Climate factors influence seasonal influenza activity in Bangkok, Thailand

activity are observed [9, 10]. Inter-tropical belt region shows highly heterogeneous timing of influenza epidemics in which some countries have an annual peak coinciding with rainy sea- son, biannual peaks of influenza activity, or no distinct seasonality at all. Previous studies have examined the meteorological factors conducive to influenza transmis- sion in different parts of the world. In temperate regions, a study in the U.K. suggests that both influenza A and B viruses preferred low temperatures and dew points [11], while a study in Canada observed that increased influenza A virus detection was associated with increased rela- tive humidity and low temperature [12]. Results from studies in tropical regions, however, are inconsistent depending on the location and the study period. In the tropical Republic of Coˆte d’Ivoire, rainfall was a good predictor of influenza activity [13], while in Uganda, influenza A (H1N1) but not influenza A(H3N2) and B activity was associated with lower precipitation [14]. In Central America, the overall influenza activity was consistently associated with increased specific humidity, but temperature and the amount of rainfall were associated with influenza activity in only certain locations [15]. Influenza pattern in tropical Southeast Asia has been less well-studied compared to that in industrialized countries. Although influenza outbreaks fluctuate seasonally, major influenza activity often coincides with the rainy season in Cambodia, Laos PDR, Myanmar and the Phil-

ippines [16–19]. Thailand experiences primary influenza activity during the rainy season, with an additional secondary peak of less magnitude at the beginning of the year [17, 20]. Mean- while, some studies in Indonesia, Singapore, and Vietnam reported no consistent pattern of influenza peak [17, 21, 22]. Circulating influenza virus strains in Southeast Asia have mostly been A(H1N1)pdm09, A(H3N2), and influenza B virus [23]. Yet, few studies have examined in-depth the seasonality of influenza with local climates. It was reported that increasing tem- perature and humidity has been linked to increased influenza activity in the Philippines and Vietnam [2, 24]. However, neither temperature nor humidity was positively associated with influenza A and B virus in Singapore [25]. In Thailand, data from suspected influenza cases suggest that temperature and relative humid- ity, but not rainfall, correlated with the overall influenza activity in central Thailand [26]. In that study, information regarding individual influenza virus (sub)types were not available, and the prediction models for influenza activity differed depending on geographical regions. It was possi- ble that the diversity of regional climates, such as in the higher elevations of the north, the sea- bound southern peninsula, and the low elevation of the central plain, may also complicate the generalization of climate influence on influenza activity within the country. Bangkok is situated in central Thailand adjacent to the Gulf of Thailand where the prevailing hot and humid weather

is disrupted by the cooler and drier air from the north a few weeks each year (typically between December and February) [27]. Since fluctuating climate factors may influence the seasonality of influenza activity in and around metropolitan Bangkok, we examined the multi-year incidence of influenza virus infection and how different influenza virus (sub)types were associated with the monthly meteorological factors of temperature, relative humidity, and rainfall. Using time series analysis, we examined potential forecasting models for influenza activity. Improved understand- ing of the seasonal pattern of influenza virus infection for a large urban area may guide public health policies on the optimal timing of influenza vaccination in the country. Materials and methods Ethical approval The study was approved by the Institutional Review Board of the Faculty of Medicine of Chu- lalongkorn University (IRB No. 127/61). Consent was not obtained because the data were ana- lyzed anonymously. PLOS ONE Climate factors influence seasonal influenza activity in Bangkok, Thailand PLOS ONE | https://doi.org/10.1371/journal.pone.0239729 September 29, 2020 2 / 15 2558) scholarship. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist.