ORIGINAL ARTICLE. Carotid Atherosclerosis and Progression of Brain Atrophy: The SMART-MR Study

Transcription

1 ORIGINAL ARTICLE Carotid Atherosclerosis and Progression of Brain Atrophy: The SMART-MR Study Majon Muller, MD, PhD, 1,2 Yolanda van der Graaf, MD, PhD, 2 Ale Algra, MD, PhD, 2,3 Jeroen Hendrikse, MD, PhD, 4 Willem P. Mali, MD, PhD, 4 and Mirjam I. Geerlings, PhD 2 ; for the SMART Study Group Objective: Atherosclerosis has been implicated in the development of brain atrophy. However, support for this association comes from cross-sectional studies. Methods: Within the Second Manifestations of ARTerial disease-magnetic Resonance (SMART-MR) study, a prospective cohort study among patients with symptomatic atherosclerotic disease (mean age 6 standard deviation, years; 80% men), magnetic resonance imaging of the brain was performed in 1,232 patients at baseline ( ) and in 663 patients at follow-up ( ). Brain segmentation was used to quantify total brain volume, cortical gray matter volume, and ventricular volume as indicators of global, cortical, and subcortical atrophy. At baseline, measurements of carotid intima media thickness (CIMT) and carotid stenosis were performed. Carotid stenosis was classified into groups 0 of 50%, 50 of 70% (moderate), and >70% (severe) and into unilateral or bilateral stenosis. Results: Cross-sectional regression analyses showed that both increased CIMT and carotid stenosis were associated with decreased relative total brain and cortical gray matter volume. Our prospective findings showed that after a mean follow-up of 3.9 years (range, years), CIMT and moderate stenosis were not related to progression of brain atrophy. Only severe or bilateral carotid stenosis was related to progression of global atrophy (b [95% confidence interval (CI)], 0.52% [ 0.84 to 0.20%], 0.94% [ 1.45 to 0.43%]), cortical atrophy (b [95% CI], 0.75% [ 1.37 to 0.13%], 1.34% [ 2.32 to 0.35%]), and subcortical atrophy (b [95% CI], 0.06% [ 0.02 to 0.16%], 0.13% [0.01 to 0.28%]). Interpretation: In a study of patients with atherosclerotic disease with 4 years of follow-up, only severe or bilateral carotid stenosis, and not moderate carotid stenosis and increased CIMT, were associated with progression of brain atrophy. ANN NEUROL 2011;70: Over the past decade, evidence has accumulated that vascular risk factors and vascular disease play an important role in the etiology of dementia and cognitive impairment. 1 Evidence exists that atherosclerotic processes underlie the association of vascular risk factors and dementia. Previous studies have shown that carotid atherosclerosis is associated with dementia and its subclinical markers. 2 8 This association may be mediated by cerebrovascular disease, such as embolic stroke or cerebral small-vessel disease. 7,9,10 The relation between carotid atherosclerosis and dementia could also be explained by hemodynamic changes in the cerebral circulation. Recent studies have shown that patients with carotid atherosclerosis have an increased risk of cerebral hypoperfusion, 11,12 which subsequently may cause brain atrophy and dementia Limited evidence exists on the direct relation between carotid atherosclerosis and brain atrophy. Cross-sectional studies have shown that increased carotid intima media thickness (CIMT) and carotid stenosis are associated with decreased total brain volume 6,8 as well as with ventricular and sulcal widening. 7 However, only 1 cross-sectional study used quantitative brain volume measurements, 6 and no studies examined the longitudinal association of carotid atherosclerosis with progression of brain atrophy. View this article online at wileyonlinelibrary.com. DOI: /ana Received Sep 15, 2010, and in revised form Jan 3, Accepted for publication Jan 28, Address correspondence to Dr Geerlings, University Medical Center Utrecht, Julius Center for Health Sciences and Primary Care, Stratenum 6.131, PO Box 85500, 3508 GA Utrecht, the Netherlands. m.geerlings@umcutrecht.nl Members of the SMART Study Group are listed in the Appendix on page xxx. From the 1 Department of Internal Medicine, VU University Medical Center, Amsterdam; 2 Julius Center for Health Sciences and Primary Care, Utrecht; 3 Department of Neurology and Neurosurgery, Rudolf Magnus Institute of Neuroscience, Utrecht; and 4 Department of Radiology, University Medical Center Utrecht, Utrecht, the Netherlands. VC 2011 American Neurological Association 237

2 ANNALS of Neurology The objective of this study was to investigate whether presence and severity of carotid atherosclerosis was related to presence and progression of global, cortical, and subcortical atrophy using quantitative brain volume measurements in individuals with symptomatic atherosclerotic disease. Patients and Methods SMART-MR Study Data were used from the Second Manifestations of ARTerial disease-magnetic Resonance (SMART-MR) study, a prospective cohort study aimed at investigating brain changes on magnetic resonance imaging (MRI) in 1,309 independently living patients with symptomatic atherosclerotic disease. Details of the design and participants have been described elsewhere. 16,17 In brief, between May 2001 and December 2005, all patients newly referred to the University Medical Center Utrecht with manifest coronary artery disease, cerebrovascular disease, peripheral arterial disease, or an abdominal aortic aneurysm, and without MR contraindications, were invited to participate. Excluded were patients with a terminal disease, those not independent in daily activities, and those referred back to the referring specialist immediately after 1 visit. During a 1-day visit to our medical center, an MRI of the brain was performed, in addition to a physical examination, ultrasonography of the carotid arteries, and blood sampling. Risk factors, medical history, and functioning were assessed with questionnaires that the patients completed before their visit to the medical center. Between January 2006 and May 2009, all participants still alive (n ¼ 1,238) were invited for follow-up measurements, including MRI of the brain, a physical examination, blood sampling, risk factors, and medical history. In total, 754 of the surviving cohort (61%) gave written informed consent and participated. The SMART-MR study was approved by the ethics committee of our institution, and written informed consent was obtained from all participants. Carotid Atherosclerosis Presence of atherosclerosis in the carotid arteries was assessed at baseline by measuring CIMT and carotid artery stenosis. Ultrasonography was performed with a 10MHz linear-array transducer (ATL Ultramark 9) by well-trained and certified ultrasound technicians at the Department of Radiology, University Medical Center Utrecht. Mean CIMT (in millimeters) was calculated for each patient based on 6 far-wall measurements of the left and right common carotid arteries as previously described. 18 The degree of the carotid artery stenosis at both sides was assessed with color Doppler-assisted duplex scanning. The severity of carotid artery stenosis was evaluated on the basis of blood flow velocity patterns. 19 The greatest stenosis observed on the right or the left side of the common or internal carotid artery was taken to determine the severity of carotid artery disease. Carotid artery stenosis 50% was defined as peak systolic velocity >150 cm/s, and a carotid artery stenosis 70% was defined as peak systolic velocity >210cm/s. 19 We classified carotid stenosis into groups of 0 to 50% (no or limited), 50 to 70% (moderate), and >70% (severe). In addition, we classified carotid stenosis >50% into unilateral or bilateral stenosis. Magnetic Resonance Protocol and Brain Segmentation At baseline and follow-up, the MR investigations were performed on a 1.5T whole-body system (Gyroscan ACS-NT, Philips Medical Systems, Best, the Netherlands). The protocol consisted of transversal T1-weighted (repetition time [TR]/ echo time [TE], 235/2 milliseconds), T2-weighted (TR/TE, 2,200/11 milliseconds and 2,200/100 milliseconds), fast fluid-attenuated inversion recovery (FLAIR; TR/TE/inversion time [TI], 6000/100/2,000 milliseconds), and inversion recovery (IR; TR/TE/TI, 2,900/22/410 milliseconds) sequences (field of view [FOV], mm; matrix size, ; slice thickness, 4.0mm; no gap; 38 slices). 16 Next, on the basis of a localizer MR angiographic slab in the sagittal plane, a 2-dimensional (2D) phase-contrast section was positioned at the level of the skull base to measure the volume flow in the internal carotid arteries (ICAs) and the basilar artery (BA). 14 The 2D phase-contrast section was positioned through the ICAs and the BA (TR/TE, 16/9 milliseconds; flip angle, 7.5 ; FOV, mm; matrix size, ; slice thickness, 5.0mm; 8 acquired signals; velocity sensitivity, 100cm/s). We used the T1-weighted gradient-echo, IR sequence and FLAIR sequence for brain segmentation. The probabilistic segmentation technique has been described elsewhere, and has been proven to be very reliable, with similarity indices >0.8 for all segmented tissue and cerebrospinal fluid (CSF) volumes. 