Kawasaki syndrome

A syndrome of unknown cause, characterised by persistent fever, conjunctivitis and mucosal changes eg strawberry tongue in a young child who has often been treated empirically with antibiotics without improvement, and is invariably miserable. Potentially complicated by coronary aneurysms, which may be fatal. Kawasaki’s is the leading cause of acquired heart disease globally after rheumatic fever, and is the leading cause in the North.

It is a vasculitis affecting medium sized arteries, and other arterial vessels down to capillary size. Was thought to be a superantigen disorder, ie with no specific infectious agent, polyclonal B activation – coronary aneurysms have been seen in other superantigen diseases eg toxic shock, have also been reported in meningococcal septicaemia. But now thought to be conventional, but unknown, infectious trigger.

Similarities with COVID PIMS-TS.

Genetics important, risk higher in East Asian immigrants, plus family history, single nucelotide polymorphisms found in 6 genes including FcγR2a, caspase 3 (CASP3), and human leukocyte antigen class II.

Clustering has been shown (Knox test significant within the space-time interval of 3 km and 3-5 days) which suggests an infectious trigger [PIDJ 27(11):981-985, Nov 2008]. TNF alpha seems important, in an animal model, knock out mice or anti TNF treatment prevents aneurysm development.

The underlying pathology is a vasculitis, and although coronary disease is the best recognized there is increasing evidence that other medium sized arteries are affected with descriptions in the literature of peripheral gangrene and cerebral infarction. Proposed model includes necrotizing vasculitis, plus chronic/subacute vasculitis, 2 weeks after onset and sometimes lasting months, with a unique type of luminal myofibroblastic proliferation. 


BPSU Kawasaki – 553 cases were notified: 389 had complete KD, 46 had atypical KD and 116 had incomplete KD!  Median time to IVIG in those with Coronary artery aneurysms (CAAs) was 10 days cf 7 days for those without.  Rate of CAA in under 1yr was 39%.  19% overall had CAA despite treatment.  Associated with low albumin, and incomplete.  Only 1.6% developed giant aneurysms, which have the worst prognosis, of course.

Risk of CAA estimated at 20-40% if untreated.  [Archives PMID 30104394]

Incidence in west seems to have plateaued, after decades of increase (?ascertainment bias). 10-20x higher rate in NE Asia (Japan, Taiwan, Korea), plus continues to increase! 1% of all Japanese kids will have had KD by age 10. But unrecognized before 1950, whereas pathological specimens in UK from over a 100yrs ago show same process.

Less complications in Asia!  Better diagnosis? Arguably all European cases could be considered high risk…

Japan has had 3 epidemics, the only country to have had them, but none since 1986.

Data from China and India indicate increase, with only a few case reports prior to 1990. Rate in Chandigarh equivalent to UK now, presumably underestimate, but already outnumbers rheumatic fever cases.  Also high rate in Kerala, but related to affluence or health care availability?    Rate in Shanghai approaching NE Asia rate. Rate in Hong Kong has tripled in 20 years. Will become the predominant cause of acquired cardiac disease?

Association between seasonal wind patterns in Pacific and KD rates in Japan, California and Hawaii! May relate to infectious agents.

[Singh S, et al. Arch Dis Child 2015;100:1084–1088. doi:10.1136/archdischild-2014-307536]

More common in males, peak age 18-24 months.Kawasaki flowchart

Case definition

(American Heart Association):

  • Fever of 5 days duration or more
  • plus 4 of the following (ie only drop 1):
    1. Conjunctivitis: Bilateral, bulbar, non-suppurative
    2. Lymphadenopathy: Cervical, >1.5 cm
    3. Rash: Polymorphous, no vesicles or crusts
    4. Changes of lips or oral mucosa: Red cracked lips; “strawberry” tongue; or diffuse erythema of oropharynx
    5. Changes of extremities:
      • Initial stage, erythema and oedema of palms and soles
      • Convalescent stage (about two weeks) periungual desquamation of fingers and toes

Images from Circulation journal 2017

In Japan, prolonged fever is one of the optional features – 1/3 of Japanese children get IVIG before day 4! McCrindle guideline (AHA, 2017) suggests complete Kawasaki disease may be diagnosed with 4 days (possibly even 3 days) fever if one of the features is peripheral erythema/swelling.

