Dementia in hereditary cerebral hemorrhage with amyloidosis-Dutch type is associated with cerebral amyloid angiopathy but is independent of plaques and neurofibrillary tangles.код для вставкиСкачать
Dementia in Hereditary Cerebral Hemorrhage with Amyloidosis-Dutch Type Is Associated with Cerebral Amyloid Angiopathy but Is Independent of Plaques and Neurofibrillary Tangles Remco Natté, MD,1 Marion L. C. Maat-Schieman, MD, PhD,1 Joost Haan, MD, PhD,1,3 Marjolijn Bornebroek, MD, PhD,1 Raymund A.C. Roos, MD, PhD,1 and Sjoerd G. van Duinen, MD, PhD2 Cerebral amyloid angiopathy is frequently found in demented and nondemented elderly persons, but its contribution to the causation of dementia is unknown. Therefore, we investigated the relation between the amount of cerebral amyloid angiopathy and the presence of dementia in 19 patients with hereditary cerebral hemorrhage with amyloidosis-Dutch type. The advantage of studying hereditary cerebral hemorrhage in amyloidosis-Dutch type is that patients with this disease consistently have severe cerebral amyloid angiopathy with minimal neurofibrillary pathology. The amount of cerebral amyloid angiopathy, as quantified by computerized morphometry, was strongly associated with the presence of dementia independent of neurofibrillary pathology, plaque density, or age. The number of cortical amyloid ␤-laden severely stenotic vessels, vessel-within-vessel configurations, and cerebral amyloid angiopathy-associated microvasculopathies was associated with the amount of cerebral amyloid angiopathy and dementia. A semiquantitative score, based on the number of amyloid ␤-laden severely stenotic vessels, completely separated demented from nondemented patients. These results suggest that extensive (more than 15 amyloid ␤-laden severely stenotic vessels in five frontal cortical sections) cerebral amyloid angiopathy alone is sufficient to cause dementia in hereditary cerebral hemorrhage with amyloidosis–Dutch type. This may have implications for clinicopathological correlations in Alzheimer’s disease and other dementias with cerebral amyloid angiopathy. Ann Neurol 2001;50:765–772 Cerebral amyloid angiopathy (CAA) refers to the presence of amyloid within cerebral vessel walls.1 In the most common form of CAA, the amyloid is primarily composed of amyloid ␤ protein (A␤).2 Although rare forms of CAA with other primary amyloidogenic proteins exist,3– 6 in the present article the term CAA is used to refer to A␤-CAA. CAA predominantly involves the meningo-cortical blood vessels. In autopsy series the prevalence of CAA increases from about 10% of persons between 60 and 69 years old to about 50% in persons over 90 years old.7,8 CAA is widely accepted as an important cause of cerebral hemorrhage and has also been associated with cerebral infarcts and diffuse white matter changes.1,9 –13 The contribution of CAA to dementia is less well established. Severe CAA in the near absence of Alzheimer’s disease (AD) pathology (neuritic plaques [NP] and neurofibrillary tangles [NFT]) has been reported in demented and nondemented persons.13–19 A major problem to determine a relation between CAA and dementia is that most patients with CAA also have AD pathology, and that extensive CAA is generally associated with more NP/NFT.1,10,19 –24 Hereditary cerebral hemorrhage with amyloidosis– Dutch type (HCHWA-D) is an autosomal dominant form of CAA caused by a point mutation at codon 693 of the A␤ precursor protein gene.12,25–29 Patients with this mutation almost invariably develop strokes and white matter changes.12,30 Dementia is a common symptom in patients over 40 years old.31,32 HCHWA-D is pathologically characterized by severe CAA and diffuse plaques,25,33 but NP are rare or absent25,33 and NFT are rarely observed, even in the hippocampus.25,27,33 Brain atrophy is uncommon.27 From the Departments of 1Neurology and 2Pathology, Leiden University Medical Center, Leiden, The Netherlands; and 3Department of Neurology, Rijnland Hospital, Leiderdorp, The Netherlands. Published online Nov 1, 2001; DOI: 10.1002/ana.10040 Received May 18, 2001, and in revised form Aug 15. Accepted for publication Aug 16, 2001. Address correspondence to Dr Natté, Department of Pathology, Leiden University Medical Center, L-1-Q, PO Box 9600, 2300 RC, Leiden, The Netherlands. E-mail: R.Natte@lumc.nl © 2001 Wiley-Liss, Inc. 765 The main objective of the present study was to investigate the