead to evaluation for other causes of fever.

●Drug reaction – Fever that occurs during or within 36 hours after the IVIG infusion may be due to a reaction to the IVIG itself [17].

●Infection – Concurrent infection is a potential causes of recurrent or prolonged fever in patients with KD. (See "Fever without a source in children 3 to 36 months of age: Evaluation and management" and "Fever of unknown origin in children: Evaluation", section on 'Overview of evaluation' and "Clinical features and diagnosis of hemophagocytic lymphohistiocytosis".)

●Alternative diagnoses – Alternative diagnoses, such as macrophage activation syndrome (a form of hemophagocytic lymphohistiocytosis) [18], autoinflammatory diseases, systemic juvenile idiopathic arthritis (sJIA), or a chronic vasculitis (eg, polyarteritis nodosa), also should be considered when findings of KD are prolonged beyond three to four weeks. (See "Systemic juvenile idiopathic arthritis: Clinical manifestations and diagnosis" and "Vasculitis in children: Incidence and classification" and "Clinical manifestations and diagnosis of polyarteritis nodosa in adults" and "The autoinflammatory diseases: An overview".)

●Multisystem inflammatory syndrome in children (MIS-C) – MIS-C is an uncommon, postinfectious complication of coronavirus disease 19 (COVID-19) that shares specific features with KD. The clinical presentation of MIS-C may include persistent fevers, gastrointestinal symptoms (abdominal pain, vomiting, diarrhea), rash, and conjunctivitis. Compared with patients with classic KD, these patients have a higher median age of onset, higher C-reactive protein (CRP) and D-dimer values, and more commonly other cardiac complications including myocardial dysfunction and shock. (See "COVID-19: Multisystem inflammatory syndrome in children (MIS-C) clinical features, evaluation, and diagnosis".)

TREATMENT APPROACH

Initial rescue therapy

Rationale — Patients who fail to respond to the initial dose of intravenous immune globulin (IVIG) are at increased risk, up to 10-fold, of developing coronary artery (CA) abnormalities and long-term sequelae of the disease [19-22]. The duration of fever correlates with the risk of developing CA aneurysms [1]. Ideally, high-risk patients should be identified prospectively, before CA abnormalities develop or when they are mild, allowing supplemental therapy to prevent or ameliorate vascular damage. However, even patients who receive augmented initial therapy may not respond. (See 'Risk factors' above and "Kawasaki disease: Initial treatment and prognosis", section on 'Identification of patients at high risk for IVIG resistance'.)

Indications — Indications for rescue therapy include persistence or recurrence of fever or other ongoing signs of inflammation consistent with KD, including progressive CA dilation [23,24]. These patients should be retreated for presumed resistant KD unless there is clear evidence that persistent or recrudescent fever is not due to KD. Fever within 36 hours of the start of IVIG therapy does not necessarily warrant retreatment, because it may represent a reaction to the medication or a slow response to therapy. Patients should also be evaluated for other causes of fever if there is clinical uncertainty regarding the cause of fever. (See 'Definition' above and 'Other causes of fever' above and "Kawasaki disease: Initial treatment and prognosis", section on 'Follow-up'.)

Choice of agent — Therapeutic options for initial rescue therapy are based upon agents that have demonstrated effectiveness in other vasculitides and include systemic glucocorticoids, biologic response modifiers such as infliximab (a tumor necrosis factor [TNF] inhibitor) [25], and/or a second dose of IVIG (algorithm 1) [26]. Anti-interleukin 1 (IL-1) agents (eg, anakinra) have also been trialed for IVIG resistance. More potent immunosuppressive agents such as cyclosporine or cyclophosphamide are typically reserved for CA disease that is more severe or refractory to numerous other treatments.

With IVIG resistance or expanding CA dimensions, the tendency is for the clinician to layer on numerous additional therapies. Consequently, it is frequently not possible to determine which agent is actually effective. In addition, fever in KD is self-limiting, and apparent response to therapy reported in uncontrolled studies may represent abatement of the disease process that is not directly related to the therapy administered.

Standard agents — The most commonly used rescue therapy agents include glucocorticoids, infliximab, and IVIG.

