Choosing the Safest Gadolinium-based Contrast Medium for MR Imaging: Not So Simple after All

See also the article by Behzadi et al in this issue.

Introduction

Medicine is an art of balancing risks and benefits. When it was discovered that gadolinium in gadolinium-based contrast media was not excreted fully from the body (1), even in patients with normal renal function (1), it threw the risk-benefit assessment of patients being considered for contrast material–enhanced magnetic resonance (MR) imaging into turmoil. Historically considered extremely safe, gadolinium-based contrast media are commonly administered and necessary for the diagnosis and risk stratification of countless serious diseases. But now we know that, regardless of renal function and apparently regardless of patient status or type of contrast agent (1–3), if a gadolinium-based contrast medium is administered, a very small quantity of that gadolinium will remain in the patient for an indeterminate and possibly long period of time.

The clinical significance of this retention is unclear but has led some, most notably the European Medicines Agency (EMA) (4), to restrict or suspend the use of linear gadolinium-based contrast media that have been shown to have greater rates of gadolinium chelate dissociation compared with their macrocyclic counterparts. Although the EMA acknowledged that there was no scientific evidence of harm related to gadolinium retention at the time of their ruling, it was assumed that the suspended agents offered no or negligible clinically relevant benefit beyond those with lesser retention (ie, those with macrocyclic structure) and unknown but potentially important risk. In other words, the potential for risk was considered to be high enough that the EMA felt justified to remove certain linear agents with documented clinical efficacy from the market. As recently as September 8, 2017 (5), the U.S. Food and Drug Administration (FDA) stated that it agreed with the EMA on the science, but disagreed with the EMA on policy, and at the time of this writing, linear agents remain in use in the United States. So, why do these agencies agree on all of the facts, but differ on their regulatory approach? Likely this has to do with different opinions on how to regulate uncertainty and how to balance competing risks and benefits.

There are several risks and benefits that must be considered when selecting which gadolinium-based contrast agent to use, with some pertaining to the patient, some pertaining to populations of patients, and others pertaining to the health care system in general. In no particular order, these are included in the Table (6). In light of these various factors, the decision making involved when selecting an agent is complicated and difficult to reduce to a single issue.

Table Risks and Benefits of Gadolinium-based Contrast Agents

The work by Behzadi et al (8) in this month’s issue of Radiology focuses on one of these considerations: whether agents differ in their risk of allergic-like reactions. Unlike gadolinium retention, allergic-like reactions are known, quantifiable harms with immediately identifiable consequences. Gadolinium-based contrast media have a very low risk of allergic-like reactions, with agent-specific reaction rates ranging from 0.015% to 0.91% overall and 0.0016% to 0.019% for severe reactions (8). Therefore, on a per-patient basis, the differential reaction risk according to agent usually is not a primary consideration. However, on a population basis, an incrementally higher reaction rate, even if that difference for severe or lethal reactions is less than 10 reactions in thousands of examinations, can be an important consideration. Experts have long suspected through anecdotal experience and preliminary retrospective investigations that gadolinium-based contrast media differ in regard to their reaction risk. What makes these differences particularly compelling is that the lowest reaction rates observed tend to be those with the least chelate stability. If that somewhat inverse relationship is generally true, then aggressive moves to minimize the unknown harms of gadolinium retention could have unintended consequences—specifically, an increase in allergic-like reactions (including severe reactions) and a potential increase in allergic-like reaction–related deaths.

According to the Organization for Economic Cooperation and Development (9), the 2016 utilization rate for MR imaging in the United States, a country with approximately 320 million inhabitants, was 120.7 examinations per 1000 individuals, indicating annual performance of approximately 38 million MR imaging examinations. For approximately 70% of those (approximately 27 million) contrast material is used (10). On the basis of the data provided by Behzadi et al (8), the severe reaction rate for gadodiamide, a nonionic linear agent with relatively greater gadolinium retention, was 0.16 of 10 000 individuals, while the severe reaction rate for the three more stable commercially available macrocyclic agents ranged from 1.2 to 1.8 of 10 000 individuals, which is significantly and approximately 10-fold greater than that of gadodiamide. One can use these point estimates to derive the annual incidence of severe reactions in the United States that might occur if the market share were 100% dominated by gadodiamide versus one of the more stable macrocyclic agents (432 reactions per year vs 3240–4860 per year, respectively). In other words, approximately 3000–4000 more people might experience a life-threatening severe allergic-like reaction in the hypothetical scenario of 100% macrocyclic use when compared with the hypothetical scenario of 100% gadodiamide use. Of course, in this myopic comparison, all of the other risks and benefits of agent selection are not taken into account (eg, agent use in patients with severe chronic kidney disease and the need for renal function screening with less stable [ie, group 1] agents [6] and rates of gadolinium retention), but the comparison does illustrate the potential effect of agent switching on a population basis when only one of these factors is considered in isolation from the others.

By using observational data to compare contrast material reaction rates is complicated and subject to biases (7) that cannot be avoided with use of meta-analysis techniques. Market share, method of data collection, reporting bias, publication bias, and years of data collection all contribute. Therefore, similar to the currently available data related to gadolinium retention, the allergic-like reaction rates reported by Behzadi et al (8) are not the end of the story. However, the signals they (8) report (considerably higher reaction rates for agents with protein binding, macrocyclic structure, and ionicity) are important additions to the literature that must be placed into the broader context of risks and benefits when selecting which agent to use for a particular population or indication. While Behzadi et al (8) have gone a long way in informing us how the risk and severity of allergic-like reactions can vary according to agent and agent type, we remain in a state of ignorance with respect to gadolinium retention and clinical harm. Clarifying the risk of gadolinium retention in patients with normal or near-normal renal function (ie, the vast majority of patients being imaged) is an urgent priority that must be pursued to better inform the decision-making process.