20,21 The segmentation program distinguishes cortical gray matter, white matter, sulcal and ventricular CSF, and lesions. The results of the segmentation analysis were visually checked for the presence of infarcts and adapted if necessary to make a distinction between white matter lesion (WML) and infarct volume. Total brain volume was calculated by summing the volumes of gray matter, white matter, WMLs, and infarcts. Total intracranial volume (ICV) was calculated by summing the total brain volume and the volumes of the sulcal and ventricular CSF. In 188 patients, the IR and T1-weighted sequences were missing due to a temporary change in MRI protocol, and the brain segmentation in these patients was based on the FLAIR sequence. Intraclass correlation coefficients between the segmentation using all 3 sequences and FLAIR only based on a subset of 740 patients were 0.995, 0.996, 0.961, 0.996, and for ICV, total brain volume, CSF, ventricular volume, and WML volume, respectively. Brain Atrophy The brain volumes that were used for this analysis were total brain volume, cortical gray matter volume, and ventricular volume as indicators of global, cortical, and subcortical atrophy. All brain volumes were normalized for ICV Volume 70, No. 2

3 Muller et al: Brain Atrophy Brain Infarcts, WMLs, Cerebral Blood Flow At baseline and follow-up, the whole brain was visually searched for infarcts by an investigator and a neuroradiologist. Discrepancies in rating were re-evaluated in a consensus meeting. Raters were blinded regarding the history and diagnosis of the patient. Infarcts were defined as focal hyperintensities on T2-weighted images of at least 3mm in diameter. Hyperintensities located in the white matter also had to be hypointense on T1-weighted and FLAIR images to distinguish them from WMLs. Dilated perivascular spaces were distinguished from infarcts on the basis of their location, form, and the absence of gliosis. The location, affected flow territory, and type were scored for every infarct. Brain infarcts were categorized as cortical infarcts, large subcortical infarcts, infarcts in the cerebellum and brainstem, and lacunar infarcts. We defined lacunar infarcts as infarcts sized 3 to 15mm in diameter and located in the subcortical white matter, thalamus, or basal ganglia. Large subcortical infarcts had the same characteristics as lacunar infarcts, but were sized >15mm and were located exclusively in subcortical areas. Volumes of WMLs obtained with the segmentation program consisted of deep and periventricular WML volumes and were summed to obtain the total volume of WMLs. WML volumes were normalized for ICV. 22 Postprocessing of the flow measurements was performed by 1 investigator. 14,23 The flow through the left and right ICAs and BA were summed to calculate the total cerebral blood flow (tcbf; ml/min). We expressed tcbf per 100ml brain parenchymal volume to obtain a measure of cerebral perfusion (pcbf). 23 Vascular Risk Factors During the patient s baseline visit to the medical center, height and weight were measured without shoes and heavy clothing, and the body mass index (BMI) was calculated (kg/m 2 ). Smoking habits and alcohol intake were assessed with questionnaires. Pack years of smoking were calculated, and alcohol intake was categorized as never, former, or current. Systolic blood pressure (SBP; mmhg) and diastolic blood pressure (DBP) were measured twice with a sphygmomanometer, and the average of the 2 measures was calculated. Hypertension was defined as mean SBP 160mmHg, mean DBP 95mmHg, or self-reported use of antihypertensive drugs. An overnight fasting venous blood sample was taken to determine glucose and lipid levels. Diabetes mellitus was defined as use of glucose-lowering agents, a known history of diabetes or a fasting plasma glucose level 7.0. Hyperlipidemia was defined as total cholesterol >5.0mmol/l, low-density lipoprotein cholesterol >3.2mmol/l, or use of lipid-lowering drugs. Study Sample Of the 1,309 patients, 19 had no MRI, and 14 had no FLAIR sequence. In addition, in 44 patients, brain volume data were missing due to motion or artifacts. As a result, baseline MRI data were available in 1,232 patients. Of the 1,232 patients, 718 patients participated in the follow-up exam. Of these, 38 had no MRI, and in 17 patients brain volume data were missing due to motion or artifacts. As a result, the analyses for the prospective analysis were performed in 663 patients. Data Analysis Missing data rarely occur completely at random, and a complete case analysis (deletion of all participants with 1 or more missing values) leads to loss of statistical power and to biased results. 