Zhenyu Jia, MD,* Wasif Mohammed, MD,* Yiru Qiu, MD, Xunning Hong, MD,* and Haibin Shi, MD, PhD*

Hypertension Increases the Risk of Cerebral Microbleed in the Territory of Posterior Cerebral Artery: A Study of the Association of Microbleeds Categorized on a Basis of Vascular Territories and Cardiovascular Risk Factors Zhenyu Jia, MD,* Wasif Mohammed, MD,* Yiru Qiu, MD, Xunning Hong, MD,* and Haibin Shi, MD, PhD* Background: It has been suggested that the etiology of cerebral microbleeds (CMBs) differs according to their location in the brain, with lobar microbleeds being caused by cerebral amyloid angiopathy and deep or infratentorial microbleeds resulting from hypertension and atherosclerosis. We hypothesized that there were associations between cerebral arterial branches, cardiovascular risk factors, and the occurrence of CMBs. We examined these relationships in the current study. Methods: Three hundred ninety-three patients with CMBs were analyzed in this study. The CMBs were listed according to the various arterial territories, and these were assessed for their relationship with cardiovascular risk factors, markers of small vessel disease, and their presence and location using multiple logistic regression. Results: Systolic blood pressure had a significant association with CMBs in the territory of the posterior cerebral artery and the deep and infratentorial locations. The presence of lacunar infarcts, hemorrhage, and white matter changes were associated with CMBs in nearly all arterial territories. Conclusions: Hypertension increases the risk of microbleeds in the territory of the posterior cerebral artery and the deep and infratentorial locations. Cerebral amyloid angiopathy may be responsible for the microbleeds in the lobar area of brain. Key Words: Cerebral arterial territory hypertension magnetic resonance imaging microbleeds posterior cerebral artery. Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of National Stroke Association. All rights reserved. From the *Department of Radiology, The First Affiliated Hospital of Nanjing Medical University, Nanjing; and Customer Service, Healthcare Sector, Siemens Ltd, Beijing, China. Received November 30, 2012; revision received December 16, 2012; accepted December 25, 2012. Supported by the priority academic program development of Jiangsu Higher Education Institutions. Address correspondence to Haibin Shi, MD, PhD, Department of Radiology, The First Affiliated Hospital of Nanjing Medical University, 300 Guang Zhou Rd, Nanjing 210000, China. E-mail: jsphradiology@hotmail.com. 1052-3057/$ - see front matter Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of National Stroke Association. All rights reserved. http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2012.12.016 Cerebral microbleeds (CMBs) are defined as rounded areas of signal loss that are seen on T2*-weighted gradient echo magnetic resonance imaging (MRI) sequences, generally 2 to 5 mm in size, some of which represent deposits of hemosiderin. 1-3 It has been suggested that the etiology of CMBs differs according to their location in the brain, with lobar microbleeds being caused by cerebral amyloid angiopathy (CAA) and deep or infratentorial microbleeds resulting from hypertension and atherosclerosis. 1,4,5 Most of these studies categorized CMBs into 1 of 3 anatomic locations: lobar (cortical gray and subcortical or periventricular white matter), deep (deep gray matter [basal ganglia and thalamus] and the white matter of the corpus callosum, the internal, external, and extreme capsule), and infratentorial (brainstem and cerebellum). 1,4 To the best of our knowledge, no study has been conducted on the relationship between CMBs in various vascular territories and the related cardiovascular risk factors. Journal of Stroke and Cerebrovascular Diseases, Vol. 23, No. 1 (January), 2014: pp e5-e11 e5

e6 Because an existence of a potential relationship between CMBs and the severity of small vessel diseases (SVDs) was suggested, 6 and because we found 4 cases with numerous CMBs restricted to the territory of the posterior cerebral artery (PCA) in the current review study (Fig 1), we hypothesized that categorizing CMBs into arterial territories was more clinically relevant. Z. JIA ET AL. Between October 2009 and September 2011, approximately 8000 patients seen in our hospital received a MRI scan of the brain for neurologic symptoms, and the MRI protocol includes a susceptibility-weighted imaging (SWI) sequence. Four hundred eighty-nine cases with CMBs were chosen. In these 489 cases, 96 cases (9 with brain tumors, 11 with brain surgery, 9 with brain trauma, 55 with chemotherapy, and 12 with immune suppression therapy) were deemed ineligible and excluded. The final sample of 393 cases with CMBs were included and analyzed in the current study. Methods Participants Brain MRI A routine MRI protocol was performed on the subjects using a 3.0-Tesla MRI scanner (Trio; Siemens, Munich, Germany). The MRI protocol included a T1-weighted sequence, a T2-weighted sequence, a fluid-attenuated inversion recovery (FLAIR) sequence, a SWI sequence, and a diffusion-weighted imaging (DWI) sequence. The SWI Figure 1. A-D, Posterior cerebral artery territory (shown in blue) and numerous cerebral microbleeds were restricted to the territory of the posterior cerebral artery. B, Isolated cerebral microbleed in the territory of the right middle cerebral artery (curved arrow). D, Hemorrhage (arrow).

