Nephrogenic systemic fibrosis: is any contrast safe in renal failure?

Contrast-induced nephropathy (CIN) has an incidence of between 2 and 15% depending on the definition of acute kidney injury (AKI), underlying renal insufficiency and the presence or absence of diabetes.1,2 Magnetic resonance imaging (MRI) and the use of iodine-free contrast agents appeared to have solved the problem of CIN in these high-risk groups. Identification of a new and potentially fatal fibrosing disorder, nephrogenic systemic fibrosis (NSF), triggered by gadolinium (Gd) exposure during MRI in patients with advanced renal insufficiency has cast doubt on this solution. NSF was first observed in 19973 and is absent in samples from two tissue repositories before that time, suggesting it is a new disorder.4 NSF is characterized by thickening and hardening of the skin overlying the extremities and trunk. Originally named Nephrogenic Fibrosing Dermopathy (NFD), involvement of skeletal muscle, fascia, lungs, liver, testes and myocardium in some patients5 has led to NSF as the preferred terminology.6 More than 215 cases of NSF have been reported to the International NFD/NSF Registry at Yale University7 and the US Food and Drug Administration (FDA) have received 90 reports of patients with moderate to end-stage renal disease who have developed FD/NSF.8 NFS occurs exclusively in patients with kidney failure, with 90% observed in haemodialysis and peritoneal dialysis patients and the remaining 10% in patients with failing renal allografts, advanced chronic kidney disease (chronic kidney disease, CKD 5, GFR <15 μm/min) or in the presence of AKI.7 Apart from restriction to patients with kidney failure, the unifying link between reported cases is exposure to Gd-based contrast agents used during MRI and in highest dose in MR angiography. All cases of NSF with available data reported to the International Registry had been exposed to Gd, more than 95% within 2–3 months of the disease onset and some within 2 days of exposure.4 Some early reported cases appeared to develop NSF following high doses of Omniscan (GE Healhcare AS, Oslo, Norway) (gadodiamide) or OptiMARK (Mallinckrodt, St Louis, MO, USA) (gadoversetamide)9,10 with an incidence as high as 55%, that is, 5 of 9 exposed patients.9 Multiple contrast exposure has been suggested as a risk factor.11 However, other reports have established no correlation to dose and severity.10 A small retrospective population-based study found an incidence of 4.3 cases per 1000 patient years and estimated that the risk of NSF was 2.4% for each radiologic exposure to Gd.6 As NSF/NFD have been reported after at least three different Gd-based contrast agents in the USA the FDA believe that the potential for NSF/NFD is present with all approved agents. Cases have been reported in the USA, Europe, Asia and now in Australia. The first Australian case reported by Cheung in the Internal Medicine Journal underwent MRI imaging for discitis in 2002, receiving 5 mmol (2.87 g) of gadodiamide (Omniscan) MR contrast agent and developed a ‘classical’ pattern with skin thickening over the hands and feet leading to flexion contractures.12 The syndrome developed 8 months after exposure.13 Interestingly, this patient had a pro-coagulopathic background before MRI, which has also been previously observed in some patients preceding development of symptoms.14 Fibrotic lesions may extend deeply into the s.c. tissue. Microscopy appearances range from subtle proliferation of dermal fibrocytes to marked thickening of the dermis with florid proliferation of fibrocytes with long dendritic processes, often associated with histiocytes and stelite factor XIIIa plus dermal dendritic cells. Immunohistochemical staining shows abundant CD34+ dermal cells, the dendritic process aligned with elastic fibres around collagen bundles of dense network. Detection of Gd in the dermis has been reported in five patients.15,16 An association with a prior pro-coagulopathic state or with high dose with recombinant erythropoietin (EPO) therapy has also been reported.14 This may reflect ligand dissociation due to transmetallization or the possibility that chronic inflammation and endothelial dysfunction, which produce resistance to EPO, also increase the propensity for extravascular Gd accumulation.11 Gd belongs to the lanthanide series of elements. The toxicity of free lanthanides including both experimental and clinical AKI nephrotoxicity is well known.17,18 The safety of Gd-based contrast agents relies on tight binding of Gd to an organic chelating ligand, which forms a stable complex around Gd3+. Chelate structure (macrocyclic vs linear) and charge (ionic vs non-ionic) are used to categorize Gd contrast agents. The macrocyclic chelates (such as gadoteric acid (Gd-DOTA; Dotarem, Guerbet, France) bind Gd3+ more tightly than linear chelates, such as gadodiamide (Omniscan) or gadopentetate dimeglumine (Magnevist, Bayer Healthcare Phamaceuticals, Wuppertal, Germany). The macrocyclic chelates are more stable in vivo and in vitro, with much lower dissociation rates.19 As Gd3+ contrast is eliminated almost entirely by the kidney, decreased renal function reduces clearance by the kidney.20 Gd-chelate dissociation is also enhanced by metabolic acidosis in renal failure.11 Transmetallization (displacing Gd3+ by competitive binding) with Zn2+, Cu2+, Ca2+ and Fe3+ may also reduce Gd–ligand complex stability leading to dissociation.21 Tissue deposition of Gd may then promote infiltration of affected tissues by circulating fibrocytes or CD68+/XIII+ dendritic cells, either pathway then promoting fibrosis.22 The differential diagnosis is systemic sclerosis (but Raynaud’s phenomenon is absent in NSF), scleromyxoedema (but sparing of the head and the absence of monoclonal gammopathy in NSF are distinguishing) and finally eosinophilic fasciitis (however, the hands and feet are usually spared in this condition in contrast to NSF where they are frequently involved). The condition appears to have an unremitting course in most patients and improvement or remission has only been described in patients who have recovered renal function including that following successful renal transplantation.10,23–25 As there is no proven therapy, prevention is essential. The question is, how to translate this into clinical practice? The FDA now recommends that patients with advanced kidney failure avoid Gd-containing contrast agents. These are defined as requiring dialysis or with GFR <15 mL/min. The FDA also advises that ‘it may be prudent to initiate prompt dialysis in patients with advanced kidney dysfunction who receive a gadolinium contrast MRA’.8 Radiologists have responded promptly to this advice with broader recommendations, including advising patients with both stages 4 and 5 CKD (i.e. GFR <30 mL/min) of the risks when obtaining informed consent.26 Some also advise planning for up to three haemodialyses in patients with CKD 5, who are exposed to Gd contrast, even when such patients have not commenced dialysis. In my hospital, radiologists have switched to more stable ionic linear ligands, such as Gd-BOPTA (Multihance, Altana Pharma AG, Singen, Germany), which also have the probable advantage of reducing the volume of Gd contrast needed to maximize relaxivity (and hence signal to noise and contrast). Clearly, these guidelines are problematic in patients with advanced kidney disease as this group of patients is characterized by extensive cardiovascular disease with young patients commencing dialysis having a fivefold higher risk of myocardial infarction than age-matched controls.27 Some groups prefer the risk of radiocontrast nephropathy with iodinated contrast media in at risk patients with GFR <30 mL/min, as most patients with GFR >30 mL/min rarely develop irreversible injury after contrast nephropathy.28 However, given the current concern that even small increases in plasma creatinine are an independent risk factor for death29 and that CIN of any degree is associated with long-term mortality,30 there is no ideal preventive strategy for the very CKD patients at highest risk for cardiovascular disease and greatest need for contrast angiographic studies. Newer techniques for binding Gd, which use nanotechnology to generate particles of known size, for example, dendrimers, are a significant advance in contrast imaging. These ‘designer particles’ allow targeting of vascular, interstitial and cellular compartments including proximal tubular cells.31 However, the in vivo safety of these agents needs to be shown and the use of Gd will remain a stumbling block if significant dissociation of Gd occurs into tissues or cells. An alternative contrast agent for MRI/MRA is super paramagnetic iron oxide (SPIO) particles, which are crystalline cores covered with dextran. At present, even the smallest of these, the ‘ultra small’ or USPIO, are too large to be filtered by glomeruli and are progressively phagocytosed by specialized cells from the monocyte–macrophage system. Because of the resulting pharmacokinetics, this group of contrast agents is devoted to perfusion studies and to labelling cells in vivo or in vitro.31 The emergence of NSF stresses our need for safe techniques for enhancing MR images in patients with renal disease. Given that one in ten in the population has some degree of kidney failure, this is both important and urgent. Critically, local guidelines need to be developed urgently with the help of nephrologists, as their patients are the susceptible ones. Recommendations such as those suggested in this issue of the journal need wide consensus and rapid dissemination.32 We also need to start collecting the evidence that will define the risk to the wider community of patients with CKD 4, that is, GFR <30 mL/min, many of whom may have not yet come to nephrological attention.