The need for repeat angiography in subarachnoid haemorrhage

This study was designed to assess the necessity for a second angiogram study in patients in whom initial angiography after primary subarachnoid haemorrhage (SAH) was negative. During a 12-year period, 122 of 694 patients (17.5 %) had negative initial angiograms. CT, available for 98 patients, showed a preponderance of subarachnoid blood in the perimesencephalic cisterns in 50 of 73 patients (68.5 %) in whom blood was visible on CT. Angiography, repeated in 67 patients, revealed an aneurysm in 4 (6 %): 2 had an aneurysm of the anterior communicating artery, 1 of the posterior inferior cerebellar artery, and 1 of the P2 segment of the posterior cerebral artery. CT showed subarachnoid blood in the interpeduncular and ambient cisterns in this last case, and a preponderance of subarachnoid blood outside the perimesencephalic cisterns in the remaining 3 patients. Received: 14 March 1997 Accepted: 2 May 1997     

Current diagnostic approaches to subarachnoid haemorrhage

Over the past decade, significant advances have been made in the field of subarachnoid haemorrhage (SAH). Prompt diagnosis with high-resolution CT and intensive critical care support remain key aspects of good patient management. Early identification and definitive treatment of underlying ruptured aneurysms is generally advocated to reduce the risk of re-bleeding, a complication with high mortality and morbidity. Although intra-arterial digital subtraction angiography (DSA) is still considered the gold standard for sourcing aneurysms, CT angiography, especially with the evolution of multi-slice technology, is slowly gaining acceptance as a rapid, accessible and minimally invasive method which appears likely to replace DSA as first-line modality in the future. Furthermore, the advent of Guglielmi detachable coils and the ISAT trial have revolutionised the treatment of ruptured aneurysms, with a significant trend towards endovascular coiling away from operative clipping. Improvements in clinical experience, coiling technology and assistive devices now allow interventionalists to potentially treat the majority of aneurysms, including wide-necked or complex lesions. The uncertain long-term results of coiling, however, still fuel strong debate and controversy. This review summarises current diagnostic approaches to SAH from a radiological perspective, with an emphasis on aneurysmal SAH and an evidence-based approach to the role of imaging and interventional radiology in diagnosis, treatment and follow-up.     

The negative angiogram in subarachnoid haemorrhage

Our purpose was to review the incidence of negative cerebral panangiography in acute nontraumatic subarachnoid haemorrhage (SAH); to document the amount and distribution of subarachnoid blood on CT and determine its relationship to findings on repeat angiography; and to study the outcome of these patients from the time of presentation to hospital discharge. From 1983 to 1992, 295 patients underwent cerebral angiography for acute SAH at our institution. The CT, angiographic and MRI findings and clinical course of patients with initially negative angiograms were reviewed retrospectively. The overall incidence of negative cerebral panangiography was 31% (92/295). An aneurysm was disclosed on a second angiogram in 4 cases, and on a third angiogram in 1, giving a total false negative rate of 5%. In 55% of cases, only a small amount of SAH was present on CT. The distribution of the subarachnoid blood was nonspecific and resembled the pattern seen in aneurysmal SAH. Ninetyfour percent, of the patients presented in Hunt-Hess grades I and II. The complications of conservative treatment were few: a rebleed rate of 4%, delayed cerebral ischemia in 4%, cerebral infarcts in 8% and hydrocephalus requiring shunting in 3%. On discharge, 93% of patients had recovered completely and the others were left with moderate disability. There were two deaths related to massive rebleeding. Patients with perimesencephalic SAH (35%) fared particularly well; none developed complications during their hospital stay and repeat angiograms never revealed an underlying aneurysm. In such cases, further angiographic investigations do not seem warranted.     

Three-dimensional CT angiography of intracranial vasospasm following subarachnoid haemorrhage

We evaluated the usefulness of three-dimensional CT angiography (3D-CTA) in the diagnosis of intracranial vasospasm following subarachnoid haemorrhage (SAH) in 13 patients suspected of having vasospasm on clinical grounds. The intracranial vessels were clearly shown by 3D-CTA in 12 patients. 3D-CTA revealed spasm in the vessels of nine patients. Catheter angiography performed in seven of these patients immediately after 3D-CTA confirmed vasospasm. A low-attenuation area was seen on CT in the other two patients, representing an ischaemic lesion due to the spasm. In nine patients, a second 3D-CTA was performed using the same technique 1 week after the first, showing no vasospasm. Initial 3D-CTA revealed no vasospasm change in three patients. Following 3D-CTA, one of these had conventional angiography, which also demonstrated no spasm.     

CT appearances of haematomas in the corpus callosum in patients with subarachnoid haemorrhage

Corpus callosum heamatoma is a rare feature in subarachnoid haemorrhage (SAH), which may result from aneurysms of the anterior communicating artery (ACoA) or pericallosal artery (PCA). In 348 patients with aneurysmal SAH, bleeding from ACoA aneurysms in 88 cases produced no abnormality on CT in 7. Blood in the cistern of the lamina terminalis was the most frequent abnormality (76/88); haematomas of the septum pellucidum, confined to patients with ACoA aneurysms, were seen in 26 (30%). Rupture of PCA aneurysms in 12 patients gave rise to blood in the pericallosal cistern, anterior interhemispheric fissure and cistern of the lamina terminalis in 11. There was no blood in the septum pellucidum or the ventricular system in any case, but haematomas in the corpus callosum occurred in 8 (67%). In all of these, blood extended into the anterodorsal aspect of the callosum and spread posteriorly along its dorsal border. An identical, supracallosal pattern was seen in 2 patients (2.5%) with ACoA aneurysms, in whom haemorrhage was more extensive, with a large frontal lobe haematoma extending up from the cistern of the lamina terminalis in 1 and a haematoma of the septum pellucidum, with intraventricular extension in the other. In 8 patients (9%) with ACoA aneurysms a corpus callosum haematoma appeared to result from passage of blood up through the cistern of the lamina terminalis into the septum pellucidum and thence into the ventral aspect of the anterior corpus callosum; blood was present within the cistern, the septum and the ventricles.     

The source of haemorrhage in traumatic basal subarachnoid haemorrhage

Traumatic basal subarachnoid haemorrhage (TBSH) following trauma to the head, face or neck is well-established as a cause of death; however it remains a heavily disputed topic as the site of vascular injury is difficult to identify. Whilst many regions within the vasculature of the head and neck have been proposed as more susceptible to rupture, the vertebral artery remains the focal point of many investigations. We present a retrospective case review of TBSH in our forensic centre at Forensic and Scientific Services in Brisbane, Australia, from 2003 to 2011. Thirteen cases of TBSH were found, one case excluded due to vasculopathy. All decedents were male, the majority of which were involved in an altercation receiving blows to the head, face, or neck and were unconscious at the scene. All victims were under the influence of alcohol, drugs, or a combination thereof. External examination revealed injuries to the head, face, and neck in all cases. Various combinations of further examination techniques were used during the post-mortem examination including brain and/or cervical spine retention, CT imaging, and angiography. Vascular injury was identified in eight of the twelve cases, all of which occurred intracranially, with seven involving the vertebral artery. Histology was most reliable in identifying the rupture site and angiography failed to reveal a rupture site. The added benefits of histology over angiography are the ability to identify the microscopic architecture of the tear and to diagnose vasculopathy that may have rendered the individual more susceptible to TBSH.     

MRI in acute subarachnoid haemorrhage; findings with a standardised stroke protocol

There is doubt as to whether acute haemorrhage is visible on MRI. We carried out MRI within 6 h of symptom onset on five patients with minor (low Hunt and Hess grades 1 or 2) subarachnoid haemorrhage (SAH) diagnosed by CT to search for any specific pattern. We used our standard stroke MRI protocol, including multiecho proton density (PD)- and T2-weighted images, echoplanar (EPI) diffusion- (DWI) and perfusion- (PWI) weighted imaging, and MRA. In all cases SAH was clearly visible on PD-weighted images with a short TE. In four patients it caused a low-signal rim on the T2*-weighted source images of PWI, and DWI revealed high signal in SAH. In the fifth patient SAH was perimesencephalic; susceptibility effects from the skull base made it impossible to detect SAH on EPI DWI and T2*-weighted images. Perfusion maps were normal in all cases. MRA and conventional angiography revealed an aneurysm in only one patient. Stroke MRI within 6 h of SAH thus shows a characteristic pattern.     

Prediction of delayed neurological deficit after subarachnoid haemorrhage: a CT blood load and Doppler velocity approach

The predictive value of cranial computed tomography (CT) blood load and serial transcranial Doppler sonography for the development of delayed ischaemic neurological deficit was assessed in 121 patients following subarachnoid haemorrhage. Of the 121 patients, 81 (67%) had thick layers of blood or haematoma, including intraventricular bleeding. The proportion of patients who developed delayed deficit was higher with increasing amounts of subarachnoid blood on the admission CT (51% of 53 cases in Fisher grade 3; 35% of 33 cases in grade 2; 28% of 7 cases in grade 1,P<0.01). Doppler velocities obtained from readings at least every 2 days following admission were higher in patients with delayed neurological deficit (peak velocity for grade 3 patients 176±6 cm/s (mean±SE), versus grade 2: 164±7 cm/s; grade 4 149±9, bothP=0.04, Mann-Whitney). Peak velocity and maximal 24-h rise tended to be higher within different CT grades in patients with a deficit than in those without; this difference was significant for grade 3 patients (P<0.01). We conclude that a combined approach with CT and Doppler sonography provides greater predictive value for the development of delayed ischaemic neurological deficit than either test considered independently. The value of Doppler sonography may be greatest for patients with Fisher grade 3 blood, in whom the risk of delayed ischaemia is greatest.     

Extravasation of contrast-enhanced blood into the subarachnoid space during computed tomography

Summary Four cases are described in which there was extravasation of contrast medium into the subarachnoid space at the time of enhanced CT. Three were being investigated for a previous subarachnoid haemorrhage which was due to an aneurysm in two and an angioma in one. The other patient had suffered a severe head injury.     

Repeat intra-arterial papaverine for recurrent cerebral vasospasm after subarachnoid haemorrhage

We assessed the prevalence of recurrent vasospasm following failure of intra-arterial papaverine and the efficacy of repeat intra-arterial infusions of papaverine for control of recurrent vasospasm. Of 24 patients treated with intra-arterial papaverine for vasospasm following aneurysm surgery, 12 did not improve clinically after the initial treatment; 9 received second or third infusions on consecutive days; 6 received only a second infusion; and 3 received a third. Superselective infusion into the intracranial arteries was performed in all nine cases. Despite angiographic improvement after the initial or second infusions, all nine patients showed varying degrees of recurrent vasospasm at the time of the second or third treatment. Within 24 h of a second infusion, three of the six patients had significant clinical improvement, and one of these showed marked improvement soon after a third infusion. Our preliminary results suggest that repeat papaverine infusion may be a way of controlling recurrent or recalcitrant vasospasm. Received: 14 October 1996 Accepted: 16 December 1996     

Magnetic resonance angiography in the management of aneurysmal subarachnoid haemorrhage: a study of 51 cases

We report preliminary experience with the use of magnetic resonance angiography (MRA) in the assessment of intracranial aneurysms. The aims of the study were to confirm that aneurysms could be detected and to assess the quality of the images with a view to early surgery if clinically indicated. We studied 51 patients with recent spontaneous subarachnoid haemorrhage with MRA, using a three-dimensional time-of-flight technique. The results of MRA were subsequently confirmed on intra-arterial digital subtraction angiography (IA-DSA) for anatomical correlation, and were considered satisfactory in 38 patients (74.5 %). In 20 patients early surgical obliteration of their aneurysm was possible on the MRA results without recourse to IA-DSA. A total of 37 aneurysms were identified in 32 patients, while the remaining 6 patients did not have an aneurysm. MRA was misinterpreted in 4 patients (7.9 %), and in 9 other patients (17.6 %) MRA was unsatisfactory due to movement artefacts. Received: 1 September 1994 Accepted: 12 May 1995     

Intraventricular blood without parenchymal clot following spontaneous subarachnoid haemorrhage

Summary The position of an intracerebral haematoma shown on computerised tomography (CT) of patients who have suffered a recent spontaneous subarachnoid haemorrhage is a powerful indicator of the source of the bleed. In the absence of such a parenchymal clot, does the distribution of intraventricular blood have similar predictive value? The scans of 22 patients showing intraventricular blood without parenchymal clot were compared with subsequent angiographic appearances. The distribution of intraventricular blood does not appear to correlate with either the clinical status of the patient, or the nature, site or size of the bleeding lesion.     

Postoperative assessment of cerebral blood flow in subarachnoid haemorrhage by means of99mTc-HMPAO tomography

Regional hypoperfusion is a very frequent complication of subarachnoid haemorrhage (SAH), being related to vasospasm in the majority of cases. Twenty-six patients who were admitted for SAH underwent follow-up with technetium-99m hexamethylpropylene amine oxime single photon emission tomography (SPECT) 3 and 8 days after surgery. Fifteen patients of these had one more examination 15 days after surgery. The degree of hypoperfusion was quantified using an index of asymmetry which allow the comparison of two symmetrical regions of interest (ROIs) on the transaxial slice which presented the greatest perfusion defect. Comparison of CT data, transcranial Doppler data and clinical signs with the perfusion as quantified by^99mTc-HMPAO SPECT indicates that a difference in counts of less than 10% between the two symmetrical ROIs is of no diagnostic value. Follow-up of the brain perfusion clearly shows that the most pronounced hypoperfusion was observed just after surgery, with progressive normalization at 8 and 15 days after surgery.^99mTc-HMPAO SPECT performed 8 days after surgery allows prediction of the clinical outcome. For these reasons,^99mTc-HMPAO SPECT, which is the only method for follow-up of cerebral perfusion in routine clinical practice, should be the first examination to be performed after surgery in patients with SAH.     

Perfusion-weighted MRI to evaluate cerebral autoregulation in aneurysmal subarachnoid haemorrhage

Introduction  The aim of this study was to evaluate autoregulatory mechanisms in different vascular territories within the first week after aneurysmal subarachnoid haemorrhage (SAH) by perfusion-weighted magnetic resonance imaging (PW-MRI). For this purpose, regional cerebral blood flow and volume (rCVF and rCBV) were measured in relation to different degrees of angiographically visible cerebral vasospasm (CVS). Materials and methods  In 51 SAH patients, PW-MRI and digital subtraction angiography were performed about 5 days after onset of SAH. Regional CBF and rCBV were analysed in the territories of the anterior cerebral artery (ACA), the middle cerebral artery (MCA) and the basal ganglia of each hemisphere in relationship to the degree of CVS in the particular territory. Correlations between rCBF, rCBV and CVS were analysed. Results  CVS was found in 22 out of 51 patients in at least one territory. In all territories, rCBV decreased with increasing degree of CVS, correlated with a decrease of rCBF. In the ACA territories, SAH patients with severe CVS had significantly lower rCBF compared to healthy subjects and to SAH patients without CVS. In the basal ganglia, rCBF and rCBV of the control group were significantly higher compared to the patients without and with moderate vasospasms. Conclusion  PW-MRI showed simultaneous decrease of rCBF and rCBV in patients with SAH. The fact that rCBV did not increase in territories with CVS to maintain rCBF reveals dysfunctional vascular autoregulation. Vasospasms in the microvasculature are most evident in the basal ganglia, showing decreased rCBV and rCBF even in SAH patients without CVS.     

Intracranial angiolipoma as cause of subarachnoid haemorrhage. Case report and review of the literature

A 33-year-old female with a longstanding history of seizures was admitted to our hospital with subarachnoid haemorrhage (SAH). Computed tomography (CT), magnetic resonance imaging (MRI) and digital subtraction angiography (DSA) depicted a vascular fat-containing lesion overlying a right frontal cortical polymicrogyria. The diagnosis of angiolipoma was established. Conservatory management was undertaken with full recovery. She has been followed for 5 years since, with neither re-bleeding nor morphologic change of the lesion. This is a rare intracranial lesion, with only 11 intracranial angiolipomas published in the literature, and is the first case reported which is associated with SAH caused by this lesion.     

New angiographic measurement tool for analysis of small cerebral vessels: application to a subarachnoid haemorrhage model in the rat

Introduction Exact quantification of vasospasm by angiography is known to be difficult especially in small vessels. The purpose of the study was to develop a new method for computerized analysis of small arteries and to demonstrate feasibility on cerebral angiographies of rats acquired on a clinical angiography unit. Methods A new software tool analysing grey values and subtracting background noise was validated on a vessel model. It was tested in practice in animals with subarachnoid haemorrhage (SAH). A total of 28 rats were divided into four groups: SAH untreated, SAH treated with local calcium antagonist, SAH treated with placebo, and sham-operated. The diameters of segments of the internal carotid, caudal cerebral, middle cerebral, rostral cerebral and the stapedial arteries were measured and compared to direct measurements of the diameters on magnified images. Results There was a direct correlation between the cross-sectional area of vessels measured in a phantom and the measurements acquired using the new image analysis method. The spread of repeated measurements with the new software was small compared to the spread of direct measurements of vessel diameters on magnified images. Application of the measurement tool to experimental SAH in rats showed a statistically significant reduction of vasospasm in the SAH groups treated with nimodipine-releasing pellets in comparison to all the other groups combined. Conclusion The presented computerized method for analysis of small intracranial vessels is a new method allowing precise relative measurements. Nimodipine-releasing subarachnoidal pellets reduce vasospasm, but further testing with larger numbers is necessary. The tool can be applied to human angiography without modification and offers the promise of substantial progress in the diagnosis of vasospasm after SAH.     

Pseudotumour of the corpus callosum due to subarachnoid haemorrhage from pericallosal aneurysm

Summary Two patients were scanned 2–5 weeks following onset of subarachnoid haemorrhage. A similar CT appearance was seen in each case. A ring enhancing mass lesion was present within the corpus callosum, suggesting an underlying tumour. However subsequent angiogrphy showed that each case was in fact due to rupture of a pericallosal artery aneurysm. The enhancing membrane of the haematoma gave the false impression of a corpus callosum tumour.     

Role of 123I-IMP SPET in the early diagnosis of borderline chronic hydrocephalus after aneurysmal subarachnoid haemorrhage

Chronic hydrocephalus after aneurysmal subarachnoid haemorrhage (SAH) is easily diagnosed in most cases. However, the diagnosis is sometimes difficult in borderline cases, in which (a) pathognomonic clinical deterioration due to hydrocephalus is masked by the neurological deficits caused in the acute stage of SAH and (b) ventricular enlargement is not so marked on CT scan. The purpose of this study was to investigate whether or not iodine-123 labelled N-isopropyl-p-iodoamphetamine (¹²³I-IMP) single-photon emission tomography (SPET) is of value for the early diagnosis of borderline chronic hydrocephalus after SAH. Fifteen patients who met the criteria of borderline chronic hydrocephalus were selected for the study, and underwent a shunt operation. The patients were divided into a shunt-effective group and a shunt-ineffective group according to neurological improvement after the shunt operation. ¹²³I-IMP SPET was performed in the acute stage of SAH, within 1 week before the shunt operation, and 2 weeks after the shunt operation. Regional cerebral blood flow was estimated by the ¹²³I-IMP autoradiographic method. Pre-shunting periventricular low-perfusion areas showed statistically significant differences between the two groups (P<0.05). In the shunt-effective group, periventricular low-perfusion areas on pre-shunting SPET were significantly enlarged compared with those in the acute stage of SAH (P<0.05), and they were significantly reduced after the shunt operation (P<0.05). In the shunt-ineffective group, periventricular low-perfusion areas showed no significant changes during the course. These results suggest that periventricular low-perfusion areas enlarge in the early stage of chronic hydrocephalus after SAH, and that ¹²³I-IMP SPET can be used for both the early diagnosis of borderline chronic hydrocephalus after SAH and the prediction of shunt effectiveness. Received 21 October and in revised form 27 December 1999     

Spontaneous dissection of the extracranial vertebral artery with spinal subarachnoid haemorrhage in a patient with Behçet's disease

A 40-year-old man with known definite Behçet's disease (BD) was admitted with confusional state which had started 4 days before admission with an acute headache and vomiting. Neurological examination revealed confusion, stiff neck, right facial weakness, left hemiparesis, dysartria and truncal ataxia. CSF was haemorrhagic and xanthochromic. Cranial CT scans were negative, but MRI showed a right pontine hyperintense lesion on T2-weighted images. Bilateral carotid angiograms were normal. Right vertebral angiogram showed findings consistent with a dissection at the V₂ segment of the artery. At the level of the fifth cervical vertebra, a radiculomedullary branch of the vertebral artery with an ancurysmal dilatation in its intradural portion was notable. This case shows that, in BD, aneurysn formation can also occur in a spinal artery and spontaneous vertebral artery dissection can be seen.     

The time course of aneurysmal haemorrhage on computed tomograms

Summary We performed serial CT scans in a prospective series of 100 patients with a ruptured aneurysm who were first scanned within 2 days of the haemorrhage. In all patients the early CT scan showed evidence of extravasated blood. In 96 patients the source of bleeding was clearly at the base of the brain, and 32 of these had a haematoma. We estimated from the results of 139 repeat scans that the probability of recognizing an aneurysmal haemorrhage on CT is 85% after 5 days, 50% after 1 week, 30% after 2 weeks (mostly patients with haematomas), and almost nil after 3 weeks.