Biochemical safety profiles of gadolinium-based extracellular contrast agents and nephrogenic systemic fibrosis


  • Hale Ersoy MD,

    Corresponding author
    1. Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA
    • Cardiovascular Imaging Section, Department of Radiology, Brigham & Women's Hospital, 75 Francis Street, ASB II, L1, Boston, MA 02115
    Search for more papers by this author
  • Frank J. Rybicki MD, PhD

    1. Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA
    Search for more papers by this author


Gadolinium (Gd)-based paramagnetic contrast agents are relatively safe when used in clinically recommended doses. However, with the rapidly expanding body of literature linking Gd-based paramagnetic contrast agents and nephrogenic systemic fibrosis (NSF), awareness of the potential side effects and adverse reactions from Gd is now an important requirement for practicing radiologists. In addition to the ongoing accumulation and analyses of clinical NSF data, it is also essential for the practicing radiologist to understand the biochemical characteristics of the extracellular Gd-chelates. The purpose of this review is to consolidate and update the available information on known side effects, adverse reactions, and toxicity of the Gd chelates, with particular emphasis on the potential mechanisms of NSF. J. Magn. Reson. Imaging 2007;26:1190–1197. © 2007 Wiley-Liss, Inc.

GADOLINIUM-BASED MR contrast agents have different molecular structures and physicochemical properties and can be classified in different ways. However, the most common approach is to differentiate contrast agents based on their tissue biodistribution: extracellular, intracellular, tissue-specific, and blood pool/intravascular contrast agents. Most gadolinium (Gd)-based contrast agents available for clinical use are extracellular agents. At present, Omniscan® 0.5 mol/liter (gadodiamide; gadolinium-diethylenetriamine-pentaacetate-bismethylamide [Gd-DTPA-BMA], Magnevist® 0.5 mol/liter (gadopentetate dimeglumine; Gd-DTPA), ProHance® 0.5 mol/liter (gadoteridol; Gd-HP-DO3A), MultiHance® 0.5 mol/liter (gadobenate dimeglumine; Gd-BOPTA), and OptiMARK® 0.5 mol/liter (gadoversetamide; Gd-DTPA-BMEA) are approved by the U.S. Food and Drug Administration (FDA), while others Dotarem® 0.5 mol/liter (gadoterate meglumine; Gd-DOTA), Primovist® 0.25 mol/liter (gadoxetic acid disodium; Gd-EOB-DTPA), Gadovist® 0.5 mol/liter (gadobutrol; Gd-BT-DO3A) and Vasovist® 0.25 mol/liter (gadofosveset trisodium; MS-325; diphenylcyclohexyl phosphonooxymethyl-Gd-DTPA) are also approved in Europe.


Unfortunately, free Gd3+ ion is toxic. Chelating with a suitable organic ligand, also known as chelator, such as diethylenetriamine pentaacetic acid (DTPA) greatly reduces free Gd3+ toxicity (1). Acute toxicity experiments demonstrate that the 50% lethal dose (LD50) of the free Gd3+ ions is 50 times higher than that of chelated Gd (2). The chelator protects the tissues from interactions with Gd3+ ions and enables rapid renal excretion of the Gd3+ ions to minimize biotransformation or accumulation in the body. The rate of renal excretion is greater for Gd-chelates than for free Gd3+ ions (2, 3).

Biochemical differences among Gd-based contrast agents are determined by the chemical structure of the chelator. The chelator can be linear or macrocyclic, and ionic or nonionic.

Biochemical properties of some Gd chelates are summarized in Table 1 (2–16).

Table 1. Biochemical Properties of Gd-Based Contrast Agents
Trade nameOther termsMolecular structureOsmolality (mOsm/kg H2O at 37°C)FormulationThermodynamic stability constant (log Ktherm)Conditional stability constant at pH 7.4Elimination half-life (minutes, mean ± SD)Dissociation half-life in 0.1N HClaElimination pathway
  • a

    Values are measured in vitro acidic conditions.

  • b

    From Ref.7.

  • c

    From Ref.9.

  • d

    From Ref.10.

  • e

    From Ref.4.

  • f

    From Ref.2.

  • g

    From Ref.3.

  • h

    From Ref.11.