HYPERTENSION AND CMB IN THE TERRITORY OF THE PCA sequence was a 3-dimensional, T2*-weighted, gradientrecalled echo sequence with a high resolution used for microbleed detection. 6,7 The parameters of SWI were as follows: TR/TE 28/20 ms, flip angle 15,matrix4483 364, number of excitations 1, field of view 18.68 3 23.0 cm, and slice thickness 2.0 mm. Rating of Cerebral Microbleeds Microbleeds were defined as focal areas of very low signal intensity in SWI (,10 mm in size). 4 Signal voids caused by sulcal vessels, symmetric calcifications in the basal ganglia, choroid plexus and pineal body, and signal averaging from bone were excluded. 8 All SWI scans were reviewed by 2 trained raters at the same time, and they recorded the presence, number, and location of microbleeds. If their opinions differed regarding the location or the nature of the CMB, a senior neuroradiologist was consulted to give a final decision. CMBs were categorized into 1 of 9 locations that were defined according to cerebral arterial territories. Cerebral Arterial Territory In this study, we categorized microbleeds into 9 different arterial territories to analyze the association between the prevalence of microbleeds and their risk factors. The various intracranial arteries that we took under consideration and their distributions are as follows: anterior cerebral artery (ACA) frontal and parietal lobe; middle cerebral artery (MCA) frontal, temporal, and parietal lobe; PCA temporal, parietal, and occipital lobe; medial lenticulostriate arteries (MSAs) arise from the ACA head of caudate; lateral lenticulostriate arteries (LSAs) arise from the MCA most part of ganglia; artery of thalamus from the PCA; artery of brain stem from the PCA and basilar artery; anterior choroideal artery (ACHA) from the internal carotid artery hippocampus; and the posterior limb of internal capsule artery of cerebellum from the basilar artery. 9,10 For assessment of arterial territory distribution of microbleeds, we first created a template scan in which the territories were labeled according to the cerebral arterial territory as described by Savoiardo. 9,10 If necessary, we modified the territories slightly according to the transverse images of time of flight (TOF) sequence for magnetic resonance angiography, so as to make the territories better match the distribution of the cerebral arteries. The 9 locations of arterial territories were named as follows: ACA, MSA, MCA, LSA, ACHA, PCA, thalamus, cerebellum, and brain stem. Cerebrovascular Disease on Magnetic Resonance Imaging Acute and subacute infarctions were defined as lesions with high signal intensity on FLAIR and DWI sequences. Old infarction was defined as lesions of low signal intensity on T1-weighted images, high signal intensity on T2- weighted images, and low signal intensity on FLAIR sequences with a hyperintense rim. The diagnosis of hemorrhage was given according to the diagnostic criteria as follows: hyperacute hemorrhage T1 high/t2 high; acute hemorrhage T1 low/t2 low surrounded by hyperintense margin; subacute hemorrhage T1 high/t2 low or high; and chronic hemorrhage T1 low/t2 low or hypointense margin surrounding hyperintense fluid cavity. White matter changes (WMCs) were assigned between 0 and 6 points using the Fazekas et al 11 visual rating scale, and we then categorized the WMCs into 4 grades (0 points 5 grade 0; 1-2 points 5 grade 1; 3-4 points 5 grade 2; and 5-6 points 5 grade 3). Cardiovascular Risk Factors Cardiovascular risk factors were determined by reviewing patients case histories in our hospital information system. The risk factors included in our analysis were systolic and diastolic blood pressure, hypertension (categorized into stage 0, 1, 2, or 3 according to the World Health Organization criteria), smoking, diabetes mellitus, antiplatelet drugs, international normalized ratio (INR), high-density lipoprotein (HDL), and low-density lipoprotein (LDL). All data were acquired at the time of admission. Data Analysis We assessed the relationship between the prevalence of microbleeds and cardiovascular risk factors using multiple logistic regression. To examine whether CMBs were Table 1. Characteristics of the study