24,25 We therefore used multiple imputation (10 datasets) to address the missing values, 26 using the statistical program R (aregimpute; version ). Data were analyzed using SPSS version 17.0 (SPSS Inc., Chicago, IL), by pooling the 10 inputted datasets. Subject characteristics were calculated for the baseline SMART-MR cohort (n ¼ 1,232) and for the SMART-MR sample with follow-up measurements (n ¼ 663). Linear regression analysis was used to investigate the cross-sectional association of CIMT and carotid stenosis with baseline relative total brain, cortical gray matter, and ventricular volume. In addition, linear regression analysis and analysis of covariance were used to investigate the prospective association of baseline carotid atherosclerosis with change in measures of brain volume by using brain volume measurements at followup as dependent variable and brain volume measurements at baseline as independent variable. Cross-sectional and prospective analyses were adjusted for age, sex, and follow-up time (model 1). To examine if the relation between carotid atherosclerosis and brain volume was explained by vascular risk factors, we further adjusted for smoking, alcohol consumption, BMI, hypertension, diabetes, and hyperlipidemia (model 2). As presence of large brain infarcts is a strong confounder in the association of carotid atherosclerosis and brain volume, the abovementioned analyses were repeated after exclusion of patients with a history of cerebrovascular disease or with 1 or more nonlacunar infarcts on MRI. Finally, to investigate whether reduced pcbf and presence of WMLs and lacunar infarcts mediate the association of carotid atherosclerosis with change in brain volume, we additionally adjusted the analyses for pcbf or for WMLs and lacunar infarcts. Results Mean (standard deviation [SD]) age of the total SMART-MR population (N ¼ 1,232) was 58 (10) years. Compared with the total SMART-MR population, patients with follow-up measurements (n ¼ 663) were younger at baseline, had hypertension and diabetes less often, had smaller CIMT, and less often had severe carotid stenosis (Table 1). The percentages of missing variables varied between 0 and 15%. Mean (SD) baseline total brain, cortical gray matter, and ventricular volumes for the total SMART-MR population were 79 (3), 36 (3), and 2.1 (1.0) %ICV and for the patients with follow-up were 79 (3), 36 (3), and 2.0 (0.8) %ICV. Cross-sectional linear regression analyses, adjusting for age and sex, showed that an increase in CIMT of 1 SD (0.31mm) and presence of moderate, severe, or August

4 ANNALS of Neurology TABLE 1: Baseline Patient Characteristics of the Total SMART-MR Population and of the SMART-MR Sample with Follow-up Measurements Characteristic Total SMART-MR Sample, N 5 1,232 SMART-MR Sample with Follow-up, n General characteristics Age, yr a 58 (10, 25 82) 57 (9, 28 79) Sex, No. (%) men 983 (80%) 541 (81%) Vascular disease category, No. (%) Coronary heart disease 735 (60%) 412 (62%) Peripheral arterial disease 268 (22%) 120 (18%) Cerebrovascular disease 284 (23%) 153 (23%) Abdominal aortic aneurysm 108 (9%) 39 (6%) Vascular risk factors Smoking (pack years) b 19 (0 50) 20 (0 50) Alcohol use, No. (%) current 924 (75%) 515 (77%) BMI, kg/m 2a 27 (4, 15 43) 27 (4, 18 39) Hypertension, No. (%) 640 (52%) 312 (47%) Diabetes, No. (%) 253 (21%) 106 (16%) Hyperlipidemia, No. (%) 968 (78%) 522 (79%) Carotid atherosclerosis Intima media thickness, mm a 0.94 (0.31, ) 0.92 (0.29, ) Carotid stenosis 50 to 70%, No. (%) 45 (4%) 24 (4%) Carotid stenosis >70 to 99%, No. (%) 83 (7%) 31 (5%) Carotid stenosis 100%, No. (%) 52 (4%) 32 (5%) Bilateral carotid stenosis >50%, No. (%) 52 (4%) 23 (3%) Cerebrovascular pathology White matter lesion volume, %ICV b 0.11 ( ) 0.09 ( ) Cortical infarcts 1, No. (%) 144 (12%) 74 (11%) Subcortical infarcts 1, No. (%) 11 (1%) 6 (1%) Cerebellar infarcts 1, No. (%) 48 (4%) 26 (4%) Brain stem infarcts 1, No. (%) 35 (3%) 17 (3%) Lacunar infarcts 1, No. (%) 232 (19%) 114 (17%) Parenchymal CBF, ml/min/100ml a 51 (11, 9 92) 52 (10, 19 91) a Mean (standard deviation, range). Median (10th 90th percentile). BMI ¼ body mass index; CBF ¼ cerebral blood flow. bilateral carotid stenosis were associated with smaller total brain and cortical gray matter volume (Table 2). CIMT was also associated with larger ventricular volume. These associations were independent of cardiovascular risk factors. When repeating the cross-sectional analyses in the 663 patients with follow-up data, the results presented in Table 2 were similar (data not shown). After a mean (SD; range) follow-up of 3.9 (0.4; ) years, CIMT, moderate carotid stenosis, and unilateral stenosis were not related to change in brain volume (Fig). Only severe or bilateral carotid stenosis were related to a decrease in total brain and cortical gray matter volume and an increase in ventricular volume. Compared with patients with no or limited carotid stenosis, the differences (b, 95% 240 Volume 70, No. 2

