Secondary Logo

Journal Logo

Original Article

A Preclinical Study to Investigate the Development of Nephrogenic Systemic Fibrosis: A Possible Role for Gadolinium-Based Contrast Media

Sieber, Martin A. PhD*‡; Pietsch, Hubertus PhD*‡; Walter, Jakob PhD*; Haider, Wolfram PhD; Frenzel, Thomas PhD*; Weinmann, Hanns-Joachim PhD*

Author Information
doi: 10.1097/RLI.0b013e31815e6277
  • Free


Nephrogenic Systemic Fibrosis (NSF) was first described in the medical literature in 2000 (however, the actual term “Nephrogenic Systemic Fibrosis” was introduced by Cowper much later), with the first reported case dating back to 1997.1 The disease is seen in patients with severe renal insufficiency, many of whom receive hemodialysis or peritoneal dialysis.2 Male and female patients are affected in approximately equal numbers and are typically in middle age.3,4 As of May 2007, over 200 cases of NSF have been reported.5 NSF causes fibrosis of the skin and connective tissues. The dermal lesions of NSF are usually distributed symmetrically, developing on the trunk and extremities but sparing the face. The following signs have been described: cutaneous thickening of the skin with or without pigment alterations, swelling (mostly of the lower extremities), progressive skin indurations comparable to “peau d'orange,” redness, pruritus, and burning sensations. The associated skin thickening may also cause joint contractures. On the cellular level, lesions are accompanied by an increase in fibroblast-like cells, dermal macrophages, and fibrocytes positive for CD34 and procollagen.6–8 Additionally, a considerable increase in collagen bundles and, in some cases, deposits of mucin are observed in skin biopsies of NSF patients.9,10

The clinical course of NSF can be progressive, involving internal organs, and may even have a fatal outcome. Currently, there is no known treatment for NSF. Improving renal function seems to slow down or arrest NSF and may result in a gradual reversal of signs and signs.2,11,12 Although the etiology of NSF is currently unknown, it is most likely multifactorial and has recently been linked to intravenous (i.v.) gadolinium (Gd)-containing contrast agents.13,14 Since the beginning of 2006 there has been increased awareness of NSF following reports in patients with severe renal impairment who had undergone a magnetic resonance (MR) scan using an i.v. Gd-containing contrast agent. The majority of cases reported in the literature to date relate to the use of 1 agent in particular, Omniscan (gadodiamide), but there is also a number of reports with regard to OptiMARK (gadoversetamide) and Magnevist (gadopentetate dimeglumine), and isolated reports of confounded and unconfounded cases with other agents. Regulatory actions up to contraindications in severely renally impaired patients have already been undertaken; however, the situation is constantly evolving. Up-to-date information on the number of reported cases for individual contrast agents and the regulatory status can be obtained on the websites of regulatory agencies, such as the FDA.15

Gd-containing contrast agents used for MR imaging (MRI) contain the paramagnetic Gd3+ ion tightly bound by a complexing agent, and are generally well tolerated16–18). Gd-containing contrast agents fall into 2 distinct classes, the linear, open-chain agents and the macrocyclic agents.19 The stability of the linear, open-chain agents, such as Omniscan and Magnevist is characterized by the thermodynamic stability constant, log Ktherm, and the conditional stability constant, log Kcond (Table 1). There are significant stability differences within the linear, open-chain compounds (in the context of MR contrast media the terms, “Omniscan” and “gadodiamide” are usually used as synonyms. However, the marketed formulation of Omniscan contains gadodiamide and 5% of excess ligand caldiamide. From now on in this study, to distinguish between the marketed formulation with excess ligand and the pure drug substance gadodiamide without excess ligand, we shall use the term “Omniscan” if the marketed product is meant, and the term “gadodiamide without caldiamide” if the pure drug substance is meant. For consistency, for all other treatments, marketed names will be used throughout). The nonionic compounds, such as Omniscan and OptiMARK, have a substantially lower complex stability compared with ionic compounds, such as Magnevist or Multihance.17,18,21 Macrocyclic Gd chelates are more kinetically inert to Gd dissociation than linear Gd chelates20–22 and for this reason log Ktherm and log Kcond are not relevant, as kinetic inertness is characterized by a different constant (kobs).20

Physicochemical Properties and Estimated Administration of Conventional Gadolinium-Containing Linear Chelates20

Most linear, open-chain, Gd-based contrast agents contain some amount of excess ligand to reduce the probability of releasing Gd3+ ions, known to be toxic, into the body.21,23 The less stable nonionic linear agents use considerably more excess ligand in their formulation: Omniscan contains about 12 mg/mL excess ligand (∼5%; in this case caldiamide). In the more stable Magnevist, only up to 0.4 mg/mL of excess ligand (∼0.1%) is present. There are limits to the extent that Gd-containing agents can be prepared with excess ligand, as the ligand itself can cause adverse effects in humans as well.24 Unchelated Gd3+ form aggregates with ions such as OH and PO43−, or binds to proteins and is practically insoluble at pH 7. The major targets of Gd3+, when injected as GdCl3, are the spleen, liver, and bone. GdCl3 is also used, experimentally, to deplete macrophages.25,26 Although Gd3+ is toxic, several milligrams of GdCl are required to elicit chronic toxic effects in rats.27 It has been speculated that Gd-containing contrast agents may elicit their toxic effects by a direct loss of the Gd3+ ion or through transmetalation—the exchange of Gd3+ with endogenous ions. The latter may result in a depletion of essential trace elements, such as Zn2+.28–32

