You are on page 1of 11

doi:10.

1093/brain/awr172

Brain 2011: 134; 2376–2386

| 2376

BRAIN
A JOURNAL OF NEUROLOGY

Acute ischaemic brain lesions in intracerebral haemorrhage: multicentre cross-sectional magnetic resonance imaging study
Simone M. Gregoire,1 Andreas Charidimou,1 Naveen Gadapa,1 Eamon Dolan,2 Nagui Antoun,3 ¨ Andre Peeters,4 Yves Vandermeeren,5,6 Patrice Laloux,5,6 Jean-Claude Baron,2,7 Hans R. Jager8 and David J. Werring1
Downloaded from http://brain.oxfordjournals.org/ by guest on September 4, 2012
1 2 3 4 5 6 7 8 Stroke Research Group, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, Queen Square, London WC1N 3BG, UK Department of Clinical Neurosciences, University of Cambridge, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK Department of Neuroradiology, University of Cambridge, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK Department of Neurology, Cliniques Universitaires UCL Saint Luc, Avenue Hippocrate, Brussels, Belgium ´ Department of Neurology, CHU Mont-Godinne, Universite Catholique de Louvain, Avenue Dr G. Therasse, Yvoir, Brussels, Belgium ´ Institute of Neuroscience, Universite Catholique de Louvain, Avenue Hippocrate, Brussels, Belgium ´ UMR 894 INSERM-Universite Paris 5, Paris, France Department of Brain Repair and Rehabilitation, Neuroradiological Academic Unit, UCL Institute of Neurology, London WC1N 3BG, UK

Correspondence to: Dr David J. Werring, UCL Institute of Neurology, National Hospital for Neurology and Neurosurgery, Box 6, Queen Square, London WC1N 3BG, UK E-mail: d.werring@ucl.ac.uk

Subclinical acute ischaemic lesions on brain magnetic resonance imaging have recently been described in spontaneous intracerebral haemorrhage, and may be important to understand pathophysiology and guide treatment. The underlying mechanisms are uncertain. We tested the hypothesis that ischaemic lesions are related to magnetic resonance imaging markers of the severity and type of small-vessel disease (hypertensive arteriopathy or cerebral amyloid angiopathy) in a multicentre, cross-sectional study. We studied consecutive patients with intracerebral haemorrhage from four specialist stroke centres, and age-matched stroke service referrals without intracerebral haemorrhage. Acute ischaemic lesions were assessed on magnetic resonance imaging (53 months after intracerebral haemorrhage) using diffusion-weighted imaging. White matter changes and cerebral microbleeds were rated with validated scales. We investigated associations between diffusion-weighted imaging lesions, clinical and radiological characteristics. We included 114 patients with intracerebral haemorrhage (39 with clinically probable cerebral amyloid angiopathy) and 47 age-matched controls. The prevalence of diffusion-weighted imaging lesions was 9/39 (23%) in probable cerebral amyloid angiopathy-related intracerebral haemorrhage versus 6/75 (8%) in the remaining patients with intracerebral haemorrhage (P = 0.024); no diffusion-weighted imaging lesions were found in controls. Diffusion-weighted imaging lesions were mainly cortical and were associated with mean white matter change score (odds ratio 1.14 per unit increase, 95% confidence interval 1.02–1.28, P = 0.024) and the presence of strictly lobar cerebral microbleeds (odds ratio 3.85, 95% confidence interval 1.15–12.93, P = 0.029). Acute, subclinical ischaemic brain lesions are frequent but previously underestimated after intracerebral haemorrhage, and are three times more common in cerebral amyloid angiopathy-related intracerebral haemorrhage than in other intracerebral haemorrhage types. Ischaemic brain lesions are associated with white

Received March 15, 2011. Revised June 9, 2011. Accepted June 22, 2011 ß The Author (2011). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: journals.permissions@oup.com

. including the basal ganglia. The remaining patients with spontaneous intracerebral haemorrhage not fulfilling the criteria for CAA [e. Prabhakaran et al. Keywords: intracerebral haemorrhage.. cavernomas. 2376–2386 | 2377 matter changes and cerebral microbleeds. Clinical information (demographics. microbleeds Abbreviations: CAA = cerebral amyloid angiopathy. for instance. especially in the elderly. 2010). Probable CAA was defined according to the Boston criteria: patients with two or more spontaneous strictly lobar intracerebral haemorrhage over the age of 55 years.. It is of practical importance to establish the underlying mechanism because one of the few promising treatments currently under investigation for intracerebral haemorrhage is aggressive blood pressure lowering to reduce haematoma expansion (Anderson et al. 2006). or due to cerebral hypoperfusion due to reduced blood pressure in the context of a failure of autoregulation. 2010). and in association with strictly lobar cerebral microbleeds.org/ by guest on September 4. CMBs = Cerebral microbleeds. Soontornniyomkij et al... Menon et al. 92%) . and the Comite d’ethique medicale of the Cliniques Universitaires UCL de Mont Godinne. Ischaemia could. DWI = diffusion-weighted imaging. but the mechanisms underlying these lesions are not clear.. WMC = white matter changes Introduction Intracerebral haemorrhage is the most disabling form of stroke. Diffusion-weighted imaging (DWI) detects acute brain ischaemia with sensitivity approaching 100% (Davis et al. 2008. We reasoned that if ischaemic DWI lesions in intracerebral haemorrhage are due to occlusive small-vessel damage. Downloaded from http://brain. Diffusion-weighted imaging lesions contribute to the overall burden of vascular-related brain damage in intracerebral haemorrhage. Cadavid et al. Queen Square (London).Acute ischaemic brain lesions in intracerebral haemorrhage Brain 2011: 134. CAA).. 2005) and is considered an important cause of lobar intracerebral haemorrhage. the Commission d’Ethique Biomedicale ´ ´ ´ Hospitalo Facultaire of the Faculte de Medicine (Universite ´ ´ ´ Catholique de Louvain). A total of 422 patients with intracerebral haemorrhage were screened. CAA is defined by amyloid-b deposition in the media and adventitia of cortical and leptomeningeal small. Attems. 2010). is an important cause of intracerebral haemorrhage. Haglund et al. Olichney et al. without any other identified cause (Knudsen et al.g.oxfordjournals. 2001). small-vessel disease.. brainstem and thalamus.g. diffusion-weighted imaging. Recent studies using DWI have detected clinically unsuspected areas of acute ischaemia in intracerebral haemorrhage (Kimberly et al. Materials and methods Participants We included patients within 3 months of spontaneous intracerebral haemorrhage consecutively referred to four specialist stroke centres in the United Kingdom and Belgium. for which acute treatments are limited and there is an urgent need to better understand the pathophysiology to help enhance the development of new treatments. All cases were ascertained using overlapping methods from prospective clinical databases and radiological reports. Age-matched controls were ascertained from a database of consecutive patients with suspected stroke attending the neurovascular clinic at The National Hospital for Neurology and Neurosurgery but with a final non-stroke/transient ischaemic attack diagnosis for their presenting symptoms after full investigation. particularly in deep brain structures. affected by two main disease processes: hypertensive arteriopathy and cerebral amyloid angiopathy (CAA). venous thrombosis or head injuries). 2006. vascular risk factors. The study received ethical approval by the National Hospital for Neurology and Neurosurgery and Institute of Neurology Joint ´ Research Ethics Committee. secondary to haemorrhagic infarction. All centres have a policy of routine MRI for investigating the cause of spontaneous intracerebral haemorrhage. and 179 because MRI with the necessary sequences of adequate quality was not available within 3 months of the acute intracerebral haemorrhage. 1987. then they should be related to the severity of the underlying small-vessel disease. Most intracerebral haemorrhages (475%) are classified as primary (spontaneous). They were matched for age with the CAA-related patients with intracerebral haemorrhage using an unbiased group-matching process blinded to their final diagnoses and imaging findings. NEX = number of excitations. of whom 129 were excluded because of diagnosis of non-spontaneous intracerebral haemorrhage (e. We identified all cases with spontaneous lobar intracerebral haemorrhage and appropriate MRI data available. Hypertensive arteriopathy. 1979. arterioles and capillaries (Vinters. we also hypothesized that ischaemic lesions would be most common in intracerebral haemorrhage attributed to probable CAA. a radiological marker for CAA. Delcourt et al. tumours. one strictly lobar haemorrhage. The hospitals were: The National Hospital for Neurology and Neurosurgery. 2009. be due to an active occlusive small-vessel arteriopathy (e. Since neuropathological analyses have identified asymptomatic ischaemic infarction as a common finding in the brain of patients with advanced CAA (Okazaki et al.5T field strength for the majority of patients (105/114. due to rupture of small arteries of the brain... mixed (deep and lobar) haemorrhages and strictly deep haemorrhage(s)] were classified as non-CAA related. 2000. use of anti-thrombotics) was obtained from prospective databases and medical records. suggesting that they result from an occlusive small-vessel arteriopathy. We tested this hypothesis using MRI to detect acute ischaemic brain lesions and relating them to two MRI markers of cerebral small-vessel disease: white matter changes and cerebral microbleeds.. Cliniques Universitaires Saint Luc (Brussels) and Cliniques Universitaires UCL de Mont Godinne (Yvoir).g. 2012 Magnetic resonance imaging protocols The MRI stroke protocol was standardized in each hospital. arteriovenous malformations. 1995. 2009. Imaging was at 1.to medium-sized arteries. aneurysms.. and may be a useful surrogate marker of ongoing ischaemic injury from small-vessel damage. Addenbrooke’s Hospital (Cambridge). characterized by lipohyalinosis and fibrinoid necrosis of small deep perforating arteries. coagulation disorders or use of oral anti-coagulants.

Axial DWI sequences were acquired with the following parameters: repetition time 4900 ms. and included T1-weighted. gap 0.9 Â 0. asymptomatic Cliniques Universitaires UCL de Mont Godinne All MRIs were carried at 1. slice gap 10%]. field of view 250 mm.) to reach consensus. intraventricular haemorrhage and superficial siderosis were also recorded.M. b = 0/1000 s/mm2. matrix 256 Â 256 (GE). DWI and gradient-recalled echo T2* sequences. Queen Square MRIs were carried at 1. echo time 15 ms.5T. field of view 22 cm. with scores ranging from 0 to 3 (Wahlund et al. flip angle 20 . with parameters as follows: 3T.org/ by guest on September 4.8 Â 2.5T field strength. The patients with intracerebral haemorrhage excluded from the study were not significantly different from those included in measures of intracerebral haemorrhage severity including: median intracerebral haemorrhage volume 17. S. in-plane resolution 1. slice thickness 5 mm. Axial gradient-recalled echo T2* images were obtained using a General Electric Genesis Signa Scanner (GE Medical Systems) [repetition time 300 ms. In our age-matched control group. Patients excluded had a higher prevalence of symptomatic lobar intracerebral haemorrhage (excluded: 82% lobar versus included 54% lobar. echo time 55 ms. matrix 256 Â 256. coronal fluid attenuation inversion recovery and axial T2-weighted magnetic resonance images. with gradient-recalled echo parameters as follows: repetition time 600 ms. slice thickness 5 mm. Axial DWI was acquired using the following parameters: repetition time 4500 ms. b = 0/1000 s/mm2. blinded to other imaging sequences and clinical details. field of view 250. Gregoire et al.6 ms. and 47 age-matched controls. blinded to clinical information. matrix 256 Â 256. gap 1 mm.5 mm..93. gap 0. DWI lesions in close proximity (520 mm) to an intracerebral haemorrhage were excluded.G. slice gap 10%. axial gradient-recalled echo T2* sequences were acquired. Gradient-recalled echo T2* sequences were obtained in the axial plane using the following parameters: repetition time 460–660 ms. The locations of DWI lesions were classified as cortical or cortical–subcortical (including cerebellum) or deep (including brainstem).5T GE Medical Systems Signa Excite scanner.5 ms. Brussels At St Luc.J.). repetition time 230 ms. slice thickness 5 mm. matrix 256 Â 160. b = 0/1000 s/ mm2. Intracerebral haemorrhage distribution was classified as lobar (including cerebellum) and deep (including brainstem). the following prevalences of vascular risk factors were found: hypertension 83%. 1. M. diabetes 13%. slice thickness 5 mm. gap 1. P = 0. Downloaded from http://brain. Gradient-recalled echo T2* sequences were obtained in the axial plane on a Siemens Magnetom Symphony System [repetition time 921 ms.5 mm. previous history of stroke 29% and use of anti-thrombotics 46%. Results The final cohort consisted of 114 patients with intracerebral haemorrhage.691. The threshold of statistical significance was set at P 5 0.4 mm. matrix 256 Â 192.002). National Hospital for Neurology and Neurosurgery. gap 0. echo time 61. 2001) on sagittal T1. repetition time 6000 ms. slice thickness 6 mm. Acute ischaemic lesions were identified as bright areas on DWI sequences and corresponding dark areas on apparent diffusion coefficient maps by S. echo time 22 ms. echo time 50–70 ms. b = 0/1000 s/mm2. We recorded the presence and measured the number of acute symptomatic intracerebral haemorrhages and the median total number of intracerebral haemorrhage on MRI. median Glasgow Coma Score on admission: 14 (range 9–15) in patients excluded versus 15 (range 6–15) in patients included.33 mm2 reconstructed to 0. Continuous variables were analysed using independent-samples t-test (when the data distribution was normal) and the Mann–Whitney U-tests (when the distribution was nonnormal). Controls had the following final diagnoses: syncope [n = 7 (17%)]. matrix 256 Â 256 (Siemens). slice thickness 6 mm. field of view 24 Â 18. 2376–2386 S.R. slice thickness 4 mm. We used univariate binary logistic regression to test for the factors associated with DWI lesions. echo time 46 ms. gap 1. field of view 22 cm.05 for all analyses. slice thickness 5 mm. and an echo planar imaging sequence: repetition time 4750 ms.2378 | Brain 2011: 134. slice thickness 3 mm. field of view 280 mm. P = 0. Statistical analysis We compared the characteristics of patients with probable CAA-related intracerebral haemorrhage and the remaining patients with spontaneous intracerebral haemorrhage.5T. 39 of whom fulfilled the criteria for probable CAA.00 cm3) in patients included. and repetition time 2907 ms. fluid attenuation inversion recovery (FLAIR). field of view 230 mm. echo time 93 ms. slice thickness 5 mm.G. b = 0/1000 s/mm2.88 Â 2.. White matter changes (WMC) were rated on a 4-point scale.5 cm3 (range 0.6 ms.M. slice thickness 5 mm. at 3T. echo time 26 ms. echo time 40 ms. Population characteristics were compared using the chi-square test and Fisher’s exact test for categorical variables. The clinical and radiological characteristics of the intracerebral haemorrhage cohort are reported in Table 1. repetition time 3312 ms. slice thickness 4 mm. Cambridge Twelve patients were scanned on a 1. and reviewed by a senior vascular neuroradiologist (H. vestibular disturbance [n = 7 (15%)].2 Â 0. . The presence of subarachnoid blood. rated the presence. slice gap 1. flip angle 20 .40–73. slice thickness 5 mm. NEX 2. repetition time 230–240 ms. Four patients were scanned on a 3T GE Medical Systems Signa HDX scanner. echo time 98.M.5 mm.94 mm2.5 mm. no gap. Four patients were scanned on a Siemens Avanto Scanner (repetition time 800 ms. echo time 94 ms. matrix 128 (2 Â 2 Â 3 mm). echo time 61.4 mm.5T field strength. P = 0. repetition time 3200 ms.27 mm2 reconstructed to 0. gap 1 mm. echo time 16 ms.5 mm. b = 0/1000 s/mm2. gap 7 mm. NEX (number of excitations) 1] for 38 patients. T2-weighted. Cliniques Universitaires Saint Luc. repetition time 10 000 ms. Axial DWI sequences were acquired using the following parameters: repetition time 1000 ms. gap 0. Image analysis All images were analysed by a clinical neurologist (S. NEX 1). slice thickness 4 mm. Axial DWI sequences were acquired with the following parameters: at 1. field of view 240 mm2.31–50.G. flip angle 90 . echo time 93 ms.7 cm3 (range 0. echo time 15 ms.56 cm3) in patients excluded versus 10. matrix 256 Â 192–224. slice thickness 4 mm. echo time 78 ms. 2009). field of view 250 mm.4 mm.4 mm). number and distribution of cerebral microbleeds on gradient-recalled echo T2* images using the microbleed anatomical rating scale (MARS) (Gregoire et al. matrix 256 (1. in-plane resolution 1. gap 0. matrix 512 Â 448.9 mm2. field of view 24 Â 18. gap 7 mm.oxfordjournals. field of view 250 mm. 2012 Addenbrooke’s Hospital.