5 Muller et al: Brain Atrophy TABLE 2: Cross-Sectional Relation between Baseline Carotid Atherosclerosis Measures and Baseline Brain Volume in 1,232 Patients Model a Total Brain Volume, % ICV Cortical Gray Matter Volume, % ICV Ventricular Volume, % ICV b b 95% CI b b 95% CI b b 95% CI CIMT, per SD to 0.17 d to 0.14 d to 0.12 e increase c to 0.08 d to 0.03 e to 0.11 e Carotid stenosis 0 to 50%, reference Moderate, to to 0.17 e to to 70% Severe, >70% to 0.51 d to 0.84 d to to 50%, reference Moderate, to to to 0.33 to 70% Severe, >70% to 0.32 d to 0.64 d to 0.26 Carotid stenosis 0 to 50%, reference Unilateral >50% to 0.35 d to 0.73 d to 0.30 Bilateral >50% to 0.32 d to 0.52 d to to 50%, reference Unilateral >50% to 0.17 e to 0.52 d to 0.26 Bilateral >50% to 0.12 e to 0.28 e to 0.34 a Model 1: adjusted for age and sex; model 2: additionally adjusted for smoking, alcohol consumption, body mass index, hypertension, diabetes, and hyperlipidemia. b b ¼ coefficient of linear regression; an increase of the independent variable by 1 unit is associated with a b increase of brain volume. c SD ¼ 0.31mm. d p < e p < ICV ¼ intracranial volume; CI ¼ confidence interval; CIMT ¼ carotid intima media thickness; SD ¼ standard deviation. confidence interval [CI]) in change in total brain, cortical gray matter, and ventricular volume adjusted for age, sex, smoking, alcohol consumption, BMI, hypertension, diabetes, hyperlipidemia, and follow-up period were 0.52 ( 0.84 to 0.20), 0.75 ( 1.37 to 0.13), and 0.06 ( 0.02 to 0.16) %ICV for patients with severe stenosis and 0.94 ( 1.45 to 0.43), 1.34 ( 2.32 to 0.35), and 0.13 (0.01 to 0.28) %ICV for patients with bilateral stenosis. Exclusion of patients with total occlusion of the carotid artery (n ¼ 32) did not change our findings (data not shown). After exclusion of patients with nonlacunar infarcts on MRI (n ¼ 80) or patients with cerebrovascular disease (n ¼ 153), the associations of carotid stenosis with brain volumes did not change (data not shown). To examine if reduced pcbf mediated the association of severe or bilateral carotid stenosis with progression of brain atrophy, we additionally adjusted the regression analyses for pcbf. The association of severe or bilateral stenosis with a decrease in total brain and cortical gray matter volume and an increase in ventricular volume were partly explained by reduced pcbf; bs (95% CI) for severe carotid stenosis were 0.47 ( 0.72 to 0.21), 0.46 ( 1.05 to 0.12), and 0.06 (0.00 to 0.13) %ICV and for bilateral stenosis were 0.92 ( 1.33 to 0.50), 0.78 ( 1.80 to 0.23), and 0.15 (0.03 to 0.27) %ICV. Adjustment for WMLs and lacunar infarcts further attenuated the association of severe or bilateral carotid stenosis with change in brain volumes; bs (95% August

6 ANNALS of Neurology FIGURE: Longitudinal relation between measures of carotid atherosclerosis and change in total brain volume (A), cortical gray matter volume (B), and ventricular volume (C). Bars represent the mean (standard error) change in brain volume after 3.9 years of follow-up. Carotid intima media thickness (CIMT) tertile 1, 0.47 to 0.78mm; tertile 2, 0.78 to 0.97mm; tertile 3, 0.97 to 4.02mm. *p < 0.05; **p < ICV 5 intracranial volume. CI) for severe carotid stenosis were 0.42 ( 0.67 to 0.16), 0.36 ( 0.95 to 0.22), and 0.05 ( 0.02 to 0.12) %ICV and for bilateral stenosis were 0.84 ( 1.26 to 0.41), 0.67 ( 1.68 to 0.35), and 0.14 (0.01 to 0.26) %ICV. Finally, post hoc analysis excluding patients with suspected cognitive impairment (Mini Mental State Examination <24) did not change the results (data not shown). Discussion In a population with symptomatic atherosclerotic disease, we observed that increased CIMT and carotid stenosis were significantly associated with decreased relative total brain and cortical gray matter volume. However, our prospective findings showed that only severe or bilateral carotid stenosis, but not increased CIMT or moderate stenosis, were related to progression of global, cortical, and subcortical atrophy over a 4-year period of patient follow-up. This relation was independent of age, sex, and vascular risk factors, such as hypertension, hyperlipidemia, and diabetes, and was partly explained by reduced cerebral perfusion and cerebral small-vessel disease. Our cross-sectional findings are in agreement with previous studies such as the Cardiovascular Health Study and the Framingham Study. 