Cluster headaches are the most painful form of primary headache and the most common of a group of headaches known as trigeminal autonomic cephalalgias (TACs) (Imai, 2013). They can be acute or chronic, but are most commonly characterised by severe, short-lasting, episodic, unilateral pain that is orbital, supraorbital, temporal or any combination of these sites, and such attacks are typically accompanied with specific autonomic symptoms (Headache Classification Subcommittee of the International Headache Society (HCS), 2005).
The need for enhanced management of primary and medication overuse headaches is driven by their perceived prevalence and the overall burden on health systems. It is estimated that ‘25 million days are lost from work or school because of migraine each year’ (National Institute for Health and Care Excellence (NICE), 2013). Additionally, it is apparent that ‘headaches are one of the most common neurological problems presented to GPs and neurologists' (NICE, 2013). McCrone et al (2011) highlight that there is a potential cost as a result of headaches of £956 million due to service use, and £4.8 billion due to lost employment. Details of primary and secondary headaches can be found in Table 1.
| Primary headache | Secondary headache |
|---|---|
| Migraine | Systemic infection |
| Tension headaches | Head injury |
| Cluster headaches | Subarachnoid haemorrhage |
| Idiopathic stabbing | Vascular disorders |
| Exertional | Space occupying mass |
Although NICE (2012) highlight the burden of headaches on the healthcare system and commonality of patient presentation at GP surgeries, it should be considered that the actual prevalence of cluster headache is relatively low. Through meta-analysis, Fischera et al (2008) identify that cluster headaches have relatively stable lifetime prevalence, and suggest that about one in 1 000 people suffer from cluster headaches. This equates to 250 000 patients with cluster headache in the UK (Bhola, 2008). Further to this, May (2012; 2013) recognises that although TACs are rare compared to some headaches, a need for enhanced capacity for recognition and treatment is justified by the fact that TACs are relatively easy to diagnose and response to treatment in most patients is considered ‘excellent’.
With the effect on the Western economy and burden on health systems in mind, there are currently a number of strategic drivers for enhancing general management of headaches. Specifically, NICE (2013) introduced a new standard for management of headaches, with the explicit goals of reducing referrals without signs of a secondary headache, enhancing accurate categorisation and diagnosis, and raising professional and public awareness around primary and medication overuse headaches. Details of the different types, causes and treatment of headaches can be found in Appendix 1.
Such a contemporary drive is that of the ‘right care’ agenda, originating with the Bradley Report (Department of Health (DH), 2005) and focusing on ‘taking healthcare to the patient’ (Association of Ambulance Chief Executives, 2011). Subsequent to this, clinical quality standards for ambulance services around treating patients on scene, and referral to destinations other than the emergency department have been implemented (DH, 2011). More recently, the Keogh Review (NHS England, 2013) re-emphasised the systemic burden associated with avoidable emergency department attendance and hospital admissions and a need to shift more care closer to home. It is estimated that 50% of ambulance calls could be managed at home, and that over 1 million admissions were considered avoidable in 2012/13 (NHS England, 2013), a proportion of which would have been related to cluster headaches.
Ahmed (2012b) states that to both exclude serious headache disorders and avoid unnecessary and expensive hospital admissions and investigations, it is important to be able to identify primary headache disorders. Given the emphasis on admission avoidance, those working within the community, such as paramedics, play a key role in early identification, safe diagnosis and differential diagnosis, and if needed onward referral for those suffering from primary cluster headaches. From this brief overview of strategic context it is clear that enhanced capacity to diagnose and manage cluster headache in the community is in line with such strategic drivers at every level, and as such a review of contemporary literature of cluster headaches has been completed. Particular focus was placed upon management, referral and treatment of acute cluster headaches in the community by ambulance and urgent care services.
Diagnosis, referral and treatment of acute cluster headaches: the literature
A review of contemporary literature on cluster headaches has been conducted to inform this article, specifically keeping in mind that any development of practice needs be carefully balanced with clinical risk and with the necessary governance structures in place.
The collated literature is used to discuss the appropriateness of non-specialist clinicians working in out-of-hospital, and in primary, urgent and emergency care (such as paramedics), safely diagnosing and managing cluster headaches in the community, while also considering the balance of possible risks and fiscal benefits in doing so.
