Cerebral autosomal-dominant arteriopathy with subcortical

Transcription

1 Extensive White Matter Hyperintensities May Increase Brain Volume in Cerebral Autosomal-Dominant Arteriopathy With Subcortical Infarcts and Leukoencephalopathy Ming Yao, MD; Eric Jouvent, MD, PhD; Marco During, MD; Ophélia Godin, PhD; Dominique Hervé, MD; Jean Pierre Guichard, MD; Yi-Cheng Zhu, MD, PhD; Andreas Gschwendtner, MD; Christian Opherk, MD; Martin Dichgans, MD, PhD; Hugues Chabriat, MD, PhD Background and Purpose The extent of white matter hyperintensities (WMH) is associated with cerebral atrophy in elderly people. WMH is a radiological hallmark of cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), but their relationship with brain volume remains poorly understood. The association between WMH and brain volume was analyzed in a large population of patients with CADASIL. Methods Demographic and MRI data of 278 patients recruited from a prospective cohort study were analyzed. Volumes of WMH and lacunar infarcts, number of cerebral microbleeds, and brain parenchymal fraction were measured. Multivariate analysis was used to study the impact of WMH on brain volume at baseline. Results In univariate analyses, brain parenchymal fraction was negatively associated with age, male sex, and all MRI markers. Multiple regression modeling showed that brain parenchymal fraction was inversely related to age, number of cerebral microbleeds, and normalized volume of lacunar infarcts but positively related to normalized volume of WMH (P<0.001). This positive relationship was independent of the presence/absence of lacunar infarcts or of cerebral microbleeds. Subgroup analysis showed that this association was significant in subjects having normalized volume of WMH 6.13 or brain parenchymal fraction 86.37% (median values, both P 0.001). Conclusion The results of the present study suggest that extensive WMH may be associated with increase of brain volume in CADASIL. In this disorder, WMH may be related not only to loss of white matter components, but also to a global increase of water content in the cerebral tissue. (Stroke. 2012;43: ) Key Words: brain tissue volume CADASIL MRI white matter hyperintensities Cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is the most frequent monogenic type of cerebral microangiopathy and is secondary to different mutations of the Notch3 gene. 1 The main clinical manifestations of the disease include attacks of migraine with aura, mood disturbances, recurrent ischemic stroke, apathy, and progressive cognitive decline. 2 MRI features in CADASIL include confluent white matter hyperintensities (WMH), lacunar infarctions, and cerebral microbleeds. 3 5 Patients with CADASIL can also present with a large number of dilated Virchow-Robin spaces, mostly located in the lentiform nuclei and within the subcortical white matter. 6 Extensive WMH are observed in the centrum semiovale and periventricular areas in all symptomatic patients. They are also detected in more than two thirds of patients with CADASIL in the anterior part of the temporal lobes. It is noteworthy that WMH in this region are associated with punctiform hypointensities on fluid-attenuated inversion recovery images mainly at the frontiers between the cortical ribbon and the underlying white-matter. 5 This highly specific aspect is presumably related to the accumulation of dilated Virchow-Robin spaces. 7 Postmortem data recently showed that dilated Virchow-Robin spaces and degeneration of myelin are the 2 main pathological lesions underlying WMH in the temporal lobes. 7 These lesions are possibly related to insufficient drainage of the interstitial fluid accumulating in the cerebral tissue secondary to blood brain barrier dysfunction and vessel wall changes. In other cerebral regions where focal ischemic lesions are detected, axonal loss with demyelination is the most prominent feature within WMH and is associated Received May 17, 2012; final revision received September 13, 2012; accepted September 19, From the University Paris Diderot, Paris, France (E.J., D.H., J.P.G., H.C.); INSERM UMR 740 (M.Y., H.C.) and UMR 708 F (O.G.), Paris, France; the Department of Neurology AP-HP, Lariboisière Hospital, Paris, France (E.J., D.H., H.C.); the Institute for Stroke and Dementia Research, Medical Centre, Ludwig-Maximilians-University, Munich, Germany (M.D., A.G., C.O., M.D.); the German Center for Neurodegenerative Diseases (DZNE, Munich), Munich, Germany (M.D., A.G., C.O., M.D.); the Department of Neuroradiology, Lariboisière Hospital, Paris, France (J.P.G.); and the Department of Neurology, Peking Union Medical College Hospital, Peking, China (M.Y., Y.