Is NSF only the tip of the “gadolinium toxicity” iceberg?

EXTRACELLULAR GADOLINIUM-BASED contrast agents (Gd-CA) are either linear or macrocyclic chelates and both chelates are available as ionic or nonionic preparations. Linear chelates are flexible open chains which do not bind gadolinium strongly, whereas the macrocyclic chelates do bind gadolinium strongly because they are rigid rings of almost optimal size to cage the gadolinium atom. Nonionic linear preparations are also less stable compared with ionic ones because binding between Gd3+ and the negatively charged carboxyl groups is stronger than with amides or alcohol in the non ionic preparations. Based on the stability constants and kinetic measurements, the least stable agents are the nonionic linear chelates gadodiamide and gadoversetamide. The stability of Gd-CA seems to be an important factor in the pathogenesis of the serious adverse reaction “nephrogenic systemic fibrosis (NSF).” Already in the early 1990s, there was concern about the stability of the nonionic linear chelates (1-4). Varadarajan et al (2) wrote in 1994 that “the benefits of high kinetic and thermodynamic stability offered by structurally preorganized and rigid metal chelates such as DOTA macrocycles for use as MRI CA are well established. DTPA-based chelates are kinetically labile, raising the possibility of in vivo transmetallation with endogenous metal ions.” Watson et al (3) stated 2 years earlier that the superiority of the cyclic over the linear agent was demonstrated by an 85% improvement in safety, thus the ratio of LD50 to effective dose was 53 for the cyclic agent versus 28 for the linear agent. The cyclic agent also showed significantly higher stability in serum as judged by gadolinium release; less than 2% gadolinium was released from the cyclic agent after 150 h compared with 10% to 20% gadolinium released from the linear agent. Today, we must ask ourselves why these warnings were not heard. Why were nonionic linear chelates chosen instead of macrocyclic chelate? Today, several end-stage renal failure patients are suffering from gadolinium toxicity. We will probably not get the answers to these questions before the legal cases against the producers of nonionic linear chelates have been settled by the courts. In contrast to the American Food and Drug Administration (FDA), the European Medicines Agency (EMEA) came to the conclusion that the stability of the different agents varies and that this is of importance for the prevalence of NSF. In their recent public assessment report (5) of June 26th 2007 the agency wrote “Are all Gd-based CA associated with the same risk of NSF? No. Current evidence suggests that the risk of developing NSF is related to the physicochemical and pharmacokinetic properties of Gd-based CA. The physicochemical properties affect the release of toxic Gd3+ from the chelate complex, and the pharmacokinetic properties influence how long the agent remains in the body”. Today, in Europe gadodiamide (Omniscan®, GE Healthcare, Chalfont St. Giles, UK), gadopentetate dimeglumine (Magnevist®, Bayer Schering, Berlin, Germany) and the recently approved gadoversetamide (OPTImark, Covidien, St. Louis, MO) are contraindicated in patients with a glomerular filtration rate below 30 mL/min/1.73 m2 including those on dialysis (CKD 4 & 5). It is also recommended that there agents should be used only with caution in patients with moderately reduced kidney function (GFR 30–60 mL/min/1.73 m2 (CKD 3)). The Japanese authorities have recently taken the same position as the Europeans. It is difficult to understand how the European and Japanese regulatory agencies could come to such different conclusions from the American agency. No unconfounded cases of NSF have been reported after gadoterate meglumine, gadoteridol, gadobenate dimeglumine, or gadobutrol in the published peer-reviewed literature. These agents have been used in many patients in imaging departments serving nephrology centers (6). A single case of stage 0 NSF (Marckmann classification (7)) has been published in the non–peer-reviewed literature after several injections of gadoteridol over a short period (> 200 mL, < 2 years), but sufficient details have not yet been published. Generally the non–peer-reviewed literature is not reliable because documentation is not sufficient to validate cases. Many diseases may mimic NSF clinically and dermatopathologically (8). However, the absence of reports in the peer-reviewed literature does not mean that it is impossible that these agents could induce NSF, but rather suggests that the risk is significantly lower than for example after gadodiamide. There have been reports of NSF in patients who have been exposed to less stable as well as to stable or high relaxivity agents. They are normally classified as confounded cases (9). What do we do with patients who have had one or more injections of less stable agents in the past? There seems to be an increased risk of NSF in these patients compared with those who have never have been exposed to an less stable agent even though a stable agent is used subsequently. The exact risk is unknown. In general, it seems appropriate to go ahead with enhanced MRI if the examination is clinically justified. Life-time exposure is correlated with severe NSF in patients with end-stage renal failure or on dialysis (10) indicating that some godolinium is left in the body, whereas the triggering dose does not differ between severe and nonsevere cases. Gadolinium has been demonstrated