Pollen exposure and hospital admissions for asthma in London, UK 2005-11

Abstract

Introduction A high proportion of the UK population has been reported as allergic to pollen. These have increased in prevalence in the UK population quickly over the last 20 years. Pollen from trees in the UK is released in the March to May period and grass pollen and output mid-May to July most years. It is highly variable on a daily basis due to a number of factors including latitude, altitude, phenology and weather. Factors that could be important in influencing pollen-asthma associations include precipitation, thunderstorms, wind, periods of dry weather, and heat. Methods Estimation was made of the association between measured pollen (grains/m3 air) (Met Office, UK) and asthma hospital admissions (Hospital Episode Statistics data, Department of Health) in an ecological time series regression for the period 2005-11 in London. Regression models were specified to estimate percentage change in risk at the 95th percentile versus the 0th percentile associated with 0-7 day lags in unconstrained distributed lag models. Within season trend was accounted for using natural cubic splines of observation days in study seasons. Models were adjusted for air pollution (PM10, NO2, SO2 and O3) and weather variables (temperature, rain and humidity). Results Increases in hospital admissions of 17.2% (95% CI 8.9, 25.5) and 14.1% (95% CI 6.2, 22.0) were noted at 4- and 5-day lags post pollen exposure, respectively, with a 0-95th percentile increase in grass pollen exposure (no increase was found on same day). There was little observed effect from any of the tree pollens. Grass pollen count correlated with O3 levels but not other air pollutants measured. Conclusions Lag effects may be important when examining the relationship between grass pollen and hospital admissions. No synergistic relationship was found between air pollution and pollen and their effect on asthma admissions to hospitals.