Response to Dr. Do

We sincerely appreciate the feedback from Dr. Quyen Do and the opportunity to respond. Gadolinium remains a singular toxicant because it serves no physiological function in the human body [1-4]. No tenable argument refutes that gadolinium causes nephrogenic systemic fibrosis, a catastrophic iatrogenesis, or the occasionally fatal gadolinium encephalopathy [2]. Our research team demonstrated that magnetic resonance imaging contrast agents stimulate bone marrow–derived fibrocytes, driving systemic fibrosis [5,6]. This process proceeds through Nox4-generated reactive oxygen species, monocyte chemoattractant protein-1, and the CCR2 receptor [5-9]. We also proved that bone marrow retains a memory of prior contrast agent exposure [10]. We were the first to report the formation of gadolinium-rich nanoparticles in rodent and human kidneys [2,8,9,11,12]. Intracellular gadolinium-rich sediment is strong evidence of in vivo degradation of magnetic resonance imaging contrast agents. We are invested in studying other potential maladies, such as symptoms associated with gadolinium exposure [1-3].

Dr. Do notes, “macrocyclic complexes remain among the safest drugs ever introduced,” [sic]. No references were provided. There are many publications on adverse events associated with macrocyclic magnetic resonance imaging contrast agents [13]. There are over fivefold more acute reactions from macrocyclic agents than linear [14]. As of 11 December 2025, we looked up the cases reported to the FDA Adverse Event Reporting System. For gadoteridol (ProHance), the FDA Adverse Event Reporting System lists 3958 reports, and 230 deaths. For gadobutrol (Gadavist), there are 5241 cases and 188 deaths. For gadoterate meglumine (Dotarem/Clariscan), 2461 cases and 54 deaths. Our research team wants to improve the safety of magnetic resonance imaging contrast agents and reduce the total number of cases and deaths by discovering the biological consequences of gadolinium exposure.

We have a recent comprehensive review on the safety of magnetic resonance imaging contrast agents, The Safety of Magnetic Resonance Imaging Contrast Agents [2]. We agree that safety must not be shortcut. The onus of demonstrating the safety of rare-earth metal-based contrast agents—and the nanoparticles they generate—should rest with the pharmaceutical industry and academic centers, not with patients too often marginalized by the medical establishment.

Professor Robert Lenkinski and Professor and Chairman Neil M. Rofsky presume that “the likelihood of meaningful human or ecological toxicity remains very low” [15]. We are very concerned that the contrast agents are entering surface waters and remaining intact as pharmaceutical agents, i.e., in the soluble form. Magnetic resonance imaging contrast agent pollution is why gadolinium concentrations have risen in waters sampled in San Francisco, Berlin, and Tokyo [15]. Drs. Lenkinski and Rofsky concede that gadolinium is present in tap water and, in Germany, in soft drinks at up to 100 ng/L. In rodents and humans, we demonstrated that intracellular gadolinium-rich nanoparticles form following systemic administration of a magnetic resonance imaging contrast agent and subsequent degradation [2,8,9,11,12]. Judicious academicians would assume that orally ingested magnetic resonance imaging contrast agents undergo the same fate. Professor Neil M. Rofsky is on the advisory boards for GE Health (Omniscan, Vueway) and Guerbet (Elucirem, Dotarem, Artirem). Readers should be aware when authors have conflicts when making such a bold claim about the likelihood of human or ecological toxicity. (Our conflict of interest is that we drink water⸮) Professors Lenkinski and Rofsky are obliged to present substantial evidence reassuring consumers that magnetic resonance imaging contrast agents are safe to ingest orally. Until then, we recommend remaining guarded.

Gadolinium-based pharmaceutical chelates sequester gadolinium to mitigate its intrinsic toxicity, yet gadolinium retention in human tissues occurs even in individuals with normal renal function [2,4,11]. Parsing the discussion in terms of thermodynamic stability constants ($K_{Therm}$) obfuscates safety concerns. The Gibbs free energy ($\Delta {\mathrm{G}}^{°}$) for the formation of the gadolinium-ligand complex for every brand of magnetic resonance imaging contrast agent is highly negative, i.e., spontaneous under standard conditions.

$$\Delta G^{°}=-RT\text{ln}K$$

All magnetic resonance imaging contrast agents have a $K_{Therm}$ equal to or greater than 3.98 × 10¹⁶ (i.e., OptiMARK). For OptiMARK, as an example of a ligand (with a $K_{Therm}$ similar to that of Vueway), the Gibbs free energy will be in the realm of −98.5 kJ·mol⁻¹, assuming a body temperature of 37° and the gas constant, R, of 8.313 J·mol⁻¹ K⁻¹. Gadolinium binding to the pharmaceutical multidentate ligands is overwhelmingly favored in the test tube. How much does the milieu intérieur impact the reverse reaction?

Although transmetallation—defined as the replacement of the gadolinium (III) cation in its pharmaceutical chelate by endogenous metal ions (e.g., Zn²⁺, Cu²⁺, Ca²⁺)—has long been posited as a mechanism of dechelation, evidence suggests it is unlikely to occur in vivo under physiological conditions. Tweedle et al. [16] and Kindberg et al. [17] detected minimal transmetallation in rodent models (even with linear agents). Schuenke et al. (inadvertently) dealt a categorical coup de grâce to ‘transmetallation’ dogma [18]: Physiological concentrations of zinc (i.e., < 0.125 mm) did not promote gadolinium dissociation from Magnevist. The transmetallation hobgoblin should now be abandoned as a concept to explain gadolinium-induced complications.

