2025 Volume 65 Issue 3 Pages 197-202
Headaches are a common complaint in the emergency department (ED). Understanding the characteristics and outcomes of headaches, especially in undiagnosed patients, is important for improving headache care in the ED. We conducted a retrospective study of 171 headache patients at the ED of Kagawa Prefectural Central Hospital, with a follow-up for the primary and undiagnosed headache groups via telephone to assess long-term outcomes. Primary, secondary, and undiagnosed headaches accounted for 15.2%, 58.4%, and 26.3% of cases, respectively. All life-threatening secondary headaches were successfully excluded with imaging tests performed on 73.7% of cases, despite the low rate of treatment in the ED (18.1%). Among the undiagnosed headache cases, the recurrence of severe headaches was low (11.9%), though awareness of chronic headaches was high (47.6%). Emergency physicians should be aware of the possibility that patients at an ED presenting an undiagnosed headache may include chronic headache patients. Seamless collaboration between the ED and headache specialists is needed to manage undiagnosed headaches in the ED.
Headaches are a major presenting symptom of patients visiting the emergency department (ED) and also the most common neurological disorder in the ED1)2). Patients presenting headaches in the ED are generally categorized according to the International Classification of Headache Disorders, 3rd edition (ICHD-3)3) as having either a primary headache (e.g., migraines, tension-type headaches, or trigeminal autonomic cephalalgias) or a secondary headache, which are caused by a wide variety of underlying etiologies. The primary goal of emergency physicians is to rule out life-threatening causes of headaches, such as haemorrhagic or ischemic stroke, meningitis, or brain tumors. To assist in identifying such dangerous secondary headaches, a red flag list known as the “SNNOOP10 list”, is frequently employed4). By utilizing this list along with appropriate imaging studies, the management of headaches in the ED generally follows an exclusion-based approach5).
A characteristic feature of headache management in the ED is that patients are sometimes simply diagnosed as having a ‘headache’ or ‘NOS’ (not otherwise specified) headache without a definitive diagnosis6)7). Once life-threatening headaches are excluded, patients are often discharged without any clinical follow-up, and it is often unclear whether their undiagnosed headache will recur or what the correct diagnosis might be. This lack of a follow-up makes it difficult to establish an appropriate approach for managing undiagnosed headaches in the ED.
The aim of this study was to investigate the clinical characteristics and outcomes of “undiagnosed headaches” in an ED.
This study was conducted at Kagawa Prefectural Central Hospital (Kagawa, Japan), known for treating a high volume of acute diseases including hemorrhagic and ischemic stroke in the Kagawa region, accepting approximately 3,500 ambulances annually with a capacity of around 500 beds. Patients can access the ED either by ambulance or a direct visit. The ED is open 24 hours a day. This study focused on patients aged 18 years and older who visited the ED at Kagawa Prefectural Central Hospital and presented headaches as their primary complaint between October 2022 and September 2023. Clinical information for the participants was retrospectively collected from medical records. Based on the diagnosing doctors’ assessment, patients were categorized into three groups: 1) those diagnosed with primary headaches, 2) those diagnosed with secondary headaches, and 3) those with undiagnosed headaches. The groups with primary and secondary headaches were further subdivided based on their headache diagnosis, according to the ICHD-3. Collected data included age, gender, ambulance use, time of visit (with daytime defined as 8:00 AM to 8:00 PM and nighttime as the remaining hours), whether blood tests or imaging studies were conducted, hospitalization status, headache treatment, and whether the patient met the SNNOOP10 criteria. The SNNOOP10 list was evaluated only if patients were clearly identified as a match from the medical record entries and patient data. We excluded cases in which headache assessment was impossible due to impaired consciousness and cases of ambulance use for hospital transfer.
Additionally, for patients diagnosed with primary or undiagnosed headaches, telephone follow-ups were conducted within 3 to 12 months of the ED visit in order to confirm outcomes of the headache. This included whether patients experienced chronic headaches at the time of the ED visit, received a new diagnosis regarding the headache post-visit, experienced headaches of similar severity post-visit, had current awareness of chronic headaches at the time of the follow-up, and details regarding headache frequency, severity, and follow-up care. Headache severity was assessed using the six-item short-form survey for measuring headache impact (HIT-6)8).