population Mean age, y (SD) 70.7 (10.8) Women, n (%) 137 (34.9) Mean systolic blood pressure, mm Hg (SD) 153.4 (24.3) Mean diastolic blood pressure, mm Hg (SD) 88.1 (14.0) Hypertension, n (%) Grade 0 82 (20.9) Grade 1 141 (35.9) Grade 2 89 (22.6) Grade 3 81 (20.6) Smoking, ever, n (%) 82 (20.9) Diabetes mellitus, n (%) 101 (25.7) Antiplatelet therapy, n (%) 94 (23.9) Mean serum HDL cholesterol, mmol/l (SD) 1.17 (.28) Mean serum LDL cholesterol, mmol/l (SD) 2.76 (.79) International normalized ratio 1.06 (.74) Acute infarct on MRI, n (%) 94 (23.9) Hemorrhage on MRI, acute or remote, n (%) 66 (16.8) Old infarct on MRI, n (%) 42 (10.7) White matter lesions on MRI (score 0-6), n (%) 0 66 (16.8) 1-2 150 (38.2) 3-4 118 (30.0) 5-6 59 (15.0) e7 Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; MRI, magnetic resonance imaging; SD, standard deviation.

e8 Z. JIA ET AL. more frequent in persons with brain infarcts or hemorrhages, we used multiple logistic regression models adjusted for age, sex, and cardiovascular risk factors. Both hypertension and WMCs were collected as grading data, so we analyzed these 2 datasets using the 2 independent sample tests method. All analyses were performed with SPSS software (version 11.0.1; SPSS Inc, Chicago, IL). Results The neurologic symptoms of the 393 cases were as follows: 144 with located signs, 72 with symptoms of transient ischemia attack (TIA), and 177 with chronic symptoms, such as headache, dizziness, and blurred vision. Table 1 shows the characteristics of all 393 patients. The mean age was 70.7 years; 137 patients were women. Of them, 101 patients were diagnosed with diabetes mellitus, 94 with acute infracts, 66 with chronic or acute intracranial hemorrhages, 125 with white matter lesions (WMLs; grade 1-3), and 66 without visible lesions. In the lobar area, we found that the number of CMBs in different arterial territories were 926 in the MCA, 497 in the (PCA), and 113 in the (ACA; Table 2). We did not find an association between the prevalence of microbleeds and age or sex for most locations except for ACHA and the cerebellum (Table 3). High systolic blood pressure was associated with presence of microbleeds in 5 locations: the MSA (odds ratio [OR] 1.02; 95% confidence interval [CI] 1.00-1.03), LSA (OR 1.01; 95% CI 1.00-1.02), PCA (OR 1.01; 95% CI 1.00-1.03), thalamus (OR 1.02; 95% CI 1.01-1.03), and brain stem (OR 1.02; 95% CI 1.01-1.03; Table 3), while with the increase of serum HDL, the prevalence of microbleeds decreased in the thalamus (OR 0.42; 95% CI 0.19-0.90) and brain stem (OR 0.40; 95% CI 0.18-0.89; Table 3). Acute and subacute infarction on MRI were related to the presence of cerebral microbleeds in the locations of the thalamus and brain stem, the hemorrhages were seen in the area of the MSA, MCA, LSA, PCA, thalamus, cerebellum, and brain stem, and old infarction lesions were seen in the area of the LSA, PCA, thalamus, cerebellum, and brain stem (Table 3). We analyzed hypertension as a grading criterion, and its significance was confirmed in our result shown in Table 3. WMCs were strongly associated with microbleeds in all locations (Table 4). Additional adjustment for cardiovascular risk factors did not change these results. Discussion A major strength of our study is its large sample size. We excluded cases with brain tumor, brain injury, brain surgery, or patients having received chemotherapy or immunosuppression therapy, because these factors could have independently increased the incidence of CMBs and the inclusion of such cases could have led to skewed results. Table 2. Distribution of cerebral microbleeds based on arterial territories Location of microbleed Total ACA MSA MCA LSA ACHA PCA Thalamus Cerebellum Brain stem Cases, n (%) 69 (17.5%) 53 (13.5%) 245 (62.3%) 152 (38.7%) 58 (14.8%) 171 (43.5%) 164 (37.2%) 98 (24.9%) 148 (37.7%) 393 Microbleeds, n (%) 113 (3.5%) 81 (2.5%) 926 (27.6%) 440 (13.1%) 85 (2.5%) 497 (14.8%) 571 (17.0%) 207 (6.2%) 430 (12.8%) 3350 Abbreviations: ACA, anterior cerebral artery; ACHA, anterior choroideal artery; LSA, lateral lenticulostriate artery; MCA, middle cerebral artery; MSA, medial lenticulostriate artery; PCA, posterior cerebral artery.

