BACKGROUND AND AIM: There is limited understanding on the short-term association between hospital admissions and ambient air pollution in sub-Sahara African countries. Therefore, this study investigated the short-term association of air pollution with daily counts of hospital admissions due to respiratory and cardiovascular diseases in Cape Town, South Africa. METHODS: Generalized additive quasi-Poisson models were used within a distributed lag linear modelling framework to estimate the cumulative effects of PM10, NO2 and SO2 up to a lag of 14 days. We further conducted multi-pollutant models and stratified our analysis by age group, sex and season. RESULTS:The relative risk (95% confidence interval (CI)) for PM10, NO2 and SO2 for all ages, both sexes, and seasons, at lag 0 – 1 for hospital admissions due to respiratory disease (RD) were 3.5% (1.8 – 5.2%), 3.2% (1.0 – 5.5%), 1.8% (0 – 3.6%), respectively. In cardiovascular disease (CVD), 2.2% (0.3% - 4%), 1.8% (-0.6% - 4.3%) and -0.5% (-2.3% - 1.3%), respectively, per inter-quartile range increase of 12 ug/m3 for PM10, 7.3 ug/m3 for NO2 and 3.6 ug/m3 for SO2. In multi-pollutant models, PM10 for associations RD remained significant despite some attenuation. The overall cumulative risk per IQR increase in PM10 for females of all ages was 2.7% (-0.2% - 5.6%), 2.9% (0.1% - 5.9%) for males and 6.8% (2.6% - 11.2%) for aged ≥ 65. However, in CVD the association were not significant for all the pollutants, the overall effect estimate for all ages and sexes were 1.4% (-0.8% – 3.7%), 1.4% (-1.4% - 4.3%) and -1% (-2.9% - 1%) for PM10, NO2 and SO2 in three-pollutant models. CONCLUSIONS:We found robust associations of daily respiratory disease hospital admissions with daily PM10 concentrations. Associations were strongest during the warm season and people aged ≥ 65. KEYWORDS: cardiovascular disease, respiratory disease, multi pollutant, short-term association, DLNM, South Africa
Read full abstract