association between CAA and dementia in patients with HCHWA-D because these patients always have severe CAA and consistently lack significant neurofibrillary pathology. Another advantage of studying HCHWA-D is its homogenous etiology of CAA. We found the amount of CAA, as quantified by computerized morphometry, to be strongly associated with the presence of dementia, independent of AD pathology or age. We present a semiquantitative score for the amount of CAA, which adequately distinguished demented from nondemented persons. Patients and Methods Nineteen HCHWA-D patients were included in the present study. In 17 patients the diagnosis of HCHWA-D was confirmed by DNA analysis. The remaining 2 patients had clinical symptoms and neuropathological alterations typical for HCHWA-D and were members of an HCHWA-D family. Apolipoprotein E genotype was available from 16 patients and was determined as described previously.34 Patients were included if (1) at least one coronal brain slice of a single hemisphere rostral to the thalamus was available and (2) if clinical information allowed dementia to be either diagnosed or excluded. Patients with signs or complaints of cognitive deterioration were only included if information on social functioning and results of neurological and neuropsychological investigation (Wechsler Adult Intelligence Scale) was all available and if, based on this information, the patient fulfilled the DSM-IV criteria for dementia. The diagnosis dementia according to the DSM-IV criteria was made in retrospect by a neurologist (RACR) who was unaware of the amount of neuropathological lesions and did not take the imaging results into account. Patients without evidence of cognitive deterioration were only considered not demented, and included as such, if the patients medical records from the general practitioner and the hospital were available and revealed no signs or complaints of deterioration in cognition or social functioning. Morphometric Quantification of CAA Only coronal slices of a single hemisphere rostral to the thalamus were selected to determine the amount of CAA morphometrically in a given patient. One (n ⫽ 5), two (n ⫽ 9), four (n ⫽ 1), five (n ⫽ 3), and six (n ⫽ 1) formalin-fixed slices were selected, depending on available material. Brain slices were embedded in paraffin and 10m-thick sections were cut from the entire hemisphere slice. CAA was shown immunohistochemically using 1 to 10 diluted monoclonal antibody 6F/3D (DAKO, Denmark) and DAB after formic acid pretreatment. Macroscopically, three to eight (mean six) frontal cortical areas where marked in each anonymized section. Only areas where the pial surface of the cortex paralleled the cortico-medullary junction were selected in order to have all cortical layers present to a similar extent in all patients. Areas with edema, hemorrhage, or infarct were excluded. Macroscopically selected cortical areas were on average 23 mm2 large. The amount of CAA was quantified using the Zeiss 766 Annals of Neurology Vol 50 No 6 December 2001 KS400 system (Zeiss, Jena, Germany), consisting of a bright field microscope equipped with a motorized stage, a DXC950P, 3-CCD video camera (Sony, Tokyo, Japan), and KS400 software (Contron Electronic, Eching, Germany). For each measurement session, the intensity of the light was standardized by measuring the intensity of the light in the center of a blank image, and set on a fixed value. A white balance and shading correction was performed on all captured images. Two measures for the amount of CAA were determined: (1) CAA load, defined as the percentage of cortical area that is occupied by A␤ positive vessel walls and (2) vessel wall thickening in A␤ positive cortical vessels, calculated by the lumen/vessel diameter ratio. For CAA load the size of the area of macroscopically selected cortex of the hemisphere sections was measured. White matter was excluded manually. To measure the size of the total area of vascular DAB staining, the image was automatically divided in positive or negative areas for DAB staining, using a fixed hue, intensity, and saturation threshold. Plaques and artefacts were excluded manually. CAA load was determined at 100⫻ magnification. We also quantified CAA by measuring the degree of vessel wall thickening of A␤ positive, crosscut vessels. We assume this reflects the extent of amyloid deposition in involved vessels. Vessel wall thickening as a