Gadolinium Contrast Agents and Brain Deposits: A Closer Look

An Interview With Dr Robert McDonald

Interviewer: Tricia Ward; Interviewee: Robert McDonald, MD, PhD


July 24, 2017


Robert J. McDonald, MD, PhD

Gadolinium-based contrast agents (GBCAs) are used in MRI and angiography. In 2014, Japanese researchers showed increased signal intensity in the brains of people who underwent repeated GBCA-imaging tests.[1] Subsequently, Mayo clinic researchers compared postmortem neuronal tissue samples from patients who underwent multiple GBCA-enhanced brain MRIs with those from patients who did not receive GBCA and found gadolinium deposits in the neuronal tissue of exposed patients.[2] Medscape spoke with Dr Robert McDonald, the lead investigator for the Mayo study.

Medscape: What prompted you to look for gadolinium in postmortem brain tissue of patients who had undergone gadolinium-enhanced MRI?

Dr McDonald: Like many people, we read the provocative paper by Dr Tomonori Kanda and colleagues[1] that suggested that the T1 signal changes that were observed in certain parts of the brain could potentially be due to gadolinium. Their paper wasn't confirmatory, so we performed an analysis to prove whether or not this was gadolinium.

Just let me be clear: If GBCAs were removed from clinical practice, we would be set back 20-30 years in terms of our diagnostic capability.

Medscape: Your first postmortem study in adults found deposits,[2] and you recently published one in the pediatric population. What did that find?

Dr McDonald: The paper in JAMA Pediatrics[3] shows the same phenomenon occurring in pediatric patients. There are lingering questions: Is it because many of the patients who are studied are really sick, or is this occurring in every patient who receives a GBCA?

The people in our original study and those in many subsequent studies were exposed to a lot of GBCAs because they have some sort of intracranial pathology. The concern was that maybe their blood/brain barrier wasn't as robust, owing to either the presence of intracranial neoplasm or to age. That is why we started studying this in other populations, including the pediatric population.

Medscape: Did the patients in your studies have brain cancer?

Dr McDonald: Yes, many of them did. This is the challenge with doing this sort of research. People are getting gadolinium for good reasons, and very few of those who die undergo autopsy; therefore, we are very limited in the tissue that is available even at a very large tertiary care center, such as the Mayo Clinic.

But we have also seen this in people with normal brain tissue.[4]

Medscape: The phenomenon of gadolinium brain deposition has been reported mostly in patients who had brain imaging for cancer or multiple sclerosis. Could it occur with imaging of other body parts?

Dr McDonald: We know from prior animal data[5] and data that we have in press[6] that gadolinium deposits in every tissue in the body. The reason it was seen in the brain is because the brain has unique architecture that makes it very different from other organs in the body. This finding was noticed in the first place because a specific part of the brain is taking up the gadolinium preferentially.

We have confirmatory evidence that gadolinium is accumulating in other organs, but our current understanding is that there is no preferential uptake in a specific part of other organs, such as the kidney or liver, so it would have been nearly impossible for someone to stumble upon this in another organ system. Kanda and colleagues observed it because of the unique preferential neuroanatomical localization of some of the deposits.

[W]e were able to conclusively prove that every macrocyclic agent we studied had some level of deposition.

Medscape: Would you assume that patients who are undergoing multiple imaging sessions over time, as occurs in multiple sclerosis or cancer, are most at risk for gadolinium deposition?

Dr McDonald: Our original study was the first to prove that not only were those deposits gadolinium, but also that there was a dose-dependent relationship. We continue to see that in all of our subsequent studies in humans and animals.

The cumulative dose affects the amount of gadolinium that is deposited. You used the term "risk," but to me, that term means that this is necessarily a bad thing. Obviously, in a perfect world, it would be best if these agents wouldn't deposit, but currently we have no concrete, scientific evidence that these deposits are harmful.

Medscape: That was the conclusion of the US Food and Drug Administration (FDA) review. What about the proposal that this is more of a concern with the linear GBCAs, particularly the linear, nonionic GBCAs?

Dr McDonald: Hospital centers usually use one agent, maybe two, and they use it for years, so comparing agents in humans is enormously challenging. To research this, you have to go to an animal model system, which is what we did.

We have an article in press in the journal Radiology that compared many of the different linear GBCAs with macrocyclic agents.[6] We did indeed find that linear agents deposit more than the macrocyclics, but unlike some of the human studies using T1-signal changes that suggested no deposition of macrocyclic agents, we were able to conclusively prove that every macrocyclic agent we studied had some level of deposition.

Even within a class—so within linear agents, for example—we saw differences between agents. Similarly, within the macrocyclic class, there were differences between agents; some deposited less, and some deposited more. The differences seemed to somewhat mirror the stabilities of the agents.

Medscape: From that, I would infer that you may not agree with the suggestion from Radbruch and colleagues[7] that switching to macrocyclics could reduce the preexisting hyperintensity?

Dr McDonald: It's true that macrocyclics deposit less. My concern with the T1-signal data that have been published is that people are putting far too much weight on the quality of these data and drawing erroneous conclusions with respect to their propensity to deposit.

Assessing T1-signal changes from MRI studies is at best a semi-quantitative method. There are so many variables: People don't standardize their MR scanner pulse sequences between scanners, let alone between institutions, so comparing data and therefore drawing robust conclusions is difficult.

