Journal of Cerebrovascular Sciences

REVIEW ARTICLE
Year
: 2021  |  Volume : 9  |  Issue : 1  |  Page : 16--21

Role of angiography in primary capsulo-ganglionic haemorrhage: A review


Saurabh Beedkar, Ajay Hegde, Girish R Menon 
 Department of Neurosurgery, Kasturba Medical College, Manipal Academy of Higher Education, Manipal, Karnataka, India

Correspondence Address:
Prof. Girish R Menon
Department of Neurosurgery, Kasturba Medical College, Manipal Academy of Higher Education, Manipal, Karnataka
India

Abstract

Spontaneous intracerebral haemorrhage (SICH) is the second most common cause of stroke, accounts in contrast to ischaemic stroke and carries a high risk of morbidity and mortality. SICH refers to brain haemorrhage in the absence of trauma and is often associated with hypertension. It is a conventional belief that hypertension is the aetiology in most of the haemorrhage in the basal ganglia region, and angiography is rarely performed in such patients. This protocol carries a risk of missing potentially curable underlying vascular conditions such as arteriovenous malformations and aneurysms, especially in younger patients without a history of hypertension. Performing an angiogram for all patients however is an overkill and unnecessary waste of resources. The role of cerebral angiography to distinguish a primary SICH from secondary SICH in all patients with SICH is thus controversial. Hence, it is the choice of the type of angiogram between digital subtraction angiogram Digital Subtraction Angiography (DSA) and computerised tomography angiogram Computerised Tomography Angiography (CTA). This article attempts to review the current diagnostic angioimaging guidelines for intracerebral haemorrhage with an aim to evolve a management protocol.



How to cite this article:
Beedkar S, Hegde A, Menon GR. Role of angiography in primary capsulo-ganglionic haemorrhage: A review.J Cerebrovasc Sci 2021;9:16-21


How to cite this URL:
Beedkar S, Hegde A, Menon GR. Role of angiography in primary capsulo-ganglionic haemorrhage: A review. J Cerebrovasc Sci [serial online] 2021 [cited 2021 Oct 22 ];9:16-21
Available from: http://www.jcvs.com/text.asp?2021/9/1/16/324812


Full Text



 Introduction



Spontaneous intracerebral haemorrhage (SICH) is the second most common cause of stroke and accounts for 7.5%–30% of all strokes.[1],[2] SICH refers to brain haemorrhage in the absence of trauma and can be either primary or secondary. Primary intracerebral haemorrhage (ICH) is more common in the elderly and is often associated with hypertension or cerebral amyloid angiopathy (CAA).[2] Secondary ICH refers to cerebral haemorrhages secondary to other underlying conditions such as arteriovenous malformations (AVM), aneurysms, coagulation abnormalities, tumours bleeds and haemorrhagic infarcts.[3] In contrast to ischaemic stroke, SICH carries a high risk of morbidity and mortality and leaves approximately half the patients dead within the first 48 hours.[2] Prompt diagnosis and timely intervention is mandatory to help in minimising deficits and improving long-term outcome.

The role of cerebral angiography to distinguish a primary SICH from secondary SICH in all patients with SICH is controversial. Hence, it is the choice of the type of angiogram between digital subtraction angiogram (DSA) and computerised tomography angiogram (CTA). Many believe that a definite history of hypertension, especially in specific ICH locations such as the thalamus, basal ganglia, pons and cerebellum could reliably exclude underlying secondary lesions and a further need for imaging.[4] Angiogram was thus recommended only for younger individuals with non-typical ICH locations without hypertension and those with primary intraventricular haemorrhage (IVH). Proponents of angiogram, however, claim that patients with typical hypertensive haemorrhage locations carry a 2%–4% prevalence of secondary vascular lesions which needs to be adequately addressed.[5],[6] Risk of rebleed with associated clot expansion, a potential threat in all patients with primary hypertensive SICHs is more common in secondary SICH. A second argument in favour of early angiographic evaluation in spontaneous hamorrhage is that it ensures early diagnosis of the underlying cause and thereby prompt treatment to improve outcome or prevent re-haemorrhage.[7] This article attempts to review the current diagnostic angioimaging guidelines for ICH with an aim to evolve a management protocol.

 Is an Angiogram Indicated Following Spontaneous Intracerebral Haemorrhage?