For the sake of patient safety, accept that gadolinium is not going to be displaced by any cation at physiological concentrations, no matter what the brand of the ligand is. Universal laws of chemistry do not permit the reversal of these enormous Gibbs free energy thresholds. We propose that trans-ligandation, or the exchange of gadolinium (III) from the pharmaceutical ligand to biological Lewis base-rich molecules (e.g., oxalate, phosphates, or proteins), may be a more plausible mechanism of Gd release [19]. Our updated model of gadolinium release reflects the known affinity of metabolites such as oxalic acid and may more accurately account for the chemically diverse species present in tissues. Transmetallation, if it occurs, happens later — and incidentally. This framing matters. If the gadolinium has already exited the pharmaceutical chelate and bound to biomolecules, bench-top transmetallation assays become moot for patient safety. Le Châtelier's principle, though, explains our observation of intracellular gadolinium-rich nanoparticles perfectly.

Our goal in publishing these results was to establish that:

1. Magnetic resonance imaging contrast agents can be decomposed actively upon exposure to endogenous compounds.

2. Biological molecules (proteins) can affect the rate of this reaction and allow it to occur at lysosomal pH.

3. Demonstrate generally that the equilibrium constant of these compounds is not necessarily indicative of real stability in vivo.

We are not proposing that the studies in this publication serve as a definitive mechanism, nor have we stated thus. Yes, we agree that 333 mM oxalate at pH ~ 1 does not occur in the human body. We reported the physical chemistry of the interactions between these compounds and select biological molecules as a first step toward building a refined mechanistic model of gadolinium deposition.

Given emerging biochemical evidence, it is now important to account for active ligand displacement. The common metabolite oxalic acid reacts with magnetic resonance imaging contrast agents. Biochemistry must be considered when discussing the stability of magnetic resonance imaging contrast agents. In the presence of oxalic acid, a common and problematic human metabolite, gadolinium precipitates into digadolinium trioxalate within seconds. We are championing the investigation of gadolinium-based contrast agent degradation in simulated physiological environments. Our Table 4 is correct; the pH ranged from 1.15 to 4.37.

Gadolinium-rich nanoparticles have been found intracellularly, within lysosomes [2,8,9,11,12]. The degree of uptake doesn't matter. Gadolinium retention is a fact. Peer-reviewed evidence confirms this [20,21]. Plasma half-lives vary depending on an individual's renal function. Our clinical investigation shows that gadolinium persists in the body for years after exposure in some individuals [1]. We remain open to alternative hypotheses regarding the genesis of intracellular gadolinium-laden nanoparticles. But to date, these structures have been found only in those exposed to magnetic resonance imaging contrast agents (in rodents and in humans).

We are pleased that Dr. Do concedes that gadolinium dissociates from the pharmaceutical chaperone under in vitro conditions. We believe that all brands of gadolinium-based contrast agents require scrutiny in vitro, in vivo, in situ, and in toto. Investigators need to study magnetic resonance imaging contrast agents in environments that model the milieu intérieur. Thermodynamic stability values are meaningless in physiological compartments where endogenous substrates have higher affinities for gadolinium than the magnetic resonance imaging contrast agent ligands. Thermodynamic stability constants are particularly irrelevant when the chemical reactions favor precipitation of gadolinium in vivo. These points, in addition to the demonstration of Le Châtelier's principle with gadolinium sedimentation in humans, are fundamentally unethical to ignore, given repeated exposures to different brands of agents.

Dr. Do speculates whether the nanoparticles observed in rodent models and human tissues are gadolinium phosphate crystals formed from the degradation of the magnetic resonance imaging contrast agent. We have already ruled this out [2,11,12]. The intracellular gadolinium-rich nanoparticles identified in our rodent models and in humans by scanning transmission electron microscopy have been analyzed using high-angle annular dark-field and bright-field modes. Based on the current observations, any ordered domains do not extend beyond a few nanometers, and the overall structures lack extended periodicity. The nanoparticles are predominantly non-crystalline. Jerrold Abraham and his co-workers presumed gadolinium phosphate crystals in skin using extended absorption fine structure (EXAFS) spectroscopy [22], yet their scanning electron microscope images are similar to the gadolinium-rich shrapnel we have reported [2,8,9,11,12]. Nonetheless, Dr. Do is conceding a key point: F-block elements are notorious for sedimentation across all physiological compartments. Again, because gadolinium from magnetic resonance imaging contrast agents precipitates and concentrates in these nanoparticles, confounded exposures matter.

If agents with high thermodynamic stability are so inert, why are manufacturers developing next-generation gadolinium-based contrast agents that require only half the dose? The marketing of “lower dose” agents undermines the premise of inertness. Also, note that the stability constants lauded vide supra fail to explain the biological findings now reported in multiple models and species.

As noted herein, thermodynamic stability and kinetic inertness are altered radically in the human body. The field of magnetic resonance imaging contrast agent safety must move forward. No one can safely state that confounded exposures are safe, because confounded exposures have been ignored for nearly 20 years. The fate of magnetic resonance imaging contrast agents in vivo requires serious inquiry. Current safety paradigms, rooted in bench chemistry, are well beyond calcified.