Cluster Headache

• Trigeminal System Activation and Neurogenic Inflammation: The primary pain generator involves activation of the trigeminal nerve (CN V), specifically the ophthalmic division, leading to unilateral orbital and periorbital pain. Activation of trigeminal neurons projecting to the superior ophthalmic nerve distribution stimulates release of neuropeptides including substance P, calcitonin gene-related peptide (CGRP), and neurokinin A from trigeminal nerve terminals in the meninges and orbital tissues. These neuropeptides trigger local neurogenic inflammation characterized by plasma extravasation, vasodilation, and recruitment of inflammatory cells. Functional imaging studies (fMRI, PET) demonstrate ipsilateral activation of the trigeminal nucleus caudalis and rostral medulla during spontaneous cluster attacks, confirming the central trigeminal involvement. The remarkable responsiveness of cluster headache to triptans and ergotamines—medications that cause cranial vasoconstriction and inhibit trigeminal neurotransmission—supports trigeminal system dysfunction as central to pathogenesis.

• Hypothalamic Dysfunction and Circadian/Circannual Rhythm Abnormality: A key distinguishing feature of cluster headache pathophysiology is the profound abnormality in circadian timing, with attacks typically occurring at similar times each day during cluster periods and demonstrating strong seasonal patterns (more common in spring and fall in many populations). Neuroimaging studies, particularly high-field fMRI, have identified specific hypothalamic hyperactivity in the lateral hypothalamus (specifically the H1 region with posterior involvement) during cluster periods, localizing an apparent "cluster headache generator." The hypothalamus controls both the suprachiasmatic nucleus (master circadian clock) and autonomic output through the posterior and lateral hypothalamic nuclei. Anatomical dysfunction in these regions would explain both the clocklike periodicity of attacks and the prominent autonomic features. Importantly, the hypothalamus also contains abundant orexinergic neurons and melanin-concentrating hormone (MCH) neurons, both implicated in pain modulation and circadian regulation. The hypothesis that cluster headache represents a disorder of hypothalamic-mediated pain timing and autonomic regulation explains why attacks occur predictably during circadian windows and seasonal periods. Supporting this mechanism, studies demonstrate altered cortisol rhythmicity and melatonin secretion patterns in cluster headache patients.

• Parasympathetic Autonomic Dysfunction: The characteristic ipsilateral autonomic symptoms (lacrimation, nasal congestion, conjunctival injection, rhinorrhea, ptosis, miosis) indicate parasympathetic hyperactivity, suggesting dysfunction in autonomic balance. These symptoms reflect activation of parasympathetic fibers via the facial nerve (CN VII), which carries preganglionic parasympathetic fibers to the lacrimal gland, nasal mucosa, and parotid gland. The anatomical pathway involves connections between the trigeminal nucleus and the superior salivatory nucleus, creating a trigeminoparasympathetic reflex arc. In cluster headache, this reflex appears profoundly dysregulated, with massive parasympathetic outflow during attacks. The sphenopalatine ganglion, which relays parasympathetic input to cranial blood vessels and mucous membranes, shows increased metabolic activity during attacks. Remarkably, blocking the sphenopalatine ganglion (via transnasal balloon compression or newer implantable devices) can terminate acute attacks, directly demonstrating the importance of parasympathetic activation. The dominance of parasympathetic over sympathetic autonomic tone is emphasized by the presence of miosis and ptosis (signs of relative sympathetic deficiency rather than parasympathetic excess alone), suggesting dysregulation of autonomic balance rather than simple parasympathetic hyperactivity.

• Vascular and Perivascular Dysfunction: Early theories emphasizing vasodilation and vasoconstriction cycles as the primary mechanism have evolved, though vascular changes remain contributory. Angiographic studies demonstrate marked dilation of the ipsilateral internal carotid artery and its branches during acute attacks, though this appears consequential rather than causative to the pain. The vascular changes likely result from trigeminal activation causing local CGRP and substance P release, which promote vasodilation through endothelial and smooth muscle effects. Importantly, the remarkable efficacy of 100% oxygen inhalation in aborting acute attacks (up to 70-80% effectiveness) suggests that vascular mechanisms remain mechanistically important, as oxygen likely works through vasoconstriction and enhanced vascular oxygen saturation. The perivascular release of inflammatory mediators creates a sterile inflammatory response in meningeal and extracranial tissues, perpetuating pain signaling.

• Genetic and Neurochemical Factors: Approximately 5-10% of cluster headache shows familial clustering, suggesting genetic predisposition, though no single causative gene has been identified. Polymorphisms in genes related to circadian rhythm regulation (CLOCK, BMAL1) and pain processing have been investigated. The hypocretin (orexin) system is particularly interesting given the hypothalamic involvement; orexin neurons regulate both pain perception and circadian function, and dysfunction of this system could theoretically produce the cluster headache phenotype. Abnormalities in serotonin and dopamine signaling have been described, with alterations in monoamine oxidase activity and serotonin transporter expression. The exceptional sensitivity to alcohol as a trigger during cluster periods may involve acetaldehyde metabolism or direct effects on hypothalamic function.