The most common location for a hypertensive ICH is the lateral basal ganglionic capsular region, referred to as putaminal haemorrhages followed by thalamus, cerebellum, brain stem and the cerebral lobes. Aetiology is unknown but believed to be due to degenerative changes of the lateral lenticulostriate arteries due to sustained hypertension and increasing age.[8] Angiographic evaluation is therefore not done in adult hypertensive patients with spontaneous putaminal bleeds. However, a structural vascular abnormality such as an AVM, aneurysm, and moyamoya disease cannot be excluded in putaminal haemorrhages. Although rupture of an aneurysm commonly presents with subarachnoid haemorrhage, few aneurysmal bleeds can result in parenchymal ICH which can be mistaken for a hypertensive haemorrhage.[8] Aneurysmal bleeds from internal carotid artery bifurcation aneurysms, middle cerebral artery lenticulostriate aneurysms and few middle cerebral artery bifurcation aneurysms can mimic a putaminal haemorrhage.[9],[10] Similarly bleeds from striate AVMs, involving the lenticular nucleus and external capsule region, supplied by the lateral lenticulostriate artery, anterior choroidal artery and insular branches of the MCA, can mimic a capsuloganglionic haemorrhage.[11]

Moyamoya disease is another rare but important cause of deep seated haemorrhages.[12] Park et al. report a rough distribution of bleed sites in moyamoya disease as follows: 40% basal ganglia haemorrhages, 30% intraventricular haemorrhages, 15% thalamic haemorrhages with intraventricular extension and 5% subcortical haemorrhages.[10]

Decisions about angiogram in SICH are generally guided by three factors: patient age, ICH location and the presence of systemic arterial hypertension.[5],[12] Generally, cerebral angiography is not recommended in hypertensive patients older than 45 years and having a thalamic, putaminal or posterior fossa ICH.[4],[6] Since the association between high blood pressure and deep-seated ICH is not conclusive,[13],[14] this strategy carries a risk of missing an underlying aneurysm, AVM, dural arteriovenous fistula or intracranial venous thrombosis if angiographic imaging is not performed. It is important to detect these underlying lesions for appropriate treatment and also to prevent re-haemorrhage. Furthermore, a delayed magnetic resonance imaging may help to identify an underlying tumour or cavernous malformation which is occult and undetectable in acute imaging. It is reasonable to avoid angiogram in hypertensive elderly patients with other comorbidities who are unlikely candidates for any major intervention. However, in all other patients with SICH, some form on angioimaging at some point in the course of illness seems a worthwhile option depending on the patient's clinical condition. Proponents for angiogram argue that decision about angiogram should not be guided only by its potential of a positive yield but should be based on patient's clinical condition and its possible impact on treatment and outcome.[15]

 Incidence of Vascular Aetiologies in Intracerebral Haemorrhage



Approximately half of the ICH cases originate in the basal ganglia (Globus pallidus, caudate, putamen, internal/external capsule, thalamus or insula), a third are hemispherical; and a sixth, in the posterior fossa (brain stem or cerebellum). It is associated with IVH in around 40% of the cases.[3]

It is a conventional belief that hypertension is the aetiology in most of the haemorrhages in the basal ganglia region.[16],[17],[18] Elderly patients are often considered unlikely to harbour secondary vascular causes and are seldom investigated. Outcome is generally poor in elderly patients and the need for aggressive investigation and treatment of the underlying lesion is questionable. Few authors have attempted to challenge this dogma, but all such studies are limited by small sample sizes and/or selection bias. The first instance of a possible underlying vascular aetiology in SICH was reported by McCormick when he found a secondary cause of ICH in 36% of a pathologic series of 144 brains.[19] Typical cases of ICH in the setting of hypertension and older patients are much less likely to have a post-mortem examination and the actual figures can be higher.