-C.Z.). The online-only Data Supplement is available with this article at /-/DC1. Correspondence to Hugues Chabriat, MD, PhD, Service de Neurologie, Hôpital Lariboisière, 2 rue Ambroise Paré, Paris, France. hugues.chabriat@lrb.aphp.fr 2012 American Heart Association, Inc. Stroke is available at DOI: /STROKEAHA

2 Yao et al Leukoariaosis and Brain Volume in CADASIL 3253 with significant neuronal loss at the latest stage of the disease. 7 At the macroscopic level, these different pathophysiological mechanisms may have opposite consequences on the cerebral volume: the accumulation of interstitial fluid may be responsible for cerebral volume enlargement, whereas demyelination and axonal loss can lead to brain atrophy. Although the negative impact of the load of WMH on brain volume has been repeatedly demonstrated in elderly populations or in subjects with subcortical ischemic vascular disease, 8 11 only few studies have examined these relationships in subjects with CADASIL. WMH were not previously found related to the brain volume in CADASIL conversely to the load of lacunar infarctions or water diffusion measured over the whole brain, which were found significantly associated with cerebral atrophy. 12,13 However, the actual relationships between WMH and cerebral volume were not specifically investigated and the samples of patients were of limited size in these studies. In the current study, we sought to evaluate the exact relationship between WMH and the brain volume in a large population of patients with CADASIL. Materials and Methods Subjects Data were obtained from a large prospective cohort study. A total of 278 patients were enrolled into the study at Lariboisière (Paris, France) or at Ludwig-Maximilians-Universitat (Munich, Germany) hospitals. In all cases, the diagnosis was confirmed by identification of a typical mutation in the Notch3 gene. 14 The complete study design has been detailed elsewhere. 15 All patients gave written consent to participate. An independent ethics committee in both participating centers approved the study. Magnetic Resonance Imaging MRI scans were performed by the use of a 1.5-T system (Vision; Siemens, Munich, Germany, or Signa General Electric Medical Systems, Paris, France). The following MRI sequences were used for analysis: 3-dimensional T1-weighted sequences (Munich: TR/ TE 11.4/4.4 ms, slice thickness 1.19 mm, no interslice gap, ; Paris: TR/TE 9/2 ms, slice thickness 0.8 mm, no interslice gap, ), fluid-attenuated inversion recovery (Munich: TR/TE/TI 4284/110/1428 ms, slice thickness 5 mm, no interslice gap, ; Paris: TR/TE/TI 8402/161/2002 ms, slice thickness 5.5 mm, no interslice gap, ), T2*-weighted gradient echo planar imaging (Munich: TR/TE 1056/22 ms, slice thickness 5 mm, no interslice gap, ; Paris: TR/TE 500/15 ms, slice thickness 5.5 mm, no interslice gap, ), and proton density images (Munich: TR/ TE 3300/16 ms, slice thickness 5 mm, no interslice gap, ; Paris: TR/TE 3300/15 ms, slice thickness 5.5 mm, no interslice gap, ). MRI data from both centers were collected and processed at BioClinica (Lyon, France) and analyzed together by trained neurologists who were blinded to the clinical data. 15 Brain Volume Assessment Determination of global brain volume from 3-dimensional T1- weighted MRI was made as previously described using Brainvisa software ( 12 High intrarater and interrater reliability was previously reported (interclass correlation coefficients and 0.922, respectively). 