Abstract: OBJECTIVES: Air pollution is an important public health concern especially for children who are particularly susceptible. Latin America has a large children population, is highly urbanized and levels of pollution are substantially high, making the potential health impact of air pollution quite large. We evaluated the effect of air pollution on children respiratory mortality in four large urban centers: Mexico City, Santiago, Chile, and Sao Paulo and Rio de Janeiro in Brazil. METHODS: Generalized Additive Models in Poisson regression was used to fit daily time-series of mortality due to respiratory diseases in infants and children, and levels of PM10 and O3. Single lag and constrained polynomial distributed lag models were explored. Analyses were carried out per cause for each age group and each city. Fixed- and random-effects meta-analysis was conducted in order to combine the city-specific results in a single summary estimate. RESULTS: These cities host nearly 43 million people and pollution levels were above the WHO guidelines. For PM10 the percentage increase in risk of death due to respiratory diseases in infants in a fixed effect model was 0.47% (0.09-0.85). For respiratory deaths in children 1-5 years old, the increase in risk was 0.58% (0.08-1.08) while a higher effect was observed for lower respiratory infections (LRI) in children 1-14 years old [1.38% (0.91-1.85)]. For O3, the only summarized estimate statistically significant was for LRI in infants. Analysis by season showed effects of O3 in the warm season for respiratory diseases in infants, while negative effects were observed for respiratory and LRI deaths in children. DISCUSSION: We provided comparable mortality impact estimates of air pollutants across these cities and age groups. This information is important because many public policies aimed at preventing the adverse effects of pollution on health consider children as the population group that deserves the highest protection.
Keywords: Air pollution; Children mortality; Infant mortality; Respiratory diseases; Time series