The pathomechanisms of NSF are not yet fully understood and, so far, NSF has only been observed in patients with severe renal impairment.33 A possible reason for this could be the prolonged retention of Gd-containing contrast agents in this patient cohort. The majority of the marketed Gd-containing MRI contrast agents are primarily excreted by glomerular filtration. In patients with normal renal function, the agents have an elimination half-life of about 1.5 hours. Within 24 hours, more than 95% is excreted from the body.34–36 In patients with severe renal dysfunction, however, the agents’ elimination half-life is considerably prolonged37 (see column 2 in Table 2). Thus, without dialysis, very high plasma levels of Gd chelates can be maintained for hours and even days due to the lack of an alternative elimination pathway, resulting in a much longer exposure of the body to Gd chelates compared with patients with normal renal function.39

Pharmacokinetic Parameters of Extracelluar Gadolinium Chelates in Humans and the Rat Model

We have conducted preclinical in vivo studies in rats to establish an experimental setting to investigate the possible link between NSF and Gd-containing MRI contrast agents. As no adequate preclinical animal model for end-stage renal disease (ESRD) exists, multiple i.v. injections of Gd-containing MRI contrast agents were used in an attempt to replicate the high level of exposure to the i.v. injection of Gd-containing MRI contrast agents seen in patients with severe renal insufficiency before hemodialysis. Specifically, this pioneering study aims to elucidate the role of unchelated Gd in NSF and to test the hypothesis that the depletion of endogenous ions can have a causative role in NSF. Finally, we intended to test the hypothesis that the stability of MRI contrast agents is an essential factor in relation to NSF by investigating 2 widely used contrast agents with significantly different thermodynamic stabilities.


Tested Compounds

Omniscan and Magnevist were obtained from GE Healthcare and Bayer Schering Pharma AG, respectively. Gadodiamide without caldiamide was prepared from commercially available Omniscan (including 5% caldiamide) by replacing the calcium ions bound to caldiamide with Gd ions. This competing complexation reduced the amount of caldiamide to less than 0.1%. Caldiamide has been synthesized from DTPA-bisanhydride and methylamine followed by complexation with calcium hydroxide and sodium hydroxide.41

We used Gd-ethylenediamine tetraacetic acid (Gd-EDTA) as a positive control because it is known to release Gd3+ ions to a high extent in the body.42 Gd-EDTA was synthesized by complexing Gd acetate (SIGMA Aldrich) with EDTA (SIGMA Aldrich) at the Chemical Department of Bayer Schering Pharma AG. The pH of Gd-EDTA was adjusted to pH 7. We used an isotonic saline (0.9%) solution as negative control.


A total of 35 male Hannover Wistar Rats, each weighing ∼190 to 240 g, were obtained from Charles River (Sulzfeld, Germany). Six animals per group were selected at random and housed with a 12-hour day-and-night rhythm. The animals were given water ad libitum and were fed a standard diet (Sniff R/M-H from Sniff, Germany). Animals were treated in accordance with the German Law on the Protection of Animals and with permission from the State Animal Welfare Committee.

Study Design

To simulate the prolonged exposure to MRI contrast agents of patients with severe renal insufficiency, the rats received multiple i.v. injections of the tested compounds. Thirty-five rats were randomly allocated to 1 of 6 treatment groups (6 rats per treatment group). Animals were dosed with Magnevist, Omniscan, or gadodiamide without caldiamide at a concentration of 0.5 mol/L. The compounds were injected into a tail vein daily, 5 times per week (not on weekends) at a dose of 2.5 mmol Gd/kg body weight over a period of 4 weeks. Control animals received the same volume of i.v. saline (0.9% weight per volume). Two treatment groups also received i.v. caldiamide and Gd-EDTA. Owing to the lower LD50, we applied both substances in reduced doses: caldiamide and Gd-EDTA were administered 5 times per week at doses of 0.5 mmol/kg (0.5 mol/L) and 0.1 mmol/kg (0.095 mol/L), respectively. The protocol planned for 20 injections per animal for all groups. To ensure that the administered Gd compounds were completely eliminated from the body, a sufficient time period, covering the respective serum elimination half-lives, had to elapse before the animals were killed. We set this time period to 5 days. The number of injections and the tested doses are listed in Table 3.

Survey of Compounds Tested, Number of Injections, and Applied Doses

Macroscopic Skin Findings and Histology

The animals were checked for macroscopic skin changes such as reddening, fur loss, scab formation, and ulceration before the first injection and then daily during the experiment. At the end of the experiment, the backs of the animals were shaved for better visualization of the skin findings.