I'm an astronomy buff, and the analogy I used at a plenary talk at the American Society of Neuroradiology meeting was inspired by a photo from NASA, where they showed the moon and they had a box around this dark field in the sky that seemed to have no stars. If you were to point your $5000 hobby telescope at that region of the sky you would see nothing—only black. You might conclude that there is nothing out there. If the Hubble telescope points at that same region of the sky, it sees hundreds of galaxies.

Our MR scanners are like high-end hobby telescopes. We can draw false conclusions if we rely on this semi-quantitative method. If mass spectrometry data from necropsy tissue from animals or humans detect gadolinium, it doesn't matter whether the MRI does not reveal T1-signal changes. The far more sensitive mass spectrometry results confirm that all GBCAs deposit to some extent.

Medscape: It depends on how hard you look.

Dr McDonald: Yes. Yet the community has put too much emphasis on those studies. The reason they keep coming out is because everyone can do it. Everyone wants to publish and weigh in on whether or not they have seen T1 shortening, but that is not the question any more. It doesn't add to our scientific knowledge when there is already evidence that macrocyclic GBCAs cause it (we are not the only ones who have published on this). Even industry has confirmed that macrocyclic agents do deposit, albeit at lower levels than linear agents.[8]

The rates of adverse allergic-like reactions (even fatal reactions) are significantly lower with linear agents. I think that is a strong argument against removing them from the market.

Medscape: What about suggestions that institutions should switch to macrocyclic agents?

Dr McDonald: I'm on the fence on this one. People are going to make that decision for financial reasons and err on the side of caution. The counterargument is that to date, there has been no conclusive scientific evidence that linear agents are more harmful or that the excess gadolinium deposited from linear agents is harmful.

We haven't established a threshold of safety where we could say that the gadolinium deposits from linear agents are unacceptably high compared with macrocyclics; we just haven't found any toxicity from that yet.

Medscape: What is the harm of erring on the side of caution and abandoning linear agents?

Dr McDonald: GCBAs are generally well-tolerated, but pharmacovigilance data show that deposition is only one part of the safety profile. Although linear agents deposit more, they tend to have lower adverse reaction rates.[9,10] On a population basis, allergic-type reactions occur more frequently than T1 shortening. The rates of adverse allergic-like reactions (even fatal reactions) are significantly lower with linear agents.

I think that is a strong argument against removing them from the market. If someone has an allergic reaction to a macrocyclic agent, the American College of Radiology (ACR) guideline recommendation is to use the most chemically dissimilar contrast agent, which would be a linear agent. (I'm on the ACR contrast safety committee, but the guidelines were written before I joined.) At the very least, we need to have linear GBCAs as a back-up for those patients who have had adverse reactions.

These types of adverse reactions are more common, so there is an argument for the continued use of linear agents. Right now, physicians are being asked to balance a theoretical risk against the real, albeit uncommon, risk associated with allergic reactions to these agents. These data must also be considered in the risk/benefit equation.

Most people are going to get very few, if any, gadolinium doses in their life, and there's already evidence that for those patients in particular, we don't see any symptoms.

Medscape: Part of the risk/benefit analysis is taking a judicious approach to using any GBCA only when really needed.

Dr McDonald: What the National Institutes of Health[11] and the FDA have suggested is spot on: You have to decide whether you really need this agent, and go forward from there. Most people are going to get very few, if any, gadolinium doses in their life, and there's already evidence that for those patients in particular, we don't see any symptoms.

Many papers are going to focus on those groups of patients because they are the easiest ones to study. To design a study on the effects of gadolinium exposure in patients who receive 20 or 30 doses is going to be a challenge, because typically, those patients are very sick. Many of them have some sort of intracranial pathology, and trying to extricate which symptoms are the result of a brain tumor versus the gadolinium is essentially impossible.

Medscape: The FDA review concluded that there is no evidence that any of the GBCAs are harmful. Do you think it's too early to say that there is no clinical effect?

Dr McDonald: First, we have already conducted the "experiment": We've given 400 million doses to a few hundred million people, and have not seen any convincing evidence of systematic symptoms after exposure.

Controlled scientific data will come soon and should help clarify things. There was a study in JAMA last year that suggested no increase in Parkinson disease in patients from Canada.[12]

We already have some clues that this is probably safe, but I think we still have to do due diligence as physicians and scientists and prove that it is safe. The burden of proof is on us, and the community is working as fast as it can.

However, I am worried that we are going to draw conclusions before those safety data exists.

Medscape: In the Canadian Parkinson disease study, most of those patients only had one scan.

Dr McDonald: Correct; there were not that many doses. On the flip side, that is a better representation of most humans who get an MRI—they get one or two doses in their life at most.

The patients getting 30 or 40 doses of gadolinium are generally pretty sick. They need their gadolinium-enhanced imaging because it provides critical diagnostic information. In the case of brain tumors, it absolutely helps us to understand whether or not the tumor is stable, growing, or reacting favorably (ie shrinking) to therapy.

Just let me be clear: If GBCAs were removed from clinical practice, we would be set back 20-30 years in terms of our diagnostic capability. The same is true in cardiac MRI.

Medscape: Are you looking for this phenomenon in cardiac MRI as well?

Dr McDonald: We're actively looking at other tissues, yes.

Medscape: Hopefully, we can speak to you again when those data are available. Thank you.

Follow Tricia Ward on Twitter: @_triciaward

For more Cardiology, follow us on Twitter: @theheartorg


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.