measure for the severity of CAA is less sensitive for changes in tissue volume and epitope preservation than surface percentage of CAA. Lumen and total vessel circumference of crosscut A␤ positive vessels in the macroscopically selected cortex were manually encircled by one technician. To compare vessel wall thickness between patients for a similar A␤ immunoreactive vessel population (in severe CAA relatively more capillaries and small arterioles are involved) and to limit the labor-intensive encircling of vessels, all vessels in the macroscopically selected cortex with a minimal diameter over 30m were measured. From the encircled lumen the central point was calculated automatically (by calculating the center of gravity), and from this point the diameter of the lumen and the vessel diameter were automatically determined in 20 directions and averaged. A vessel was considered crosscut when the maximum vessel diameter was not more than 1.5⫻ the minimum vessel diameter. Vessel wall thickening was measured at 400⫻ magnification. Intraindividual reliability was tested by the repeated measurement of five randomly selected areas in five different randomly selected patients. For CAA load, single measure intraclass correlation coefficient for intraindividual reproducibility was 0.993 (95% confidence interval [CI]: 0.934 – 0.999). For the ratio lumen-vessel diameter, this was 0.959 (95% CI: 0.670 – 0.995). Quantification of Plaque Load and Braak Staging Plaque load was defined as the number of plaques per mm2 cortex. Plaques were visualized by methenamine silver stain (MS)35 in coronal, single hemisphere sections adjacent to the immunostained sections. In the MS stained sections the same cortical areas were macroscopically selected as in the immunostained sections. In these areas the size of the cortical area was measured using the KS400 system. The number of plaques in these areas was counted manually. Inter- and intraobserver reliability were tested by the repeated measure- Fig 1. Examples of vascular pathology in the present study. (A) Amyloid laden severely stenotic vessels (ALSSV) (arrows) in a demented patient with one of the highest amounts of cerebral amyloid angiopathy (CAA). Vessels indicated by arrowheads were not counted as ALSSV because they were not considered to be round. (Magnification ⫻100 before 32% reduction.) (B) Representative sample of amyloid laden crosscut vessels in a nondemented hereditary cerebral hemorrhage with amyloidosis-Dutch type (HCHWA-D) patient with one of the mildest degrees of CAA. Vessel wall thickening is limited. (Magnification ⫻50 before 32% reduction.) (C) Vessel-within-vessel configurations in which the inner and outer structures were distinguishable at 100% of the circumference (arrows). Partially split vessels (arrowheads) were not included to provide clear definitions for counting, although they probably represent different severities of the same lesion. (Magnification ⫻100 before 32% reduction.) (D) Microaneurysm with eosinophilic hyalinization (arrows) and foamy macrophages (arrowheads) in a demented patient. (Magnification ⫻200 before 32% reduction.) ment of five randomly selected cortical areas in five different randomly selected patients. Single measure intraclass correlation coefficient for interindividual reproducibility (RN and MLCM-S) of the plaque counts was 0.98 (95% CI: 0.82– 0.99). This was 0.96 (95% CI: 0.69 – 0.99) for intraindividual observer reproducibility of the plaque counts by the investigator performing all plaque counts (RN). Neurofibrillary degeneration was assessed by two investigators (RN and MLCM-S) on anonymized sections using the Braak criteria.36 Of each patient, all available formalin-fixed and paraffin-embedded blocks (one or two) of hippocampus were used. Neurofibrillary degeneration was detected by a modified Bielschowsky’s stain.37 Association of Morphologic Features of CAA with Dementia and with the Amount of CAA as Quantified by Morphometry Three CAA-related vessel wall changes were investigated for their relation with the morphometrically assessed amount of CAA and for their relation with dementia: (1) cortical amyloid laden severely stenotic vessels (ALSSV), (2) vesselwithin-vessel configurations (V-W-V configurations), and (3) cerebral amyloid angiopathy-associated microvasculopathies (CAA-AM). ALSSV were defined as A␤ positive, crosscut vessels in which the vessel wall thickness was