1. Classification

Episodic vs chronic cluster headache

CH is defined by ICHD-3 (International Classification of Headache Disorders, 3rd edition) as attacks of severe, strictly unilateral orbital, supraorbital and/or temporal pain lasting 15–180 minutes untreated, occurring between once every other day and eight times a day, accompanied by ipsilateral cranial autonomic signs (conjunctival injection/lacrimation, nasal congestion/rhinorrhoea, eyelid oedema, forehead/facial sweating, miosis/ptosis) and/or restlessness or agitation peer-reviewed12. (Editor’s note: the verbatim ICHD-3 criteria are reproduced in Part I, §2.1.)

• Episodic cluster headache (ECH): attacks occur in bouts of 7 days to 1 year (untreated), separated by pain-free remission periods of 3 months or more; requires at least two such bouts peer-reviewed3.
• Chronic cluster headache (CCH): attacks persist for a year or longer with no remission, or with remissions lasting less than 3 months peer-reviewed4. CCH can arise de novo (“primary chronic”) or evolve from ECH (“secondary chronic”) peer-reviewed4.
• The 1-month → 3-month criterion change matters and is a real source of confusion in circulating literature. ICHD-2 and the ICHD-3-beta draft used a ≥1-month remission cutoff for episodic CH; the final ICHD-3 (2018) text raised this to ≥3 months, on the reasoning that shorter “remissions” often represented under-treated chronic disease rather than a true separate episodic bout peer-reviewed773. Many older papers, some patient-facing summaries, and even some newer secondary sources still quote the 1-month figure — this chapter uses the current ≥3-month ICHD-3 standard throughout, and any 1-month figures cited below are flagged as pre-2018 sourcing.
• Proportions vary more across studies and countries than pop-science summaries usually admit. ICHD-3 commentary and most reviews put episodic CH at roughly 80–90% and chronic at 10–15% peer-reviewed15. But cohort-level numbers scatter: a German dissertation cohort found 82.8% episodic / 9.7% primary-chronic / 7.5% secondary-chronic peer-reviewed58; a German bibliographic dissertation found 72.9% episodic / 27.1% chronic peer-reviewed59; a large German-population epidemiological summary (Ärzteblatt) states 85% episodic / 15% chronic peer-reviewed60; and a Japanese clinic cohort found chronic CH prevalence as low as 2.8%, one of several markers of a possible East-Asian phenotype difference discussed further below peer-reviewed57.

Conversion between forms — the numbers, precisely

This is an area where the older, frequently-repeated round numbers (“about 10–15% convert”) turn out to understate how genuinely fluid CH phenotype is. The best current data come from the Danish Headache Center (Rigshospitalet), using structured interview-based follow-up of a large, well-characterised cohort:

• An interview-based follow-up study of 430 CH patients from the Danish Headache Center found a total transition-incidence of 20.7% — one-fifth of the whole cohort experienced at least one phenotype change during the disease course. Of the whole cohort, 14.4% transitioned from episodic to chronic, and 6.3% transitioned from chronic to episodic. Side-shifting attacks (pain moving to the other side over time) predicted transition (p = 0.007) peer-reviewed61.
• A follow-up study of the same/related cohort (430 patients, re-interviewed) found: 1-year transition rate 6.5%, 5-year transition rate 19.8% for the whole cohort. Specifically, the risk of episodic→chronic conversion was 4.0% at 1 year and 12.3% at 5 years; the risk of chronic→episodic conversion was markedly higher — 11.1% at 1 year and 25.0% at 5 years. Side-shifting attacks were reported in 32% of chronic patients and carried an odds ratio of 2.24 for being chronic rather than episodic peer-reviewed6263.
• An older but influential ten-to-twenty-five-year natural-history follow-up (Manzoni et al.) of 123 episodic and 9 chronic patients found: of initially episodic patients, 80.7% remained episodic, 12.9% shifted to chronic (“secondary chronic”), and 6.4% shifted to a “combined” pattern. Of initially chronic patients, 52.4% remained chronic, 32.6% reverted to episodic, and 14.3% shifted to combined peer-reviewed6465. This 32.6%/33% chronic-to-episodic reversion figure recurs across several reviews as “about a third,” sometimes rounded up to “as many as 50%” in older secondary literature peer-reviewed66 — treat the 50% figure as an upper-bound outlier rather than a central estimate.
• A large multicentre South Korean study specifically tracking frequent relapse (not just chronic conversion) found a relapse rate of 0.29 per person-year (95% CI 0.27–0.32) among episodic patients followed for a mean of 4.2 years, and separately found that only 3.8% of first-onset CH newly became chronic, and 1.4% of those with an episodic history newly became chronic, during prospective observation — notably lower than the Danish/Italian retrospective figures, possibly reflecting shorter follow-up windows or a different at-risk population peer-reviewed67.
• The European Academy of Neurology’s 2023 guideline states plainly: up to 12% of episodic patients progress to chronic; primary chronic presentations make up ~15% of all CH; and reversion from chronic to episodic “can also be seen, although rarely” peer-reviewed68 — this “rarely” framing sits in direct tension with the Danish Headache Center’s 25% five-year reversion figure and Manzoni’s 32.6%, and is presented here as an open discrepancy rather than resolved.
• Total, spontaneous remission (permanent cessation of all cluster activity) is documented but not well quantified: one source states remission after a single cluster period occurred in only 17% of a small cohort after 3+ years of follow-up, while most patients experience recurrence peer-reviewed69; a separate source notes “total remission of the disease has been described” without giving a rate peer-reviewed70.

Refractory chronic cluster headache (rCCH) is not a formally separate ICHD-3 entity but has working criteria from the European Headache Federation: at least three severe CCH attacks per week impacting quality of life despite treatment, and failure of at least three evidence-based prophylactic agents at maximum tolerated dose, with symptomatic chronic CH ruled out via MRI/MRA peer-reviewed6.

Where CH sits within the TACs

The term “trigeminal autonomic cephalalgia” was coined by Goadsby and Lipton in their landmark 1997 Brain paper, grouping short-lived, unilateral, trigeminally-distributed headaches with prominent cranial parasympathetic autonomic features, and separating them from headaches with sparse/no autonomic activation (trigeminal neuralgia, hypnic headache, cough headache, etc.) peer-reviewed89. ICHD-2 (2004) formally established the TAC category with three members (CH, paroxysmal hemicrania, SUNCT); ICHD-3 (2013/2018) expanded the group by adding SUNA as a sibling to SUNCT (both now under the umbrella “short-lasting unilateral neuralgiform headache attacks”, SUNHA) and reclassifying hemicrania continua from “other primary headaches” into the TACs, on the basis of shared hypothalamic activation and cranial autonomic features peer-reviewed810. The unifying pathophysiological thread across the TAC family is thought to be excessive reflex activation of the trigeminal-autonomic reflex arc, permitted by dysfunction in the posterior hypothalamic grey matter peer-reviewed9.