Abstract
Prevalence of nephrogenic systemic fibrosis
Most reports have been based on pathological or nephrological registers rather than the prospective examination of the skin of exposed patients. The prevalence of nephrogenic systemic fibrosis (NSF) after exposure to gadodiamide has been reported to be in the range of 3–7% in patients with reduced renal function (1). In patients with CKD 5 (GFR less than 15 mL/min 1.73 m) reviewed prospectively, it may be as much as 18% (2); all patients with suspicious lesions had a skin biopsy. The prevalence was higher after two or more injections (36%) than after a single injection (12%), indicating a cumulative effect (2). In Boston 30% of patients on dialysis had developed NSF based on a systematic examination of the patients in five dialysis centers, but skin biopsies were only taken in a few patients (3).
In the peer-reviewed literature only one center has reported a large number (>10) of NSF cases after gadopentetate dimeglumine (3), but many centers have reported more than 10 cases after gadodiamide (1). This difference is not just a reflection of the market share of the two products, because gadopentetate dimeglumine has been administered to as many as 4–5 times the number of patients who have had gadodiamide. In the Four American University Study, the overall incidence was 0.039% after gadodiamide and 0.003% after gadopentetate dimeglumine (4). The benchmark incidence of NSF was one in 2913 patients who underwent gadodiamide-enhanced magnetic resonance imaging (MRI) and one in 44,224 patients who underwent gadopentetate dimeglumine-enhanced MRI (P < 0.001). The study was based on patient records from databases of dermatology, pathology, internal medicine, nephrology, transplant surgery, and radiology departments and not on a systematic examination of patients with reduced renal function exposed to a gadolinium-based contrast agent.
By 1 February 2008, 190 cases (confirmed by biopsy and clinical examination) had been reported in the peer-reviewed literature: 157 had had gadodiamide, eight gadopentetate, three gadoversetamide, and in five no exposure could be verified. In 18, the agent could not be identified and four received several agents (5). Up to 2012, a total of 711 NSF cases were published (6), but it is likely that they are only a small percentage of all cases. The true number of patients who developed NSF is not known because only two studies included systematic inspection of the skin in patients with reduced renal failure or on dialysis, so patients may have died with or of undiagnosed NSF between 1996 and 2006. Simple review of patient records is not enough to determine the true prevalence.
Patients at risk of NSF
Patients with CKD 4 and 5 (GFR < 30 mL min−1), and those on hemo- or peritoneal dialysis have an increased risk of developing NSF after being exposed to a less stable gadolinium-based contrast agent (7,8). Patients at lower risk are those with CKD 3 (GFR 30–59 mL/min). To date, no cases of NSF have been reported in patients with normal renal function, CKD 1 and 2 (GFR > 60 mL/min 1.73 m), in the peer-reviewed literature. Patients with acute renal failure are at particular risk because the reduced renal function may be overlooked by a single determination of their estimated glomerular filtration rate (eGFR). If they receive a less stable gadolinium agent when they have low renal function and develop NSF, NSF does not disappear when the renal function improves (9).
The lessons of the last 8 years
Since the first 2–3 years after the link between gadolinium and the development of NSF was recognized, very little has been added to the four key clinical features of NSF: (i) exposure to a non-ionic linear gadolinium-based contrast agent (not necessarily in high doses, as NSF has occurred after 0.1 mmol/kg); (ii) renal insufficiency including treatment with dialysis; (iii) great variability in involvement, from a small skin plaque to generalized involvement of the skin and internal organs; and (iv) no effective treatment.
There is very likely to be a further unknown factor, or factors. Many co-factors have been suggested: high doses of erythropoietin (EPO), metabolic acidosis, iron and ferritin, chronic inflammation, hypercoagulability, thrombotic events, recent vascular surgery, recent renal transplant failure, recent surgery, anion gap, or increased phosphate. However, no universal co-factor apart from renal failure has been identified (10,11).
A few cases of NSF have been reported in patients who had only received gadobutrol, but there is uncertainty about the histopathologic changes (12–14). In 141 patients on long-term dialysis who had received gadoteridol, no case of NSF was found (15). Of 255 patients being treated with dialysis who had enhanced MRI with gadoterate, none developed NSF within 4 months of exposure (16). Today almost the same number of injections of one of the three macrocyclic agents has been given worldwide as of the non-ionic linear chelates. It is now apparent that the safety profile of the more stable agents is better than that of the less stable agents.
Over recent years, interest in NSF has decreased, as indicated by the number of publications indexed under nephrogenic systemic fibrosis in PubMed. After a peak in 2009 the number of papers is slowly decreasing (Fig. 1). Between 1 January 2000 and 31 December 2015, a total of 911 manuscripts have been published under the keyword “nephrogenic systemic fibrosis”.
Number of publications listed under “nephrogenic systemic fibrosis” per year in PubMed 2001–2014.
Has NSF been eliminated?
The FDA has received more than 1600 reports of NSF, all of patients who had an injection of gadolinium-based contrast medium before 2010 (17). According to Pirovano et al., 711 NSF cases had been published in the peer-reviewed literature up to the beginning of 2012 (6). These cases probably represent only the top of the iceberg. Many patients with NSF may have died between 1996 and 2006 without their diagnosis being recognized.