  • j

    From Ref.8.

  • k

    From Ref.15.

  • l

    Note that the concentration of Primovist and Vasovist is 0.25 mol/liter, not 0.50 mol/liter as in the other contrast agents.

  • m

    From Ref.6.

  • n

    From Ref.16.

  • o

    From Ref.5.

  • p

    From Ref.13.

  • q

    From Ref.12.

  • r

    From Ref.14.

  • N/A = not available.

Magnevist® 0.5 mol/liter (Bayer HealthCare, Morristown, NJ, USA)Gd-DTPA (gadopentetate dimeglumine)Linear, ionic1960bFree DTPA 0.2% (1 mmol/liter)22.1c17.7d94 ± 11ef10 minutesefkidney
Omniscan® 0.5 mol/liter (GE Healthcare, Princeton, NJ, USA)Gd-DTPA-BMA (gadodiamide)Linear, non-ionic650bCa-DTPA-BMA (Na+ salt) 5% (25 mmol/liter)16.8c14.9d77.8 ± 16fg35 secondsfgKidney
OptiMARK® 0.5 mol/liter (Mallinckrodt, St Louis, MO, USA)Gd-DTPA-BMEA (gadoversetamide)Linear, non-ionic1110bCa-DTPA-BMEA (Na+ salt) (50 mmol/liter)16.6b15.0b103.6 ± 19.5hjN/AKidney
MultiHance® 0.5 mol/liter (Bracco Diagnostics Inc., Princeton, NJ, USA)Gd-BOPTA (gadobenate dimeglumine)Linear, ionic1970bNo excess ligand22.6b18.4b72 ± 5kN/A93% kidney, 3% bile
Primovist® 0.25 mol/liter (Bayer Schering Pharma AG, Berlin, Germany)lGd-EOB-DTPA (gadoxetic acid disodium)Linear, ionic688mCa-EOB-DTPA (concentration N/A)23.5cN/AN/AN/A50% kidney, 50% bile
Dotarem® 0.5 mol/liter (Guerbet Research, Aulnay Sous Bois, France)Gd-DOTA (gadoterate meglumine)Macrocyclic, ionic1350bNo excess ligand25.4c19.0d91 ± 14fj>1 monthfjKidney
ProHance® 0.5 mol/liter (Bracco Diagnostics Inc., Princeton, NJ, USA)Gd-HP-DO3A (gadoteridol)Macrocyclic, non-ionic630b[Ca-HP-DO3A]2 (Ca2+ salt) 0.1% (0.5 mmol/liter)22.8c17.1cn94.2 ± 4.8fo3 hoursfoKidney
Gadovist® 0.5 mol/liter (Bayer Schering Pharma AG, Berlin, Germany)Gd-BT-DO3A (gadobutrol)Macrocyclic, non-ionic1603bCa-BT-DO3A (Na+ salt) (1 mmol/liter)21.8bN/AN/AN/AKidney
Vasovist® 0.25 mol/liter (Bayer Schering Pharma AG, Berlin, Germany)lGd-DTPA (gadofosveset trisodium, MS-325)Linear, ionic700–950pFosveset (concentration N/A)p22.0qN/A1110rN/A91% kidney, 9% bile

Following intravenous administration, water-soluble Gd chelates initially distribute into the intravascular space and then rapidly diffuse across capillary membranes into the interstitial space. Gadobenate dimeglumine has partial hepatobiliary excretion, which gives it advantages for certain liver applications, while the blood pool agents have extended vascular presence and hence undergo continuous recirculation. Nevertheless, an ideal agent would be excreted without biotransformation or dechelation. For extracellular Gd chelates, the equilibrium phase occurs approximately 10 minutes after the intravenous injection. The steady-state volume of distribution of these agents is consistent with an extracellular distribution, in the range of 0.2–0.3 liters/kg (20–30% of body weight). Tissues with leaky capillaries and large interstitial spaces (e.g., malignant tumors, inflammation, etc.) enhance in the arterial phase and nearly all tissues enhance to some degree in the equilibrium phase. Exceptions are brain and testicular tissues due to their impermeable capillaries to the extracellular contrast agents. When this “blood-tissue barrier” is intact, there is no interstitial space distribution.

The plasma half-life of a Gd chelate is determined by the volume of distribution and the glomerular filtration rate (GFR) of that particular agent. In healthy adults, the plasma elimination half-life for extracellular contrast agents following intravenous injection is approximately 70–90 minutes, and 98% of the injected dose of the contrast agent is cleared from plasma via passive glomerular filtration within 24 hours (17). However, in patients with renal impairment, the plasma half-life of the Gd chelate is lengthened, inversely proportional to the residual glomerular function (18, 19). Gd chelates can be removed via hemodialysis or continuous ambulatory peritoneal dialysis (18, 20), although the peritoneal dialysis less effective than the hemodialysis. Reported average clearance rates of the Gd chelates via hemodialysis are 78%, 96%, and 99% at the first, second, and third sessions, respectively (18, 21). For most Gd chelates, the excretion via liver is negligible, even in patients with severely decreased renal function (19). Two exceptions are gadoxetic acid disodium and gadobenate dimeglumine, which have 57.0 ± 2.49% and 3% to 5% biliary excretion, respectively (15, 22). In patients with renal impairment, the amount excreted into the bile increases to 7% to 8% of the injected dose for gadobenate dimeglumine. Gadofosveset trisodium reversibly binds to albumin, providing extended intravascular enhancement compared to extracellular Gd chelates, and demonstrates 91% renal and 9% biliary excretion.


Gd chelates have an extremely low overall incidence of adverse events, most of which are minor. These adverse events can be classified into two groups: nonallergic reactions (e.g., headache, fatigue, arthralgia, taste perversion, flushed feeling, nausea, or vomiting) and idiosyncratic allergy-like reactions (e.g., hives, diffuse erythema, respiratory distress, chest tightness, respiratory distress, periorbital edema). Adverse reactions, while rare, are more common in patients with a history of asthma or allergy, in patients who received the contrast agent injected at a faster rate, and in patients with history of allergic reaction to a Gd-based or iodinated contrast agents (23). There may be cross reactivity among different Gd chelates. Cochran et al (24) published an adverse reaction rate of 0.07% with Gd-based contrast agents in 28,340 administrations, with only one case of bronchospasm (0.0035%). Murphy et al (25) have reported an overall adverse event rate of 0.17% with two anaphylactic reaction (0.01%) in 21,000 contrast-enhanced MR studies. Neindorf et al (26) have reported severe adverse event rate of 0.0003% in an estimated total of more than 2,000,000 applications. In 2006, Knopp et al (27) reported that the adverse event incidence following 45 million intravenous administrations of gadopentetate dimeglumine over the 15 years of clinical use was less than 0.01% of procedures, and within the total adverse events reported serious adverse event prevalence was 9.3%. Shellock et al (28) evaluated the safety and tolerability of gadobenate dimeglumine relative to that of gadopentetate dimeglumine in patients and volunteers undergoing MRI for various clinical conditions, and reported that the safety profile of gadobenate dimeglumine is similar to gadopentetate dimeglumine in patients and volunteers. In various comparative studies of Gd chelates, similar overall incidences and type of adverse events have been noted between the different contrast agents (7).

The safety profile of Gd-based contrast agents in the pediatric population under two years of age and in pregnant women is not known. While data supports the use of Gd-based contrast agents in pregnancy and lactation (29), agents can cross the placenta into the fetal circulation to subsequently be excreted by the fetal kidneys into the amniotic fluid. Consequently, in our practice Gd-based contrast agents are only used after a case-by-case analysis of the potential risks and benefits.

Contrast administration site and injection rate are important. In animal models, the effects of the rapid intravenous administration of the Gd chelates on the cardiovascular system have been quite extensively studied. Gd chelates injected directly into the central circulation have been reported to cause transient blood pressure disturbances, nonspecific electrocardiogram (ECG) changes, prolonged PR interval, tachycardia, atrioventricular conduction defects, and atrial and ventricular arrhythmias (30–33). While Muhler et al (31, 32) have concluded that injecting an ionic, hyperosmolar contrast medium at high dose as a fast bolus probably induces more transient disturbances in the cardiovascular system than low osmolar, nonionic agents, Idee et al (30) have suggested transmetallation as possible mechanism of the cardiovascular depressive effects of the Gd chelates. Transmetallation is discussed in detail below. It is important to note that the data from these animal studies may not predict those effects in humans. Pirovano et al (34) could not show a difference between the gadobenate dimeglumine and placebo. However, regardless of the translation from animal studies and the actual interplay of mechanisms, it is advised to avoid rapid administration of the Gd chelates directly into the central circulation, particularly at high doses (33, 34).

Although Gd chelates are not considered nephrotoxic in clinically recommended doses (35–40), increased off-label use of Gd chelates at high doses (e.g., for computed tomography [CT] angiography and digital subtraction angiography) have led to increased safety concerns. Animal studies demonstrate that Gd chelates can be nephrotoxic at high doses (>0.3 mmol/kg), and cause transient osmotic (albuminuria) and chemotoxic (renal enzyme excretion in urine) injury to the kidneys (41, 42). In fact, the experimental data in animals suggest that some Gd chelates may be more nephrotoxic than iodinated contrast agents at doses required for equivalent X-ray attenuation (43, 44). There is also supportive human data (45). Despite reports of negligible nephrotoxicity with clinically recommended doses, Gd chelates can cause acute renal failure in patients with underlying chronic renal insufficiency (46). Therefore, some authors do not recommend Gd chelates for radiographic examinations, particularly in patients with renal insufficiency (33, 45, 47, 48).

Contrast extravasation into tissues may cause edema, inflammation, and necrosis. Tissue damage appears to be more severe with Gd chelates that have high osmolality (33, 49). Early in vitro studies have shown that deoxygenated sickle erythrocytes align perpendicular to the magnetic field (50). In theory, increased local magnetic moments following contrast agent administration may increase the perpendicular alignment, and thus may induce vasoocclusive complications. However, later work in patients with sickle cell anemia has not provided supportive evidence of this (51). To date, vasoocclusive complications induced by Gd chelates have not been reported in patients with sickle cell anemia, and as such the proposed risk in contrast agent package inserts has never been a clinical concern in routine applications.


Various studies have been conducted to test the acute tolerance to different Gd chelates. Toxicity of the Gd chelates mainly comes from free Gd3+ ions and uncomplexed ligands (2). Once released, free Gd3+ ions are rapidly sequestered within the bone and the liver with a biologic half-life of several weeks (52). In rat and canine experiments, Bartolini et al (53) have shown that free Gd3+ ions may cross the blood-brain barrier. However, the concentration of Gd ions in the brain tissue was insignificant in comparison to the highest concentration of Gd3+ ions detected in the kidneys. Free Gd3+ ion is mainly cleared via blood in the liver and lungs. Yoneda et al (54) have demonstrated that the biologic half-life of Gd3+ ions in the lung of Wister rats is approximately 136 days.

The consequence is long-term retention of free Gd3+ ions. Complications of this include interactions with physiologic systems such as the reticuloendothelial system, and/or inhibition of the activity of some enzymes such as Ca2+-activated Mg-adenosine triphosphatase (ATPase) in the sarcoplasmic reticulum of skeletal muscle fibers, some dehydrogenases, kinases, glutathione S-transferases and aldolase via noncompetitive inhibition of Ca2+ binding (55, 56). In vitro studies demonstrate that Gd3+ competitively inhibits Ca2+ ion binding to the transport sites on purified sarcoplasmic reticulum Ca-ATPase (57), perhaps explaining the hemodynamic disturbances observed in animal toxicity studies (30). The calcium channel inhibition effect of the free Gd3+ impacts on subsequent physiological processes that depend upon Ca2+ influx, such as neural transmission and blood coagulation. Gadolinium chloride inhibits phagocytosis in normal and activated Kupffer cells and decreases the total cytochrome P450 content of hepatocytic microsomes (6). Free Gd ions forms complex with hydroxides and phosphates that are insoluble at a pH higher than 6.2 (58). Phagocytosis of these complexes might depress the reticuloendothelial system (59, 60), and might lead to foreign body reactions and fibrosis by inhibiting certain enzymes in the dermis (36, 61). It has been shown that the linear compounds exhibit more toxic effects on the reproductive functions and the skin than the cyclic compounds (36, 62).

The chelator must be highly selective for Gd3+ ion and tightly bound to it in order to prevent its release into the circulation. This process of releasing Gd3+ ions and binding a different cation is referred to as transmetallation. The stability of Gd chelates is determined by the thermodynamic stability constant, conditional stability constant, selectivity constant, and kinetic stability of the complex (2, 63). The in vitro energy required for the metalloligand to release Gd3+ ion determines the thermodynamic stability constant (64) and indicates the affinity of the unprotonated ligand for a metal ion. When the thermodynamic stability is weak, the chelator more readily releases Gd3+ ions. Whereas the thermodynamic stability constant does not take the pH into account, the conditional stability constant is a measure of the stability of a complex at physiological pH. Therefore, the conditional stability constant should be considered a more relevant parameter when evaluating the stability of a contrast agent at physiologic pH. The conditional thermodynamic stability at pH 7.4 is substantially lower in all cases than the overall thermodynamic constant. The selectivity constant describes the transmetallation from a thermodynamic point of view (i.e., at equilibrium) and corresponds to the difference between the thermodynamic stability constants of the Gd-chelate and the other metalloligands. This is especially important for the in vivo stability of the Gd chelates since there will be other endogenous cations (e.g., Fe3+, Mg2+, Cu2+, Zn2+, or Ca2+) competing with Gd3+ ions for the ligand, and endogenous anions (e.g., phosphate) competing for the Gd3+ ions (6, 65). Normally, if the ligand releases Gd3+ ions it will quickly rebind that same ion. But if the ligand prefers other readily available cations, free Gd3+ ions will be released into the tissues. Precipitation of metal salts or colloids could occur in vivo if ligands have too weak an affinity for a metal ion. The fourth, and probably the most important physicochemical criterion for in vivo stability is the kinetic rate of the metalloligands. The kinetic rate can be estimated from the dissociation half-life of the complex (Table 1).

Numerous in vitro and in vivo animal studies aimed at investigating the differences in transmetallation potential among the Gd chelates have suggested that the molecular structure of the Gd chelate plays an important role in the tendency for transmetallation. The in vitro studies demonstrate that linear Gd complexes such as gadodiamide, gadoversetamide, and gadopentetate dimeglumine are more prone to transmetallation when compared with the macrocyclic Gd complexes such as gadoterate meglumine and gadoteridol (5, 9, 30, 66–73). In vivo transmetallation may occur with endogenous Zn2+, Cu2+, and Ca2+ ions (2, 5, 9, 69). Among these endogenous metal ions, Zn2+ appears to be the strongest competitor for Gd chelators, based on the fact that subchronic animal toxicity studies have produced findings characteristic of zinc deficiency (2, 36). These results are consistent with the thermodynamic stability constants of the various Gd chelates tested.

The practical relevance of Gd chelate stability in humans has been demonstrated, and it is important to review this data. A retrospective study by Prince et al (68) revealed an interference between certain Gd chelates and colorimetric serum calcium measurements. The authors reported a spurious decrease in serum calcium measurements from normal to hypocalcemia (<8.5 mg/dL) in 16% of patients receiving gadodiamide. The decrease in calcium levels was greater in patients who received high dose (>0.2 mmol/kg) of gadodiamide, and in patients with renal insufficiency. Subsequent in vitro experiments revealed reduced serum calcium measurements in serum samples mixed with different concentrations of gadodiamide and gadoversetamide, the two agents with the lowest stability constants (68), but not with gadoteridol, gadopentetate dimeglumine, and gadobenate dimeglumine. Gadodiamide and gadoversetamide may interfere with angiotensin-converting enzyme, and zinc, magnesium, and total iron binding capacity assays. There is a positive interference with colorimetric measurements of Mg and total iron binding capacity, and both positive and negative interference with Fe assays (74, 75). Gadopentetate dimeglumine may transiently increase the serum ferritin level and produce negative interference with a colorimetric zinc assays (75).