6,7 These large-scale population-based studies reported that increased CIMT as well as moderate and severe carotid stenosis were associated with sulcal and ventricular widening 7 and with smaller total brain volume. 6 To our knowledge, this is the first study that also examined the prospective relation of measures of carotid atherosclerosis with progression of brain atrophy. An explanation of the disparity between the cross-sectional and prospective findings could be that carotid stenosis and CIMT reflect different stages and severity of the atherosclerotic process. 18,19 Carotid stenosis and plaques are irreversible focal manifestations of atherosclerosis, whereas increased CIMT represents mainly hypertensive medial hypertrophy. 27 In fact, studies have shown that CIMT could be reduced by antihypertensive and lipid-lowering treatment. 28,29 However, because CIMT has proven to be a strong predictor of future cardiovascular events, 30 it is unlikely that treatment could have explained the discrepancies between the cross-sectional and prospective findings. It is, however, conceivable that compared with severe or bilateral carotid stenosis, increased CIMT or moderate stenosis may require a longer follow-up time than 4 years to achieve similar effects on progression of brain atrophy. The disparity between the cross-sectional and prospective findings could not be explained by a healthy survivor effect, because the cross-sectional results were similar within the 663 patients with complete follow-up. Another explanation for the cross-sectional results could be that carotid atherosclerosis and brain atrophy are independent processes with common elements in the causal pathway, because many of the contributory factors in atherosclerosis have emerged as potential contributors in dementia. 31 Common pathophysiological elements could include hypertension, hyperlipidemia, inflammation, or genetic factors. 31 Although adjustment for cardiovascular risk 242 Volume 70, No. 2

7 Muller et al: Brain Atrophy factors did not change the observed associations between carotid atherosclerosis and brain atrophy, this explanation cannot be excluded. Several mechanisms may explain an association between severe or bilateral carotid stenosis and progression of brain atrophy. Cerebrovascular disease may mediate the association between atherosclerosis and atrophy. Because excluding individuals with large brain infarcts on MRI did not affect the estimate, it is not likely that the association with brain atrophy is explained by cortical or large subcortical infarcts. However, adjusting the analyses for presence of cerebral small-vessel disease, such as WMLs and lacunar infarcts, partly mediated the association between atherosclerosis and progression of brain atrophy. Also, severe or bilateral stenosis may induce cerebral hypoperfusion leading to cerebral hypoxia especially in individuals with reduced or impaired collateral circulation. 32 A good collateral circulation can generate normal cerebral hemodynamics in patients with carotid stenosis. When collaterals are not adequate, the perfusion pressure distal to the lesion begins to fall, which might trigger regional brain microcirculatory disturbances that can evolve into a neurodegenerative process, eventually leading to brain atrophy. 13,32 The cortical gray matter is especially vulnerable to hypoperfusion, because the gray matter has higher oxygen and glucose requirements than the white matter. 33 This is confirmed by our findings that adjusting for pcbf, as an indicator of cerebral perfusion, only attenuated the association of carotid stenosis with progression of cortical atrophy. Future studies are needed, using measures of flow in the circle of Willis and cerebral perfusion at brain tissue level, to further elucidate the relation between carotid artery disease, collateral circulation, cerebral hypoperfusion, and neurodegeneration. 