Blood rheology and microcirculation are strongly influenced by red blood cell (RBC) aggregation. The aggregability of RBCs can vary significantly due to factors such as their mechanical and membrane surface properties, which are affected by cell aging in vivo. In this study, we investigate RBC aggregability as a function of their density, a marker of cell age and mechanical properties, by separating RBCs from healthy donors into different density fractions using Percoll density gradient centrifugation. We examine the dissociation rates of aggregates in a controlled medium supplemented with Dextran, employing an extensional flow technique based on hyperbolic microfluidic constrictions and image analysis, assisted by a convolutional neural network (CNN). In contrast to other techniques, our microfluidic experimental approach highlights the behavior of RBC aggregates in dynamic flow conditions relevant to microcirculation. Our results demonstrate that aggregate dissociation is strongly correlated with cell density and that aggregates formed from the denser fractions of RBCs are significantly more robust than those from the average cell population. This study provides insight into the effect of RBC aging in vivo on their mechanical properties and aggregability, underscoring the importance of further exploration of RBC aggregation in the context of cellular senescence and its potential implications for hemodynamics. Additionally, it suggests that this technique can complement existing methods for improved evaluation of RBC aggregability in health and disease. Previous article