J O’Neill, S McLaggan, R Gibson*
Department Clinical Radiology, Royal Infirmary Edinburgh *Department Clinical Neuroradiology, Western General Hospital, Edinburgh
Correspondence to: John O’Neill, Radiology Department, St Josephs Hospital, 50 Charlton Avenue East, Hamilton, Ontario, Canada. Tel: 905 5216009 Fax: 905 521 Email: johnandal101@hotmail.com
SMJ 2005 50(4): 151-153
Abstract
Aim: Assess and correlate the clinical presentation, cranial CT and lumbar puncture (LP) findings in patients presenting with acute headache, clinically suspicious of subarachnoid haemorrhage, from the Accident and Emergency (A/E) department of a main teaching hospital. Method: We retrospectively reviewed over a 1-year period all patients referred for cranial CT from the A/E department, with clinically suspected nontraumatic SAH. Patients with a negative CT, with a diagnosis of SAH on LP, had their initial CT examination reassessed in a double blind review. Results: 116 consecutive patients were included in the study. Patients were divided into 3 groups post CT results: diagnostic of SAH (19), abnormal without evidence of SAH (16) and normal (81). The clinical management of the latter group was analysed with particular emphasis on the performance and results of lumbar puncture (LP). 81 patients, clinically suspicious of SAH in our study had a normal cranial CT and were eligible for LP. 49% had an LP, which was positive in 15% and contributed 24% to the total number diagnosed with SAH. LP was not performed in 51% of eligible patients. Conclusion: This study reviews the varying clinical presentations of SAH and the difficulty in clinically diagnosing SAH where headache is the only symptom. We review the wide differential diagnosis on cranial CT and on discharge of patients presenting with symptoms clinically suspicious of SAH. In addition this study identifies a high proportion of patients without detailed clinical notes and discharged without a diagnosis. The study reaffirms that a significant proportion of patients are diagnosed on LP and stress the importance of performing a LP in all cases with clinically suspected SAH with a negative CT. This latter point requires highlighting in the medical literature as over half of eligible patients did not have a LP performed despite this being accepted diagnostic protocol.
Introduction
Acute severe headache of recent onset is a relatively common presenting complaint to the A/E department.1,2,3 Although most headaches are benign in origin, a significant number have a serious underlying disease. Studies estimate 1.2% of patients presenting to A/E with a complaint of headache and up to 21% in those requiring admission because of headache have a serious underlying pathology.3,4,5 Subarachnoid haemorrhage (SAH) is one of the more serious pathologies and in the latter studies accounted for between 3% and 6% of cases respectively.
SAH has an incidence rate in the general population six cases per 100,000 person years.6 The consideration of SAH in the differential diagnosis of new onset severe headache is essential as it is related to a high morbidity and mortality. In a review of population-based studies case fatality had a weighted average of 51%, with patients surviving haemorrhage remaining independent in only one third of cases.7,8
Classically SAH is described as a sudden onset of severe, worst pain ever, headache. Two-thirds have associated neurological symptoms or clinical findings, which add considerable weight to the clinical diagnosis of SAH. These patients will proceed to a cranial CT examination. The difficulty arises in the third of patients presenting solely with headache. Only 50% describe it as instant, with the remainder coming on over a period of seconds to minutes.9
In explosive headache alone, only 10% will be secondary to SAH.10 Cranial CT is the first line investigation if other benign causes are not excluded clinically. Sensitivity of 98% within the first 12 hours progressively decreases thereafter.11,12
Patients with negative CT with the clinical suspicion of SAH require lumbar puncture prior to discharge.13,14 A delay of at least six and preferably 12hrs should exist between the spinal tap and the onset of headache.15 This is essential to allow diffusion of blood into the spinal subarachnoid space and sufficient lysis of the red blood cells within the CSF to form bilirubin and oxyhaemoglobin. These pigments give the CSF the characteristic yellow tinge after centrifugation (xanthochromia). In our study we retrospectively reviewed, over a one-year period, cranial CT, clinical features and LP results in all patients referred from a major A/E department with clinically suspected SAH.
Materials and methods
CT records were retrospectively reviewed over a one year period. All cases with acute headache, with a duration of less than 24 hours, and clinically suspicious of acute nontraumatic SAH, referred for cranial CT from the A/E department were included in the study. The cranial CT report, A/E clinical report and in-patient records, in those patients admitted, were reviewed. A data file was composed for each patient. Double blind review of the hard copy cranial CT studies in those patients initially reported as normal, and who subsequently had a positive LP, were reviewed by two independent neuroradiologists. Results CT record review revealed 127 patients referred from the A/E department who had a cranial CT performed for suspected SAH. 116 were eligible for inclusion in the study, the remaining 11 patients were excluded due to incomplete records. These 116 patients were divided into three groups depending on their cranial CT results:
a) diagnostic for SAH (19),
b) abnormal, but without evidence SAH (ABN) (16)
c) normal (NML) (81) (Table I).
The latter group was reviewed with respect to the performance and result of LP.
