Calculating the fraction of Kawasaki disease potentially attributable to seasonal pathogens: a time series analysis

Kawasaki disease is an acute, febrile, systemic vasculitis of children that primarily affects medium-sized blood vessels with a tropism for the coronary arteries. Although the etiological factors remain unknown, infections have been suggested as the trigger of Kawasaki disease. We sought to calculate the fraction of Kawasaki disease potentially attributable to seasonal infections.

This cohort study used a population-based time series analysis from the French hospitalisation database (Programme de Médicalisation des Systèmes d’Information), which includes all inpatients admitted to any public or private hospital in France. We included all children aged 0–17 years hospitalised for Kawasaki disease in France over 13 years. The monthly incidence of Kawasaki disease per 10,000 children over time was analysed by a quasi-Poisson regression model. The model accounted for seasonality by using harmonic terms (a pair of sines and cosines with 12-month periods). The circulation of eight common seasonal pathogens (adenovirus, influenza, metapneumovirus, Mycoplasma pneumoniae, norovirus, rhinovirus, rotavirus, respiratory syncytial virus, and Streptococcus pneumonia) over the same period was included in the model to analyse the fraction of Kawasaki disease potentially attributable to each pathogen. Infections were identified on the basis of polymerase chain reaction or rapid antigen testing in hospital laboratories.

Between Jan 1, 2007, and Dec 31, 2019, we included 10,337 children with Kawasaki disease and 442,762 children with the selected infectious diseases. In the Kawasaki disease cohort, the median age [IQR] was 2 [0–4] years, 6164 [59.6%] were boys. Adenovirus infection was potentially responsible for 24.4% [21.5–27.8] (p < 0.001) of Kawasaki diseases, Norovirus for 6.7% [1.3–11.2] (p = 0.002), and RSV 4.6% [1.2–7.8] (p = 0.022). Sensitivity analyses found similar results.

This cohort study of data from a comprehensive national hospitalisation database indicated that approximately 35% of Kawasaki diseases was potentially attributable to seasonal infections.