Thermoregulation and periodically forced SEIR model: Understanding asthma seasonality in South Florida

The impact of changing weather and air quality conditions on cardio-respiratory diseases is very important. Epidemiological as well as biometeorology studies have documented the existence of exacerbating factors for asthma, different from often cited allergens. Economic conditions, educational backgrounds, stress, diets, along with weather and air quality conditions appear to compete and complement the allergen hypothesis. Motivated by these facts and aimed at understanding the inner working of asthma, time series of Emergency Department (ED) visits due to asthma in Miami Dade and Broward Counties were provided by the Florida Asthma Coalition for six years (2005-2011) with daily temporal resolution. They show a peak of attendance between the months of November and January and a minimum between the months of April and June every year. Such seasonality represents an increase (decrease) about 38 % above (below) the mean number of visits per day during this period. Weekends appear to be the most common time for ED visits, leading us to think about the influence of the exposition time to pollutants, cultural and socio-economic impacts, as well as possible lagging effects. Besides that, the spatial distribution of cases (Zip code breakdown) revealed that the proximity to highways or heavy-loaded transportation routes increases the occurrence of cases. The time series of weather parameters (Temperature, Humidity, Wind Speed and direction, Pressure, Diurnal range of both temperature and humidity, and extreme of both, temperature and humidity) were obtained from the Weatherbug mesonet in South Florida. Surface ozone, and particulate matter were obtained from EPA registered stations. Time series and correlation analyses between weather variables and the number of cases resulted in weak to moderate associations mainly with the minimum and mean temperature, the mean humidity, and some derived thermal indexes. Lagging effects were analyzed up to seven days previous to ED visit. A General Additive Model (GAM) for Multilinear regression (MLR) was implemented based on the most sensitive lags. Thermal homeostasis seems to explain variance in the number of cases no more than a 30 %. When air quality (Ozone and particulate matter) time series is used, direct associations are less evident. It is noteworthy that large levels of allergens and surface ozone tend to occur in spring and early summer due to the release of pollens, the end of the dry season, which produces many wildfires, and the sun activated photolysis. Both, ozone and particulate matter peak after the “winter asthma season” and the location of measuring stations might play a role underestimating the amount of ozone that really impacts communities nearby highways. Indirect associations with temperature and humidity might result also from the sensitivity of prospective patients to upper and lower respiratory tract infections due to rhinoviruses and syncytial viruses. From a synoptic point of view, this seasonal peak is registered in Canada first in late August and then, as first cold air masses reach southern portions of US the position in time of this seasonal effect shifts to December and January. Weather conditions in South Florida are favorable for these viruses to proliferate during the later fall and early winter. In this end, a periodically forced SEIR model is considered to account for the periodic outbreaks of respiratory infections as triggers of asthma and its further exacerbation. The time series resulting from such a model follows very closely the time series of recorded cases, showing additionally a lagging effect between 3 and 7 days after the cold temperatures are established. Clear winter skies combined with low temperatures may produce temperature inversions and shallow boundary layers effects. This combination may inhibit the atmospheric mixing both vertically (convection) and horizontally (advection) and the concentration of pollutants near the surface increases. Persistent low temperatures and dry conditions together are factor-stimulating molds and viruses spreading, which combined with moderate levels of ozone appears as the primary triggering mechanisms. A combined effect of thermoregulation and “flu season” seems to polarize the immune system in the direction of an inflammatory response that ignites the asthma attack.