12 Automated determination of the intracranial cavity was performed on proton density images from the base to top of the skull using a dedicated algorithm. The corresponding mask was visually checked and marginally corrected if necessary. The volume of intracranial cavity (cerebral parenchyma and cerebrospinal fluid) was then calculated. Brain parenchymal fraction was defined as the ratio of brain tissue volume to intracranial cavity volume (brain parenchymal fraction=brain tissue volume/intracranial cavity volume). WMH Quantification WMH were analyzed on fluid-attenuated inversion recovery images as previously described. 15 All fluid-attenuated inversion recovery axial slices from the base of the cerebellum to the vertex were analyzed with a high interrater reliability (intraclass correlation coefficient, 0.995). 15 The total volume of WMH was normalized to the intracranial cavity in each patient (normalized volume of WMH=[volume of WMH/intracranial cavity volume] 100). 15 Lacunar Volume The volume of lacunar infarcts was determined on 3-dimensional T1- weighted images as previously detailed. 15 Hypointense lesions with signal intensity identical to that of cerebrospinal fluid and of diameter >2 mm, clearly distinct from Virchow-Robin space, were selected and delineated. The normalized lacunar volume was calculated individually as: (volume of lacunes/intracranial cavity volume) 100. Good interrater reliability was previously shown (intraclass correlation coefficient, 0.830). 15 Microbleeds Cerebral microbleeds were defined as rounded hypointense foci 5 mm in diameter on gradient echo sequence distinct from vascular flow voids, leptomeningeal hemosiderosis, or nonhemorrhagic subcortical mineralization. The number of cerebral microbleeds was recorded with a high interrater reliability (intraclass correlation coefficient, 0.962). 15 Statistical Methods Statistical analyses were performed using SPSS Version 20.0 for Windows (SPSS Inc). All P values were 2-tailed and criteria for significance were P<0.05. Normalized lesion volumes were always used for statistical analysis. We first explored the relationships between brain parenchymal fraction and different demographic and MRI variables (including age, sex, WMH volume, lacunar volume, and number of cerebral microbleeds) using univariate analysis. Nonparametric Spearman correlation tests were used for continuous variables. T test was used for comparison of mean brain parenchymal fraction values between men and women. Second, multivariate linear regression models were used to investigate correlation between brain parenchymal fraction and WMH volume. Candidate covariates were those associated with brain parenchymal fraction in univariate analysis (P<0.05). In the final model, variables were selected by stepwise regression analysis with entry and removal P values set at Thereafter, to explain the relationships between brain parenchymal fraction and WMH volume, we removed the different covariates (sex, lacunar volume, number of cerebral microbleeds, and age) from the original model one by one. The interaction term (age*wmh volume) was also added into the original regression model. To further understand the impact of WMH volume on brain volume, additional stratified multivariate stepwise analyses were performed according to the median value of WMH volume and brain parenchymal fraction as well as according to the presence or absence of lacunar infarcts and cerebral microbleeds, respectively. In the corresponding model, the same dependent variable and covariates were included as mentioned previously. Results Characteristics of the 278 patients with CADASIL are shown in Table 1. WMH volume, lacunar volume, and the number of cerebral microbleeds were all found to increase significantly with age (data not shown, all P values <0.001). In univariate analysis, brain parenchymal fraction was negatively related to age (Spearman ρ= 0.574, P<0.001) and

