• 
• 

Three Part Question

In [children with suspected Kawasaki disease] is [high dose aspirin] better than [low dose aspirin] in combination with IVIG for [reduced incidence of coronary artery aneurysm formation]?

Clinical Scenario

You are the Subspecialty Paediatric Emergency Medicine Trainee working in the Paediatric Emergency Department review clinic. The next patient is a four-year-old girl who has been brought back for review of throat swab results, as prior to being seen in the PED last week she had been on amoxicillin from the GP for four days without clinical effect. Her mother tells you she has now had an additional week of Penicillin V without improvement. She has had intermittent fever for nine days, is miserable with red eyes and a cracked, sore mouth, a transient rash, and this morning her mother noticed that her hands and feet were sore with peeling skin. The throat swab is negative. You realise that Kawasaki disease is a significant possibility in this case, and wonder whether you should start aspirin prior to urgent paediatric cardiology review – and if so, at what dose?

Search Strategy

Ovid MEDLINE® 1948 to September 09 2011
EMBASE® 1980 to present
Cochrane Database
exp Mucocutaneous Lymph Node Syndrome/ OR Kawasaki disease.mp OR Kawasaki.mp
AND
exp Aspirin/ OR aspirin.mp OR exp Salicylic Acids/ OR exp Salicylic Acid/ exp Salicylates/ OR salicyl*.mp
AND
exp Dose-Response Relationship, Drug/ OR dose.mp
LIMIT TO
Humans, English Language, “All Child” (0-18 years)

Search Outcome

The MEDLINE search produced 101 papers, of which four were relevant. The search also identified a relevant Cochrane review, which will be discussed in the comment section.
The EMBASE search produced no additional relevant papers. 

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Koren, Rose et al. 1985 Canada	54 cases of Kawasaki disease over 26 months. Group A: children diagnosed with Kawasaki were given high dose aspirin (80-180mg/kg/day in divided doses). Group B: children not diagnosed with Kawasaki disease during febrile phase were given PRN aspirin at <30mg/kg/day. Least square linear regression for serum salicylate analysis; 2 test for coronary involvement rate.	Controlled clinical trial, level IIIb	Two-dimensional echocardiography at time of diagnosis. Selective coronary angiography in those with positive echo.	Group A: 6/36 (16.7%) coronary involvement.	Poor study design; no clear protocol, non-uniform administration of aspirin in group B. Major confounder in heterogeneity between Group A and Group B; Group B received later diagnosis of Kawasaki disease, demonstrated in a later study to be independently associated with higher rates of coronary artery involvement.
			Serum salicylate concentration vs therapeutic range of 20-30mg/dL.	Few cases achieved serum concentration >20mg/dL even when dose increased to 180mg/kg/day. Levels during non-febrile phase were twice as high as those achieved with the same dose during febrile phase.
Akagi, Kato et al. 1990 Japan	64 children with Kawasaki disease diagnosed within 7 days of start of fever. Quasi-randomised (alternate allocation by age and sex) to two groups. Group 1: High dose aspirin (100mg/kg/day) Group 2: Low dose aspirin (30mg/kg/day). Administered TDS 8 hourly in both groups for 14 days then reduced to 5mg/kg/day. No treatment with gamma-globulin within study groups. 2-tailed students t-test for duration of fever and laboratory data; chi-squared test for coronary involvement rate	Quasi-randomised controlled trial; level IIb	Fever duration (days) >38degC	Duration of fever in Group 1 (high dose) significantly shorter than Group 2 (low dose). P<0.05	No evidence of true randomisation. No power calculation, ?small study numbers. No further group demographic analysis (loosely age and sex matched).
			Laboratory test abnormalities; admission, day 4,day 7 & day 14.	No significant difference in FBC, ESR or Albumin.
			Presence or absence of coronary lesions determined by two-dimensional echocardiography, on admission then twice weekly for one month and once every one-two months thereafter.	No significant difference in incidence of coronary artery lesions between the two groups; 5 patients (17%) in the high dose group, 7 patients (23%) in low dose group.
Furusho, Kamiya et al. 1991 Japan	Three-armed multicentre study; patients admitted with Kawasaki disease over three time periods were enrolled at sixteen centres within seven days of illness onset. Only Study 3 relevant to search question. (Study 1: random allocation to aspirin and IVIG versus aspirin alone. Study 2: random allocation to one of three IVIG dosing regimens plus aspirin.) Study 3: 151 patients randomised to three treatment groups: IVIG plus aspirin (49 subjects), IVIG alone (52 subjects), aspirin alone (49 subjects). 2 for comparison of rates of coronary artery involvement. Where aspirin was given, regimen was oral administration at 30-50mg/kg in three divided doses while febrile. From fever disappearing to end of “acute reaction” aspirin given as 10-30mg/kg once daily.	Multicentre randomised controlled trial, level Ib	2-dimensional echocardiography three times per week for 60 days from onset of illness. Diagnostic criteria of Kawasaki Disease Research Committee used. Selective angiography following positive echocardiography.	No significant difference in coronary artery lesion incidence between IVIG alone and IVIG plus aspirin groups before 30 days and at 30 days.	No detail regarding method of randomisation. No power calculation. No specific data given; broad summary table only. ?relevance of 30 days as endpoint? No analysis of differing aspirin dosing nor information regarding variance in length of time of higher dose administration within groups. No blinding.
				Significant difference in coronary artery lesion incidence between aspirin alone (38.8%) and IVIG alone (18.9%) before 30 days (P<0.05). No significant difference at 30 days.
				Significant difference in coronary artery lesion incidence between aspirin alone (38.8%) and aspirin plus IVIG (18.4%) before 30 days (P<0.05). No significant difference at 30 days.
Saulsbury 2002 USA	72 children over 13 years, 40 boys (56%) and 32 girls (44%). 70 patients received aspirin. Patients diagnosed based on criteria, all patients had echocardiography at the time of diagnosis and within 4-8 weeks. All patients received intravenous immunoglobulin. Two data subsets: if high dose aspirin had not been commenced, low dose aspirin was recommended. 24/70 received high dose aspirin (34%), all switched to low dose aspirin after varying interval. 46/70 patients were commenced on low dose aspirin from outset (66%). Unpaired t-test and Fisher’s exact test to determine difference in incidence of coronary artery involvement between groups.	Retrospective cohort study; level IIb	Duration of fever (oral or rectal temperature >38degC)	Duration of fever was significantly shorter in patients on high dose aspirin (P=0.04), with no evidence for duration of fever as a function of aspirin or IVIG dosing.	Small sample size, no power calculation. No randomisation or blinding. No clear treatment protocol applied. Considerable heterogeneity between high and low dose aspirin groups. Potential for investigator bias as author of paper (a Paediatric Immunologist) was consulted on the treatment of 60/72 patients.
			Presence or absence of coronary artery abnormalities.	Coronary artery abnormalities in 12/72 (17% overall).