in the skin of NSF patients (11). Abraham et al (12) showed that amount of gadolinium in the skin may increase up to at last three years after last injection of gadodiamide. Where does the gadolinium come from? Very likely from the bone, but other tissues may be involved. Gadolinium accumulates in the bone of patients with normal renal function. Four times more accumulates after an less stable nonionic linear agent than a nonionic cyclic agent (13). Bone has a slow turnover. Nowadays patients undergo multiple enhanced MRI examinations, e.g., women with increased risk of breast cancer are encouraged to undergo annual enhanced MRI. After each examination some gadolinium will accumulate in the bone (and probably other tissues) and it will stay there for many years (12-14). What will happen when the gadolinium is released? It is known that osteoporosis increases bone turnover. The other heavy metal lead also accumulates in the bone and many years ago we saw patients who developed signs of lead toxicity when their bones became osteoporotic. Are we going to see a similar phenomenon with gadolinium within the next 10, maybe 20 years? It is impossible to say that this will not be the case. As significantly less amounts of ligand free gadolinium accumulate after administration of a cyclic agent than after administration of a linear agent, the use of a cyclic agent seems to be preferable. The release of high amounts of gadolinium over a short period, for example, in the case of osteoporosis, might cause the classic symptoms of gadolinium toxicity rather than NSF. Gadolinium alone is highly toxic in vivo because it is distributed to bone, lymph nodes and to the liver, where it rapidly produces liver necrosis. It obstructs calcium ion passage through muscle cells (reducing neuromuscular transmission), and interferes with intracellular enzymes and cell membranes by the process of transmetallation, a phenomenon whereby Gd3+; replaces endogenous metals such as zinc and copper. Copenhagen University Hospital Herlev has not seen a single case of NSF after it changed from gadodiamide to a cyclic agent in mid March 2006 in all patients independent of renal function (15). Other institutions, for example, University of Wisconsin, University of North Carolina, have reported the same finding at the European Congress of Radiology (Vienna 2008). Institutions that have used stable agents report that they have no cases of NSF (16, 17). Copenhagen University Hospital Herlev has not changed their imaging protocols and does not require assessment of renal function before enhanced MRI, because administration of the cyclic agent must be clinically justified. Use of a cyclic agent will also reduce the potential future risk of seeing other aspects of gadolinium toxicity. Have we come to the end for the nonionic linear agents? Probably. They offer no diagnostic advantages compared with the other extracellular agents. In Denmark, the market share of the nonionic linear agent has dropped from 58% in the third quarter of 2005 to below 8% in the third quarter of 2007. The European Society of Urogenital Radiology has recently released a guideline on NSF (Table 1). It has already been adopted by several national societies. However, it does not take the problem of accumulation of the heavy metal gadolinium in bone and other tissue after release from the chelate into consideration. When is it appropriate to conclude that a gadolinium-based contrast agent is safe? For example contrast agent X has been administered Y × 106 times during the last 10 years. Does this information suggest that it is safe? No. In reality the statement says nothing. How is the number of injections reached? Is the figure reached the simple way: the number of liters sold divided by the dose for a standard person? How many of those included in the calculation received a triple dose? In what types of patient was the agent administered? When we talk about less stable agents there is certainly a difference between whether the agent was given in, for example, 30 million patients once or 10 times in 3 million patients over a short period. The latter group is the group at higher risk of heavy metal poisoning than the first group. Therefore long-term observation (e.g., 20 years) will be necessary before we can conclude anything about safety of gadolinium-based contrast agents, in particular those with the lowest stability. It is important to remember that testing before approval by the health authorities only included a single injection and whether or not acute nonrenal and acute renal adverse reactions occurred. No gadolinium-based agent has been tested with multiple injections and for long-term effects (heavy metal poisoning) before approval. Recent studies (10, 12-14) have indicated that low stability agents may have a problem. Good research on the long time effects is strongly warranted. In the meantime, we need gadolinium-based agents to use for clinically well justified enhanced MRI and agents which leave the least amount of gadolinium in the body seem preferable for all patients. Based on both the older and the most recent literature, it is possible that with NSF we have only seen the tip of the “gadolinium toxicity” iceberg. It will take years before we get the final answer. A lot of research on the stability of the various agents and the long-term effects of gadolinium must be undertaken. Until firm facts are available, agents that leave the least amount of free gadolinium in the body should be preferred.