3 3254 Stroke December 2012 Table 1. Main Characteristics of Patients With CADASIL Obtained at Baseline (N=278) Age, y, mean±sd (range) ± (23 78) Male sex 124/278 (44.6%) MRI markers Absolute volume of WMH, mm 3, median (range) Normalized volume of WMH, median (range) Absolute volume of lacunar infarcts, mm 3, median (range) Normalized volume of lacunar infarcts, median (range) Number of cerebral microbleeds,* median (range) Brain parenchymal fraction, median (range) ( ) 6.13 ( ) ( ) ( ) 2 (1 141) 86.37% (65.33% 98.02%) CADASIL indicates autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy; WMH, white matter hyperintensities. *Calculated only in patients with cerebral microbleeds. to the different MRI lesions (ρ= 0.249, 0.510, and 0.371, respectively, for WMH volume, lacunar volume, and number of cerebral microbleeds; P<0.001). Men were also found to have a smaller mean brain parenchymal fraction than women (83.60%±6.74% versus 86.76%±5.98%, P<0.001). The stepwise multivariate analysis including sex, age, WMH volume, lacunar volume, and number of cerebral microbleeds as independent variables showed a negative association between brain parenchymal fraction and age, male sex, lacunar volume, and number of cerebral microbleeds (Table 2). It is noteworthy that an independent positive relationship was detected between brain parenchymal fraction and WMH volume (standardized coefficients β=0.216, P<0.001). Stepwise removal of sex, lacunar volume, and number of cerebral microbleeds from the original model did not alter the significant positive relationships between brain parenchymal fraction and WMH volume as long as age was kept in the model (data not shown). The interterm indicated that age was not a significant modifier of the relationship between brain parenchymal fraction Table 2. Stepwise Multivariate Analysis of Factors Related to Brain Parenchymal Fraction at Baseline Model R 2 Standardized Coefficients Beta P Value 1. Age < Male sex < Number of cerebral microbleeds < Normalized volume of lacunar infarcts < Normalized volume of WMH <0.001 Dependent variable: brain parenchymal fraction; candidate covariates: age, sex, normalized volume of WMH, normalized volume of lacunar infarcts, and number of cerebral microbleeds. WMH indicates white matter hyperintensities. and WMH volume (P=0.386). To further investigate the relationships between WMH volume and brain parenchymal fraction, multivariate stratified regression analyses were performed according to the median value of WMH volume (6.13) and brain parenchymal fraction (86.37%) as well as according to the presence or absence of cerebral microbleeds and lacunar infarct. After controlling for confounding factors, we observed a strongly positive correlation between WMH volume and brain parenchymal fraction regardless of the presence or absence of cerebral microbleeds or of lacunar infarcts (Table 3). Conversely, such a significantly positive association was only detected in the subjects with baseline brain parenchymal fraction larger than 86.37% (β=0.319, P=0.001). WMH volume was positively correlated with brain parenchymal fraction in patients with WMH volume >6.13 (β=0.216, P<0.001); this association was inverse in patients with WMH volume inferior to 6.13 (β= 0.159, P=0.026). The Figure further illustrates that subjects with large WMH volume can present with large brain parenchymal fraction and with obvious shrinkage of the ventricular system. Discussion To our knowledge, this is the largest study analyzing the relationship between brain volume and WMH volume in CADASIL. This is the first report showing a positive and independent correlation between WMH volume and brain parenchymal fraction at baseline in subjects with CADASIL. In contrast, age, male sex, and the volume of lacunar infarcts were found to be predictors of brain atrophy in line with results previously obtained in smaller samples of patients. 12,13 In the multivariate analysis, removing of confounders one by one showed that the positive association between brain parenchymal fraction and WMH remained unchanged as long as age was maintained in the model. This association is modest but highly significant and probably hidden at the crosssectional level by the strong but inverse effect of age on brain volume. Peter et al 13 did not find a significant link between the load of T2 hyperintensities and brain volume in a cohort of 76 patients with CADASIL. The different MRI sequences, quantitative methods, and sample sizes might partly explain these discrepant results. However, in a smaller group of patients from the same cohort evaluated in the present study, Jouvent et al 12 did not either observe a significant