Abstract
More than 6 years have passed since Grobner (1) and Marckmann et al. (2) reported their seminal observations of a previously unrecognized link between exposure to gadodiamide and the development of nephrogenic systemic fibrosis (NSF). It is time to ask what we have learned.
Was it a surprise?
Yes, for the overwhelming majority of physicians, but not for a small minority. In 1983, researchers at University of Pittsburgh had already reported that exposure to lanthanides resulted in changes in the collagen structure (3). In 1990, the textbook Biochemistry of Lanthanides, a comprehensive source of information about these ions, was published (4). In 1991, a workshop on contrast-enhanced magnetic resonance organized by the Society of Magnetic Resonance in Medicine was held in Napa Valley. Abstracts and discussions were published in Magnetic Resonance in Medicine (5), but few individuals read the abstracts and the related discussions. The stability of non-ionic linear chelated gadolinium was heavily debated. It was noted that the reported pathological changes after exposure to a non-ionic linear chelating gadolinium agent were not seen after exposure to an ionic linear agent. In 1992 pre-clinical findings with the non-ionic linear chelate gadolinium agent, gadopendiamide, was reported; due to the concerning safety findings, which were similar to those seen after other non-ionic linear chelates, the agent was never processed to clinical trial (6). Three years later, a Belgian group showed using relaxometry that gadodiamide releases gadolinium with time in serum (7); unfortunately the report was not released until 2012, and it was the Danish Ministry of Health, who released this report. Frenzel et al. (8) showed the same findings in 2009 including the observation that non-ionic linear agents differed significantly from the other agents in serum with regard to their ability to hold the toxic gadolinium ion. We will never know why these safety concerns never received much notice outside the small group of experts, but it illustrates the importance of dissemination of all information.
Why did it take so long time to identify the link?
It is always a question about how to interpret unusual clinical events; it is even more difficult if one does not know all the relevant pre-clinical findings (9). In some instances there are no safety signs in the pre-clinical studies, but safety signs have also been misinterpreted; the issue whether there were truly safety signs or not is often solved in courts of law.
The first case of NSF was diagnosed in California in 1997, although at that time the syndrome was termed nephrogenic fibrosing dermopathy (NFD). During the next 3 years, three other university centers reported NFD cases and the first case series was published (10). Administration of contrast media is a single event, which in most instances are not listed together with other medications in the patient's medical record. Patients with reduced renal function receive numerous pharmaceuticals as well as they are subject to a wide range of skin lesions due to their uremia. Thus there could be several factors that may have led to the occurrence of a severe dermopathy in these patients. The facts are (i) NSF does not occur in all patients exposed to the same agent; (ii) the lifetime dose prior to development of NSF does vary; (iii) NSF varies from a single small plaque to severe disability and death; and (iv) NSF represents a systemic illness, hence the syndrome was renamed in 2003 NSF from NFD. In many instances sales representatives are queried by clinicians whether they have heard about untoward reactions to pharmaceuticals that they are responsible for marketing to clinicians, but the sales people responsible for contrast media marketing never visit clinicians in clinical departments and the clinicians did not inform the radiologists when unexpected dermopathies developed among patients with chronic kidney disease (CKD). However, even if the sales representatives had been called by clinicians, it would probably never have changed anything. When a toxicity is not listed in the regulatory approved summary of product characteristics, it would be considered unlikely that an exposure to a particular drug would be associated with or even the cause of the suspected adverse reaction. There must be a certain number of cases of a particular syndrome before one hypothesizes a link between a drug and an unusual clinical syndrome – although as few as five, seven, and 11 cases led astute clinicians to initially identify causal relationships among gadodiamide, aspirin, and epoetin as causes of NSF, Reye's syndrome, and pure red cell aplasia (9).
