Abstract
Background: A delay in diagnosing aneurismal subarachnoid haemorrhage (SAH) occurs in a substantial proportion of patients who present with headache as the only symptom. Objective: To identify determinants for a delay in referral in patients with SAH, who present with isolated headache to the general practitioner (GP). Methods: For all 112 patients with SAH admitted to the hospital between October 2008 and June 2009, we sent a questionnaire to the GPs asking for details presented during the initial GP visit. In this retrospective study, we included 31 patients with SAH who initially presented with isolated headache. We assessed acuteness of headache onset, history of headaches and a patient delay as determinants for delayed referral (> 2 h after a visit to the GP), by calculating risk ratios (RRs) with corresponding 95% confidence intervals (CIs). Results: Referral was delayed in 18 of these 31 patients. The delay occurred in all 10 patients in whom the GP was unaware of the acute onset of headache and in 8 of 21 patients in whom the GP was aware of this symptom (RR: 2.6; 95% CI: 1.5–4.5). A history of headaches (RR: 1.8; 95% CI: 1.1–3.0) and a patient delay (RR: 2.1; 95% CI: 1.0–4.5) also increased the probability of delayed referral.
Conclusion: In patients with SAH who presented with isolated headache to the GP, GP's unawareness of the acute onset of the headache, a history of headaches and late presentation by the patient increased the probability of delayed referral.
KEY MESSAGE:
In patients with aneurysmal subarachnoid haemorrhage who present to the general practitioner with isolated headache, unawareness by the GP of the acute onset of the headache, a history of headaches and late presentation after headache onset increase the probability of delayed referral by the general practitioner.
Background
Aneurysmal subarachnoid haemorrhage (SAH) has a case fatality rate of approximately 35% in Western countries (Citation1). Delayed diagnosis leads to re-bleeding in about half of the patients (Citation2,Citation3), resulting in death or dependency in 80% of cases (Citation4). Immediately referred patients deteriorate only in 2% of instances (Citation3). Delayed diagnosis and referral occurs in around 20–25% of patients with SAH (Citation2,Citation3,Citation5,Citation6).
In the Netherlands, virtually all initial medical contacts are with the general practitioner (GP), except in case of obvious emergencies, such as severe trauma or coma. In general practice, headache complaints account for 4% of consultations. Only 0.1% of these patients have SAH (Citation7), making SAH a diagnostic challenge for GPs. The risk of initial misdiagnosis is greatest in the 30% of patients who present with headache as the only symptom of SAH (Citation3,Citation5,Citation8,Citation9), whereas such patients have a relatively good prognosis (90% good or excellent outcome at six weeks) when referred immediately (Citation3). When patients with SAH and headache as the only symptom, are referred with a delay their chances of a good outcome are reduced to 60% (Citation2–4). The cardinal, though non-specific, feature of SAH is the acute onset of the headache, see Box I (Citation10). Although probably few patients will omit to mention the severity of their headache, they may not realize the significance of the speed of onset and forget to mention this (Citation10).
Identified reasons for delayed diagnosis of SAH were a preserved consciousness and isolated headache (i.e. without other neurological signs or symptoms) at presentation (Citation3,Citation9,Citation11). Patients with preserved consciousness and isolated headache will often present to the GP in the Netherlands. There are no identified predictors of delayed diagnosis of SAH by the GP in this subset of patients. If there were, these predictors could increase the GP's awareness of possible SAH in patients presenting with isolated headache and prevent delayed diagnosis and referral in these patients.
In a prospectively collected database of patients with SAH who were referred to a neurosurgical centre, we investigated three likely determinants for a delayed referral in patients with SAH who present with isolated headache to the GP: unawareness by the GP regarding the acute onset of the headache, a patient's history of headaches and a patient delay.
Methods
Study setting and subjects
We used data from patients included in a prospectively collected database of patients with SAH. These patients were referred to the Neurosurgical Centre of the University Medical Centre of Utrecht, between October 2008 and June 2009. The presence of subarachnoid bleeding had to be confirmed by computed tomography (CT) or xanthochromia of the cerebrospinal fluid. The presence of an aneurysm was confirmed by CT-, magnetic resonance-, or catheter angiography.