estimated to exceed the lumen diameter in all directions (Fig 1A). Crosscut vessels were defined as round-shaped vessels. A␤ positive capillaries were excluded. ALSSV were counted manually in the A␤ immunostained hemisphere sections that had been analyzed by morphometry. V-W-V configurations were defined as vessels in which the “inner vessel” was completely distinguishable from the “outer vessel” (see Fig 1C). The following features of CAA-AM have been assessed: (1) hyalinosis, (2) compact homogenous fibrosis of the tunica media, (3) histiocytes within the vessel wall, (4) microaneurysms, (5) fibrinoid necrosis, (6) vessel wall calcifications, (7) thrombosis, (8) presence of perivascular lymphocytes, and (9) perivascular multinucleated giant cells. The morphology of these changes in HCHWA-D has been described.38 The number of vessels with one or more CAA-AM was counted. Thus, a vessel with a microaneurysm with hyalinosis and histiocytes was counted once (see Fig 1D). V-W-V configurations and CAA-AM were counted in HE stained, coronal, single hemisphere sections adjacent to the immunostained sections used for morphometry. The entire frontal cortex in a given section was assessed, thus not only the selected areas for morphometric measurement. The numbers of ALSSV, V-W-V configurations, and vessels with CAA-AM in the large hemisphere sections were divided by the estimated times the cortical area of each large section represented the cortical area Natté et al: Dementia in HCHWA-D Is Associated with CAA 767 Table 1. Demographics, the Amount of CAA and Plaque Counts for Demented and Nondemented HCHWA-D Patients All Patients Gender (M:F) Age, yr (SEM) CAA-load, % (SEM) Lumen/vessel D, % (SEM) Vessel D, m (SEM) Plaques, n/mm2 (SEM) Patients 50–60 yr Demented (n ⫽ 8) Nondemented (n ⫽ 11) pa Demented (n ⫽ 4) Nondemented (n ⫽ 8) p 6:2 63.5 (3.70) 1.28 (0.21) 45.0 (2.20) 49.6 (0.70) 8.4 (2.50) 3:8 52.0 (1.80) 0.32 (0.03) 59.0 (1.20) 50.1 (0.80) 10.2 (1.30) 0.008 0.001 ⬍0.001 0.778 0.493 4:0 55.0 (1.30) 1.16 (0.26) 43.3 (3.40) 50.6 (0.80) 11.4 (4.60) 2:6 54.9 (1.40) 0.33 (0.05) 59.1 (1.60) 49.9 (0.90) 8.9 (1.50) 0.957 0.027 0.007 0.808 0.528 a p values were calculated by Mann-Whitney U test, except for age and plaques for which p values were calculated by Student’s t test. CAA⫽cerebral amyloid angiopathy; HCHWA-D⫽hereditary cerebral hemorrhage with amyloidosis–Dutch type; M⫽male; F⫽female; CAAload⫽surface of cortical vessel wall A␤ immunoreactivity as a percentage of the surface of cortex measured, lumen/vessel D⫽lumen diameter as a percentage of vessel diameter. of five standard cortical sections. All sections were anonymized. A semiquantitative score based on the number of ALSSV was tested on a second sample of frontal cortex of the same patients (see Results). Of each patient, five standard blocks of paraffin-embedded frontal cortex were randomly selected from our archive. These blocks mostly originated from the hemisphere opposite to the hemisphere from which large sections were analyzed by morphometry. Immunolabeling for A␤ was performed using 6F/3D antibody as described.39 Statistical Methods Differences between demented and nondemented patients were analyzed by Mann-Whitney U test for CAA load, lumen/vessel diameter ratio, number of ALSSV, number of V-W-V configurations, and number of vessels with CAAAM. Age and plaque counts were normally distributed and analyzed by Student’s t test. The association between the morphometric measurements of CAA and the manually counted morphological features of CAA was calculated using Spearman’s correlation. Results Eight HCHWA-D patients were demented and 11 patients were not (Table 1). All patients showed CAA with circular and transmural involvement of the vessel wall in nearly all 10⫻10 cortical fields. Eighty-nine percent showed cracking40 of vessel walls and 84% showed at least one vessel with CAA-AM (see Fig 1). In line with earlier findings in HCHWA-D,34,41 the number of patients with an APOE4 allele was not significantly different between demented and nondemented HCHWA-D patients. Patients with an APOE4 allele did not have a higher amount of CAA than patients without an APOE4 allele (data not shown). The Amount of CAA Is Strongly Associated with Dementia Independent of AD Pathology and Age Demented patients had a higher CAA load and more vessel wall thickening than nondemented patients (Fig 2, see Table 1). There was little