Gadolinium Deposition and Retention in the Brain: Should We Be Concerned?

Introduction

In this article, we will use the descriptions regarding the process by which gadolinium-based contrast agents (GBCAs) enter and are cleared by the brain provided by Le Fur and Caravan (1). The terms residual gadolinium and the presence of gadolinium are used in a general way when no precise mechanism is known or suggested. Retention means that a gadolinium species is retained but slowly eliminated over time, in opposition to deposition for which no excretion occurs. The terms in the title, namely, gadolinium deposition and retention in the brain, were selected because, as we will discuss later, there is evidence that both retention and deposition can occur in the brain, depending on which type of GBCA was administered.

The clinical use of gadolinium-based contrast material–enhanced MRI began in the late 1980s. Approximately 450 million doses of these contrast agents have been administered to date, with approximately 30 million doses administered annually. Along the way, there have been two major health care issues that have been related to the administration of GBCAs. The first, nephrogenic systemic fibrosis (NSF), was identified as a result of the administration of GBCAs in renally impaired patients. The implementation of clear guidelines for the use of GBCAs in renally impaired patients by the Food and Drug Administration (FDA) and American College of Radiology (ACR) in the United States and other advisory and regulatory agencies in other countries has virtually reduced the number of NSF cases following GBCA administration to zero. It is important to understand that the safety issues and relevant guidelines for use of GBCAs regarding NSF were addressed based on the development and adoption of standardized criteria for the clinical and pathologic diagnosis of NSF, which have very strong epidemiologic correlations with renal impairment, not the establishment of an unambiguous biologic mechanism relating the properties of the GBCAs to the physiologic development of NSF. Although there still is no validated mechanism relating gadolinium exposure to NSF, there is compelling evidence that dechelation (ie, the dissociation of gadolinium from the GBCA) occurred, and insoluble forms of gadolinium, most probably gadolinium phosphate, were deposited as a solid in the skin of patients who developed NSF. When discussing the stability of GBCAs, the terms thermodynamic stability and kinetic stability are often used to describe some of the chemical properties of the GBCAs. For clarity, it has been recommended that the more precise chemical term lability be used in place of kinetic stability (2). Because dechelation occurs in the course of the deposition of insoluble solids, the labilities, which are the inverse of the rates of dissociation, are more important determinants of the extent of solid deposition than thermodynamic stability constants or conditional stability constants. This was pointed out by Rofsky et al (3) in 2008 and Sherry et al (4) in 2009, providing a chemical rationale that explained the fact that epidemiologic studies showed that more than 98% of the cases of NSF occurred in patients who had undergone studies with the rapidly dissociating acyclic (linear) chelates gadodiamide (Omniscan), gadoversetamide (Optimark), or gadopentetate dimeglumine (Magnevist). It is important to note that different levels of gadolinium deposition observed in vivo in early tissue deposition studies performed in mice in 1995 (5) and humans in 2004 (6) and 2006 (7) were correlated with the rates of dissociation of gadolinium from the GBCAs, not their thermodynamic stability constants. As pointed out previously, once these agents stopped being administered to renally impaired patients, the number of NSF cases were dramatically reduced.

In 2014, the report by Kanda et al (8) showing unanticipated, dose-dependent, hyperintensity in deep gray matter regions on non–contrast-enhanced T1-weighted MR images in patients who had been injected with either gadopentetate dimeglumine or gadodiamide with normal renal function raised new safety concerns about the use of GBCAs in clinical practice. Over the intervening almost 5 years, there have been numerous original scientific publications, clinical case studies, research investigations performed in animal models, review articles, websites, and regulatory meetings and public forums dealing with a variety of issues related to these safety concerns. We will specifically refer to the FDA Medical Imaging Drugs Advisory Committee (MIDAC) meeting held on September 8, 2017, which has a comprehensive list of materials including briefing documents and a transcript of its proceedings online (https://www.fda.gov/advisory-committees/medical-imaging-drugs-advisory-committee/briefing-information-september-8-2017-meeting-medical-imaging-drugs-advisory-committee-midac). We will also refer to a publication resulting from a 2018 National Institutes of Health (NIH)/ACR/Radiological Society of North America (RSNA) workshop on gadolinium chelates that describes a research road map to address knowledge gaps about gadolinium retention and deposition (2) and recent review articles by La Fur and Caravan (1) and Gianolio et al (9) that summarize both what is known today about this topic and what might be learned from future research studies. Rather than provide a comprehensive review of all of these efforts, we will distill the information provided by these sources into what is known about gadolinium retention and deposition and the relevant knowledge gaps that still exist in this field. It is also important to apply some of the lessons learned from the investigations into the underlying causes of NSF to safety concerns regarding gadolinium deposition and retention in the brain.

Four questions were posed to the MIDAC panel at the 2017 meeting (see page 207–208 of the transcript, https://www.fda.gov/media/108935/download). We begin by focusing on question 2, “Is there a causal relationship between (gadolinium) retention and symptoms in patients with normal renal function?” Literature data and presentations on patients reporting symptoms potentially related to GBCA exposure indicated MRI examinations were performed for a variety of medical conditions; 132 case reports were identified with various symptoms often centering around pain. Most were self-reported and lacked details such as the specific GBCA product, number of GBCA administrations, and time to onset of symptoms. The MIDAC panel concluded that no causal relationship could be established at this juncture between gadolinium retention and patient symptoms, but that further study was needed. This conclusion was supported in the publication summarizing knowledge gaps discussed at the NIH/ACR/RSNA workshop held in February 2018 (2).

