In an open‑label pilot in chronic cluster headache, intranasal ketamine did not meet the predefined 50% pain reduction at 15 minutes but produced a significant mean 59% reduction at 30 minutes (69% of evaluable patients ≤4/10) with no serious adverse events. These results indicate intranasal ketamine may be an effective acute treatment at 30 minutes and warrant larger controlled trials, with caution regarding ketamine’s abuse potential.
Objective
To investigate the safety and efficacy of intranasal ketamine for the treatment of a single cluster headache (CH) attack.
Background
Acute treatment options for patients with CH who have an insufficient response to oxygen and triptans are limited. Intranasal ketamine has anecdotally been successful in treating a CH attack.
Methods
We conducted an open‐label pilot study enrolling 23 patients with chronic CH ( International Classification of Headache Disorders , 3rd edition), and of these, 20 patients treated a single CH attack with intranasal ketamine. Under in‐hospital observation, patients received 15 mg of intranasal ketamine every 6 min a maximum of five times. The primary endpoint was a 50% reduction in pain intensity within 15 min after initiating treatment.
Results
The primary endpoint was not met; 15 min after the first ketamine administration, the mean reduction in pain intensity was 1.1 (95% confidence interval [CI]: −0.6 to 2.7, p = 0.188) on the numeric rating scale (NRS), equivalent to a 15% reduction in pain intensity. However, 30 min after the first application, the pain intensity was reduced by 59% on an 11‐point NRS (mean difference: 4.3, 95% CI: 2.4–6.2, p < 0.001, N = 16) and 11 out of 16 (69%) scored 4 or below on the NRS. Four patients received rescue medication 15 min after the first ketamine application and were therefore excluded from the analysis at 30 min. Half of the patients preferred ketamine to oxygen and/or sumatriptan injection. No serious adverse events were identified during the trial.
Conclusion
Intranasal ketamine may be an effective acute treatment for CH at 30 min but should be tested in a larger controlled design. Patients and physicians should be conscious of the abuse potential of ketamine.
Cluster headache (CH) attacks are typically treated acutely with triptans (subcutaneous or intranasal) and high-flow oxygen, but an estimated 20%–30% of patients gain insufficient relief from these first-line options. Previous trials concentrated mainly on episodic CH, and patients with chronic cluster headache (cCH) often respond less well to triptans and oxygen. A few alternative acute modalities exist (for example intranasal lidocaine, octreotide, or neurostimulation) but each has practical or safety limitations and some are unavailable, leaving a treatment gap for severely affected, refractory patients. Petersen and colleagues conducted a proof-of-concept, open-label pilot study to evaluate whether intranasal racemic ketamine could be an effective and safe acute treatment for single CH attacks in people with cCH. The stated aim was to assess effect and tolerability under supervised, in-hospital conditions with an emphasis on short-term pain reduction after administration.
This was an open-label inpatient trial enrolling patients with chronic cluster headache according to the International Classification of Headache Disorders, 3rd edition. Eligible participants were aged 18–60 years, weighed more than 49 kg, and had body mass index under 31. Exclusion criteria included significant cardiac, hepatic, neurological or psychiatric disease (including severe depression and psychosis), hypersensitivity to ketamine, upper respiratory tract allergy risk, pregnancy or breastfeeding, nasal abnormality, recent illegal drug use or history of drug abuse, and baseline hypertension (>140/90 mm Hg). Patients could remain on stable preventive medications. Recruitment occurred at a single specialised centre between December 2019 and May 2020. The investigational product was an intranasal spray formulation of racemic R,S-ketamine (150 mg/ml as hydrochloride). Administration was 15 mg (100 µl) every 6 minutes for a minimum of three doses and up to five doses (maximum total 75 mg). Participants applied the spray to the nostril contralateral to the pain side while tilting the head back and were observed for at least 3 hours. Absolute bioavailability was reported as approximately 45%. Rescue medication was offered 15 minutes after the first ketamine dose if pain relief was inadequate. Observers recorded pain intensity and cranial autonomic and related symptoms at prespecified time points (0, 6, 12, 15, 18, 24, 30, 60, 120 and 180 minutes) and monitored blood pressure, heart rate and oxygen saturation at these same times. Follow-up contact occurred by phone 1–2 weeks after discharge. The primary efficacy endpoint was a ≥50% reduction in pain intensity on an 11-point numeric rating scale (NRS, 0–10) at 15 minutes after the first ketamine administration. Safety assessments included prespecified adverse events (nausea, vomiting, euphoria, hallucination, palpitations, double vision) and vital-sign monitoring because ketamine can provoke hypertension and tachycardia. The sample size calculation targeted 20 treated participants to detect a between-time difference greater than about 3 NRS units (power 90%, alpha 0.05, assumed SD 4) using paired two-tailed t-tests. Participants who received rescue medication were excluded from the 30-minute analyses. Statistical analyses used R; no adjustments for multiple testing were performed.