Comment(s)

No good quality randomised controlled trials regarding the value of aspirin dose in Kawasaki disease have been published, as identified by the Cochrane review which was updated in 2009. It looked only at the Furusho paper, deeming the others lacking in sufficient quality to meet the Cochrane Collaboration’s rather more stringent standards, and gave the following implication for practice; “The existing evidence from RCTs does not permit any recommendation for or against treatment of children with Kawasaki disease with salicylate in combination with IVIG.” (Baumer, Love et al. 2009) The included studies contain serious methodological omissions and as such provide no evidence for a difference in efficacy between high and low dose aspirin. Is there any information of value within the studies discussed? No study could demonstrate a clear advantage of aspirin with the exception of the Koren paper which was not significant at the 10% level (P=0.14), and where the groups were so clinically different the comparison between them is almost completely invalid. Two studies (Saulsbury, Agaki) found that aspirin tended to reduce the length of the febrile phase of illness, though there is no evidence that this links directly to improvement in morbidity or mortality outcomes. The Furusho paper found significant reduction in the incidence of coronary artery aneurysms in both the with-aspirin and without-aspirin groups receiving intravenous immunoglobulin. The value of aspirin in Kawasaki disease is as an agent with anti-inflammatory and anti-platelet properties, assuming it will act both to reduce associated fever and coronary artery inflammation and also prevent a pro-thrombotic state which might precipitate coronary artery thrombosis. Besides the lack of evidence of efficacy, though, there are further problems with this hypothesis. A Canadian investigative paper found that in order to achieve “therapeutic” serum salicylate levels in children with Kawasaki disease, high doses were required. These high doses, however, were unpredictable, with 55% of children receiving 100-110mg/kg/day in subtherapeutic range, and 28% of children on 120mg/kg/day with salicylate levels in the toxic range (Koren and MacLeod 1984). Koren’s later clinical trial (Koren, Rose et al. 1985) further corroborates this, with a finding of variable salicylate levels despite consistent dosing within the same patient, depending on the stage of the clinical disease. Use of aspirin in childhood is associated with development of Reye syndrome (Starko, Ray et al. 1980; Waldman, Hall et al. 1982), a syndrome of encephalopathy and liver failure which carries a case fatality rate quoted at 31% (Belay, Bresee et al. 1999). Without compelling evidence for benefit of aspirin in Kawasaki disease it is difficult to justify exposure to an additional disease process. No case of Reye syndrome was described among the papers, but the risk of side effects and toxicity of salicylate is dose-dependent, which should be taken into clinical consideration (Hirsh 1985).

Clinical Bottom Line

There is insufficient evidence to suggest that high dose aspirin is any more or less effective than low-dose aspirin in preventing coronary artery aneurysm development or cardiac death associated with Kawasaki disease. It is worth considering that the Cochrane review of aspirin in Kawasaki disease found no evidence of benefit at all.

Level of Evidence

Level 3 - Small numbers of small studies or great heterogeneity or very different population.

References

1. Koren, G., V. Rose, et al. "Probable Efficacy of High-Dose Salicylates in Reducing Coronary Involvement in Kawasaki Disease." JAMA: The Journal of the American Medical Association 1985; 254(6): 767-769.
2. Akagi, T., H. Kato, et al. "A study on the optimal dose of aspirin therapy in Kawasaki disease." The Kurume Medical Journal 1990; 37(3): 203-208
3. Furusho, K., T. Kamiya, et al. "Intravenous γ-Globulin for Kawasaki Disease." Pediatrics International 1991; 33(6): 799-804.
4. Saulsbury, F. T. "Comparison of High-Dose and Low-Dose Aspirin Plus Intravenous Immunoglobulin in the Treatment of Kawasaki Syndrome." Clinical Pediatrics 2002; 41(8): 597-601.