All Gd-based contrast agents that are approved for clinical use have been thought of as having exceptionally high in vivo stability in humans and to have a wide safety margin when used at relatively low clinical doses (0.1–0.3 mmol/kg). However, this confidence has been shaken following reports of an association between Gd-containing contrast agent administration and the incidence of nephrogenic systemic fibrosis (NSF) in patients either on dialysis or with end-stage renal failure (76–87). As a result, the relevance of transmetallation in humans has gained considerable attention.


NSF was first described in the medical literature in 2000 (88). This disorder predominantly involves the skin, but it may also affect other organs such as the lungs, liver, muscle, and the heart in patients with renal failure. Grobner (83) was the first to suggest that MR contrast media containing Gd might trigger NSF. Recently, researchers have documented the presence of Gd in biopsy specimens of patients with NSF (80, 89). The incidence of NSF among patients with GFR < 30 mL/minute is 3% to 5% (76, 82, 85)

Although it was not clear whether the Gd detected was in its “free” ionic form or bound to its chelator molecule, these observations led to the proposal that NSF could be a late adverse reaction to Gd chelates. This suggestion was indirectly supported by the fact that NSF was not described until 10 years ago.

So far, the majority of associations with NSF have been described in patients with moderate to severe renal failure who received gadodiamide, gadobenate dimeglumine followed by gadodiamide, and gadopentetate dimeglumine (Table 2) (76, 78, 80, 82, 83, 85–87, 89, 90), although editorials by Kuo et al (77) and Bongartz (62) mention that fewer cases have been reported in the United States and Europe after the administration of other linear Gd-based contrast agents (gadoversetamide, gadopentetate dimeglumine, and gadobenate dimeglumine). Of note, for the case associated with gadobenate dimeglumine, the patient received gadodiamide as well. The FDA refers to this as “confounded data.” The majority of NSF cases occurred after high doses of Gd agents for indications such as magnetic resonance angiography (MRA) or in patients exposed to single, standard doses of contrast several times in a short period of time. Thus, it is important in clinical practice to develop MR protocols that minimize the volume of contrast for the specific clinical indication, particularly in patients with moderate to severe renal failure.

Table 2. NSF Case Reports Associating the Gd-Chelate Administration With the Incidence of NSF
ReferenceYearNo. of patientsGd in the biopsy specimenContrast agent
  1. N/A = not available.

Boyd et al (80)20061YesN/A
Marckmann et al (82)200613Not investigatedGadodiamide
Grobner (83)20065Not investigatedGadodiamide
Broome et al (85)200612Not investigatedGadodiamide
Sadowski et al (76)200713Not investigatedGadodiamide (12), gadobenate dimeglumine followed by gadodiamide (1)
Deo et al (90)20073Not investigatedGadopentetate dimeglumine
Khurana et al (78)20076Not investigatedGadodiamide
High et al (89)20077YesN/A
Lim et al (86)20072Not investigatedGadodiamide
Marckmann et al (87)200719Not investigatedGadodiamide
TotalGadodiamide (77), gadopentetate dimeglumine (3), gadobenate dimeglumine followed by gadodiamide (1)

There is no definite consensus among the authors regarding the most important in vivo stability parameter, and it would be wrong to speculate that some Gd chelates are more stable than the others by taking a single parameter into account. Other factors, including the concentration of competing ions or ligands, and the interaction time between the Gd chelate and the competitors are important as well (68, 69). The association with renal failure can be explained by the fact that dissociation of Gd chelates is more likely to occur when it remains inside the body for an extended period of time.

Under normal circumstances, in vivo transmetallation and Gd release from Gd chelates are substantially slower than the renal clearance rate, resulting in significantly lower toxicity than the predicted toxicity by solution thermodynamics. A slower clearance rate from the body would be likely to increase the toxicity of Gd complexes. Factors that decrease the clearance rate of Gd complexes from the body have great impact on the kinetic stability of the complex (63, 68). If the elimination half-life is in the range of hours or days, then the reaction kinetics can be more important determinant of the in vivo stability of the Gd chelates. Therefore, rapid elimination of the contrast agent from the body is important to minimize the time period for Gd chelate decomplexation and subsequent toxicity.

