Rotavirus activity and meteorological variations in an Asian subtropical city, Hong Kong, 1995-2009.

Rotavirus is a leading cause of severe infectious diarrhea in infants and young children aged <5 years. Rotavirus infections have minimal to strong seasonality depending on geographical locations. In this study, a comprehensive retrospective analysis was performed to evaluate the association between rotavirus admission and multiple key meteorological variables, including air temperature, relative humidity, atmospheric pressure, and solar radiation over a 15-year period from 1995 to 2009 in Hong Kong. Rotavirus infections were found to show a distinct cyclical pattern with an annual peak in cold season. The weekly number of cases showed the strongest correlation with average air temperature of the previous 7 days (rho=-0.69; P<0.0001), followed by atmospheric pressure (rho=+0.67; P<0.0001); whereas only weak correlation with relative humidity (rho=-0.252; P<0.0001) and solar radiation (rho=-0.312; P<0.0001) was observed. Curve fitting regression analysis suggested that the correlation was nonlinear in nature in which the effect was more profound towards lower air temperature and higher atmospheric pressure conditions. In binary logistic regression analysis, a final model that included air temperature (≤ 20°C) and atmospheric pressure (≥ 1,013 hPa) predicted correctly 85.3% and 82.6% of weeks with rotavirus activity above and below the baseline level, respectively. In multivariate Poisson model, air temperature and solar radiation were independent factors associated with the weekly number of rotavirus cases, adjusted for seasonal variation. In summary, the current study provides evidence suggesting that local seasonal activity of rotavirus correlated strongly with air temperature, followed by atmospheric pressure but only minimally with relative humidity in pre-vaccine era.