Vascular Dementia in a Population-Based Autopsy Study


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Vascular Dementia in a Population-Based Autopsy Study

  1. 1. ORIGINAL CONTRIBUTIONVascular Dementia in a Population-BasedAutopsy StudyDavid S. Knopman, MD; Joseph E. Parisi, MD; Bradley F. Boeve, MD; Ruth H. Cha, MS; Hulya Apaydin, MD;Alessandro Salviati, MD; Steven D. Edland, PhD; Walter A. Rocca, MD, MPHBackground: The validity of the clinical diagnosis of available in 89 (21%) with median age at onset of 80 yearsvascular dementia (VaD) remains suboptimal. (range, 50-96 years; 52 [58%] women). Pathological di- agnoses were AD in 45 patients (51%), pure VaD in 12Objective: To investigate clinicopathologic correla- (13%), combined AD and VaD in 11 (12%), and othertions in VaD. diagnoses in the remaining 21 patients. Criteria for VaD that required either a temporal relationship between aMethods: We used the medical records-linkage sys- stroke and dementia onset or worsening, or bilateral in-tem of the Rochester Epidemiology Project to identify in- farctions in specified locations demonstrated on imag-cident cases of dementia in Rochester, Minn, from Janu- ing results (Mayo Clinic criteria) had 75% sensitivity andary 1, 1985, through December 31, 1989. Dementia and 81% specificity for pure VaD (positive likelihood ratio,Alzheimer disease (AD) were defined by the criteria of 3.9; 95% confidence interval, 2.2-6.7). Five cases of purethe Diagnostic and Statistical Manual of Mental Disor- VaD lacked the temporal relationship and accounted forders, Fourth Edition. Vascular dementia was defined by the imperfect sensitivity of the criteria.criteria including imaging results. Pathological charac-teristics of AD were quantified by means of standard scor- Conclusions: In this population-based autopsy study,ing methods for neurofibrillary tangles and neuritic the presence of vascular pathological characteristics inplaques. Vascular pathological findings were assessed by the absence of major AD pathological findings was com-expert neuropathological opinion. mon. Pure VaD without overt clinical strokes remains a challenge for antemortem diagnosis.Results: Of 419 patients with dementia who died be-fore the study, neuropathological examination results were Arch Neurol. 2003;60:569-575 T HE CONTRIBUTION of vascu- (present study) of VaD. The unique ac- lar pathologic characteris- cess to clinical data on dementia and stroke tics to dementia is com- in a geographically defined population, to- plex and remains unclear. gether with autopsy results on 21% of in- This is reflected in the poor cident cases of dementia, enabled us to validity of clinical diagnostic criteria when study the relationships between clinical compared with postmortem diagnoses.1-7 features and neuropathological diag- However, the interpretation of clinico- noses. pathologic studies has been hampered byFrom the Departments of different sources of patients across stud- METHODSNeurology (Drs Knopman, ies. For example, the frequency of vascu-Parisi, Boeve, and Rocca), lar dementia (VaD) in autopsy series was CASE ASCERTAINMENT ANDHealth Sciences Research much higher when recruitment was from CLINICAL DIAGNOSIS(Ms Cha and Drs Edland and well-defined populations1,8 compared withRocca), and Pathology dementia clinics.5,9 The relative rarity of We used the medical records-linkage system of(Dr Parisi), Mayo Clinic and prospective studies of patients with de- the Rochester Epidemiology Project to identifyMayo Foundation, Rochester, mentia who prove to have VaD neuro- all subjects residing in Rochester who devel-Minn; Department of oped dementia during the 5-year period from pathologically limits our ability to im-Neurology, Istanbul University, January 1, 1985, through December 31, 1989.Cerrahpasa Medical School, ¸ prove clinical diagnostic criteria. Details about the study population and the iden-Istanbul, Turkey (Dr Apaydin); Our study of dementia from 1985 tification of incident cases were reported else-and Department of Neurololgy, through 1989 in Rochester, Minn, al- where.10,12-14 Age- and sex-specific incidence ratesUniversity of Verona, Verona, lowed us to describe the incidence,10 sur- for dementia, AD, and VaD in the RochesterItaly (Dr Salviati). vival,11 and clinicopathologic correlation population were reported previously.10,12 To be (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 569 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  2. 