overlap between de- 768 Annals of Neurology Vol 50 No 6 December 2001 mented and nondemented patients for these two morphometric measures of CAA. Plaque density was similar in demented and nondemented patients (see Table 1). In 4 patients (2 demented, 2 nondemented), no NFT were observed in the temporal cortex, CA region, or enthorhinal cortex. Fourteen patients had Braak stage I/II and 1 demented patient had Braak stage III. Thus, in the present patient series, the number of plaques or NFT did not confound the relation between CAA and dementia. Demented patients were older than nondemented patients. If patients younger than 50 and older than 60 years old were excluded, demented patients still had a higher CAA load and thicker vessel walls than nondemented patients. Both groups had the same mean age (see Table 1). Morphologic Features of CAA Are Associated with Dementia and the Amount of CAA as Quantified by Morphometry The numbers of cortical ALSSV and V-W-V configurations were associated with the amount of CAA as quantified by morphometry. There was a marginally significant association between the number of vessels with CAA-AM and the two morphometric measures for the amount of CAA (Fig 3). The numbers of ALSSV, V-W-V configurations, and vessels with CAA-AM were higher in demented than in nondemented HCHWA-D patients (see Fig 3 and Table 2). Counts in large sections revealed that above 15 ALSSV per cortical area (that corresponds with five standard cortical sections), all patients were demented; below 15 ALSSV, all patients but 1 were not (see Fig 3A and B). In a second sample of frontal cortex of the same patients (this time five standard paraffinembedded blocks), all demented patients had more than 15 ALSSV, whereas all nondemented patients had less ( p ⬍ 0.001, 2 test). The one demented patient with a CAA load and number of ALSSV in the lower area is representative for the amount of CAA in the entire brain. To test these assumptions, the amount of CAA was measured in coronal sections throughout one hemisphere in 4 patients (2 demented and 2 nondemented). The amount of CAA was similar at all coronal levels from the frontal pole to the caudal end of the corpus callosum. More caudally, CAA load increased but the order of CAA severity among these 4 patients was similar at all coronal levels (Fig 4A). Vessel wall thickening tended to increase at the occipital end of the brain (see Fig 4B). In some sections of these 4 patients, areas of temporal cortex were also analyzed but there was no consistent difference in the amount of CAA between temporal cortex and frontal/parietal cortex at the different coronal levels. Fig 2. Association between the amount of cerebral amyloid angiopathy (CAA) and age and dementia in hereditary cerebral hemorrhage with amyloidosis–Dutch type. (A) CAA load is the area of cortical vessel wall A␤ immunoreactivity as a percentage of the area of cortex. (B) Lumen/vessel D ⫽ lumen diameter as a percentage of vessel diameter. range (see Figs 2 and 3A and B) showed 30 ALSSV in this second sample. One nondemented patient was not assessed because only one standard block of frontal cortex was available. Counts were performed separately by two investigators (RN and SGvD) who classified all patients identically. The Amount of CAA in Coronal One-Hemisphere Sections in the Fronto-Occipital Direction For the tissue-sampling scheme used in the present study we assumed that (1) rostral to the thalamus, coronal brain slices at different levels have similar amounts of CAA and (2) the amount of CAA in this Discussion In the present study we show that a high amount of CAA in the frontal neocortex is associated with the presence of dementia in patients with HCHWA-D. This association is independent of neurofibrillary pathology, plaque density, or age. The number of ALSSV, the number of V-W-V configurations, and the number of vessels with CAA-AM were associated with (1) the amount of CAA as measured by computerized morphometry and (2) dementia. All patients with more than 15 ALSSV in five standard sections of frontal cortex were demented, whereas all patients with less than 15 ALSSV were not. CAA appears to be important for dementia in HCHWA-D in quantities found in the most severely involved HCHWA-D patients, having more than 15 ALSSV per five sections of frontal cortex. The mechanisms by which CAA may cause dementia are partly unclear. The present finding of a