Oscillatory fluctuations in the incidence of rotavirus infections by serotypes 1, 2, 3, and 4.

The statistical evidence for regularity in the epidemic cycles of rotavirus infection for serotypes 1, 2, 3, and 4 was examined. Hospitalization longitudinal data of the monthly incidence of rotavirus infections from the city of Melbourne, Australia during 1977-1993 were used. Periodograms were used for exploring seasonal and longer-term cycles (interepidemic periods) of rotavirus infection. There was a satisfactory agreement between the interepidemic period estimated by means of periodograms with the one predicted by theoretical epidemiological studies. Thus, there is a clear evidence of a biennial peak in the epidemiology of rotavirus. Results of the study show an evidence of the likely existence of an interepidemic cycle of 4.6-5.2 years of duration. The finding of this interepidemic cycle was unexpected, and does not arise from the alternating incidence of the 4 serotypes since this peak appears in the periodogram of each serotype.