Information has been drawn from key NHS clinical guidance and strategic white papers via a search on NHS Evidence. Literature searches were carried out on CINAHL and MEDLINE databases using the search term ‘cluster headache’. A particular focus has been set on any papers written after the release of NICE Clinical Guidance 150 (NICE, 2012). Literature over 10 years old has been excluded and only papers written in the English language have been considered. Where relevant and applicable to the UK healthcare system, international literature has been included. In addition to this, further searching was conducted by reviewing reference lists of relevant key papers, and checking search strategies in relevant systematic reviews, to inform the following areas of practice.
Diagnosis
The diagnosis of cluster headaches is a clinical task, combining medical history and clinical examination (Edvardsson, 2014). In addition to undertaking specific tests to detect meningeal irritation (Kernig's and Brudzinski's signs) (Holle and Oberman, 2013), physical examination should essentially include a comprehensive neurological examination, with fundoscopy (Bal and Hollingworth, 2005)—something which is now being included in undergraduate paramedic education.
Diagnostic criteria are clearly set out by the International Classification of Headache Disorders (HCS, 2005). The HCS criteria (2005) were used and reworded for interpretation by non-headache specialists in NICE Clinical Guideline 150 and recommended for use in creating a ‘clearer pathway for the non-specialist’ (NICE, 2012: 76), such as paramedics. These include a requirement for at least five attacks fulfilling specific criteria of ‘severe or very severe’ unilateral orbital, supraorbital and/or temporal pain lasting 15–180 minutes if untreated.’ NICE (2012) and HCS (2015) state that this should be accompanied with at least one autonomic symptom, making adequate history taking equally as important as physical examination to such patients.
Such autonomic symptoms include ‘one-sided forehead and facial sweating’ and ‘one-sided constricted pupil and swollen eyelid’. However, Imai (2013) suggests that headaches presenting with all but the autonomic symptoms can be just as debilitating and should be considered equally for treatment. The characteristic of ‘a sense of restlessness or agitation’ (HCS, 2005) is supported by Kernick et al (2006), who highlight the fact that often cluster headache sufferers will want to move about, whereas migraine sufferers will avoid movement.
Age is not highlighted by NICE (2012), or the HCS (2005) as a diagnostic factor; however, there is some low class empirical evidence (based on a cohort of 68 only) to suggest enhanced diagnostic accuracy in the age group 40 to 50 years (Edvardsson, 2014). This indicates that a higher index of suspicion around secondary cause should be adopted in older and younger age groups. Further to this, NICE takes into account the fact that headaches do present differently in children, for example, ‘occipital headache in children, whether unilateral or bilateral, is rare and calls for diagnostic caution; many cases are attributable to structural lesions’ (NICE, 2012: 75).
Red flags
In order to carefully consider risk, it is worth discussing red flags. The first bout of cluster-type headaches should be perceived with a high degree of clinical suspicion given the variety of possible causes, and the degree of clinical risk should be considered high, making it likely that patients presenting with the first episode of cluster-type headache should be admitted, in order to seek specialist assessment.
NICE (2012) recommend a series of symptoms, such as vomiting and exercise as indications for consideration of additional investigation or referral for people presenting with a headache (NICE, 2012: 78), and paramedics should take this into consideration prior to discharging a patient presenting with such symptoms. When considering red flags associated with cancer, NICE (2012) guidance differs in subtle ways from other contemporary literature on the subject. NICE focuses on the confirmed history of malignancy, whereas other contemporary literature refers to specific signs such as persistent early morning headaches and progressive headache worsening over weeks (British Association for the Study of Headache, 2010). The reason for this is unclear; however, it is possibly due to the fact that headache is mainly associated with late stage brain tumour (Ahmed, 2012a).
HCS (2005) include the caveat that the headache should not be attributable to another disorder. Thus, rather than simply referring to a list of signs and symptoms that indicate ongoing referral, a degree of understanding around high-risk differentials is also valuable, and should be included in undergraduate paramedic education. Bal and Hollingworth (2005) and Ahmed (2012b) also highlight meningitis, sub-arachnoid haemorrhage and cerebral venous thrombosis as high-risk differentials. Thus symptoms such as fevers, rashes, sudden onset, new onset neurological deficit, cognitive dysfunction, a history of pregnancy and early morning headaches should all be considered red flags.