(A) Bleeds with mixed distribution. Examples of gradient-recalled echo T2* MRI studies of two non-CAA patients are shown in Fig. Probable CAA (n = 39) 74 14 24 11 6 8 5 11 20 8 9 2 9 35 19 1 15 (49–90) (36) (62) (28) (15) (21) (13) (28) (51) (0–76) (2–26) (5) (23) (90) (49) (0–5) (38) All other intracerebral haemorrhagea (n = 75) 69 30 64 23 13 16 7 26 61 5 9 4 21 42 14 1 11 (49–93) (40) (85) (31) (17) (21) (9) (35) (81) (0–84) (0–20) (5) (28) (56) (19) (1–3) (15) P-value 0. (B) Deep bleeds only. There were 33 patients with .004** Downloaded from http://brain. memory problems [n = 2 (4%)]. 2012 A B Figure 1 Gradient-recalled echo T2* images of non-CAA cases. median (range) Presence of cortical infarcts (%) Presence of lacunar infarcts (%) Presence of microbleeds (%) Presence of multiple intracerebral haemorrhage (%) Total number of intracerebral haemorrhage on MRI. median (range) (years) Sex.505 1. The gradient-recalled echo T2* image shows an intracerebral haemorrhage in the right thalamus (red arrow).094 0.05. and one with exclusively deep haemorrhages.859 1. and other [n = 7 (15%)]. 2376–2386 | 2379 Table 1 Characteristics of patients with probable CAA compared with patients with other causes of spontaneous intracerebral haemorrhage Clinical and radiological characteristics of the patients Clinical demographics Age. median (range) Presence of subarachnoid or intraventricular haemorrhage (%) a Possible CAA and non-CAA.884 0.Acute ischaemic brain lesions in intracerebral haemorrhage Brain 2011: 134. 1: one with a mixed distribution of haemorrhages (a large deep symptomatic haemorrhage with both deep and lobar cerebral microbleeds).001** 50.284 50.001** 0. transient global amnesia [n = 4 (9%)].549 0.013* 0. and microbleeds in the left internal capsule and right occipital lobe (white arrows). incidental cerebral infarction [n = 1 (2%)].000 0. carotid stenosis [n = 6 (13%)]. functional [n = 4 (9%)]. migraine [n = 4 (9%)]. small-vessel disease without acute stroke [n = 4 (9%)].001** 0. **P 5 0.300 0.01.537 50. *P 5 0. female (%) Hypertension (%) Smoking history (%) Diabetes (%) Statins on admission (%) Previous history of ischaemic stroke or transient ischaemic attack (%) On anti-thrombotics (%) First ever intracerebral haemorrhage (%) Interval between intracerebral haemorrhage and MRI. median (range) (days) Radiological characteristics Total score of white matter change. The gradient-recalled echo T2* image shows a deep intracerebral haemorrhage in the right basal ganglia (red arrow).oxfordjournals.669 0.org/ by guest on September 4.002** 0.000 0.

Two patients had two DWI lesions. (B) corresponding T2-weighted sequence. except in two patients with simultaneous multiple intracerebral haemorrhages at presentation.42–1. Representative examples of DWI lesions are shown in Fig. and controls = 0/47 (0%). (D) a small infarct is visible on the corresponding T2-weighted sequence (white arrow).328). (C and D) Non-CAA-related intracerebral haemorrhage.25–1. Gregoire et al.2380 | Brain 2011: 134. 2012 C D Figure 2 Examples of positive DWI lesions.org/ by guest on September 4. (8%). (A and B) Probable CAA-related intracerebral haemorrhage. In these. especially the frontal lobes [n = 6 (38%)]. frontal and parietal. and multiple (52) intracerebral haemorrhages (29%). the rest had a single DWI lesion. DWI lesions were all round or ovoid.46 cm diameter in patients with probable CAA. 2. suggestive of CAA.024). DWI lesions were most commonly found in lobar cortical–subcortical areas [n = 12 (75%)]. and 0. their size was 0.oxfordjournals. Prevalence of diffusion-weighted imaging lesions The prevalence of patients with DWI lesions was as follows: intracerebral haemorrhage overall = 15/114 (13%). all intracerebral haemorrhages identified on MRI (apart from the symptomatic lesion) occurred at previous remote times. (A) DWI sequence showing presence of small ischaemic lesion in the right parietal lobe (white arrow) in the presence of two lobar intracerebral haemorrhage. 2376–2386 S. all other intracerebral haemorrhage = 6/75 A B Downloaded from http://brain.50 cm in patients with other causes of spontaneous intracerebral haemorrhage (P = 0. M. Half of the MRI scans were done within 1 week of the presenting intracerebral haemorrhage [63/114 (55%)]. probable CAA = 9/39 (23%). (C) DWI sequence showing presence of small lacunar ischaemic lesion in the left lentiform nucleus (white arrow) in the presence of deep intracerebral haemorrhage. DWI lesions were more common in patients with probable CAA compared with patients with other causes of intracerebral haemorrhage (P = 0. .

