2. Transitional / Bridge Therapy

2.1 Why bridging exists at all

peer-reviewed Verapamil is the first-line preventive, and it works — in the randomised trial 80% of patients had a halving of attack frequency and 26% became attack-free (Aust Prescr 2022). But it must be titrated slowly under ECG surveillance:

• Start 80 mg three times a day for at least 2 weeks, then increase by 80 mg every 2 weeks, range 240–960 mg/day (Aust Prescr 2022).
• EAN specifies ≥240 mg/day as the effective threshold (EAN 2023).
• ECG before starting and at every dose change, repeated at a stable dose after 10 days, then every 1–2 months, then every 6 months.
• One in five patients will develop an arrhythmia — first- or second-degree heart block, junctional rhythms, right bundle branch block, bradycardia — and delayed-onset arrhythmias have been reported (Aust Prescr 2022).

Do the arithmetic. Starting at 240 mg/day and increasing by 80 mg every two weeks, reaching 480 mg takes six weeks; reaching 960 mg takes three months. During that entire period the patient is having attacks at full frequency, capped at two triptan injections a day.

That gap is what bridge therapy exists to fill. It is not an optional refinement. peer-reviewed Australian Prescriber: “Bridging therapies are frequently used at the start of a cluster to control attacks while up-titrating preventive therapy” (Aust Prescr 2022).

A note on the chronic case. Bridge therapy is conceptualised around episodic CH — get the patient through the first weeks of a bout. In chronic CH there is no bout to get through; the same intervention is being used to buy relief while a preventive is optimised, or simply to buy relief. The evidence base reflects the episodic framing, and this shows up in the response rates below.

2.2 Corticosteroids

The only RCT

peer-reviewed Obermann M, Nägel S, Ose C, Sonuc N, Scherag A, Storch P, Gaul C, Böger A, Kraya T, Jansen J-P, Straube A, Freilinger T, Kaube H, Jürgens TP, Diener H-C, Katsarava Z, Kleinschnitz C, Holle D. “Safety and efficacy of prednisone versus placebo in short-term prevention of episodic cluster headache: a multicentre, double-blind, randomised controlled trial.” Lancet Neurol. 2021;20(1):29–37 (PMID 33245858, DOI 10.1016/S1474-4422(20)30363-X). EudraCT 2011-006204-13; DRKS00004716. German multicentre trial — 10 German headache centres, April 2013 to January 2018. Stopped early due to slow recruitment and expired funding.

Regimen: prednisone 100 mg orally for 5 days, then taper by 20 mg every 3 days — total 17 days. All patients simultaneously started verapamil 40 mg three times daily, titrated to 120 mg three times daily by day 19 — i.e. the trial tested exactly the real-world bridging scenario.

118 enrolled, 116 randomised (57 prednisone / 59 placebo), mITT 109 (53/56).

Outcome	Prednisone	Placebo	Difference
Mean attacks, first week	7.1 (SD 6.5)	9.5 (SD 6.0)	−2.4 (95% CI −4.8 to −0.03), p=0.002
Patients with adverse events	37/52 (71%)	39/55 (71%)	—
Total adverse events	135	135	—
Serious adverse events	0	2 (inguinal hernia; severe CH deterioration)	—

Most common prednisone adverse effects: headache, palpitations, dizziness, nausea.

Read the effect size honestly. A reduction of 2.4 attacks in the first week — from 9.5 to 7.1 — is statistically significant and clinically worth having, but it is not dramatic. The confidence interval nearly touches zero. This is the entire randomised evidence base for oral steroids in cluster headache, and it is in episodic CH only. Anyone describing oral prednisone as a reliably effective bridge is going beyond what this trial shows.

Regimens in use

Source	Regimen
EAN 2023 peer-reviewed	250 or 500 mg IV prednisone in the morning, or 60–100 mg orally for 5 days then reduce 10 mg every 4 days
Obermann 2021 RCT peer-reviewed	100 mg × 5 days, then −20 mg every 3 days (17 days total)
Australian Prescriber peer-reviewed	Prednisolone 1 mg/kg, max 75 mg daily, with gastric ulcer prophylaxis, down-titrated over two weeks
Leroux & Ducros review peer-reviewed	“at least 50 mg prednisone or equivalent daily for 1 week then a taper,” cumulative ~500 mg or more

Sources: EAN 2023; Obermann 2021; Aust Prescr 2022; Leroux & Ducros, Curr Pain Headache Rep 2013;17:325.