Skin samples were taken from the animals immediately after necropsy. Three skins probes were taken from the back of every animal. The samples were taken from similar areas in each animal. The probes were fixed in 4% neutral buffered formalin. After fixation and routine dehydration, all tissue samples were embedded in paraffin and sectioned (4–6 μm) for hematoxylin and eosin and immunohistochemical staining. The histopathologic reading was performed by 2 veterinary histopathologists (European College of Veterinary Pathologists certified) in a blinded fashion according to best practice guideline of the Society of Toxicologic Pathology.43 In general, structure and classification follow the logical framework described elsewhere.44,45 The severity of the findings was graded using scores from 1 to 5 based on the increases in cellularity (namely increases in fibroblast-like cells, dendritic-like cells, and lymphocytes): 1 = minimal; 2 = slight; 3 = moderate; 4 = marked; 5 = massive.


Immunohistochemical analysis was performed with a polyclonal rabbit anti-CD34 antibody (C18; sc-7045 Santa Cruz) using the alkaline phosphatase, antialkaline phosphatase (APAAP) method. The sections were developed in new fuchsin-naphthol (Sigma, St. Louis, MO) and counterstained with hematoxylin. In each section, staining was assessed on 10 adjacent fields at 400× magnification. Two sections, taken from 3 animals per group, were analyzed at 2 different magnifications.

Gd and Zinc Measurements

The Gd and Zn concentrations in the skin (taken from the back, next to the histology probes) and in the liver and femur were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) at a wavelength of 342.247 nm and 206.200 nm, respectively (ICP-AES, IRIS Advantage, Thermo).46 The method does not, however, distinguish between chelated or unchelated Gd or Zn, and accordingly the Gd and Zn levels obtained with this method may be composed of both chelated and unchelated ions. For the measurement of the Gd and Zn concentrations in the femur, we analyzed samples of the femur that included the bone marrow. The limit of detection (LoD) for the Gd measurement in the skin and femur was 7 nmol/g tissue, and in the liver the LoD was 3 nmol/g tissue. For the Zn concentration, the LoD in the skin and femur was 14 nmol/g tissue, and in the liver it was 4 nmol/g tissue. Gd and Zn measurements were performed for each animal taking 1 sample per organ.

Transmission Electron Microscopy and Energy Dispersive x-ray Analysis

For transmission electron microscopy (TEM), the probes were prepared as described elsewhere.47 In brief, the samples were fixed with 2.5% glutaraldehyde in 0.1 M Na-cacodylate buffer (containing 0.05% CaCl2), pH 7.4, for 24 hours at 4°C. The probes were secondarily fixed in osmium tetroxide/Na-cacodylate buffer for 4 hours at 4°C and then stained with 1% uranyl acetate in 0.05 M maleate-buffered solution, pH 5.2, for 1 hour at 4°C. The probes were subsequently dehydrated in graded ethanol series, infiltrated, embedded in Spurr epoxy resin, and polymerized for 24 hours at 70°C. The sections were sampled on uncoated 300 mesh thin-bar grids.

TEM was performed on Philips CM 300 UT with an acceleration voltage of 200 kV. As a test holder, a standard Philips double tilting holder was used. For element specific analysis of the observed skin depositions, we used energy dispersive x-ray (EDX) analysis. The EDAX DX 4 system was used with (Li) Si crystal 10 mm2 and a resolution of 140 eV. The probes were investigated with a tilting corner of 20 degree in the Nanoprobe mode.

Statistical Evaluation

Gd measurements were evaluated for the skin, femur, and liver (excluding saline control animals, where no Gd was detected) and tested globally for differences in medians between at least 2 groups with a nonparametric Kruskal-Wallis test. In case of significance at a two-sided 5% level of significance, pairwise comparisons were performed using a nonparametric multicomparison Kruskal-Wallis test (Dunn test) to ensure an experimentwise two-sided level of significance of 5%.


Model for Gd Exposure

Based on the pharmacokinetic data after a single injection of contrast agent, the chosen dosing regimen resulted in an extrapolated systemic exposure in rats similar to that seen after application of twice a triple dose of contrast agent in severely renally impaired patients (Table 2).

Macroscopic and Microscopic Findings and Histology

Macroscopic changes of the skin, such as reddening, fur loss, scab formation and ulceration, were observed in animals i.v. injected with Gd-EDTA, gadodiamide without caldiamide, and Omniscan (Figs. 1 and 2, observed after 8 and 16 days).

Observed macroscopic skin lesions after treatment with Omniscan, Magnevist, gadodiamide without caldiamide, Gd-EDTA, caldiamide, and saline. Each day of the experiment is depicted as a box (treatment-free weekends were omitted). Days in which no skin lesions in a group were observed are marked by white boxes; days in which small skin lesions in a group were observed are marked by pink boxes; days in which some animals in a group showed substantial skin defects are marked by blue and red striped boxes; days in which all animals showed substantial skin defects are marked by red boxes. The total number of animals and the number of animals showing skin defects are shown on the right side.
Gross lesions in the skin of animals 5 days after the last injection of the respective contrast media. Overview of the backs of animals. One representative animal per group is shown. For better visibility of the skin lesions the animals were shaved. Skin lesions were observed in the animals treated with (A) Gd-EDTA, (B) gadodiamide without caldiamide, and (C) Omniscan.

Animals receiving gadodiamide without caldiamide had to be killed 5 days after the 10th injection because of the severity of the skin lesions. There were no macroscopic changes seen in animals treated with Magnevist, caldiamide, and saline.

Histopathologic findings were noted for Gd-EDTA, gadodiamide without caldiamide, and Omniscan, but not for Magnevist, caldiamide, or saline (Table 4). Minimal-to-moderate increases in cellularity (namely an increase in fibroblast-like cells, dendritic-like cells, and lymphocytes) and fibrosis were observed in the dermis of animals after treatment with Gd-EDTA, gadodiamide without caldiamide, and Omniscan, resembling the NSF findings seen in humans. None of the other compounds (Magnevist, caldiamide, and saline) induced such lesions in any of the examined animals (Fig. 3). [Bayer Schering Pharma has retained an independent company to select 5 external veterinary and human pathologists to serve as members of a Pathology Working Group (PWG). The PWG conducted a blinded review of skin specimens from animals treated in certain of our preclinical studies. The results will be published in due course.]

Overview of the Observed Macroscopic and Microscopic Findings in the Dermis 5 Days After the Last Injection
Fibrosing dermopathy in rat skin. Hematoxylin and eosin stain of skin probes taken 5 days after the last injection of the respective contrast media. Images (A–F) (upper row) shown with ×5 magnification, whereas images (G–L) (lower row) show ×20 magnification of the dermis. Fibrosis of the dermis was observed in the animals treated with (A/G) Gd-EDTA, (B/H) gadodiamide without caldiamide, and (C/I) Omniscan.


The highest CD34 positive expression was observed in positive mesenchymal spindle-shaped cells (most likely fibroblasts) in the gadodiamide without caldiamide group, in which macroscopic and microscopic skin lesions were also observed. CD34 positive cells were present in clusters, preferentially in the superficial dermis. CD34 fibroblasts were also clearly evident in the Omniscan group in the superficial dermis, especially in multifocal clusters and in the Gd-EDTA animals (Fig. 4). No CD34 positive mesenchymal spindle-shaped cells were observed in the Magnevist, saline, and caldiamide groups (Fig. 4).

CD34 positive cells in the skin of rats. Immunohistology with an antibody specific for CD34 positive cells on paraffin sections. The skin probes were taken 5 days after the last injection of the respective contrast media. CD34 positive cells were found in the dermis of animals treated with gadodiamide without caldiamide, Gd-EDTA, and Omniscan.

Gd Levels in the Skin

High concentrations of Gd were measured in the skin probes from animals treated with Gd-EDTA, gadodiamide without caldiamide, and Omniscan (Fig. 5A). In animals treated with 10 injections of gadodiamide (the animals had to be killed after the 10th injection because of the severity of the observed skin lesions), a Gd concentration of 2.2 ± 0.2 μmol/g was present in the skin. Similar concentrations were quantified after 20 injections of Omniscan (1.7 ± 0.2 μmol/g) and Gd-EDTA (1.7 ± 0.1 μmol/g; injected at a lower dose of just 0.1 mmol/kg body weight). The observed Gd concentrations in animals treated with 20 injections of Magnevist were significantly lower (0.21 ± 0.1 μmol/g). No Gd in the skin was measured within the detection limit of the ICP-AES analysis in the caldiamide and saline control groups.

Gadolinium concentrations in the skin, liver, and femur. Gadolinium concentration of the skin, liver, and femur 5 days after the last injection of the respective compound. Standard deviations are indicated.

Gd Levels in the Liver and Femur

High Gd concentrations were observed in the liver of rats treated with 20 injections of Gd-EDTA (2.5 ± 0.2 μmol/g) (Fig. 5). After injection of gadodiamide without caldiamide, the concentration was even higher, amounting to 2.9 ± 0.2 μmol/g. After 20 injections of Omniscan, a Gd concentration of 0.4 ± 0.1 μmol/g was measured. In rats treated with 20 injections of Magnevist, a Gd concentration of 0.2 ± 0.02 μmol/g was observed (Fig. 5). The Gd concentrations in the liver of rats from the caldiamide and the saline controls were below the LoD by ICP-AES analysis.

In the femur, a Gd concentration of 2.3 ± 0.02 μmol/g was measured after Gd-EDTA. Gd concentrations of 0.6 ± 0.1 μmol/g were measured after 10 injections of gadodiamide without caldiamide (Fig. 5). In animals treated with Omniscan and Magnevist, the concentration of Gd observed in the femur was 0.6 ± 0.1 μmol/g and 0.2 ± 0.1 μmol/g, respectively (Fig. 5). No Gd was measured within the LoD of the ICP-AES analysis in the femur of the rats from the caldiamide and the saline control groups.

From one of the Omniscan-treated animals, which had shown severe skin lesions, a skin sample was investigated by electron microscopy (Fig. 6). In the dermis, it showed dark, electron-dense deposits in the vacuoles of macrophage-like cells. The EDX spectrum of those deposits showed 2 characteristic peaks for Gd at 1.1 m/z and 6.0 m/z (Fig. 6, arrows). The EDX spectrum in those areas also showed a characteristic peak for phosphorus at 2.0 m/z (Fig. 6, arrowhead).

Electron microscopy of the dermis. Animals received 20 injections of Omniscan. Two corresponding EDX spectra from the same area are marked by circles (yellow area with electron dense deposit; green area without deposits served as a control). The spectra demonstrate characteristic peaks indicative for gadolinium (arrows) and phosphorous (arrowhead) emitted from the deposits. EDX = energy dispersive x-ray.

Zinc in the Skin, Femur, and Liver

No significant differences in the concentration of Zn in the skin (each about 0.2 μmol/g), the femur (about 18 μmol/g each), and the liver (about 0.6 μmol/g) were observed between all treatment groups, including the saline control (Fig. 7).

Zinc concentrations in the skin, liver, and femur. Zinc concentration of the skin, liver, and femur 5 days after the last injection of the respective compound. Standard deviations are indicated.


After recent concerns surrounding the development of NSF in patients with severe renal impairment and the association with the use of Gd-containing contrast agents in MR examinations,13,14,48 we established a preclinical dosing regimen in rats to investigate this further.

The chosen multiple injection i.v. scheme results in prolonged systemic exposure of the organism to the contrast agent, which resembles the clinical situation found in patients with severe renal impairment after an application of twice a triple dose of contrast agent (Table 2), and thus covers the prolonged systemic exposure to Gd-containing contrast agents in human patients with severely impaired renal function. Using this model, we observed, in rats, macroscopic lesions, microscopic changes, and histopathologic alterations correlating highly to findings in human NSF patients. Rats treated with Gd-EDTA, gadodiamide without caldiamide, and Omniscan showed macroscopic skin lesions, increased dermal cellularity composed of dendritic-like cells, and spindle-shaped fibrocytes (probably fibroblasts).6–8 We also observed an infiltration of CD34 positive fibrocytes into the dermis as seen in human NSF patients.49 In humans, the pathologic dermal lesions deeply infiltrate the subcutis, whereas in rats the lesions were noted in the dermis only. This might be caused by the panniculus carnosus that separates the dermis from the subcutis in rats, an anatomic structure missing in humans.

Cases of NSF reported to date have been associated with the use of Gd-containing contrast agents in patients with severe renal failure, and in particular with 1 agent, Omniscan, although cases with OptiMARK and Magnevist have also been reported. In this study, Omniscan caused severe, macroscopic, NSF-like skin lesions in some animals, whereas Magnevist did not. However, owing to the residual Gd concentration found in the skin after Magnevist exposure, it cannot be excluded that there is a residual, even though significantly smaller, risk for the induction of NSF with Magnevist in this model.

Omniscan and Magnevist differ in 2 principal characteristics—the thermodynamic and conditional stability and the amount of excess ligand. The higher amount of excess ligand in Omniscan has been introduced to compensate for the lower thermodynamic stability.34,50 It has been suggested that the excess ligand might play a role in the occurrence of NSF.51 Our experiments indicate that the excess ligand in the Gd-containing contrast agents is unlikely to be responsible for the skin lesions, as the application of the pure calcium-bound ligand caldiamide did not result in any NSF-like signs. Furthermore, by decreasing the amount of caldiamide we observed stronger NSF-like signs.

Therefore, the occurrence of the NSF-like lesions seems to be correlated to the stability of the Gd-chelates and, thus, to the dissociation of the respective Gd complexes. This dissociation of the Gd complex could generate toxic effects either by toxicity of the Gd3+ ions or via transmetalation (Gd3+ vs. Zn2+) and depletion of endogenous trace elements in the skin51 especially as enhanced Zn excretion is observed in patients after application of Gd-containing contrast media,28–31 and a loss of Zn2+ has been reported in subchronic toxicity studies in rats.34 Incidentally, skin lesions in male rats were also observed in the same subchronic toxicity study and were attributed to “a disturbance of Zn metabolism.”34

In this study, we used Zn levels in the skin as a surrogate marker for the depletion of endogenous ions. We did not detect a significant difference in Zn levels in the skin (or in the liver or femur) between treated animals compared with the saline control group, which indicates that depletion of endogenous ions is not the cause of NSF-like signs in our model. We, therefore, also regard the original explanation by Harpur et al for their skin lesions to be incorrect. We did find considerable amounts of Gd in the skin (and to a much lesser extent in the femur and liver), especially after i.v. administration of Gd-EDTA, gadodiamide without caldiamide, and Omniscan. No Gd was detected in the saline control group in accordance with the fact that Gd is not endogenously found in rats (or humans). We found a qualitative correlation between Gd concentrations in the skin and the occurrence of NSF-like signs. The highest Gd concentration was observed after application of gadodiamide without caldiamide and Gd-EDTA, the groups with the most severe skin lesions. We were also able to conduct in a separate experiment an electron microscopy of a representative sample. This electron microscopy of an Omniscan-treated animal, which exhibited NSF-like signs, showed that at least parts of the Gd found in the skin had been deposited as small particles in macrophage-like cells. A high concentration of phosphorus was also found in the observed deposits, suggesting that at least some Gd is deposited as Gd-phosphate (Gd-PO4). Similar Gd-PO4 particles were described in the skin of NSF patients. The formation of such particles is strongly enhanced by the high PO4 level observed in patients with ESRD undergoing dialysis.52 The PO4 plasma values of those patients are comparable to the values in rats.53 On the other hand, one has to remember that our Gd-detection method (ICP-AES) cannot distinguish between chelated and unchelated Gd, leaving the possibility that parts of the detected Gd in the skin are still bound to the ligand and might be depleted at a later time point.

Nevertheless, our data suggest that in our model there is a qualitative correlation between the degree and onset of NSF-like signs and release of free Gd3+ ions. Given the limitations of extrapolation of animal results to the human situation, however, it cannot be concluded with absolute certainty that Gd3+ ions play a causative, let alone a sole role in the pathogenesis of NSF in humans. The concentration of Gd in the skin and to a lesser degree in femur and liver (Magnevist less than Omniscan) inversely correlates with the thermodynamic and conditional stability of the linear agents (Magnevist being more stable than Omniscan). A similar inverse correlation between residual Gd, and the stability of the contrast agents was found by Tweedle et al.54 To further investigate the differences between marketed Gd contrast agents, we are conducting a follow-up investigation comparing a range of marketed Gd contrast agents (Sieber MA, Lengsfeld P, Walter J, et al. Gd-containing contrast agents and their potential role in the pathogenesis of Nephrogenic Systemic Fibrosis. Manuscript in preparation).55

We have found very pronounced differences in the Gd concentrations between Magnevist and Omniscan in the skin, but lesser differences in the femur and liver. We do not have a strong explanation for this effect, but we would speculate that this observation is related to the effects of distribution kinetics and that the Gd observed in femur and liver originates to a higher extent from chelated Gd. Further studies are needed to fully explain this observation.

When interpreting the findings of this study to evaluate their predictive value regarding a possible association between the administration of gadolinium-based contrast agents and NSF, the chosen animal model and the number of animals involved in the study have to be taken into account. We are undertaking further studies to address these limitations.

In summary, we have established an animal dosing regimen that covered the prolonged systemic exposure to Gd-containing contrast agents in human patients with severely impaired renal function. Using this setup we were able to find NSF-like signs in animals treated with a Gd-containing contrast agent. From the results we conclude that the occurrence of the NSF-like signs in this model most likely correlates with the release of Gd3+ ions. A significant role of the depletion of essential trace elements by transmetalation or other toxic effects of the excess ligand in Gd-containing agents for the occurrence of NSF-like signs in the animal model can be ruled out. The release of Gd found in this experiment is correlated to the stability of the compounds, with the less stable compound Omniscan leading to significantly higher levels of Gd, especially in the skin, and to the occurrence of NSF-like signs, as opposed to the more stable Magnevist, for which more than 10-fold lower Gd levels in the skin and no macroscopic or microscopic skin lesions were observed.


The authors thank Philipp Lengsfeld, PhD, Christiane Pering, MD, and Daniel Grosu, MD, for the critical reading of the manuscript, and Ines Heinzelmann, Stefanie Runge, and Andrea Baumgart for their excellent technical assistance. The authors also thank Leigh Prevost, BSc Pharm, Medicus International, for his editorial assistance. For help with the ELM, the authors thank Wilfried Bleiss, PhD, Humboldt University, Berlin.


