Leuko-Araiosis and cerebral hypoperfusion compared between patients with ischemic vascular dementia and normal elderly volunteers

J Stroke Cerebrovasc Dis

1992;2:180-188 © 1992 National Stroke Association

Leuko-Araiosis and Cerebral Hypoperfusion Compared Between Patients with Ischemic Vascular Dementia and Normal Elderly Volunteers [un Kawamura, M.D., John Stirling Meyer, M.D., Yasuo Terayama, MD., and 'Susan Weathers, M.D.

Lesions in white matter of unknown origin among the elderly have been of interest for many years, and their pathogenesis and clinical significance need clarification. Local cerebral blood flow was measured among patients with ichemic vascular dementia (IVD, n = 38) and compared with age-matched normal volunteers (n = 18) utilizing xenon-enhanced computed tomography (CT). Volume ratios for total leuko-araiosis, as well as volume ratios for apparently normal white and gray matter, were determined by plain, noncontrasted CT densitometry throughout slices of brain examined later during stable xenon inhalation enabling perfusion values for each compartment to be compared. Volume ratios for totalleukoaraiosis to total brain parenchyma were twice as large among patients with IVD (12.0 ± 5.6%) compared with elderly normal volunteers (6.0 ± 2.7%). However, cerebral perfusion values within regions of leuko-araiosis compared to "normal" white matter were decreased to the same degree among patients with IVD (14 ± 6 mlllOO g brain/min) and among elderly normal volunteers with leukoaraiosis (13 ± 5). Local cerebral blood flow values were reduced for all regions of brain examined among IVD patients compared with age-matched normals. Among patients with IVD, multiple regression analyses correlated increased volumes of leuko-araiosis with (a) advancing age, (b) hypertension, and (c) reduced perfusion in vascular territories supplying the putamen. Hypoperfusion within deep cerebral territories correlates with pathogenesis of leuko-araiosis among patients with ischemic vascular dementia. Key Words: Cerebral blood f1owLeuko-araiosis-Ischemic vascular dementia.

Abnormalities within white matter of unknown origin detected by neuroimaging among the normal elderly and the elderly with dementia have been of interest for many years. Hachinski suggested the From the Cerebral Blood Flow Laboratory and INeuroradiology Service, Department of Veterans Affairs Medical Center and Departments of Neurology and lRadiology, Baylor College of Medicine, Houston, TX, U.S.A. Address correspondence and reprint requests to Dr. J. S. Meyer at Cerebral Blood Flow Laboratory, Department of Veterans Affairs Medical Center, 2002 Holcombe Boulevard, ISlA, Houston, TX 77030, U.S.A. 180

J STROKE CEREBROVASC DIS, VOL.

2, NO.4, 1992

designation "leuko-araiosis" to stress that they may have possible but unknown clinical relevance (l). Later it was reported that regions of leuko-araiosis detected by computed tomography (CT) or magnetic resonance imaging (MRI) correlated with risk factors for stroke, aging, and cognitive impairments (2-9). It was then postulated that the pathogenesis of leukoaraiosis among patients with ischemic vascular dementia may be related to cerebral hypoperfusion and ischemia (10,11). However, leuko-araiosis also occurs among elderly normals and patients with senile dementia of the Alzheimer type, suggesting that other

LEUKO-ARAIOSIS AND HYPOPERFUSION IN DEMENTIA

factors, including Wallerian degeneration and amyloid angiopathy, may also contribute to leuko-araiosis (12,13) . In order to best define relationships between cerebral hypoperfusion and leuko-araiosis, quantitative measurements of volumes of leuko-araiosis should be directly correlated with measurements of their local cerebral blood flow (LCBF). To date, few quantitative investigations of leuko-araiosis have been correlated with measures of cerebral blood flow (14-17), and none are available where LCBF has been measured within total volumes of leuko-araiosis estimated throughout entire brain slices. The present investigation was designed to measure total volumes of leuko-araiosis in brain tissue slices according to newly defined CT criteria and to compare totalleuko-araiosis perfusion values with similar measurements made in volumes of "normal" white and gray matter utilizing the CT-CBF stable xenon inhalation method in order that the results could be compared between groups of age-matched normal volunteers and patients with ischemic vascular dementia (IVD).

