[Air pollution and infant mortality from pneumonia]

This study examines the relationship between air pollution, measured as concentration of suspended particulates in the atmosphere, and infant mortality due to pneumonia in the metropolitan area of Rio de Janeiro. Multiple linear regression (progressive or stepwise method) was used to analyze infant mortality due to pneumonia, diarrhea, and all causes in 1980, by geographic area, income level, and degree of contamination. While the variable "proportion of families with income equivalent to more than two minimum wages" was included in the regressions corresponding to the three types of infant mortality, the average contamination index had a statistically significant coefficient (b = 0.2208; t = 2.670; P = 0.0137) only in the case of mortality due to pneumonia. This would suggest a biological association, but, as in any ecological study, such conclusions should be viewed with caution. We believe that air quality indicators are essential to consider in studies of acute respiratory infections in developing countries.