It is well known that there are social inequalities in health. Following the ecological approach, unemployment has been one of the most used indicators to study social inequalities. The aim of the present study was to investigate the relationships between indicators of extreme poverty and social unrest, along with unemployment, and mortality in Barcelona, during the years 1989 to 1993. A cross-sectional ecological study was carried out using Primary Health Care Areas (PHCAs) as the unit of analysis. The study population consisted of residents in Barcelona City. The indicators studied as dependent variables were the age-standardized mortality rates of the following causes of death: total mortality; lung cancer; bronchitis, emphysema, and asthma; cirrhosis; cerebrovascular disease; ischemic heart disease; breast cancer; traffic accidents; acquired immunodeficiency syndrome (AIDS); and drug overdose. Independent variables were male unemployment rate of the primary health care areas and indicators of extreme poverty and social conflict. A descriptive analysis, a bivariate analysis using Spearman correlation coefficients, and a multivariate analysis fitting Poisson regression models were carried out. For the main results, one group of causes of death was associated only with unemployment: bronchitis, emphysema and asthma, cerebrovascular disease, and ischemic heart disease (both men and women); lung cancer (only among men); total mortality and cirrhosis (only among women). Among men, another group of causes of death was associated with extreme poverty and/or social unrest, as well as unemployment: total mortality, cirrhosis, and drug overdose. AIDS in men was only associated with extreme poverty and social unrest. We concluded that we see different types of relationships between deprivation and mortality. Unemployment has been related to mortality because of pathologies with socially accepted risk factors (tobacco and alcohol). Causes of death with risk factors not socially accepted (illegal drug use) have been related to indicators of marginality as well as unemployment.