The article by Kim et al1 brings to light a larger topic— namely, how does one “prove” safety, either in general or specificallywithin themagnetic resonance imaging (MRI) environment?For example, in the caseof cochlear implants, several studieshavedocumented the “safety” ofMRI exposure by numerous patients with various cochlear implants. Why, then, did the patients in the study by Kim et al experience adverse events or pain, even to the point of terminating the examination? It is important to recognize that it is a mathematical impossibility to prove something to be safe, per se. All that can be done is to document that when specifically tested, some stimulus, agent, process, orother influencing factorwas found to not be associated with any recognized adverse events to a specifically tested threshold of detectability or statistical significance. For example, if 100 patients were studied following exposure to a specific drug, one might claim that following 100 exposures, the incidence of hives or other recognized cutaneous allergic reactionswas0.Clearly this is not the same as proving safety—it merely documents that in their experience, the recorded incidence of a specifically studied adverse event was 0. This might provide a statistical window of confidence to safety, but it is far from proving that this agent is, per se, safe. For example, it is entirely possible that the 101st patient exposedmight experience a severe anaphylactic reaction from exposure to that same drug. Additionally, it may be that allergic reactionswere not observed, but perhaps the patient did experience a severe but clinically asymptomatic hypokalemiaor thrombocytopenia—adverseevents thatmaynot have been anticipated, specifically prospectively studied, or clinically recognized in the study cohort. Not having observed that which one did not study is exceedingly common andmust be consideredbeforebroadlypainting the testedhypothesis itself as entirely safe. Furthermore, many studies purport to document safety merely by exposing a finite number of patients to a specific stimulus, environment, or situation and failing to observe—or recognize?—any adverse reactions. The current literature is repletewith such cases: for example,MRI exposure andpatients with cardiac pacemakers. What remains unrecognized, however, is that failing to identify an adverse reaction or event is not the same as proving safety. Yet a single publication documenting a life-threatening adverse event on a single patient with a cardiac pacemaker exposed to MRI successfully disproves the entire inappropriately broad safety claim. As has beenpreviously noted, “...failing to identify an adverse event isnotequivalent todemonstratingsafety—especiallywhenonly a limitednumberofpatientsare studied.”2(p1326)This solidprinciple is the basis of theUSFood andDrugAdministration’s rejectionof the“wesawnoproblems”methodologyand for their having required far more rigorous science before they would approve the first truly MRI conditionally safe cardiac pacemaker.3 Finally, while MRI scanners operating at magnetic fields of 1.5 T to 3.0 T raise concerns of potential translational and rotational forces, as detailed by Kim et al,1 scanners operatRelated article page 45 Research Original Investigation Adverse Events DuringMRIWith Cochlear Implants
Read full abstract