No DWI lesions were associated with any new clinical symptoms. and a higher prevalence of lobar cerebral microbleeds [14/15 (93%) versus 57/99 (58%).Acute ischaemic brain lesions in intracerebral haemorrhage 99/114 (87%) within 1 month.25 cm3 (range 0. median (range) Presence of subarachnoid or intraventricular haemorrhage (%) Presence of strictly lobar microbleeds including cerebellum (%) Presence of deep microbleeds including brainstem (%) Total number of strictly lobar microbleeds.org/ by guest on September 4.000 0.327 0.284 0.008].11–12.229 0. female (%) Hypertension (%) Interval between intracerebral haemorrhage and MRI.1 silent Downloaded from http://brain.028* 0.28. 95% CI 1. median (range) Total number of deep microbleeds. Factors associated with diffusionweighted imaging lesions In univariate regression analysis.4 silent infarctions per person-year.973 0. In multivariate analysis.736 0. About one-third of DWI lesions were detected within 1 week [6/17 (35%)]. significant predictors of DWI lesions were smoking history. P = 0. DWI ( + ) lesion (n = 15) 73 4 10 10 162 85 8 4 4 3 3 9 12 0 4 5 1 5 14 4 4 0 (49–88) (27) (67) (0–64) (37) (21) (53) (27) (27) (20) (20) (60) (3–17) (0) (27) (33) (1–3) (33) (93) (27) (0–14) (0–14) DWI ( À ) lesion (n = 99) 71 10 78 4 161 87 26 15 20 9 34 30 8 6 26 28 1 21 57 28 1 1 (19–93) (10) (80) (0–84) (29) (16) (27) (16) (21) (9) (36) (30) (0–26) (6) (26) (29) (0–5) (21) (58) (28) (0–294) (0–14) P-value 0. 2012 Table 2 Characteristics of patients with and without DWI lesions Clinical and radiological characteristics of the patients Clinical demographics Age. For DWI-negative patients. mean (SD) (mmHg) Smoking history (%) Diabetes (%) Statins on admission (%) Previous history of ischaemia or transient ischaemic attack (%) On anti-thrombotics (%) Diagnosis of probable CAA (%) Radiological characteristics Total score of white matter change. 95% CI 1.93. The longest interval between presenting intracerebral haemorrhage and MRI was 84 days.57.85.067 0.034).308 0. then the estimated annual incidence can be calculated as: [Total number of DWI lesions/number of patients]Â [365 (days per year)/duration of DWI lesions (days)] For a consistent incidence of DWI lesions per year.63 cm3). total score of white matter change.024] and the presence of strictly lobar cerebral microbleeds (OR 3.353 0. median (range) *P 5 0.008** 1.693 0. the estimated annual incidence in the whole cohort of patients with intracerebral haemorrhage is therefore (17/114) Â (365/10) = 5.764 0. diagnosis of probable CAA.983 .15–12.. 14/17 (82%) were detected within 1 month.14 per unit increase. median (range) Presence of cortical infarcts (%) Presence of lacunar infarcts (%) Presence of multiple intracerebral haemorrhage (%) Total number of intracerebral haemorrhage on MRI. There was no statistically significant difference in the median intracerebral haemorrhage volume between DWI-positive and DWI-negative groups. P = 0. median (range) (days) Systolic blood pressure. The latest DWI lesion detected was at 64 days after the presenting intracerebral haemorrhage. Adjusting the analysis for variation in magnetic field strength (1. and the presence of strictly lobar cerebral microbleeds (Table 3). **P 5 0. These associations remained significant after adjustment for age and gender.202 0.31– 37.000 0.05. 1998). while in DWI-positive patients. but not deep cerebral microbleeds [4/15 (27%) versus 28/99 (28%)] (Table 2).024* 0.581 0. 2376–2386 | 2381 infarctions per person-year.029) (Table 4).307 0. mean (SD) (mmHg) Diastolic blood pressure.007** 1.840 0. Patients with DWI lesions had more severe white matter change [mean total score 12 (range 3–17) versus 8 (range 0–26).007]. There was no difference in admission systolic or diastolic blood pressure between patients with and without DWI lesions.02–1.5 or 3T) did not substantially weaken the association between strictly lobar microbleeds and the presence of DWI lesions (adjusted OR 3. Brain 2011: 134. median intracerebral haemorrhage volume was 10.oxfordjournals.471 0. median (range) (years) Sex. P = 0. 95% confidence interval (CI) 1.73. P = 0. P = 0. median intracerebral Estimated incidence of diffusionweighted imaging lesions Assuming that DWI lesions remain detectable for $10 days after stroke (Burdette et al.01. in probable CAA-related intracerebral haemorrhage the estimated annual incidence is (9/39) Â (365/ 10) = 8. DWI lesions were associated with both mean white matter change score [odds ratio (OR) 1.

only one imaging study has investigated the presence of silent ischaemic lesions in CAA (Kimberly et al.92 (0.33) Diabetes 1.724 0. Discussion In this study. 2009).39–3.70) Classification of intracerebral haemorrhage mechanism Diagnosis of probable CAA 3. and found that DWI lesions were associated with aggressive blood pressure lowering (Prabhakaran et al. To our knowledge. matter change and the presence of lobar cerebral microbleed. they probably represent acute small-vessel infarcts related to small-vessel arteriopathy. In agreement with our findings..13 (1.03) (0. What are the implications of our findings on the understanding of intracerebral haemorrhage pathophysiology? We demonstrated a high prevalence of silent acute ischaemic events within the first 3 months after CAA-related intracerebral haemorrhage at a mean time of 8 days.93) intracerebral haemorrhage Presence of lacunar infarcts 4. while no lesions were found in a control group of 55 subjects with Alzheimer’s disease or mild cognitive impairment (P = 0. M.438 0.54 (0.55–6.49) On anti-thrombotics 0. male 0.85 0.01–63. The lack of association with admission blood pressure suggests that hypoperfusion resulting from low systemic blood pressure is unlikely to have caused these ischaemic lesions.51–1. DWI-hyperintense lesions were identified in 12 out of 78 subjects (15%) with probable or definite CAA.58–50. We found no DWI-hyperintense lesions in our age-matched controls despite them having a similar or higher prevalence of conventional vascular risk factors to the patients with intracerebral haemorrhage.38) Presence of strictly lobar 10.01–9.941 0. Since DWI lesions are positive only for $10 days (Burdette et al. supporting this conclusion (Reinhard et al. Table 3 Univariate regression analysis testing the factors associated with DWI lesions Factors associated with DWI lesions OR (95% CI) Clinical characteristics Age (years) 1.238 0.13–10.56) Radiological characteristics Total score of white matter change 1. 2006).39–1. they were about three times more common in probable CAA-related intracerebral haemorrhage compared with the other cases of intracerebral haemorrhage.897 Table 4 Multivariate regression analysis showing the factors associated with DWI lesions Factors associated with DWI lesions OR (95% CI) Mean white matter change score Strictly lobar microbleeds Age (years) Sex 1. but this study did not include any intracerebral haemorrhage cases not attributed to CAA.029 0. 1998) and were still detectable in our study up to $2 months (Table 2). we frequently detected subclinical acute ischaemic brain lesions in patients with intracerebral haemorrhage.28) (1. mainly in the cortical–subcortical areas of the frontal and temporal lobes..024 0. DWI lesions were associated with markers of small-vessel damage including total white .049* 0.648 0..234 0. Since these lesions were associated with markers of small-vessel damage severity (white matter change and cerebral microbleeds).32 (1.. 1979.05. the authors did not find an association between DWI lesions and conventional risk factors. including cerebellum Presence of deep microbleeds. P-value 0.030* 0. About a third of silent ischaemic lesions were detected within 1 week after intracerebral haemorrhage.96–1. 2010).62 (1. Ischaemia may therefore be a previously unrecognized cause of progressive disability with possible relevance for clinical outcome after intracerebral haemorrhage.00 (0.09 cm3 (range 0.08 (1.16–1.314 0.374 0.00 (1. 2000. 2376–2386 S. they found no association between DWI lesions and white matter change.88 (0. 2010).59) Number of blood pressure 0. 0. They were small and mostly ovoid or round in shape. but this study did not investigate the relationship of DWI lesions to MRI markers of small-vessel disease and did not use formal criteria to define CAA... and $80% of them within 1 month.02–1.99) High cholesterol 0.047* 0.298 0. More than 80% of the lesions occurred within 1 month post-intracerebral haemorrhage.31–81.org/ by guest on September 4. especially CAA. Imaging data suggest that autoregulation is largely preserved in acute intracerebral haemorrhage.. but we were unable to investigate this further as we did not acquire repeat imaging studies in this cohort.96 (0.17–2.2382 | Brain 2011: 134.26) Presence of multiple 1. 2012 0.14 3.45 (1.15–12. DWI-positive lesions were predominately located in lobar areas.027* Downloaded from http://brain.001). and therefore was not representative of the full spectrum of spontaneous intracerebral haemorrhage.77 (0.97 0. rather than conventional risk factors.892).53) medications at time of intracerebral haemorrhage Smoking history 3.27–3. Cadavid et al. and a common feature of the underlying arteriopathy. Haglund et al.12–1.80) Hypertension 0. although we do not have detailed information on blood pressure control.47–42.05) Sex.469 haemorrhage volume was 12.49 (0.93) (0.55) microbleeds.00 cm3) (P = 0. these lesions are likely to be a dynamic persisting phenomenon in patients with intracerebral haemorrhage. The distribution and morphology of the detected DWI lesions correlates well with the ‘microinfarcts’ identified in neuropathological studies of patients with CAA (Okazaki et al.29) P-value 0. Overall.02–1.5) Presence of microbleeds Presence of microbleeds 8.15–1.45 (0.194 0.oxfordjournals.46 (0. Gregoire et al.92–1. which suggests a possible early peak followed by a descending incidence over time. In contrast to our results.14) including brainstem *P 5 0. Another recent study found DWI lesions in 23% of patients in a cohort of 118 patients with intracerebral haemorrhage scanned within 1 month.019* 0.23 (0. but not with admission blood pressure.

2012 B Artwork: Dr A Charidimou Figure 3 Schematic diagram of microvascular effects of CAA and possible underlying pathophysiological pathways. CAA is associated with cerebral microbleeds in a mainly cortical distribution and with white matter changes (WMC). CAA preferentially affects the small arteries and arterioles of the cerebral cortex and grey-white matter junction by the deposition of amyloid-b in the vessel walls (purple). 2376–2386 | 2383 A Downloaded from http://brain.org/ by guest on September 4.Acute ischaemic brain lesions in intracerebral haemorrhage Brain 2011: 134. (A) The different brain lesions associated with small-vessel disease in patients with sporadic CAA. Micro-infarcts appearing as DWI-positive (continued) .oxfordjournals.

2009. and an intriguing cross-talk between the haemorrhagic and ischaemic aspects of small-vessel diseases.2384 | Brain 2011: 134.oxfordjournals. 2008). Arima et al. and endothelial/vascular smooth muscle dysfunction (Dotti et al. In our study. Jouvent et al. The association of these lesions with cerebral microbleeds suggests shared pathophysiological pathways between the haemorrhagic process and the ischaemia-prone vasculopathy. Another potential mechanism through which CAA-affected cortical and leptomeningeal vessels can damage white matter is by the entrapment of amyloid-b in perivascular spaces. 2001. 2009) (Fig. with a negative predictive value of 39%. These changes can not only cause vessels to become brittle and prone to microaneurysm formation and blood leakage (Attems et al. 2010) (Fig. 3). so that some lobar patients with intracerebral haemorrhage not fulfilling the Boston criteria are likely to have a degree of CAA pathology (Knudsen et al. but the effects on vascular damage from silent ischaemic lesions is unknown. 3). It suffers from limitations due to the lack of pathological confirmation of the ischaemic nature of the DWI lesions and of the CAA pathology as a cause of intracerebral haemorrhage. Finally. The association between asymptomatic acute infarcts and hypertensive small-vessel disease was suggested in previous studies. Thus. The use of PET with amyloid-binding ligands (such as Pittsburgh B-compound) will help to determine whether a regional correlation exists between amyloid deposition and ischaemia.. The association between DWI lesions and white matter change burden suggests that at least some white matter change could be due to accumulated microinfarcts (Smith. 2009). we found no significant differences in measures of intracerebral haemorrhage severity (including haematoma volume and Glasgow Coma Scores on admission) between included and excluded subjects. 2010). although the Boston criteria have been shown to have excellent specificity for CAA. amyloid-b is deposited in small arteries and arterioles.. Kang et al. 2003. (B) Inset of a cortical artery and penetrating arterioles and capillaries from the region in the box asterisked (A).org/ by guest on September 4. suggesting a dynamic continuum between ‘microbleeding’ and ‘microinfarction’. This group had a higher prevalence of hypertension and a mixed pattern of cerebral microbleeds (lobar and deep) suggesting Downloaded from http://brain. In all hospital-based intracerebral haemorrhage MRI studies. 2003).. the concomitant use of DWI and apparent diffusion coefficients maps to assess the presence of silent ischaemia and the review of all positive scans by an experienced neuroradiologist. Since white matter change severity increases with increasing microbleed burden in CAA cohorts and thus with the severity of CAA-related microvasculopathy (Greenberg et al. 2004. 3). a further large study is underway (INTERACT 2) (Delcourt et al. lowering blood pressure is one of the most promising available treatments in acute intracerebral haemorrhage. Gregoire et al. but also to impair local regulation of cerebral blood flow and thus small-vessel or capillary occlusion (Smith et al. 2011). our data suggest a previously unsuspected and intriguing interplay between the haemorrhagic and ischaemic components in small-vessel diseases (Kidwell et al. we could not rule out the confounding effect of a direct influence of the acute intracerebral haemorrhage in causing Indeed. In CAA. 2011).. Patients excluded had a higher prevalence of symptomatic lobar intracerebral haemorrhage. these DWI lesions were located in cortical–subcortical areas of the frontal and temporal lobes... putative mechanisms of vascular dysfunction and downstream brain lesions in CAA. and white matter change severity adds to the evidence that. it is possible that some of the ischaemic lesions in the non-CAA intracerebral haemorrhage group may have in fact been due to the presence of CAA. Soontornniyomkij et al. Our findings of a high prevalence of small DWI-hyperintense lesions indicative of acute ischaemic infarction in patients with CAA can be attributed to multiple potential mechanisms: loss of smooth muscle cells and autoregulation. .. although no definite correlation with hypertension was found (O’Sullivan et al. WMC = white matter changes. causing thickening of the vessel wall and lumen restriction (Olichney et al. Taken together with recent reports of the development of cerebral microbleeds after ischaemic stroke. wall thickening or direct vessel occlusion.. Furthermore. although generally considered to be a haemorrhagic disorder. BBB = blood–brain barrier. However. they were small and ovoid or round in shape. that microinfarction can also result from hypertensive arteriopathy. this could have weakened the association we found between probable CAA and DWI lesions. 1995). impairing interstitial drainage pathways from white matter otherwise not directly affected by CAA pathology (Roher et al. Moreover. endothelial dysfunction. 2010). having been shown to reduce haematoma expansion in a randomized trial (Anderson et al. there is likely to be ascertainment bias towards a less severely impaired cohort. in the present study.. 2010). 2001). or chronic hypoperfusion. 2010). which could lead to an underestimation of the association between DWI lesions and CAA-related intracerebral haemorrhage... DWI lesions were found in only 8% of patients with intracerebral haemorrhage who did not fulfil the criteria for probable CAA. CAA is also characterized by frequent small infarcts (Menon et al. In our study. infarcts seem to relate to the severity of the underlying small-vessel damage. Our study has several strengths: the systematic evaluation of MRI scans by one trained rater using validated scales.. 2009). the relation between DWI lesions and presence of cerebral microbleeds suggests shared pathophysiological pathways for haemorrhagic process and ischaemia (Fig.. as in CAA-related haemorrhage (Dierksen et al. 2376–2386 S. M. Our finding of a significant association between DWI lesions and lobar cerebral microbleeds. 2012 Figure 3 Continued lesions in the MRI scans of these patients are the latest addition to the spectrum of brain lesions in CAA. Anti-hypertensive treatment is also an important chronic treatment for patients following intracerebral haemorrhage (PROGRESS Collaborative Group.. as hypertensive arteriopathy and CAA often co-exist. it is not possible to completely rule out bias as the most severe cases will suffer early mortality or be unable to tolerate an MRI examination. 2004). the sensitivity is low. The deposition of amyloid-b in the vessel wall can disrupt both vessel morphology and function. Other mechanisms such as disruption of the blood–brain barrier and active inflammation could also contribute to this active vasculopathic process.

Neuropathol Appl Neurobiol 2011. 40: e675–7. Greenberg SM. Dotti CG. Ghebremedhin E. Kidwell CS. Kang DW. Poupon C. Hansen LA. Arima H. Smith EE. 41: 394–6.. 75: 693–8. et al. making this unlikely.G. Yousry TA.W. The presence of DWI lesions may identify a subgroup of patients with intracerebral haemorrhage who might have greater benefit from anti-thrombotic treatment. J Neuropathol Exp Neurol 2000. Ricci PE. J Comput Assist Tomogr 1998. Effects of perindopril-based lowering of blood pressure on intracerebral hemorrhage related to amyloid angiopathy: the PROGRESS trial.R. Our study clearly points to a previously unsuspected high prevalence of acute ischaemia in patients within 3 months of intracerebral haemorrhage. et al. Rich PM. Heeley E. Stapf C.J. Neurology 2009. Alzheimer’s dementia by circulation disorders: when trees hide the forest. Le Bihan D..C. Biffi A. Jouvent E. The second (main) phase of an open. and found no relationship between intracerebral haemorrhage volume and the presence of DWI lesions. Dierksen GA. Davis DP. the risk versus benefit of anti-platelet agents is less clear (Rosand et al. Woodward M. Mayo Clin Proc 1979.J. Towfighi A. Attems J. et al. Paquet C. Huang Y. Diffusion-weighted magnetic resonance imaging versus computed tomography in the diagnosis of acute ischemic stroke. Stroke 2010. Haglund M. 7: 391–9.M. Neurology 2010. Grundman M. 24: 1348–54. 52: 702–8. Knudsen KA.). Nat Cell Biol 2009. Jellinger K. Rosand J. Jeng J. Further studies will determine whether the detection of DWI-positive lesions in CAA can be used as a surrogate marker with value in diagnosis or assessing the severity and progression of the disease. Tzourio C. O’Sullivan M. Rost N. randomised. J Emerg Med 2006. Spatial relation between microbleeds and amyloid deposits in amyloid angiopathy. Huang Y. 35: 2616–9. Englund E. Gray F. Lancet Neurol 2008. Rosand J. Chalela JA.org/ by guest on September 4.). Menon RS. Dr Cecile Grandin and Dr Beatrice de Coene. 73: 1759–66. Mena H. Olichney JM. Bousser MD. Cerebral amyloid angiopathy and cortical microinfarcts as putative substrates of vascular dementia. et al. Clinical diagnosis of cerebral amyloid angiopathy: validation of the Boston criteria. De Strooper B. et al. Biffi et al. et al. et al. Stroke 2011. Department of Health/Higher Education Funding Council for England (Clinical Senior Lectureship Award) and the Stroke Association (research support to D. 2376–2386 | 2385 Acknowledgements We would like to thank Mrs Adrienne Wallis and Mrs Helen Green for their technical assistance. Review: sporadic cerebral amyloid angiopathy. 2000. ¨ Gregoire SM. Rost NS. We are also grateful to Prof. Greenberg SM. 2012 . AJNR Am J Neuroradiol 2003. 54: 22–31. 110: 345–59. 56: 537–9. Arima H.). Wang JG. Robertson T. Macmahon S. References Anderson CS. Okazaki H. Lindley R. Evidence for a prolonged stroke-prone state? Neurology 2004. Markus HS. Greenberg SM. Brain 2011: 134. Intracortical infarcts in small vessel disease. Sporadic cerebral amyloid angiopathy: pathology. Clark CA. Early and late recurrence of ischemic lesion on MRI. Rosand J. 22: 792–4. Aspirin and recurrent intracerebral hemorrhage in cerebral amyloid angiopathy. but we excluded any DWI lesions in close proximity to the symptomatic intracerebral haemorrhage. Anderson C. Gilson A. Sjobeck M. Neurology 2001. Silent ischemic infarcts are associated with hemorrhage burden in cerebral amyloid angiopathy. Stroke Association (TSA 2006/08). 5: 110–6. Amyloid angiopathy-related vascular cognitive impairment. Dambrosia JA. Clinicopathologic studies of primary cerebral amyloid angiopathy. Thal DR. The Microbleed Anatomical Rating Scale (MARS): reliability of a tool to map brain microbleeds. Cadavid D. Jager HR. Thierry Duprez. Neuroimaging demonstration of evolving small vessel ischemic injury in cerebral amyloid angiopathy. Viswanathan A. et al. Karluk D. Gilson A. Int J Stroke 2010. Delcourt C. 37: 75–93. Downloaded from http://brain. Nandigam RN. Brown MM.). 2010).Acute ischaemic brain lesions in intracerebral haemorrhage ischaemia as we did not include a group of patients with CAA without symptomatic lobar haemorrhage. Cerebral infarction in Alzheimer’s disease is associated with severe amyloid angiopathy and hypertension. Kimberly WT. 42: e27–30. Khan MA. Neurology 2009. Van Nostrand W. Int J Geriatr Psychiatry 2006. Koeller K.oxfordjournals. Funding Stroke Association (research support to S. 59: 768–73. Stroke 2009. neuroradiologists at Cliniques UCL Saint Luc and Cliniques UCL de Mont Godinne in Belgium for MRI assistance. A case control study in human brain biopsies. but clinical studies are needed to address this question. Our findings raise clinical questions. Frequency of subclinical lacunar infarcts in ischemic leukoaraiosis and cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy. Attems J. Passant U. Hofstetter CR. Latour LL. Elster AD. especially the dilemma about the use or avoidance of anti-thrombotics in patients with intracerebral haemorrhage. A combined 7-T postmortem MRI and neuropathological case study in cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy. 73: 1614–5. Intensive blood pressure reduction in acute cerebral haemorrhage trial (INTERACT): a randomised pilot trial. clinical implications. Becker JA. Smith EE. 72: 1230–5. Cerebral beta amyloid angiopathy is a risk factor for cerebral ischemic infarction. Wang J. Stroke 2004. Chaudhary UJ. Frommelt RA. Katzman R. et al. Thal LJ. Acta Neuropathol 2005. 68: 545–8. Although the risk of anti-coagulation likely overweighs the benefit in most patients with intracerebral haemorrhage due to CAA. Imbesi SG. This work was undertaken at UCLH/ UCL that received a proportion of funding from the Department of Health’s National Institute for Health Research Biomedical Research Centre’s funding scheme. Mangin JF. Arch Neurol 1995. 11: 114–6. Busl K. Greek State Scholarship Foundation (research support to A. Samantha Dickson Brain Tumour Trust and the Brain Research Trust (research support to H. Ann Neurol 2010. Campbell RJ. Halpin A. Burdette JH. 21: 681–7. Warach S. and possible pathomechanisms. Skehan ME. 31: 269–77. Barrick TR. Peng B. Neal B. associations and clinical impact. but longitudinal studies with serial clinical examinations and imaging are also needed to further characterize the dynamics. Red meets white: do microbleeds link hemorrhagic and ischemic cerebrovascular disease? Neurology 2009. Reagan TJ. Kidwell CS. multicentre study to investigate the effectiveness of an intensive blood pressure reduction in acute cerebral haemorrhage trial (INTERACT2). 63: 2261–5. Calculation of apparent diffusion coefficients (ADCs) in brain using two-point and six-point methods. Gurol ME. Kallis C.