HYPERTENSION AND CMB IN THE TERRITORY OF THE PCA Table 3. Cardiovascular risk factors, cerebral vascular disease, and the presence of cerebral microbleeds* e9 Location of microbleed ACA MSA MCA LSA PCA ACHA Thalamus Cerebellum Brain stem Age per year.031; 1.04.014; 1.03 (1.01-1.07) (1.01-1.05) SBP per SD increase.026; 1.02 (1.00-1.03).050; 1.01 (1.00-1.02).024; 1.01 (1.00-1.03).001; 1.02 (1.01-1.03).007; 1.02 (1.01-1.03) HDL per SD increase.025; 0.42 (.19-.90).024; 0.40 (.18-.89) Acute and subacute.002; 2.14 (1.31-3.48).000; 2.44 (1.49-4.00) infarction Hemorrhage.018; 2.23.007; 2.33.000; 3.81.012; 2.01.000; 3.63.019; 1.98.000; 3.64 (1.14-4.36) (1.26-4.32) (2.18-6.67) Old infarction.020; 2.18 (1.12-4.23) (1.17-3.43).028; 2.09 (1.08-4.03) (2.05-6.44).003; 2.90 (1.45-5.78) (1.12-3.50).015; 2.30 (1.18-4.50) (2.06-6.43).000; 3.60 (1.80-7.22) Abbreviations: ACA, anterior cerebral artery; ACHA, anterior choroideal artery; CI, confidence interval; HDL, high-density lipoprotein; LSA, lateral lenticulostriate artery; MCA, middle cerebral artery; MSA, medial lenticulostriate artery; OR, odds ratio; PCA, posterior cerebral artery; SBP, systolic blood pressure; SD, standard deviation. Note: Some factors (sex, diastolic blood pressure, smoking, diabetes mellitus, antiplatelet drugs, international normalized ratio, and lowdensity lipoprotein cholesterol) do not show a significant association with cerebral microbleeds in all 9 locations, so results were not shown in the table. *All values presented as P value; OR (95% CI). Results with P..05 are not shown. Another strength of this study is the use of a 3.0-Tesla magnetic field strength. The SWI sequence, a fully velocity-compensated, high-resolution, 3-dimensional gradient-echo sequence that uses magnitude and filteredphase information, has a high sensitivity in detecting cerebral microbleeds when compared with conventional T2*-weighted gradient echo sequences 7 was also a major asset. Our study was limited to being restricted to the patients who presented with neurologic symptoms and received MRI scans. Our results might not be an accurate representation of the risk or prevalence in the general population. The definition of the border between the arterial territories played an important role in our study. Although we used 2 methods to get an accurate delineation, 9,10 there remains a small chance that the microbleeds located at the watershed area between 2 different arterial territories might have influenced the results. In contrast to previous studies, 1,3 we only found a slight association between age and the lesions of ACHA and the cerebellum. In addition, there was no sex predilection in CMB prevalence and men did not have a higher risk of CMBs than women. It might not be a true representation of the general population, given that our sample was comprised of CMB patients and, in such situations, the association between age/sex and lesions may not be significant. The significant association between hemorrhage and CMBs seen in our study suggests that microbleeds may reflect an increased risk of hemorrhagic stroke or hemorrhagic transformation of ischemic stroke. 4,12 The Boston criteria for diagnosis of CAA rely on the tendency for CAA-related hemorrhages to be multiple and to occur primarily in cortical and cortical subcortical brain regions. 13 The relationship between hemorrhage and CMBs in the arterial territories of lobar areas, such as the MCA and PCA, reflect an association between CMBs and CAA. The association between WMCs and the incidence of microbleeds is significant in all 9 Table 4. White matter changes, hypertension, and incidence of cerebral microbleeds* Location of cerebral microbleed ACA MSA MCA LSA PCA ACHA Thalamus Cerebellum Brain stem White matter changes (grade 0-3) P 5.00 P 5.00 P 5.00 P 5.00 P 5.00 P 5.00 P 5.00 P 5.00 P 5.00 Hypertension (grade 0-3) P 5.00 P 5.00 P 5.00 P 5.00 Abbreviations: ACA, anterior cerebral artery; ACHA, anterior choroideal artery; LSA, lateral lenticulostriate artery; MCA, middle cerebral artery; MSA, medial lenticulostriate artery; PCA, posterior cerebral artery. *Results with P..05 not shown in the table.

