Abstract The depression of cellular immunity in children suffering from severe protein-energy malnutrition was characterized by a decreased number of peripheral T lymphocyte subpopulations and an increased number of immature T lymphocytes. As previously studied, the T lymphocyte immaturity is linked to a deficit of thymulin from the thymic epithelium and to a critical atrophy of the thymus. Hence, the thymus might be considered as the first target for immuno-nutritional studies. During the treatment of severely malnourished children, nutritional recovery, based on anthropometric criteria such as weight for height, was reached in one month. Immunological recovery, based on percentages of the T lymphocyte subpopulations and thymic area, required two months. Because of the lag between nutritional and immunological recoveries, a discharge based only on anthropometrical parameters generated immune-depressed children. The great variation between thymic atrophy on admission and thymic recovery on discharge suggested that the initial diagnosis and the follow-up of the immunological status of malnourished children can be indirectly assessed in situ by a non-invasive method such as ultrasonography of the thymus.