Safety update on the possible causal relationship between gadolinium‐containing MRI agents and nephrogenic systemic fibrosis

ATTENTION IS DRAWN to recent reports identifying a possible link between a new scleroderma-like disorder and exposure to gadolinium-containing agents used in MRI in patients with end-stage renal disease. A press release dated May 29, 2006 issued by the Danish Medicines Agency presented a possible causal connection between this rare, potentially life-threatening condition referred to as nephrogenic systemic fibrosis (NSF), also known as nephrogenic fibrosing dermopathy (NFD), and gadolinium-containing MRI agents. This advisory alert was stimulated by a publication in January 2006 by Dr. T. Grobner (1) and information given by nephrologists at Copenhagen University Hospital, which was later published (2). This press release was followed by a Public Health Advisory notification by the U.S. Food and Drug Administration (FDA) on June 8, 2006, in which safety procedures were recommended in the use of gadolinium-containing agents (http://www.fda.gov/cder/drug/infopage/gcca/default.htm). The European Society of Urogenital Radiology conducted a rapid survey to collect NSF incidences in patients subjected to dialysis or presenting with a severely reduced glomerular filtration rate. It was found that more than 150 patients with NSF had been exposed to gadolinium-containing agents (3). Simultaneously, the FDA, on December 22, 2006, announced an updated advisory, revealing a total of 90 patients with NSF who had received a gadolinium-based contrast agent for an MRI examination (http://www.fda.gov/cder/drug/advisory/gadolinium_agents_20061222.htm). Subsequently, authorities urged and continue to urge caution in the administration of gadolinium-containing agents in renally impaired patients, particularly as the awareness and recognition of NSF as a clinical entity had evolved over the same period of time that gadolinium-containing agents have been used clinically in patients with end-stage renal disease (2). Findings from Broome et al (4) have shown that dialysis following the administration of a gadolinium-containing agent did not appear to prevent the development of NSF in their patients. However, it should be stressed that exposure to gadolinium-containing agents is not currently a proven prerequisite for the onset of NSF. Several recent publications have reviewed the clinical and laboratory characteristics of NSF (5, 6), and we refer the reader to these for more detailed information about the demographic, epidemiological, and histopathological characteristics of this disease. The exact pathogenesis of NSF is currently unknown, but dermatologists and pathologists have focused on the hypothesis that NSF is mediated by a circulating cell denoted as the circulating fibrocyte, involved in wound healing and tissue remodeling, since histological examinations of skin from patients with NSF have demonstrated infiltrates of cells with the same unique immunohistochemical fingerprint as fibrocytes (7, 8). However, the factors responsible for differentiation of this cell into terminal fibroblast-like cells are not clear (6). Because of the close association of surgery and thrombosis in many NSF patients, thrombosis and/or endothelial damage may be associated with the initiation of a series of events that culminate in NSF. Another possible factor in the development of NSF is the administration of erythropoietin to renally compromised patients. As erythropoietin has profibrogenic potential, it could possibly trigger an exaggerated would-healing response that is histologically similar to NSF (9). Erythropoietin is the principal therapy for anemia in both dialysis and predialysis patients. Its use dramatically increased after the publication of the Dialysis Outcomes Quality Initiative guidelines for anemia management in 1997 (http://www.kidney.org/professionals/KDOQI) (10). Alternatively, a report correlating the decreasing plasma levels of transforming growth factor (TGF)-β after plasmapheresis with the amelioration of clinical signs of NSF in a few patients (11) is consistent with the hypothesis that a lack of treatment with an angiotensin converting enzyme (ACE) inhibitor may at least partially be linked to the development of this disease (12). All commercially available gadolinium-containing MR agents are designed as a single gadolinium atom chelated to a ligand with a high stability in vivo. They all possess equivalent paramagnetic properties, but differ to some extent in their viscosity, osmolality, and ionic characteristics (13). These agents are usually excreted rapidly and almost completely via glomerular filtration. In renally diseased patients, however, the clearance of gadolinium-containing contrast agents is exceedingly prolonged compared to healthy humans (14). It has been speculated that the prolonged half-time of these agents may facilitate dissociation of the gadolinium ligand into a free metal-ion and ligand. Hence, several recent studies have addressed this issue by measuring gadolinium within the tissue of NSF patients exposed to gadolinium-containing agents. High and colleagues have recently reported results in a study commenced to identify gadolinium in blocks of paraffin-embedded tissue from patients with a clinical and histopathological diagnosis of NSF (15). Measurements were conducted with electron microscopy and electron dispersion spectroscopy, revealing detectable particles of gadolinium in four of seven patients. Similar observations have been shown in a case report by Boyd et al (16). These interesting and novel observations corroborate an epidemiological association between exposure to gadolinium-containing agents and development of NSF. However, these findings do not establish causation since accumulation of gadolinium within body tissue has also been demonstrated in healthy humans; such as tissue retention of gadolinium in bone in healthy patients exposed to a clinical acceptable dosage of a gadolinium-containing agent (17). It has not yet been established whether these remote accumulations of gadolinium occur as a free ion or bound to the chelate. However, it is generally known that the stability of a gadolinium chelate is related to the affinity between the chelate and metal, the affinity of the chelate for gadolinium compared to other