Cluster Headache: High-Flow Oxygen's Real Access Problem Against a Cardiac Risk Profile
High-flow oxygen is one of the best-evidenced, side-effect-free acute treatments in headache medicine, yet genuinely hard to get covered because cluster headache patients have normal oxygen saturation between attacks. The option easiest to get covered instead carries real cardiovascular caution for a patient with his specific risk profile.
D.K. timed tonight’s attack the way he’s learned to time all of them, without wanting to: eleven minutes from the first stab behind his right eye to the point where he was rocking in the ED waiting room chair, unable to sit still, tears running from just the one eye. He is 44, drives a delivery route he could run from memory by now, has smoked a pack a day for about twenty-five years and has quit twice, both times during a cluster period, and has had episodic cluster headache for a decade — this bout, his second cluster period this year, running about three weeks in with one or two attacks most days. His blood pressure, checked at triage, was 152/94, in line with what his chart already shows: hypertension, documented for four years, controlled inconsistently more by refill gaps than by his amlodipine actually failing him.
High-flow oxygen — twelve to fifteen liters a minute through a non-rebreather mask, not the low-flow nasal cannula most people picture — is genuinely one of the better-evidenced acute treatments in headache medicine, Cohen’s 2009 randomized trial aborted 78% of attacks at fifteen minutes on 12 liters a minute against 20% on air, with essentially no systemic side effect to weigh against that — and it enrolled cluster patients on their own attacks, not a mixed headache population, so the result describes his condition directly. It is also, for reasons that have very little to do with the pharmacology, one of the hardest treatments in headache medicine to actually get: home oxygen concentrators and cylinders are typically covered by insurance for hypoxemic patients, and a cluster headache patient’s oxygen saturation is normal between and often during attacks, so a straightforward durable-medical-equipment order routinely gets denied without a specific letter of medical necessity most primary care visits don’t have time to write.
Sumatriptan injection is the practical fallback that gets covered without a fight. The triptan label names documented ischemic coronary disease and uncontrolled hypertension as contraindications, and he has neither on paper — no cardiac workup has ever been done, and his hypertension carries a diagnosis rather than an uncontrolled label. By the label’s own terms he is eligible, though the terms are doing less work than they appear to: he has no documented coronary disease because no one has ever looked for it in a forty-four-year-old with a twenty-five-pack-year history.
In the ED, mid-attack, deciding tonight's treatment and the longer plan at once
His long-term acute treatment should be high-flow oxygen. Cohen's 2009 randomized trial in JAMA showed 78% of attacks aborted at fifteen minutes on 12 liters a minute against 20% on air, with essentially no systemic side effect — genuinely valuable for a patient with his cardiovascular profile. I know it won't help him in this room tonight, but I want the letter of medical necessity started today, not after we've defaulted to something else and never circled back.
Oxygen is genuinely the better answer longer-term, I'm not disputing the trial. But he needs relief from tonight's attack, and home oxygen isn't sitting in his apartment waiting for him even if the LOMN clears next week. His risk factors are real, but the actual labeled contraindication for sumatriptan is documented CAD or truly uncontrolled hypertension — not risk factors in the abstract. One elevated reading tonight isn't the same as an uncontrolled-hypertension diagnosis.
I want to give him sumatriptan injection now so he isn't sitting here in pain for another hour while we build the longer plan.
I'd push back on reading tonight's number as just one reading. He's a twenty-five-pack-year smoker with four years of documented, inconsistently-controlled hypertension, and tonight's reading is consistent with that pattern, not an outlier from it. That combination is different in kind from risk factors we're treating in the abstract.
Let's do a same-visit BP recheck and a brief cardiac risk assessment, and use non-invasive vagus nerve stimulation as the bridge tonight instead — it's FDA-approved for acute cluster headache and carries none of the vasoconstrictive question. If it doesn't abort the attack and his risk assessment comes back reassuring, sumatriptan is still there.
One caveat about that recheck, and it cuts against my own proposal: a cluster attack is an autonomic event. Whatever we measure while he is still in it, or minutes after it breaks, is partly the attack talking rather than his baseline. A reassuring number tonight shouldn't be read as evidence his hypertension is controlled — that question belongs to the primary care follow-up. Tonight's reading can only tell us whether he is frankly hypertensive enough to stop us.
Agreed: non-invasive vagus nerve stimulation tried tonight in the ED, home oxygen ordered with a letter of medical necessity documenting his specific cardiovascular risk profile, blood pressure rechecked before discharge, and a short-interval primary care follow-up scheduled for his hypertension control.
The Emergency Medicine Physician's threshold is to proceed straight to sumatriptan rather than wait further, given how severe cluster attacks are to sit through untreated.
The Clinical Pharmacologist's threshold is to wait for the same-visit risk assessment to fully return first. Not agreed between them — left as a real branch point for whoever is in the room if it comes to that.