relationship between brain parenchymal fraction and WMH measured using the same quantitative methods and MRI sequences. There was no technical difference between that study and the present one, but some other discrepancies do indeed exist between these 2 studies. First, the sample size (n=129) was much smaller than ours (n=278). Second, the relationship between brain parenchymal fraction and WMH was only assessed by univariate analysis in the article by Jouvent et al that showed only a negative trend between these 2 parameters without any statistical significance. Accordingly, WMH was not included in the multivariate model. We showed that only when controlling for age did the positive association between brain parenchymal fraction and WMH become visible. WMH have been repeatedly found associated with reduction of gray matter 8,9,16 or brain volume 10,11,17 in elderly populations or in subjects with other types of subcortical ischemic

4 Yao et al Leukoariaosis and Brain Volume in CADASIL 3255 Table 3. Association Between Normalized WMH Volume and Brain Parenchymal Fraction in Different Stratified Analysis vascular disease. These findings were previously obtained with or without adjustment for age and using different rating or segmentation methods. The contrasting results of the present study appear unlikely related to the different quantification of WMH. In addition, the load of WMH in CADASIL largely exceeds that usually detected in the elderly populations. Altogether, these discrepancies support that the pathogenesis of WMH is not strictly identical in CADASIL and in age-associated small vessel disease. Different pathophysiological mechanisms have been attributed to the development of WMH associated with small vessel disease. Ischemic insults of white matter, blood brain barrier dysfunction, and chronic leakage of fluid and macromolecules into the brain tissue or a combination of these mechanisms may be involved The present results suggest that WMH may not merely be related to ischemic lesions Model R 2 Beta Standardized Coefficients Normalized volume of WMH <median 1. Age <0.001 P Value 2. Normalized volume of lacunar infarcts Male sex Normalized volume of WMH median 1. Age < Normalized volume of lacunar infarcts Male sex Normalized volume of WMH Number of cerebral microbleeds <0.001 Brain parenchymal fraction <median 1. Age < Number of cerebral microbleeds Male sex median 1. Age < Normalized volume of WMH Normalized volume of lacunar infarcts Lacunar infarct Absence (N=84) 1. Age < Normalized volume of WMH Presence (N=194) 1. Age < Normalized volume of WMH < Male sex < Number of cerebral microbleeds < Normalized volume of lacunar infarcts Cerebral microbleeds Absence (N=190) 1. Age < Normalized volume of lacunar infarcts < Normalized volume of WMH < Male sex <0.001 Presence (N=88) 1. Age < Normalized volume of WMH Male sex Number of cerebral microbleeds In each model, dependent variable was brain parenchymal fraction; candidate covariates include age, sex, normalized volume of WMH, normalized volume of lacunar infarcts, and number of cerebral microbleeds. Only the variables significantly correlated with brain parenchymal fraction are listed. WMH indicates white matter hyperintensities. in CADASIL. First, WMH of ischemic origin underlain by demyelination and axonal loss would have been related to cerebral atrophy as observed with lacunar infarcts. Second, WMH are poorly related to the clinical severity in CADASIL and can be detected at the late stage of the disease in the total absence of lacunar infarction. 21,22 Third, WMH are detected early in the temporal lobes during the course of the disease, a long time before the onset of ischemic stroke usually occurring around the 50s. 2 5 All of these further support that WMH and focal ischemic lesions might have a different pathogenesis. The large number of dilated Virchow-Robin spaces observed in CADASIL and their strong association with the early development of WMH in the temporal lobes 7 support that blood brain barrier dysfunction and insufficient drainage of interstitial fluid presumably related to vessel wall structural changes might play a key role in the extension of WMH.