competing electrolytes, and its dissociation constant. Not only is the relative contribution of each of these three factors in the overall toxicity of gadolinium complexes unclear, the in vitro measurements of several of these chemical properties may not reflect the stability or toxicity of the individual complexes in vivo. Complex stability measurements can be expressed in several different ways, the most common being the thermodynamic stability. The evaluation of thermodynamic properties is based on the assumption of chemical equilibrium, which is hardly ever attained in vivo. Thus, the thermodynamic properties of the gadolinium chelates should be considered as only one of the factors that determine the in vivo stability of these agents. A comparison of the thermodynamic stability and the conditional stability, the degree of metal chelation at physiological conditions (pH = 7.4), indicates that the thermodynamic stability constant greatly underestimates the likely stability of the complex in vivo, and therefore should not be used to predict the stability of the complex in biological systems. Extending the concept of examining the behavior of the gadolinium complex in a more physiologically appropriate environment, studies of the stability of gadolinium chelates in solutions at physiological pH containing those electrolytes present in the extracellular fluid; i.e., zinc, calcium, and copper, have shown that this in vitro index of stability (denoted as the selectivity constant: log Ksel) correlates much better with toxicity (18, 19). The zinc ion is particularly important in this respect, as it has an in vivo concentration of approximately 0.5 to 1.5 μg/mL (7.6–22.9 μM) (20), and although approximately 35% is bound to α2-macroglobulin, the rest is predominantly bound loosely to albumin and is readily available for chelation. If the ligand selectivity is low, as in the case of gadolinium-ethylene diamine tetraacetic acid (EDTA) (log Ksel = 4.23), displacement of the bound gadolinium ion in the chelate complex by the endogenous zinc ion will occur more readily, resulting in an increased toxicity. The log Ksel of the two MRI-available agents gadolinium-diethylene triamine pentaacetic acid (DTPA)-BMA (gadodiamide) and gadolinium-DTPA (gadopentetate dimeglumine) over their corresponding zinc complexes under physiological conditions (pH = 7.4) are 9.04 and 7.04, respectively, indicating that gadolinium-DTPA is approximately 100 times more likely to dissociate in the presence of zinc than gadolinium-DTPA-BMA (21-23). Thus, despite the greater affinity (denoted as log Kf), based on thermodynamic selectivity, between metal and ligand in the gadolinium-DTPA molecule (log Kf = 22.46) compared to gadolinium-DTPA-BMA (log Kf = 16.85), the stability of gadolinium-DTPA-BMA is greater under physiological conditions than gadolinium-DTPA. The third major factor influencing the stability of gadolinium chelates is dissociation kinetics. Dissociation of the complex into a free gadolinium ion and a free chelate is sufficiently slow in relation to the rapid rate of excretion from the body of healthy humans or animals, and, thus, is not considered to significantly affect the in vivo stability or toxicity of any of the commercially available gadolinium-containing MRI agents. Nonetheless, under conditions of extended circulation in the body, as in patients with severely compromised renal function, including those who have developed NSF, this factor may play a greater role in possible adverse effects. In patients who are dialyzed following exposure to the gadolinium-containing agent, the amount of gadolinium in the body will be significantly reduced during dialysis. These difficulties in extrapolating the pharmacokinetic properties of gadolinium chelates from healthy humans to patients with renal disease encouraged Joffe et al (14) to evaluate the pharmacokinetics, dialyzability, and safety of gadolinium-DTPA-BMA injected in patients with severely reduced renal function and not given renal replacement therapy, and in patients undergoing hemodialysis or continuous ambulatory peritoneal dialysis. In 27 patients, the results of serum and urine analysis showed no metabolism or transmetallation of the gadolinium agent. Similar findings were observed in selected serum and dialyzate samples from patients undergoing hemodialysis or continuous ambulatory peritoneal dialysis. These findings agree with the numerous clinical investigations reported about the safety of gadolinium-containing agents. Arsenault et al (24) analyzed 21,078 MRI examinations in 15,830 patients, reporting retrospectively that 151 patients presented impaired glomerular filtration following exposure to gadolinium agents; the overall incidence of adverse events was 3.6%. None of these were classified as severe or life threatening. These findings support the conclusions from another study in 2154 volunteers and patients (25), concluding that a delayed excretion of gadolinium-containing chelates from the body had no adverse effects. In conclusion, the number of cases known today strengthens the association between the administration of gadolinium-containing agents suitable for MRI and the development of the rare scleroderma-like disease NSF. This has prompted immediate actions from involved parties. It is important to stress that NSF has only been observed in a small percentage of exposed patients, predominantly involving patients with severe to end-stage renal insufficiency. Until the exact pathogenesis has been understood in parallel with thorough epidemiological surveillance and the development of NSF-specific animal models and investigation of causal effects of endogenous agents, therapeutic approaches will remain empirical. It can be counterproductive if early and possibly incorrect speculations are perpetuated that are not based on authoritative documentation and sound clinical evidence. Importantly, based on the available information, it is advisable to follow the guidelines from the regulatory authorities (FDA and European medicine agencies) and professional bodies like the European Society of Urogenital Radiology. They recommend that approved gadolinium-containing agents only should be used if clinical essential, counting both patients with severe renal impairment and neonates, the latter because of their immature renal function.