34 Strengths of our study are the large number of patients investigated and the volumetric assessment of measures of brain atrophy, which made it possible to obtain precise estimates of progression of brain atrophy, and resulted in a large power to detect associations. Volumetric segmentation methods have several advantages compared with visual rating and 2D assessment of brain atrophy. Contrary to visual rating scales, volumetric assessments are not impeded by ceiling effects. Furthermore, volumetric segmentation methods are more precise, and therefore more likely to detect small differences in brain volumes that cannot be detected by visual rating and 2D techniques. In addition, the segmentation of different brain tissue types and CSF spaces allowed us to differentiate between cortical and subcortical brain atrophy. Finally, the extensive information on markers of cerebrovascular pathology and cardiovascular risk factors allowed us to investigate whether the association between carotid atherosclerosis and brain atrophy was independent of these possible confounders or modifiers. Our interpretation of the results may be limited by a few factors. First, individuals who participated in the follow-up examination represent a relatively healthy group. If anything, however, this may have led to an underestimation of the true association. Second, because the majority of our study sample consisted of male patients with atherosclerotic disease, we do not know to what extent our results also apply to women or can be generalized to the general population. Although we adjusted for vascular risk and disease, our population consisted of patients with manifest arterial disease, and it is thus still possible that other factors than atherosclerosis may explain our findings, such as heart failure or hypoxia, which may have occurred during myocardial infarction or cardiovascular interventions. Nonetheless, our cross-sectional findings were similar to those found in populationbased samples. 6,7 In summary, in this prospective study in patients with symptomatic atherosclerotic disease, only high grade or bilateral carotid artery narrowing led to progression of brain atrophy. Acknowledgment Supported by a program grant from the Netherlands Organization for Scientific Research-Medical Sciences (NWO-MW: project No ) and a grant from the Netherlands Organization for Scientific Research (NWO: project No ; M.I.G.). Potential Conflicts of Interest A.A.: grants, Netherlands Heart Foundation, Netherlands Brain Foundation, Thrombosis Foundation Holland, Netherlands Organization for Scientific Research, and Netherlands Organization for Health Research and Development; consulting fee/honorarium, Boehringer Ingelheim; other, a principal investigator of ESPRIT, the European/Australian Stroke Prevention in Reversible Ischemia Trial, a trial that was run independently of any pharmaceutical company; after completion and full analysis of ESPRIT, the study group accepted financial support from Boehringer Ingelheim for post hoc exploratory analyses of the ESPRIT trial data; for this purpose a contract was signed ensuring complete scientific freedom. Appendix We thank the members of the SMART Study Group of University Medical Center Utrecht: A. Algra, MD, PhD, Julius Center for Health Sciences and Primary Care, Rudolf Magnus Institute for Neurosciences, Department August

8 ANNALS of Neurology of Neurology; P. A. Doevendans, MD, PhD, Department of Cardiology; Y. van der Graaf, MD, PhD, D. E. Grobbee, MD, PhD, and G. E. H. M. Rutten, MD, PhD, Julius Center for Health Sciences and Primary Care; L. J. Kappelle, MD, PhD, Department of Neurology; W. P. T. M. Mali, MD, PhD, Department of Radiology; F. L. Moll, MD, PhD, Department of Vascular Surgery; and F. L. J. Visseren, MD, PhD, Department of Vascular Medicine. References 1. Viswanathan A, Rocca WA, Tzourio C. Vascular risk factors and dementia: how to move forward? Neurology 2009;72: Johnston SC, O Meara ES, Manolio TA, et al. Cognitive impairment and decline are associated with carotid artery disease in patients without clinically evident cerebrovascular disease. Ann Intern Med 2004;140: Muller M, Grobbee DE, Aleman A, et al. Cardiovascular disease and cognitive performance in middle-aged and elderly men. Atherosclerosis 2007;190: Rice Wendell C, Zonderman AB, Metter EJ, et al. Carotid intimal medial thickness predicts cognitive decline among adults without clinical vascular disease. Stroke 2009;40: van Oijen M, de Jong FJ, Witteman JC, et al. Atherosclerosis and risk for dementia. Ann Neurol 2007;61: Romero JR, Beiser A, Seshadri S, et al. Carotid artery atherosclerosis, MRI indices of brain ischemia, aging, and cognitive impairment: the Framingham study. Stroke 2009;40: Manolio TA, Burke GL, O Leary DH, et al. Relationships of cerebral MRI findings to ultrasonographic carotid atherosclerosis in older adults: the Cardiovascular Health Study. CHS Collaborative Research Group. Arterioscler Thromb Vasc Biol 1999;19: Kin T, Yamano S, Sakurai R, et al. Carotid atherosclerosis is associated with brain atrophy in Japanese elders. Gerontology 2007;53: Barnett HJ, Gunton RW, Eliasziw M, et al. Causes and severity of ischemic stroke in patients with internal carotid artery stenosis. JAMA 2000;283: Pico F, Dufouil C, Levy C, et al. Longitudinal study of carotid atherosclerosis and white matter hyperintensities: the EVA-MRI cohort. Cerebrovasc Dis 2002;14: Sojkova J, Najjar SS, Beason-Held LL, et al. Intima-media thickness and regional cerebral blood flow in older adults. Stroke 2010;41: Bokkers RP, van der Worp HB, Mali WP, et al. Noninvasive MR imaging of cerebral perfusion in patients with a carotid artery stenosis. Neurology 2009;73: de la Torre JC. Critically attained threshold of cerebral hypoperfusion: the CATCH hypothesis of Alzheimer s pathogenesis. Neurobiol Aging 2000;21: Muller M, van der Graaf Y, Visseren FL, et al. Blood pressure, cerebral blood flow, and brain volumes. The SMART-MR study. J Hypertens 2010;28: Poels MM, Ikram MA, Vernooij MW, et al. Total cerebral blood flow in relation to cognitive function: the Rotterdam Scan Study. J Cereb Blood Flow Metab 2008;28: Geerlings MI, Appelman AP, Vincken KL, et al. Brain volumes and cerebrovascular lesions on MRI in patients with atherosclerotic disease. The SMART-MR study. Atherosclerosis 2010;210: Simons PC, Algra A, van de Laak MF, et al. Second manifestations of ARTerial disease (SMART) study: rationale and design. Eur J Epidemiol 1999;15: Simons PC, Algra A, Bots ML, et al. Common carotid intimamedia thickness and arterial stiffness: indicators of cardiovascular risk in high-risk patients. The SMART study (Second Manifestations of ARTerial disease). Circulation 1999;100: Elgersma OE, van LM, Buijs PC, et al. Changes over time in optimal duplex threshold for the identification of patients eligible for carotid endarterectomy. Stroke 1998;29: Anbeek P, Vincken KL, van Bochove GS, et al. Probabilistic segmentation of brain tissue in MR imaging. Neuroimage 2005;27: Muller M, Appelman AP, van der Graaf Y, et al. Brain atrophy and cognition: interaction with cerebrovascular pathology? Neurobiol Aging 2009; June 9 [Epub ahead of print]. 22. Hachinski V, Iadecola C, Petersen RC, et al. National Institute of Neurological Disorders and Stroke-Canadian Stroke Network vascular cognitive impairment harmonization standards. Stroke 2006; 37: Appelman AP, van der Graaf Y, Vincken KL, et al. Total cerebral blood flow, white matter lesions and brain atrophy: the SMART- MR study. J Cereb Blood Flow Metab 2008;28: van der Heijden GJ, Donders AR, Stijnen T, et al. Imputation of missing values is superior to complete case analysis and the missing-indicator method in multivariable diagnostic research: a clinical example. J Clin Epidemiol 2006;59: Donders AR, van der Heijden GJ, Stijnen T, et al. Review: a gentle introduction to imputation of missing values. J Clin Epidemiol 2006;59: Rubin DB, Schenker N. Multiple imputation in health-care databases: an overview and some applications. Stat Med 1991;10: Spence JD. Measurement of intima-media thickness vs. carotid plaque: uses in patient care, genetic research and evaluation of new therapies. Int J Stroke 2006;1: Kang S, Wu Y, Li X. Effects of statin therapy on the progression of carotid atherosclerosis: a systematic review and meta-analysis. Atherosclerosis 2004;177: Wang JG, Staessen JA, Li Y, et al. Carotid intima-media thickness and antihypertensive treatment: a meta-analysis of randomized controlled trials. Stroke 2006;37: Lorenz MW, Markus HS, Bots ML, et al. Prediction of clinical cardiovascular events with carotid intima-media thickness: a systematic review and meta-analysis. Circulation 2007;115: Casserly I, Topol E. Convergence of atherosclerosis and Alzheimer s disease: inflammation, cholesterol, and misfolded proteins. Lancet 2004;363: Vernieri F, Pasqualetti P, Matteis M, et al. Effect of collateral blood flow and cerebral vasomotor reactivity on the outcome of carotid artery occlusion. Stroke 2001;32: Kirkness CJ. Cerebral blood flow monitoring in clinical practice. AACN Clin Issues 2005;16: van Laar PJ, van der Grond J, Mali WP, et al. Magnetic resonance evaluation of the cerebral circulation in obstructive arterial disease. Cerebrovasc Dis 2006;21: Volume 70, No. 2