1. Cowper SE, Robin HS, Steinberg SM, et al. Scleromyxoedema-like cutaneous diseases in renal-dialysis patients. Lancet. 2000;356:1000–1001.
2. DeHoratius DM, Cowper SE. Nephrogenic systemic fibrosis: an emerging threat among renal patients. Semin Dial. 2006;19:191–194.
3. Cowper SE. Nephrogenic fibrosing dermopathy: the first 6 years. Curr Opin Rheumatol. 2003;15:785–790.
4. Scheinfeld N. Nephrogenic fibrosing dermopathy: a comprehensive review for the dermatologist. Am J Clin Dermatol. 2006;7:237–247.
5. Cowper SE. Nephrogenic fibrosing dermopathy [NFD/NSF Website]. 2001–2007. Available at
6. Cowper SE, Bucala R. Nephrogenic fibrosing dermopathy: suspect identified, motive unclear. Am J Dermatopathol. 2003;25:358.
7. Kucher C, Xu X, Pasha T, et al. Histopathologic comparison of nephrogenic fibrosing dermopathy and scleromyxedema. J Cutan Pathol. 2005;32:484–490.
8. Ortonne N, Lipsker D, Chantrel F, et al. Presence of CD45RO+ CD34+ cells with collagen synthesis activity in nephrogenic fibrosing dermopathy: a new pathogenic hypothesis. Br J Dermatol. 2004;150:1050–1052.
9. Edsall LC, English JC III, Teague MW, et al. Calciphylaxis and metastatic calcification associated with nephrogenic fibrosing dermopathy. J Cutan Pathol. 2004;31:247–253.
10. Lewis KG, Lester BW, Pan TD, et al. Nephrogenic fibrosing dermopathy and calciphylaxis with pseudoxanthoma elasticum-like changes. J Cutan Pathol. 2006;33:695–700.
11. Kuo PH, Kanal E, Abu-Alfa AK, et al. Gadolinium-based MR contrast agents and nephrogenic systemic fibrosis. Radiology. 2007;242:647–649.
12. Piera-Velazquez S, Sandorfi N, Jimenez SA. Nephrogenic systemic fibrosis/nephrogenic fibrosing dermopathy: clinical aspects. Skinmed. 2007;6:24–27.
13. Grobner T. Gadolinium—a specific trigger for the development of nephrogenic fibrosing dermopathy and nephrogenic systemic fibrosis? Nephrol Dial Transplant. 2006;21:3604–3605.
14. Marckmann P, Skov L, Rossen K, et al. Nephrogenic systemic fibrosis: suspected causative role of gadodiamide used for contrast-enhanced magnetic resonance imaging. J Am Soc Nephrol. 2006;17:2359–2362.
15. FDA.
16. Knopp MV, Balzer T, Esser M, et al. Assessment of utilization and pharmacovigilance based on spontaneous adverse event reporting of gadopentetate dimeglumine as a magnetic resonance contrast agent after 45 million administrations and 15 years of clinical use. Invest Radiol. 2006;41:491–499.
17. Bourrinet P, Martel E, El Amrani AI, et al. Cardiovascular safety of gadoterate meglumine (Gd-DOTA). Invest Radiol. 2007;42:63–77.
18. Herborn CU, Honold E, Wolf M, et al. Clinical safety and diagnostic value of the gadolinium chelate gadoterate meglumine (Gd-DOTA). Invest Radiol. 2007;42:58–62.
19. Caravan P, Ellison JJ, McMurry TJ, et al. Gadolinium(III) chelates as MRI contrast agents: structure, dynamics, and applications. Chem Rev. 1999;99:2293–2352.
20. Idee JM, Port M, Raynal I, et al. Clinical and biological consequences of transmetallation induced by contrast agents for magnetic resonance imaging: a review. Fundam Clin Pharmacol. 2006;20:563–576.
21. Schmitt-Willich H. Stability of linear and macrocyclic gadolinium based contrast agents. Br J Radiol. 2007;80:581–582; author reply 584–585.
22. Tweedle MF. “Stability” of gadolinium chelates. Br J Radiol. 2007;80:583–584; author reply 584–585.
23. Weinmann HJ, Brasch RC, Press WR, et al. Characteristics of gadolinium-DTPA complex: a potential NMR contrast agent. AJR Am J Roentgenol. 1984;142:619–624.
24. Lauffer RB. Paramagnetic metal complexes as water proton relaxation agents for NMR imaging: theory and design. Chem Rev. 1987;87:901–927.
25. Ide M, Kuwamura M, Kotani T, et al. Effects of gadolinium chloride (GdCl(3)) on the appearance of macrophage populations and fibrogenesis in thioacetamide-induced rat hepatic lesions. J Comp Pathol. 2005;133:92–102.
26. Oyaizu T, Okada Y, Shoji W, et al. Reduction of recipient macrophages by gadolinium chloride prevents development of obliterative airway disease in a rat model of heterotopic tracheal transplantation. Transplantation. 2003;76:1214–1220.
27. Spencer AJ, Wilson SA, Batchelor J, et al. Gadolinium chloride toxicity in the rat. Toxicol Pathol. 1997;25:245–255.
28. Kimura J, Ishiguchi T, Matsuda J, et al. Human comparative study of zinc and copper excretion via urine after administration of magnetic resonance imaging contrast agents. Radiat Med. 2005;23:322–326.
29. Puttagunta NR, Gibby WA, Puttagunta VL. Comparative transmetallation kinetics and thermodynamic stability of gadolinium-DTPA bis-glucosamide and other magnetic resonance imaging contrast media. Invest Radiol. 1996;31:619–624.
