Ethylenediaminetetraacetic acid (EDTA) is a chelating agent commonly used in protein purification, both to eliminate contaminating divalent cations and to inhibit protease activity. For a number of subsequent applications EDTA needs to be exhaustively removed. Most purification methods rely in extensive dialysis and/or gel filtration in order to exchange or remove protein buffer components, including metal chelators. We report here that dialysis protocols, even as extensive as those typically employed for protein refolding, may not effectively remove EDTA, which is reduced only by approximately two-fold and it also persists after spin-column gel filtration, as determined by NMR and by colorimetric methods. Remarkably, the most efficient removal was achieved by ultrafiltration, after which EDTA became virtually undetectable. These results highlight a potentially widespread source of experimental variability affecting free divalent cation concentrations in protein applications.