A systematic review of the efficacy of ketamine for craniofacial pain

ABSTRACT Background Craniofacial pain (CFP) poses a burden on patients and health care systems. It is hypothesized that ketamine, an N-methyl-d-aspartate (NMDA) receptor antagonist, can reverse central sensitization associated with causation and propagation of CFP. This systematic review aims to assess the role of ketamine for CFP. Methods Databases were searched for studies published up to September 26, 2022, investigating the efficacy of ketamine for adults with CFP. Primary outcome was the change in pain intensity at 60 min postintervention. Two reviewers screened and extracted data. Registration with PROSPERO was performed (CRD42020178649). Results Twenty papers (six randomized controlled trials [RCTs], 14 observational studies) including 670 patients were identified. Substantial heterogeneity in terms of study design, population, dose, route of administration, treatment duration, and follow-up was noted. Bolus dose ranged from 0.2–0.3 mg/kg (intravenous) to 0.4 mg/kg (intramuscular) to 0.25–0.75 mg/kg (intranasal). Ketamine infusions (0.1–1 mg/kg/h) were given over various durations. Follow-up was short in RCTs (from 60 min to 72 h) but longer in observational studies (up to 18 months). Ketamine by bolus treatment failed to reduce migraine intensity but had an effect by reducing intensity of aura, cluster headache (CH), and trigeminal neuralgia. Prolonged ketamine infusions showed sustainable reduction of migraine intensity and frequency of CH attacks, but the quality of the evidence is low. Conclusion Current evidence remains conflicting on the efficacy of ketamine for CFP owing to low quality and heterogeneity across studies. Ketamine infusions are suggested to provide sustained improvement, possibly because of prolonged duration and higher dosage of administration. RCTs should focus on the dose–response relationship of prolonged ketamine infusions on CFP.


Description of the Condition
Craniofacial pain (CFP) is common, with a lifelong incidence of 96%, and poses a significant burden on patients' quality of life and the health care system. [1][2][3][4] Despite the availability of a variety of treatment modalities, many patients experience unsatisfactory pain relief and/or adverse effects from existing pharmacological interventions or cannot afford expensive treatments. 3,5 It is hypothesized that CFP syndromes share a mechanism of central sensitization as a cause for pain. Central sensitization is characterized by an increase in neuronal excitability secondary to repetitive stimulation of the nociceptive C-fibers in the trigeminocervical complex and the brain, which is mediated by the activation of N-methyl-D-aspartate (NMDA) receptors. [6][7][8] Activation of the NMDA receptors plays a major role in ongoing pain, opioid-induced hyperalgesia, and mood dysregulation, and it is considered the principal receptor involved in the phenomena of central sensitization and "wind-up," resulting in hyperalgesia, allodynia, and spontaneous pain. 9,10 This central sensitization can be reversed by blockade of these receptors by noncompetitive NMDA antagonists such as ketamine. 6,7,10,11 Ketamine is a chemical derivative of phencyclidine with analgesic, dissociative, and psychomimetic properties. 12 Its primary mechanism of action is as a noncompetitive antagonist of the NMDA receptors residing in the central nervous system. 10,13 Ketamine is a versatile drug that can be administered via many routes, including intravenous (IV) and intramuscular (IM) but also oral, intranasal, inhalation, topical, and rectal, rendering it an easy agent for out-of-hospital care. It can be given as a single bolus or infusion or a combination of both. Although its potency is comparable to that of opioids, ketamine has a much better safety profile and is less likely to lead to development of tolerance. Possible adverse effects of ketamine include hypertension and tachycardia, hallucinations, and hepatic toxicity with chronic exposure. 14 Ketamine has successfully been used in the treatment of complex chronic pain states such as complex regional pain syndrome and neuropathic pain, 11 but its therapeutic role in CFP has not been completely established. Therefore, the objective of this systematic review was to evaluate the efficacy of ketamine for the treatment of CFP and examine its effects on painassociated domains.

Registration
This systematic review was conducted according to the recommendations of the Cochrane Collaboration and was reported as per the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. The protocol of this systematic review was registered with PROSPERO (ID CRD42020178649).