The recurrence problem

peer-reviewed “Recurrence of attacks upon weaning is a common problem” (Leroux & Ducros 2013). Ambrosini et al. put it more bluntly: “Oral steroids can interrupt bouts of cluster headache attacks, but recurrence is frequent and may lead to steroid-dependency” (Ambrosini et al., Pain 2005).

peer-reviewed Australian Prescriber: “Prolonged use should be minimised because of adverse effects” (Aust Prescr 2022).

This is the central problem with oral steroids in chronic CH. In episodic CH you take a two-week course and the bout ends anyway. In chronic CH there is no natural endpoint, so the patient hits the taper and the attacks come back — creating pressure for repeated or continuous courses, which is where avascular necrosis, osteoporosis, diabetes, weight gain and adrenal suppression live. Oral steroids are a poor fit for chronic daily CH specifically because the disease does not stop when the course does. This is the strongest argument for the injection route below.

citizen science Interestingly, in the systematic review of survey studies, corticosteroids were rated the most effective conventional prophylactic treatment, closely followed by verapamil (Rusanen et al. 2022). Patients rate steroids highly — which is consistent both with them working and with them being remembered fondly by people whose bouts ended.

2.3 Greater occipital nerve (GON) / suboccipital steroid injections

This is, in my reading, the most under-appreciated intervention in the whole acute-and-bridge space: two positive randomised controlled trials, a clean safety record, and effect sizes that dwarf oral steroids.

RCT 1 — Ambrosini (Italy/Belgium)

peer-reviewed Ambrosini A, Vandenheede M, Rossi P, Aloj F, Sauli E, Pierelli F, Schoenen J. “Suboccipital injection with a mixture of rapid- and long-acting steroids in cluster headache: a double-blind placebo-controlled study.” Pain. 2005;118(1–2):92–6 (PMID 16202532, DOI 10.1016/j.pain.2005.07.015). Headache Clinic, INM Neuromed IRCCS, Pozzilli, Italy. Italian/Belgian collaboration.

23 outpatients (16 episodic, 7 chronic), one-week run-in, randomised to 13 verum / 10 placebo (physiological saline). Episodic patients were in a new bout of no more than one week’s duration. Single suboccipital injection of a mixture of long- and rapid-acting betamethasone, ipsilateral to the attacks, in the region of the greater occipital nerve. Injections by three investigators; follow-up by four blinded investigators at 1 and 4 weeks.

Outcome	Verum	Placebo	p
Attack-free in the first week	11/13 (85%) — including 3 chronic patients	0/10	0.0001
Remained attack-free at 4 weeks	8 of those 11	—	0.0026
Remission lasting 4–26 months	5 patients	—	—

Authors’ conclusion: “A single suboccipital steroid injection completely suppresses attacks in more than 80% of CH patients,” maintained for at least 4 weeks in the majority.

Placebo response was zero. In a disease where drug-trial placebo response runs 14–43% (EAN 2023), a 0/10 placebo arm and an 85% verum arm is a striking separation — though with n=23 the confidence intervals are wide.

RCT 2 — Leroux (France)

peer-reviewed Leroux E, Valade D, Taifas I, Vicaut E, Chagnon M, Roos C, Ducros A. “Suboccipital steroid injections for transitional treatment of patients with more than two cluster headache attacks per day: a randomised, double-blind, placebo-controlled trial.” Lancet Neurol. 2011;10(10):891–7 (PMID 21903477, DOI 10.1016/S1474-4422(11)70186-7, NCT00804895). Emergency Headache Centre, Lariboisière Hospital, Paris. French source. Received no funding. Editorial comment: PMID 21903478.

The eligibility criterion is precisely the population this chapter is about: adults aged 18–65 with more than two cluster headache attacks per day. 43 randomised (15 chronic, 28 episodic), stratified, blocks of four. Injectors were unblinded; patients and the evaluating physician were blinded. ITT analysis.