Clinical presentation, Table II, was of a classical headache for SAH in the majority. Classical headache were those described as sudden onset, worst pain ever headaches. This group had a sensitivity of 60%, specificity of 58% and a positive predictive value of 0.33 for SAH in those patients presenting with headache. Patients with meningism/ photophobia and collapse had sensitivities of 33% and 37.5% with PPV of 0.26 and 0.47 respectively.
The second group within headaches, described as “other”, included non-classical headache and those documented in the clinical notes as headache without detailed description. Patients presenting with collapse were included in the study only if SAH was a principal clinical diagnosis. Our review of clinical notes did in some cases demonstrate an inadequate description of the patients’ presentation and clinical examination.
Patient records were reviewed with respect to previous history. Thirteen percent of patients in group C had a history of migraine, with 2% each in group A and B. The clinical presentation in each case presented new features compared with previous episodes of migraine. There was no clinical documentation of sentinel headaches in any group. In addition no predisposing factors reaching statistical significance were identified. The mean ages were 55, 53 and 42 years for groups A, B and C respectively.
Thirty percent of patients had abnormalities detected on their CT examinations, which could account for the presenting clinical features (Table III). A total of 70%(81) had normal examinations and were eligible for LP. 49.4% (40) had a LP performed. Within this subset of patients results were unavailable for 13 patients. 19 patients were classified as “other” and included traumatic LP, unsuccessful LP (3), patient refused examination or self-discharged (4), and one patient who was on wafarin therapy (Table IV).
The time of headache onset to performance of lumbar puncture was poorly documented in the patients’ clinical notes. The six patients who had a positive LP results for xanthochromia, had a double blind review of their initial cranial CT studies on hardcopy, by two independent neuroradiologists, and confirmed that the studies were normal. All six patients positive for xanthochromia had cerebral angiography. Four patients had normal angiography and two had middle cranial artery aneurysms, one of which was coiled.
Seventeen patients had no final diagnosis on discharge with 22 cases described as headache of unknown origin or without evidence of SAH (Table V). The four described as “other” included one case of aortic dissection with extension into carotid arteries, one each of encephalitis, vasovagal collapse and acute headache secondary to anxiety. Review of the timing of CT requests demonstrated that 50% of CT examinations for suspected SAH were performed out of normal working hours, between 5pm and 8am.
Discussion
SAH is classically described as presenting with severe sudden onset, worst ever, occipital headache. In our study this presentation alone had a low sensitivity, specificity and PPV for SAH. The combinations however of classical headache with signs of meningism, nausea and vomiting, collapse or focal neurology significantly increases sensitivity. Collapse and classical headache, for example, has a sensitivity of 96% and a PPV of 0.37. All patients, however, presenting with such a combination of symptoms would normally receive a cranial CT examination. Some combinations of non-classical headache with, for example features of meningism, without features of raised intracranial pressure, would be clinically suspicious of meningitis and would have an LP and/or cranial CT. These groups of patients are within the clinical boundaries of patients requiring further investigation.
The difficulty arises when patients present solely with headache and deciding which patients require further examination with CT and/or LP. A detailed description of the presenting complaint and clinical history is essential. Sudden onset severe headache has a differential diagnosis of SAH, meningitis, first attack migraine, thunderclap headache, central venous sinus thrombosis, internal carotid artery dissection and those outlined above (Tables IV, V).
History and a detailed clinical examination will usually narrow the clinical diagnosis. A history of multiple previous episodes of similar headache with a normal clinical examination usually indicates an underlying benign process. The limitation to this hypothesis is that sentinel headaches do occur, and although not documented in our study group have been shown to occur in up to 39% of patients with SAH.15 It is however not possible on clinical grounds alone to separate this group with sentinel headaches from thunderclap headache.
Eighty-one patients, clinically suspicious of SAH in our study had a normal cranial CT and were eligible for LP. A total of 49.4% had a LP, which was positive in 15% and contributed 24% to the total number of patients diagnosed with SAH. This is a significant percentage and is an ominous finding considering that 50.6% of this group of patients did not receive a LP for a multitude of reasons as outlined (Table IV). In those cases, where LP was classified as “not performed”, the clinical notes gave a limited reason including review by a senior colleague who felt the examination was not required, or it simply stated that the headache had improved. Without a detailed clinical note it was not possible to assess in more detail why these patients did not undergo a LP.
Discharge diagnosis did not document a cause for the patients’ symptoms in 48% of the cases within group C. The social and economic impact of patients discharged without a diagnosis requires further investigation. Both of the latter deficiencies in record keeping have been noted in previous studies and we emphasise the need to keep accurate and detailed clinical notes.4
Cranial CT demonstrates subarachnoid blood in up to 97.5% in the first 24 hours and declines thereafter.12 Additional CT features include intracerebral in 20-40% and intraventricular and subdural blood in 15-35% and 5% respectively. Our study demonstrated higher secondary features with 73% and 31.5% of cases with ventricular and intracerebral extension respectively. Seventy-nine percent presented with hydrocephalus. Twenty-one percent had no secondary findings. Two patients had non-aneurysmal perimesencephalic haemorrhage. This normally constitutes 10% of SAH and two-thirds of those with normal angiograms and is associated with a significantly better prognosis.7
MRI, using fluid attenuated inversion recovery sequence (FLAIR) has been shown in some studies to be as sensitive as CT in the acute phase.16-19 Patients, however, are often restless without anaesthesia in the acute stage and added to the limited availability and cost of MRI with respect to CT, the later examination is preferred initially in our centre. We do not see any benefit from advocating MRI as a first line investigation in this group of patients presenting acutely as in our study.