Subjects As summarized in Table 1,38 elderly patients with IVD (aged 68.0 ± 10.6 years) were admitted to the study, and results were compared with 18 neurologically and cognitively normal, age-matched volunteers (aged 67.4 ± 9.9 years). All subjects signed informed consent according to protocols approved annually by Institutional Review Boards of Baylor College of Medicine and the Department of Veterans Affairs Medical Center, Houston, TX. All subjects underwent similar assessments, which included general medical and neurological examinations, Cognitive Capacity Screening Examinations (CCSE) (18,19), Hachinski Ischemic Scores (20,21), and clinical and laboratory tests to determine whether any associated risk factors for stroke were present,

Age,gender, and cognitive capacity screening examination (CCSE) scores amollg IVDpatients and age-matched normal volunteers

Table 1.

No. of cases Mean age Range Gender (M/F) CCSE scores

IVD patients

Normal volunteers

38 68.0 ± 10.6 37-92 23/15 20.6 ± 6.8

18 67.4 ± 9.9 57-86 6/12 28.8 ± 1.7

IVD, ischemic vascular dementia.

because risk factors for stroke have been implicated in the pathogenesis of leuko-araiosis. Diagnosis of dementia was made according to recommendations of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) (22). CCSE scores for demented patients were also required to remain below 24 on serial examinations. It has been shown that CCSE scores 25 and below provide reliable, quantitative indices for dementia, which correlate well with impairments documented by more extensive neuropsychological test batteries, including the Wechsler Adult Intelligence Scale (Revised), mini-mental status questionnaire, and other neurobehavioral test scores (19). CCSE scores have proven particularly useful for tracking the cognitive impairments in patients with IVD (23). Diagnosis for probable IVD met criteria proposed by the State of California Alzheimer's Disease Diagnostic and Treatment Centers (24) plus the presence of focal neurological signs, Hachinski ischemic scores above 7, and CCSE values below 25 (23). Risk factors analyzed in patients with IVD were hypertension (n = 28), heart disease (n = 21) , hyperlipidemia (n = 22), diabetes mellitus (n = 14), and smoking (n = 9). Types of strokes among patients with IVD classified according to recommendations of the National Institute of Neurological Disorders and Stroke (25) were as follows: lacunar stroke (n = 28), lacunar stroke plus stenosis or occlusion of internal carotid artery (n = 8), atherothrombotic stroke plus stenosis or occlusion ofinternal carotid artery (n = 1), and vasculitis (n = 1). Stenosis or occlusion of the internal carotid arteries was considered to be present when ultrasonographic or angiographic procedures disclosed occlusion or stenosis by 80% or more. Criteria for normal volunteers included (a) normal neurological and CCSE examinations; (b) absence of history of stroke or neurological or psychiatric disorders; and (c) exclusion of cerebral abnormalities other than leukoaraiosis and age-related atrophy by CT and MRI.

Methods LCBF was measured in all subjects utilizing the xenon-enh ance d CT-CBF method, which consisted of serial CT scanning for 8 min while inhaling 27% xenon gas as the contrast agent utilizing a commercially available xenon gas delivery system (Enhancer 3000, Diversified Diagnostic Products Inc., Houston, TX). Methodological details have been reported previously (26). One of two high-resolution, rapid CT scanners (Siemens DR version H, Siemens Medical

J STROKE CEREBROVASC DIS, VOL. 2, NO.4,

1992

181

J.

KAWAMURAIT AL

Systems Inc., Iselin, NJ, or Synerview SX-1200, Picker International Inc., Highland Heights, OH) was utilized for LCBF measurements. The CT scanning parameters were 96 kVp, 540 rnA, 8-mm slice thickness with 5-s scanning times for the Siemens Somatom DR-H, and 130 kVp, 140 rnA, 10 mm with 2-s scanning times for the Picker Synerview. Patients and volunteers fasted for 6 h prior to CBF measurements. In all subjects, CT levels for CBF measurements were selected to include frontal, temporal, and occipital cortex , caudate nucleus, putamen, and thalamus in parallel with the canthomeatal line. With the faster CT scanner (Synerview SX-1200) available later, LCBFvalues were measured at a second level 10 mm above the basal ganglia to include the centrum semiovale, bodies of the lateral ventricles, parietal cortex, and white matter. Measurements at the second, higher level were obtained in 16 of 38 patients with IVD and 10 of 18 elderly normal volunteers. After two baseline, noncontrasted CT scans for each CT level were obtained, seven progressively enhanced CT scans at each level were recorded at J-min intervals between the second and eighth minutes of xenon gas inhalation. End-tidal partial pressures for xenon gas and carbon dioxide were recorded on a polygraph. LCBFvalues were generated as color-coded images for each brain slice utilizing a desktop computer programmed so that two noncontrasted control scans acted as baseline and seven postenhancement scans determined local xenon tissue saturation curves according to Kety's formula (27). The original CT images (512 X 512 pixels) were compressed to 128 X 128 before calculating LCBF values, with pre- and postcalculation smoothing so that the size of each voxel measured was 1.82 X 1.82 X mm utilizing Siemens DR-H and 1.88 X 1.88 X 10 mm utilizing the Picker SX-1200. By identifying specific anatomical locations on the plain CT images and utilizing the cursor, LCBF values were separated into nine representative regions (including frontal, temporal, and occipital cortex, caudate nucleus, putamen and thalamus, frontal and occipital white matter) for each hemisphere at the level of the basal ganglia and two additional regions (including parietal cortex and white matter) at the level of the bodies of lateral ventricles (28,29). Pooled mean LCBF values were calculated for each representative region. EEG and EKG were monitored throughout the CBF measurements. Methodological limitations of the xenon CT-CBF method include anesthetic effects of xenon gas, errors due to CT noise, motion artifact, tissue overlap, and radiation exposure. However, these limitations and 182