Data Sources and Search Strategy
We conducted a comprehensive search of the literature from inception to February 1, 2020. Updated searches were conducted over the same databases and clinical trial registries on November 3, 2020, and September 26, 2022, with the assistance of a medical information specialist (M.E.).
The following databases were searched: Embase (1947-), MEDLINE (1946-), MEDLINE ePubs, In-Process and Other Non-Indexed Citations, Cochrane Database of Systematic Reviews (2005-), Cochrane Central Register of Controlled Trials (1991-), and PubMed-NOT-MEDLINE. We also searched Web of Science Core Collection (1900-; Clarivate Analytics) and Scopus (1960-;Elsevier). Clinical trial registries, ClinicalTrials.Gov, and the World Health Organization International Clinical Trials Registry Platform were searched to identify trials.
We restricted our search to human subjects with moderate-to-severe pain. For Embase, MEDLINE, Cochrane CENTRAL, and Scopus, both controlled vocabulary terms (Embase-Emtree; MEDLINE-MeSH) and text word searching were conducted for each of the following segments: ("ketamine" or related synonyms) AND ("headache" or "migraine" or "cluster headache" or "facial pain" or "short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT)" or "short-lasting unilateral neuralgiform headache attacks with cranial autonomic symptoms (SUNA)" or "trigeminal neuralgia" or "trigeminal autonomic cephalalgia (TAC)" or "occipital neuralgia" or "cephalalgia" or other related terms). The most recent Ovid Medline search strategy and the MEDLINE search strategy are provided in Appendix 1.

Inclusion/Exclusion Criteria
The studies were screened for eligibility based on title, abstract, and subsequently full manuscript. Studies were included if they met the selection criteria as listed below.

Population
This review included studies of human subjects ≥18 years of age with a CFP (primary headache or neuropathic facial pain) diagnosis that fit the criteria of the International Classification of Headache Disorders third edition. 15 Studies that investigated effects of ketamine on a mixed population of patients with pain for which data on CFP could not be extracted separately were excluded. No restrictions were put in terms of chronicity, frequency, or duration of the attack. An initial restriction of including participants with only moderate-to-severe pain was waived at the start of the title screening process because of fear of selection bias.

Intervention
The intervention was defined as bolus/infusion administration of ketamine of any dose or administration type (IV, IM, subcutaneous, intranasal, epidural, sublingual, rectal, oral). Studies in a perioperative setting and/or with mixed combinations of ketamine with another drug were excluded. There was no limit on the duration of treatment or number of treatments.

Comparator
Comparators included no treatment, placebo treatment, or conventional medical management, which could include pharmacological, physical, psychological, and/ or interventional therapies.

Outcome
The primary outcome was the change in intensity of pain assessed on a numeric rating scale (NRS)/visual analog scale (VAS) at 60 min after the intervention. Secondary outcomes included (1) positive response (defined as a reduction in pain score by ≥30% from baseline at 60 min after the intervention) and effect of ketamine infusion on (2) pain intensity (NRS/VAS) at any time after the intervention up to 6 months posttreatment, (3) adverse effects, (4) functional outcome, (5) quality of life, (6) mood, and (7) patient satisfaction. The threshold of >30% pain relief has been demonstrated to constitute clinically meaningful improvement. 16 The initial restriction of 6 months' follow-up time was extended to 18 months at the time of data collection.

Study Selection Process
All citations were independently screened on title and abstract for eligibility by two reviewers (Y. H. and P.K.) as per the inclusion criteria. Covidence 17 was used as a systematic review management tool. Papers of interest were then full-text screened. Of the selected papers, data were independently extracted by two reviewers (T.M. and M.P.). Any disagreement was resolved through discussion with the senior author (P.K.).

Data Extraction
The reference data, populations, and outcomes were extracted from the articles into prespecified tables on a standardized data collection form in Word that was pilot tested before use. Extracted data for each study included general characteristics (publication year, design, number of arms), patient characteristics (number, demographics, and sample size), clinical information (diagnosis, duration, pain intensity), details of intervention and comparator (dose and administration regimen), data on primary and secondary outcomes of interest, follow-up time points, and adverse effects.

