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


Data Source

VAERS is a national postmarketing spontaneous reporting system for vaccine adverse events, jointly operated by the FDA and Centers for Disease Control and Prevention (CDC). The system was established in November 1990 and receives reports from health professionals, manufacturers, patients, parents, and others. VAERS is subject to limitations usually associated with passive surveillance systems, including underreporting, unconfirmed diagnoses, inadequate data regarding numbers of doses administered, and lack of direct and unbiased comparison groups.[18–20] In contrast, one of its major strengths is that it can identify rare adverse events after immunization.

Case Definition and Identification

Signs, symptoms, and medical conditions reported to VAERS were coded using the Medical Dictionary for Regulatory Activities, an internationally standardized coding terminology.[21] To identify potential KD cases, 2 search procedures were performed on October 15, 2007: (1) Medical Dictionary for Regulatory Activities Preferred Term search for Kawasaki disease, and (2) all-text string search for kaw. Records obtained from the 2 search procedures were compared and merged. The KD case definition used in the VAERS review was adapted from CDC[1] and the American Academy of Pediatrics' Red Book's[2] criteria and included 4 categories: classic, atypical, possible, and noncase (Table, Supplemental Digital Content 1, https://links.lww.com/A1384). Reports with insufficient data for classification were further investigated, and additional information was requested. To improve the accuracy of case classification, all KD reports and accompanying medical records were reviewed by 3 medical epidemiologists and classified according to the case definitions described in Table (Supplemental Digital Content 1, https://links.lww.com/A1384). Final classification for all cases was reached by consensus.


Classic, atypical, and possible cases were included in analyses. Descriptive analysis was used to characterize the age, gender, time from vaccination to adverse event onset, and vaccine type of KD cases. The proportional reporting ratio (PRR) is a method commonly used in pharmacovigilance to detect potential association between a certain adverse event and a drug or vaccine.[22,23] To evaluate the KD signal from the RotaTeq clinical trials, PPR was used to compare the proportion of reports for a given vaccine that were KD with the proportion of reports for all other vaccines that were KD. Because KD is mainly a childhood disease and because the clinical trial data suggested a 30-day postvaccination risk window, PRR analysis was restricted to cases with age <5 years and onset ≤30 days postvaccination, and an additional analysis was restricted to cases with age <18 years and onset ≤30 days. Conventional criteria for possible association, including number of KD cases ≥3, PRR ≥2, and χ2 ≥4, were used to select vaccines for further evaluation.[22] χ2 tests were performed using the Mantel-Haenszel method, adjusting for age (<3 months, 3–5 months, 6–59 months, 5–17 years, and unknown) and gender.

Reporting rates were calculated as follows. The number of KD cases within 30 days after vaccination was divided by the estimated person time of observation (ie, number of doses distributed multiplied by 30 days of observation). Because data on the actual number of doses administered were not available, the number of net doses distributed was used as an estimate. Dose distribution data were obtained from the CDC's Biologics Surveillance System (CDC unpublished data) and from the manufacturer.


Business Objects XI R2 (Business Objects Americas, San Jose, CA) software were used for data management. All data analyses were performed using SAS 9.0 (SAS Institute, Cary, NC).


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.