/ STROKE CEREBROVASCDIS, VOL. 2, NO.4, 1992

measurement errors were minimized by utilizing modern, high-resolution, rapid CT scanners, short exposure times, and low concentrations of xenon gas as described previously (26). Computerized densitometry was utilized to determine volumes of leuko-araiosis compared to volumes of "normal" white and gray matter and cerebrospinal fluid from their Hounsfield unit numbers determined for each CT voxel obtained from the baseline preenhanced CT images. Previously, we have reported densitometric estimates for frontal leuko-araiosis (28,29). At that time, however, only leuko-araiosis surrounding the anterior horns of lateral ventricles was measured. In the present study, the same densitometric methods were utilized for the first time to analyze volumes of leuko-araiosis present throughout entire CT images. By these means, it was possible to quantitate total volumes of leukoaraiosis throughout each brain slice and correlate results with volumes of "normal" white and gray matter corrected for brain atrophy and differences in head size by subtracting cerebrospinal fluid volumes from parenchymal volumes and expressing tissue volumes as ratios of remaining brain parenchyma to intracranial volumes. Independent determinations for volumes ofleukoaraiosis by two of the authors utilizing this new method and comparing their results for estimated volumes of "normal" white and gray matter and cerebrospinal fluid, agreed well when threshold values for leuko-araiosis were set between 17 and 29 Hounsfield numbers when threshold values for "normal" white matter were set between 30 and 36, for gray matter were set between 37 and 60, and for cerebrospinal fluid were set between 0 and 17. This held true for both CT scanners, which were calibrated in the same manner using a standard phantom. Threshold values for white matter hypodensity (i.e., Ieuko-araiosis) among patients with Binswanger's disease have been reported previously by other investigators to be below 30 Hounsfield numbers, which agrees well with independent estimates reported here (30). As stated above, volumes for each compartment were expressed as percentage ratios to total volumes of brain parenchyma for each CT level, which were derived by subtracting measured volumes of cerebrospinal fluid from total brain volumes at each level. Standard deviations for Hounsfield numbers were ±0.2 when a standard CT phantom was scanned rapidly 11 times. Similar calibration values in Hounsfield numbers and the same reproducibility errors were determined for each of the two CT scanners (26). Based on their Hounsfield numbers, cerebral

LEUKO-ARAIOSISAND HYPOPERFUSION IN DEMENTIA

compartments including leuko-araiosis, if present, and "normal" white and normal gray matter, were identified and their cerebral perfusions were then measured and compared. The validity of CT densitometry for identifying white matter lesions has been confirmed by earlier comparisons of magnetic resonance imaging (MRI) with CT imaging among patients with both IVD and senile dementia of the Alzheimer's type (31-40). MRI is more sensitive for identifying white matter lesions from T2-weighted MR scans, however, densitometric measurements are not possible utilizing MRI, since T2 settings for optimal discriminations of leuko-araiosis vary from patient to patient (41,42). Nevertheless, when CT evidence of leuko-araiosis was identified and MR scans were available for comparison in the same individuals, zones of leukoaraiosis were always confirmed by MRI imaging. Previously published multivariate analyses of 25 patients with IVD, which included 6 patients who participated in the present study, revealed that the presence of diffuse periventricular high-intensity lesions in white matter (leuko-araiosis) detected by MRI and CT is one of a number of discriminant functions that may be considered for differentiating IVD from senile dementia of the Alzheimer's type (42). In the present study, MRI was performed in 16 of the patients with IVD. The MR patterns and extent of leuko-araiosis were consistent in all with CT interpretations of leuko-araiosis.

Biostatistical Analyses In order to estimate the relative contributions of aging, levels of mean arterial blood pressure, and reductions in LCBF to the pathogenesis of leukoaraiosis, multiple linear regression analyses were applied. The severity of leuko-araiosis was regarded as the dependent variable, whereas age, mean arterial blood pressure, and local perfusion values for nine representative cerebral regions (frontal, temporal, and occipital cortex, caudate nucleus, putamen, thalamus, frontal, occipital, and capsular white matter) served as 11 explanatory variables. LCBF values for parietal cortex and parietal white matter were not included among the explanatory variables because of insufficient numbers for statistical analysis, for technical reasons described in the first paragraph of the "Methods" section. A statistical software package (SPSS/PC +, SPSS Inc., Chicago, 11) was used for the analyses. Stepwise protocols designed for examining independent variables were applied, assigning a standardized score for each variable. Criteria in F values for entry to and removal from the equation were 3.84 and 2.71, respectively. Data are presented as mean ± SD. Statistical analyses were performed by Student's t test.

Results Figure 1 illustrates regions of leuko-araiosis at the level of the basal ganglia in a 68-year-old man with

~igure 1. . Plain CTimage (left) alldCTdensitometri~ image (~jddle) showextensiveleuko-araiosis ofthewhitematter withcorrespond~ng red~lctlOns of LCBF m~p supenmposed on the CT Image (nght) at the levelof the basal ganglia in a severelyhypertensive man with Is~hemlc vascular den:entla and multiple lacunar strokes (arrows). Hesuffersfromsubcortical arteriosclerotic leukoencephalopathy of the BI~lswanger t}lpe. H~ IS 6~ yea:~ ofage, nowIIIl.able to w~lk and u:heel~hair-ridden with ahistoryofmultiplestrokes, incontinence, apraxia of gait, anda.bu!la. RegrOlls I~entified by CTdens.rtometry, III the middle ,~nage, as leuko-araiosis aredisplayed in white. Theratio oftiolumesof

leuko-araiosis to total bralll parenchyma at this levelwas 22%. CBFIS severely reduced throughout the brainbut particularly in tilewhite matter.

J STROKECEREBROVASCDIS, VOL.2, NO.4, 1992

183

J. KAWAMURAIT AL.

NORMALS

D

NORMALS

IVO

_

IVO

LA

LA

WM

WM

GM

GM 1 - - - - - - - - - '

o

20

40

60

80

100

(%)

VOLUME PERCENTAGE RATIOS

Figure 2. Volume percentage ratios for leuko-araiosi s to total brain volume measured at the levelsofbasal ganglia and centrum semiovale among patients witll IVDcompared with age-matched normal volunteers. Values for ratios of leuko-araiosis to brain volumeamongpatients with IVD weretwiceas large as those estimated amongage-matched normal volunteers. Significantdifferencescompared withage-matched nonnalvolunteersare indicated by ,,*"p < 0.001.

severe IVD, considered by both clinical and MR and CT neuroimaging criteria to be suffering from Binswanger's type of subcortical arteriosclerotic leukoencephalopathy (43,44). LCBF measurements disclose severe, patchy, and confluent reductions of cerebral perfusion throughout both hemispheres primarily affecting white matter. Figure 2 displays the pooled volume ratios for leuko-araiotic and normal white and gray matter for all CT slices examined among patients with IVD compared to pooled values among age-matched normal volunteers. Ratios for leuko-araiosis to brain parenchyma among patients with IVD are much greater than among age-matched normals (p < O.OOl); however, volumes for normal white and gray matter were not significantly different. Figure 3 exhibits pooled cerebral perfusion values for each of the above three compartments examined among patients with IVD and compared to elderly normal volunteers. Cerebral perfusion in volumes of leuko-araiosis were reduced to the same degree among both IVD patients and normal volunteers. However, LCBF values within remaining gray and white matter among patients with MID were signif184

J STROKECEREBROVASCDIS, VOL. 2, NO.4, 1992

o

20

40

60

C8F VALUES (ml/l DOg/min)

Figure 3. CBFvalues for gray matter, "normal" tohite matter, and volumes of leuko-araio sis compared between patients with IVD and elderly normal volunteers. CBF in volumes of leukoaraio sis are significantly decreased compared to "normal" uihlte matteramongbothIVD patientsand elderly normals. In patients uiith IVD, CBFvaluesforgrayand "nonnal" tohite matteraresignificantly reduced compared to nonnal oolunteers. Statistically significantdifferences are indicated as follows: ,,*"p < 0.001 compared with age-matched normal volunteers; u#p < 0.001 compared witll "normal" wlzitematter.