Assessment of Quality as Risk of Bias
Two review authors (Y.H. and M.P.) independently assessed the risk of bias for randomized controlled trials using the Cochrane Risk of Bias tool 2.0. 18 Any disagreement was resolved through discussion with the senior author (P.K.). The Robins-I was used to assess the risk of bias in observational studies. 19 For case reports/case series, the Quality Appraisal Tool for Case Series was used. 20

Data Synthesis and Analysis
We narratively synthesized the characteristics of all studies that met inclusion criteria. Study characteristics and treatment details were summarized. For continuous data, means (or medians) and standard deviations (or interquartile ranges or ranges) were extracted. No meta-analysis was performed because of the heterogeneity of data, low quality, and small sample sizes of included studies.

Search Results
A total of 2080 unique articles were retrieved from all searches, of which 1956 were excluded at the screening stage. During the initial search, 72 full texts were assessed for eligibility, of which 14 papers were deemed to meet all eligibility criteria for inclusion in this review ( Figure 1). During the following searches, 9 additional full texts were selected and assessed, of which 6 papers met all eligibility criteria for inclusion in this review ( Figure 2). The 20 papers included in our review, reporting on 670 patients in total, included the following studies: Six randomized controlled trials (RCTs) 21-26 and 14 observational studies, of which 5 were prospective cohort studies, 27-31 3 were retrospective cohort studies, 32-34 3 were case series, [35][36][37] and 3 were case reports, were included [38][39][40] (Tables 1 and 2).

Risk of Bias for Included Studies
The results for assessment of the risk of bias for the included RCTs are listed in Appendix 3. The overall risk of bias was deemed to have "some concerns" for two of the RCTs 24,25 and "high risk of bias" for the remaining four RCTs. [21][22][23]26 The risk of bias assessment of the included nonrandomized trials showed one study of low risk of bias, 30 one study of moderate risk of bias, 33 two studies deemed to have a serious risk for bias, 27,32,34 and two studies that demonstrated a critical risk of bias 28,29,31 (Appendix 4). The quality of the two case series was moderate 35,36 and one was of high quality 37 (Appendix 5).

Details of the Ketamine Treatment.
Route of Administration and Dosing. Treatment with a single [23][24][25][26]31 or double 22 or five 30 boluses of ketamine was investigated in seven papers. Five papers (three RCTs, one cohort study, one case series; n = 245) looked into the effect of and intranasal bolus of ketamine compared to midazolam, 23 diphenhydramine, 22 ketorolac, 26 or no comparator. 30,35 Bolus dose ranged from 0.25 to 0.75 mg/kg. Ketamine was administered as an IV bolus in two RCTs. 24,25 Bolus dose ranged from 0.2 to 0.3 mg/ kg and comparators were placebo 25 or prochlorperazine with diphenhydramine. 24 The effect of IV ketamine infusions was reported in ten observational studies (including 283 patients). 29,[31][32][33][34]37,40 Patient population in these studies was mixed and included patients with CM and CH. Infusions were given over dosages ranging from 0.1 to 1 mg/kg/h over various durations (ranging from a single 1-h infusion up to infusions for up to 6 h on nine consecutive days). Two studies had a comparator (lidocaine). 31,34 Duration of Follow-Up. The follow-up time in the RCTs ranged from 60 min 25,26 to 72 h posttreatment, 22,24 with one additional trial specifying only that follow-up occurred for six consecutive migraine attacks. 23 The follow-up time for the observational studies was longer, ranging from immediately posttreatment, 30,34-36 to 1 week, 37     studies (all observational studies) investigated ketamine infusions. [31][32][33][34][35][36] Of the four RCTs investigating the analgesic efficacy of ketamine bolus treatment, only one RCT demonstrated superiority of ketamine compared to the comparator (ketorolac), providing significant pain relief up to 2 h posttreatment. 26 The three other RCTs failed to demonstrate superiority of ketamine compared to placebo, metoclopramide, or midazolam 22,23,25 (Table 2).
In stark contrast, all of the five observational ketamine infusion studies (n = 233) reported significant pain improvement at the end of the infusion period. [31][32][33][34]36 Infusions of solely ketamine were given for 8 h 36 to 5 days in a row. 32,33 In two studies, the effect of multiday ketamine infusion was compared to multiday lidocaine infusion. 31,34 In the studies with lidocaine as a comparator, both groups were associated with significant pain reduction at the end of the infusion as compared to baseline, although the difference between groups was significant in favor of lidocaine in one study 34 and in favor of ketamine in the other study. 31 Only three out of five papers reported long-term outcomes. Two studies noted a positive effect up to 1 32 and 3 months. 33 In the third paper, patients had returned back to their baseline pain at 1-month follow-up. 31 Impact on Aura Attack during Migraine. Only one RCT investigated the effect of ketamine and midazolam on length and severity of aura compared to placebo. 23 The severity of the aura was significantly improved by ketamine compared to placebo. However, both agents were equally effective in reducing the median duration of the attack compared to placebo. In a case series of 11 patients with familial hemiplegic migraine treated with 25 mg intranasal ketamine, 5 patients experienced an improvement of aura duration and severity for all 14 attacks treated. Three patients experienced a return of aura symptoms after initial improvement. 35 Impact on Pain Relief and Attack Frequency of TAC Headaches. All seven studies including patients with TAC (n = 76) reported a resolution or reduction in the attack frequency after ketamine treatment. An immediate but short-lasting effect was noted after bolus treatment 30 and an immediate to sustained effect was noted after ketamine infusion. 29,34,[37][38][39][40] Patients were reported to have more than a 50% reduction in attack frequency for 6 weeks 38 to complete resolution of attacks for 6 weeks 34,37,38 up to 3 months. 39,40 Effect on Pain-Associated Domains (Functional Outcome, Sleep, Quality of Life). Only one RCT reported on the effect of ketamine on pain-associated domains. No significant difference in functional disability scores (rated on a 4-point scale), measured 30 min posttreatment, between the IV ketamine bolus and placebo groups was identified. 25 Patient Preference/Satisfaction. Only three papers reported on patient satisfaction and/or preference for ketamine versus the comparator using an NRS. 22,24,25 In only one study was a significant difference identified, in favor of prochlorperazine. 24 Impact of Ketamine on Intake of Pain Medication. A few studies commented on change in use of rescue medication during or following ketamine treatment. Although not statistically significant, two RCTs demonstrated that a smaller proportion of patients in the ketamine group needed rescue medication during the infusion compared to the comparator. 22,25 Two case series mentioned a reduction in dose and/or discontinuation of analgesics following ketamine infusion, but further details were not provided. 38,39 Adverse Effects and Complications of Ketamine Infusion. Approximately 70% of patients across all studies experienced side effects related to ketamine infusions; however, most were mild and resolved with decreased rate or ending the ketamine infusion. Side effects included dizziness, sedation, and blurred vision'     One patient's ketamine infusion was stopped because the patient experienced suicidal thoughts 32 and one patient preferred to stop ketamine because of the side effects. 33 One additional patient developed an asymptomatic elevation in liver enzymes. 32 None of the patients needed to be hospitalized because of intolerable side effects such as feelings of insobriety, confusion, nausea and vomiting, hallucinations, and tachycardia.

Facial Pain Demographics and Pain Profiles of Participants in the Included Studies.
Of three studies including 54 patients (85% females, 29-89 years), one crossover RCT (n = 30) 21 and one small observational study (n = 7) 28 examined the use of ketamine for patients with trigeminal neuropathy. Another small observational study (n = 17) included patients with neuropathic orofacial pain (not specified). 27 Baseline pain intensity was moderate to severe 21,28 or not provided. 27 The duration of pain ranged from 6 months to 28 years.

Details and Outcome of Ketamine Treatment.
In the three studies, all patients received a single bolus of ketamine 0.4 mg/kg. One study compared ketamine to IM pethidine 1 mg/kg. 21 In two studies, the bolus treatment was followed by 4 days of oral ketamine. 21,27 Follow-up duration was short in these studies and ranged from 60 min after ketamine IM bolus 21,27,28 to 3 days after oral treatment. 21,27 Ketamine demonstrated a significant immediate improvement of facial pain in two studies for the duration of follow-up (3 days). 21,27 It was noted that those responding to IM ketamine also reported pain relief with oral ketamine. 21,27 The adverse events profile was similar as reported in the primary headache studies.

Discussion
Despite recent evolutions in pharmacological and interventional management of patients with CFP, many patients continue to experience a significant reduction in quality of life, thereby imposing a significant burden on the health care system. We therefore conducted this systematic review to evaluate the evidence on the role of ketamine for patients with headache and/or facial pain.

Impact on Headache and Facial Pain
Although we evaluated 5 RCTs and 12 observational studies, evidence remains conflicting on the efficacy of ketamine for treatment of primary headaches. This is likely due to the limited study quality, moderate to high risk of study bias, confounding and substantial heterogeneity across studies in terms of study design, patient populations, details of ketamine treatment, and (lack of) follow-up. Only 2 of the 5 RCTs demonstrated a significant immediate effect of ketamine on the intensity of pain and aura of migraine relief in the emergency room, 23,26 and the remaining RCTs failed to demonstrate a significant effect of ketamine on headache  intensity compared to various comparators. On the other hand, all observational studies demonstrated a significant reduction in migraine pain intensity and duration, as well as decreased frequency and intensity of CH attacks immediately postintervention that lasted for up to 3 months postinfusion. The difference in these findings may be explained as follows. First, it may be attributed to the shorter administration times and/or lower dose of ketamine bolus treatment in the RCTs, because nearly all observational studies explored single-/multiple-day ketamine infusions. The typical doses for chronic pain IV ketamine treatment are 0.2 to 0.75 mg/kg (bolus), 0.5 to 2 mg/kg/h (infusion), 0.1 to 0.5 mg/kg for IM administration, and 0.2 to 1 mg/kg for intranasal administration, which is higher than what was administrated in the RCTs. 10 A high dosage and extended administration (infusions) of ketamine have been associated with better pain relief in patients with chronic pain. 41 Similar conclusions were made in the consensus guidelines on the use of IV ketamine infusions for chronic pain. 10 Future research in patients with primary headache should therefore focus on investigating the effect of highdose and repeated administration of ketamine in a randomized controlled setting. Second, most RCTs had a mixed patient population, including patients with acute and chronic headache of migraine, cluster, and MOH types, which increased study heterogeneity, potentially affecting the outcome. Lastly, the 5 papers that did show a positive outcome of ketamine bolus treatment investigated patients solely with TAC, and bolus treatment failed in studies on patients with chronic migraine. This difference in outcome could be explained by a difference in pathophysiology, where TAC is driven by changes in the sphenopalatine ganglion, which potentially is more susceptible for the ketamine effect. The evidence on ketamine for neuropathic facial pain is scarce but appears promising. Intramuscular ketamine was associated with significant improvement in two out of three studies on patients with trigeminal neuropathy, 21 and a tendency for continued pain relief with oral ketamine after initial IV/IM treatment was also noted. 21,27 This treatment option of oral ketamine should be further explored because this could be a valuable therapy for patients in remote areas with limited access to health care facilities.

Effect on Pain-Associated Domains
Although the International Classification of Headache Disorders third edition diagnostic criteria for CFP are widely accepted and recommended, 42 a majority of the papers did not mention use of this validated instrument for diagnosis. This is a weakness that needs to be addressed to avoid misclassification bias and improve generalizability of the study results. Though a handful of studies examined patient satisfaction/preference and the need for rescue medications, most papers in this review did not mention use of validated tools to evaluate the effect of ketamine on pain-associated domains (sleep, mood, quality of life) in addition to its analgesic efficacy. This is surprising because the importance of evaluating pain-related domains in clinical and research settings is established, and validated tools to evaluate these domains are available. 43 Ketamine has been demonstrated to have an effect on sleep and mood and could therefore potentially improve patients' quality of life even if pain intensity is unchanged. 44,45 Safety It is widely accepted that ketamine is associated with adverse psychomimetic, cardiovascular, and gastrointestinal effects resulting from its activity on a variety of substrate receptors including NMDA, acetylcholine, opioid, monoamine, and histamine. 10 The incidence of reported ketamine-induced side effects in the included papers was high regardless of type of administration or dose and only included psychomimetic effects. Most side effects were mild and resolved after decreasing the dose or ending treatment. Although central nervous system effects appear to be dose dependent when ketamine is used in anesthetic doses, the evidence is not as clear for subanesthetic regimens, beyond a yet-to-be-determined threshold. 10 In our review, no clear difference was noted in the incidence of side effects when comparing high to low doses. Most patients experienced hallucinations, which resolved after cessation of treatment, and one patient experienced suicidal thoughts. Based on the America Psychiatric Association 44 guidelines on the use of ketamine, a history of psychosis is a contraindication for administration of subanesthetic IV ketamine.

Strengths and Limitations
We found one other systematic review on the efficacy of ketamine for headaches, with evidence found in RCTs that were primarily focused on bolus administration of ketamine for acute headache pain relief. 46 The authors concluded that the benefit of ketamine for headache treatment is unclear but that long-term follow-up and different ketamine dosages in patients with chronic pain should be explored. Therefore, our systematic review attempted to take a closer look at the larger amount of evidence in observational trials that focused on ketamine infusions. Our review demonstrated that prolonged ketamine infusions can indeed significantly and sustainably reduce migraine intensity and frequency of CH attacks. This provides promising evidence, and this temporal and dose-dependent relationship should be further explored in new high-quality trials. To our knowledge, no attempt has been made to systematically gather evidence regarding the effects of ketamine on facial pain. This review has several limitations. Most of the studies included in this review reported observational data, with inherent high risk of confounding and bias. Further, the heterogeneity across papers was substantial, with large differences in dosing, duration and route of administration, and comparators, which made comparison challenging and precluded a meta-analysis. Lastly, this review did not assess the long-term benefits and harms of ketamine for the treatment of CFP owing to lack of data in the literature.
A high-quality placebo-controlled RCT investigating the effect and safety of high-dose and/or prolonged ketamine infusion treatment on refractory headache pain and pain-associated domains would be the logical next step to investigate ketamine's potential in the treatment of CFP. Recent systematic reviews and metaanalyses of this evidence show that infusions of ketamine 1 mg/kg/h are associated with greater and sustained pain relief compared to one-off bolus and lowdose infusion regimens. 10 However, adverse effects of ketamine are common and could limit dose escalation, and the potential long-term risks of repeated administration of high doses of ketamine need to be further investigated.

Conclusions
Current evidence on the efficacy of ketamine for treatment of CFP remains conflicting, precluding the ability to make any recommendations. This is likely due to the limited study quality, moderate to high risk of study bias, and substantial heterogeneity across studies in terms of study design, patient populations, details of ketamine treatment, and (lack of) follow-up. Ketamine bolus treatment showed significant reduction of migraine aura severity, TAC intensity and frequency, and trigeminal neuropathic pain but failed to reduce migraine intensity or duration. The included observational trials suggest that ketamine infusion treatment has an immediate and sustained benefit on headache pain intensity, possibly because of the prolonged duration of treatment and higher dose of administration. Most papers failed to evaluate the effect of ketamine on pain-associated domains. RCTs are required that focus on the dose-response relationship and immediate and long-term effects of prolonged ketamine treatment on CFP and pain-associated domains.

Disclosure Statement
YH received the Early Career Investigator Pain Research Grant from the Canadian Pain Society for the KetHead study "A Multi-Center Randomized Controlled Trial of Efficacy and Safety of Intravenous Ketamine for Chronic Daily Headaches: The KetHead Study." All other authors have no conflicts of interest to report.

Authorship Statement
Study conception and design: YH and ME; drafting of the article: YH and MP; research into, design of, and execution of database search strategies: ME; management and curation of database results, write-up of search methodology; ME; data acquisition, analysis, and interpretation: YH, TM, MP, PK; review and editing of manuscript: all authors.