Pediatric patients who had experienced an emergency departmentvisit or an inpatient hospitalization at Cincinnati ChildrensHospital for treatment of acute asthma were studied.
The number of emergency department visits and hospitalizationsfor treatment of asthma were determined by review of emergencydepartment logs and a hospital computer database. Air qualitydata were obtained from a centrally located monitoring station.Ozone concentrations were continuously monitored, and data wererecorded as daily averages and the highest 1-hour average concentrationfor each day. Concentrations of airborne particulates <10µm in diameter were obtained by using a volumetric airsampler with a size-selective inlet, and 24-hour average valueswere calculated. Pollen and fungal counts were obtained by usinga Rotorod sampler (Multidata, Inc, Plymouth Meeting, PA). Multiple-regressionmodels were developed to examine all potential exposure measuresas predictors of the number of daily asthma visits. Poissonregression analysis was used to model the daily number of asthmavisits as a function of air quality data and temporal variables.In the data analyses, air quality measures from 0 to 5 daysbefore the asthma visit date were used, to account for delayedeffects.
A series of Poisson regression models was used to identify predictorsof changes in the number of asthma visits. Initially, the logarithmof pollen counts and the month of the year (April to October)were significant predictors of the number of asthma visits.The number of asthma visits per day was associated with pollencounts reported for the same day (P = .014). The effect wasincreasingly strong, however, for pollen counts recorded 1,2, and 3 days before the visit. The logarithm of the pollencounts lagged 3 days was the most significant predictor of asthmavisits (P < .001). This effect was very strong during thesummer and spring months; however, in the autumn, when pollencounts and asthma visits were both high, daily variations inpollen counts did not account for the variations in daily asthmavisits as they did during other seasons. The analyses also showeda synergistic effect between pollen and particulate levels,in that the exposure-response to pollen counts was moderatelyhigh on days when particulate matter levels were low but wassignificantly higher on days when particulate matter levelswere >33 µg/m3. Fungal spore counts and average ozoneconcentrations were not significant predictors of asthma visits.
Ambient concentrations of pollens and small particles were stronglyassociated with emergency visits for treatment of pediatricasthma in Cincinnati, Ohio. Concentrations of ozone did notappear to be associated with pediatric asthma exacerbations.
Several studies have demonstrated associations between particulatematter levels and emergency department visits, and several haveshown correlations between pollen counts and asthma symptoms.This study shows the added effects of both on asthma symptoms.It would be interesting to evaluate particulate matter levelsand pollen counts in various urban, suburban, and rural settings,to assess their influence. In addition, examination of particulatematter levels inside and outside households, schools, and officesmight give us a better understanding of the conditions thatinfluence asthma. The fact that pollen counts influenced asthmaadmissions in the spring and summer but not the autumn mightbe secondary to other factors that dominate during that season(eg, cold weather and respiratory infections).