Data collection
In all patients admitted during the study period with SAH, information on date and time of arrival in the hospital was collected from the hospital administrative records, and information on age, gender, in hospital mortality and clinical presentation on admittance to the hospital was retrieved from the hospital patient records. To obtain information on the symptoms and signs of patients at the time of their contact with the GP, questionnaires were sent to the attending GPs of all patients admitted during the study period, including GPs on call during out of office hours. The GP who had first contact with the patient, was asked to classify the clinical condition of the patient at the time of presentation on an adapted scale from the World Federation of Neurosurgical Societies (WFNS) SAH grading scale. In this five-point scale, grade one is ‘only headache with no mental alteration or focal neurological symptoms,’ grade two is ‘slightly decreased consciousness with no focal neurological symptoms,’ grade three is ‘slightly decreased consciousness with focal neurological symptoms present,’ grade four is ‘moderately decreased consciousness (with or without focal neurological symptoms)’ and grade five is ‘coma’ or ‘in need of resuscitation.’
Box I. Clinical features of subarachnoid haemorrhage (SAH).
Although not specific to SAH, the cardinal feature of SAH is a sudden, explosive headache, a so-called ‘thunderclap headache’. In half of the patients this headache develops instantaneously, in other patients the headache develops in seconds to even a few minutes (Citation10).
Other signs of SAH are focal neurological phenomena or a blurring of consciousness, accompanying the onset of headache. These signs occur in about 70% of patients with SAH, but are absent in the remaining 30% (Citation3,Citation5,Citation9). Furthermore, neither these additional phenomena nor the sudden onset discriminate SAH from innocuous causes of headache, let alone from other emergencies that present with thunderclap headache (such as cerebral venous sinus thrombosis, cervical artery dissection, pituitary apoplexy, or a colloid cyst of the third ventricle) (Citation10,Citation13).
Another sign of SAH is neck stiffness, which is common but develops only several hours after the onset of headache. Therefore, in the acute phase, the absence of neck stiffness does not rule out SAH.
Finally, subhyaloid haemorrhages, which can be seen at fundoscopy, occur in one out of every 5 to 10 SAH patients, but only seldom, if at all, in patients with isolated headache and no loss of consciousness at onset (Citation14).
Take-home messages
1. The main alarm signs of SAH are the acute onset and the severity of the headache.
2. The absence of other associated signs and symptoms in a patient with a sudden, severe headache should not cause hesitation in immediate referral of this patient to a hospital for additional diagnostic evaluation.
With this retrospective survey, we also retrieved further information on the date and time of preliminary contact with the GP, date and time of onset of symptoms, whether at first contact with the patient the GP knew that the headache started acutely, a patient history of headaches and the date and time of referral of the patient to a hospital. No definition of acute onset of a headache was given and the term was left free for interpretation by the GPs. An invitation to contact us in case of difficulties with answering the questions was written under the questionnaire and the GPs were actively contacted by telephone when no response was forthcoming, or when incomplete and ambiguous answers were provided.
In this study, we included patients with SAH who sought first medical contact with their GP because of isolated headache, i.e. headache without any other neurological or alarming signs (WFNS grade one).
Variables and statistical analysis
Outcome. The purpose of this study was to identify determinants for a delayed referral in patients with SAH presenting with isolated headache to the GP. Therefore, after including patients with SAH who presented with isolated headache to the GP, we distinguished patients with delayed referral from patients with immediate referral. Delayed referral was defined as referral by the GP more than two hours after the first medical contact or by no referral by the GP at all.
Determinants. In our sample we expected to find between 20 and 30 patients with a delayed referral. Therefore, we included only three potential determinants for delayed referral in our study: GP's unawareness of the sudden onset of the headache, a patient's history of headaches and a patient delay. Patient delay was defined as a nine or more hour interval between onset of SAH-related symptoms and the time of the patient visiting the GP.
Analysis. To determine the relationship between the determinants and a delayed referral we divided the proportion of the patients with the determinant who had a delayed referral by the proportion of the patients without the determinant who had a delayed referral and calculated corresponding 95% confidence intervals (CIs). We chose this method instead of odds ratios because we expected high proportions of delayed referral and these high proportions may lead to exaggerated odd ratios.
Results
During the eight month study period, 112 patients with SAH were admitted to our service; their baseline characteristics are given in . One patient had no GP. 47 of the 112 patients with SAH, had first contact with the GP. The GPs who had primary contact with the patients were all identified; the response rate was 100%. 16 patients presented to the GP with a decreased consciousness and/or neurological deficits (WFNS grade 2–5); they were immediately referred to the hospital.
Table I. Patients with subarachnoid haemorrhage (SAH) as present in a neurosurgical database (eight months data collection).
31 patients presented to the GP with isolated headache (WFNS grade 1) and were included in the analyses. The mean age of these 31 patients was 53 (range: 15–84) years; 9 patients were men. Referral was delayed in 18 of the 31 patients with SAH who presented with isolated headache to the GP.
A delay in referral occurred in all 10 patients in whom the GP was unaware, and in 8 of the 21 patients in whom the GP was aware of the acute onset of headache (risk ratio (RR) 2.6; 95% CI: 1.5–4.5). A history of headaches (RR: 1.8; 95% CI: 1.1–3.0) and a patient delay (RR: 2.1; 95% CI: 1.0–4.5) also increased the likelihood of delayed referral in patients with SAH who presented with isolated headache to the GP ().
Table II. Determinants of delayed versus immediately referred patients with subarachnoid haemorrhage (SAH) who presented with isolated headache to the GP (n = 31).
Discussion
Main results
Our study shows that unawareness of the sudden onset of the headache is an important determinant of delayed referral in patients with SAH who presented with isolated headache to the GP. None of the patients in whom the GP was unaware of the acuteness of headache onset were referred immediately. Other patients presenting with isolated headache to the GP, who are at risk of delayed referral, are those with a history of headaches and those who seek medical advice late after the onset of headache.
Our study also shows that GPs who are aware of the sudden onset of headache, defer referral in one third of patients with SAH who present with isolated headache, and in whom the diagnosis SAH is established in the hospital later. Apparently, knowledge of the acuteness of onset is essential but not always sufficient to raise a GP's suspicion of SAH.
Strengths and limitations
Patient selection. Within our database of SAH cases, we restricted the study population to patients who presented with isolated headache to the GP, because these patients are most at risk for misdiagnosis and these patients have the most to lose when re-bleeding occurs (Citation3,Citation5,Citation8,Citation9). The consequence of our patient selection is that our results will only apply to this subset of patients, being approximately 30% of patients with SAH.
Study design. We chose a retrospective study design for our study because of time-efficiency reasons: a hospital database of well-documented SAH cases was available and the complementary primary care data were relatively easy to collect retrospectively. Strength of our study is that we identified and questioned all GPs who had primary contact with the patients in our study. A prospective study design would require a high number of GPs, who would have to register virtually all patients presenting with headache during a long period of time. Furthermore, in a prospective design the mere knowledge of the study might influence the GP's behaviour. Therefore, it is our opinion that the retrospective design is appropriate, although this design may have caused a retrieval bias, in the sense that the speed of onset had to be retrieved from the GPs’ records in retrospect. It may be that GPs were aware of the acute onset of the headache, but did not record it. If this is true, the search for other determinants for lack of immediate referral in patients known to have had an acute onset headache becomes even more important than it is already. The date and time of onset of symptoms were also retrospectively retrieved from GP notes during the survey. The reliability of this information is therefore dependent on the recall of the patient and accuracy of history taking and reporting of the GP at that time.
Variable selection. Due to the sample size of our study, we were restricted to only three variables as potential determinants for a delayed referral. Other potential determinants for a delay in referral, such as a short duration of headache or good reaction to analgesics could, therefore, not be studied.
Interpretation
The three determinants we studied are probably not independent. Patients who seek advice after a delay may be less likely to tell (or to be asked about) the acuteness of onset of the headache than those patients who seek advice immediately.
We could not find other studies investigating reasons for delayed referral of patient with SAH in general practice. In our study, half of the patients with SAH who presented with isolated headache to the attending GP were not immediately referred to the hospital. Other studies, which focused on the percentage of delayed diagnosis within all admitted patients with SAH (regardless of the nature of first medical contact and the degree of neurological symptoms of the patient), found proportions of delayed diagnosis and referral between 12 to 51% (Citation2,Citation3,Citation5,Citation6). The high proportion of delayed referral in our study can probably be explained by our restriction to patients with SAH presenting with isolated headache and by our study setting of GP practices without immediate access to imaging facilities. SAH is less likely to be misdiagnosed in the emergency department where these facilities are at hand. In one study in the emergency department of 1507 patients with SAH, only 5.4% were misdiagnosed (Citation11).
Implications
Preventing delayed referral in patients with SAH who present with isolated headache, will only have a minor impact on overall outcome in patients with SAH (Citation8). However, these patients are relatively young and in good clinical condition, and if re-bleeding by early treatment is prevented, they may have many healthy years ahead of them (Citation6). Moreover, the impact of a poor outcome associated with a delayed referral is enormous for these patients and treating physicians.