icantly reduced compared with age-matched normals. Figure 4 illustrates pooled perfusion values for 11 representative regions of the brain among patients with IVD compared to normal volunteers. LCBF values for all regions examined were reduced among patients with IVD compared to normal volunteers. Associations between volume ratios for leukoaraiosis at the level of basal ganglia to blood pressure, to age, and to LCBFvalues for the nine representative cerebral regions were assessed by multiple linear regression analyses, and positive results are summarized in Table 2. CCSE scores were not included in the explanatory variables, since significant differences in CCSE scores between IVD patients and elderly normals were brought about by the selection criteria. Furthermore, range distributions for CCSE scores among IVD patients were narrow. Among 11 explanatory variables tested in patients with IVD, advanced age, elevated mean arterial blood pressure, and reductions of putaminal LCBF values correlated with severity of leuko-araiosis, Among age-matched normal volunteers, correlations were

LEUKO-ARAIOSISAND HYPOPERFUSION IN DEMENTIA

mean arterial blood pressures between the two groups (normals, 97 ± 9 mm Hg; MID, 98 ± 11). There were no EEG changes during inhalation of xenon gas .

•••

FC

TC PC

•••

•••

OC

Discussion

•••

CAU

••

•••

PUT

•••

THA

••

FW

PW

-

•

OW INT

o

20

NOR

0

••• •••

IVO

40

60

80

LCBF VALUES (mJ/l DOg/min) LCBF values for 11 representat iue braill regions ill patientswitltIVDcompared witltelderly normal volunteers. LCBF values forallcerebral regions measured weresignificantlydecreased ill patients witlt IVD compared to age-matched normals. FC, [ronia! cor/ex; TC,temporal cor/ex; pc, parietal cortex; OC, occipital cor/ex; CAU,caudate nucleus: PUT, putamen; THA, thalamus; FW, [rontal white matter; PW, parietal white matter; OW, occipital white matter; INT, internal capsule. Statistically sigllificallt differellces fromage-matched normal volunteers are indicated as follows: "p < 0.01; '''p < 0.001. Figure 4.

not found between the severity of leuko-araiosis and any explanatory variables of those tested. During CBF measurements, partial pressures for end-tidal carbon dioxide (PEC0 2) did not differ between normals (33.4 ± 4.3 mm Hg) and patients with IVD (32.2 ± 2.4), nor were there differences in

Table 2.

Abnormal white matter lesions detected by CT or MRI have been termed "leuko-araiosis" to stress their unknown cause and to emphasize the need for investigations concerning their nature (1-4) . A number of studies have been concerned with clinical and pathological correlates of leuko-araiosis. However, up to the present report, quantitative analyses of leuko-araiosis utilizing standard densitometric limits in Hounsfield number have not been available (28,29). Recently, LCBF values were correlated in this laboratory with the severity of leuko-araiosis in the frontal lobe, and it was concluded that in IVD patients hypoperfusion of the basal ganglia is associated with leuko-araiosis (29) . However, there were some methodological problems concerned with our earlier report. Because densitometric analysis was limited to white matter in the frontal lobe, this allowed for the possibility of contamination of estimated volumes of leuko-araiosis by "normal" white matter and cerebrospinal fluid because of partial volume effects. In the present study, all available white matter on each CT slice was analyzed by densitometry and compared with remaining volumes of"normal" white and gray matter. Furthermore, brain atrophy was adjusted for by measuring cerebrospinal fluid volumes and determining volume ratios to remaining intracranial parenchyma. Comparison of volumes of leuko-araiosis with "normal" white and gray matter confirmed that cross-contamination due to partial volume effects was minimized.

Multiple regression analyses relating to severity ofleuko-araiosis at the basal ganglia level

Factor

B

SEB

Beta

Statistical significance

Age Blood pressure LCBF for putamen

0.220 0.132 -0.341

0.065 0.061 0.102

0.433 0.278 -0.416

P < 0.002 P < 0.05 P < 0.005

B, partial regression correlation coefficient for each variable; SEB, standard error of B; Beta, partial regression correlation coefficients expressed in standardized Z-score form s. The coefficient for determination (R2) , 0.475; adjusted for degrees of freedom (adjusted R2), 0.429; multiple correlation coefficient, 0.689. Results were statistically significant (p < 0.0001, F-test).