• Male Gender: Males demonstrate a 3-4:1 predominance over females for cluster headache, with some studies suggesting ratios as high as 6:1. The biological basis for this striking sex difference remains incompletely understood but likely involves sex hormone influences on pain processing and autonomic regulation. Some evidence suggests women may present with later onset and less frequent attacks, possibly reflecting protective effects of estrogen on nociceptive pathways or differential autonomic regulation.

• Age of Onset: Typical onset occurs in the 20s-40s age range, with peak incidence in the 30s. Approximately 85-90% of cluster headache cases begin before age 50. Onset in childhood or after age 60 is uncommon but well-documented, and later-onset cases may have different prognostic features. The timing of typical onset in early adulthood suggests involvement of developmental maturation of hypothalamic and trigeminal systems.

• Smoking: Active smoking is associated with increased cluster headache prevalence and possibly increased attack frequency, with some series showing smoking in up to 50-80% of cluster patients compared to population prevalence of 15-20%. Whether smoking is causative or represents symptom self-medication remains debated. Smoking cessation may improve attack frequency and is universally recommended.

• Alcohol Consumption: Alcohol is a powerful attack trigger in the majority of episodic cluster patients during active cluster periods, with beer and whiskey being particularly provocative. The trigger effect is exquisitely time-dependent and period-dependent—alcohol reliably triggers attacks within 30-60 minutes during active cluster periods but has no effect during remission periods, highlighting the cyclical vulnerability that defines cluster headache. The mechanism involves acetaldehyde metabolites, direct hypothalamic effects, or vasodilation, though the precise pathway remains unclear. Patients in remission phases can often consume alcohol freely without triggering attacks.

• Stimulant Use: Cocaine, methamphetamine, and other sympathomimetic stimulants can trigger cluster headaches, likely through trigeminal stimulation and vascular effects. Caffeine has mixed effects; some patients report triggering while others experience benefit during attacks.

• Circadian and Seasonal Factors: While not modifiable risk factors, temporal patterns dramatically influence cluster headache presentation. The circadian predictability (attacks at similar clock times) and seasonal clustering (spring and fall predominance in many populations) reflect underlying hypothalamic dysfunction. Some patients show clear seasonal patterns while others demonstrate random clustering across the year.

• Familial Aggregation: Approximately 5-10% of cluster headache patients have first-degree relatives with cluster headache, suggesting genetic predisposition. Monozygotic twin concordance data are limited but suggest partial genetic determination. Autosomal dominant inheritance with incomplete penetrance has been proposed in some families.

• Oxygen Tension and Altitude: Some evidence suggests that high altitude or lower atmospheric oxygen pressure may trigger cluster periods. This could relate to the effectiveness of oxygen therapy and suggests that oxygen metabolism may be relevant to cluster pathophysiology.

• Head Trauma: Occasional reports describe cluster headache onset following head trauma, though this association remains anecdotal and separate from post-traumatic headache syndromes. The temporal relationship remains unclear.

• Unilateral Orbital/Periorbital Pain: The cardinal feature is severe, strictly unilateral pain centered in the orbital and periorbital region, typically localized to the distribution of the ophthalmic division of the trigeminal nerve. Pain is often described as "boring," "sharp," "burning," or "throbbing"—though the exact descriptor varies—and is distinctly different from the throbbing bilateral quality of migraine. The pain is consistently ipsilateral (always on the same side during a cluster period, though the side can switch between cluster periods in rare cases). Pain intensity is excruciating, rated 8-10/10 on visual analog scales, often described as "the worst pain imaginable" or "worse than childbirth or surgery." The pain may radiate to the temporal region, cheek, jaw, teeth, or nose on the affected side. Importantly, patients may report pain radiating posteriorly to the occiput or neck, but the primary epicenter is always in or behind the eye. This severity and unilateral orbital location immediately distinguish cluster from migraine or tension-type headache.

• Duration and Frequency of Attacks: Individual cluster attacks typically last 15-180 minutes, with a mean duration of 30-45 minutes. Attacks typically occur 1-8 times per day during active cluster periods, with many patients experiencing nightly attacks at the same or similar time. This remarkable temporal periodicity—often to within 30 minutes of the previous night's attack—is virtually pathognomonic for cluster headache. Patients often report awakening from sleep at a characteristic time (e.g., 2 AM) with cluster pain, a distinctive feature that helps differentiate from migraine (which may awaken patients but without the rigid periodicity). The frequency distribution follows a Poisson distribution, with some attacks closer together and others more spaced, but the overall temporal clustering during active periods is unmistakable.

