Functions like sensing and pacing in implantable cardioverter defibrillators (ICDs) tend to resist interference from the energy fields generated by MRI, as long as device programming is properly adjusted before the scan.
That applies even to patients with older “legacy” devices implanted before the 2015 advent of MRI-conditional ICDs despite, in practice, prevalent but misguided resistance to obtaining MRI scans in such cases.
Less is known whether such non–MRI-conditional devices, once exposed to MRI, will then reliably deliver antiarrhythmic shocks or antitachycardia pacing (ATP) when needed.
A new cohort study has tried to fill in some of that knowledge gap. It showed no evidence of an excess risk for death or ICD failure to deliver therapy within about 2 years of clinically indicated MRI scans in 629 patients with non–MRI-conditional devices.
The findings, published online in the Annals of Internal Medicine, come with caveats. For example, they’re based on the experience of one, albeit major, center and on MRIs that were for varied indications using 1.5-tesla equipment only.
Despite such safety evidence for appropriately adjusted non–MRI-conditional ICDs, many patients with the devices don›t receive clinically indicated MRI scans due to “perceived risk” that the ICDs won’t then reliably deliver appropriate therapy, observe the authors, led by Joshua Ra, MD, University of California, San Francisco.
Any such risks are “largely theoretical,” but may still explain “why some institutions are shying away from offering MRI exams” to patients with non–MRI-conditional ICDs, Dr. Ra told this news organization.
Many such hospitals refer such patients to more experienced centers, creating “significant logistical barriers in terms of patient access to these MRIs,” he said. “That seems to still be prevalent, unfortunately.”
The current findings “provide another layer of reassurance” that MRI scans in patients with non–MRI-conditional ICDs don’t impair a device’s ability to deliver shocks or ATP, Dr. Ra said.
The cohort consisted of 629 patients with non–MRI-conditional ICDs who underwent 813 clinically indicated MRI exams from 2003 to early 2015 at Johns Hopkins University, Baltimore.
Scans performed within 4 weeks of device implantation were excluded because, the report notes, that’s when spontaneous lead dislodgements or changes to device parameters are most likely to occur. Also excluded were patients with permanent epicardial leads, abandoned leads, or subcutaneous ICD lead systems, the report states.
Still, Dr. Ra said, the cohort is fairly representative of “the modern patient population” of non–MRI-conditional ICD recipients.
A total of 4,177 arrhythmia episodes were documented during a median 2.2 years between scans and last device interrogation prior to pulse-generator change-out or lead exchange.
Of note, Dr. Ra observed, the arrhythmias were confirmed in only 85% of the cohort. Most of the remainder were referral patients who were lost to follow-up whose devices were unavailable for interrogation.
Device therapy terminated “nearly all” documented spontaneous arrhythmias in that 85% of patients, the report states. They included 757 episodes of ventricular tachycardia (VT) or ventricular fibrillation (VF), including 130 that were shocked and the remainder that were managed with ATP. There were also 105 supraventricular tachycardias, all successfully terminated with shocks.
There were no cases of VT or VF detection delay from undersensing or instances of syncope because of “abnormalities” in device detection of arrhythmias, the report states.
Of the 210 known deaths, which occurred a median 1.7 years after the scan, about half were noncardiac and more than a third were cardiac but nonarrhythmic.
Ten patients died from arrhythmia-related cardiac causes, representing 5% of deaths; but 7% of deaths were of undetermined cause.
“No direct relationship of deaths attributable to prior MRI exposure was found or reported,” the report states.
The researchers informally compared outcomes between older and more recently implanted non–MRI-conditional ICDs, the latter presumably with more modern design features. Their data, based on device interrogations, Dr. Ra said, “seem to suggest there were no differences.”
The study was supported by Johns Hopkins University and the National Institutes of Health. Author disclosures are available at apconline.org.
A version of this article first appeared on Medscape.com.