/ STROKECEREBROVASCDIS,VOL. 2, NO.4, 1992

185

J. KAWAMURAETAL. It became apparent that total volumes of leukoaraiosis in patients with IVD are significantly greater than among elderly normals and that associations exist among leuko-araiosis, aging, and hypertension. Results are in agreement with previous publications emphasizing the frequent association of leukoaraiosis with vascular dementia (2,3,5,9,31,32), advancing age (32-34), and hypertension (4,14,35). These reports causally related cerebral vascular disease and ischemia to leuko-araiosis. The present study confirms that cerebral perfusion is reduced in volumes of leuko-araiosis and that hypoperfusion correlates with the extent of leuko-araiosis particularly if hypoperfusion involves the vascular territories supplying the putamen. The etiology and pathogenesis of leuko-araiosis is still a matter of dispute. This is because several factors probably contribute to leuko-araiosis. These include ischemia and hypoperfusion within the territories of the deep penetrating cerebral arterioles resulting in ischemic demyelination in addition to Wallerian degeneration secondary to cortical neuronal loss (11). Kobari et al. (16) correlated the severity of leukoaraiosis detected by both MRI and CT with reductions of LCBF values for cortical and subcortical gray matter. Fazekas and co-workers (14) reported reductions of slow flow components (F2) , measured by the intravenous t33Xe method, among asymptomatic elderly volunteers with white matter lesions. Herholz et a1. (17) demonstrated that large white matter lesions in deep brain structures were associated with diffuse decreases of cortical CBF among patients with atherosclerotic disease of the internal carotid arteries. Delpla and associates (36) correlated neuropsychological test results with positron emission tomographic (PET) measurements of (CBF) and oxygen consumption among elderly volunteers with leukoaraiosis and risk factors for stroke (36). They concluded that neuropsychological test results indicating subtle frontal lobe dysfunctions correlated with CBF reductions for both white and gray matter in the frontal lobe associated with reduced oxygen consumption in the same locations. Several pathological studies have reported associations between leuko-araiosis and ischemic changes in subcortical gray and white matter, which were accounted for by the special vulnerability of periventricular border zones, which have poor collateral arterial supply (37-39). In the present study, correlations between severity of leuko-araiosis and hypoperfusion in subcortical gray matter among IVD patients were confirmed. Considering all reported results together, the isch186

J STROKE CEREBROVASCDIS, VOL. 2, NO.4,

1992

emic-hypoperfusion hypothesis appears to play an important part in the pathogenesis of leuko-araiosis among patients with IVD. Cortical infarctions present in our patients with IVD produced focal LCBF reductions within the cortex, but these apparently contributed little to the development of leukoaraiosis, since reductions of cortical gray matter flow did not correlate with the severity of leukoaraiosis. Perfusion values measured in regions of leukoaraiosis among normal volunteers revealed similar reductions as in IVD, but they were strictly limited to small volumes of leuko-araiosis present, whereas in IVD patients volumes of leuko-araiosis were much larger. Further investigations, including local cerebrovascular responsiveness to hypercapnia, within volumes of leuko-araiosis, should be tested to confirm the ischemic nature of leuko-araiosis or whether hypoperfusion is also brought about by reduced metabolic demands of gray and white matter resulting from disconnections (11,40). Longitudinal studies are also under way in this laboratory to determine whether leuko-araiosis is a progressive disorder that correlates temporally with cognitive declines. It is concluded that leuko-araiosis among patients with IVD correlates with advancing age, hypertension, and hypoperfusion within affected zones of white matter and neighboring putaminal territories. Hypoperfusion within volumes of leuko-araiosis are similar among IVD patients and age-matched normals, but volumes of leuko-araiosis are much greater in IVD. Taken together, reported evidence suggests that ischemia and hypoperfusion supplying deep brain structures playa part in the pathogenesis of leuko-araiosis, particularly among patients with vascular dementia. A similar, but mild process, also occurs in many elderly normal volunteers, and this may put them at risk for later cognitive declines. Acknowledgment: Dianne B. Bailey processed the manuscript. Ada Hinds, CRT, and James Simon, CRT, provided technical assistance with CT scanning. John Thornby, Ph.D., Professor of Psychiatry, Baylor College of Medicine, and Biostatistical Consultant, Houston VAMC, provided statistical assistance.

