Dexamethasone in Benign Headaches

This week in journal club, Matt reviewed a nice little trial submitted by a group of Texans to the Canadian Journal of Emergency Medicine. They studied IV dexamethasone in preventing benign headache recurrence (Can J Emerg Med 2006;8(6):393-400, PDF) – something I had never tried, but apparently has been bouncing around the neurology and EM literature for 20 years.

It turns out that migraines may not be simply a vascular disorder, but rather an inflammatory disease. And, as Matt pointed out, it’s very difficult to diagnose migraines; it might be simpler for us ED folk to say headaches exist on a continuum between tension and migraine, and maybe ED patients with primary headache would benefit from a steroid.

The authors had two aims: to see if patients discharged from the ED with benign headache had less recurrence if they received dex, and to see if dexamethasone could reduce the need for future doctor’s visits and allow a return to normal daily activity. To answer these questions, they designed a randomized, double-blind, placebo-controlled trial of adults presenting to the ED with benign headache (as diagnosed by the ED doc). Exclusions were reasonable: pregnancy, fever, stiff neck, focal neuro deficit, or cause to worry about immunosupression (active PUD, DM I, systemic fungal infection, or already taking steroids).

Of note, the dexamethasone was given IV, and the IV could not be placed just for the purposes of enrolling into the study. So those presenting with headaches had to be sufficiently worrisome so as to warrant IV placement, either for labs or fluids or other meds. Also of note - this was a convenience sample, and it took them ten months to enroll just 57 patients – it’s unclear whether they just didn’t get that many headaches requiring IV, or if the authors were seeking a higher level of convenience than I’m used to seeing.

The patients got 10 mg IV dex before discharge, rated their pain, and were followed up between 48 and 72 hours by phone. They thought they’d need 94 patients to achieve a clinically significant decrease in recurrence, but they achieved that early and so, unfortunately, stopped the study before the secondary goal of assessing quality-of-life was completely answered.

What did they find? Just 9.7% of patients who received dexamethasone experienced headache recurrence, whereas 58.3% of placebo patients had recurrence (a number consistent with previous studies of headache recurrence). Pretty awesome, right? We calculated the NNT (number needed to treat) for dex at around two, which is incredible. But quizically, when the information was collected in a different way, 12.9% of dexamethasone patients reported either headache severe enough to require a physician visit, or precluding a return to normal life (compared to 33.3% of placebo patients). The vast majority of dexamethasone patients (87.1%) needed no further treatment, or only OTC treatment – compared to 66.7% of placebo patients. Why don’t these “secondary goal” numbers match up with the primary goal data? It’s unclear, and tarnished the otherwise compelling results.

Who were these patients, anyway? Well, both placebo and dexamethasone patients were predominantly female, in their early thirties, and just under half reported a history of migraines). Both groups had about a 13% LP rate. For a randomized group, the dexamethasone patients seemed luckier – compared to controls, they got more of every class of pain medication (more NSAIDs, more opioids, more antiemetics, and more tylenol, caffeine, etc). Their initial pain rating (VAS) and pain at discharge were comparable to the placebos.

Side effects were negligible in this tiny study – with the placebo group reporting nausea, cramps, and auditory hallicuinations, and the dex group mostly complaining of tingling. The feared reaction of disseminated strongyloides did not materialize from this one-time dex dose.

There are a number of little things in this study that might seem like drawbacks, but probably aren’t – such as the small size of the study, and the use of a young, generally healthy population. And the lack of distinction between migraine and other headaches, and the lack of standardizing treatment – all of these things better reflect practice, and I think make the results more impressive. But it’s not clear why the investigators limited themselves to one dose of IV dex when IM would be just as easy to randomize and might have allowed more enrollment.

Still, many of us were impressed by this well-executed study of a simple intervention to a common complaint. It shows that RCTs don’t need to be multimillion dollar projects, and can still generate findings that can change our practice. I think of this study like flomax for nephrolithiasis  –  it’s using a common medication for a common complaint in a new way. The positive results let us feel like we’re actually doing something novel and effective for these otherwise hurting and often hard-to-satisfy patients. So, along with encouraging some caffeine consumption, many of us at journal club were looking forward to administering dexamethasone to our future benign headache patients.

More resources: ACEP policy on headaches (focus on diagnosis, neuroimaging and LP), and EMCrit.org has a great page on headache differential and workup.