• Ipsilateral Autonomic Symptoms: Autonomic symptoms occur in 80-100% of cluster attacks and are invariably ipsilateral (same side as pain), representing a defining diagnostic feature. The classic autonomic symptoms include:
• Conjunctival injection (80-90% of attacks): Redness of the conjunctiva on the affected side, often dramatic and clearly unilateral
• Lacrimation (up to 80%): Profuse unilateral tearing, often soaking tissues
• Nasal congestion/rhinorrhea (up to 80%): Unilateral or ipsilateral-predominant nasal symptoms, with clear rhinorrhea described in many attacks
• Ptosis (up to 65%): Drooping of the ipsilateral eyelid, though typically incomplete (partial ptosis) and mild
• Miosis (up to 60%): Constricted ipsilateral pupil, often subtle and requiring careful comparison to the contralateral side
• Facial flushing/sweating (up to 40%): Ipsilateral facial redness or hyperhidrosis, reflecting parasympathetic activation
• Eyelid edema (up to 40%): Unilateral periocular puffiness from local inflammation and vasodilation

• Behavioral/Functional Features: During acute cluster attacks, patients cannot remain still and characteristically pace, rock, or engage in repetitive movements—in stark contrast to migraine patients who seek dark, quiet rooms and remain motionless. This restlessness and agitation during attacks is a reliable clinical sign. Many patients describe the need to apply pressure or cold to the eye region during attacks, suggesting an attempt to block nociceptive input. Some patients report that physical exertion worsens attacks, and they avoid exercise during cluster periods.

• Cluster Period Characteristics: The cluster period (weeks to months during which attacks occur regularly) is separated from remission periods (weeks to months of complete freedom from attacks) by clear temporal boundaries. During active cluster periods, patients experience remarkable predictability—many can state within 30 minutes when the next attack will occur. The timing of clusters shows seasonal patterns in many patients (spring and fall predominance), though others show random clustering. Remission periods are typically characterized by complete freedom from headaches; patients in remission do not experience warning symptoms or mild versions of cluster pain.

• Premonitory Symptoms: Unlike migraine, cluster headache does NOT have classic premonitory symptoms or aura. However, some patients report a brief (seconds to minutes) warning sensation before cluster pain begins, described as a tingling, pressure, or mild sensation in the orbital region. Others report that they simply awaken from sleep in pain. The absence of aura (visual, sensory, or motor phenomena) is important for differential diagnosis.

• Important Clinical Variants:
• Episodic Cluster Headache (80-90% of cases): This is the typical presentation with clearly demarcated cluster periods separated by headache-free remission periods, lasting weeks to months. Patients are completely pain-free between cluster periods.
• Chronic Cluster Headache (10-20% of cases): Characterized by cluster periods lasting >1 year without remission of 1 month or greater. Pain-free intervals are absent or very brief (<1 month between "mini-clusters"). Chronic cluster headache is more resistant to preventive therapy and may require escalation to more aggressive interventions.
• Chronic Remitting Cluster Headache: Chronic cluster with clear remissions, representing an intermediate pattern.
• Episodic Chronic Cluster Headache: A case where episodic cluster transitions to chronic, a troubling evolution that occurs in approximately 10-15% of episodic patients.
• Primary Stabbing Headache Coexisting with Cluster: Some cluster patients report brief (1-5 seconds) sharp, stabbing pains in the same orbital region between cluster attacks, distinct from the longer cluster attacks; this may represent related pathophysiology.

• Diagnostic Criteria: The International Classification of Headache Disorders, 3rd Edition (ICHD-3) provides explicit criteria for cluster headache diagnosis:

Cluster Headache Diagnostic Criteria (ICHD-3)

Acute (abortive) therapy — must be fast-acting, because attacks peak within minutes

• High-flow oxygen: 100% O2 by non-rebreather mask at roughly 12–15 L/min for 15–20 minutes, patient seated and leaning forward. Works via cranial vasoconstriction and inhibition of trigeminal–parasympathetic outflow. First-line, no drug interactions, repeatable. Nasal cannula flow rates are inadequate — a classic distractor. Use judiciously in severe COPD with chronic hypercapnia, where high inspired oxygen can blunt hypoxic respiratory drive and worsen CO2 retention.
• Triptans, parenteral or intranasal: sumatriptan 6 mg subcutaneously is the best-studied; intranasal zolmitriptan is an alternative. 5-HT1B/1D agonism constricts cranial vessels and blocks CGRP release. Labeling limits subcutaneous sumatriptan to two 6 mg injections per 24 hours, separated by at least 1 hour. Oral triptans are too slow for an attack lasting 30–45 minutes.
• Intranasal lidocaine or dihydroergotamine are second-line adjuncts.