higher amount of CAA-AM in demented than nondemented patients suggests that CAA-related dementia is largely mediated by (small) cerebral hemorrhages/infarcts, as CAA-AM have been associated with the occurrence of brain hemorrhages/infarcts in CAA patients.19,42 The higher number of severely stenotic vessels in demented patients points in the same direction. Cerebral hemorrhages and infarcts as a result of CAA are likely to contribute to cognitive decline but other mechanisms may also play a role.31 Neuron loss has been associated with CAA in the absence of ischemic/hemorrhagic brain damage.43 Furthermore, CAA may affect cognition by inducing white matter changes.13,31 Even if CAA results in dementia by causing brain hemorrhage/infarcts, from a clinical point of view it is important to realize that the absence of a history of stroke or focal radiological lesions in demented persons does not exclude the possibility that the dementia is purely CAA related.17,30,34,44 For instance, in the present patient series 1 patient presented with dementia in the absence Natté et al: Dementia in HCHWA-D Is Associated with CAA 769 Fig 3. Correlation between the morphometrically determined amount of cerebral amyloid angiopathy (CAA) and manually counted features of CAA severity and dementia. n ⫽ counts in single hemisphere sections per cortical area corresponding to the cortical area in five standard cortical sections. V-W-V config ⫽ vessel within vessel configurations; CAA-AM ⫽ cerebral amyloid angiopathyassociated microvasculopathies. For definitions of amyloid laden severely stenotic vessels, V-W-V config and CAA-AM, see Patients and Methods. For description of CAA load and lumen/vessel D see Figure 2. (A) Correlation coefficient (CC) ⫽ 0.887, p ⬍ 0.001. (B) CC ⫽ 0.917, p ⬍ 0.001. (C) CC ⫽ 0.758, p ⬍ 0.001. (D) CC ⫽ 0.434, p ⫽ 0.063. Correlation between V-W-V config and lumen/vessel D: CC ⫽ ⫺0.725, p ⬍ 0.001, and between CAA-AM and lumen/vessel D: CC ⫽ ⫺0.457, p ⫽ 0.049 (not shown). CC and p values for correlations between the different parameters for CAA severity were calculated using Spearman’s correlation. of strokes, transient ischemic attacks (TIAs), or focal lesions on computed tomography scan. He was initially diagnosed as having AD, but at autopsy his brain showed Braak stage I/II NFT. Another patient had complaints of memory loss and problems with spatial orientation and daily tasks before he suffered TIAs, followed by strokes. No NFT were observed in the hippocampus or enthorhinal cortex of this patient. Two patients suffered a single stroke after which a good recovery was reported, but they became demented 7 and 3 years later, respectively, without noticed additional 770 Annals of Neurology Vol 50 No 6 December 2001 strokes within these intermediate periods. At the time of death, the former patient had Braak stage I/II (at age 81 years), the latter had Braak stage 0. The present findings underline the potential importance of other (non-CAA) vascular disease in dementia. CAA predominantly involves cerebral microvessels, suggesting that non-CAA vascular dementia may be particularly related to microvascular disease. Indeed, small ischemic lesions caused by microvascular disease have been suggested as the most consistent pathological correlate for (non-CAA) vascular dementia.45,46 Table 2. Results of the Manual Counts of Different Features of CAA Severity ALSSV, (n) (SEM) V-W-V config, (n) (SEM) CAA-AM, (n) (SEM) Demented (n ⫽ 8) Nondemented (n ⫽ 11) pa 63.3 (13.60) 27.6 (8.40) 5.3 (2.10) 3.3 (0.96) 2.2 (0.89) 1.1 (0.40) ⬍0.001 ⬍0.001 0.004 a p values were calculated by Mann Whitney U test. CAA⫽cerebral amyloid angiopathy, ALSSV⫽amyloid laden severely stenotic vessels. n⫽counts in single hemisphere sections per cortical area corresponding to the cortical area in five standard cortical sections, V-W-V config⫽vessel within vessel configuration, CAA-AM⫽cerebral amyloid angiopathy-associated microvasculopathies. We conclude that in HCHWA-D, the amount of CAA is strongly associated with dementia independent of AD pathology. This finding suggests that CAA in Fig 4. Amount of CAA at different coronal levels in the fronto-occipital direction in one hemisphere in 4 patients (solid square, open square, triangle, and asterisk). FH ⫽ beginning of frontal horn; TH ⫽ beginning of thalamus; LGN ⫽ lateral geniculate nucleus; OCC ⫽ occipital end of the corpus callosum; OOH ⫽ occipital