4.4 Deep brain stimulation (DBS) of the posterior hypothalamus

4.4.1 History and rationale

historical + peer-reviewed DBS for CH grew directly out of functional imaging: PET showed the posterior hypothalamus becoming active during cluster attacks. Massimo Leone and colleagues in Milan, Italy published the first case in 2001 — Leone M et al., N Engl J Med 2001;345:1428-1429 (cited in Leone & Proietti Cecchini, Cephalalgia 2016, DOI 10.1177/0333102415607176; also Karger systematic review). Leone’s review states DBS of the posterior hypothalamic area “was first introduced in 2000 to treat drug-refractory chronic cluster headache”.

Leone (Istituto Neurologico Carlo Besta, Milan) is the field’s central figure; the Italian series is by far the largest and longest-followed. This is an Italian-origin therapy and much of the foundational work is non-Anglophone in origin, though published in English.

4.4.2 Leone’s pooled series — 79 patients

peer-reviewed Leone M, Proietti Cecchini A. “Deep brain stimulation in headache.” Cephalalgia 2016 (online 7 December 2015), DOI 10.1177/0333102415607176 (SAGE full text).

• 79 patients with hypothalamic stimulation, 88.6% chronic cluster headache; remainder SUNCT, one paroxysmal hemicrania, one symptomatic trigeminal neuralgia, one symptomatic CH-like syndrome.
• Mean follow-up 2.2 years.
• ≥50% improvement in headache frequency and/or intensity: 55 patients (69.6%).
• In the study-by-study table totals: pain-free 24 (30.4%), ≥50% improvement 31 (39.2%).

Note an internal discrepancy in the source: the abstract states 69.6% (55/79) achieved ≥50% improvement, while the compiled table totals give 24 pain-free (30.4%) plus 31 with ≥50% improvement (39.2%) — 55 patients combined, i.e. the 69.6% figure counts pain-free patients within the responder group. Read the 69.6% as “pain-free OR ≥50% improved” (Leone & Proietti Cecchini 2016).

peer-reviewed The remarkable long-term finding. Leone’s own longest-followed cohort: 17 drug-resistant chronic CH patients at a median follow-up of 8.7 years — improvement in 70% (12/17); six almost pain-free; six transformed into episodic CH. Critically: “Improvement could be maintained with the stimulators off but only after years of continuous stimulation”, which the authors say “suggested that hypothalamic stimulation can change disease course” (Leone & Proietti Cecchini 2016).

That claim — that DBS may be disease-modifying rather than merely suppressive — is unique in the CH literature and comes from a single uncontrolled series by the group that invented the therapy. It is the strongest reason DBS has not been abandoned, and it remains unreplicated. Treat as intriguing and unproven.

4.4.3 Series-by-series results and adverse events

peer-reviewed From Leone & Proietti Cecchini’s compiled table (SAGE):

Series	n	Follow-up (yrs)	Pain-free	≥50% improved	Side effects / complications
Leone (2001/2004/2006/2013)	19	8.7	6	6	Electrode displacement ×2; infection ×4; electrode malpositioning ×1; transient non-symptomatic third-ventricle haemorrhage ×1; slight unilateral muscle weakness ×1; seizure ×1
Schoenen (2005)	6	4	2	1	Fatal haemorrhage; panic attack; oculomotor disturbances
D’Andrea (2006)	3	2.5	2	0	—
Benabid (2006)	1	1	1	0	—
Starr (2007)	4	1	0	2	Transient ischaemic attack
Owen (2007)	1	0.7	1	0	—
Mateos (2007)	2	1	1	1	—
Black (2007)	2	2.6	0	2	—
Bartsch (2008)	6	1.4	2	1	—
Piacentino (2008)	4	0.4	3	1	—
Fontaine (2010) — RCT	11	1	3	3	Subcutaneous infection; transient loss of consciousness with hemiparesis; micturition syncopes
Hidding & May (2011)	1	NR	0	0	Headache; high-frequency tremor
Seijo (2011)	5	2.8	2	3	Euphoria; well-being; dizziness and oculomotor disturbances; concentration difficulties; headache; cervical dystonia; increased appetite
Kovacs (2014)	2	2	NR	2	Disrupted sleep pattern in both
Total (all indications)	79	2.2	24 (30.4%)	31 (39.2%)