5 3256 Stroke December 2012 A 45 year-old 48 year-old B 69 year-old 66 year-old Figure. A B, Two pairs of fluid-attenuated inversion recovery sequences of age- and sex-matched patients without lacunar infarct and cerebral microbleeds at baseline are illustrated. In each pair, 1 patient presents with load of white matter hyperintensities (WMH) <median value and low brain parenchymal fraction (BPF; left) and the other with load of WMH >median value and high BPF (right). These signal abnormalities might be, more or less, related to increased water content in cerebral tissue that can cause brain volume enlargement. An alternative explanation for the positive association between brain parenchymal fraction and WMH might be that large white matter volume owing to large brain volume makes it possible to develop more WMH. However, we found an inverse relationship between brain parenchymal fraction and WMH in the univariate analysis. Moreover, no correlation between WMH and the intracranial cavity was found even after controlling for other confounders (data not shown). Therefore, such an artifactual result appears very unlikely. Interestingly, although the interaction term indicated that there was no significant global age effect on the relationship between WMH and the brain volume, the stratified multivariate regression analysis showed that the association between WMH and the brain volume was not identical in the whole sample. A positive link between these 2 parameters was found significant only in subjects with extensive WMH and in those with high brain parenchymal fraction. These findings are difficult to interpret. Because the extent of WMH is strongly related to age in CADASIL, the positive but modest association between WMH and brain parenchymal fraction is probably concealed by the opposite effect of age on brain volume in most subjects. The present data suggest that the increase of brain volume becomes visible only at a certain degree of extension of WMH. The opposite relationships between WMH and brain parenchymal fraction according to the initial volume of WMH further support that water accumulation in cerebral tissue may accelerate with extension of WMH during the course of the disease. The methodological strengths of this study include the large sample size of cross-sectional data and the quantitative and validated assessment of lesion volumes. Our study also has potential limitations. First, we just compared the brain volume between subjects with CADASIL having more or less WMH, but we did not compare the results with those obtained in an age- and sex-matched healthy population using the same methods. Second, because of the cross-sectional design of this study, the exact contribution of WMH volume growth to the brain volume changes was not assessed. Third, we could not exclude that the negative findings in subgroup analysis did not result from insufficient power and limited sample size. Fourth, the brain volume measurements were limited to the whole brain in all subjects of our cohort because separate segmentation of white and gray matter is not accurate and reliable enough in CADASIL. Indeed, extensive changes of T1 or T2 signal are detected not only in the white matter, but also in the subcortical gray matter of most patients with CADASIL. 23 Such signal changes can dramatically alter the contrast needed for segmentation of the white and gray matter at the subcortical level, but also for isolating the cortex from the rest of the brain. Thus, correlations with gray matter volume and white matter volume were not evaluated. Finally, our research did not include any regional analysis of WMH, but WMH may have distinct pathogenesis in different cerebral areas with variable impact on brain volume. 8,16 In conclusion, the extent of WMH was found independently associated with some increase of brain volume in CADASIL. This positive relationship was modest and probably masked by the atrophy effect of age on brain volume at the cross-sectional level. These results may contribute to a better understanding of the mechanisms responsible for WMH in CADASIL. Prospective studies with long duration of follow-up and specific regional investigations may help to further elucidate the intimate mechanisms leading to WMH in CADASIL. Sources of Funding This work was supported by PHRC grant AOR (DRC/APHP) and performed with the help of ARNEVA (Association de Recherche