2. included in the study, patients with dementia were required to study neuropathologist (J.E.P.), and, if necessary, additional slideshave resided in Rochester in the year of onset of dementia and were made using tau, -amyloid, or -synuclein immunohisto-for at least 1 preceding year. Patients who had moved to Roch- chemical techniques.ester for the management of a preexisting dementing illness were The pathological diagnosis of AD was made using a com-excluded. A series of nondemented referent subjects from the same bination of the criteria of the Consortium to Establish a Reg-population was also identified through the records-linkage sys- istry for AD (CERAD) (based on the count and location of neu-tem11; however, because the referent subjects were not followed ritic plaques)19 and the staging system of Braak and Braak (basedup prospectively, their dementia status at the time of death was on the count and location of neurofibrillary tangles).20 The patho-unknown. Therefore, these referent subjects could not serve as logical diagnosis of AD was made when either a brain metcontrols for neuropathological studies. CERAD probable or definite AD criteria or the Braak and Braak Information for the diagnosis of dementia and for the clas- stage was IV or higher.sification of cases by type of dementia was abstracted from the The neuropathological diagnosis of VaD was made if theremedical history, neurological examinations results, and neuro- were multiple infarctions in gray matter structures from the thala-imaging study findings as collected historically in the patient dos- mus rostrally. Frank infarctions in white matter were also in-siers of the records-linkage system. Diagnostic criteria for de- cluded. Bilaterality of infarctions, but not necessarily in the samementia and for Alzheimer disease (AD) were those of the Diagnostic structures, was required. Neurofibrillary tangles that were lessand Statistical Manual of Mental Disorders, Fourth Edition (DSM- than or equal to Braak and Braak stage III and neuritic plaquesIV).15 On the basis of our previous work, we defined VaD as ei- that were considered sparse or less by CERAD criteria were re-ther (1) dementia onset or worsening within 3 months of clini- quired for a diagnosis of pure VaD. In addition, the absence ofcal stroke or (2) bilateral gray matter infarctions shown by imaging other pathological characteristics was required for a diagnosisthat fulfilled specified location criteria (critical imaging lesions).10 of VaD. When other pathological features were present in theWe refer to these as the Mayo Clinic criteria.10 We also indepen- setting of substantial vascular pathological findings, multipledently applied the following 4 sets of published diagnostic crite- diagnoses were made (eg, the combination of VaD and AD).ria for VaD: the DSM-IV criteria,15 the National Institute of Neu- The diagnosis of dementia with Lewy bodies was based on Lewyrological Disorders and Stroke and the Association Internationale body pathological changes in the substantia nigra, amygdala,pour la Recherche et l’Enseignement en Neurosciences and neocortex.21(NINDS-AIREN) probable VaD criteria,16 the International Clas-sification of Diseases and Related Health Problems, 10th Revision of STATISTICAL METHODSthe World Health Organization (ICD-10) criteria,17 and the Stateof California Alzheimer’s Disease Diagnostic and Treatment Cen- The 2 and rank sum tests were used to compare cases withters (ADDTC) criteria.18 Imaging findings were only used for the and without autopsy results, and to compare groups with dif-classification of dementia if they were obtained from 1 year be- ferent pathological diagnoses. Sensitivity, specificity, positivefore through 3 years after the onset of dementia.10 We excluded predictive value, negative predictive value, and positive like-imaging lesions that clearly postdated the onset or the early phase lihood ratio (with 95% confidence interval) were calculated toof dementia. compare clinical features or clinical diagnoses with pathologi- cal diagnoses (standard for comparison). Only cases of demen- AUTOPSY AND PATHOLOGICAL DIAGNOSIS tia were included in the calculations of diagnostic accuracy; the study did not involve a group of controls (subjects who diedPatients from our incidence series who died before the time of without dementia).this study underwent autopsy according to routine clinical prac-tice. No special brain donation program was available for de- RESULTSmentia in the study population at the time of this study. Patho-logical samples pertaining to residents of Rochester are routinelyarchived and are retrievable for additional studies. There were 482 incident cases of dementia identified in Neuropathological examinations were carried out under the Rochester from January 1, 1985, through December 31,supervision of one of 2 board-certified neuropathologists (J.E.P. 1989. Eighteen percent of these were diagnosed as hav-and Haruo Okazaki, MD). Written records of the original neu- ing VaD by the Mayo Clinic criteria, and 4% were diag-ropathological examination were reviewed for descriptions of gross nosed as having probable VaD by the NINDS-AIREN cri-pathologic findings. These records often contained reference to teria.10 Of the 419 incident patients with dementia whoclinical information such as a history of stroke; therefore, our were deceased at the time of this study, neuropathologi-pathological diagnoses cannot be considered independent of clini-cal information. For the cases that underwent autopsy in the late cal examination results were available in 89 cases (21%1980s and early 1990s, the original routine neuropathological ex- of all deceased; 52 [58%] women). At onset of demen-aminations included only hematoxylin-eosin staining plus Biel- tia, 9 of the autopsied cases were younger than 70 years,schowsky silver staining. Brain regions sampled included fron- 7 were 90 years and older, and the remaining 73 weretal, temporal, parietal, and occipital neocortices, and striatum, between 70 and 90 years.hippocampus, amygdala, midbrain, pons, and medulla. Because We compared patients with dementia who under-the brains were acquired in the context of routine clinical activi- went autopsy with those who died but did not undergoties, there was no additional standardization for location of the autopsy. Autopsied patients were younger at onset of de-tissue samples within a particular region or for thickness of the mentia (median age, 80 years; range, 50-96 years) com-slices. Both hemispheres were available for gross and micro- pared with those who died and were not autopsied (me-scopic inspection. At later dates, some cases also underwent tauor -amyloid immunostaining. All pertinent histopathological re- dian age, 84 years; range, 52-100 years; P=.004, rank sumports, photographs of gross specimens, and histologic sections test). Patients with a temporal relationship between de-were first examined by three of us (J.E.P. and A.S. [2 neuropa- mentia and stroke underwent autopsy more frequentlythologists] and H.A. [a neurologist]), then they were all re- than did others (15 [17%] vs 35 [11%]); however, theviewed by a dementia expert (D.K.). All cases with vascular or difference was not statistically significant. The autop-uncertain pathological features were reexamined by the primary sied patients had 2 or more strokes more often than did (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 570 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  3. 3. Table 1. Neuropathological Diagnoses and Their Relationship to Clinical Features of Cerebrovascular Disease No. With Cerebrovascular Feature Median Age No. of No. (%) at Onset Temporal Relationship Critical Infarctions Both Features Either Feature Pathological Diagnoses Patients of Women (Range), y of Stroke and Dementia* on Imaging† Present Present Alzheimer disease 45 29 (64) 80 (53-95) 0 2 0 2 Alzheimer disease plus trauma 1 0 86 0 1 0 1 Dementia with Lewy bodies 9 6 (67) 78 (66-87) 0 1 0 1 Pure vascular dementia 12 4 (33) 80 (62-93) 7 4 2 9 Combination of vascular dementia 11 7 (64) 85 (71-94) 3 4 1 6 and Alzheimer disease Insufficient pathological findings 4 3 (75) 80 (72-96) 4 1 1 4 Other nonvascular pathological 7 3 (43) 82 (50-96) 1 0 0 1 findings‡ Total 89 52 (58) 80 (50-96) 15 13 4 24 *Onset or marked worsening of dementia within 3 months of a clinical stroke. †Bilateral gray matter infarctions shown by imaging in frontal, parietal, or temporal lobe cortex; in thalamus; or in basal ganglia. ‡Includes glioma (n = 1), progressive supranuclear palsy (n = 1), argyrophilic grain disease (n = 1), hydrocephalus without other definitive pathological findings(n = 1), and nonspecific degenerative changes (n = 3).the nonautopsied patients (10 [11%] vs 14 [4%]; 2 =6.35; pure VaD, the Mayo Clinic criteria were modestly sensi-P=.01). There were no differences between the autop- tive and specific; the ICD-10 criteria and the DSM-IV cri-sied group and the nonautopsied group with regard to teria (including white matter lesions) had similar sensitiv-sex, number of critical infarcts by imaging findings, or ity but lower specificity than the Mayo Clinic criteria, andpresence of focal neurological signs. Finally, the dura- the NINDS-AIREN criteria (including or excluding whitetion of dementia at the time of death was similar in the matter lesions) were less sensitive but more specific. Theautopsied patients (median, 4 years; range, 1-12 years) ADDTC criteria were less sensitive than the Mayo Clinicand nonautopsied patients (median, 5 years; range. 1-13 criteria but equally specific. For pure VaD and combinedyears; P =.46, rank sum test). VaD and AD together, the Mayo Clinic criteria were less The pathological diagnoses and demographic char- sensitive but more specific than the ICD-10 and the DSM-IVacteristics of the 89 patients are given in Table 1. There criteria (including white matter lesions). The positive pre-were 45 patients (51%) in whom AD was considered the dictive value of the Mayo Clinic criteria for pure VaD andsole cause of the dementia and 9 (10%) with Lewy body combined VaD and AD was 63% and the negative predic-disease. Twelve patients (13%) had pure VaD and 11 tive value, 88%.