In headaches, NICE (2012) also includes consideration of giant cell (temporal) arteritis, which would most likely be indicated in patients over 50 years with history of visual disturbances, fever, malaise, night sweats, loss of appetite and weight loss, plus visibly thick, tender and nodular temporal arteries (Dasgupta et al, 2010), indicating palpating of the temporal arteries as part of the physical exam (Bal and Hollingworth, 2005). Narrow angle glaucoma is also a differential diagnosis that needs to be considered in all patients presenting with headaches. This includes symptoms such as red painful eyes, visual field deficits, and cloudiness of the cornea, and further justifies the need for an optical exam in patients with headache (NICE, 2012), thus validating its inclusion in undergraduate paramedic education, among other reasons.
One of the high-risk differentials and with a common presentation to the first bout of cluster headache is dissection of the carotid artery (NICE, 2012). Although presentation is variable, it can also present with unilateral headache and pain in the face. The index of suspicion is enhanced in young adults. In this demographic, carotid artery dissection accounts for accounts for 15–20% of strokes and is usually associated with a history of trauma, or connective tissue disease (el Khashab, 2012).
Onward referral
It is suggested that clinicians refer people with cluster headaches for further investigations such as computed tomography (CT) based on the risk of significant secondary pathology (Sun et al, 2013), which may account for some admissions of patients suffering recurrent episodes by paramedics, who think that such scans are likely to be carried out. Due to a paucity of empirical evidence and a degree of subjectivity around the risk, historically decisions made around primary headaches vary, with context having a significant bearing on referral pathways. Referral and management decisions out of hospital have been based on the evaluation of a number of complex factors, these include:
‘Therapeutic and economic value, clinical confidence, time constraints within the consultation, availability of imaging, practitioner's and patient's approach to risk and uncertainty, reassurance and medico-legal concerns’
It is important to balance costs and risks associated with imaging with the perceived benefits of referral, thus making the enhancement of non-specialist community practitioner headache education increasingly important. Imaging using CT, or magnetic resonance imaging (MRI) carries with it the risk associated with exposure to radiation doses, in conjunction with the enhanced anxiety associated with referral and incidental findings. Considering this, NICE state that ‘the few abnormal cases detected by the tests do not appear to be cost-effective’ (NICE, 2012: 87). Additionally, it is worth noting that when comparing imaging with neurologist referral, the cost savings associated with the latter are statistically significant. This is not, however, specific to cluster headache, but should perhaps be considered by paramedics who would admit a patient purely on the basis that they might receive a CT scan, prior to discussion with a specialist.
In answer to the question: ‘Should young people and adults with suspected primary headaches undergo brain imaging to rule out serious pathology?’ NICE (2012) make the recommendation to ‘not refer people diagnosed with tension-type headache, migraine, cluster headache or medication overuse headache for neuro-imaging solely for reassurance’ (NICE, 2012: 94). This approach is supported by Holle and Oberman's (2013) decision tree for decision-making in patients with headaches, which it could be argued should be incorporated into guidelines specific to community practitioners, such as paramedics. This provides a simplified approach, where patients presenting with typical symptoms and age at onset of a primary (idiopathic) headache and a normal neurological exam do not require further diagnostic tests or neuro-imaging. Where history and neurological exam indicate a suspected secondary headache, further diagnostic tests and neuro-imaging are required.
This approach could be applicable by non-specialists to chronic, or episodic cluster headache; however, Holle and Oberman (2013: 370) also suggest that ‘MRI should be performed once in each patient for accurate elimination of secondary cause,’ although in the absence of red flags and first presentation, this could be organised though primary care as an outpatient, and not through acute admission to accident and emergency departments. NICE also recommend that non-specialists ‘Discuss the need for neuro-imaging for people with a first bout of cluster headache with a GP with a special interest in headache or a neurologist,’ (NICE, 2012: 96), something which depending on area, service provision, and agreed pathways, could be possible for paramedics working within the community.
This need for specialist referral in the first bout of a cluster headache is based on consensus opinion and justified by the need to exclude vascular abnormalities, specifically carotid artery dissection (Snezana et al, 2011; Divjak et al, 2013). The fact that carotid artery dissection carries a substantial risk of disabling stroke and is a potentially life-threatening condition, suggests that it should be treated as time critical and transferred as an emergency for MRI and angiography scan of the head and neck, with the potential for anticoagulant therapy and stenting to reduce the risk of carotid thrombosis and embolism (Bazari et al, 2010).
