In the last decade, uremic toxicity as a potential cause for the excess of cardiovascular disease and mortality observed in chronic kidney disease gained more and more interest. This review focuses on uremic toxins with known cardiovascular effects and their removal. For protein-bound solutes, for example, indoxylsulfate and the conjugates of p-cresol, and for small water-soluble solutes, for example, guanidines, such as ADMA and SDMA, there is a growing evidence for a role in cardiovascular toxicity in vitro (e.g., affecting leukocyte, endothelial, vascular smooth muscle cell function) and/or in vivo. Several middle molecules (e.g., beta-2-microglobulin, interleukin-6, TNF-alpha and FGF-23) were shown to be predictors for cardiovascular disease and/or mortality. Most of these solutes, however, are difficult to remove during dialysis, which is traditionally assessed by studying the removal of urea, which can be considered as a relatively inert uremic retention solute. However, even the effective removal of other small water-soluble toxins than urea can be hampered by their larger distribution volumes. Middle molecules (beta-2-microglobulin as prototype, but not necessarily representative for others) are cleared more efficiently when the pore size of the dialyzer membrane increases, convection is applied and dialysis time is prolonged. Only adding convection to diffusion improves the removal of protein-bound toxins. Therefore, alternative removal strategies, such as intestinal adsorption, drugs interfering with toxic biochemical pathways or decreasing toxin concentration, and extracorporeal plasma adsorption, as well as kinetic behavior during dialysis need further investigation. Even more importantly, randomized clinical studies are required to demonstrate a survival advantage through these strategies.