References 1. Hachinski VC, Potter P, Merskey H. Leuko-araiosis. Arch Nellrol1987;44:21-3. 2. Steingart A, Hachinski VC, tau C, et al. Cognitive and neurological findings in subjects with diffuse white matter lucencies on computed tomographic scan (leuko-araiosis). Arch Nellrol1987;44":32-S.

LEUKO-ARAIOSISAND HYPOPERFUSION IN DEMENTIA 3. Steingart A, Hachinski VC, Lau C, et aI. Cognitive and neurologic findings in dementia patients with diffuse white matter lucencies on computed tomographic scan (leuko-araiosls). Arcll NellroI1987;44:36-9. 4. Inzitari D, Diaz F, Fox A, et al. Vascular risk factors and leuko-araiosis. Arcli NellroI1987;44 :42-7. 5. Gupta SR, Naheedy MH, Young IC, Ghobrial M, Rubino FA, Hindo W. Periventricular white matter changes and dementia. Clinical, neuropsychological, radiological, and pathological correlation. Arch Neurol 1988 ;45 :637-41. 6. Lechn er H, Schmidt R,Bertha G, Iustich E,Offenbacher H, Schneider G. Nuclear magnetic resonance image white matter lesions and risk factors for stroke in normal individuals. Stroke 1988;19:263-5. 7. Rao SM, Mittenberg W, Bernardin L, Haughton V, Leo GJ. Neuropsychological test findings in subjects with leuko-araiosis. Arch NeuroI1989;46 :40-4. 8. Janota I, Mirsen TR, Hachinski VC, Lee DH, Merskey H. Neuropathologic correlates of leuko-araiosis, Arch NeuroI1988;46:1124-8. 9. Kertesz A, Polk M, Carr T. Cognition and white matter changes on magnetic resonance imaging in dementia. Arch NeuroI1990;47:387-91. 10. Chimowitz MI, Awad lA, Furlan AJ. Periventricular lesions on MRI. Facts and theories. Stroke 1989;20: 963-7. 11. Hastak SM, Hachinski vc: Leuko-araiosis: current concepts. In: Bes A, Geraud G, eds. Current problems ill Ilellrology, 12. Circulation cerebrale et uieillissement. Paris : John Libbey Eurotext, 1990:81-9. 12. Brun A, Englund E. A white matter disorder in dementia of the Alzheimer type: a pathoanatomical study. AIlIlNellro11986; 19:253-62. 13. Leys D, Pruvo P, Parent M, et aI. Could Wallerian degeneration contribute to "leuko-araiosis" in subjects free of any vascular disorder? / Neurol Neurosurg Psychiatry 1991;54:46-50. 14. Fazekas F, Niederkorn K, Schmidt R, et al. White matter signal abnormalities in normal individuals: correlation with carotid ultrasonography, cerebral blood flow measurements, and cerebrovascular risk factors. Stroke 1988;19:1285-8. 15. Kobari M, Meyer JS, Ichijo M. Leuko-araiosis, cerebral atrophy, and cerebral perfusion in normal aging. Arcli NellroI1990;47:161-5. 16. Kobari M, Meyer JS, Ichijo M, Oravez WT. Leukoaraiosis: correlation of MR and CT findings with blood flow, atrophy, and cognition. A/NR 1990;11:273-81. 17. Herholz K, Heindel W, Rackl A, et al, Regional cerebral blood flow in patients with leuko-araiosis and atherosclerotic carotid artery d iseas e. Arch Neurol 1990;47:392-6. 18. Jacobs JW, Bernhard MR, Delgado A, Strain JJ. Screening for organic mental syndromes in the medically ill. AIm Intern Med 1977;86:40-6. 19. Hershey LA, Jaffe DF, Greenough PG, Yang FL Validation of cognitive and functional assessment instruments in vascular dementia. lilt] Psyclliatr Med 1987; 17:183-92. 20. Hachinski VC, Iliff LD, Zilhka E, et al. Cerebral blood flow in dementia. Arcn NellroI1975;32:632-7. 21. Rosen WG, Terry RD, Fuld PA, et al. Pathological verification of ischemic score in differentiation of dernentias . A,m NeuroI1980;7:486-8 .