Therefore, all patients presenting with headache to the GP should be asked about the nature of onset of their headache. Subsequently, all patients with an acute onset headache should be referred immediately to the hospital for additional diagnostic evaluation. This is in accordance with the Dutch GP guideline on headache (Citation12). The discomfort and cost of referral of the large proportion of patients with ‘benign thunderclap headache’ is outweighed by the gain of survival associated with prevention of re-bleeding by early treatment of the ruptured aneurysm (Citation6). This approach of referring all patients with sudden onset headache to a hospital for further diagnostic evaluation has been estimated to cost a thousand pound per quality-adjusted life year gained, in a retrospective study in the UK (Citation6).
Further research is needed on why patients known to have had an acute onset of headache sometimes are not referred. This could be done in a prospective survey in general practices where the determinants for (deferred) referral of patients presenting with sudden onset headache could be investigated, although such a survey may already change the proportion of referred patients.
Conclusion
This study shows that unawareness by the GP of the acute onset of headache, a history of headaches and late presentation increase the probability of delayed referral in patients with SAH who present with isolated headache to the GP.
Ethical standards
This study does not involve human subjects.
Funding
This research was run without any external funding.
Acknowledgements
The authors should like to thank all the participating general practitioners for their time and effort invested in this study. The authors also should like to thank Professor Algra for his help with the statistical analysis and Miss Collington and Mrs Wilson for their help in reviewing the text.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Nieuwkamp DJ, Setz LE, Algra A, Linn FH, de Rooij NK, Rinkel GJ. Changes in case fatality of aneurysmal subarachnoid haemorrhage over time, according to age, sex, and region: A meta-analysis. Lancet Neurol. 2009;8:635–42.
- Neil-Dwyer G, Lang D. ‘Brain attack’—aneurysmal subarachnoid haemorrhage: Death due to delayed diagnosis. J R Coll Physicians Lond. 1997;31:49–52.
- Mayer PL, Awad IA, Todor R, Harbaugh K, Varnavas G, Lansen TA, . Misdiagnosis of symptomatic cerebral aneurysm. Prevalence and correlation with outcome at four institutions. Stroke 1996;27:1558–63.
- Roos YB, de Haan RJ, Beenen LF, Groen RJ, Albrecht KW, Vermeulen M. Complications and outcome in patients with aneurysmal subarachnoid haemorrhage: a prospective hospital based cohort study in the Netherlands. J Neurol Neurosurg Psychiatry 2000;68:337–41.
- Kowalski RG, Claassen J, Kreiter KT, Bates JE, Ostapkovich ND, Connolly ES, . Initial misdiagnosis and outcome after subarachnoid hemorrhage. JAMA 2004;291:866–9.
- Tolias CM, Choksey MS. Will increased awareness among physicians of the significance of sudden agonizing headache affect the outcome of subarachnoid hemorrhage? Coventry and Warwickshire Study: Audit of subarachnoid hemorrhage (establishing historical controls), hypothesis, campaign layout, and cost estimation. Stroke 1996;27:807–12.
- Kernick D, Stapley S, Goadsby PJ, Hamilton W. What happens to new-onset headache presented to primary care? A case-cohort study using electronic primary care records. Cephalalgia 2008;28:1188–95.
- Linn FH, Wijdicks EF, van der Graaf Y, Weerdesteyn-van Vliet FA, Bartelds AI, van Gijn J. Prospective study of sentinel headache in aneurysmal subarachnoid haemorrhage. Lancet 1994;344:590–3.
- Fridriksson S, Hillman J, Landtblom AM, Boive J. Education of referring doctors about sudden onset headache in subarachnoid hemorrhage. A prospective study. Acta Neurol Scand. 2001;103:238–42.
- van Gijn J, Rinkel GJ. Subarachnoid haemorrhage: Diagnosis, causes and management. Brain 2001;124:249–78.
- Vermeulen MJ, Schull MJ. Missed diagnosis of subarachnoid hemorrhage in the emergency department. Stroke 2007;38: 1216–21.
- Knuistingh Neven A, Bartelink MEL, De Jongh TOH, Ongering JEP, Oosterhuis WW, Van der Weerd PCM, . NHG-standaard Hoofdpijn. Huisarts Wet. 2004;46:411–22.
- Ferrane E, Tassorelli C, Rossi P, Lisotto C, Nappi G. Focus on the management of thunderclap headache: From nosography to treatment. J Headache Pain 2011;12:251–8.
- Fountas KN, Kapsalaki EZ, Lee GP, Machinis TG, Grigorian AA, Robinson JS, . Terson hemorrhage in patients suffering aneurysmal subarachnoid hemorrhage: Predisposing factors and prognostic significance. J Neurosurg. 2008;109:439–44.