Neuroimaging is performed in a minority of patients with migraine in the emergency department. The finding contradicts prevailing opinion that unnecessary neuroimaging is frequent, according to researchers.
PHILADELPHIA—Neuroimaging is used in 23% of patients with migraine who present to the emergency department and is justified in 89% of such cases, according to Deborah I. Friedman, MD, MPH, Professor of Ophthalmology and Neurology at the University of Rochester in New York, and colleagues. Most patients with migraine went to the emergency department after their usual self-administered therapy failed, and about half had had a contraindication to receiving migraine-specific therapy in the emergency department, the researchers found.
As reported at the 14th Congress of the International Headache Society, Dr. Friedman’s group observed a large cohort of adult patients who were diagnosed with migraine and seen at the University of Rochester’s affiliated emergency departments from January 1, 2005, to December 31, 2005. The researchers analyzed history of migraine and type, indication of drug-seeking behavior, contraindications to receiving migraine-specific therapy, current headache characteristics, laboratory and neuroimaging tests performed, and other data. Categorical and continuous variables were compared using the chi-square test and Student’s t-test.
“The most commonly used treatment was parenteral prochlorperazine (n = 95),” Dr. Friedman and coauthors stated. Many patients received more than one medication, including parenteral ketorolac (n = 65), opioids (n = 63), another dopamine antagonist (n = 21; 18 promethazine, two metoclopramide, one ondansetron), parenteral oral opioids (n = 19), oral phenothiazine (n = 10), oral simple analgesics (n = 9), nasal lidocaine (n = 1), and IV diphenhydramine (n = 1). IV fluids (n = 74) and oxygen (n = 6) were also administered.
A total of 78 patients (50% of all patients) received justified nonspecific therapy, including 44 patients (28% of total patients) who used dihydroergotamine or a triptan prior to their emergency department visit, and 68 patients received nonspecific therapy, Dr. Friedman and colleagues reported. Ten patients were not treated.
Imaging Indications and Results
Various indications for imaging were observed: new headache with focal neurologic symptoms or visual aura, worst headache of one’s life, new headache, history of headache and new focal neurologic symptoms, and neurologic symptoms without headache, were the most common (four patients each). There was no apparent reason documented for requesting imaging studies in four patients. Additional indications included history of migraine with change in headache pattern, escalating headache pattern with or without focal neurologic symptoms, recent trauma, history of cluster headache, and other.
At discharge, 46 (23%) patients were headache free; 34 (22%) had marked improvement or mild residual pain; five (3%) had moderate improvement or residual pain; 42 (27%) were not documented; and five (3%) were admitted to the hospital.
“Neuroimaging studies were performed in the minority of cases (23%), with documented justification in 89%; even so, neuroimaging was the largest contributor to overall charges,” Dr. Friedman’s group stated. The average hospital charge was $1,799 per patient. Among eligible patients, 11.5% received migraine-specific treatment, a finding consistent with a previous study.
Neuroimaging and Treatment
“Our study, conducted in the same fashion as a typical quality assurance review, does not support the prevailing notion that neuroimaging is overutilized in the emergency department in patients with migrainous head pain,” Dr. Friedman and colleagues concluded. Patients who were treated for a typical or usual migraine did not undergo an imaging study. “The neuroimaging results did not change the management of patients in this cohort; none of the patients with a history of migraine who presented with symptoms consistent with migraine had an abnormal imaging study.”
Dr. Friedman’s group recognized some study limitations reflective of the retrospective design, including accuracy of coding, nonuniform charting, and lack of a unified pain rating scale. “However, the detailed review of a large cohort provides further insight into ‘what’s really happening’ in the emergency department,” they stated. Prospective studies and treatment protocols may improve patient care in this setting.
—Laura Sassano