A patient came to see me last month with a folder of cluster headache diagnoses from two different clinics. She had been on verapamil for four months, tried high-flow oxygen at home, and kept a sumatriptan injection in her bag — none of it touched the pain. What caught my attention wasn’t her medication list. It was the pattern: fifteen to twenty short attacks a day, each lasting well under twenty minutes, always on the right side, always with a watery eye and a blocked nostril. That’s not a typical cluster headache pattern. It’s paroxysmal hemicrania — and unlike cluster headache, it has a treatment that works completely, usually within days, once you’re on the right drug.
Why Paroxysmal Hemicrania Gets Mistaken for Cluster Headache
Paroxysmal hemicrania (PH) and cluster headache belong to the same family of disorders — the trigeminal autonomic cephalalgias. Both cause severe, one-sided pain around the eye or temple, paired with autonomic signs on the same side: tearing, redness, a stuffy or runny nostril, eyelid drooping or swelling. On the surface they look like the same disease, and it’s an easy mix-up even for experienced clinicians who don’t see PH often, since it’s far rarer than cluster headache.
The difference is in the rhythm of the attacks. Cluster headache attacks typically last 15 to 180 minutes and occur one to eight times a day, often waking patients at the same time each night. Paroxysmal hemicrania attacks are shorter — usually 2 to 30 minutes — but far more frequent, often more than five times a day and sometimes 20 to 40 times a day. PH also skews toward women, while cluster headache is roughly three times more common in men. Both can run in bouts with remission (episodic) or persist without a break for more than a year (chronic).
The Indomethacin Test Is the Diagnosis
What makes paroxysmal hemicrania distinct isn’t just its pattern — it’s that it responds completely to indomethacin, an older nonsteroidal anti-inflammatory drug. This response is so specific that it’s written directly into the diagnostic criteria: if indomethacin doesn’t fully resolve the attacks, the diagnosis isn’t PH.
A typical trial starts at 25 mg three times a day, increasing over one to two weeks toward 75 mg daily or higher if there’s no response, with some patients needing up to 225 mg a day before the effect is judged adequate. When the dose is right, the change is usually obvious within days — not the gradual, partial improvement you might see with a preventive medication, but attacks stopping almost entirely. Because indomethacin carries real risks of stomach irritation, ulcers, and kidney strain, I pair it with a stomach-protecting medication and check kidney function and blood pressure before and during the trial, especially if it’s needed long-term.
If a patient gets partial relief but not full resolution, I don’t stop there — it usually means the diagnosis needs another look, most often toward hemicrania continua, which is also indomethacin-responsive but has a different attack pattern of continuous background pain with flares, rather than PH’s discrete attacks.
What It Means If Verapamil or Triptans Already Failed You
Verapamil and oxygen are the backbone of cluster headache prevention and abortive care, and injectable triptans can shorten a cluster attack within minutes. If you’ve had a fair trial of these — proper verapamil dosing with an ECG to check the heart’s conduction, oxygen at a high enough flow rate, a triptan used correctly — and you’re still getting little to no relief, that failure is informative rather than a sign you simply need to push doses higher. I’ve written before about why verapamil fails for cluster headache and what I try next in true cluster headache; paroxysmal hemicrania is one of the reasons a workup that starts from a cluster headache diagnosis sometimes needs to be revisited entirely, not escalated.
Triptans can occasionally blunt a PH attack partially, since the pathways overlap, but they rarely produce the complete relief they give in true cluster headache. That inconsistent, partial response is itself a clue worth mentioning to your doctor.
If you’ve been treated for months without real relief, it’s worth asking for a second look at the diagnosis rather than a fourth medication trial. I go through this reasoning in more detail in why your headache isn’t getting better.
Getting the Diagnosis Right Before Starting Indomethacin
Before diagnosing any trigeminal autonomic cephalalgia as primary, I want brain imaging — usually MRI with attention to the pituitary and posterior fossa — because pituitary lesions and structural problems near the skull base can produce an identical attack pattern. A normal scan doesn’t mean nothing is wrong; it means the more dangerous explanations have been ruled out, which I discuss further in what a normal MRI actually rules out.
It’s also worth distinguishing PH from two close relatives: SUNCT/SUNA, where stabs last only seconds rather than minutes, and hemicrania continua, which I’ve written about separately because it causes a continuous daily headache rather than discrete attacks — both are also in the indomethacin-responsive family, and the distinctions matter for how the trial is run. The single most useful thing you can bring to that appointment is a diary logging exactly how long each attack lasts and how many happen per day. That detail, more than anything else, is what separates these diagnoses.
Not sure whether your symptoms need urgent attention first? Start with our free headache red-flags checklist. If you’d like a diagnosis reviewed by a neurologist, you can talk to a specialist directly.
This article is for educational purposes and is not a substitute for medical care. If you have sudden, severe, or unusual headache symptoms, seek emergency medical attention immediately.
From the textbook. Chapter 28, “Paroxysmal Hemicrania,” of my clinical textbook Headache Medicine covers this topic in full for doctors, with the evidence and decision pathways behind it. About the book →