A sample of 20 participants was needed to detect a difference greater than 3 NRS units (defined as 3.1 NRS units in the powercalculation) in a paired two-tailed t-test assuming a standard deviation of 4, a significant level of 0.05, and a power of 90%. Pain intensities before the first administration of ketamine and at 15 or 30 min were compared using a paired t-test. The participants who chose to use rescue medication were excluded from the analysis at 30 min. Significance was assessed at the 0.05 level. R 3.6.1 and R studio were used for the statistical analyses. No additional tests were executed to avoid multiple testing. A spaghetti plot of pain intensity (Figure) was created by ggplot2 package. Smoothed conditional means and 95% confidence interval (CI) of pain intensity over time were visualized by stat_smooth.
Create a free account to open full-text PDFs.
Of 58 patients assessed, 23 were enrolled and 20 treated a single cluster attack in hospital; three enrolled patients did not experience an attack during admission and did not complete the trial. The treated cohort had mean age 47 years, 70% were male, and mean disease duration was 20 years. Baseline attack characteristics included typical intensities ≥7 on the NRS and a mean self-reported frequency of 3.3 attacks per day. Fourteen of 20 (70%) used concomitant preventive medication; 11 of these were taking verapamil, with a reported mean total daily dose of 540 mg (range 200–960 mg), although the extraction contains a minor notation uncertainty about that dose. All patients had prior experience with oxygen and/or triptans and were ketamine‑naïve. At 15 minutes after the first ketamine dose (n = 20) the mean pain intensity fell from 7.2 ± 1.3 to 6.1 ± 3.1, a mean decrease of 1.1 NRS units (95% CI: -0.6 to 2.7), which corresponded to approximately a 15% reduction and did not meet the primary endpoint (p = 0.188). Four of 20 participants (20%) achieved ≥50% pain reduction at 15 minutes. Four participants chose rescue medication at 15 minutes and were therefore excluded from the 30-minute analysis. At 30 minutes after the first dose (n = 16), pain intensity was substantially lower: the extraction reports a mean difference of 4.3 NRS units (95% CI: 2.4–6.2), corresponding to a 59% reduction. Eleven of 16 patients (69%) scored 4 or below on the NRS at 30 minutes. The extracted text reports the p-value for this improvement inconsistently—an earlier summary line gives p < 0.001 while another passage prints 'p > 0.001'—but the authors characterise the 30-minute reduction as statistically significant. Exactly half of the participants expressed a preference for ketamine over oxygen and/or sumatriptan injection. No serious adverse events were reported in the 20 treated patients. Observers recorded expected transient side effects of ketamine, and cardiac measurements were described as stable after administration, although the authors note these readings may have been confounded by the CH attack itself. The investigators did not perform additional statistical tests to correct for multiple comparisons.
Petersen and colleagues interpret their findings as showing that intranasal racemic ketamine did not achieve the predefined primary outcome at 15 minutes but produced a clinically meaningful and larger reduction in pain intensity by 30 minutes in this small, open-label sample of patients with refractory chronic cluster headache. They caution that the absence of a placebo or active-control arm after 30 minutes is the major limitation, so ketamine cannot be recommended for routine acute therapy of CH until confirmatory controlled trials are conducted. The authors position their results relative to prior, sparse evidence on intranasal ketamine for CH and note that timing of effect may relate to pharmacokinetics: prior paediatric data showed peak plasma concentrations 8.5–20 minutes after nasal dosing, and with a stepwise dosing schedule the full effect of the initial three doses may occur around 25 minutes. They compare the 30-minute effect with nasal triptan data and argue that an effect within the first half hour is clinically relevant because most CH attacks last longer than 45 minutes. Nevertheless, intranasal ketamine is unlikely to supplant oxygen or triptans as first-line therapy because it produced little effect at 15 minutes and has a less favourable side-effect profile. Safety and longer-term concerns receive detailed attention: no serious adverse events occurred in this trial, but the investigators emphasise ketamine’s abuse potential and the uncertain long-term safety outside controlled dosing. They reference observations of ulcerative cystitis and cortical changes in people with high‑dose recreational use, and note that controlled intranasal esketamine trials for depression have not demonstrated comparable severe effects over 52 weeks; still, the risk of dependence argues for restricted dispensing and supervised dose titration. The possibility of medication‑overuse headache with repeated ketamine is raised as plausible but unresolved. Mechanistic explanations are discussed cautiously: ketamine’s NMDA receptor antagonism and downstream effects on nitric oxide synthesis, and putative hypothalamic/neuropeptide effects (orexin, melanin-concentrating hormone) are proposed as biologically plausible contributors to an anti‑CH effect. The investigators acknowledge multiple study limitations beyond the lack of blinding and control: small size, single-centre recruitment, a homogeneous sample restricted to chronic, largely refractory patients, and potential bias introduced by excluding those who received rescue medication from the 30‑minute analysis. Strengths included supervised administration, standardised observation timepoints, and focus on a clinically refractory population likely to need new options. For future work the authors recommend randomised, placebo-controlled trials that emphasise outcomes at 30 minutes, consider different dosing strategies (including forced titration or larger initial doses) and explore S‑ketamine as an alternative with potentially fewer cognitive side effects. They also suggest extending evaluation to episodic CH populations and studying optimal dosing to balance speed of onset and tolerability.