Migraine in the Emergency Department: A Prospective Multinational Study of Patient Characteristics, Management, and Outcomes

Headaches account for 1–3% of all emergency department (ED) visits, of which migraine is the most common discrete primary headache disorder diagnosed [1, 2]. Migraine is the leading cause of disability worldwide among people below the age of 50 years [3, 4]. Despite this, it has been suggested that migraine continues to be one of the most neglected and badly managed medical disorders [3, 5, 6]. The role of ED is primarily in abortive management, but there may be some role for recommending or referring patients for initiation of preventive therapy [7].

Migraine headaches can exhibit s similar clinical features to potentially life-threatening conditions, such as infectious and inflammatory conditions, vascular pathologies (including subarachnoid hemorrhage), and space-occupying lesions. They typically lack specific neurologic signs [8, 9]. This can make accurate diagnosis challenging. Diagnostic practices and acute management pathways may vary both between and within countries, depending, in part on individual hospital treatment pathways. Several regional and international guidelines exist for the assessment and management of acute migraine attacks [10-12]. They are broadly similar. The adherence to such guidelines, especially across geographical regions has not been well described. We embarked on the Headache in Emergency Departments (HEAD) study to understand the epidemiology of atraumatic headache in the EDs of ten countries. The aim of this planned sub-study is to describe the real-world demographics, clinical features, investigations, treatment, and outcome of patients with an ED final diagnosis of migraine.

This was a planned sub-study of a prospective multicentre observational study conducted over one calendar month in 2019 (for most sites March 2019), the HEAD study. It includes patients coded in the HEAD study cohort as having migraine as their final ED diagnosis.

For the parent study, participants were adult patients (aged ≥18 years) with nontraumatic headache as their presenting complaint [13]. Exclusion criteria were history of trauma to the head within 48 h of presentation, missing records, interhospital transfers, representation with the same headache as a recent previous visit and headache as an associated symptom rather than a main complaint. Determination of whether headache was a primary complaint was at the discretion of the local researcher based on all available data.

Qualifying adult patients presenting during the study period were identified from ED data management systems. Patients were identified prospectively but, depending on- site resources, some data were collected retrospectively.

Data collected included demographics, clinical assessment, investigation, treatment, diagnosis, disposition, and outcome. Data were collected from clinical records by local researchers onto piloted data forms or directly into the study database (depending on local processes and resources). Study data were collected and managed using REDCap electronic data capture tools hosted at the Joseph Epstein Centre for Emergency Medicine Research (Melbourne, Australia). For the majority of sites, the data were collected in March 2019, with some minor variations following delays in ethical/governance approval. Outcomes of interest for this study are demographics, clinical features, patterns of investigation, treatment, disposition and outcome of patients diagnosed with migraine as their final ED diagnosis.

Data analyses were predominantly descriptive (counts and proportions with confidence intervals where appropriate). It was undertaken using Stata v15 (College Station, TX, USA).

The lead ethics approval was by Melbourne Health Human Research Ethics Committee (HREC/43148/MH-2018). Ethics approval was subsequently obtained for each participating site according to local institutional requirements. In most jurisdictions, the study was conducted under waiver of consent. Patient consent was required in a few jurisdictions including the UK, where an opt-out consent approach was used, and approval obtained through the Health Research Authorization following application and review by committee (REC reference: 19/SW/0089). The study was registered with Australia and New Zealand Clinical Trials Registry (trial number 376695).

A total of 4,536 patients presenting with headache were included in the parent study across 67 hospital groups (74 EDs) in 10 countries (Australia 28, New Zealand 9, Turkey 9, UK 7, Singapore 4, Belgium, 4, France 3, Romania 1, Hong Kong 1, and Israel 1). For hospitals that provided caseload data, headache patients accounted for 1% of ED caseload (95% CI 1.0–1.1%). Migraine was the final diagnosis in 1,101 (24.3%) of patients.

The proportion of ED headache patients diagnosed with migraine varied between 18.0% and 32.2% of all patients presenting with headache in countries that contributed to at least 100 patients to the study (i.e., Australia, Turkey, New Zealand, Singapore, the UK, and France) (online suppl. Table A; for all online suppl. material, see www.karger.com/doi/10.1159/000520548).

Demographic features, past medical history, regular medication, disposition, and reattendance within 48 h are shown in Table 1. The median age was 37 (interquartile range 28–48) years. Females constituted 73.7% of patients, 3.5% of whom were pregnant.