end of the occipital horn. For description of CAA load and lumen/vessel D, see Figure 2. the absence of NP/NFT is sufficient to cause dementia. Thus, CAA alone may be a valid pathological substrate for dementia,13–19 and routine grading for CAA in addition to grading NFT and NP may improve clinicopathological correlations in AD and other dementing illnesses with CAA. In HCHWA-D, the number of ALSSV is a practical tool to assess if CAA is severe enough to be significant for dementia. Whether this is also true for sporadic CAA needs to be investigated. R. Natté was supported by the Internationale Stichting Alzheimer Onderzoek (ISAO 96506). C. Welling-Graafland performed all stainings. A. Van Der Wal patiently encircled all vessels. H. A. M. Middelkoop, PhD, assisted with the retrospective interpretation of neuropsychological testing results. ApoE genotypes were determined by E. A. Bakker, PhD. References 1. Vinters HV. Cerebral amyloid angiopathy. A critical review. Stroke 1987;18:311–324. 2. Glenner GG, Wong CW. Alzheimer’s disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 1984;120: 885– 890. 3. Ghetti B, Piccardo P, Spillantini MG, et al. Vascular variant of prion protein cerebral amyloidosis with tau-positive neurofibrillary tangles: the phenotype of the stop codon 145 mutation in PRNP. Proc Natl Acad Sci U S A 1996;93:744 –748. 4. Olafsson I, Thorsteinsson L, Jensson O. The molecular pathology of hereditary cystatin C amyloid angiopathy causing brain hemorrhage. Brain Pathol 1996;6:121–126. 5. Ushiyama M, Ikeda S, Yanagisawa N. Transthyretin-type cerebral amyloid angiopathy in type I familial amyloid polyneuropathy. Acta Neuropathol (Berl) 1991;81:524 –528. 6. Vidal R, Frangione B, Rostagno A, et al. A stop-codon mutation in the BRI gene associated with familial British dementia. Nature 1999;399:776 –781. 7. Masuda J, Tanaka K, Ueda K, et al. Autopsy study of incidence and distribution of cerebral amyloid angiopathy in Hisayama, Japan. Stroke 1988;19:205–210. 8. Tomonaga M. Cerebral amyloid angiopathy in the elderly. J Am Geriatr Soc 1981;29:151–157. 9. Greenberg SM. Cerebral amyloid angiopathy: prospects for clinical diagnosis and treatment. Neurology 1998;51:690 – 694. 10. Ellis RJ, Olichney JM, Thal LJ, et al. Cerebral amyloid angiopathy in the brains of patients with Alzheimer’s disease: the CERAD experience, Part XV. Neurology 1996;46:1592–1596. Natté et al: Dementia in HCHWA-D Is Associated with CAA 771 11. Olichney JM, Hansen LA, Hofstetter CR, et al. Cerebral infarction in Alzheimer’s disease is associated with severe amyloid angiopathy and hypertension. Arch Neurol 1995;52:702–708. 12. Bornebroek M, Haan J, Maat-Schieman ML, et al. Hereditary cerebral hemorrhage with amyloidosis-Dutch type (HCHWAD): I. A review of clinical, radiologic and genetic aspects. Brain Pathol 1996;6:111–114. 13. Gray F, Dubas F, Roullet E, et al. Leukoencephalopathy in diffuse hemorrhagic cerebral amyloid angiopathy. Ann Neurol 1985;18:54 –59. 14. Cohen DL, Hedera P, Premkumar DR, et al. Amyloid-beta protein angiopathies masquerading as Alzheimer’s disease? Ann N Y Acad Sci 1997;826:390 –395. 15. Greenberg SM, Vonsattel JP, Stakes JW, et al. The clinical spectrum of cerebral amyloid angiopathy: presentations without lobar hemorrhage. Neurology 1993;43:2073–2079. 16. Mandybur TI. Cerebral amyloid angiopathy: the vascular pathology and complications. J Neuropathol Exp Neurol 1986; 45:79 –90. 17. Nadeau SE, Bebin J, Smith E. Nonspecific dementia, cortical blindness, and Congophilic angiopathy. A clinicopathological report. J Neurol 1987;234:14 –18. 18. Okazaki H, Reagan TJ, Campbell RJ. Clinicopathological studies of primary cerebral amyloid angiopathy. Mayo Clin Proc 1979;54:22–31. 19. Vonsattel JP, Myers RH, Hedley-Whyte ET, et al. Cerebral amyloid angiopathy without and with cerebral hemorrhages: a comparative histological study. Ann Neurol 1991;30:637– 649. 20. Joachim CL, Morris JH, Selkoe DJ. Clinically diagnosed Alzheimer’s disease: autopsy results in 150 cases. Ann Neurol 1988;24:50 –56. 21. Esiri MM, Wilcock GK. Cerebral amyloid angiopathy in dementia and old age. J Neurol Neurosurg Psychiatry 1986;49: 1221–1226. 22. Greenberg SM, Rebeck GW, Vonsattel JP, et al. Apolipoprotein E epsilon 4 and cerebral hemorrhage associated with amyloid angiopathy. Ann Neurol 1995;38:254 –259. 