Associations between meteorological factors, air pollution and Legionnaires’ disease in New Zealand: Time series analysis

Abstract

BackgroundPrior studies have shown that meteorological factors may be associated with increases in legionellosis (Legionnaire's disease, (LD)), caused by Legionella, a globally ubiquitous bacterium found naturally in aquatic habitats, soils, and compost. The aim of this retrospective time series analysis was to examine the association between meteorological factors and air pollution parameters and the incidence of sporadic, community-acquired, laboratory confirmed LD. MethodsDaily cases of community-acquired legionellosis, meteorological and air pollution data from two urban areas, Auckland (North Island) and Christchurch (South Island) were collected from January 1, 1997 until December 31, 2020. Using Quasi-Poisson regression, associations between symptom onset and meteorological and air pollution variables were investigated using an interrupted time series analysis. ResultsThe two cities had different meteorological conditions and LD epidemiology and seasonal patterns of Legionella spp. LD incidence rates (per 100,000 population) were higher in Christchurch than Auckland for L.pneumophila (25.8 vs 10.8) and L.longbeachae (78.2 vs 4.9). Seasonal patterns were detected in Christchurch with a higher risk of LD during spring and summer (RR: 1.87, 95% CI: 1.42, 2.49) compared to autumn and winter months. In Auckland, the level of particulate matter 9–10 days prior to the onset date was positively associated with LD occurrence (RR: 1.02, 95% CI: 1.00, 1.04) compared to Christchurch, where Tmax recorded one day prior the onset (RR: 1.03, 95% CI: 1.00, 1.07) and sulphur dioxide 6 days prior to the onset date (RR: 1.27, 95% CI: 1.10, 1.45) were positively associated with LD occurrence. Atmospheric pressure 12 days prior (RR: 0.95, 95% CI: 0.90, 1.00) and wind speed 13 days prior (RR: 0.94, 95% CI: 0.89, 0.99) to the onset date were negatively associated with LD risk. In both cities, no association was detected between the level of precipitation and LD risk. ConclusionsMeteorological factors and air pollutants were associated with the risk of LD. However, different seasonal patterns and prevalent LD species seem to have distinct patterns of association between the two categories of exposures. These findings suggest the importance of considering meteorological and air quality conditions in conjunction with the source of exposure and the LD species involved. They also imply potential climate change impacts on LD and benefits from reducing air pollution, though findings need to be replicated in other geographical regions.