Intervention: three suboccipital injections of cortivazol 3.75 mg, 48–72 hours apart, as add-on to oral verapamil (episodic) or add-on prophylaxis (chronic). Primary outcome assessed over the 72 hours 2–4 days after the third injection.

Outcome	Cortivazol	Placebo	Effect
Mean ≤2 attacks/day after injections	20/21	12/22	OR 14.5 (95% CI 1.8–116.9), p=0.012
Mean attacks in first 15 days	10.6 (95% CI 1.4–19.9)	30.3 (95% CI 21.4–39.3)	difference 19.7 (95% CI 6.8–32.6), p=0.004
Serious adverse events	none	none	—
Other adverse events	18/21	14/22	32/43 (74%) overall, p=0.162

Most common adverse events: injection-site neck pain and non-cluster headache.

Authors’ conclusion: “Suboccipital cortivazol injections can relieve cluster headaches rapidly in patients having frequent daily attacks, irrespective of type (chronic or episodic).”

That last clause is the important one. Unlike oxygen, unlike zolmitriptan, unlike oral steroids, the Leroux trial explicitly found the benefit did not depend on episodic vs chronic status. A reduction from 30.3 to 10.6 attacks over 15 days is a two-thirds reduction in attack burden in a population averaging two attacks a day at baseline.

Contextual note: all episodic patients started verapamil at trial entry; mean verapamil dose was 420 mg (active) vs 564 mg (placebo), and additional prophylactics were used by 5% (active) vs 27% (placebo) — the placebo group needed more of everything else.

Large prospective observational study — Germany

peer-reviewed Gaul C, Roguski J, Dresler T, Abbas H, Totzeck A, Görlinger K, Diener H-C, Weber R. “Efficacy and safety of a single occipital nerve blockade in episodic and chronic cluster headache: A prospective observational study.” Cephalalgia. 2017;37(9) (online 16 June 2016), DOI 10.1177/0333102416654886. German source. n=101 (61 episodic, 40 chronic). Single unilateral injection of triamcinolone 10 mg + bupivacaine 0.5%.

Outcome	Result
Complete or partial response	83.2%
Complete response (pain-free ≥1 day), episodic	67.2%
Complete response, chronic	50%
Overall duration of effect	mean 27.3 ± 26.2 days; median 14 days
Pain-free duration, episodic	33.6 ± 26.3 days
Pain-free duration, chronic	14.3 ± 21.3 days (p<0.001 vs episodic)
Subjective benefit, episodic vs chronic	70% vs 43% (p<0.001)
Side effects	10.9% — tension-type headache 4.6%, injection-site pain 1.8%, retro-orbital pain 1.8%, nausea 0.9%, pressure 0.9%
Serious adverse events	none

The most mechanistically interesting finding in this section: numbness occurred in 27.7% (restricted to the GON/LON territory) and 41.6% (more extended distribution), but there was no association between numbness or local anaesthesia and treatment response (p=0.88).

That result argues strongly that the therapeutic agent is the steroid, not the nerve block. If blocking conduction in the greater occipital nerve were the mechanism, patients with successful anaesthesia should do better. They don’t. This aligns with Leroux & Ducros’ conclusion that anaesthetic-only injections are not supported peer-reviewed (Leroux & Ducros 2013). The implication is that “GON block” is a misleading name — it is really a depot steroid injection in the suboccipital region, and accurate nerve localisation may matter less than getting an adequate dose of long-acting steroid into the right tissue plane.

A conference-abstract version of this work appears at PMC3620288, DOI 10.1186/1129-2377-14-S1-P60.

Long-acting steroid, no anaesthetic — Italy

peer-reviewed Merli E, Asioli GM, Favoni V, Zenesini C, Mascarella D, Sartori A, Cortelli P, Cevoli S, Pierangeli G. “Great occipital nerve long-acting steroid injections in cluster headache therapy: an observational prospective study.” J Neurol. 2022;269(4):2193–2199 (PMC8940833, DOI 10.1007/s00415-021-10884-0, PMID 34820736). IRCCS Bologna. Italian source. 60 enrolled, 58 analysed (46 episodic, 12 chronic).

