*GBCA – gadolinium-based contrast agent. In this column, volume of GBCA administered is provided alongside the estimated concentration of gadolinium per gram of brain. Calculated using 1400 g as the average weight of the human brain in order to be consistent with calculations from Arlt et al.[24]
**Intrathecal administration was unintentional.
Ψ Unintentional intrathecal administration was due to drug error.
Φ Adapted from Provenzano et al. 2019[5]

To understand the neurological risks of intrathecal gadolinium, it is critical to understand gadolinium’s chemistry, pharmacology, and proposed mechanisms of toxicity. Gadolinium is toxic in its free, ionic state (Gd3+).[6] In an effort to limit the toxicity of free gadolinium, the gadolinium ion is bound to a chelating ligand.[7] These chelating ligands are characterized based on their shape as either linear or macrocyclic. GBCAs can be further classified as ionic or nonionic depending on if the compound is electrically neutral (nonionic) or charged (ionic). Macrocyclic GBCAs form a tight “cage” around the gadolinium ion, making them more stable (less likely to release the gadolinium ion) than linear GBCAs. Ionic GBCAs have a higher thermodynamic stability constant (stronger bond between the gadolinium ion and chelating ligand) than nonionic GBCAs. [1] There are nine different brands of GBCAs: gadofosveset trisodium, gadoxetate disodium, gadopentetate dimeglumine, and gadobenate dimeglumine (linear ionic); gadodiamide and gadoversetamide (linear nonionic); gadoterate meglumine (macrocyclic ionic); gadobutrol and gadoteridol (macrocyclic nonionic) (Table 2). The prescribing information from eight out of the nine brands mention the risk of serious neurological adverse events after GBCA administration. Of these, gadodiamide is the only brand that mentions the neurological risks of intrathecal administration, and its prescribing information specifically states that gadodiamide is not to be used intrathecally.

Table 2: Prescribing information of different brands of GBCAΦ

*Specifically indicated the risk of neurotoxicity due to intrathecal administration
Source: 2012 Hao and prescribing info[1]
Φ Adapted from Provenzano et al. 2019[5]

Transmetalation is the major method of in vivo dechelation. Dechelation of the gadolinium ion occurs when endogenous ions (namely, Cu2+, Fe3+, Zn2+, and Ca2+) compete for gadolinium’s chelating ligand. [1,8] Research has mainly focused on in vivo gadolinium transmetalation with intravenous use, not intrathecal use. Therefore the influence of transmetalation on intrathecal gadolinium neurotoxicity is unclear. [1]

There have been several proposed mechanisms for acute gadolinium-induced neurotoxicity. Feng et al.[9] found that gadolinium exposure quickly led to disruption of mitochondrial function in rat cortical neurons. Mitochondrial dysfunction resulted in an accumulation of reactive oxygen species, which inhibited cellular function and induced apoptosis. Xia et al.[10] also studied gadolinium neurotoxicity in rat cortical neurons and demonstrated that oxidative stress from mitochondrial disruption led to strain in the endoplasmic reticulum. Alterations in the endoplasmic reticulum resulted in the activation of the apoptotic pathway. In another study, Feng et al.[11] observed dramatic extracellular calcium ion influx in astrocytes after gadolinium exposure. Elevated calcium levels may play an important role in apoptosis. Gadolinium disrupts multiple calcium-dependent pathways since its ionic radius is similar in size to calcium, allowing it to compete with calcium in vivo. Gadolinium has been shown to block voltage-gated calcium channels, inhibit calcium-activated enzymes, and function as an agonist in calcium receptors. These qualities give gadolinium a wide range of detrimental effects, from mitochondrial disruption and oxidative stress, to inhibition of muscle contraction and nerve signal transmission.[8]

