The study investigates the magnitude of differences in child and adult mortality by wealth in Uzbekistan, a former soviet country of Central Asia, and compares it with similar indicators from sub-Saharan Africa. Data were derived from Demographic and Health Surveys. An “Absolute Wealth Index” was built from data on goods owned by households and quality of housing, and scaled from 0 to 12. Wealth was distributed evenly in Uzbekistan, with a symmetric distribution around a mean of 5.5 modern goods. In sub-Saharan Africa, on the contrary, the wealth distribution had a lower mean (2.5) and was highly skewed towards the left, revealing a high proportion of very poor people. Adult and child mortality levels were lower in Uzbekistan. Despite these major differences, the relationships between mortality indicators and the wealth index were similar in the two cases. The magnitude of mortality differentials by wealth was of the same order in both cases, with gradients ranging from 2.5 to 1 for child mortality and 1.5 to 1 for adult mortality (poorest versus richest). However, mortality levels remained lower in Uzbekistan than in sub-Saharan Africa at the same level of wealth for both children and adults. A similar relationship was found between nutritional status and wealth index in both cases. On the contrary, there were no differences by wealth in use of health services and level of education in Uzbekistan, whereas wealth gradients were steep for the same variables in sub-Saharan Africa. The study suggests that mortality differentials were primarily due to nutritional status, and not to access and use of health services or to education. The discussion focuses on health and social policies during the colonial and post-colonial period that have produced these patterns.