Technique, described precisely: methylprednisolone 60 mg in 2 mL saline, NO local anaesthetic, 25-gauge needle, three injections on alternate days, ipsilateral. Injection point: one-third laterally along the line from inion to mastoid, 2 cm below the occipital nuchal ridge; aspirate before injecting.

Outcome	All	Episodic	Chronic
Complete responders	26/58 (44.8%)	22/46 (47.8%)	4/12 (33.3%)
Partial responders (≥50% reduction)	9/58 (15.5%)	10.8%	33.3%
Combined	60.3%	—	—

Follow-up findings:

• 50% of episodic complete responders were still pain-free at one year; mean latency to recurrence 242 ± 130.49 days (~8 months).
• All chronic patients relapsed; median pain-free duration 94 ± 7.85 days (~3 months) — which the authors note is “substantially longer than previously reported (13–65 days).”

For a chronic daily patient, this is the most directly relevant number in the section: roughly three months of meaningful benefit per treatment course, in the third of chronic patients who respond completely, using a course of three injections without local anaesthetic. Three months is long enough to make this a repeatable strategy rather than a one-off.

Historical case series

peer-reviewed/historical From the Leroux & Ducros review table (Leroux & Ducros 2013):

• Anthony 1985, n=20: depot methylprednisolone 120 mg + lignocaine 2%, 2–5 mL → more than 20 headache-free days in 14/20; repeat injections benefited 9/15.
• Anthony 2000, n=20 chronic CH: 120 mg + lignocaine 1%, 3–4 mL → good response reported.

Dosing and technique guidance

peer-reviewed From Leroux & Ducros (source):

• Recommended total: 80–160 mg methylprednisolone equivalent.
• Injection site: medial third of the line between inion and mastoid, 2 cm under the ridge.
• Feel the occipital bone to avoid going too deep near the foramen magnum.
• Retract the plunger to rule out vascular puncture.
• Anaesthetic-only injections are not supported.
• Repeating three times per site per year is probably safe — but note this is inferred from the intra-articular steroid literature, not from trial data in CH. peer-reviewed label, but the safety claim itself is extrapolation and should be treated as such.

The dose-sparing argument, which is the clinical case for injection over tablets: the Leroux three-injection cortivazol regimen equals a prednisone-equivalent of 187.5 mg; the Ambrosini betamethasone regimen equals 135 mg. A standard oral bridging course is ~500 mg or more of prednisone equivalent (Leroux & Ducros 2013). You get equal or better effect for roughly a quarter to a third of the systemic steroid exposure. For anyone facing repeated bridging over years — i.e. anyone with chronic CH — that ratio is the whole argument.

Where the guidelines land

peer-reviewed CNS Drugs 2020 names GON injection the first-choice transitional treatment (Brandt et al. 2020, PMC7018790).

peer-reviewed Australian Prescriber lists GON block with local anaesthetic plus depot methylprednisolone as an alternative bridging strategy which “can reduce attacks for an average of four weeks” and “avoids the adverse effects of a course of oral steroids” (Aust Prescr 2022).

One unresolved technical question, stated as such: Australian and German practice adds local anaesthetic; the Bologna group deliberately omits it; the German Gaul study found no relationship between achieved anaesthesia and outcome. Whether local anaesthetic adds anything beyond immediate comfort is unknown. It is not a reason to refuse an injection either way.

2.4 Bridge therapy — practical summary

Option	Best evidence	Typical effect	Duration	Main drawback
Oral prednisone/prednisolone	1 RCT, episodic only (Obermann 2021)	−2.4 attacks in week 1	~2–3 weeks, then recurrence on weaning	Systemic steroid load; recurrence; dependency risk
Suboccipital / GON steroid injection	2 RCTs (Ambrosini 2005; Leroux 2011) + 2 large prospective cohorts	85% attack-free week 1; 30.3→10.6 attacks/15 days; 83% any response	mean 27 days (Gaul); ~94 days in chronic responders (Merli)	Needs a clinician who does the procedure; ~⅓ complete-response rate in chronic

If I were compressing this section to one sentence for a chronic daily patient: ask specifically about a course of suboccipital/GON steroid injections rather than accepting an oral prednisone taper by default, and cite Leroux 2011 — it is the trial that enrolled exactly people with more than two attacks a day.