Agreeing on definitions for the NSF diagnosis is not easy; in 2011 a grid for clinical and histopathological diagnosis of NSF was proposed, 14 years after the syndrome was first identified in clinical practice (11). Finally, it was generally thought for several years that gadolinium-based agents as a class were safe. As Caravan et al. (12) wrote in 1999: “In the rushed world of modern medicine, radiologists, technicians, and nurses often refrain from calling the agents by their brand names, preferring instead the affectionate ‘gado’. They trust this clear, odorless ‘magnetic light’, one of the safest classes of drugs ever developed. Aside from the cost ($50–80/bottle), asking the nurse to ‘Give him some gado’ is as easy as starting a saline drip or obtaining a blood sample.” In such an environment it is difficult to identify a causal relationship between a gadolinium-based contrast agent and NSF without substantial evidence. It requires some careful engagement by clinicians.
What are the clinical facts?
Nearly all of the clinical knowledge about NSF was available within 2 years following the seminar reports of Grobner and Marckman et al. (1, 2, 13). Nearly nothing new has been reported since 2008. Severely reduced renal function and gadolinium-based contrast medium particularly gadodiamide turned out to be mandatory factors. It was also clear that a third factor should be present, the clinical evidence of a specific or some factors is not there and we will probably never know. Among the proposed factors are erythropoietin (EPO), metabolic acidosis, anion gap, iron, increased phosphate, zinc loss (transmetallation), inflammation, and ACE inhibitors. We never get clinical evidence for the third factor.
Has NSF been erased?
FDA has received more than 1600 reports of NSF; all reported patients had an injection of gadolinium-based contrast media before 2010 (13). About 600 unconfounded cases (i.e. cases that received only one specific brand of a gadolinium-based contrast agent) have gone to court. A total of 711 NSF cases have been published (14). These cases probably represent the tip of the iceberg. Many patients may have died from 1996 through 2006 without having an NSF diagnosis. A French group (15) studied more than 500 patients undergoing dialysis; 56% were exposed to a gadolinium-based contrast agent after 2009. They did not find a single case of NSF. The overwhelming majority received gadoterate, but six received a gadolinium-based contrast agent that was contraindicated according to the Summary of Product Characteristics. It could be that information about dialysis use was not included in the patient's record, the technician administered the wrong vial by mistake, or kidney function tests were not recorded or were incorrectly recorded. As long as the specific contraindicated agents are easily available on the shelf, mistakes will happen. If the mistake occurs, will the patient be closely followed and if skin changes occur, will this finding be reported to the health authorities? We do not know. As administration of gadodiamide, gadopentetate, and gadovertisamide in patients with reduced renal function has been contraindicated in Europe since 2007, the responsible radiologist will be in a difficult legal situation if the patient mistakenly receives this product and develops NSF. Thus it is premature to conclude that NSF has been totally erased. In the process of erasing NSF, introduction of guidelines (16) for use of gadolinium-based contrast agents have been beneficial.
What are the potential long-term effects?
NSF may occur from hours to years after exposure to gadolinium-based agents. It is unclear where the gadolinium is during a long latent period, but the bone is one potential reservoir (17). Gadolinium can replace calcium in the hydroxyapatite of bone and the remodeled bone has a slow turnover. Diseases like osteoporosis affect bone turnover; this could cause the release of retained gadolinium accumulated from several examinations over many years. Gadolinium is not found in the body of patients who never received a gadolinium-based contrast media. It can be found in bone of patients with normal renal function. Both in man and rodents more gadolinium is found in bone after exposure to a non-ionic linear chelate than a macrocyclic agent. The long-term effects of this finding of a toxic ion accumulating in bone are unknown. Will it stay in the bones forever? When released will it cause necrosis in the liver? Induce hematological diseases? One should never administer more contrast than necessary to reach a diagnostic result; an agent that leaves the smallest amount for the gadolinium ion in the body should be used. The warnings have once been ignored with drastic results. On the other hand the consequence of not using an agent may be even worse: overlooking a treatable disease. That is not good practice.