Migraine with atypical features may require neuroimaging, according to the guideline. These include an unusual aura; change in clinical features; a first or worst migraine; a migraine that presents with brainstem aura, confusion, or motor manifestation; migraine accompaniments in later life; headaches that are side-locked or posttraumatic; and aura that presents without headache.
Assessing the evidence
The recommendation to avoid MRI or CT in otherwise neurologically normal patients with migraine carried a grade A recommendation from the American Headache Society, while the specific considerations for neuroimaging was based on consensus and carried a grade C recommendation, according to lead author Randolph W. Evans, MD, of the department of neurology at Baylor College of Medicine in Houston, and colleagues.
The recommendations, published in the journal Headache (2020 Feb;60(2):318-36), came from a systematic review of 23 studies of adults at least 18 years old who underwent MRI or CT during outpatient treatment for migraine between 1973 and 2018. Ten studies looked at CT neuroimaging in patients with migraine, nine studies examined MRI neuroimaging alone in patients with migraine, and four studies contained adults with headache or migraine who underwent either MRI or CT. The majority of studies analyzed were retrospective or cross-sectional in nature, while four studies were prospective observational studies.
Dr. Evans and colleagues noted that neuroimaging for patients with suspected migraine is ordered for a variety of reasons, such as excluding conditions that aren’t migraine, diagnostic certainty, cognitive bias, practice workflow, medicolegal concerns, addressing patient and family anxiety, and addressing clinician anxiety. Neuroimaging also can be costly, they said, adding up to an estimated $1 billion annually according to one study, and can lead to additional testing from findings that may not be clinically significant.
Good advice, with caveats
In an interview, Alan M. Rapoport, MD, editor-in-chief of Neurology Reviews, said that while he generally does not like broad guideline recommendations, the recommendation made by the American Headache Society to avoid neuroimaging in patients with a normal neurological examination without any atypical features and red flags “takes most of the important factors into consideration and will work almost all the time.” The recommendation made by consensus for specific considerations of neuroimaging was issued by top headache specialists in the United States who reviewed the data, and it is unlikely a patient with a migraine as diagnosed by the International Classification of Headache Disorders with a normal neurological examination would have a significant abnormality that would appear with imaging, Dr. Rapoport said.
“If everyone caring for migraine patients knew these recommendations, and used them unless the patients fit the exclusions mentioned, we would have more efficient clinical practice and save lots of money on unnecessary scanning,” he said.
However, Dr. Rapoport, clinical professor of neurology at the University of California, Los Angeles, founder of the New England Center for Headache, and past president of The International Headache Society, said that not all clinicians will be convinced by the American Headache Society’s recommendations.
“Various third parties often jump on society recommendations or guidelines and prevent smart clinicians from doing what they need to do when they want to disregard the recommendation or guideline,” he explained. “More importantly, if a physician feels the need to think out of the box and image a patient without a clear reason, and the patient cannot pay for the scan when a medical insurance company refuses to authorize it, there can be a bad result if the patient does not get the study.”
Dr. Rapoport noted that the guideline does not address situations where neuroimaging may not pick up conditions that lead to migraine, such as a subarachnoid or subdural hemorrhage, reversible cerebral vasoconstriction syndrome, or early aspects of low cerebrospinal fluid pressure syndrome. Anxiety on the part of the patient or the clinician is another area that can be addressed by future research, he said.
“If the clinician does a good job of explaining the odds of anything significant being found with a typical migraine history and normal examination, and the patient says [they] need an MRI with contrast to be sure, it will be difficult to dissuade them,” said Dr. Rapoport. “If you don’t order one, they will find a way to get one. If it is abnormal, you could be in trouble. Also, if the clinician has no good reason to do a scan but has anxiety about what is being missed, it will probably get done.”
There was no funding source for the guidelines. The authors reported personal and institutional relationships in the form of advisory board memberships, investigator appointments, speakers bureau positions, research support, and consultancies for a variety of pharmaceutical companies, agencies, institutions, publishers, and other organizations.