22. American Psychiatric Association. Diagnostic alld statistical manual of melltal disorders , 3rd ed . Washington, DC: American Psychiatric Association, 1987:103-7. 23. Hershey LA. Dementia associated with stroke. Stroke 1990;21 (1I):1I.9-II.11. 24. Chui HC, Victoroff]l, Margolin D, et al. Criteria for the diagnosis of ischemic vascular dementia proposed by the State of California Alzheimer's Disease Diagnostic and Treatment Centers. Nellrology 1992;42:473-80 . 25. Whisnant IP, Basford JR, Bernstein EF, et al. Special report from the National Institute of Neurological Disorders and Stroke. Classification of cerebrovascular disease III. Stroke 1990;21:637-76. 26. Meyer IS, Shinohara T, Imai A, et al, Imaging local cerebral blood flow by Xe-enhanced computed tomography-technical optimization procedures. Neuroradiology 1988;30:283-92 . 27. Kety SS. The theory and applications of the exchange of inert gas at the lungs and tissu es. Pharmacol Rev 1951;3:1-41. 28. Meyer JS, Kawamura I, Ichijo M, Kobari M, Terayama Y. Leuko-araiosis, cerebral atrophy, and blood flow in elderly patients with dementia. AIm Neurol 1990; 28:254 . 29. Kawamura J, Meyer IS, Terayama Y, Weathers S. Leukoaraiosis correlates with cerebral hypoperfusion in vascular dementia. Stroke 1991;22:609-14 . 30. McQuinn BA, O'Leary DH . White matter lucencies on computed tomography, subacute arteriosclerotic encephalopathy (Binswanger's d isease), and blood pressure. Stroke 1987;18:900-5. 31. George AE, De Leon MJ, Gentes CI, et al. Leukoencephalopathy in normal and pathologic aging: 1. CT of brain lucencies. A]NR 1986;7:561-6. 32. Valentine AR, Moseley IF, Kendall BE. White matter abnormality in cerebral atrophy: c1inicoradiological correlations. / Neurol Neurosurg Psychiatry 1980;43: 139-42. 33. Goto K, Ishii N, Fukasawa H. Diffuse white-matter disease in the geriatric population. Radiology 1981; 141:687-95. 34. Kinkel WR, Jacobs L, Polachini I, Bates V, Heffner RR. Subcortical arteriosclerotic encephalopathy (Binswanger's disease). Computed tomographic, nuclear magnetic resonance, and clinical correlations. Arch NellroI1985;42:951-9. 35. Awad lA, Spetzler RF, Hodak lA, Awad CA, Carey R. Incidental subcortical lesions identified on magnetic resonance imaging in the elderly. I. Correlation with age and cerebrovascular risk factors . Stroke 1986;17: 1084-9. 36. Delpla PA, Zatorre R, Meyer E, et al. Leucoaraiose et dysfonctionnement frontal precoce chez Ie sujet age non dement: approche neuropshychologique et par la camera a positons. In: Bes A, Geraud G, cds, Current problems ill neurology, 12. Circlliat ioll cerebrale et uieillissemeni. Paris: John Libbey Eurotext, 1990;123-39. 37. Awad lA, Johnson PC, Spetzler RF, Hodak JA. Incidental subcortical lesions identified on magnetic resonance imaging in the elderly. II. Postmortem pathological correlations Stroke1986;17:1090-7 . 38. Braffman BH, Zimmerman RA, Trojanowski IQ, Gonatas NK, Hickey WF, Schlaepfer WW. Brain MR: pathologic correlation with gross and histopathology.

] STROKECEREBROVASCDIS, VOL. 2, NO.4, 1992

187

I. KAWAMURAET AL. 2. Hyperintense white-matter foci in the elderly. AJNR 1988;151:559-66. 39. De Reuck J. The human periventicular arterial blood supply and the anatomy of cerebral infarctions. Ellr NellroI1971;5:321-34. 40. Meyer JS. Does diaschisis have clinical correlation? Editorial. Mayo Clill Proc 1991;66:430-2. 41. Erkinjuntti T, Ketonen L, Sulkava R, Siponen J, Vuorialho M, Iivanainen M. Do white matter changes on MRI and CT differentiate vascular dementia from Alzheimer's disease? I Neurol Neurosurg Psyc1liatry 1987;50:37-42.

188

I STROKE CEREBROVASCDIS, VOL. 2, NO.4,

1992

42. Reed KM, Rogers RL, Meyer J5. Cerebral magnetic resonance imaging compared in Alzheimer's and multi-infarct dementia. I Neuropsqchiair Clill Neurosci 1991;3:51-7. 43. Caplan LR,Schoene We. Clinical feature of subcortical arteriosclerotic encephalopathy (Binswanger disease). Nellrology 1978;28:1206-15; Clill Neurosci 1991;3: 51-7. 44. Roman Gc. Senile dementia of the Binswanger type. A vascular form of dementia in the elderly. lAMA 1987;258:1782-8.