Time series analysis of air pollution and mortality: eVects by cause, age and socioeconomic status

Objective—To investigate the association between outdoor air pollution and mortality in São Paulo, Brazil. Design—Time series study Methods—All causes, respiratory and cardiovascular mortality were analysed and the role of age and socioeconomic status in modifying associations between mortality and air pollution were investigated. Models used Poisson regression and included terms for temporal patterns, meteorology, and autocorrelation. Main results—All causes all ages mortality showed much smaller associations with air pollution than mortality for specific causes and age groups. In the elderly, a 3–4% increase in daily deaths for all causes and for cardiovascular diseases was associated with an increase in fine particulate matter and in sulphur dioxide from the 10th to the 90th percentile. For respiratory deaths the increase in mortality was higher (6%). Cardiovascular deaths were additionally associated with levels of carbon monoxide (4% increase in daily deaths). The associations between air pollutants and mortality in children under 5 years of age were not statistically significant. There was a significant trend of increasing risk of death according to age with eVects most evident for subjects over 65 years old. The eVect of air pollution was also larger in areas of higher socioeconomic level. Conclusions—These results show further evidence of an association between air pollution and mortality but of smaller magnitude than found in other similar studies. In addition, it seems that older age groups are at a higher risk of mortality associated with air pollution. Such complexity should be taken into account in health risk assessment based on time series studies. (J Epidemiol Community Health 2000;54:750–755) Early documented air pollution episodes such as the London smog of 1952 led to the premature deaths of thousands of people, and prompted governments of many countries to create and enforce environmental policies that contributed to reduce their urban levels of air pollution. A relation with mortality at lower levels of exposure has only become apparent with recent improvements in analytical tools, including the use of time series methods. A number of recent time series mortality studies have provided evidence of positive associations between air pollution and all cause mortality, cardiovascular mortality, respiratory mortality 11 or even more specific causes such as pneumonia and chronic obstructive pulmonary disease (COPD) mortality. 13 Fine particulate matter (PM10) has been implicated in most studies and many consider this pollutant the most important. Less frequently, pollutants such as sulphur dioxide (SO2), nitrogen dioxide (NO2) and ozone (O3) have also been found to be related to health events. As yet few of these studies have focused on developing countries. There remains a need for studies in cities of developing countries, where levels of air pollution and meteorological conditions are diVerent from North America and Western Europe. In addition, despite early evidence indicating that the eVect of air pollution was possibly higher in more vulnerable sub-populations such as young children and the elderly, few studies conducted so far have attempted to explore this issue in depth. Moreover, no studies to date have examined the role of socioeconomic status in the vulnerability of subpopulations to air pollution. Several studies, including one in São Paulo, Brazil have shown that social and physical deprivation places people at a higher risk of mortality. Therefore, it seems reasonable to explore if socioeconomically deprived populations are also more vulnerable to air pollution. A preliminary time series study carried out in São Paulo, Brazil reported a significant association between daily respiratory mortality in children under 5 years old and levels of oxides of nitrogen. In a more recent study, the same investigators found an association between levels of PM10 and mortality in the elderly. The present study of a larger dataset extends the investigation of the eVects of air pollution in São Paulo. Daily mortality for all age groups is explored, but with emphasis on children under 5 years old and the elderly. We documented eVects on cause specific as well as all cause mortality and sought to identify the most vulnerable population sub-groups defined in terms of age and socioeconomic status that may be more sensitive to short-term health eVects of air pollution. Methods The study covered a period of three years (1991 to 1993) and was carried out in São Paulo, the most urbanised, industrialised and aZuent city in Brazil with a population of about 9.5 million inhabitants according to the 1991 national census. It encompassed all but one administrative district of the city, which was excluded because it has a low population density and is the furthest from an air quality J Epidemiol Community Health 2000;54:750–755 750 Departamento de Medicina Preventiva, Faculdade de Medicina da Universidade de São Paulo-FMUSP, Av Dr Arnaldo 455, Sao Paulo, 01246–903-SP, Brazil