There are differences in the in vivo kinetic stability of the Gd chelates. In vitro studies have demonstrated that macrocyclic chelates (gadoterate meglumine and gadoteridol) demonstrate a much slower kinetic rate, and therefore a longer decomplexation and transmetallation half-life compared to linear chelates (gadopentetate dimeglumine and gadodiamide) (16, 67, 70). The elimination half-life of these agents is prolonged 20 times in renal insufficiency compared with the elimination half-life (approximately 90 minutes) in people with normal renal functions. This is more profound (34.2 hours) for gadodiamide in case of terminal renal insufficiency (18, 37, 91–93). This leads prolonged interaction of the Gd chelates with endogenous anions and cations, which can cause transmetallation, particularly in case of unstable complexes.

Grobner (83) has associated metabolic acidosis with the development of NSF. Although their conclusion requires further investigation, it is known that acidic conditions enhance dechelation of the metalloligands (16, 67), and thus coincidence of metabolic acidosis and slow clearance of Gd chelates due the decreased or absent renal excretion might promote Gd accumulation within the tissues. A case control study by Marckmann et al (87) demonstrated that increasing cumulative gadodiamide exposure, high-dose erythropoietin treatment, and higher serum concentrations of ionized calcium and phosphate increase the risk of gadodiamide-related NSF in renal failure patients. Severe cases seem to develop primarily among patients in regular hemodialysis therapy at exposure. Their findings support the previous chemical theories on the possibility of in vivo interaction (transmetallation) between gadodiamide and endogenous cations, including Ca2+ and Fe2+ and on competition between the Gd ligand in gadodiamide and endogenous anions, including phosphates. High serum levels of ionized calcium might be increasing the risk of transmetallation and higher phosphate levels might be increasing the chance of Gd ion retention. Toxic Gd ions can cross the plasma membranes resulting in serious consequences (6, 65).

The commercial formulations of gadodiamide includes excess calcium-bound chelator (naphthalene acetic acid [NAA]) to minimize transmetallation and ensures the absence of free Gd3+ in the solutions during contrast shelf lives (5, 71) (see Table 1). Experiments in mice demonstrate that the addition of 5% calcium (DTPA-BMA) into the solution substantially reduces the toxicity of the gadodiamide; the LD50 of the nonformulated gadodiamide and formulated gadodiamide solution are 14.8 mmol/kg and 38.3 mmol/kg, respectively (2). Although it is obvious that excess chelator significantly decreases the acute toxicity and early transmetallation following intravenous administration, an animal study by Tweedle et al (71) showed that gadodiamide leaves more Gd residue in the body compared to the other Gd chelates. The authors measured residual 153Gd in liver and bone at various times up to 14 days following administration of 153Gd-labeled gadopentetate, gadoteridol, gadoterate, and gadodiamide. Gadodiamide left significantly higher 153Gd compared to the other Gd chelates at long residence times. The authors concluded that the excess chelate only conferred an advantage in the early stage following the intravenous administration, since other ions such as copper, zinc, and calcium in vivo can bind strongly to the chelates used for formulation of this class of MR contrast agent. These findings have been validated with a recent study in humans by Gibby et al (94). The authors demonstrated that commercially available gadodiamide solution left 2.5 times more Gd in the bone compared with gadoteridol.


The safety profiles and biochemistry of Gd-based contrast agents have been extensively described. There are numerous remaining questions concerning the mechanisms involved with respect to transmetallation, especially in patients with renal impairment who receive high doses of Gd-based contrast agents. However, it is important for the practicing radiologist to understand the underpinnings of these questions to best evaluate the increasing volume of data. With respect to NSF, the association with Gd accumulation in the skin has been described only in patients who had moderate to severe renal failure.

While the purpose of this article is to review the scientific basis of adverse effects such as NSF, it is important to recognize that most of the clinical cases are not well described. While low thermodynamic stability and fast kinetic ratio of the gadodiamide may not be the exclusive etiology of the complex disease process in NSF patients, evidence from work to date remains substantial and thus Gd accumulation under acidic/alkaline conditions or under the influence of a competitor ligands in NSF remain critical in ongoing investigations. With respect to policy in contrast delivery, we are obliged to follow the recommendations of the FDA and other leadership groups in our field. We hope that all radiologists will maintain a working understanding of the material presented in order to better understand, interpret, and implement these recommendations on a patient-by-patient basis.