In conclusion we outline the clinical presentation of patients to A/E with acute severe headache clinically suggestive of SAH and demonstrate the difficulties in diagnosing clinically SAH where headache is the only symptom. We emphasise the need for detailed clinical notes, in particular with the history of the presentation, timing of lumbar puncture and the clinical decisions taken thereafter. We review the differential diagnosis on cranial CT and on discharge and highlight the high proportion of patients discharged without a diagnosis.
The study reaffirms that a significant proportion of patients are diagnosed on LP and stresses the importance of performing a LP in all cases with clinically suspected SAH with a negative CT. This latter point requires highlighting in the medical literature as over half of eligible patients did not have a LP performed despite this being accepted diagnostic protocol.
REFERENCES
1 Craig JJ, Patterson VH. Headaches in the accident and emergency department. Br J Hospital Med1997; 57(5): 202-206
2 Barton CW. Evaluation and treatment of headache patients in the emergency department: a survey. Headache; 34(2): 91-94
3 Dhopesh V, Anwar R, Herring C. A retrospective assessment of emergency department patients with complaint of headache. Headache 1979; 19:37-42.
4 Stevenson RJ, Dutta D, Macwalter RS. The management of acute headache in adults in an acute admissions unit. Scot Med J 1998; 43:173-176
5 Frodden DI, Peatfield RC, Milsom PL. Beware the patient with a headache in the accident and emergency department. Arch Emergency Med 1989; 6:7- 12.
6 Linn FH, Rinkel GJ, Algra A, van Gijn J. Incidence of subarachnoid hemorrhage: role of region, year, and rate of computed tomography:a metaanalysis. Stroke 1996;27: 625-9.
7 van Gijn J, Rinkel GJ. Subarachnoid haemorrhage: diagnosis, causes and management. Brain 2001 Feb; 124(Pt 2): 249-78.
8 Hopp JW, Rinkel GJE, Algra A. Case fatality rates and functional outcome after subarachnoid haemorrhage systematic review. Stroke 1997; 28:660-4
9 Linn FH, Rinkel GJ, Algra A, van Gijn J. Headache characteristics in subarachnoid haemorrhage and benign thunderclap headache. J Neurol Neurosurg Psychiatry 1998; 65:791-3
10 Linn FH, Wijdicks EF, van der Graaf Y, Weerdesteyn-van Vliet FA, Bartelds AI, van Gijn J. Prospective study of sentinel headache in aneursymal subarachnoid haemorrhage. Lancet 1994; 344:590-3
11 van der Wee N, Rinkel GJ, Hasan D, Van Gijn J. Detection of subarachnoid haemorrhage on early CT: is lumbar puncture still needed after a negative scan? J Neurol Neurosurg Psychiatry 1995; 58:357-9
12 Morgenstern LB, Luna-Gonzales H, Huber JC, Wong SS, O Uthman M, Gurian JH, Castillo PR, Shaw SG et al. Worst headache and subarachnoid haemorrhage: prospective, modern computed tomography and spinal fluid analysis. Ann Emerg Med 1998 Sept; 32:3:297-304
13 Lledo A, Calandre L, Martinez-Menendez B, Perez-Sempere A, Portera-Sanchez A. Acute headache of recent onset and subarachoid haemorrhage: a prospective study. Headache 194; 34:172-174
14 Vermeulen M, van Gijn J. The diagnosis of subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 1990; 53:365-72
15 Leblanc R. The minor leak preceding subarachnoid haemorrhage. J Neurosurg 1987; 66:35-39
16 Noguchi K, Ogawa T, Inugami A, Toyoshima H, Sugawara S, Hatazawa J. Acute subarachnoid haemorrhage: MR imaging with fluid attenuated inversion recovery pulse sequences. Radiology 1995; 196:773-7
17 Noguchi K, Ogawa T, Seto H, Inugami A, Hadeishi H, Fujita H. Subacute and chronic subarachnoid haemorrhage: diagnosis with fluid attenuated inversion recovery MR imaging. Radiology 1997; 203:257-62.
18 de Paepe A, van Landegem W, de Keyser F, de Reuck J. Association of multiple intracranial aneurysms and collagen type 111 deficiency. Clin Neurol Neurosurg 1988; 90:53-60
19 Woodcock RJ, Short J, Do HM, Jensen ME, Kallmes DF. Imaging of acute subarachnoid hemorrhage with a fluid-attenuated inversion recovery sequence in an animal model: comparison with non-contrast enhanced CT. Am J Neuroradiol 2001; 22:1698-1703.
20 Wijdicks EF, Kerhoff H, Van Gijn J. Longterm follow-up of 71 patients with thunderclap headache mimicking subarachnoid haemorrhage. Lancet. 1988 Jul 9;2(8602):68-70.