In a French study of more than 500 patients undergoing dialysis, 56% received a gadolinium-based contrast agent after 2009 (16), and no cases of NSF were found. Six patients had received a lower stability gadolinium-based contrast agent that was contraindicated according to the Summary of Product Characteristics. Possible explanations are: (i) that information about dialysis treatment was not included on the request for the examination; (ii) that the technician administered the wrong vial by mistake; or (iii) that kidney function tests were not recorded or were incorrectly recorded. If agents which are contraindicated in some patients are available in a department, mistakes will happen. Such errors may not be recognized, but, if they are, will any skin changes which occur be reported to the health authorities? The administration of gadodiamide, gadopentetate, and gadoversetamide in patients with reduced renal function has been contraindicated in Europe since 2007, so the radiologist responsible will be in a difficult legal situation if the patient mistakenly receives a contra-indicated product and develops NSF (18).
ESUR guidelines on nephrogenic systemic fibrosis – from ESUR Guidelines on Contrast Media, version 9.0, Vienna, Austria 2014 (19).
Confounded cases: If two different Gadolinium-based contrast emdia have been injected, it is impossible to determine with certainty which agent triggered the development of NSF and the situation is described as “confounded”; Unconfounded cases: The patient has never been exposed to more than one agent.
Gadolinium and patients with normal renal function
Gadolinium has been demonstrated in the skin of patients with NSF (20,21) and in the bone of patients who have received a gadolinium-based contrast agent but have not developed NSF (22). The amount of gadolinium in the skin of patients with NSF seems to increase up to 3 years after the last exposure to a gadolinium-based contrast agent (17). Where does it come from? In view of the low turnover of bone, this is the likely source. The long-term implications of gadolinium deposition in bone, which is likely to occur more commonly in pediatric patients because of the more active bone creation in this population, have yet to be determined (4).
Another risk group could be patients who undergo multiple enhanced MRI examinations, for example, women with an increased risk of breast cancer who may follow recommendations to undergo annual enhanced MRI. After each examination some gadolinium will accumulate in the bone and will remain there for many years (23). What will happen when the gadolinium is released from bone, for example, when osteoporosis increases bone turnover? The release of an overload of gadolinium might cause classic toxicity symptoms, not NSF. Free gadolinium may produce liver necrosis, obstruct calcium ion passage through muscle cells, and interfere with intracellular enzymes and cell membranes by the process of transmetallation, when Gd3+ replaces endogenous metals such as zinc and copper.
Another possible risk group is patients with diabetes mellitus. After about 10 years, 50% may develop diabetic nephropathy, but these patients cannot be identified before their renal function deteriorates. If some gadolinium remains in the body from an MR examination when the patient had normal renal function, it is possible that this patient will develop more severe NSF after contrast-enhanced MRI with a less stable agent when their GFR becomes severely reduced. In one study, independent of renal function, patients with severe NSF had had a higher lifetime dose of the gadolinium-based agent than those developing non-severe NSF (24).
The safety of multiple injections of gadolinium-based contrast agents has never been studied; the phase I–III studies that led to approval of these agents by the health authorities included only a single injection in man and in most cases with a dose of 0.1 mmol/kg body weight. Since then there has been a major change in the use of enhanced MRI. The chance of having several enhanced MRI examinations within a shorter period is much higher now than 15 years ago when there were only 1000–2000 MRI units worldwide and each had a lower throughput. The phase I–III studies did not foresee these changes and therefore cannot be used to document the safety of these products as they are now used.
What should be done?
In practice, it is much easier to use one of the more stable gadolinium agents, for which eGFR measurement before administration is not mandatory in all patients. No differences in diagnostic efficacy have been demonstrated among the six extracellular agents (25).
There are several conditions where alternative imaging is diagnostically inferior and cannot replace enhanced MRI. The risk of the NSF is low if the non-ionic linear chelates are avoided, and if the most stable agents are used in the smallest dose consistent with a diagnostic result in at-risk patients. The risk of complications (procedural, allergy-like reactions, contrast-induced nephropathy, radiation) following conventional or computed tomography (CT) arteriography with iodine-based contrast medium must be weighed carefully against performing MR using a stable gadolinium agent. In most cases there is no better alternative to contrast-enhanced MRI (26,27).
Despite the American College of Radiology recommendation that hemodialysis should be used in at- risk patients (28), starting dialysis should be considered with care, since the morbidity associated with hemodialysis in a patient not already adjusted to hemodialysis is higher than the risk of NSF after exposure to a macrocyclic gadolinium agent (29).
Legal actions
In 2007, the European Medicines Agency (EMA) contraindicated the use of gadovertisamide, gadodiamide, and gadopentetate in patients with reduced renal function (<30 mL/min/173 m2) or on dialysis. The US-based Food and Drug Administration (FDA) followed EMA in 2010. The other agents may be used with caution. Regulations have not been changed since 2010 (30).
Conclusion
NSF is an important very late adverse reaction to lower stability gadolinium-based contrast agents, which occurs in patients with impaired renal function (31). The recognition of this reaction to agents previously considered to be very safe emphasizes the need to have a good clinical indication for all enhanced MRI examinations, to choose an agent that leaves the smallest amount of gadolinium in the body, and to keep complete records of the type and dose of agent given. On the other hand, one should never deny a patient with reduced renal function a clinically well indicated enhanced MRI; a stable agent (low or intermediate risk agent of causing NSF) should always be used.
Research into the etiology of NSF has drawn attention to the retention of gadolinium in the body tissues long after an enhanced MR examination, even in patients with normal renal function. The safety implications of this are as yet unclear and further research is necessary.