apoe in dementia

The Effect of APOE-e4 on Dementia is Mediatedby Alzheimer Neuropathology

James A. Mortimer, PhD,* David A. Snowdon, PhD,w and William R. Markesbery, MDz

Abstract: The degree to which the association of e4 with dementiais mediated by AD lesions in comparison with vascular lesions iscontroversial. The present study was undertaken to determine theroles of Alzheimer disease (AD) and vascular pathology inmediating the effect of apolipoprotein E (APOE)-e4 alleles ondementia. Clinicopathologic correlations were studied in 267Catholic sisters participating in the Nun Study. The extent towhich AD and vascular pathologies mediated the effect of APOE-e4 on dementia was investigated using multiple logistic regression.Adjusted for age at death and education, possession of 1 or more e4alleles was an important risk factor for dementia (oddsratio=2.98; 95% confidence interval, 1.62-5.48). This associationwas lost (odds ratio=1.38; 95% confidence interval, 0.68-2.80)when an index of the severity of AD-related neuropathology wasadded to the model, but changed little when measures of theseverity of vascular pathology were added. The findings suggestthat the effect of e4 on dementia is mediated by the severity of ADpathology. Although infarcts and atherosclerosis contribute to theoccurrence of dementia, this contribution seems unrelated toAPOE genotype.

Key Words: apolipoprotein E, Alzheimer disease, AD neuropathol-

ogy, vascular lesions, clinicopathologic correlation

(Alzheimer Dis Assoc Disord 2009;23:152–157)

At autopsy, demented patients show considerable overlapin pathologic lesions in the brain, with Alzheimer

disease (AD) and vascular lesions commonly found.1,2 Theheterogeneity of pathology raises the question of how thesedifferent lesions might mediate the effects of the e4 allele ofapolipoprotein E (APOE), an important risk factor fordementia3 and AD.3–8

In the Religious Orders Study (ROS), when the effectof AD pathology was controlled in regression analyses, theassociation of the e4 allele with clinical AD was reduced by50%.9 This finding suggests that much of the effect of thisallele on AD was through an association with the classicneuropathologic lesions of this illness. The ROS comparedparticipants with probable AD to controls and excludedpersons with other types of dementia, in whom e4 may also

have influenced non-AD pathology. In the present study,all cases of dementia were included.

The degree to which the association of e4 withdementia is mediated by vascular lesions in comparisonwith AD lesions is unknown. The possibility that vascularlesions may play a role in mediating the effects of the e4allele is raised by studies showing associations betweenAPOE-e4 and increased risk of vascular dementia.10–13 andalso by a recent study in which APOE-e4 was shown tomore than double the odds of cortical and subcorticalinfarction.14

We used data from 267 participants in the Nun Studywho came to autopsy to examine the roles of AD andvascular pathology in mediating the effect of APOE-e4alleles. Participants were included regardless of the presenceor cause of dementia. Our analyses closely parallel thosefrom the previous ROS publication9 in determining theeffect of adding measures of pathology on the associationbetween the e4 allele and dementia in regression models.

METHODS

ParticipantsParticipants in the Nun Study15,16 are members of the

School Sisters of Notre Dame congregation and live incommunities in the Midwestern, Eastern, and SouthernUnited States. In 1991 and 1992, 678 Catholic sisters, aged75 to 102 were enrolled who agreed to annual evaluationand brain donation after death. Participants includedindividuals who were initially demented and those free ofdementia and other cognitive impairments. Cognitive andphysical function was assessed annually. All participants ortheir guardians gave informed consent to their participationin this study. Ninety-two percent of those enrolledcontinued participation and 96% of those who died wereautopsied. The present study is based on the first 412participants who were autopsied. Data on both APOEgenotype and ratings of amyloid angiopathy were availablefor 267 of these participants, who constituted the samplefor the present analyses.

Assessment During LifeCognitive function was evaluated annually with the

Consortium to Establish a Registry for Alzheimer Disease(CERAD) neuropsychologic battery.17 This battery as-sesses memory, concentration, language, visuospatial abil-ity, and orientation to time and place. Performance-basedtesting was used to evaluate basic and instrumentalactivities of daily living.18,19 Participants were consideredto be demented when all of the following conditionswere met: (1) impairment in memory (defined by a score<4 on the Delayed Word Recall Test); (2) impairment in 1or more other areas of cognition (defined by scores of <11on Verbal Fluency, <13 on Boston Naming, or <8 onCopyright r 2009 by Lippincott Williams & Wilkins

Received for publication June 18, 2008; accepted September 13, 2008.From the *Department of Epidemiology and Biostatistics, College of

Public Health, University of South Florida, Tampa, FL; Depart-ments of wNeurology; and zNeurology and Pathology, SandersBrown Center on Aging, University of Kentucky, Lexington, KY.

