Long-term exposure to ambient air pollution with PM2.5 and associations with asthma incidence in children born in Denmark 1998-2016
Marie Pedersen1, Shuo Liu1, Zorana Jovanovic Andersen1, Anne-Marie Nybo Andersen1, Hans Bisgaard2, Jørgen Brandt3, Esben Budtz-Jørgensen1, Klaus Bønnelykke2, Lise Marie Frohn3, Leslie Stayner4, Jibran Khan3, Matthias Ketzel3, Bert Brunekreef5, Steffen Loft1
1University of Copenhagen, Copenhagen, Denmark
2Gentofte hospital, Copenhagen, Denmark
3Aarhus University, Roskilde, Denmark
4University of Illinois at Chicago, US
5Utrecht University, Utrecht, the Netherlands
Background. Long-term exposure to ambient air pollution with particulate matter £2.5 mm (PM2.5) has been associated with onset of asthma, but outcome definitions differ between studies.
Objectives. We examined the associations between early-life exposure to ambient PM2.5 from all sources and asthma incidence in children aged 0-18 years.
Methods. We included all live-born singletons born in Denmark during 1998-2016. We used the unique personal identification number for data linkage and to extract information about asthma, personal characteristics and residential address history from national registers. We defined incidence of asthma as the first registered diagnosis according to ICD-10 codes J45* or J46* from hospital admission, emergency room and outpatients records for all children and after restricting to children with 4 or more years of follow-up and defining asthma according to the first diagnosis after 4 years of age. Monthly mean concentrations of PM2.5 from all sources were modelled using the multiscale DEHM-UBM-AirGIS dispersion model system. Time weighted prenatal and postnatal mean exposures were calculated taking all residential addresses into account and using the gestational age and date of diagnosis. We used Cox proportional hazard models with age as the time dimension and adjustment for calendar year, sex, parity, season, household income, maternal education and smoking during pregnancy to estimate hazard ratios (HRs) and 95% confidence intervals (95% CIs) fitted with time-weighted average exposures for the prenatal period (fixed exposure means) and time-varying annual average exposure a year before the event for the postnatal follow-up period.
Results. Out of the 1,060,154 children, 6.1% had an asthma diagnosis during the mean follow-up of 8.8 years. The prenatal mean concentrations of PM2.5 was 10.5 (SD: 1.8) mg/m3. A 5 mg/m3 increase in prenatal mean PM2.5 from all sources was associated with a HR of 1.07 (95% CI: 1.04, 1.10) for all. The association strengthened to 1.23 (95% CI: 1.18, 1.28) after restricting to children with 4 or more years of follow-up and defining asthma according to the first diagnosis after 4 years of age. The associations from models with postnatal exposures were almost identical to those of prenatal exposure. Additional adjustment for maternal asthma did not make a significant difference.
Conclusions. The findings strengthen the evidence that exposure to ambient PM2.5 from all sources contribute to asthma development and that the association is stronger at older ages of more certain diagnosis.