Acute intrathecal gadolinium-induced neurotoxicity has been demonstrated in multiple animal models. Ray et al.[12] injected gadopentetate dimeglumine into the lateral ventricles of rats. Dose-dependent neurotoxicity was seen with symptoms beginning at gadolinium concentrations of 5 μmol/g brain. Affected rats demonstrated behavioral changes and focal lesions in the thalamus, brain stem, and spinal cord. No evidence of neurological decline was observed in rats that received gadolinium concentrations of 2.5 or 3.3 μmol/g brain. In a second study, Ray et al. [13] performed a similar experiment, this time including the GBCA gadodiamide. Once again, dose-dependent neurotoxicity was seen, with morphological changes beginning at 1.25 μmol/g brain and behavioral changes beginning at 2.5 μmol/g brain. Toney et al. found that 2.5 μmol/g brain intrathecal gadopentetate dimeglumine caused no significant signs of neurotoxicity in rats. [14] However, documented histological changes were seen including gliosis and inflammation.

Numerous documented cases of intrathecal GBCA acute neurotoxicity have been reported in humans with outcomes ranging from transient signs and symptoms with no long-term sequelae to death (Table 1). Neurotoxic effects resulted from both intentional and unintentional intrathecal delivery of GBCAs. Neurotoxic intrathecal gadolinium concentrations ranged from 1.07 μmol/g brain to 7 μmol/g brain (Table 1). In human case series and reports using lower concentrations (i.e. 0.07 to 0.36 μmol/g brain) of intrathecal gadolinium, no signs of acute neurotoxicity were seen (Figure 1). [15-21] Alarmingly, dose ranges found to be relatively safe in animal models (e.g. rats) have been demonstrated to be neurotoxic and even fatal in humans. For example, doses as high as 3.33 μmol/g brain have caused no neurotoxic effects in rats, but doses as low as 2.3 μmol/g brain have been fatal in humans. [5,12] Notably, doses lower than 2.3 μmol/g brain (1.07, 1.43, and 2.14 μmol/g brain) have been neurotoxic in humans, but not fatal (Table 1). The disparity between animal and human neurotoxicity at similar concentrations undermines the efficacy of directly comparing animal and human responses to intrathecal gadolinium. This highlights the need for further research into the specific mechanism and risk factors for human gadolinium-induced neurotoxicity.

Specific volumes and dose limits for GBCAs have been suggested as viable visualization alternatives in interventional pain procedures when there is the possibility of inadvertent or deliberate intrathecal administration. [3,22]  First, we will remind the reader that the administration of a GBCA via routes other than intravenous is an off label use and that significant nervous system-related warnings are listed in the prescribing information. [5] Second, specific formulations of GBCAs have differing chelating agents and concentrations of gadolinium and thus the toxicity profile of one type is not directly applicable to another. When looking at the limited published clinical reports of acute GBCA neurotoxicity to date, physicians may desire to stay below a gadolinium dose of 1 μmol/g brain (ie, theoretical number based on case reports suggesting neurological risks). Based on a theoretical dose limit of 1 μmol/g brain, restricted GBCA volumes based on the GBCA’s gadolinium molar concentration would be feasible. GBCA volume limits would range between 1.4 mL (1 mol/L of gadolinium) to 5.6 ml (0.25 mol/L of gadolinium). The relative radiographic conspicuity of a GBCA is directly correlated to the gadolinium concentration of the agent selected.[23] Gadolinium-based contrast agents that have a higher gadolinium concentration are more visible under fluoroscopic imaging, but lower volumes would have to be used to stay below the theoretical neurotoxic intrathecal administration level.

Intrathecal gadolinium poses a serious risk of neurotoxicity, even at small doses. GBCAs are not FDA approved for use in the intrathecal or epidural spaces. Furthermore, if GBCAs are being considered as an alternative to iodine-based contrast agents for visualization in interventional pain procedures, one must remember that this is an off-label use with significant risk of acute neurotoxicity when intrathecal administration occurs intentionally or unintentionally. Since different types of GBCA have different chelating agents and concentrations of gadolinium, the toxicity profile of one type is not directly applicable to another. Extensive research is required to better understand the risk of GBCA neurotoxicity. Until further basic and clinical research is performed, GBCAs should not be considered a safe and viable alternative to iodine-based contrast in neuraxial interventional pain procedures for which there is the risk of intrathecal administration.

References