Glucocorticoid therapy — Use of glucocorticoid therapy for refractory KD is supported by observational studies, limited clinical trial data, and indirect evidence from other vasculitides [7,19-21,27-33]. In these studies, many patients had a good clinical response, with resolution of fevers in the majority of patients and rates of CA aneurysms similar to those reported in patients who received an additional dose of IVIG. Use is also indirectly based upon the response of high-risk patients to initial therapy that is augmented with glucocorticoids [10]. Severe adverse effects from a two- to three-week course of glucocorticoids are rare, whereas IVIG retreatment is associated with a dose-dependent risk of hemolytic anemia in patients with non-O blood types. There is greater clinical experience with glucocorticoids than with more costly biologic response modifiers (eg, infliximab, etanercept, anakinra), but data directly comparing these treatments are lacking. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Glucocorticoids'.)

In a meta-analysis of six studies (two small randomized trials and four observational studies) involving a total of 383 patients with refractory KD, rates of CA abnormalities were similar with glucocorticoid therapy or repeat IVIG (odds ratio [OR] 0.85, 95% CI 0.57-1.56) [28]. Two other meta-analyses reported similar findings [27,29]. The first meta-analysis also found that the effects of glucocorticoids in KD were inversely proportional to the amount of time that has passed since disease onset [28]. One meta-analysis reported that more patients had fever resolution with glucocorticoid therapy compared with repeat IVIG (83 versus 62 percent; relative risk [RR] 1.36, 95% CI 1.09-1.72), based on a pooled estimate from four studies (n = 127) reporting this outcome [29].

A prospective, single-center, randomized controlled trial assessed the efficacy of second-dose IVIG versus methylprednisolone (15 mg/kg/day for three days) [34]. The methylprednisolone-treated group had fewer days of fever and more rapid normalization of laboratory studies. There was no difference in CA abnormalities at one or three months. However, the patients treated with methylprednisolone were more likely to be readmitted.

Various glucocorticoid regimens are used in patients who fail to respond to IVIG therapy, generally paralleling glucocorticoid regimens for augmented initial therapy of KD. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Additional therapy for patients at high risk for IVIG resistance'.)

We usually initiate treatment with 1 mg/kg of intravenous methylprednisone twice daily. We do not extend patients' hospitalization solely to administer glucocorticoids intravenously. Depending upon the degree of CA dilatation, patients may be observed longer in the hospital or may have frequent outpatient lab tests and echocardiograms. Patients who do not continue to improve or whose CA diameters increase on oral glucocorticoids are considered to have failed glucocorticoid therapy and are usually treated with infliximab or cyclosporine. (See 'Infliximab' below and 'Cyclosporine' below.)

Patients who have defervesced and have stable to improving CA dilatation are switched to oral prednisone (1 mg/kg twice daily) for home treatment one day before the anticipated hospital discharge. They continue this dosing until their first outpatient echocardiogram (which is typically within one week of discharge). If the C-reactive protein (CRP) is normal, the prednisolone is decreased to 1 mg/kg/day for five days, followed by 0.5 mg/kg/day for five days, then discontinued. If the CRP does not normalize and/or the coronary dimensions significantly increase, the patient is admitted for further treatment.

Our approach differs from practice in Japan, where children are usually hospitalized for the entire glucocorticoid course and rescue therapy is based upon the RAISE protocol [10]. The RAISE approach typically consists of prednisolone at a starting dose of 2 mg/kg/day intravenously [IV] divided three times daily for five days. Patients are switched to oral prednisolone if the fever defervesces after the first five days of treatment. The prednisolone taper (1 mg/kg/day to 0.5 mg/kg/day in five-day steps) is started once CRP is within the normal range (≤5 mg/L) and the patient has received at least five days of the 2 mg/kg/day dose.

An alternative to the 15-day glucocorticoid taper is IV pulsed-dose methylprednisolone (30 mg/kg per day) for up to three consecutive days, but the data do not support that this regimen is as effective as the lower, daily dosing of glucocorticoids in regard to CA outcomes.