Statistical analysis was performed using GraphPad Prism 10 (GraphPad Software, Inc., San Diego, CA, USA). Data were analyzed using the Chi-squared test, Fisher’s exact test, the Mann-Whitney test, or the Kruskal-Wallis test, as appropriate. Statistical significance was set at P < 0.05. This study was approved by the Ethics Committee of Kagawa Prefectural Central Hospital (Approval Number: 1196, titled Reconfirming of Undiagnosed headache Outcome in Kagawa: RUOK survey). This study used patients’ clinical data retrospectively, and an opt-out method was employed via the hospital website (https://www.chp-kagawa.jp/) in accordance with the hospital’s ethical guidelines.
During the study period, 181 patient records with headaches as the chief complaint were identified, and after exclusions, 171 records were analyzed. The proportion of female patients was 66.1%, the average age was 50.0 years (ranging from 18 to 101), the ambulance use rate was 32.1% (n = 55), and the rate of visits during daytime hours was 29.8% (n = 51). Blood tests were conducted in 80.1% of cases (n = 137), computed tomography (CT) scans in 50.9% (n = 87), magnetic resonance imaging (MRI) scans in 17.0% (n = 29), and both CT and MRI scans in 5.8% (n = 10). The hospitalization rate was 21.6% (n = 37), symptomatic treatment was administered in 18.1% of cases (n = 31), and 78.4% of patients (n = 134) met the SNNOOP10 list. Primary headaches were diagnosed in 15.2% of cases (n = 26), secondary headaches in 58.4% (n = 100), and undiagnosed headaches in 26.3% (n = 45) (Fig. 1). Among secondary headaches, the most common diagnosis was headaches attributed to infection (n = 40), followed by headaches attributed to cranial and/or cervical vascular disorder (n = 21) and headaches attributed to non-vascular intracranial disorder (n = 12), as presented in Fig. 1. The characteristics of each headache group are shown in Table 1. The primary headache group had the highest proportion of females, at 80.8% (n = 21), and was significantly youngest with an average age of 35.6 years (P < 0.01). The rate of symptomatic treatment was significantly highest, at 57.6% (n = 15, P < 0.01), and the rate of meeting the SNNOOP10 list was significantly lowest at 26.9% (n = 7, P < 0.01). The secondary headache group had the highest hospitalization rate, at 31.0% (n = 31), the lowest rate of symptomatic treatment in the ED, at 16.0% (n = 16), and the highest rate of meeting the SNNOOP10 list, at 89.0% (n = 89). The undiagnosed headache group had the highest rate of blood tests, at 82.2% (n = 37), and the rate of CT or MRI scans was significantly highest, at 88.9% (n = 40, P < 0.05). The outcomes of the primary headache and undiagnosed headache groups were confirmed via a telephone follow-up (Table 2). Responses were obtained from 23 out of 26 patients in the primary headache group and from 42 out of 45 patients in the undiagnosed headache group. The rate of chronic headaches at the time of the ED visit was significantly higher in the primary headache group, at 87.0% (n = 20), compared to 47.6% (n = 20) in the undiagnosed headache group (P < 0.01). Only one case in the undiagnosed headache group received a new diagnosis after the ED visit, which was herpes zoster and had a good prognosis. There was no significant difference between the groups in terms of experiencing severe recurrent headaches after the ED visit (21.7%, n = 5 vs. 11.9%, n = 5). At follow-up, the rate of chronic headaches was significantly higher in the primary headache group, at 91.3% (n = 21), compared to 45.2% (n = 19) in the undiagnosed headache group (P < 0.01). A comparison was made between patients with primary headaches and those with chronic headaches in the undiagnosed headache group (Table 3). The rate of receiving headache treatment was 28.6% (n = 6) vs. 26.3% (n = 5), the average number of headache days per month was 5.1 vs. 6.3, and the average HIT-6 score was 50.5 vs. 47.7, with no significant differences observed in any of these measures.