58.4% (643/1,101) had a previous migraine diagnosis Half (50.1%) presented to ED within 24 h of symptom onset. Over a third (34.2%) of patients were taking chronic medications, of which 35.8%, 13.0%, and 11.9% were taking triptans, β-blockers, and tricyclic antidepressants, respectively. The majority were either directly discharged from the ED (64.4%) or from the ED observation unit (29.2%).

Clinical features and severity are shown in Table 2. Most patients presented with gradual onset of headache (57.5%) that was either generalized (44.6%) or unilateral (43.6%) in location and of moderate (40.6%) to severe (35.6%) pain (Table 2). Predominant symptoms were nausea or vomiting (60.2%), photophobia (42.8%), and patient report of current or transient new visual disturbance (22.2%). Neurological signs on clinical examination were uncommon.

Investigations are shown in Table 3. Head computed tomography (CT) was performed in about one quarter (25.9%) of patients. In 2 cases, CT raised suspicion of a vascular abnormality which was excluded on further testing. There was significant variation in the proportion of head CT performed among different countries (p < 0.001) (online suppl. Table B). Lumbar puncture was performed infrequently (2.1%), and none were diagnostic.

Self-administered and ambulance-initiated treatment is shown in Table 4. Common prehospital medications that were self-administered by patients include paracetamol (64.7%), nonsteroidal anti-inflammatory drugs (NSAIDs) (32.8%), triptans (18.9%), and codeine (14.0%) (Table 4). Among the 200 (18.2%) patients conveyed to hospital by ambulance, antiemetics (64.6%), paracetamol (35.9%), fentanyl (32.3%), and intravenous morphine (14.2%) were most frequently given by paramedics.

Pharmacological treatment in ED is shown in Table 5. 84.9% of patients received medications for their headache. Paracetamol (46.3%) and NSAIDs (42.7%) were the most commonly. Metoclopramide (22.8%), ondansetron (19.2%), chlorpromazine (12.8%), and prochlorperazine (12.8%) were also used. Chlorpromazine was only used in Australia and New Zealand. Over two-thirds (67.2%) of patients received intravenous fluids. Rescue medications (given >30 min after initial treatment) were required in 41.3% of patients. Almost a third (30.6%) received chlorpromazine as second-line therapy. Of the 437 patients who presented within 24 h of headache onset (39.7%) and reported moderate or severe headache, only 88 (20%) received initial treatment with a triptan or phenothiazine.

Almost a quarter of patients in the HEAD study who presented with atraumatic headache were ultimately classified at ED discharge as having migraine, underscoring the high prevalence of this condition [14]. The epidemiology of patients diagnosed with migraine in our study cohort is congruent with a previous large study in the USA where the predominant age-group was between 30 and 39 years, with female to male ratio of 3:1 [15]. A non-negligible proportion of females (3.5%) in our study were pregnant at the time of ED presentation, stressing the need for emergency physicians to be well aware of the principles of acute migraine management in the context of pregnancy [16, 17].

In our cohort, the classification as migraine was mostly in line with the International Classification of Headache Disorders-3 criteria for both variants with and without aura [18]. A high proportion of patients had recognized migraine features such as location (unilateral), severity (moderate to severe intensity), symptoms (nausea, vomiting, and photophobia), and aura (especially visual disturbance) [7]. The fact that 58.4% (643/1,101) of migraineurs were previously diagnosed to have suffered from migraine probably contributed to diagnostic considerations in the ED since migraine is a disorder of recurrent attacks.

Neuroimaging is not recommended in most instances, unless there are abnormal neurological findings on clinical examination, sudden severe headache or red flag features that do not fit into the diagnosis of migraine [19]. Our data could be interpreted to suggest overperformance of head CT (25.9%) and magnetic resonance imaging (2.0%) despite the known very low yield among migraineurs. That said, about 20% of patients did not have a confirmed prior diagnosis of migraine and a similar proportion had concerning clinical features so a cautious approach to ruling out a serious cause is probably justified. When comparing countries and regions, there were significant differences in the proportion of head CT performed, ranging from 14.2% (Singapore) to 50.0% (France) (online suppl. Table B). This study was not designed to explore reasons for this variation in practice. It could be postulated that issues such as structure of the health system, reimbursement and financing mechanisms, the medicolegal environment and health-seeking behaviors could vary between countries and contribute to this variation in practice. Alternatively, it may be due, in part, to random variation as the number of patients diagnosed with migraine in some countries was very low.

