Retention of gadolinium (Gd) in biological tissues is considered an important cofactor in the development of nephrogenic systemic fibrosis (NSF). Research on this issue has so far focused on the stability of Gd-based contrast media (GdCM) and a possible release of Gd3+ from the complex. No studies have investigated competing chelators that may occur in vivo. We performed proton T1-relaxometry in solutions of nine approved GdCM and the macromolecular chelator heparin (250 000 IU per 10 ml) without and with addition of ZnCl2. For the three linear, nonspecific GdCM complexes, Omniscan®, OptiMARK® and Magnevist®, 2 h of incubation in heparin at 37 °C in the presence of 2.0 mm ZnCl2 led to an increase in T1-relaxivity by a factor of 7.7, 5.6 and 5.1, respectively. For the three macrocyclic complexes, Gadovist®, Dotarem® and Prohance®, only a minor increase in T1-relaxivity by a factor of 1.5, 1.6 and 1.7 was found, respectively. Without addition of ZnCl2, no difference between the two GdCM groups was observed (factors of 1.4, 1.2, 1.1, 1.3, 1.5 and 1.4, respectively). The increase in T1-relaxivities observed for linear GdCM complexes may be attributable to partial transchelation with formation of a macromolecular Gd–heparin complex. For comparison, mixing of GdCl3 and heparin results in a 8.7-fold higher T1-relaxivity compared with a solution of GdCl3 in water. Heparin is a glycosaminoglycan (GAG) and as such occurs in the human body as a component of the extracellular matrix. GAGs generally are known to be strong chelators. Gd3+ released from chelates of GdCM might be complexed by GAGs in vivo, which would explain their retention in biological tissues. Plasma GAG levels are elevated in end-stage renal disease; hence, our results might contribute to the elucidation of NSF. Copyright © 2012 John Wiley & Sons, Ltd.