Serious side effects from even a two- to three-week course of glucocorticoids are unusual. Rarely, a child may develop a dramatic leukemoid reaction, with absolute neutrophil counts spiking to 80,000/mm³, but the leukocytosis has not been associated with identifiable ill effects. The reaction rapidly reverses when the glucocorticoids are tapered or discontinued. Other side effects are discussed separately. (See "Major adverse effects of systemic glucocorticoids".)

Infliximab — Anti-tumor necrosis factor (TNF) alpha agents have been studied for the treatment of KD because elevated levels of TNF are a characteristic of KD [35]. Treatment of IVIG-resistant cases of KD with infliximab, a chimeric monoclonal antibody that binds to TNF-alpha, is associated with a rapid, and apparently safe, lowering of fever and normalization of proinflammatory cytokines but not markers of cell-mediated vasculitis [36-41].

In an early, small, randomized, multicenter trial, patients who failed an initial dose of IVIG were randomly selected to receive either an IV dose of infliximab (5 mg/kg) or a second dose of IVIG (2 g/kg) [36]. Fever resolved within 24 hours in 11 of 12 patients treated with infliximab and 8 of 12 who received a second dose of IVIG. Two of the four patients who failed to respond to the second dose of IVIG responded with fever resolution after crossing over to infliximab, and the one patient who initially failed to respond to infliximab responded to a second IVIG infusion. Progression of echocardiographic changes in CAs was similar between the two groups. Findings were similar in a subsequent larger retrospective review [37].

Of possible rescue therapies, the most studied comparison is second-dose IVIG versus infliximab. KIDCARE was a randomized, multicenter comparative effectiveness trial comparing second-dose IVIG (2 g/kg) and infliximab (10 mg/kg) in patients with IVIG resistance with a primary outcome of fever resolution at 24 hours after initiation of study treatment with no fever recurrence [42]. The infliximab treated group met the primary outcome more often (OR 0.31 [0.13, 0.73], p = 0.0076). Furthermore, the infliximab treated group had fewer days of fever, shorter hospitalizations, and fewer adverse events. However, there were no differences in CA outcomes or resolution of inflammatory marker elevation. These findings were consistent with previously published smaller trials and observational studies [36,37,43-45]. As with glucocorticoids, a Korean study found that benefits of infliximab are more pronounced when the TNF inhibitor is administered earlier in the patient's course [46].

The potential adverse events associated with long-term use of TNF inhibitors, including infection and malignancy, are reviewed in greater detail separately. Few serious adverse effects have been reported in association with use of infliximab in KD. Presumably this is because KD is a self-limited condition, so patients do not need prolonged, repeat dosing. Unlike long-term use in chronic diseases, children have brief exposure to the medications and therefore have less profound suppression of cytokines, lymphocytes, and neutrophils. Single-dose infliximab has not been associated with adverse reactions to live vaccines that were given prior to treatment [43,47,48]. (See "Tumor necrosis factor-alpha inhibitors: An overview of adverse effects" and "Polyarticular juvenile idiopathic arthritis: Treatment and prognosis", section on 'Tumor necrosis factor inhibitors'.)

Intravenous immune globulin — The first agent examined for therapy of children with persistent fevers after receiving IVIG was retreatment with additional IVIG. At the time, IVIG was the only agent shown to be effective for treatment of KD, and it was known to have a positive dose-response [49]. A study of >1600 patients enrolled in the International Kawasaki Disease Registry (IKDR) revealed that approximately one-third of patients with CA aneurysms received a second dose of IVIG [50]. Nonetheless, cumulative doses of IVIG above 4 g/kg typically do not demonstrate additional benefit in KD [1,49,51]. Thus, IVIG retreatment of persistently febrile children is usually limited to one additional infusion of IVIG (up to 2 g/kg) for a maximum total cumulative IVIG dose of 4 g/kg (algorithm 1).

IVIG retreatment also carries a dose-related risk of causing hemolytic anemia, with up to 15 percent of children receiving 4 g/kg requiring transfusions [52]. The hemolysis is due to antibodies against major blood types in IVIG. Children with type AB blood are at highest risk, while those with type O do not develop hemolysis.