Since the publication by Kanda et al (8), there have been numerous reports describing hyperintensities observed in a variety of patient cohorts after the administration of GBCAs (for a recent review, see Pullicino et al [10]). The authors concluded that deposition and retention were much more likely to happen with the acyclic agents. These data expose a serious knowledge gap; namely, that we do not know what chemical form of gadolinium present in the human brain causes the hyperintensities observed. The techniques by which the chemical form of gadolinium is determined are called speciation studies. The various techniques used in speciation are reviewed in McDonald et al (2) and La Fur and Caravan (1). Very few, if any, of these techniques can be used noninvasively in the human brain, thus limiting studies to either small numbers of human tissue samples obtained at autopsy or preclinical studies using animal models. Interestingly, Gianolio et al (9) provided a discussion of what species of gadolinium could be present at sufficient concentration and with sufficient relaxivities to account for the hyperintensities observed. It is theoretically possible that the intact GBCA might cause these hyperintensities if it were present at high enough concentration in brain tissue. Gianolio et al and others (see references cited in [9]) had provided evidence that there could be a soluble, macromolecular complex formed (with a size estimate of about 300 kDa and a relaxivity of about 100 L ⋅ mmol⁻¹ ⋅ sec⁻¹). In their recent review, Gianolio et al (9) speculate about the possibilities that insoluble, solid forms of gadolinium containing species that could have high relaxivities might also contribute to the observations of hyperintensities with T1-weighted MRI sequences. These speculations are based on the facts that it is unlikely that the intact GBCA is present at sufficient concentration because only a small fraction of the injected dose of the GBCA reaches the brain and similarly only a fraction of the dechelated gadolinium would form a macromolecular complex. Our own studies presented at the 19th Annual International Society for Magnetic Resonance in Medicine meeting in Montreal (abstract #475) and at a lunch-time symposium on gadolinium safety presented at the Society of Computed Tomography and Magnetic Resonance meeting held in 2016 showed that solid forms of gadolinium, particularly the surface of gadolinium phosphate, had extremely high relaxivities and might cause the hyperintensities observed in human brain. Preclinical studies have shown that the macrocyclic GBCAs enter and leave the brain largely intact even after 24 days. The acyclic GBCAs have a much higher likelihood of dechelating to form both insoluble presumably inorganic salts and macromolecular complexes. If we apply the definitions provided at the beginning of this article, there can be both gadolinium retention (in the form of the intact chelate) with the macrocyclic GBCAs and gadolinium deposition and retention with the acyclic GBCAs. We emphasize again that at this time no causal relationship between gadolinium deposition or gadolinium retention and clinical symptoms has been established. However, it also important to note that “absence of evidence is not evidence of absence.” Because these discussions have provided no certain risk-free course of action for using GBCAs, it may be unclear to the MRI community as to how we should proceed. One hint was provided in the first of the four questions posed to the panel at the 2017 MIDAC meeting. The panel members were asked, “How do you characterize the risks of gadolinium retention? In other words, FDA has determined that benefit-risk is favorable for all approved GBCAs, but we have a new finding of gadolinium retention in general, and in the brain in particular, in patients with normal renal function associated with an unknown risk. How is this unknown risk related to the known risk of gadolinium retention in certain patients with renal failure?” It suggests that the FDA was asking for methodology for establishing a rational benefit-risk ratio for the clinical use of the individual FDA-approved GBCAs in a variety of clinical use cases. This should be addressed by the MRI community. We should encourage the development of methods for noninvasively determining the concentration of gadolinium in tissues and organs that may be more accessible than the human brain. Promising preliminary results have been reported by Lord et al in a comparison of noninvasive x-ray fluorescence measurements of gadolinium concentrations in the tibia (see [11] and technical references cited within) with higher concentration found in symptomatic self-reported exposed subjects as compared with nonsymptomatic exposed subjects.

Murata et al (12) have published gadolinium concentrations measured by using inductively coupled plasma mass spectroscopy in the globus pallidus and bone tissue obtained at autopsy from patients who had received GBCAs. Although the concentrations of gadolinium in the brain were much lower than in bone and even lower in patients who were administered a macrocyclic agent rather than an acyclic agent, a linear relationship between the concentration of gadolinium in the globus pallidus and bone in patients was observed. The slope indicated that the concentration of gadolinium in the globus pallidus was about 3% of that observed in cortical bone. The approach of using noninvasive measurements of gadolinium concentration in accessible bones as a surrogate for brain deposition or tissue deposition of gadolinium in general might prove to be a useful method to monitor methods to prevent or reverse gadolinium deposition in the brain and other organs. Investigation into the development of non–contrast-enhanced MRI methods that can perform as well clinically as contrast-enhanced scans are already underway. These should be encouraged. We should support research in the development of alternate MRI contrast agents such as those based on manganese or iron, both of which we either ingest or are present in our bodies. Finally, because we know that the approved macrocyclic agents exhibit much lower levels of gadolinium deposition in tissues than the approved acyclic agents, one strategy of risk reduction is wider use of these agents. The sales of these macrocyclic agents have increased over the past few years, suggesting that the market has already shifted in this direction. One major caveat still remains. Because we have no current evidence that gadolinium retention and deposition cause clinical symptoms, we have no direct method to determine whether any measures taken are effective at reducing clinical symptoms in patient populations receiving GBCAs. This remains a major concern in this field.