Many people would also diagnose if only 3 of these features plus coronary artery aneurysms detected. Lymphadenopathy is the least common feature (?easy to miss) – esp uncommon in younger children. Perineal desquamation is also quite characteristic.

There may also be abdominal pain, diarrhoea, arthralgia/arthritis, aseptic meningitis, and pneumonitis (with or without pulmonary nodules). There may be a murmur (mitral incompetence), and myocarditis can occur but it is rarely severe. Renal involvement, encephalopathy have been described, macrophage activation syndrome has been reported rarely.

Pitfalls –

  • infants, adolescents (“glandular fever”).
  • Signs change over time – rash fades, desquamates, then nail changes but similarly lymph nodes, mouth changes.  So at any one time only a few features may be evident – especially beyond first week.
  • Normal or low platelets (actually a high risk feature in some scoring systems)
  • Sterile pyuria (“UTI”), “aseptic meningitis”
  • Presentations with myocarditis, surgical abdomen (GI vasculitis)
  • Single dose of IVIG without response does not mean diagnosis is wrong, in fact means more aggressive treatment required.
  • Beware positive cultures putting you off diagnosis! Might be triggering infection! Even Strep, adenovirus

Tips –                           

  • BCG reactivation as a clue
  • Arthritis as a clue
  • Axillary/inguinal aneurysms on examination? [Janet G-M]
  • Repeat echo early, especially where diagnosis still not clear cut. Cardiology need chasing!?
  • IVIG raises ESR! Don’t interpret as treatment failure!
  • Incomplete/uncertain have highest risk of complications! So consider if fever >=5/7 with 2 criteria, or fever >= 7/7 without necessarily any criteria! CRP>30 and ESR>40 should trigger further assessment eg echo and use of additional lab criteria eg high platelets, high ALT, high WCC, low albumin

Many of the clinical features of the disease are outbreak dependent with a different spectrum of clinical findings in one mini-outbreak compared with another, and with cases having similar clinical phenotypes clustering temporally.

So consider echo in any young child with persistent fever.

The term atypical or incomplete Kawasaki disease is used for cases without the required number of features. Whether this is the same disease or not is unclear; there will undoubtedly be some cases that overlap with other systemic vasculitides eg polyarteritis nodosa (PAN). In PAN, mucocutaneous changes are uncommon, whereas renal disease is common. Gall bladder hydrops appears to be unique to Kawasaki’s. On the other hand, having a rigid case definition is perhaps unhelpful since incomplete cases are often seen, particularly in infants, and are associated with a delay in diagnosis and worse prognosis. Under the age of 3 months, the majority of cases with coronary aneurysms have atypical presentations.

McCrindle AHA guideline suggest incomplete KD should be diagnosed where:

Differential diagnosis

  • Scarlet fever
  • Toxic shock syndrome – some cases of “Kawasaki shock syndrome” have been described
  • Scalded skin syndrome
  • Measles
  • Drug reaction
  • Systemic onset juvenile idiopathic arthritis
  • Steven-Johnson syndrome
  • Rickettsial infection eg Rocky Mountain spotted fever
  • Leptospirosis

Investigations mainly help exclude alternative diagnoses. Typically with the disease itself they simply indicate systemic inflammation and can be useful for monitoring response to treatment. Hence there are usually elevated white cells, platelets, ESR and CRP. Liver function tests are often mildly deranged. Low sodium and albumin suggest vascular leak. The ECG may have PR interval changes, and ST segment or T wave changes. Echocardiography may reveal dilated, ectatic coronary arteries or frank aneurysms.


Eleftheriou (2013) guidelines differ in highlighting high risk patients eg failed IVIG, infants, Kobayashi score >5) and recommending methylprednisolone in addition to IVIG. Infliximab is included as an option for resistant or recrudescent disease. Of course, diagnosis should be reconsidered if response to treatment is poor.

Kobayashi criteria used to predict IVIG failure (5+ points), but more sensitive in Japanese populations – just 33% in Non-Japanese, with 87% specificity:

  • Na<=133 (2 points)
  • ALT>=100 (1 point)
  • Plts <300 (1 point)
  • CRP>10 (1 point)
  • >80% neutrophils (2 points)
  • Plus age<12 months (2 points) and fever <=4 days (1 point)

Standard treatment is with intravenous immunoglobulin (IVIG) 2g/kg over 12 hours, ideally within the first 7-10 days of the illness, and with aspirin 30-50 mg/kg/day (in 4 divided doses) orally during the acute phase. This combination reduces the risk of aneurysm formation from 25% to 9%. Most will respond to a single dose, but about 20% will require a second dose. Of these, only a half will then defervesce.

IVIG side effects – fever, headache, joint pain, aseptic meningitis, BBV, allergic reaction (?), raised ESR. Remember to defer immunisations for 3 months.

Once defervescence has occurred, the aspirin dose can be reduced to an anti-platelet dose of 2-5mg/kg/day. Aspirin is stopped after 6 weeks unless aneurysms found.

Duration of fever is the most powerful predictor of poor coronary outcome (one additional day of fever increasing the odds of aneurysm development by 3-5x). Delayed diagnosis is usually a reflection of slow evolution of criteria rather than atypical presentation – in that study, diagnosis after 10 days had a 2.8x higher risk of aneurysms (although they also had higher platelet counts). [Pediatrics. 115(4):e428-33, 2005. PMID 15805345]

Methylprednisolone treatment is 0.8mg/kg BD IV for 5-7 day or until CRP normal, followed by Prednisolone 2mg/kg weaning over 2-3 weeks. Other regimes are Methylpred 10-30mg/kg IV OD for 3 days followed by prednisolone.

In Europe steroids reserved for worse cases – but by epidemiology, all European cases could be considered high risk?? 39% of under 1yrs had coronary aneurysms (BPSU study). Europe Kawasaki trial in progress (KDCAAP) – Adding immediate corticosteroid treatment to standard of care IVIG and aspirin.


Several scorings systems have been developed to predict IVIG resistance and poor outcome. Kobayashi score relates to low sodium, low platelets (<300), short duration of illness (<4 days), high ALT/CRP, high percentage neutrophils (>80%) and age under 1yr. It works well in Japan, but when used in a US study had low sensitivity (but good specificity).


ECG and echo should be done as soon as possible but should not delay treatment. If first echo is normal and CRP normal after 1 week, repeat scans recommended at 2 and 6 weeks. But chase cardiology to repeat early if diagnosis unclear.

Those with Z score more than 10 (“large”) have 25% risk of coronary event within 10yrs (girls), 50% (boys)!

Most aneurysms will resolve over time, unless they are giant (>8mm). Serial echocardiography should be done to monitor resolution. Evidence of subacute chronic vasculitis for months (post-mortem cases) so move now to infliximab treatment etc after initial immunosuppression.

Warfarin should be considered for giant aneurysms, with initial heparinization to prevent paradoxical thrombosis, although its potential for complications in young children is significant. Stress testing and angiography may be appropriate. Aspirin can be discontinued if aneurysms resolve, but it is likely that the atherosclerosis risk remains high and life long follow up to address other risk factors is sensible.

Mortality in the UK has been as high as 3.7%, but is much lower in Japan.

[2020 ArchDisChild Ed and Practice Kelly]

[2017 AHA guidelines]

[Eleftheriou Arch Dis Child 2014;99:74–83, J Paed and Child Health 49 (2013) 614–623]