Options for higher-risk patients — For higher-risk patients with refractory KD (ie, patients with giant CA aneurysms [Z-score ≥10] or smaller CA aneurysms that are increasing in size), we suggest combination therapy rather than single-agent rescue therapy. Combination therapy typically consists of a second dose of IVIG plus a 15-day glucocorticoid taper or infliximab. In addition, more potent immunosuppressive agents (eg, cyclosporine or cyclophosphamide) are used in some cases.

Combination therapy — Limited data suggest that combination therapy (use of two of the following: IVIG, glucocorticoids, infliximab) may have a greater impact on CA outcomes than monotherapy, although this benefit diminishes as time from onset of illness to start of salvage therapy increases.

One multicenter, retrospective study showed improved outcomes in small numbers of children receiving either glucocorticoids plus IVIG or infliximab plus IVIG compared with those receiving IVIG alone [23].

A retrospective study of 359 consecutive patients in Japan between 2000 and 2011 compared outcomes in children with IVIG resistance treated with 2 mg/kg of prednisolone alone, tapered over at least 15 days; a second 1 or 2 g/kg dose of IVIG; or both tapering prednisolone and a second dose of IVIG. CA abnormalities up to one month after treatment were evident in 28.7 percent of children who received a second IVIG dose, in 30.6 percent of those who were treated with prednisolone alone, and in 15.9 percent who received both IVIG plus prednisolone [31]. Patients in the groups differed significantly in mean age and date of treatment, factors known to affect the risk of developing CA abnormalities. Unfortunately, the results are not definitive due to the nature of retrospective observational studies. Additionally, there was no concurrent placebo-control group, so it is not possible to know whether IVIG or prednisolone alone had any beneficial effect.

As noted above, a systematic review and meta-analysis of 16 studies involving 2746 patients treated with IVIG plus glucocorticoids (initial or rescue therapy) versus IVIG alone found that the efficacy of glucocorticoids in protecting CAs was inversely related to the duration of illness before glucocorticoids were administered [28].

Cyclosporine — Among genetic variants associated with susceptibility to KD are the inositol-trisphosphate 3-kinase C (ITPKC) and caspase 3 (CASP3) genes, which are thought to increase signaling through the calcium nuclear factor of activated T cells (NFAT) pathway. This led to interest in possible therapeutic effects of calcineurin inhibitors such as cyclosporine, which target this pathway [24]. One study in Japan showed that 22 of 28 patients resistant to two doses of IVIG defervesced within five days of treatment with cyclosporine [53]. In an open-label, blinded-endpoints trial of initial therapy, 174 patients predicted to be at high risk of having IVIG resistance (Kobayashi score >5) were randomly assigned to 5 mg/kg cyclosporine orally divided twice daily for five days in addition to routine IVIG 2 g/kg and aspirin 30 mg/kg/day or routine therapy without cyclosporine [11]. A lower rate in the primary endpoint of incidence of CA abnormalities, per the Japanese Ministry of Health, from treatment day 3 to week 12 was seen in the group treated with cyclosporine (RR 0.46, 95% CI 0.25-0.86), but this is not a validated outcome measure. Secondary validated endpoints did not confirm a CA benefit in the patients treated with cyclosporine, with no change in mean Z-scores in each CA segment and no difference in CA abnormalities between the two groups at weeks 1, 4, and 12.

Cyclophosphamide — At least one group with considerable experience in the diagnosis and treatment of KD has reported treating severe refractory cases with cyclophosphamide [54]. In view of its efficacy in other pediatric vasculitides of medium-sized arteries, we have treated >10 patients refractory to multiple courses of IVIG, glucocorticoids as well as other agents with 10 mg/kg cyclophosphamide [55]. In seven of the patients, fever and coronary dilatation ceased to continue after one dose of cyclophosphamide, and the other three improved after a second cyclophosphamide dose one week later. In all cases, improvement occurred shortly after the dose of cyclophosphamide, though it is not possible to know if these patients might have resolution of signs and symptoms without addition of cyclophosphamide.

Subsequent rescue therapy

Rationale — Children who continue to have evidence of active vasculitis despite initial rescue therapy are at particularly high risk of developing severe CA changes. The degree of CA dilatation corresponds with morbidity and mortality in KD.