[(12%) had combined VaD and AD. The diagnoses of the The imperfect sensitivity of the diagnostic criteriaremaining patients are described in Table 1. The distri- for pure VaD was due to 5 patients with pure VaD at neu-bution by temporal relationship between dementia and ropathological examination who lacked a temporal re-stroke and by presence of critical infarctions on imag- lationship between a clinical stroke and the onset of theiring is also given in Table 1. Both these clinical features dementia. Rereview of the medical records of these 5 pa-of vascular disease were infrequent in autopsy-proved AD tients showed no evidence of stroke temporally relatedand in patients having dementia with Lewy bodies. By to dementia onset in 4, all of whom had family care-contrast, both vascular features were common among givers who were interviewed by behavioral neurolo-those with pure VaD or the combination of VaD and AD. gists. A major language barrier in the fifth case may have Details about the pathological findings of the pa- obscured the documentation of a stroke temporally re-tients with pure VaD and those with combined VaD and lated to the onset of dementia. Thus, 4 (33%) of 12 ofAD are given in Table 2. The data in Table 2 allow the our patients with pure VaD at pathological examinationreader to observe the effects of varying the diagnostic cut unequivocally lacked a temporal relationship between apoints for AD and VaD. Of the 12 patients with pure VaD, clinical stroke event and their dementia. The patients with7 had large vessel–distribution infarctions with or with- combined VaD and AD at pathological examination lackedout additional lacunar infarctions, 4 had exclusively la- a temporal relationship between dementia onset andcunar infarctions, and 1 had predominant leukoencepha- stroke even more frequently (13 of 23 patients).lopathy. In the 45 patients diagnosed as having pure AD We also calculated the sensitivity and specificity ofand the 9 with Lewy body disease, some cerebrovascu- the DSM-IV diagnosis for AD in this cohort, excludinglar disease was present: 15 (17% of the 89 who under- cases of combined VaD and AD (n=45). The sensitivitywent autopsy) had at least 1 microscopic or small infarc- was 98%, but the specificity was only 41%; the positivetion in the thalamus or more rostrally. Four of the 15 predictive value was 63%, with a negative predictive valuepatients had more than 1 small infarction (total infarct of 95%.volume, 0.5 cm3). Table 3 shows the sensitivity, specificity, and posi- COMMENTtive likelihood ratio for individual features and for severalsets of diagnostic criteria in the patients with pure VaD and The principal observations from our study were as fol-in those with pure VaD or VaD and AD. For patients with lows: (1) Vascular dementia was common neuropatho- (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 571 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  4. 4. Table 2. Description of Vascular and AD Pathological Findings in Patients With Pure VaD and Combined VaD and AD Plaque Pathological Findings Sex/Age at Death, y by CERAD Criteria* Braak and Braak Stage† Vascular Pathological Findings Patients With Pure VaD F/95 None II 4 Lacunar infarcts M/65 None I-II 5 Infarcts, including large L frontal, moderate R frontal M/77 None II 6 Lacunar infarcts M/89 Sparse II 4 Lacunar infarcts M/87 None II 4 Lacunar infarcts M/77 None II 6 Infarcts, including large R temporal M/87 None II 6 Infarcts, including moderate L frontal M/69 None II 6 Infarcts, including large R frontoparietal and L PCA M/74 None II Multiple small cortical and deep infarcts plus moderate R PCA F/90 Sparse II Extensive leukoencephalopathy, 3 lacunar infarctions F/86 Sparse II Moderate R PCA infarct F/83 Sparse II 5 Lacunar infarcts plus moderate L parietal Patients With Combined VaD and AD M/100 Sparse IV 8 Infarcts, extensive leukoencephalopathy M/91 Sparse IV 5 Infarcts F/94 Sparse III-IV 5 Large infarcts, including massive L MCA, large R PCA F/93 Moderate IV 2 Infarcts, including large L temporal F/92 Sparse IV 4 Lacunar infarcts F/78 Frequent IV 4 Infarcts, including large R parietal F/87 Moderate IV 3 Lacunar infarcts plus hippocampal infarct F/89 Sparse III-IV 6 Infarcts including several large cortical infarcts M/86 None V 2 Small infarcts F/90 Moderate V 3 Lacunes and vascular insult to CA1 M/82 Moderate IV 2 Lacunar infarcts, including caudate and DM thalamus Abbreviations: AD, Alzheimer disease; CA1, CA1 region of the hippocampus; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; DM thalamus,medial dorsal nucleus of the thalamus; L, left; MCA, middle cerebral artery; NFTP, neurofibrillary tangle pathological findings; PCA, posterior cerebral artery; R,right; VaD, vascular dementia. *Abundance of cored or neuritic neocortical plaques by the criteria of CERAD.19 †From Braak and Braak.20 Stage I indicates NFTP confined to the transentorhinal cortex (layer IV); stage II, NFTP in entorhinal cortex (layer II); stage III, NFTP inhippocampus (CA1 and subiculum); stage IV, NFTP in association neocortex (mild); stage V, NFTP in association neocortex (moderate to severe); and stage VI,NFTP in primary cortices.logically. (2) The feature of dementia temporally re- series, and another 20% were patients with a combina-lated to stroke was the best clinical predictor of pure tion of AD and VaD.1vascular pathological findings; however, this feature had One reason for our high proportion of VaD may beimperfect sensitivity. (3) At least a few cases with pure the population-based origin of our series. We were ableVaD at neuropathological examination lacked a history to capture patients with stroke and subsequent demen-of clinical stroke temporally related to the onset of de- tia more completely than most other clinicopathologicmentia. studies. Patients with stroke who become demented may In this population-based sample of incident cases be less likely to be evaluated in dementia clinics and lessof dementia who underwent routine pathological exami- likely to be given the diagnostic label of dementia. Thisnation, VaD (or more precisely, etiologically relevant ce- may explain why VaD is less common in referral-basedrebral infarctions) without AD, VaD combined with AD, clinicopathologic series9,22 compared with population-and AD with small infarctions were common. The pro- based series.1,8 A second reason was the increased au-portion of patients with dementia and etiologically im- topsy rate among cases with evidence of cerebrovascu-portant cerebral infarctions ranged from 13% to 25%. De- lar disease.spite the higher rate of autopsy for patients with clinical A third reason for our high rate of dementia attrib-features of cerebrovascular disease that was observed in uted to VaD neuropathologically may be the presence ofthis study, and despite the imprecision of the clinical di- incidental cerebrovascular disease.23 The cerebrovascu-agnosis of VaD, our estimates of the proportion of pa- lar disease observed in some patients with dementia maytients with VaD (those with pure VaD plus those with actually have been etiologically irrelevant. However, therecombined VaD and AD) are consistent with the results is evidence against this hypothesis. We have studied theof our incidence study.10 We previously reported that the neuropathological findings of 39 elderly subjects who wereproportion of all incident cases of dementia classified clini- evaluated cognitively within 1 year of death and foundcally as VaD was 18% by the Mayo Clinic criteria.10 Our to be free of dementia at the Mayo Clinic Alzheimer’s Dis-observed proportion of VaD at autopsy was also compa- ease Research Center, Rochester. None of these 39 non-rable to that from a population-based autopsy series from demented subjects had a burden of chronic cerebral in-England. Patients with pure VaD constituted 11% of that farctions greater than that of the patients diagnosed as (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 572 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  5. 5. Table 3. Validity of Diagnostic Formulations for Pure VaD and for Broadly Defined VaD (Pure VaD Plus Combined VaD and AD) Pure VaD Broadly Defined VaD Sensitivity, Specificity, Positive Likelihood Sensitivity, Specificity, Positive Likelihood Diagnostic Criteria Operational Features* % (n)† % (n)† Ratio (95% CI)‡ % (n)† % (n)† Ratio (95% CI)‡ Criteria Used in the Present Study Temporal link Clinical stroke temporally 58 (7/12) 90 (69/77) 5.6 (2.5-12.6) 43 (10/23) 92 (61/66) 5.7 (2.2-15.0) between stroke related to dementia and dementia (within 3 mo)§ Imaging findings Critical imaging lesions on 33 (4/12) 88 (68/77) 2.9 (1.0-7.8) 35 (8/23) 92 (61/66) 4.6 (1.7-12.6) strictly defined imaging VaD, Mayo Clinic Dementia temporally related 75 (9/12) 81 (62/77) 3.9 (2.2-6.7) 65 (15/23) 86 (57/66) 4.8 (2.4-9.4) criteria10 to stroke or critical imaging lesions (excluding white matter lesions) VaD, Mayo Clinic Dementia temporally related 75 (9/12) 73 (56/77) 2.8 (1.7-4.5) 70 (16/23) 79 (52/66) 3.3 (1.9-5.6) criteria, including to stroke or critical white matter imaging lesions or lesions10 bilateral extensive subcortical white matter lesions on imaging Criteria From the Literature NINDS-AIREN Dementia temporally related 17 (2/12) 97 (75/77) 6.4 (1.0-41.4) 13 (3/23) 98 (65/66) 8.6 (0.9-78.7) probable VaD, to a stroke and critical excluding white imaging lesions matter lesions16 (excluding white matter lesions) and 1 focal sign NINDS-AIREN Dementia temporally related 25 (3/12) 96 (74/77) 6.4 (1.5-28.2) 22 (5/23) 98 (65/66) 14.3 (1.8-116.5) probable VaD, to a stroke and critical including white imaging lesions matter lesions16 (including white matter lesions) and 1 focal sign DSM-IV, excluding Critical imaging lesions 67 (8/12) 69 (53/77) 2.1 (1.3-3.6) 70 (16/23) 76 (50/66) 2.9 (1.7-4.8) white matter (excluding white matter lesions15 lesions) or 1 focal sign DSM-IV, including Critical imaging lesions 75 (9/12) 64 (49/77) 2.1 (1.3-3.2) 74 (17/23) 70 (46/66) 2.4 (1.6-3.8) white matter (including white matter lesions15 lesions) or 1 focal sign ICD-10 17 Dementia temporally related 75 (9/12) 74 (57/77) 2.9 (1.8-4.8) 70 (16/23) 80 (53/66) 3.5 (2.0-6.2) to a stroke and 1 focal sign or critical imaging lesions (including white matter lesions) ADDTC criteria18 Dementia and 1 infarct at 67 (8/12) 79 (61/77) 3.2 (1.8-5.8) 57 (13/23) 83 (55/66) 3.4 (1.8-6.5) imaging and (1 additional infarct at imaging or 1 stroke temporally related or 2 strokes) Abbreviations: AD, Alzheimer disease; ADDTC, criteria by the State of California Alzheimer’s Disease Diagnostic and Treatment Centers; CI, confidence interval;DSM-IV, Diagnostic and Statistic Manual for Mental Diseases, Fourth Edition; ICD-10, International Classification of Diseases and Related Health Problems,10th Revision of the World Health Organization; n, numerator and denominator from which sensitivity or specificity is derived; NINDS-AIREN, National Instituteof Neurological Disorders and Stroke and the Association Internationale pour la Recherche et l’Enseignement en Neurosciences; VaD, vascular dementia. *Features from history or neuroimaging findings that constitute operational definition for specific diagnostic criteria. †For sensitivity, numerator indicates number of true positives; denominator, sum of true positives and false negatives. For specificity, numerator indicatesnumber of true negatives; denominator, sum of true negatives and false positives. ‡Positive likelihood ratio is sensitivity divided by the result of 1 minus specificity. §Onset or marked worsening of dementia within 3 months of a clinical stroke. Bilateral gray matter infarctions shown by imaging in frontal, parietal, or temporal lobe cortex; thalamus; or basal ganglia.having pure AD in this study (D.K., J.E.P., B.F.B., R.H.C., cerebrovascular lesions were etiologically important. Em-A.S., Misericordia Floriach-Robert, MD, Robert Ivnik, pirical demonstrations that cerebrovascular disease wors-PhD, Glenn Smith, PhD, Dennis Jackson, MD, Kris A. ens the expression of AD further suggest that cerebralJohnson, Lindsay Petersen, William McDonald, Ronald infarction is relevant to the expression of late-life cog-Petersen, PhD, MD, unpublished data, 2002). In addi- nitive decline.24,25tion, the lack of other explanations for the dementia in Our results provided a disappointing but expectedour patients with pure VaD suggested that the observed view regarding the diagnostic validity of clinical criteria (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 573 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  6. 6. for VaD. Similar findings have been reported by oth- dence of critical infarctions is present in a patient withers.1-7 None of the criteria, including the Mayo Clinic cri- dementia, survival is only slightly worse than that for AD,teria, had sufficient sensitivity to allow confident clini- and the neuropathologic findings are likely to include acal detection of VaD. The specificity and negative substantial vascular component.predictive value of the criteria were much better for such A strength of this study was the availability of andetection. The feature of dementia that was temporally autopsy series derived from all incident cases of demen-related to stroke had a high positive likelihood ratio. This tia in a defined population over a defined time interval.feature is part of the NINDS-AIREN16 and ICD-10 crite- Clinical stroke is accurately detected by the medical re-ria17; however, its contribution to sensitivity was se- cords-linkage system serving our study population.31verely reduced by the inclusion of other clinical fea- Weaknesses of the study include the retrospective col-tures (such as focal neurological signs or imaging criteria) lection of clinical information, which may have led to ain the NINDS-AIREN criteria. higher rate of identification of dementia in patients with A new approach to the problem of imperfect sensi- strokes compared with those without strokes. In addi-tivity of VaD criteria is needed. Several patients with patho- tion, because only a few patients underwent routine au-logically determined pure VaD and most of those with topsy, our estimates of sensitivity and specificity may becombined VaD and AD did not have clinical strokes that biased. However, this limitation is common to any au-were temporally related to the onset of their dementia. topsy study. Indeed, the fact that we were able to char-The existence of these patients and our confidence in the acterize some of the differences between autopsied andabsence of clinical strokes before the onset of dementia nonautopsied cases was unique. Another weakness of ourled us to conclude that there are 2 types of VaD. One type investigation was the lack of accepted and reliable patho-emerges from a clinical stroke event; the other develops logical criteria for VaD; it is possible that we overdiag-in an insidious fashion without any clinically apparent nosed pure VaD as a result of our neuropathological cri-stroke. The former type is readily diagnosed by the tem- teria. Finally, our pathological diagnoses were obtainedporal profile of the dementia relative to clinical strokes by pathologists who had access to clinical information;and imaging findings of new infarctions. The latter type therefore, clinical diagnosis and pathological diagnosislacks a characteristic clinical profile but can be sus- were not completely independent.pected on the basis of imaging study findings. Several prospective imaging studies have demon- Accepted for publication October 25, 2002.strated the occurrence of clinically silent infarc- Author contributions: Study concept and design (Drstions.26,27 A gradually accumulating burden of infarc- Knopman, Salviati, and Rocca); acquisition of data (Drstions or white matter ischemia could lead to a dementia Apaydin, Salviati, and Rocca); analysis and interpreta-that appears insidious and gradual. In addition, clini- tion of data (Drs Knopman, Parisi, Boeve, Salviati, Ed-cally silent cerebrovascular disease in the form of cere- land, and Rocca and Ms Cha); drafting of the manuscriptbral hypoperfusion could produce hippocampal neuro- (Drs Knopman, Parisi, and Rocca and Ms Cha); criticalnal loss28 or severe white matter pathological changes.29,30 revision of the manuscript for important intellectual con-We observed 1 case with ischemic leukoencephalopa- tent (Drs Knopman, Parisi, Boeve, Apaydin, Salviati, Ed-thy among the 4 patients without a clinical stroke. To land, and Rocca); statistical expertise (Drs Knopman, Ed-improve the sensitivity of the diagnostic criteria for the land, and Rocca and Ms Cha); administrative, technical,insidious type of VaD will require better tools to distin- and material support (Drs Salviati and Rocca); and studyguish nonspecific white matter lesions from severe is- supervision (Dr Rocca).chemic leukoencephalopathy. The use of magnetic reso- This work was supported in part by grants AG 06786nance imaging should increase the detection of lacunar (Mayo Alzheimer’s Disease Patient Registry) and AG 16574infarctions that may be missed with computed tomog- (Mayo Alzheimer’s Disease Research Center) from the Na-raphy. Differentiation of ischemic vs AD pathologic tional Institute on Aging, Bethesda, Md, and was made pos-changes in the hippocampus may also improve recogni- sible by grant AR 30582 (the Rochester Epidemiology Project)tion of VaD. Finally, a reliable biomarker for AD would from the National Institutes of Health, Bethesda.also be advantageous for VaD diagnosis. This study was presented in part at the Eighth Inter- Despite imperfect sensitivity in neuropathological national Conference on Alzheimer’s Disease and Related Dis-terms, clinicians may find the Mayo Clinic criteria for VaD orders, Stockholm, Sweden, July 24, 2002.useful. The 2 features of dementia onset or worsening This work is dedicated to the memory of Emre Kok-temporally related to a stroke and imaging evidence of men, MD, who was the principal founder of the Mayo Alz-bilateral gray matter infarctions are informative for pre- heimer’s Disease Patient Registry.dicting neuropathological findings, as demonstrated Corresponding author and reprints: David S. Knop-herein, and for predicting survival, as reported else- man, MD, Department of Neurology, Mayo Clinic, 200 Firstwhere.11 When both features are present (equivalent to St SW, Rochester, MN 55905 (e-mail: NINDS-AIREN criteria but ignoring the focal sign re-quirement), survival is decreased compared with those REFERENCESwith neither feature, and the autopsy diagnosis is likelyto be either pure VaD or combined VaD and AD. When 1. Holmes C, Cairns N, Lantos P, Mann A. Validity of current clinical criteria for Alz-only the temporal relationship feature is present, sur- heimer’s disease, vascular dementia and dementia with Lewy bodies. Br J Psy-vival is nearly as poor, but the diagnostic likelihood of chiatry. 1999;174:45-50.VaD at autopsy is slightly less. When only imaging evi- 2. Moroney JT, Bagiella E, Desmond DW, et al. Meta-analysis of the Hachinski Is- (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 574 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.
  7. 7. chemic Score in pathologically verified dementias. Neurology. 1997;49:1096- the diagnosis of ischemic vascular dementia proposed by the State of California 1105. Alzheimer’s Disease Diagnostic and Treatment Centers. Neurology. 1992;42: 3. Wade JPH, Mirsen TR, Hachinski VC, Fisman M, Lau C, Merskey H. The clinical 473-480. diagnosis of Alzheimer’s disease. Arch Neurol. 1987;44:24-29. 19. Mirra SS, Heyman A, McKeel D, et al. The Consortium to Establish a Registry for 4. Victoroff J, Mack WJ, Lyness SA, Chui HC. Multicenter clinicopathological cor- Alzheimer’s Disease (CERAD), part II: standardization of the neuropathologic as- relation in dementia. Am J Psychiatry. 1995;152:1476-1484. sessment of Alzheimer’s disease. Neurology. 1991;41:479-486. 5. Galasko D, Hansen LA, Katzman R, et al. Clinical-neuropathological correlations 20. Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta in Alzheimer’s disease and related dementias. Arch Neurol. 1994;51:888-895. Neuropathol (Berl). 1991;82:239-259. 6. Gold G, Giannakopoulos P, Montes-Paixao C Jr, et al. Sensitivity and specificity 21. McKeith IG, Galasko D, Kosaka K, et al. Consensus guidelines for the clinical and of newly proposed clinical criteria for possible vascular dementia. Neurology. pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consor- 1997;49:690-694. tium on DLB international workshop. Neurology. 1996;47:1113-1124. 7. Gold G, Bouras C, Canuto A, et al. Clinicopathological validation study of four 22. Hulette C, Nochlin D, McKeel D, et al. Clinical-neuropathologic findings in sets of clinical criteria for vascular dementia. Am J Psychiatry. 2002;159:82-87. multi-infarct dementia: a report of six autopsied cases. Neurology. 1997;48: 8. Lim A, Tsuang D, Kukull W, et al. Clinico-neuropathological correlation of Alz- 668-672. heimer’s disease in a community-based case series. J Am Geriatr Soc. 1999;47: 23. Neuropathology Group of the Medical Research Council Cognitive Function and 564-569. Ageing Study. Pathological correlates of late-onset dementia in a multicentre, 9. Nolan KA, Lino MM, Seligmann AW, Blass JP. Absence of vascular dementia in community-based population in England and Wales. Lancet. 2001;357:169- an autopsy series from a dementia clinic. J Am Geriatr Soc. 1998;46:597-604. 175.10. Knopman DS, Rocca WA, Cha RH, Edland SD, Kokmen E. Incidence of vascular 24. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. dementia in Rochester, Minn, 1985-1989. Arch Neurol. 2002;59:1605-1610. Brain infarction and the clinical expression of Alzheimer disease: the Nun Study.11. Knopman DS, Rocca WA, Cha RH, Edland SD, Kokmen E. Survival study of vas- JAMA. 1997;277:813-817. cular dementia in Rochester, Minnesota. Arch Neurol. 2003;60:85-90. 25. Heyman A, Fillenbaum GG, Welsh-Bohmer KA, et al. Cerebral infarcts in patients12. Edland SD, Rocca WA, Petersen RC, Cha RH, Kokmen E. Dementia and Alzhei- with autopsy-proven Alzheimer’s disease: CERAD, part XVIII. Neurology. 1998; mer’s disease incidence rates do not vary by sex in Rochester, Minn. Arch Neu- 51:159-162. rol. 2002;59:1589-1593. 26. Longstreth WT Jr, Bernick C, Manolio TA, Bryan N, Jungreis CA, Price TR. La-13. Rocca WA, Cha RH, Waring SC, Kokmen E. Incidence of dementia and Alzhei- cunar infarcts defined by magnetic resonance imaging of 3660 elderly people: mer’s disease: a reanalysis of data from Rochester, Minnesota, 1975-1984. Am the Cardiovascular Health Study. Arch Neurol. 1998;55:1217-1225. J Epidemiol. 1998;148:51-62. 27. Brott T, Tomsick T, Feinberg W, et al. Baseline silent cerebral infarction in the14. Kokmen E, Beard CM, O’Brien PC, Kurland LT. Epidemiology of dementia in Roch- Asymptomatic Carotid Atherosclerosis Study. Stroke. 1994;25:1122-1129. ester, Minnesota. Mayo Clin Proc. 1996;71:275-282. 28. Fein G, Di Sclafani V, Tanabe J, et al. Hippocampal and cortical atrophy predict15. American Psychiatric Association. Diagnostic and Statistical Manual of Mental dementia in subcortical ischemic vascular disease. Neurology. 2000;55:1626- Disorders, Fourth Edition. Washington, DC: American Psychiatric Asssociation; 1635. 1994. 29. Caplan LR. Binswanger’s disease—revisited. Neurology. 1995;45:626-633.16. Roman GC, Tatemichi TK, Erkinjuntti T, et al. Vascular dementia: diagnostic cri- ´ 30. Esiri MM, Wilcock GK, Morris JH. Neuropathological assessment of the lesions teria for research studies: report of the NINDS-AIREN International Workshop. of significance in vascular dementia. J Neurol Neurosurg Psychiatry. 1997;63: Neurology. 1993;43:250-260. 749-753.17. World Health Organization. Application of the International Classification of Dis- 31. Whisnant JP, Melton LJ III, Davis PH, O’Fallon WM, Nishimaru K, Schoenberg eases to Neurology (ICD-NA). 2nd ed. Geneva, Switzerland: World Health Orga- BS. Comparison of case ascertainment by medical record linkage and cohort fol- nization; 1997. low-up to determine incidence rates for transient ischemic attacks and stroke.18. Chui HC, Victoroff JI, Margolin D, Jagust W, Shankle R, Katzman R. Criteria for J Clin Epidemiol. 1990;43:791-797. (REPRINTED) ARCH NEUROL / VOL 60, APR 2003 WWW.ARCHNEUROL.COM 575 Downloaded from on December 6, 2010 ©2003 American Medical Association. All rights reserved.