Temes RE. 2012 . Prabhakaran S. Kalback W. Mermash S. Ouyang B. Warfarin-associated hemorrhage and cerebral amyloid angiopathy: a genetic and pathologic study. Stroke 2001. 41: S139–43. Mohammad Y. PROGRESS Collaborative-Group. 2376–2386 S.org/ by guest on September 4. et al. John S. O’Donnell HC. Leukoaraiosis and stroke. Buttler KJ. 40: 2601–6. Kuo YM. Downloaded from http://brain. Reinhard M. 18: 311–24. Mol Med 2003. et al. 41: 89–94. Acute brain infarcts after spontaneous intracerebral hemorrhage: a diffusion-weighted imaging study. Smith EE. Neunhoeffer F. Pomakian J. Clare R. Neurology 2000. Sjogren M. et al. Secondary decline of cerebral autoregulation is associated with worse outcome after intracerebral hemorrhage. Barkhof F. ¨ Wahlund LO. Knebel C. et al.2386 | Brain 2011: 134. Rosand J. et al. Hylek EM. Vinters HV. Cerebral microinfarcts associated with severe cerebral b-amyloid angiopathy. 55: 947–51. Greenberg SM. A critical review. 9: 112–22. Cortical and leptomeningeal cerebrovascular amyloid and white matter pathology in Alzheimer’s disease. Fazekas F. Stroke 1987.105 individuals with previous stroke or transient ischaemic attack. 36: 264–71. Timmer J. Augustin M. Randomised trial of a perindopril-based blood pressure-lowering regimen among 6. Smith EE. Brain Pathol 2010. Stroke 2010. blood vessels. Gupta R. and brain function. Cerebral amyloid angiopathy. 20: 459–67. Gregoire et al. 32: 1318–22. Roher AE. Gerds TA. A new rating scale for age-related white matter changes applicable to MRI and CT. Esh C. M. Stroke 2009. Greenberg SM. Beta-amyloid. Stroke 2010. Lancet 2001. Badkoobehi H. 358: 1033–41. Lynch MD.oxfordjournals. Niesen WD. Weiss N. Intensive Care Med 2010. Bronge L. Soontornniyomkij V.