30. Puttagunta NR, Gibby WA, Smith GT. Human in vivo comparative study of zinc and copper transmetallation after administration of magnetic resonance imaging contrast agents. Invest Radiol. 1996;31:739–742.
31. Sarka L, Burai L, Brucher E. The rates of the exchange reactions between [Gd(DTPA)]2- and the endogenous ions Cu2+ and Zn2+: a kinetic model for the prediction of the in vivo stability of [Gd(DTPA)]2-, used as a contrast agent in magnetic resonance imaging. Chemistry. 2000;6:719–724.
32. Tweedle MF, Hagan JJ, Kumar K, et al. Reaction of gadolinium chelates with endogenously available ions. Magn Reson Imaging. 1991;9:409–415.
33. Saab G, Abu-Alfa A, Sadowski EA, et al. Are patients with moderate renal failure at risk for developing nephrogenic systemic fibrosis? Radiology. 2007;244:930–932.
34. Harpur ES, Worah D, Hals PA, et al. Preclinical safety assessment and pharmacokinetics of gadodiamide injection, a new magnetic resonance imaging contrast agent. Invest Radiol. 1993;28(Suppl 1):S28–S43.
35. Staks T, Schuhmann-Giampieri G, Frenzel T, et al. Pharmacokinetics, dose proportionality, and tolerability of gadobutrol after single intravenous injection in healthy volunteers. Invest Radiol. 1994;29:709–715.
36. Weinmann HJ, Laniado M, Mutzel W. Pharmacokinetics of GdDTPA/dimeglumine after intravenous injection into healthy volunteers. Physiol Chem Phys Med NMR. 1984;16:167–172.
37. Jimenez SA, Artlett CM, Sandorfi N, et al. Dialysis-associated systemic fibrosis (nephrogenic fibrosing dermopathy): study of inflammatory cells and transforming growth factor beta1 expression in affected skin. Arthritis Rheum. 2004;50:2660–2666.
38. VanWagoner M, O'Toole M, Worah D, et al. A phase I clinical trial with gadodiamide injection, a nonionic magnetic resonance imaging enhancement agent. Invest Radiol. 1991;26:980–986.
39. Joffe P, Thomsen HS, Meusel M. Pharmacokinetics of gadodiamide injection in patients with severe renal insufficiency and patients undergoing hemodialysis or continuous ambulatory peritoneal dialysis. Acad Radiol. 1998;5:491–502.
40. Schuhmann-Giampieri G, Krestin G. Pharmacokinetics of Gd-DTPA in patients with chronic renal failure. Invest Radiol. 1991;26:975–979.
41. Cacheris WP. Chelate Compositions. 1989.
42. Wedeking P, Kumar K, Tweedle MF. Dissociation of gadolinium chelates in mice: relationship to chemical characteristics. Magn Reson Imaging. 1992;10:641–648.
43. Crissman JW, Goodman DG, Hildebrandt PK, et al. Best practices guideline: toxicologic histopathology. Toxicol Pathol. 2004;32:126–131.
44. Greaves P, ed. Integumentary System. Amsterdam: Elsevier; 2000:1–54.
45. Klein-Szanto A, Conti C. Skin and oral mucosa. In: Handbook of Toxicologic Pathology. Vol. 2. Haschek W, Rousseaux C, Wallig M, eds. San Diego: Academic Press; 2002:85–119.
46. Noelte J. ICP Emissionsspektometrie für Praktiker. Weinheim: Wiley-VCH Verlag GmbH; 2002.
47. Franzen C, Fischer S, Schroeder J, et al. In vitro cultivation of an insect microsporidian Tubulinosema ratisbonensis in mammalian cells. J Eukaryot Microbiol. 2005;52:349–355.
48. Sadowski EA, Bennett LK, Chan MR, et al. Nephrogenic systemic fibrosis: risk factors and incidence estimation. Radiology. 2007;243:148–157.
49. Quan TE, Cowper S, Wu SP, et al. Circulating fibrocytes: collagen-secreting cells of the peripheral blood. Int J Biochem Cell Biol. 2004;36:598–606.
50. Cacheris WP, Quay SC, Rocklage SM. The relationship between thermodynamics and the toxicity of gadolinium complexes. Magn Reson Imaging. 1990;8:467–481.
51. Khurana A, Runge VM, Narayanan M, et al. Nephrogenic systemic fibrosis: a review of 6 cases temporally related to gadodiamide injection (omniscan). Invest Radiol. 2007;42:139–145.
52. Habicht A, Watschinger B. [Sympathetic overactivity and the kidney.] Wien Klin Wochenschr. 2003;115:634–640.
53. Krinke G, Naylor DC, Skorpil V. Pyridoxine megavitaminosis: an analysis of the early changes induced with massive doses of vitamin B6 in rat primary sensory neurons. J Neuropathol Exp Neurol. 1985;44:117–129.
54. Tweedle MF, Wedeking P, Kumar K. Biodistribution of radiolabeled, formulated gadopentetate, gadoteridol, gadoterate, and gadodiamide in mice and rats. Invest Radiol. 1995;30:372–380.
55. Sieber MA, Walter J, Frenzel T, et al. Are Gadolinium contrast agents a possible trigger for the development of nephrogenic fibrosing dermopathy? Eur Radiol. 2007;17(Suppl 1):158.

gadolinium; nephrogenic systemic fibrosis; gadodiamide; gadopentetate dimeglumine; magnetic resonance; contrast agent

© 2008 Lippincott Williams & Wilkins, Inc.