Transitional therapy

• Corticosteroids: an oral prednisone burst-and-taper, or suboccipital/greater occipital nerve injection with steroid — the injection carries the strongest AAN-graded evidence and avoids systemic exposure.

Preventive therapy (start on day one of the cluster period)

• Verapamil (non-dihydropyridine calcium channel blocker) is the mainstay; high doses are often needed. Obtain a baseline ECG and repeat with each dose escalation to detect PR prolongation and AV block.
• Lithium (favored in chronic cluster), topiramate, and melatonin are alternatives.
• Galcanezumab, a CGRP monoclonal antibody, is FDA-approved for episodic cluster headache.

Disease-related

• Suicidal ideation and completed suicide: the pain severity earns the label "suicide headache"; rates of suicidal thinking are substantially higher than in migraine. Signaled by expressed hopelessness, planning, or escalating self-medication. This is a psychiatric emergency — screen at every visit.
• Persistent partial Horner syndrome: repeated attacks can leave residual ptosis and miosis between attacks from sympathetic fiber injury along the carotid. Anhidrosis is typically absent because sudomotor fibers travel with the external carotid.
• Transformation from episodic to chronic cluster — ICHD-3 defines chronic cluster headache as attacks occurring for ≥1 year without remission, or with remissions lasting <3 months; this transformation predicts poorer treatment response.
• Misdiagnosis of a secondary mimic is the dangerous "complication" of premature closure: carotid artery dissection (neck pain, Horner syndrome, stroke risk — an emergency), pituitary adenoma or apoplexy (visual field loss, endocrinopathy), and cavernous sinus or posterior fossa lesions. Any cranial nerve deficit persisting between attacks, or a first attack after age 50, mandates neuroimaging.

Treatment-related

• Verapamil: PR prolongation progressing to high-grade AV block and bradycardia — detected only by serial ECGs, since patients are often asymptomatic; symptomatic bradyarrhythmia is an emergency. Also constipation, peripheral edema, gingival hyperplasia.
• Corticosteroids: hyperglycemia, HPA-axis suppression, and with repeated bursts avascular necrosis of the femoral head — signaled by new groin pain on weight-bearing.
• Lithium: narrow therapeutic index; tremor, nephrogenic diabetes insipidus, hypothyroidism; toxicity (confusion, ataxia, seizures) is precipitated by dehydration, NSAIDs, thiazides, or ACE inhibitors and is an emergency.
• Topiramate: nephrolithiasis, metabolic acidosis, cognitive slowing, oligohidrosis, and acute angle-closure glaucoma — sudden eye pain and blurred vision, an ophthalmologic emergency easily mistaken for a cluster attack.
• Triptans/ergots: coronary vasospasm, myocardial ischemia, and with ergot overuse ergotism with limb ischemia.
• Medication-overuse headache and opioid dependence from inappropriate analgesic escalation.

• The stem's giveaway triad: severe strictly unilateral orbital/retro-orbital pain, ipsilateral autonomic features (lacrimation, conjunctival injection, rhinorrhea, partial ptosis/miosis), and a restless, pacing patient. Migraine patients lie still in a dark room — restlessness versus immobility is the single most tested discriminator.
• The clock is diagnostic: attacks lasting 15–180 minutes that wake the patient at the same hour nightly, in bouts recurring seasonally, reflect the hypothalamic generator. Alcohol triggers attacks only during an active bout, not in remission.
• Best next step for an acute attack: 100% oxygen by non-rebreather at high flow, or subcutaneous sumatriptan 6 mg. The classic wrong answer is an oral triptan or an NSAID — the attack peaks and resolves faster than they absorb. Opioids are never the answer.
• Best next step for prevention: verapamil, started at bout onset — and the paired question is almost always "what do you check first?" → a baseline ECG, repeated with dose escalations, to catch PR prolongation and AV block.
• The bridge: a prednisone taper or a greater occipital nerve block with steroid covers the weeks before verapamil takes effect. Examiners like the occipital nerve block as the steroid-sparing choice.
• The association tested: male predominance and heavy smoking, with a hypothalamic "generator" on functional imaging — cluster is the trigeminal autonomic cephalalgia whose pathology is hypothalamic, not vascular.
• The mimic to exclude: unilateral orbital pain with Horner syndrome can be carotid artery dissection; persistent cranial neuropathy, papilledema, or first onset after age 50 buys an MRI with vascular imaging before calling it primary.
• Distractor to avoid: attacks lasting seconds to minutes, dozens per day, with conjunctival injection and tearing are SUNCT/SUNA; attacks lasting minutes that abolish completely with indomethacin are paroxysmal hemicrania — cluster does not respond to indomethacin.