23. Vinters HV, Gilbert JJ. Cerebral amyloid angiopathy: incidence and complications in the aging brain. II. The distribution of amyloid vascular changes. Stroke 1983;14:924 –928. 24. Yamada M, Tsukagoshi H, Otomo E, et al. Cerebral amyloid angiopathy in the aged. J Neurol 1987;234:371–376. 25. Maat-Schieman ML, van Duinen SG, Bornebroek M, et al. Hereditary cerebral hemorrhage with amyloidosis-Dutch type (HCHWA-D): II. A review of histopathological aspects. Brain Pathol 1996;6:115–120. 26. Wattendorff AR, Bots GT, Went LN, et al. Familial cerebral amyloid angiopathy presenting as recurrent cerebral haemorrhage. J Neurol Sci 1982;55:121–135. 27. Wattendorff AR, Frangione B, Luyendijk W, et al. Hereditary cerebral haemorrhage with amyloidosis, Dutch type (HCHWAD): clinicopathological studies. J Neurol Neurosurg Psychiatry 1995;58:699 –705. 28. Levy E, Carman MD, Fernandez-Madrid IJ, et al. Mutation of the Alzheimer’s disease amyloid gene in hereditary cerebral hemorrhage, Dutch type. Science 1990;248:1124 –1126. 29. Van Broeckhoven C, Haan J, Bakker E, et al. Amyloid beta protein precursor gene and hereditary cerebral hemorrhage with amyloidosis (Dutch). Science 1990;248:1120 –1122. 772 Annals of Neurology Vol 50 No 6 December 2001 30. Bornebroek M, Haan J, Van Buchem MA, et al. White matter lesions and cognitive deterioration in presymptomatic carriers of the amyloid precursor protein gene codon 693 mutation. Arch Neurol 1996;53:43– 48. 31. Bornebroek M, Van Buchem MA, Haan J, et al. Hereditary cerebral hemorrhage with amyloidosis-Dutch type: better correlation of cognitive deterioration with advancing age than with number of focal lesions or white matter hyperintensities. Alzheimer Dis Assoc Disord 1996;10:224 –231. 32. Haan J, Lanser JB, Zijderveld I, et al. Dementia in hereditary cerebral hemorrhage with amyloidosis-Dutch type. Arch Neurol 1990;47:965–967. 33. Maat-Schieman ML, Radder CM, van Duinen SG, et al. Hereditary cerebral hemorrhage with amyloidosis (Dutch): a model for congophilic plaque formation without neurofibrillary pathology. Acta Neuropathol (Berl) 1994;88:371–378. 34. Haan J, Van Broeckhoven C, van Duijn CM, et al. The apolipoprotein E epsilon 4 allele does not influence the clinical expression of the amyloid precursor protein gene codon 693 or 692 mutations. Ann Neurol 1994;36:434 – 437. 35. Yamaguchi H, Haga C, Hirai S, et al. Distinctive, rapid, and easy labeling of diffuse plaques in the Alzheimer brains by a new methenamine silver stain. Acta Neuropathol (Berl) 1990; 79:569 –572. 36. Braak H, Braak E. Neuropathological stageing of Alzheimerrelated changes. Acta Neuropathol (Berl) 1991;82:239 –259. 37. Mirra SS, Hart MN, Terry RD. Making the diagnosis of Alzheimer’s disease. A primer for practicing pathologists. Arch Pathol Lab Med 1993;117:132–144. 38. Vinters HV, Natte R, Maat-Schieman ML, et al. Secondary microvascular degeneration in amyloid angiopathy of patients with hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA- D). Acta Neuropathol (Berl) 1998;95:235–244. 39. Maat-Schieman ML, Yamaguchi H, van Duinen SG, et al. Agerelated plaque morphology and C-terminal heterogeneity of amyloid beta in Dutch-type hereditary cerebral hemorrhage with amyloidosis. Acta Neuropathol (Berl) 2000;99:409 – 419. 40. Greenberg SM, Vonsattel JP. Diagnosis of cerebral amyloid angiopathy. Sensitivity and specificity of cortical biopsy. Stroke 1997;28:1418 –1422. 41. Bornebroek M, Haan J, van Duinen SG, et al. Dutch hereditary cerebral amyloid angiopathy: structural lesions and apolipoprotein E genotype. Ann Neurol 1997;41:695– 698. 42. Natte R, Vinters HV, Maat-Schieman ML, et al. Microvasculopathy is associated with the number of cerebrovascular lesions in hereditary cerebral hemorrhage with amyloidosis, Dutch type. Stroke 1998;29:1588 –1594. 43. Zarow C, Zaias B, Lyness SA, et al. Cerebral amyloid angiopathy in Alzheimer disease is associated with apolipoprotein E4 and cortical neuron loss. Alzheimer Dis Assoc Disord 1999;13: 1– 8. 44. Haan J, Bakker E, Jennekens-Schinkel A, et al. Progressive dementia, without cerebral hemorrhage, in a patient with hereditary cerebral amyloid angiopathy. Clin Neurol Neurosurg 1992; 94:317–318. 45. Vinters HV, Ellis WG, Zarow C, et al. Neuropathologic substrates of ischemic vascular dementia. J Neuropathol Exp Neurol 2000;59:931–945. 46. Esiri MM, Wilcock GK, Morris JH. Neuropathological assessment of the lesions of significance in vascular dementia. J Neurol Neurosurg Psychiatry 1997;63:749 –753.