6 Yao et al Leukoariaosis and Brain Volume in CADASIL 3257 en Neurologie VAsculaire), Hôpital Lariboisière, France, the German Center for Neurodegenerative Diseases (DZNE), and the Vascular Dementia Research Foundation. Dr Yao is funded by the French Chinese Foundation for Science and Applications (FFCSA) and the China Scholarship Council (CSC). Sponsors were not involved either in the design of the study or in the data analyses or article elaboration. None. Disclosures References 1. Joutel A, Corpechot C, Ducros A, Vahedi K, Chabriat H, Mouton P, et al. Notch3 mutations in CADASIL, a hereditary adult-onset condition causing stroke and dementia. Nature. 1996;383: Chabriat H, Vahedi K, Iba-Zizen MT, Joutel A, Nibbio A, Nagy TG, et al. Clinical spectrum of CADASIL: a study of 7 families. Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy. Lancet. 1995;346: Chabriat H, Levy C, Taillia H, Iba-Zizen MT, Vahedi K, Joutel A, et al. Patterns of MRI lesions in CADASIL. Neurology. 1998;51: van den Boom R, Lesnik Oberstein SA, Ferrari MD, Haan J, van Buchem MA. Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy: MR imaging findings at different ages 3 rd -6 th decades. Radiology. 2003;229: Singhal S, Rich P, Markus HS. The spatial distribution of MR imaging abnormalities in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy and their relationship to age and clinical features. AJNR Am J Neuroradiol. 2005;26: Cumurciuc R, Guichard JP, Reizine D, Gray F, Bousser MG, Chabriat H. Dilation of Virchow-Robin spaces in CADASIL. Eur J Neurol. 2006;13: Yamamoto Y, Ihara M, Tham C, Low RW, Slade JY, Moss T, et al. Neuropathological correlates of temporal pole white matter hyperintensities in CADASIL. Stroke. 2009;40: Wen W, Sachdev PS, Chen X, Anstey K. Gray matter reduction is correlated with white matter hyperintensity volume: a voxel-based morphometric study in a large epidemiological sample. Neuroimage. 2006;29: Fein G, Di Sclafani V, Tanabe J, Cardenas V, Weiner MW, Jagust WJ, et al. Hippocampal and cortical atrophy predict dementia in subcortical ischemic vascular disease. Neurology. 2000;55: Schmidt R, Ropele S, Enzinger C, Petrovic K, Smith S, Schmidt H, et al. White matter lesion progression, brain atrophy, and cognitive decline: the Austrian stroke prevention study. Ann Neurol. 2005;58: Enzinger C, Fazekas F, Matthews PM, Ropele S, Schmidt H, Smith S, et al. Risk factors for progression of brain atrophy in aging: six-year follow-up of normal subjects. Neurology. 2005;64: Jouvent E, Viswanathan A, Mangin JF, O Sullivan M, Guichard JP, Gschwendtner A, et al. Brain atrophy is related to lacunar lesions and tissue microstructural changes in CADASIL. Stroke. 2007;38: Peters N, Holtmannspötter M, Opherk C, Gschwendtner A, Herzog J, Sämann P, et al. Brain volume changes in CADASIL: a serial MRI study in pure subcortical ischemic vascular disease. Neurology. 2006;66: Chabriat H, Joutel A, Dichgans M, Tournier-Lasserve E, Bousser MG. CADASIL. Lancet Neurol. 2009;8: Viswanathan A, Guichard JP, Gschwendtner A, Buffon F, Cumurcuic R, Boutron C, et al. Blood pressure and haemoglobin A1c are associated with microhaemorrhage in CADASIL: a two-centre cohort study. Brain. 2006;129(pt 9): Godin O, Maillard P, Crivello F, Alpérovitch A, Mazoyer B, Tzourio C, et al. Association of white-matter lesions with brain atrophy markers: the three-city Dijon MRI study. Cerebrovasc Dis. 2009;28: Grinberg LT, Thal DR. Vascular pathology in the aged human brain. Acta Neuropathol. 2010;119: Black S, Gao F, Bilbao J. Understanding white matter disease: imaging-pathological correlations in vascular cognitive impairment. Stroke. 2009;40(3 suppl):s48 S Fernando MS, Simpson JE, Matthews F, Brayne C, Lewis CE, Barber R, et al. MRC Cognitive Function and Ageing Neuropathology Study Group. White matter lesions in an unselected cohort of the elderly: molecular pathology suggests origin from chronic hypoperfusion injury. Stroke. 2006;37: Awad IA, Johnson PC, Spetzler RF, Hodak JA. Incidental subcortical lesions identified on magnetic resonance imaging in the elderly. II. Postmortem pathological correlations. Stroke. 1986;17: Viswanathan A, Godin O, Jouvent E, O Sullivan M, Gschwendtner A, Peters N, et al. Impact of MRI markers in subcortical vascular dementia: a multimodal analysis in CADASIL. Neurobiol Aging. 2009;30: Liem MK, van der Grond J, Haan J, van den Boom R, Ferrari MD, Knaap YM, et al. Lacunar infarcts are the main correlate with cognitive dysfunction in CADASIL. Stroke. 2007;38: Jacqmin M, Hervé D, Viswanathan A, Guichard JP, During M, Dichgans M, Chabriat H. Confluent thalamic hyperintensities in CADASIL. Cerebrovasc Dis. 2010;30: