Evaluating the seasonality of human influenza A/H5N1 in Egypt and the effects of weather conditions

The seasonality of human influenza in tropical regions is not well understood. Although H5N1 has caused over 250 human deaths worldwide, and several studies have evaluated environmental risk factors for H5N1 in poultry, there has been no systematic study of climate or seasonality of influenza A/H5N1 in either humans or poultry. In order to better understand influenza seasonality, a secondary data analysis was conducted on human influenza A/H5N1 cases in Egypt (n=50), using cases reported to the World Health Organization through May 1, 2008. In addition, the distribution of human cases was compared to reported poultry outbreaks in these countries. Usual regression methods were determined to be inappropriate since the pattern of seasonality of H5N1 was unknown and the H5N1-climate dataset violated key assumptions of these methods. Instead, Fourier analysis was used to determine the pattern of seasonality of cases and to assess the contributions of weather variables. This method allowed direct evaluation of seasonality or periodicity, by determining the frequency that best fit the data. Human H5N1 cases occurred with a period of 1.18 years/cycle, and were linearly related to climate (κ²≥0.94). Incidence peaked when precipitation was high; temperature and absolute humidity, low; and relative humidity, intermediate — conditions outside the range typically described as optimal for influenza transmission and survival. Finally, occurrence of large poultry outbreaks was not a good predictor of human cases. The results will inform models of influenza spread in tropical areas and allow better targeting of preventive interventions for human H5N1.