Kawasaki Disease After Vaccination: Reports to the Vaccine Adverse Event Reporting System 1990-2007

Wei Hua, MD, PhD; Hector S. Izurieta, MD, MPH; Barbara Slade, MD; Ermias D. Belay, MD; Penina Haber, MPH; Rosemary Tiernan, MD, MPH; Emily Jane Woo, MD, MPH; John Iskander, MD, MPH; M. Miles Braun, MD, MPH; Robert Ball, MD, MPH, ScM


Pediatr Infect Dis J. 2009;28(11):944-947. 

In This Article


We investigated 97 Kawasaki cases received by VAERS through October 14, 2007. Consistent with KD epidemiology in the general population, these reported confirmed and possible KD cases were predominantly children <5 years and more frequently males. We had limited data to describe the disease pattern in terms of race and ethnicity, because such information is not requested by the VAERS reporting form.

In the present review, infants <6 months old and <3 months old constituted 39% and 14% of KD cases. Although the peak age incidence varies in different countries, and KD usually occurs in children younger than age 5 years, the disease is less frequent in infants younger than age 6 months.[24–27] In a 1997 Japanese nationwide KD survey, approximately 11.2% of cases occurred in infants aged <6 months.[28] In another study involving 105,755 KD patients in Japan during 1970–1995, only 1.7% were aged <3 months.[26] There are limited data on age distribution of KD in young infants in the United States; estimates range from 3%[25] to 8%.[24] Although the higher proportion of KD cases in infants aged <6 months and <3 months in our review might suggest disproportional occurrence of KD in young infants after vaccination, this observation may simply reflect required vaccinations at ages 2 and 4 months, providing more opportunity for KD to be reported to VAERS at ages <6 months.

There was no clustering of onset intervals after day 1 postvaccination. Although the majority of KD cases after Pediarix and Prevnar had onset on days 0 to 1 postvaccination, onset intervals for RotaTeq cases spread out between 0 and 54 days postvaccination, with most occurring within 2 to 30 days. Additional analysis also suggested that there was no temporal clustering for RotaTeq for the entire time period, or for Pediarix or Prevnar after day 1 postvaccination. Clustering of onset intervals on days 0 to 1 postvaccination may be only coincidental or reflect a reporting bias. A possible additional explanation is that the Pediarix and Prevnar reports were mostly not stimulated whereas for RotaTeq they were. Stimulated reports may be more likely to be spread out in time.

The KD patients in the present review received a variety of vaccines. The disproportionality analyses showed that, compared with all other vaccines, KD constituted a higher proportion of reports for RotaTeq, Pediarix, and Prevnar for the entire study period. These findings, however, were not reproducible when the analysis was stratified by reporting time period before and after the RotaTeq label revision. While Pediarix was the only vaccine that met the conventional criteria for disproportionality for the time period before the label revision, after the label revision RotaTeq became the only vaccine for which a higher proportion of KD was reported. Pediarix combines DTaP, hepatitis B, and inactivated polio vaccines. KD and polyarteritis nodosa following hepatitis B vaccination have been reported;[15,29] however, a causal relationship has not been established.

The RotaTeq label revision for KD stimulated reporting of KD for RotaTeq, as well as for vaccines coadministered with RotaTeq, including Pediarix and Prevnar; this could explain partially, if not entirely, the inconsistent findings in disproportionality analysis for the entire time period and the time periods stratified based on whether reports were received before or after the RotaTeq label revision. It might be noteworthy that the PRR was elevated only for Pediarix compared with all other vaccines before stimulated reporting for RotaTeq occurred. After the RotaTeq label revision, a marked increase in KD reports for RotaTeq was received by VAERS, which may have masked the disproportional reporting of KD for Pediarix, because RotaTeq reports were included in the comparison group; this may explain why the PRR for Pediarix was no longer elevated, while RotaTeq became the only vaccine that met the conventional criteria for elevated PRR after the label was revised. The PRR for Prevnar was no longer elevated for either time period after the analysis was stratified. In the present review, because RotaTeq, Pediarix, and Prevnar were coadministered in most cases, it was not feasible to calculate the PRR for one vaccine while excluding the other 2 from the comparison. Continued monitoring is needed.

In this review, the RotaTeq KD reporting rate was 0.65 per 100,000 person-years before the label revision and 2.78 per 100,000 person-years after the label revision. For Pediarix, the reporting rate was 0.37 per 100,000 person-years before the label revision and 2.44 per 100,000 after the label revision. Compared with the KD background incidence rate of 9 to 19/100,000 person-years for children <5 years (as an estimate for specific age groups for RotaTeq and Prevnar) in the United States, the KD reporting rates in VAERS for either RotaTeq or Pediarix were substantially lower. The reporting rates in this review cannot be interpreted as incidence rates, because VAERS is subject to underreporting, unconfirmed diagnoses, and other limitations inherent to passive reporting systems.[18–20]

Overall, in response to a potential signal of KD suggested by the RotaTeq clinical trial data, we reviewed all KD reports to VAERS for all US licensed vaccines and evaluated the signal using the FDA guidance on safety signals.[30] Although the role of vaccine in the pathogenesis of KD is hypothetically biologically plausible,[9,11] epidemiological characteristics of the cases, disproportionality analysis, and the comparison of reporting rates to background rates did not indicate consistently that KD may be related to vaccines. An ongoing observational study will help clarify the lack of evidence for a KD safety signal in this review.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.