Choice of therapy — The general practice is to add agents thought to be more potent in children with larger aneurysms. Thus, after rescue treatment with IVIG plus glucocorticoids, a child with a moderate-sized CA aneurysm (Z-score 3 to <10) may be given infliximab or cyclosporine. More severely affected children with giant CA aneurysms (Z-score ≥10) may receive cyclophosphamide in an attempt to rapidly arrest the vasculitis.

Neither the efficacy of such rescue therapies nor the optimal sequence of their use have been compared in controlled trials, and only small numbers of case reports are in the literature. In addition, some of the treatments have the potential to cause more serious toxicity compared with initial rescue therapy agents. Thus, these therapies should be given under the supervision of a clinician with expertise in treating children with KD and managing immunosuppressed patients.

●Cyclosporine/cyclophosphamide – Therapy with cyclosporine or cyclophosphamide is a commonly chosen option for a second course of salvage treatment. These agents are discussed in greater detail above. (See 'Cyclosporine' above and 'Cyclophosphamide' above.)

●Plasmapheresis – Plasmapheresis is generally most effective in acute inflammatory conditions, whereas it seems to lack long-term benefit in chronic diseases. Thus, KD appears to be an ideal disorder for this intervention. On the other hand, plasmapheresis is a technically complex and medically hazardous intervention, restricting its utility in a self-limited condition such as KD [56]. In view of the risks associated with plasmapheresis, as well as the potential efficacy of alternative salvage regimens, we rarely if ever utilize plasmapheresis but would consider it in children who have active vasculitis despite multiple doses of IVIG, glucocorticoids, other agents, and cyclophosphamide. (See "Therapeutic apheresis (plasma exchange or cytapheresis): Indications and technology".)

Despite its limitations, rare case reports of apparent dramatic responses of KD to plasmapheresis appeared in the literature during the time period in which glucocorticoids were viewed as contraindicated [57,58]. In a nonrandomized study of 75 children with KD who were unresponsive to two doses of IVIG, 27 children underwent plasmapheresis, and 48 received an additional dose of IVIG [59]. A lower percentage of children in the plasmapheresis group developed CA lesions (11 versus 48 percent in the IVIG group). In another such study, six patients unresponsive to two doses of IVIG and one dose of infliximab were treated with plasmapheresis [38]. All six patients defervesced, had resolution of clinical symptoms, and improvement in laboratory abnormalities. CA outcomes, however, were not reported.

●Interleukin 1 (IL-1) inhibition – KD and systemic juvenile idiopathic arthritis (sJIA) share several phenotypic and epidemiologic characteristics, as well as evidence of elevated levels of IL-1 [60]. In addition, IL-1-induced inflammation has been shown to play a role in acute myocardial infarction and in the development of CA vasculitis in the Lactobacillus casei cell-wall extract mouse model of KD [61]. This has led to several trials using anakinra, a short-acting competitive inhibitor of IL-1 binding to the IL-1 type 1 receptor (IL-1R1), or canakinumab, a monoclonal antibody against IL-1-beta [62]. Kone-Paut assessed 16 children with IVIG resistance treated with escalating doses of anakinra. The primary outcome was abatement of fever, which was achieved in 75 percent of the intention-to-treat group. It was well tolerated. The size of the cohort precluded robust analysis of the effect on CA outcomes [63].

●IL-6 inhibition – Tocilizumab, a novel anti-IL-6 monoclonal antibody, has been used in several cases of IVIG-nonresponsive KD, although its effectiveness has been inconsistent [64]. In one case series, two of four patients with refractory KD who were treated with tocilizumab developed giant CA aneurysms despite rapid improvement in fever, other clinical symptoms, and laboratory measures [65]. This finding raises the concern that IL-6 blockers such as tocilizumab may disrupt the process of reparative inflammation, thereby increasing the risk of CA aneurysms.

●Other tumor necrosis factor (TNF) inhibition – Etanercept is a recombinant protein that binds TNF alpha and blocks its interaction with TNF receptors. In a randomized, controlled trial of adjunctive therapy for high-risk patients, children over one year of age had less IVIG resistance if they received etanercept immediately after the initial dose of IVIG than those who did not receive etanercept [66]. There is one case report of its use in addition to methotrexate in a refractory case of KD [67]. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Tumor necrosis factor inhibition'.)