e10 territories, a finding that was in line with other studies. 4,14,15 It may also be evidence to support the association between CMBs and CAA in the lobar area. 1 Another relationship was found between hypertension and CMBs in the arterial territories of the LSA, PCA, thalamus, and brain stem. This relationship might be explained by a close association between small-vessel diseases (SVDs) and CMBs. 3,16 SVDs, such as WMCs and infarctions, were more frequent in patients with hemorrhage closely related to hypertension. 17 Therefore, CMBs in such territories related to hypertension might reflect target organ damage of the entire brain caused by chronic hypertension. Aside from the deep region of brain (e.g., the LSA, thalamus, and brain stem), we found an association between hypertension and CMBs in the territory of PCA. In other cross-sectional studies, 4,14 an association between cardiovascular risk factors, the presence of lacunar infarcts and white matter lesions, and prevalent microbleeds in a deep or infratentorial region was found, but not in a lobar location like the PCA territory. Although the pathogenesis of pure lobar microbleeds was more closely related to CAA and less closely related to hypertension, 1 Yakushiji et al, 17 in their study, noted that all patients with strictly lobar CMBs (n 5 6) had a history of hypertension, and both of their patients who met the Boston criteria for possible CAA also had a history of hypertension. They concluded that even in patients with CMBs in the lobar area, hypertension may also be a causative factor for future CAA-related hemorrhages. Histopathologic analyses of CMBs have shown that CMBs reflect widespread deterioration of the arterioles caused by hypertension, amyloid deposition, or both pathogenic mechanisms. 18 We believe that because the PCA and the arteries of the thalamus and pons originate from the same artery (i.e., the basilar artery), and because the PCA is straight and short, its territory might be more prone to be affected by systolic blood pressure. What s more, in studies by McKinney et al 19 and Hefzy et al, 20 hemorrhages and microbleeds were detected in posterior reversible encephalopathy syndrome, in which a unique pattern of brain vasogenic edema predominates in the parietal and occipital regions. This was attributed to be caused by immunosuppressant drugs used in solid organ transplants and by severe hypertension, 19,20 which we think may also be an explanation for the mechanism of CMB occurrence in the PCA territory. There was a lower prevalence of CMBs in the area of thalamus and brain stem in patients with a high HDL level, and this was partially in agreement with previous reports. 4,21,22 Although it seems possible that serum cholesterol plays a role in vessel wall integrity, the underlying mechanism of this association is still unknown. An association between acute and subacute infarction and the lesion in the arterial territories of thalamus and brain stem was found, but we did not find obvious infarctions in these 2 areas, so this association may not be accurate. In summary, hypertension increases the risk of microbleeds in the territory of the PCA and the deep/infratentorial locations. CAA may be responsible for microbleeds in the lobar area of brain. Our study offers new insights into the risk factors for microbleeds and warrants further investigation into the prognosis of microbleeds according to arterial territories. References Z. JIA ET AL. 1. Jeerakathil T, Wolf PA, Beiser A, et al. Cerebral microbleeds: Prevalence and associations with cardiovascular risk factors in the Framingham Study. Stroke 2004;35:1831-1835. 