Respiratory hospital admissions are associated with ambient airborne pollen in Darwin, Australia 2004-2005
Hanigan, I and Johnston, F, Respiratory hospital admissions are associated with ambient airborne pollen in Darwin, Australia 2004-2005, Clinical and Experimental Allergy, 37, (10) pp. 1556-1565. ISSN 0954-7894 (2007) [Refereed Article]
Background: Although the role of pollen and fungus in specific allergic disorders has been well established, the public health impacts of ambient concentrations of airborne pollen and fungal spores; the shapes of concentration-response relationships; and the relative effects of different taxa are gaps in current knowledge. Objective: To investigate associations between daily average ambient pollen and fungal spore concentrations with hospital admissions for total respiratory diseases; asthma; chronic obstructive pulmonary disease (COPD); and respiratory infections in Darwin, Australia, during the period from April 2004 to November 2005. Methods: We assessed these relationships in a two-stage modelling approach designed to quantify potential non-linear relationships. First, generalized additive models determined the shapes of concentration-response relationships. Second, linear associations were examined using generalized linear models. Non-linear relationships were analysed by categorizing pollen and fungal spore concentrations based on their distributions. Results: Positive linear associations were found between total pollen concentrations and hospital admissions for total respiratory diseases and COPD. While our exploratory first-stage analysis suggested non-linear relationships for total pollen with asthma and respiratory infections, no convincing evidence for these relationships was found in the second-stage analysis. When individual taxa were investigated, associations were the strongest in relation to Myrtaceae pollen (the dominant tree taxa in the region), while positive associations not attaining statistical significance were observed for Poaceae, Cyperaceae and Arecaceae. No associations were evident for any conditions with fungal spores. Conclusions: Our finding of an association between pollen count and respiratory hospital admissions that could not be explained by asthma admissions suggests that ambient airborne pollens might have a wider public health impact than previously recognized.