4.4.4 The death — Schoenen 2005

peer-reviewed Schoenen J, Di Clemente L, Vandenheede M, Fumal A, De Pasqua V, Mouchamps M, Remacle JM, de Noordhout AM. “Hypothalamic stimulation in chronic cluster headache: a pilot study of efficacy and mode of action.” Brain 2005;128(Pt 4):940-947, PMID 15689358.

Six patients with refractory chronic CH, ipsilateral ventroposterior hypothalamic stimulation, mean follow-up 14.5 months, mean voltage 3.28V with “diplopia being the major factor limiting its increase”.

• Clinical outcome excellent in 3: two pain-free, one with fewer than three attacks per month.
• One patient had only transient remissions.
• “Another patient died” of “an intracerebral haemorrhage.”
• In one pain-free patient, switching the stimulator off led to attacks resuming after 3 months, until it was switched on again.

This is the index death in the CH DBS literature and it happened in a six-patient pilot study. It is the single fact that most shapes how this therapy is regarded.

4.4.5 Systematic review and meta-analysis — 108 cases

peer-reviewed “Deep Brain Stimulation for Chronic Cluster Headaches.” Stereotact Funct Neurosurg 2023;101(4):232-243 (Karger). PRISMA 2020 systematic review and random-effects meta-analysis; PubMed + Scopus searched 3 January 2022; 797 records screened → 16 included studies → 108 unique cases. All reports graded “good” on the Newcastle-Ottawa Scale.

Population: average age 46.6 (range 24–71); 76.9% male (n=83); 86.3% unilateral headaches; baseline 33.8 attacks/week (SD 16.7); baseline intensity 8.5/10 (SD 1.8); mean 9.5 years (SD 6.7) from diagnosis to DBS.

Targets: posterior hypothalamus 58.3% (n=63); ventral tegmental area 21.3% (n=23); pre-rubral tegmentum/mammillothalamic tract 13.9% (n=15); posterior inferior third-ventricle floor 6.5% (n=7). Unilateral leads 82.4%; local anaesthesia with minimal sedation 82.4%.

Efficacy:

Outcome	Result
Attack frequency before → after	33.8/week → 10.1/week (−70.1%)
Intensity before → after	8.5 → 4.2 (−50.6%)
Responders (≥50% improvement in frequency)	78/105 (74.3%)
Frequency mean difference (meta-analysis)	MD 20.97, 95% CI 13.58–27.01, p=0.0001
Intensity mean difference (meta-analysis)	MD 5.00, 95% CI 3.16–6.84, p<0.0001
Overall follow-up	Mean 45.4 months (SD 38.7), range 1–144
Feasibility	>99% of cases

Complications — the definitive numbers:

Category	Rate	n
Mortality	<1%	1/108 — intracerebral haemorrhage, death on postoperative day 3, “due to catastrophic bleeding after MER”
Major complications, total	16.67%	18
Electrode misplacement or breakage	7.41%	8 (4 migration/misplacement requiring replacement; 4 intracranial breakage requiring revision)
Surgical-site infection requiring additional surgery	4.63%	5
Neurological deficits	3.70%	4 (seizure — resolved; intraoperative TIA with ipsilateral hemiplegia — resolved in 5 min; irreversible dysarthria — not resolved; persistent diplopia — not resolved)
Contralateral / side-shift cluster attacks	2.8%	3
Intracerebral haemorrhage	0.93%	1
Minor complications, total	4.63%	5 (infection resolved with antibiotics 2.78%; asymptomatic third-ventricle-wall haemorrhage 0.93%; micturition syncope 0.93%)
Transient stimulation-related — diplopia	22.2%	24
Transient — changes in satiety/hunger, sexual drive/hormones	17.6%	19
Transient — vertigo or dizziness	6.5%	7

Transient stimulation-related effects were amplitude-related and mostly resolved with lower intensity (Karger 2023).

