Time-Series Analysis of Coxsackievirus B Serotype Surveillance Data in Japan

ObjectiveStudies have identified serotypes of Coxsackievirus B (CVB) enterovirus as a cause of type 1 diabetes. Studies have also identified cyclical variations in type 1 diabetes incidence—peak incidences occurring in 4- to 6-years periods in two regions in England, a 5-year period in Western Australia, and 5.33-year period in Poland. To date, no studies have investigated whether CVB infection rates demonstrate similar cyclical variation characteristics. The purpose of this study was to characterize periodicity in CVB surveillance data.ResultsMaximum entropy spectral analysis was performed on monthly CVB surveillance data. In addition to demonstrating a 1-year cycle for all serotypes, spectral peaks demonstrated dominant cycles—6.9-, 3.8-, 4.3-, 9.5-, and 7.8- year periods for CVB1, CVB2, CVB3, CVB4, and CVB5, respectively. Pearson correlation was used to compare the least-squares fit curves based on periods estimated from the analysis with the original data. The results for all five serotypes—CVB1, CVB2, CVB3, CVB4, and CVB5—demonstrated good correlation—ρ = 0.96, ρ = 0.60, ρ = 0.90, ρ = 0.88, and ρ = 0.67, respectively. This method could be a useful tool for the efficient investigation of CVB as a pathogen of type 1 diabetes.