Treatment
In the out-of-hospital environment it is clear that consistent, direct access to a specialist is perhaps not realistic and patients may have protracted waiting times under such referral. In 2012 in the UK, one neurologist served a population of between 117 000 and 200 000. When compared with other European countries, such as Italy, where one neurologist serves a population of 10 000 people, we can clearly see a disparity (Ahmed, 2012b). Further to this, Ahmed (2012b) highlights the fact that no more than 20 consultant neurologists consider themselves to have a specialist interest in headache disorders, there are 12 specialist headache nurses and only a handful of GPs with a specialist interest in headaches. Limited access to specialist services, combined with the fact that often cluster headache sufferers may find themselves incorrectly referred to ENT, ophthalmic, or maxilla facial departments, may contribute to delayed, or protracted diagnosis (Kernick et al, 2006).
The key treatment options for acute treatment of cluster headaches are nasal triptans and oxygen, the evidence for this is good (class I). There is a paucity of sufficiently robust evidence to support any other treatment regime, in fact NICE (2012) explicitly warn against offering paracetamol, non-steroidal anti-inflammatories, opioids, ergots or oral triptans for the treatment of cluster headache. Given the emphasis on such treatments, paramedics may be best placed for delivery within the community, and is something which should be actively explored.
From meta-analysis integrating six randomised, double-blind, placebo-controlled studies, Law et al (2013) confirm that triptans are effective in the management of pain within 15 minutes of onset of cluster headaches. It is interesting to note that the evidence suggests that the use of subcutaneous triptans has greater efficacy than nasal administration (Law et al, 2013). It may be that as a first-line treatment nasal triptans are preferable due to the need for ongoing prophylactic/self-treatment and risk of needles. Individually the studies involved in the meta-analysis are class II, with any bias limited by blinding and exclusion of small sample sizes (below 10 participants). However, Law et al do note that the number of studies included constitutes a limitation, and sample size puts this analysis on the ‘limit of acceptability’ (Law et al, 2013: 20). This low sample size, combined with variability in doses and route of administration makes the quality of the evidence low, and suggests an overall need for further research in the use of triptans for cluster headache. Law et al (2013) suggest a need for a minimum of 200 participants in each treatment arm in at least two separate studies to be completely confident in any conclusions.
NICE (2012) also recommend the use of 100% oxygen at a flow rate of at least 12 litres per minute with a non-rebreathing mask and a reservoir bag. Through meta-analysis of nine randomised trials, Bennett et al (2008) identify that there is reasonable evidence to support hyperbaric oxygen treatment, and weak evidence to support the use of normobaric oxygen. Thus there is a need for further research into the efficacy of normobaric oxygen. However, due to the limited access and prohibitive cost of hyperbaric oxygen, it is still recognised that normobaric oxygen is a suitable treatment, and one which may be able to be provided by paramedics within the community, for those suffering recurrent episodes. More recently, there have been a number of empirical studies, including Cohen et al (2009), that support the use of high-flow oxygen therapy for acute and recurrent cluster headaches and possibly for all headache presentations (Ozkurt et al, 2012). This is with the knowledge that the true mechanism of action of oxygen in cluster headache is not well understood (Haan et al, 2012).
Due to the ‘clustered’ nature of episodic attacks in cluster headache, acute management of an isolated event needs to be combined with appropriate prophylaxis. It is also clear that every patient with active cluster headache requires frequent follow-up, both to ensure that optimum treatment is maintained and to monitor for treatment toxicity (Steiner and Martelletti, 2007). NICE (2012) do suggest the arrangement of provision of home and ambulatory oxygen; however, there is limited evidence associated with the cost effectiveness of this. Despite this, Edwards (2012) states that it is not excusable to use the cost of oxygen and triptans not to prescribe as the evidence, and NICE (2012) support their use for this debilitating illness. NICE (2012) suggest that home oxygen services cost around £175 per new person and around £69 per 6-month check-up; however, these costs are not specifically associated with cluster headaches. Ambulatory oxygen has been applied in certain areas in the UK and it is suggested that the application of a standardised approach can be beneficial. It is worth noting that the requirement for a specialist to prescribe ambulatory oxygen therapy may result in delayed access to this treatment.