Primary headaches accounted for 15.2% of cases (n = 26), secondary headaches for 58.4% (n = 100), and undiagnosed headaches for 26.3% (n = 45). Secondary headaches were diagnosed based on ICHD-3, with the preceding numbers indicating their classification.
| Primary headache (n = 26) |
Secondary headache (n = 100) |
Undiagnosed headache (n = 45) |
All (n = 171) |
|
|---|---|---|---|---|
| Female (%) | 80.8 | 58.0 | 73.3 | 66.1 |
| Age (%) | 35.6** | 50.4 | 57.6 | 50.0 |
| Used ambulance (%) | 26.9 | 35.0 | 35.7 | 32.1 |
| Daytime access (%) | 15.4 | 35.0 | 26.2 | 29.8 |
| Received blood test (%) | 76.9 | 80.0 | 82.2 | 80.1 |
| Received CT and/or MRI (%) | 76.9 | 66.0 | 88.9* | 73.7 |
| Hospitalization (%) | 7.7 | 31.0 | 8.9 | 21.6 |
| Treatment in ED (%) | 57.6** | 16.0 | 22.2 | 18.1 |
| Meet SNNOOP10 (%) | 26.9** | 89.0 | 86.7 | 78.4 |
A Kruskal-Wallis test was performed to assess statistical differences across the groups, followed by Dunn’s post-hoc test to identify specific group differences. Significant differences between groups are indicated as follows: *P < 0.05, **P < 0.01. ED: emergency department.
| Primary Headache (n = 23) |
Undiagnosed Headache (n = 42) |
P-value | |
|---|---|---|---|
| Chronic headache at ED visit (%) | 87.0 (n = 20) | 47.6 (n = 20) | 0.004 |
| Other diagnosis after ED visit (%) | 0 | 2.4 (n = 1) | 1.0 |
| Experience of severe headache after ED visit (%) | 21.7 (n = 5) | 11.9 (n = 5) | 0.478 |
| Chronic headache at follow up (%) | 91.3 (n = 21) | 45.2 (n = 19) | <0.001 |
ED: emergency department.
| Primary Headache (n = 21) |
Undiagnosed headache with chronic headache (n = 19) | P-value | |
|---|---|---|---|
| Already receiving headache medication (%) | 28.6 (n = 6) | 26.3 (n = 5) | 0.873 |
| Frequency of headache (days per month) | Mean 5.1 (SD 7.0) | Mean 6.3 (SD 5.4) | 0.275 |
| HIT-6 | Mean 50.5 (SD 9.6) | Mean 47.7 (SD 10.6) | 0.377 |
HIT-6: headache impact test-6, SD: standard deviation.
In this study, we presented an overview of the clinical characteristics of patients with headaches in our ED, and notably, we also clarified the previously unknown prognosis of patients with undiagnosed headaches.
Clinical characteristics and management of headache patients in the EDIn our study, primary headaches accounted for 15.2% of cases, significantly fewer than the 58.4% reporting secondary headaches. This is an atypical result, as primary headaches are often considered the predominant reason for ED visits, despite regional and institutional differences. For example, studies in Italy and Austria reported primary headache diagnoses in 77.3% of cases, while in Singapore, it was 73.6%1)9). The lower proportion of primary headaches in our hospital may be biased by the presence of other comorbidities leading to secondary headaches or by physicians’ hesitation to diagnose primary headaches, resulting in undiagnosed headaches. The correct diagnosis rate for headaches may be lower among non-specialist physicians compared to neurologists or headache specialists, mainly due to a lack of understanding of the ICHD-36). A Canadian cohort study reported a correct diagnosis of 35.7% for primary headaches in their ED10). Therefore, it is not recommended to manage patients exclusively based on the ED diagnosis after discharge from the ED, without verifying the accuracy of the diagnosis.
Excluding “life-threatening secondary headaches” is a major focus for physicians in many EDs, which is why screeners like SNNOOP10 are widely used, and thorough examinations are recommended for patients with red flags. In our cases, all dangerous secondary headaches were successfully excluded at the time of ED presentation, and some patients classified with cerebrovascular headaches had severe outcomes, including death in the ED, highlighting the importance of diagnosing potentially life-threatening secondary headaches11). In this context, our study showed that a head CT was performed in 50.9% of patients and an MRI in 15.0%, whereas in previous reports CT imaging was performed in 29–53% of patients1)9)12)13). Since conditions like reversible cerebral vasoconstriction syndrome (RCVS), intra/extracranial artery dissection, and subarachnoid hemorrhage (SAH), which can have severe outcomes, are often difficult to diagnose with CT alone14). MRI may be a viable option for high-risk patients, especially in countries with high MRI availability, like Japan. Moreover, the SNNOOP10 list is known for its high sensitivity5), and its use is recommended for patients fitting the criteria to conduct examinations with differential diagnoses in mind. Retrospective examination of our cases showed that 89.0% of secondary headaches fit the SNNOOP10 list, compared to 26.9% of primary headaches, highlighting the list’s importance and effectiveness in identifying high-risk patients in the ED. However, migraine patients can present sudden pain exacerbations or pattern changes that mimic secondary headaches15)16), making it difficult to diagnose these headaches using the SNNOOP10 list.
