- Top of page
Background and purpose: To find a specified diagnosis for every patient investigated in the hospital emergency room for acute headache suspicious of subarachnoid haemorrhage (SAH), and to describe similarities and differences between the diagnostic groups.
Methods: We used a standardized set of questionnaires and supplementary tests, including cerebral computed tomography (CT) and if needed lumbar puncture, in the investigation of the patients. Two neurologists diagnosed the same cases independently.
Results: We found 30 different diagnoses as the cause of acute headache. Sixteen per cent had a SAH, and 57% had a primary headache. Patient characteristics, conditions at headache onset and accompanying symptoms were surprisingly similar in the diagnostic groups. For three SAH patients, it took 30 min to reach maximum pain intensity. In all diagnostic groups, a large proportion of the patients reached maximum pain within 60 s.
Conclusions: To distinguish between benign and malignant causes of acute headache is difficult based on clinical features. The consistent use of CT and lumbar puncture is valuable when investigating sudden onset ‘first or worst headache ever’. This can reduce the risk of missing a SAH diagnosis, and make it possible to give more exact diagnoses to patients suffering from both primary and secondary headaches.
- Top of page
Acute headache is a symptom frequently presented in the emergency room, and is caused by a great variety of clinical conditions, the most dreaded being subarachnoid haemorrhage (SAH). The notion that all SAH patients experience an instantaneous gunshot- or sledgehammer like headache has been proven false .
There is no definition of acute or sudden onset headache in the International Headache Society (IHS) guidelines , and no author has attempted to draw an absolute limit to what can be regarded as acute, although in several studies different time-limits have been employed. These headache onset limits range from <10 s , via <1 min , <1 h , to no specified limit . The presence or lack of accompanying symptoms like nausea, vomiting, photophobia and collapse at onset does not seem reliable to distinguish between SAH and benign causes for acute headache [1,3,4].
Terms like ‘acute headache’, ‘benign thunderclap headache’ and ‘non-SAH headache’ are used for lack of a specific diagnosis. These are terms that reveal nothing about the cause of the headache, whether it will recur or how it is best treated. Many secondary headaches besides SAH can be missed if the appropriate investigations are not made, with prolonged discomfort or even disability as the result.
Our study aimed at specifying the many possible causes for acute headache, and to observe similarities and differences in demography and clinical presentation to elucidate the many diagnostic considerations necessary when dealing with an acute headache patient.
- Top of page
A total of 433 consecutive adult patients seen in the emergency room at Akershus University Hospital in Norway 1998–2002 with acute headache were included in this prospective study. All had what they described as ‘the first or worst headache ever’, reaching maximum pain within seconds to hours, without a known cause and with clinical features suspicious of SAH. The hospital serves a population of about 350 000, and has the only emergency room in this region. Thus, practically all patients above 16 years of age with an acute headache are received in our emergency room, and as there are no separate Emergency Departments in Norway, all headache patients were examined directly by doctors from the Neurology Department. Patients were mostly referred from general practitioners, but quite frequently patients were brought directly from home to the hospital emergency room by relatives or by ambulance. Children were not included, as they are treated in the Pediatric Department.
We developed a set of questionnaires intended to comprise data relating to conditions at headache onset, headache characteristics and accompanying symptoms, results of clinical and supplementary tests and patient characteristics and history.
On arrival, all patients went through a clinical and full neurological examination. Blood tests, electrocardiograms and computed tomography (CT) were performed without delay, and if the CT was normal or inconclusive, a lumbar puncture was done when 12 h had elapsed from headache onset. The cerebrospinal fluid (CSF) was analysed with a white blood cell count and protein and glucose levels, and for 270 patients we measured the opening intraspinal pressure. Spectrophotometry for detection of bilirubin and oxyhaemoglobin was performed for 260 patients. Magnetic resonance imaging and cerebral angiography were not part of the routine investigations, but applied when considered necessary.
Within 24 h from admission, the patients were given a second questionnaire relating to data such as marital and occupational state, previous headache conditions or complaints and family history. They were at the same time examined by one of two neurologists (EMD, SHB) for neck mobility and head and neck muscle tenderness. By this time most of the patients with a diagnosed SAH were already referred to the neurosurgical unit serving our region.
Two neurologists (EMD, SHB) went through all the case histories and test results and made separate diagnoses, using as guidelines the International Classification of Headache Disorders . Both are experienced neurologists with a special interest in headache. The primary evaluation of diagnoses showed a 75% concordance. In the remaining 25% of cases, the procedure was repeated without discussing the diagnostic criteria, and the next evaluation showed a further 75% concordance. The still diverting diagnoses (n = 22) were then discussed openly and final, common diagnoses were attained (Table 1). Data were analysed with spss version 14.0 (SPSS Inc., Chicago, IL, USA).
Table 1. Some characteristics of secondary versus primary headaches
| || Subarachnoid haemorrhage (n = 71)||Other secondary headaches (n = 113)|| Primary headaches (n = 249)|
| Male||30 (42)||55 (49)||95 (38)|
| Female||41 (58)||58 (51)||154 (62)|
| Mean||53.3||45.2|| 41.1|
| Range||19–90||15–84|| 15–88|
|State at ictus|
| Activity||32 (45)a||36 (33)||117 (47)|
| Rest||20 (28)||77 (67)||126 (51)|
|Maximum pain intensity|
| Within seconds||50 (70)||62 (55)||140 (57)|
| Within 60 min||16 (23)||34 (30)||74 (30)|
| Over 60 min||0 (0)||13 (12)||31 (13)|
| Moderate||27 (38)||40 (35)||111 (45)|
| Severe||32 (45)||25 (22)||70 (28)|
| Moderate||24 (34)||19 (17)||49 (20)|
| Severe||26 (37)||19 (17)||50 (20)|
| Moderate||25 (35)||39 (35)||92 (37)|
| Severe||8 (11)||13 (12)||42 (17)|
The study was approved by the Regional Ethics Committee.
- Top of page
Patients were included in the emergency room when SAH could not be ruled out and no other diagnosis was evident. Thus, the distribution of diagnoses does not reflect all patients admitted with headache as one of their symptoms. Obviously in the 5-year study period, we admitted more than one patient with e.g. cerebral venous thrombosis or internal carotid or vertebral artery dissection, but the others had characteristic symptoms and clinical findings of neurological deficits pointing more directly to a diagnosis other than SAH and were consequently not included in the study. A typical sinusitis is rarely missed, but inflammation of the ethmoid and sphenoidal sinuses especially can sometimes present with acute headache and very few infection symptoms. The fact that the patients were directly examined by doctors from the Neurology Department may have had an influence on the inclusion, but this could go either way because neurologists are perhaps more aware of the risk of missing a SAH but at the same time could be better at taking a good history of headache patients and thus decide upon a likely diagnosis.
Sixteen per cent of the patients in our study had SAH. Other studies show some variety in the proportion of patients with SAH, from 7% , 11% , 19%  to 64% . The inclusion criteria in these studies were not similar. Given a proportion of SAH between those of other observers, we think our findings are fairly representative of patients being investigated for acute headache at a large general hospital.
The most prominent finding of this study is that there is considerable overlap between the diagnostic groups regarding headache onset characteristics and accompanying symptoms. There is no single observation that clearly separates particular diagnoses or primary from secondary headaches, and even for SAH the symptom description is so little different from those of the other diagnoses that we maintain that a CT of the brain has to be performed in all cases of acute headache admitted to hospital. This was also the conclusion of Linn et al. , albeit in a study with a much larger proportion of SAH patients.
A large proportion of the patients (n = 163) had a primary headache that did not exactly fit the IHS criteria (IHSC) for either ETTH or migraine without aura. According to the IHSC, ETTH pain is mild to moderate and not accompanied by nausea and vomiting. However, a large epidemiological study  found that 13% of patients classified as ETTH reported severe headache during some attacks, and in a treatment study  several ETTH patients reported pain intensity on a visual analogue scale (VAS) to be 7–8 (10 being the worst pain imaginable). In the American Migraine II Study, the proportion of TTH patients reporting nausea was 49%, vomiting 19% and photophobia 70% . Some authors have suggested that the IHSC are difficult to use in an emergency room setting, and that here up to 40% of primary headache patients could not be given a specific diagnosis according to IHSC [10–12].
In our study, 53% of the presumed ETTH patients reported nausea along with 22% with vomiting and 39% with photophobia. The patients were asked specifically if they felt nauseous on arrival, and we recorded both moderate and severe nausea, as well as moderate and severe vomiting. Eleven per cent of presumed ETTH patients reported severe nausea, in comparison to 50% of the migraine without aura patients. Nausea and vomiting are possibly more likely to be present in a moderate degree during an acute attack of ETTH in the emergency room, compared to the conditions in an outpatient clinic where diagnoses are made strictly according to IHSC.