The simplest and most effective preventative treatment for cluster headache is steroid use; however, long-term use brings with it severe adverse side effects (Leroux et al, 2011). There is class II evidence to support the use of calcium channel blockers (verapamil) in treating episodic cluster headache, as such NICE (2012) recommend this as the preferred treatment option. Further to this, NICE also recommend the need for research into topiramate to prevent recurrent cluster headache. Where chronic sufferers prove resistant to all options for prophylactic medication, this brings with it consideration of new, specialist treatment possibilities, such as occipital nerve blocks, for which there is good (class II) evidence, but little in the way of cost-benefit analysis (Bakar et al, 2013; Shimizu, 2013).
It is identified that medical professionals should highlight the fact that cluster headaches are a valid, medical condition with biological cause (NICE, 2012). Provision of information and support for people with headache disorders is essential in ensuring good outcomes, and can form part of paramedic health promotion activities within the community. To which end it is important for non-specialists (such as paramedics) to increase their knowledge, both around the presentation and management of the condition itself and around available information and support services. Such information is available from NHS Choices and the Organisation for the Understanding of Cluster Headaches (OUCH), and found in Steiner and Martelletti's ‘Information for people affected by cluster headache’ as part of the ‘Campaign to reduce the burden of headache worldwide’ (Steiner and Martelletti, 2007).
Conclusions
Cluster headaches are relatively rare. Despite this, the fact that acute episodes are severely painful and debilitating, and diagnosis and treatment is identified as being relatively easy, suggests that a focus on enhancing management of cluster headaches is valid, and by non-specialists within the community. In support of the NICE (2013) standards, enhanced awareness of decision-making for non-specialist clinicians, combined with consistent, direct access to specialist services, is essential to ensure timely diagnosis and management and minimise unnecessary suffering and pain associated with this debilitating illness (Kernick et al, 2006). An understanding of differentials and the assessment skills to appropriately elicit the necessary information is essential in categorising headaches and ruling out high risk, secondary causes. It is clear that this may require training and competence assurance, specifically around neurological examination and fundoscopy, for those who have not undertaken undergraduate paramedic education.
Recent literature and clinical guidelines are clear in their recommendations to not refer for imaging solely for reassurance and to refer to a specialist during the first bout of cluster headache (NICE, 2012; Holle and Oberman, 2013). Possibly due to the fact that the latter is purely based on consensus, there is, however, little indication of risk, or urgency of referral. The fact that the differential of dissected carotid artery is potentially life threatening and direct access to specialists is not necessarily a reality, suggests strongly that until proven otherwise the presentation of the first bout of cluster headache should be treated as a time-critical emergency by non-specialists. Thus it is likely these patients will be transported straight to hospital for further diagnosis.
However, for recurrent bouts of cluster headache, it seems reasonable to avoid unnecessary referral or imaging and enhance management using triptans and oxygen. It is clear that the use of ambulatory oxygen is effective, but there is limited understanding currently as to how effectively this is being applied for cluster headaches and its cost effectiveness. When compared to the potential for repeated daily callouts for ambulance service provision of oxygen, it is likely that the cost effectiveness of prescribed oxygen is easily demonstrated, however there is limited empirical evidence to support this. With the necessary awareness of the condition it is clear that the ambulance service could support patients should there be a lag between first diagnosis and access to ambulatory oxygen; however, more investigation would be required to determine if this would be cost effective, especially given the effect this increase in work may have on already stretched ambulance service delivery. Where cluster headache is the most likely diagnosis, it could also be valuable if ambulance services had direct access to triptans for first-line treatment, and local guidelines and patient group directives ought to be considered.
Taking the risk to patients during the first bout of cluster headache into account, it appears there is insufficient evidence to support autonomous decision-making out of hospital without specialist referral, or imaging. However, in cases of recurrent cluster headache it is appropriate for non-specialists to manage such cases in consultation with specialists. It is important that such management forms part of an integrated care plan for individuals suffering from recurrent cluster headache. From this it appears that further studies are required into the feasibility of implementing an integrated, whole system approach to managing acute and recurrent attacks of cluster headache in the UK.
Key Points
Conflict of interest: none declared