RESUMO
Effects of meteorological variables and air pollutants on child respiratory morbidity are investigated during two consecutive summers (December-March 1992/1993 and 1993/1994) at the Metropolitan Area of São Paulo (MASP), Brazil. The MASP, with almost 17 million inhabitants, is considered the most populous region in South America. Due to warmer temperatures, increased rainfall and consequent low levels of air pollutants during the summer compared to winter, less attention has been paid to epidemiological studies during this season, especially in tropical urban areas such as São Paulo. In the present work, principal component analysis (PCA) is applied to medical end environmental data to identify patterns relating child morbidity, meteorological variables and air pollutants during the summer. The following pollutant concentrations are examined: SO2, inhalable particulate matter (PM10), and O3. The meteorological variables investigated are air temperature, water vapor (water vapor density) and solar radiation. Although low correlation between respiratory morbidity and environmental variables are, in general, observed for the entire dataset, the PCA method indicates that child morbidity is positively associated with O3 for the 1992/1993 summer. This pattern is identified in the third principal component (PC3), which explains about 19% of the total variance of all data in this summer. However, the 1993/1994 summer shows a more complex association between both groups, suggesting stronger ties with meteorological variables. Marked changes in synoptic conditions from the end of January to end of March of the 1993/1994 summer seem to have played an important role in modulating respiratory morbidity. A detailed examination of meteorological conditions in that period indicates that prefrontal (postfrontal), hot (cold) and dry (wet) days favored the observed decrease (increase) of respiratory morbidity.