2. Fisher M, French S, Ji P, et al. Cerebral microbleeds in the elderly: A pathological analysis. Stroke 2010;41:2782-2785. 3. Roob G, Schmidt R, Kapeller P, et al. MRI evidence of past cerebral microbleeds in a healthy elderly population. Neurology 1999;52:991-994. 4. Vernooij MW, van der Lugt A, Ikram MA, et al. Prevalence and risk factors of cerebral microbleeds: The Rotterdam Scan Study. Neurology 2008;70:1208-1214. 5. Koennecke HC. Cerebral microbleeds on MRI: Prevalence, associations, and potential clinical implications. Neurology 2006;66:165-171. 6. Vernooij MW, Ikram MA, Wielopolski PA, et al. Cerebral microbleeds: Accelerated 3D T2*-weighted GRE MR imaging versus conventional 2D T2*-weighted GRE MR imaging for detection. Radiology 2008;248:272-277. 7. Haacke EM, Mittal S, Wu Z, et al. Susceptibility-weighted imaging: Technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol 2009;30:19-30. 8. Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: A guide to detection and interpretation. Lancet Neurol 2009;8:165-174. 9. Savoiardo M. The vascular territories of the carotid and vertebrobasilar systems. Diagrams based on CT studies of infarcts. Ital J Neurol Sci 1986;7:405-409. 10. Chan J, D Souza D. Cerebral vascular territories. Available from http://radiopaedia.org/articles/cerebral_ vascular_territories. Accessed January 18, 2013. 11. Pantoni L, Simoni M, Pracucci G, et al. Visual rating scales for age-related white matter changes (leukoaraiosis): Can the heterogeneity be reduced? Stroke 2002;33:2827-2833. 12. Nighoghossian N, Hermier M, Adeleine P, et al. Old microbleeds are a potential risk factor for cerebral bleeding after ischemic stroke: A gradient-echo T2*-weighted brain MRI study. Stroke 2002;33:735-742. 13. Vinters HV. Cerebral amyloid angiopathy. A critical review. Stroke 1987;18:311-324. 14. Poels MM, Ikram MA, van der Lugt A, et al. Incidence of cerebral microbleeds in the general population: The Rotterdam Scan Study. Stroke 2011;42:656-661. 15. Goos JD, Henneman WJ, Sluimer JD, et al. Incidence of cerebral microbleeds: A longitudinal study in a memory clinic population. Neurology 2010;74:1954-1960. 16. Tsushima Y, Aoki J, Endo K. Brain microhemorrhages detected on T2*-weighted gradient-echo MR images. AJNR Am J Neuroradiol 2003;24:88-96. 17. Yakushiji Y, Yokota C, Yamada N, et al. Clinical characteristics by topographical distribution of brain microbleeds, with a particular emphasis on diffuse microbleeds. J Stroke Cerebrovasc Dis 2011;20:214-221. 18. Fazekas F, Kleinert R, Roob G, et al. Histopathologic analysis of foci of signal loss on gradient-echo T2*-weighted MR images in patients with spontaneous intracerebral

HYPERTENSION AND CMB IN THE TERRITORY OF THE PCA e11 hemorrhage: Evidence of microangiopathy-related microbleeds. AJNR Am J Neuroradiol 1999;20:637-642. 19. McKinney AM, Sarikaya B, Gustafson C, et al. Detection of microhemorrhage in posterior reversible encephalopathy syndrome using susceptibility-weighted imaging. AJNR Am J Neuroradiol 2012;33:896-903. 20. Hefzy HM, Bartynski WS, Boardman JF, et al. Hemorrhage in posterior reversible encephalopathy syndrome: Imaging and clinical features. AJNR Am J Neuroradiol 2009;30:1371-1379. 21. Segal AZ, Chiu RI, Eggleston-Sexton PM, et al. Low cholesterol as a risk factor for primary intracerebral hemorrhage: A case-control study. Neuroepidemiology 1999;18:185-193. 22. Amarenco P, Bogousslavsky J, Callahan A 3rd, et al. High-dose atorvastatin after stroke or transient ischemic attack. N Engl J Med 2006;355:549-559.