For most of the patients, the headache that brought them to hospital was their first severe sudden onset headache. As such, it can be argued that for instance the criteria for a migraine diagnosis are not fulfilled, because a recurrence of attacks is necessary. However, we believe that many of our patients suffered their first attack of migraine and they were consequently diagnosed as migraine with or without aura. Both migraine and TTH are primary headache disorders and the classification is based solely on clinical features and the presence or absence of associated symptoms. The difficulty in distinguishing ETTH from migraine headache is acknowledged, and some authors even contend that these disorders are the extremes of a continuum with a common pathophysiology [9,13]. Our diagnoses were based on an over-all evaluation of the clinical presentation, taking into account which was the dominating symptom, how the symptoms progressed and whether there was palpable tenderness of head and neck muscles. A further scrutiny of the onset and progression of symptoms resulted in more concordant diagnoses. We later compared the first round of diagnostic concordance to the second, and found that in the diagnostic groups approximately the same proportion of cases had to be revised, with the exception of orgasmic and exertional headache. Based on clinical judgment and the method of two neurologists separately reviewing the same case histories, we think our diagnoses are adequate. We did not employ the IHS diagnostic code ‘Probable migraine’ but chose to designate the diagnostic groups as either ETTH or migraine without aura, based on the considerations discussed above.
In our study, 58% of all patients included were women, and the female proportion of SAH patients was also 58%. A higher incidence of SAH in women, especially after the age of 55, has been shown in a meta-analysis . The large preponderance of women we found in the groups with migraine with and without aura is consistent with prevalence numbers in the general population . Next to migraine, the group with idiopathic low CSF pressure had the highest female ratio (71%), in accordance with previous studies . Orgasmic headache has been considered to affect males predominantly , but we found the female percentage to be as high as 47%.
The diagnostic groups were quite homogeneous with respect to age (mean age within 35–45 years), the exceptions being the groups with stroke and SAH where the mean age was higher, and the group with lymphocytic meningitis where the mean age was lower. Several studies have shown that age >50 years is a significant independent predictor of serious intracranial pathology [18,19], and that patients diagnosed with SAH have a higher mean age than patients with a benign cause for sudden onset headache [1,3].
We found that 70% of SAH patients reached headache maximum within 60 s, and none after more than 30 min. However, only three diagnostic groups had <50% of patients reaching pain maximum within 60 s, and the groups orgasmic headache had 100%, primary exertional headache 71% and lymphocytic meningitis 67% of patients in this category. Thus, the time elapsed to reach maximum headache clearly is a very unreliable variable to distinguish SAH from other causes of acute headache.
The appearance or lack of accompanying symptoms seems even less reliable. In most diagnostic groups, the majority of patients experienced nausea. In all groups, 35% or more of the patients stated photophobia. Vomiting was the accompanying symptom more likely to occur with SAH, also shown in previous studies , but was equally frequent in patients suffering from migraine without aura. Interestingly, the groups with primary exertional headache and orgasmic headache, with their known association to migraine, had some of the lowest frequencies of vomiting, photophobia and nausea.
In only three diagnostic groups did the number of patients arriving at hospital within 24 h of headache onset amount to 75% or more, the groups being SAH, migraine without aura and migraine with aura. Eighteen patients with SAH arrived later than 24 h after headache onset, and seven of these cases were due to a delayed referral from a general practitioner. This suggests the need for an increased alertness towards sudden onset headache, both amongst physicians and in the general public. Through public health measures and educational programmes for physicians, the notion of sudden onset first or worst headache should create the same attention and level of investigation as sudden chest pain. A study from Canada  found that patients who were admitted to a non-teaching hospital were more than twice as likely to have a missed SAH as those at teaching sites, and that there was a 2.7-fold risk of misdiagnosis for patients with less acute presentations. They found the rate of missed SAH diagnoses to be 5.4%, whereas other studies have shown this rate to be 12%  and 20% , and there are reports of even higher rates . The patients with few accompanying symptoms and no neurological deficits are the most likely to be misdiagnosed, which is a regrettable fact, as these SAH patients are the most likely to gain from early treatment and make a full recovery. In this light, we maintain that patients who are referred to hospital with their first or worst headache should undergo investigation with CT, and if this is normal or inconclusive a lumbar puncture. To the best of our knowledge, no SAH patient was missed when this procedure was followed. A diagnosis based solely on clinical presentation is in most cases not reliable, as shown also by Linn et al. . The cost–benefit aspect of this approach can clearly be discussed, but the fact remains that both economical costs and human misery are tremendous in a case of poor outcome in a SAH patient, and a missed or delayed diagnosis is one of the most important factors for a poor outcome [20,21,24].