Rates of premature mortality among adults are important measures of the economic and psychosocial well-being of human populations. In many countries, such rates are, as a rule, inversely related to the level of attained education. We examined changes in educational group-specific mortality rates among urban adults in Poland during the country's rapid transition in the 1990s from a socialist command economy to a free market system. Two census-based analyses of individual death records of urban dwellers aged 35–64 years were compared. We utilized all records of death, which occurred during the 2-year periods 1988–89 and 2001–02. Population denominators were taken from the censuses of 1988 and 2002. The age-specific mortality rates were used to evaluate absolute differences in mortality. To assess relative differences between educational levels, mortality rate ratios (MRRs) with 95% CI (confidence interval) were calculated using Poisson regression. A regular educational gradient in mortality persisted in each 10-year age group throughout the period covered by our data. Moreover, age-specific mortality rates declined steadily in all educational groups, and this decline was most marked in the two oldest age groups (45–54 and 55–64 years). The trend was accompanied by widening of educational differences in mortality as expressed by MRRs. Systemic political transformation in Poland has brought a mixture of beneficial and detrimental effects on the well-being of society. With regard to the changes in rates of premature mortality among adults, the benefits have prevailed, although individuals with the lowest educational level benefited less than those with the highest education.