Investigation of Air Pollution and Otitis Media in a Population-based Birth Cohort

Abstract

P-727 Introduction: Middle ear infection, or otitis media, is one of the most common childhood diseases in the developed world and its incidence is increasing. It is presently the second most common reason children receive prescription medication. Due to increasing antibiotic resistance in otitis media pathogens, the public health focus has shifted from treatment to prevention for the management of this disease. Environmental tobacco smoke (ETS) is an accepted risk factor but ambient air pollution has not been thoroughly investigated. We hypothesize that ambient air pollutants can influence development of otitis media through suppression of the immune system. Due to the multi-factorial nature of otitis media it is important that a study investigating this disease adequately control for potential confounders. As part of the Border Air Quality Study (BAQS) we took advantage of Canada's universal healthcare administrative databases and linked multiple datasets to control for potential confounders during investigation of this hypothesis in a population-based cohort. Methods: The study population was identified using postal code and date of birth from the vital statistics registry. These individual-level data were linked, by personal health number, to hospitalization and outpatient physician visit records to obtain information on clinical diagnoses (ICD-9 codes: 381, 381.0, 382, 382.0, 382.9). The date and billing status of each file entry was used to collect additional confounder information. These files were further linked with a perinatal database to ascertain breastfeeding initiation and the smoking status of both parents. Exposure histories were linked to each child based on their residential address (six-digit postal code) and neighborhood socioeconomic indicators were obtained from the 2001 Canadian Census. Results: We created a database of exposure, health outcome and risk factor data for every child born in the southwestern British Columbia study area from January 1999–December 2000 (n=59,917; 48.5% female). Age, gender, ethnicity (First Nations status), season, parity, ETS, non-breastfeeding and socioeconomic status were all considered in the analysis as potential confounders. There was complete two year follow-up for 87% of the cohort. Exposures were measured as mean monthly concentrations based on inverse distance weighting of hourly ambient monitoring data (CO, NO/NO2, PM10, PM2.5, O3, SO2) and were also estimated for traffic-related air pollution and wood smoke using land use regression models. Discussion: This database has enabled us to conduct the first population-based study examining the relationship between air pollution and otitis media while controlling for many potential confounders at the individual level.