●Miscellaneous – Rituximab, a B cell-depleting monoclonal anti-CD20 antibody, was administered in a single case involving a child with KD refractory to IVIG and glucocorticoids [68]. The child had rapid resolution of clinical signs and symptoms of KD, including CA findings and laboratory abnormalities.

Many other agents, including pentoxifylline [69], antioxidants [70], and neutrophil-elastase inhibitors [71], have been studied for the treatment of refractory KD, but their efficacy remains unclear.

FOLLOW-UP — 

The complications associated with KD, prognosis, and recommended follow-up (including echocardiography to evaluate for coronary artery [CA] involvement) are reviewed in detail separately. (See "Kawasaki disease: Initial treatment and prognosis", section on 'Complications' and "Kawasaki disease: Initial treatment and prognosis", section on 'Prognosis' and "Kawasaki disease: Initial treatment and prognosis", section on 'Follow-up'.)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Kawasaki disease".)

SUMMARY AND RECOMMENDATIONS

●Definition of and risk factors for refractory Kawasaki disease (KD) – Some patients with KD have persistent or recurrent fever despite treatment with intravenous immune globulin (IVIG) and aspirin with or without glucocorticoids. Refractory KD refers to patients with any of the following (see 'Refractory disease' above):

•Persistent or recurrent fever 36 hours after completion of initial intravenous immune globulin (IVIG) therapy up to two weeks after the start of treatment

•Other signs of failed initial therapy such as progressive coronary artery (CA) dilation or persistent hypotension in the setting of Kawasaki disease shock syndrome

Risk factors for IVIG resistance include male sex, age under six months, elevated C-reactive protein (CRP), CA dilatation at presentation, and low sodium or platelet levels. Of these, the most predictive are young age and baseline CA dilation or aneurysm. In patients with prolonged or recurrent fever, other causes of fever should be considered. (See 'Refractory disease' above.)

●Treatment for refractory KD – Patients with refractory KD require additional therapy because they are at increased risk of developing CA aneurysms and CA sequelae. The ultimate goal of rescue therapy is to prevent or minimize CA complications, but no rescue agent unequivocally has been shown to achieve this goal in all children who have failed to respond to initial IVIG. Choice of therapy depends upon degree of CA involvement, prior treatments used, and the clinician's preference. Our general approach is as follows (algorithm 1) (see 'Initial rescue therapy' above):

•No prior glucocorticoid therapy – For most patients with refractory KD who have not received prior glucocorticoid therapy, we suggest glucocorticoid treatment rather than other rescue therapies (Grade 2C). We typically treat with a 15-day tapering course of prednisone or prednisolone. Reasonable alternatives are a single dose of infliximab (5 mg/kg), a tumor necrosis factor (TNF) alpha inhibitor, or retreatment with a single dose of IVIG (2 g/kg). We typically do not choose a repeat dose of IVIG over glucocorticoid therapy in patients with blood type AB, A, or B, because of the risk of hemolytic anemia in this setting. (See 'Glucocorticoid therapy' above and 'Infliximab' above and 'Intravenous immune globulin' above.)

•Higher-risk patients – For higher-risk patients with giant CA aneurysms [Z-score ≥10] or smaller CA aneurysms that are increasing in size, we suggest combination therapy rather than single-agent rescue therapy (Grade 2C). Combination therapy typically consists of a second dose of IVIG plus a 15-day glucocorticoid taper or infliximab. In addition, more potent immunosuppressive agents (eg, cyclosporine or cyclophosphamide) are used in some cases. (See 'Initial rescue therapy' above.)

•Patients who failed initial rescue therapy – For patients who fail the first course of rescue therapy, subsequent treatment options include other immunosuppressive agents (eg, cyclophosphamide or cyclosporine). The choice among these depends upon degree of CA involvement and prior treatments used. There are limited data concerning the risks and benefits of these agents for refractory KD and only a few, small trials with which to evaluate them. (See 'Subsequent rescue therapy' above.)

ACKNOWLEDGMENT — 

The UpToDate editorial staff acknowledges Robert Sundel, MD, who contributed to earlier versions of this topic review.