The global impact of influenza on morbidity and mortality.

In ̄uenza in zones with a temperate climate is characterized by the occurrence of one annual epidemic during the winter months. In the northern hemisphere, the in ̄uenza season falls in November through April, while in the southern hemisphere in ̄uenza occurs during May through September [1]. In contrast, in countries with a tropical climate, the timing of periods of in ̄uenza epidemics is less distinct, and signi®cant amounts of in ̄uenza virus are isolated throughout the year. In several tropical countries, a biannual pattern has been reported, with epidemics occurring both in spring and autumn, in between the seasons of in ̄uenza epidemics in temperate zones (Fig. 1). The virological basis for recurrent epidemics is continual antigenic drift among circulating in ̄uenza viruses [1,2]. Worldwide pandemics of in ̄uenza occur infrequently, in association with the unpredictable emergence of new in ̄uenza A virus subtypes [1±4]. Pandemics have occurred three times this century (Fig. 1): The 1918 `Spanish ̄u' A(H1N1) pandemic was particularly severe, causing about 20 million deaths worldwide, while the more recent pandemics, A(H2N2) `Asian ̄u' in 1957 and A(H3N2) `Hong Kong ̄u' in 1968, were associated with moderately increased mortality [1]. Since 1968, strains of in ̄uenza A(H1N1), A(H3N2) and B viruses have co-circulated and caused discrete or overlapping epidemics each season. The recent outbreak of 18 human cases of avian A(H5N1) in ̄uenza in Hong Kong and the high fatality were a reminder that another pandemic with heavy global morbidity and mortality is a very real possibility. Indeed, as the pattern of recurrence of pandemics since the mid-eighteenth century indicates that pandemics occur every 30 years or so [1], it may be only a few years before the next pandemic occurs. How the world community chooses to respond to the annual recurrence of in ̄uenza epidemics and to the infrequent threat of pandemics depends in large part on the perceived disease burden. The burden of in ̄uenza has not been well established in tropical countries, in contrast to temperate countries. In some tropical countries, no signi®cant disease burden appears to be associated with in ̄uenza epidemics, but only with pandemics. Further, as many tropical countries are also in the process of developing, in ̄uenza epidemics may be considered of little signi®cance in relation to the severe burden of other infectious diseases and the economic issues of survival. Nevertheless, as witnessed by several of the contributions to these proceedings, many Asian countries are in fact conducting surveillance of in ̄uenza viruses and are documenting the annual activity, thus assisting the world community in the early identi®cation of new viruses with pandemic potential. The challenge lies in documenting and quantifying severe morbidity and mortality due to in ̄uenza in these countriesÐwithin the context of the pattern of year-round in ̄uenza activity and biannual epidemics in tropical countries.