The reported incidence of a vascular lesion in a basal ganglia haemorrhages are as low as 4% but can vary from 0% to 53%.[4],[6],[10],[13],[20],[21],[22] In a prospective study evaluating the yield of angiography in SICH in 82 patients, Halpin et al. detected underlying vascular lesions in 51% of all patients and 13% of hypertensive patients. Basal ganglia bleeds picked up 31% of the lesions, while posterior fossa demonstrated 18% (AVM in 23% and aneurysm in 8%).[13] A subset of patients with ICH without suspicious features on noncontrast computed tomography (NCCT) underwent delayed angiography, revealing a lesion in 24% of patients. Later, in a retrospective SICH study, Toffol et al. reported an almost similar incidence and observed that DSA detected lesions I 44% (22/50) of non-hypertensive patients compared to 12.7% (6/47) in hypertensive patients (6/47).[4] Griffith et al., in their prospective study of 100 patients with ICH, reported angiographic abnormalities in 58% of normotensive patients and 25% of hypertensive patients.[20] They recommend to maintain a suspicion of vascular aetiology in all cases since hypertension and vascular malformations can coexist in older patients without a clear relation to clinical and NCCT findings.

Zhu et al. detected lesions in 34% of patients in a cohort of 206 patients.[6] Angiographic abnormalities were higher in patients younger than 45 years (50%) compared to those older (18%). Positivity was 45% in normotensive individuals compared to 9% in hypertensives. In patients with lobar haemorrhage, 49% had positive findings on DSA and 19% of these patients were hypertensives. Amongst older normotensive patients with basal ganglia bleeds 7% had an underlying vascular lesion. Thirteen per cent of positive DSA patients had basal ganglia haemorrhages. This study, however, had ×2.5 more non-hypertensives and it is possible that the inclusion of fewer hypertensive patients underestimated the incidence of secondary lesions. A more recent study similarly showed a lower DSA yield of 39% in older patients compared to 55% in patients younger than 45 years.[5]

Almandoz et al. retrospectively studied the role of CTA in 623 patients with ICH. Only 15% (91/623) of patients had a positive CTA with a detectable vascular lesion.[23] Notably, 11% (70/623) were older than 45 years of age and had pre-existing hypertension. The incidence was very low at 4% (11/273) in patients older than 71 years (11/273). Overall, the highest CTA yield was 34% (210/623), in female non-hypertensive patients younger than 45 years with lobar or infratentorial haemorrhage or IVH.

Studies have attempted to compare the efficacy of DSA with CTA. A prospective study of lobar ICH in 78 patients compared the diagnostic yield of CTA with DSA. Secondary lesions were detected in 28% of patients and both CTA and DSA were found to have equal efficacy in diagnosing vascular lesions.[24] Park et al. in contrast report a DSA yield for putaminal haemorrhage of 15%, which was more than that inferred from CTA or magnetic resonance angiography (MRA).[10]

 Choice of Angioimaging



A plain computed tomography (CT) scan has low sensitivity and may not pick up secondary causes in patients with preexisting hypertension.[5],[6],[13] Few studies have systematically examined the performance of NCCT in detecting secondary lesions. In one prospective study, patients with SICH were classified according to the presence or absence of suspicious features on NCCT. However, NCCT alone demonstrated a poor sensitivity and specificity of 77% and 84%, respectively, relative to DSA.[13] Delgado et al. in a retrospective analysis categorised 623 patients with ICH into low, intermediate and high probability of vascular lesions on the basis of NCCT characteristics.[21] More than two-thirds of patients were assigned to an intermediate-risk category as they could not be confidently classified with NCCT. Only 3% of patients demonstrated features of an underlying suspicious lesion on NCCT. In a follow-up study, if intermediate NCCT appearances were considered together with low-risk patients, the sensitivity and specificity for detecting lesions were 20% and 99%, respectively. If intermediate- and high-risk groups were considered together, they were 95% and 35%, respectively. A scoring system to stratify the risk for underlying lesions with clinical features of NCCT has shown a positive correlation between higher score and secondary lesions.[21],[23] However, low prevalence of suspicious features on NCCT remains a significant limitation of such scores.

There are no clear guidelines to choose between DSA, CTA and MRA. Digital subtraction angiography (DSA) remains the cornerstone for the detection of secondary lesions.[25] DSA is costly, not readily available, requires expertise, has potential complications and is thus not an ideal screening test. Moreover, the routine application of catheter angiography is limited by age-related vascular disease. CTA and MRA, although quick and valuable for diagnosing vascular abnormalities,[26],[27] may not detect subtle vascular lesions. Computerised tomography angiography (CTA) with its wide availability, ease of doing and an acceptable track record in subarachnoid haemorrhage[3],[28] is increasingly being considered as a useful non-invasive screening technique in acute stroke and trauma.[3],[24] Few authors believe that the diagnostic yield of CTA and MRA for detecting vascular abnormalities in patients with SICH is not inferior to that of DSA and recommend its use as a part of routine imaging in the initial diagnostic work-up of patients with acute SICH.[3],[25],[29] Others, however, believe that until further studies of the diagnostic accuracy of non-invasive investigations such as CT angiography or MRA have been performed, these techniques cannot replace DSA for the investigation of ICH.[15]

One major concern with both, DSA and CTA is the use of contrast. Majority of the patients with SICH may have some form of renal impairment making the use of contrast risky. Contrast-induced nephropathy (CIN) is the iatrogenic renal injury in susceptible individuals following intravascular administration of radio-opaque contrast media. CIN is defined as a 25% relative increase, or a 0.5 mg/dL absolute increase, in serum creatinine levels within 72 h of contrast exposure, without an alternative explanation.[30] The potential for deleterious effects on renal function applies equally to CTA and DSA. However, the risk of acute renal impairment following the administration of contrast for CT angiography has often been overestimated by extrapolating data from cardiology patients who underwent conventional angiography.

Thus, radiologists are often overcautious and do not perform contrast studies without baseline serum creatinine levels. However, several retrospective and prospective reviews of patients undergoing CTA for acute stroke have reported a low incidence of 2%–5% of CIN.[31],[32] The most effective prophylactic intervention is adequate hydration of the patient with intravenous or oral fluids before and after contrast administration in addition to withholding nephrotoxic medications such as nonsteroidal anti-inflammatory drugs, furosemide and metformin (prevent lactic acidosis).[33]

 Current Guidelines



The American Heart Association/American Stroke Association (AHA/ASA) guidelines indicate that either NCCT or MRI may be used in the initial diagnostic evaluation of ICH.[2],[25],[34] In most centres, however, NCCT remains the technique of choice due to its availability, patient tolerability, reduced cost and scan duration.

The 2015 AHA/ASA guidelines recommended additional imaging in the following instances (Class IIa; Level of Evidence B).

Lobar haemorrhageAge <55 yearsNormotensive patientsPrimary isolate IVHAssociated subarachnoid haemorrhage.

Catheter angiography (DSA) is not recommended in older hypertensive patients with a typical hypertensive bleed within the basal ganglia, thalamus, cerebellum or brain stem. This guideline was based on the low diagnostic yield of angiography in this subset weighed against the risks of conventional angiography. With the wide availability of non-invasive vascular imaging such as CTA and MRA, the recent AHA/ASA guidelines (2015) are more flexible in terms of in ICH and have left the decision to obtain vascular imaging to the treating clinical team.

 The Way Forward



Available literature suggests that it is difficult to exclude an underlying aetiology even in older patients with a 'characteristic hypertensive capsuloganglionic haemorrhage'.[35] Current evidence also suggests that although DSA is the gold standard, CTA is an equally useful diagnostic tool. It is also reasonable to argue that, while appropriate application of diagnostic imaging modalities can have a meaningful impact on initial management, indiscriminate use of the same can be counterproductive. The answer is in a judicious use of vascular imaging to search for lesions in patients previously thought least likely to have them like older hypertensive patients. However, it will be prudent to exclude elderly hypertensive capsuloganglionic haemorrhage location with poor clinical grade and significant comorbidities from additional imaging.

A practical approach would be as follows [Figure 1]. NCCT scan as the first-line imaging modality to confirm bleed. Once the bleed is confirmed, proceed with a CTA as it is a useful marker of haematoma expansion. CTA also helps to detect the presence of any underlying secondary lesions which would assist the surgeon in case the patient needs emergency decompression. Patients with positive findings on CTA need to follow the specific protocol as recommended for the underlying disease. In the absence of CTA findings, patients are then dichotomised to hypertensive (older patients with typical bleed locations) or normotensive groups according to the current practice. Hypertensives (older patients) are investigated further depending on their clinical status comorbidities and only if they can be taken up for any invasive procedures. Patients, with lobar ICH, need to undergo an MR imaging to look for microbleeds (e.g., CAA) or other aetiologies (neoplasm), angiographically occult lesions such as cavernoma. CTA-negative normotensive patients younger than 55 years, patients with isolated IVH and those with lobar haemorrhage in the absence of microbleeds on MR imaging should be strongly considered for a DSA. CTA-negative normotensive patients older than 55 years may occasionally be considered for DSA in case of strong suspicion of any underlying lesion. Follow-up imaging, in the form of a remote MR imaging at 8–12 weeks, is useful to document the evolution of the ICH and to exclude an abnormality masked by the haematoma.{Figure 1}

 Summary



Spontaneous primary ICH especially capsuloganglionic bleeds may have an underlying vascular aetiology other than hypertension. Angiography to identify this small percentage of patients' needs is a customised judicious approach.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1Qureshi AI, Mendelow AD, Hanley DF. Intracerebral haemorrhage. Lancet 2009;373:1632-44.
2Broderick J, Connolly S, Feldmann E, Hanley D, Kase C, Krieger D, et al. Guidelines for the Management of Spontaneous Intracerebral Haemorrhage in Adults: 2007 Update: A Guideline from the American Heart Association/American Stroke Association Stroke Council, High Blood Pressure Research Council, and the quality of care and outcomes in research interdisciplinary working group. Stroke 2007;38:2001-23.
3Khosravani H, Mayer SA, Demchuk A, Jahromi BS, Gladstone DJ, Flaherty M, et al. Emergency noninvasive angiography for acute intracerebral haemorrhage. AJNR Am J Neuroradiol 2013;34:1481-7.
4Toffol GJ, Biller J, Adams HP Jr., Smoker WR. The predicted value of arteriography in non traumatic intracerebral haemorrhage. Stroke 1986;17:881-3.
5Yeung R, Ahmad T, Aviv RI, de Tilly LN, Fox AJ, Symons SP. Comparison of CTA to DSA in determining the etiology of spontaneous ICH. Can J Neurol Sci 2009;36:176-80.
6Zhu XL, Chan MS, Poon WS. Spontaneous intracranial haemorrhage: Which patients need diagnostic cerebral angiography? A prospective study of 206 cases and review of the literature. Stroke 1997;28:1406-9.
7Stapf C, Van der Worp HB, Steiner T, Rinkel GJ, Nedeltchev K, Mast H, et al. Stroke research priorities for the next decade – A supplement statement on intracranial haemorrhage. Cerebrovasc Dis 2007;23:318-9.
8Caplan LR. Putaminal haemorrhage. In: Kase CS, Caplan LR, editors. Intracerebral Haemorrhage. Boston: Butterworth-Heinemann; 1994. p. 309-27.
9Fasulakis S, Andronikou S. Comparison of MR angiography and conventional angiography in the investigation of intracranial arteriovenous malformations and aneurysms in children. Pediatr Radiol 2003;33:378-84.
10Park J, Hwang YH, Baik SK, Kim YS, Park SH, Hamm IS. Angiographic examination of spontaneous putaminal haemorrhage. Cerebrovasc Dis 2007;24:434-8.
11Tew JM Jr., Lewis AI, Reichert KW. Management strategies and surgical techniques for deep-seated supratentorial arteriovenous malformations. Neurosurgery 1995;36:1065-72.
12Saeki N, Nakazaki S, Kubota M, Yamaura A, Hoshi S, Sunada S, et al. Hemorrhagic type moyamoya disease. Clin Neurol Neurosurg 1997;99 Suppl 2:S196-201.
13Halpin SF, Britton JA, Byrne JV, Clifton A, Hart G, Moore A. Prospective evaluation of cerebral angiography and computed tomography in cerebral haematoma. J Neurol Neurosurg Psychiatry 1994;57:1180-6.
14Jackson CA, Sudlow CL. Is hypertension a more frequent risk factor for deep than for lobar supratentorial intracerebral haemorrhage? J Neurol Neurosurg Psychiatry 2006;77:1244-52.
15Cordonnier C, Klijn CJ, van Beijnum J, Al-Shahi Salman R. Radiological investigation of spontaneous intracerebral haemorrhage: Systematic review and trinational survey. Stroke 2010;41:685-90.
16Douglas MA, Haerer AF. Long-term prognosis of hypertensive intracerebral haemorrhage. Stroke 1982;13:488-91.
17Fisher CM. Pathological observations in hypertensive cerebral haemorrhage. J Neuropathol Exp Neurol 1971;30:536-50.
18Russell DS. Spontaneous intracranial haemorrhage. Proc R Soc Med 1954;47:689-93.
19McCormick WF, Rosenfield DB. Massive brain haemorrhage: A review of 144 cases and an examination of their causes. Stroke 1973;4:946-54.
20Griffiths PD, Beveridge CJ, Gholkar A. Angiography in non-traumatic brain haematoma. An analysis of 100 cases. Acta Radiol 1997;38:797-802.
21Delgado Almandoz JE, Schaefer PW, Goldstein JN, Rosand J, Lev MH, González RG, et al. Practical scoring system for the identification of patients with intracerebral haemorrhage at highest risk of harboring an underlying vascular etiology: The Secondary Intracerebral Haemorrhage Score. AJNR Am J Neuroradiol 2010;31:1653-60.
22Laissy JP, Normand G, Monroc M, Duchateau C, Alibert F, Thiebot J. Spontaneous intracerebral hematomas from vascular causes. Predictive value of CT compared with angiography. Neuroradiology 1991;33:291-5.
23Delgado Almandoz JE, Jagadeesan BD, Moran CJ, Cross DT 3rd, Zipfel GJ, Lee JM, et al. Independent validation of the secondary intracerebral haemorrhage score with catheter angiography and findings of emergent hematoma evacuation. Neurosurgery 2012;70:131-40.
24Yoon DY, Chang SK, Choi CS, Kim WK, Lee JH. Multidetector row CT angiography in spontaneous lobar intracerebral haemorrhage: A prospective comparison with conventional angiography. AJNR Am J Neuroradiol 2009;30:962-7.
25Morgenstern LB, Hemphill JC 3rd, Anderson C, Becker K, Broderick JP, Connolly ES Jr, et al. Guidelines for the management of spontaneous intracerebral haemorrhage: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2010;41:2108-29.
26Evans AL, Coley SC, Wilkinson ID, Griffiths PD. First-line investigation of acute intracerebral haemorrhage using dynamic magnetic resonance angiography. Acta Radiol 2005;46:625-30.
27Murai Y, Takagi R, Ikeda Y, Yamamoto Y, Teramoto A. Three-dimensional computerized tomography angiography in patients with hyperacute intracerebral haemorrhage. J Neurosurg 1999;91:424-31.
28Smith WS, Roberts HC, Chuang NA, Ong KC, Lee TJ, Johnston SC, et al. Safety and feasibility of a CT protocol for acute stroke: Combined CT, CT angiography, and CT perfusion imaging in 53 consecutive patients. AJNR Am J Neuroradiol 2003;24:688-90.
29Eshwar CN, Khandelwal N, Bapuraj JR, Mathuriya SN, Vasista RK, Kak VK, et al. Spontaneous intracranial hematomas: Role of dynamic CT and angiography. Acta NeurolScand 1998;98:176-81.
30Mehran R, Nikolsky E. Contrast-induced nephropathy: definition, epidemiology, and patients at risk. Kidney international. 2006;69:S11-5.
31Krol AL, Dzialowski I, Roy J, Puetz V, Subramaniam S, Coutts SB, et al. Incidence of radiocontrast nephropathy in patients undergoing acute stroke computed tomography angiography. Stroke 2007;38:2364-6.
32Oleinik A, Romero JM, Schwab K, Lev MH, Jhawar N, Delgado Almandoz JE, et al. CT angiography for intracerebral haemorrhage does not increase risk of acute nephropathy. Stroke 2009;40:2393-7.
33Owen RJ, Hiremath S, Myers A, Fraser-Hill M, Barrett BJ. Canadian Association of Radiologists consensus guidelines for the prevention of contrast-induced nephropathy: Update 2012. Can Assoc Radiol J 2014;65:96-105.
34Steiner T, Kaste M, Forsting M, Mendelow D, Kwiecinski H, Szikora I, et al. Recommendations for the management of intracranial haemorrhage - Part I: Spontaneous intracerebral haemorrhage. The European Stroke Initiative Writing Committee and the Writing Committee for the EUSI Executive Committee. Cerebrovasc Dis 2006;22:294-316.
35Kidwell CS, Wintermark M. Imaging of intracranial haemorrhage. Lancet Neurol 2008;7:256-67.