This paper analyses 14 epidemiologic investigations (published from January 1980 to September 1991) about the relationship between urban air pollution and mortality. Air pollution indicators and mortality indicators are examined. Methods to analyse the relationships between these two kinds of indicators are classified according as they bring "qualitative information" trying to answer the question: "is there any relationship between air pollution and mortality?" or according as they try to quantify this relationship. Results are presented by author. Confounding factors and means to take them into account are described. This paper ends by a discussion about interest and limits of these studies. It emphasizes the importance of the collaboration between metrologists of urban air pollution and epidemiologists.