‘Gadolinium toxicity’ is a generic term that had been ascribed to a myriad of signs and symptoms resulting from exposure to GdCAs, some of which have been confirmed – hypersensitivity reactions and NSF, whilst others are more debatable – PC-AKI and chronic pain. After having been deemed at an increased risk of gadolinium toxicity, the administration of GdCAs to patients with renal insufficiency has declined. Contrast-enhanced MRI scans may still be performed in this population if the benefit of information to be gained outweighs the potential risk, thus investigations may still continue retrospectively.
Through health data linkage, the anonymised patient data from 39,078 MRI scans, performed locally between 2004 and 2016, was linked with patient-specific electronic health records. Gadoteric acid was administered during 22,897 of these scans, 3,209 of which were performed on RI individuals. A low number of hypersensitivity reactions were observed within our adult cohort (0.01%) and none within our paediatric patients post-contrast exposure. After following these patients via their health records for an average of 6 years, zero cases of NSF were identified. Serum creatinine levels for the study cohort were largely unaffected (Fig. 1), with no statistically higher rate of PC-AKI following exposure to gadoteric acid when compared to non-contrast examinations, regardless of baseline renal function. Across all patients, there has been no significantly higher rate of chronic pain onset, a symptom that has been associated with GDD, when compared with patients who had undergone non-contrast scans.
To the best of our knowledge, this study is one of the largest and longest investigations into NSF diagnoses in cohorts with RI following GdCA administration. It is amongst the first to investigate for an array of potential signs and symptoms of gadolinium toxicity. No study has previously used data linkage in this manner to identify hypersensitivity reactions to gadoteric acid. Traditional adverse event monitoring methods have estimated reaction rates to gadoteric acid at between 0.12%  and 0.97% . Rates of NSF and PC-AKI have been researched previously, with which this study agrees well. In particular, no unconfounded cases of NSF have been confirmed following gadoteric acid administration [2, 29, 31] and no significantly higher rates of PC-AKI were observed between the cohort of patients with RI exposed to gadoteric acid and those who underwent non-contrast scans only . As GDD is a recently hypothesised clinical entity, few studies have investigated for potential symptoms. Whilst this study has investigated for the incidence of chronic pain onset, similar data linkage methodologies have been used to assess for cerebellar sequelae [32, 33]. In particular, Perrotta et al failed to observe the occurrence of clinical cerebellar syndrome in their gadoteric acid-exposed population .
Conversely, other GdCAs have been associated with some form of gadolinium toxicity. Although the vast majority of NSF cases have been seen following the administration of linear agents, there have been unconfounded cases reported with other macrocyclic agents [12, 34, 35] . In cases of suspected GDD, chronic pain is a frequent symptom  for which chronic pain medication has been administered in order to ease symptoms . Available data on GdCA-induced PC-AKI varies widely, and with small sample sizes it is difficult to draw definitive conclusions , but PC-AKI has been observed following administrations of both gadodiamide  and gadopentetate dimeglumine .
Demonstrating that GdCAs can be used safely in this population may improve outcomes for patients with RI. By being able to examine patients using the full capabilities of MRI, co-morbidities and complications may be better diagnosed and treated . Although an MRI-based clinical trial is likely to be deemed unethical at this time, health data linkage has allowed MRI-based research to continue in this cohort and further help to identify any potential clinical consequences of gadolinium toxicity. It was noted in the EMA’s report  that, regardless of renal function at the time of GdCA administration, it remains unknown whether later declines in patient health may trigger gadolinium toxicity symptoms. Thus, it is imperative to employ a surveillance technique that can accurately monitor patient populations over long periods of time. Ramahlo et al also suggests that reactions to gadolinium may depend somewhat on genetics . As such, there is a potential for further studies to use health data linkage, incorporating genetic data, where available, to help develop understanding of how this disease may progress, if at all.
There are limitations to this study. Regarding the collection of hypersensitivity reaction data, it is possible that by only examining A&E records some mild adverse events may have been overlooked. Patients who suffer a mild reaction may present to their GP or other healthcare providers (e.g. community pharmacy) and, although there is potential for these records to also be examined, they were not available for this study. As NSF awareness increased within the time period of this study, precautionary methods, such as immediate post-MRI dialysis, may have been employed thus reducing any potential effects. The retrospective nature of this study allowed for NSF diagnoses to be determined only via dermatology records. It is appreciated, however, that NSF is a multi-factorial pathology and a definitive diagnosis includes clinical, biochemical and histopathological features , which were unavailable. Serum creatinine changes may have resulted from other factors that were not controlled for, such as cardiovascular disease, renal artery stenosis, advanced age, dehydration and use of nephrotoxic drugs. Averaging serum creatinine levels at precise time points pre- and post-MRI would have provided a clearer indication of renal function; however, these data were not available. Following the completion of this study, it was noted that new guidelines regarding PC-AKI were published  in which the definition of PC-AKI was amended. As access to the data had expired, the analysis could not be repeated using the new definitions. Unfortunately, height data were not recorded and thus the Schwartz formula could not be used to categorise paediatric renal function. Cohort demography differences are unlikely to have impacted results as gadolinium toxicity has been shown to have no age, gender or racial predisposition [11, 42]. It cannot be guaranteed that patients were administered no other GdCAs, nor sought medical attention in other institutions. In some cases, patients may not have consulted their general practitioners or sought any other form of medical attention. Beyond NSF, a disease profile relating to gadolinium deposition is unconfirmed so these results are merely conjecture. Chronic pain and PC-AKI have, however, all been associated with the administration of GdCAs, and using large datasets such as this one may be the only way to identify unconfounded symptoms, if any, caused by the agents.
Cases of potential gadolinium toxicity to gadoteric acid (Dotarem®) were investigated through the electronic linkage of health data. None of the 22,897 contrast-enhanced scans performed on adult and paediatric patients, including the 3,209 scans performed on patients with renal insufficiency, resulted in a diagnosis of NSF. A low number of hypersensitivity reactions post-contrast were observed. There was no statistically higher rate of chronic pain or cases of PC-AKI observed between the contrast- and non-contrast-enhanced scans investigated.