A possible approach to neuroimaging in patients with migraine is using the so-called SNNOOP10 list of flag items to screen for secondary headaches [20] (online suppl. Table C). These have been shown to have high negative predictive value [21]. They have not however been validated in a large ED cohort of patients with headache.

Abortive or symptomatic treatment of migraine remains the cornerstone of therapy in the ED [7]. Stratified treatment strategies based on the severity of the episode have been shown to be more effective than a step-care approach [22, 23]. An initial single large dose of medication given early during the headache has been demonstrated to be more effective than small, repeated doses [24]. Generally, simple analgesics such as paracetamol and NSAIDs are considered first-line agents for mild attacks due to the superior cost and side effect profile compared to migraine-specific medications [23]. In our cohort, very few patients had mild headache. In moderate to severe attacks, triptans combined with NSAIDs and an antiemetic agent (to treat nausea and assist absorption) or use of a phenothiazine such as chlorpromazine or prochlorperazine is suggested [25, 26]. These were prescribed infrequently (Table 5). When coupled with the finding that about half of patients have taken some analgesia before attending ED and the high proportion of patients requiring rescue therapy (>40%), this suggests that suboptimal treatment in ED is common and strengthens the case for stratified therapy based on severity and appropriate initial dosing [27, 28].

The proportion of patients treated with opiates in concerning. Opiates have been shown to be less effective than alternative agents and carry the risks of addiction (Dodson). Opiate use for treatment of headache more generally has been challenged. Despite this, their use is common (Pettat). Interestingly, although ergot alkaloids have been suggested for treatment of status migrainosus none were used as initial or rescue therapy in our patient population [29].

It is important where possible to avoid medication overuse headache or analgesic rebound headache. The risk is highest with opioids, moderate with triptans, and lowest with NSAIDs [27, 28, 30]. The number of patients with migraine in our study on regular triptans, codeine, and other opioid medications and who self-administered similar class of drugs, use of fentanyl and morphine in ambulances, and use of opioids in the ED may warrant further investigation to gain further insight into the potential risk of medication overuse headache.

The strengths of this sub-study include the size of the cohort and the incorporation of broad geographical regions in different continents, allowing some inferences in clinical practice related to variation in health care systems to be made. We were able to establish face validity of our methodology as the demographics of migraine patients in our cohort closely resembles previous epidemiological studies from the USA. The limitations include the cross-sectional nature of this study may introduce confounding issues that could lead to bias and subsequent misinterpretation [30, 31]. Information on discharge pain score, recurrence, and number of episodes of headaches as well as dose and timing of treatment were not collected. The number of cases from certain participating countries is very low, which could have led to imprecise estimates, although we attempted to circumvent this issue by only considering regions that contributed significant to the dataset in between country analyses. The diagnosis of migraine was made by the treating clinician and not confirmed independently. It is possible that some cases were miscoded; this is the nature of ED practice.

Migraine is one of the commonest neurological conditions in the ED. This study suggests that therapeutic practices are not congruent with current guidelines, especially for patients with severe symptoms. Efforts to improve and sustain compliance with existing management best practices are required urgently.

HEAD study group (Appendix, published as online suppl. material).

This study was conducted in accordance with the World Medical Association Declaration of Helsinki. The lead ethics approval was by Melbourne Health Human Research Ethics Committee (HREC/43148/MH-2018). Ethics approval was subsequently obtained for each participating site according to local institutional requirements. In most jurisdictions, the study was conducted under waiver of consent. Patient consent was required in a few jurisdictions including the UK, where an opt-out consent approach was used, and approval obtained through the Health Research Authorization following application and review by committee (REC reference: 19/SW/0089). The study was registered with Australia and New Zealand Clinical Trials Registry (trial number 376695).

The authors have no conflicts of interest to declare.

No funding was received for this study.

T.W., A.M.K., K.H.C., and W.S.K. conceived the idea of the HEAD study. T.W., A.M.K., W.S.K., K.H.C., F.B.K., G.K., M.A.K., S.K., C.A.G., T.R., D.H., and S.L. developed the concept further through a series of international teleconferences and agreed on the main study protocol that was drafted by A.M.K. T.W. drafted the original manuscript. All the authors contributed with revisions and agreed with final submission version.

Data are available from the corresponding author upon reasonable request.