From a therapeutic standpoint, headache treatment was provided to only 18.1% of headache patients. Previous reports6)17) showed a much higher rate of treatment (57.0–58.1%), while another study showed that headaches in 26.7–80.2% of patients did not improve upon ED discharge18). Since patients visit the ED not only to identify the causes of their headaches but also to alleviate unbearable pain, providing effective pain relief and prescribing analgesics for home use seem crucial. According to the “Stratified care strategy”, which aims to select treatment based on the severity of pain, it may be effective for the acute phase of a headache; not only NSAIDs and acetaminophen, but also triptans, ergotamines, and antiemetic drugs, should be considered as treatment options in the ED19).
Prognosis of Undiagnosed HeadachesIt has already been reported that 24–44% of patients in the ED do not receive a specific diagnosis and are simply recorded as having a “headache”1)12)13)20), without a clear understanding of their etiology and outcomes. In our study, 86.7% of the undiagnosed headache group fitted the SNNOOP10 list at the time of the ED visit and had high rates of blood and imaging tests, yet no significant abnormalities were found. A follow-up revealed the effectiveness of the exclusion process for secondary headaches, and if primary or secondary headache indicators were absent during ED evaluation, the recurrence rate of headaches was low (21.7%), regardless of the actual pathology of the headache.
Additionally, about half of the undiagnosed group were aware of chronic headaches. While their hospital visit rates for headaches were lower than those with primary headaches, the severity estimated by the HIT-6 showed no significant difference, and their frequency was even higher. This group may include patients with a primary headache, medication overuse headache (MOH) or an undiagnosed secondary headache. Previous reports indicated that two-thirds of patients were not referred to headache clinics upon discharge, leading to the lack of a specific headache treatment21)22). Our results underscore the need for appropriate referrals from the ED to headache clinics.
Our study has several limitations. First, the data were from a single center and a single year and may not be generalizable due to the small number of cases. Second, the group names did not strictly indicate the correct diagnosis of headaches. The diagnosis of headaches was entirely dependent on the ED doctors’ assessment, and even a follow-up through telephone interviews alone could not confirm the correct diagnosis. For precise diagnostic data collection, further prospective studies are needed. Third, secondary headaches accounted for the highest proportion in our study, with infections being the most common cause, including many COVID-19 patients, indicating potential variations due to the study period. Fourth, as the SNNOOP10 list was only evaluated in terms of confirmed retrospective matches, it may be underestimated as it includes cases that are actually matches but not in the medical record data. It would thus be desirable to evaluate the SNNOOP10 list prospectively.
This study revealed that patients with “undiagnosed headaches” who underwent appropriate testing and exclusion of secondary headaches had a low recurrence rate of severe headaches but a high likelihood of underlying chronic headaches. To improve headache management in the ED, the following challenges were identified:
1. A high diagnostic rate for secondary headaches needs to be maintained through testing, while adequately providing pain relief, as needed.
2. Recognition that the undiagnosed group may include chronic headache patients and offer referrals to headache clinics.
Emergency physicians managing headaches should aim to appropriately classify headaches into primary, secondary, and undiagnosed categories, and provide seamless treatment plans from acute to chronic phases to improve patients’ quality of life. This requires an accurate understanding of the patient population in each medical institution and collaboration between emergency and headache specialists in each region.
The authors disclose no potential conflicts of interest.
The authors thank the Emergency Department crew for their thoughtful support in this study.
Abstract of this work was presented at the 115th Chugoku-Shikoku Regional Meeting of the Japanese Society of Neurology and recommended by the conference chairperson for the publication to Rinsho Shinkeigaku.
computed tomography
EDemergency department
ICHD-3the International Classification of Headache Disorders 3rd edition
MRImagnetic resonance imaging
MOHmedication overuse headache
NOSnot otherwise specified
RCVSreversible cerebral vasoconstriction syndrome
SAHsubarachnoid hemorrhage