peer-reviewed An interesting technical finding: microelectrode recording (MER) was associated with better postoperative intensity scores (3.05±2.53 with MER vs 5.18±3.17 without, p=0.006) — but the single fatality was caused by bleeding after MER (Karger 2023). Better targeting, higher haemorrhage risk. That trade-off is unresolved.

peer-reviewed Conflicting haemorrhage estimates, stated explicitly: the Karger meta-analysis gives intracerebral haemorrhage at 0.93% (1/108). The EAN 2023 guideline states a bleeding risk of approximately 2% (EAN 2023). A 2014 review put “the collated risk of serious hemorrhage for DBS in CCH” at 3%, noting this was within the range reported for DBS in movement disorders (PMC4119587). The true figure is probably somewhere in 1–3%; the range is honest uncertainty from small numbers, not a discrepancy anyone has resolved.

4.4.6 The negative randomised trial

peer-reviewed Fontaine et al. (2010) ran the only randomised, prospective, crossover, double-blind study: 11 patients with severe refractory CCH, no significant difference between active and sham stimulation during the randomised phase; three serious adverse events (subcutaneous infection, transient loss of consciousness, micturition syncopes) (summarised in PMC8665918; table in Leone & Proietti Cecchini 2016).

peer-reviewed Leone’s defence of the negative result: “There is only one double-blind, randomised, controlled study investigating DBS efficacy in chronic CH… the short blind period, one month, [is] too short to give clinically significant information on DBS efficacy” (Leone & Proietti Cecchini 2016). Other reviews agree the negative result likely reflects the short blinded phase, because reported delays to response range from 1 to 86 days, with a mean delay of ~42 days (PMC4119587).

peer-reviewed The EAN also notes that a randomised trial of endoventricular tegmental stimulation was negative (EAN 2023).

So the state of evidence is: ~70–75% responder rates across 108 uncontrolled cases with long follow-up, and the only controlled trial was negative. This is precisely the pattern that ONS also shows. The difference is that DBS carries a ~1% mortality and ~17% major complication rate, so the burden of proof is much higher.

4.4.7 Acute use — doesn’t work

peer-reviewed In a larger case series only 23% of attacks improved with direct DBS stimulation during an attack, so DBS “cannot be regarded as effective in the treatment of attacks” (PMC8665918). DBS is purely preventive.

4.4.8 Current status and guideline position

peer-reviewed The EAN 2023 guideline makes no formal recommendation on DBS. It records that a death has been reported, states a bleeding risk of approximately 2%, notes reports of secondary worsening, notes the negative endoventricular tegmental RCT, and states that ONS should be attempted before DBS is planned (EAN 2023 PDF).

peer-reviewed The 2021 narrative review’s caution is the sharpest sentence in this chapter: “Only DBS electrodes should be implanted with very high caution since fatal outcome of the operation has been reported” (PMC8665918).

peer-reviewed The European Headache Federation has recommended that neuromodulation be considered only after all other medical treatments have failed, performed by tertiary headache centres, using the least invasive methods before considering DBS (PMC4119587).

How rare is it now? No source I could retrieve gives an annual worldwide figure. What can be said with confidence: the cumulative published world literature to a 2022 search date is 108 cases across roughly 22 years (Karger 2023) — under five published cases per year globally. It is performed at a handful of tertiary centres (Milan, Liège, Oxford, Grenoble, Oviedo, Barcelona, Hannover). Whether it continues in routine use anywhere in 2026 is unknown from the sources I could verify; the direction of travel in guidelines is clearly away from it. There is no realistic Australian pathway to hypothalamic DBS for cluster headache.

4.4.9 Community experience

community report I found no substantive patient-community discussion of hypothalamic DBS for cluster headache in the sources retrieved — no Reddit threads, no ClusterBusters threads, no patient blogs surfaced in searching. Given roughly 108 published patients worldwide, this is unsurprising, but it means there is effectively no community sentiment data for DBS, and any claim otherwise should be treated with suspicion. The absence should not be read as either endorsement or condemnation.