Funded by grants P50-AG025711 (J.A.M.), R01-AG09862 (D.A.S.)and 1P30-AG28383 (W.R.M.) from the National Institute onAging.

Reprints: James A. Mortimer, PhD, Department of Epidemiology andBiostatistics, University of South Florida, 13201 Bruce B. DownsBoulevard, MDC-56, Tampa, FL 33612-3805 (e-mail: [email protected]).

ORIGINAL ARTICLE

152 | www.alzheimerjournal.com Alzheimer Dis Assoc Disord � Volume 23, Number 2, April–June 2009

Constructional Praxis); (3) impairment in basic or instru-mental activities of daily living (defined by inability to use atelephone, handle money, or dress oneself); and (4) evidenceof a decline in ability from a previous level attributable tocognitive impairment. Cut scores selected to define cogni-tive impairment on tests were based on scores less than thefifth percentile for normal controls in CERAD.20

APOE GenotypingAPOE genotyping was performed using genetic

material from buccal cells collected from living participantsor from frozen or paraffin-embedded brain specimens usinglaboratory methods described previously.21

Neuropathologic EvaluationGross and microscopic evaluations of the brains of

participants who died during the follow-up period ofthe study were performed by a board-certified clinicalneuropathologist (W.R.M.) who was blinded to diagnosis,cognitive test scores, and functional assessments.Diffuse and neuritic plaques (NPs) and neurofibrillarytangles (NFTs) were counted in the 5 most severely affectedfields of the middle frontal gyrus (Brodmann area 9),inferior parietal lobule (areas 39/40), middle temporal gyrus(area 21), and the CA1 and subiculum of the hippocampusafter staining with the modified Bielschowsky method. TheGallyas stain, which is better for detecting neuropilthreads and argyrophilic grains and slightly better fordetecting NFT in medial temporal lobe structures, was usedto count the neurofibrillary pathology in CA1 andsubiculum of the hippocampus. All sections were cut at8-mm thickness.

To meet the study’s neuropathologic criteria for AD,participants were required to have: (1) abundant senileplaques in the frontal, temporal, or parietal lobes, that is,16 or more senile plaques per mm2; (2) NPs in at least 1lobe; and (3) NFT in at least 1 lobe and abundant NFT inthe entorhinal cortex, hippocampus, and amygdala. Parti-cipants meeting the study’s neuropathologic criteria allsatisfied CERAD criteria for neuropathologic AD and hada high or intermediate likelihood of neuropathologic ADaccording to the National Institute on Aging/ReaganInstitute guidelines.

Cortical and subcortical infarcts were identified at thegross examination of the intact brain and of 1.5 cm thickcoronal sections of the cerebral hemispheres, brain-stem, and cerebellum. They were categorized as large(Z1.5 cm) or lacunar (<1.5 cm). The degree of athero-sclerosis in the circle of Willis was rated in 4 grades fromabsent to severe. Because there is no universally acceptedstaging of atherosclerosis in the circle of Willis, the scaleused reflects the subjective impression of the studyneuropathologist (W.R.M.) regarding the overall averageof involvement of the number of the affected vessels at thebase of the brain and the degree of atherosclerotic deposi-tions in these vessels. In the present study, 13% of partici-pants had absent to mild atherosclerosis (plaques present in0% to less than 25% of the vessel walls of the circle ofWillis), 33% moderate (25% to 50% inclusive) and 54%severe (>50%).

Chronic microinfarcts were identified from slides ofthe caudate nucleus, putamen, globus pallidus, thalamus,internal capsule, frontal lobe (area 9), temporal lobe, parietallobe, occipital lobe (areas 18 and 19), and hippocampus.

The severity of amyloid angiopathy was rated from 1(absent) to 4 (severe). The number of blood vesselscontaining b-amyloid by immunostaining was counted in6 regions, including the leptomeninges and the cortex of thefrontal, parietal, temporal, and occipital lobes and hippo-campus. When 6 or more blood vessels contained amyloidin greater than 50% of the wall in 3 or more brain regionsand leptomeninges, it was considered as severe. When 2 ormore regions and leptomeninges had 6 or more affectedvessels it was considered as moderate, and when 1 regionshowed 6 or more affected vessels it was considered as mild.If the severity was less than this, amyloid angiopathy wasconsidered to be absent. Amyloid angiopathy determined inthis manner was absent in 24%, mild in 43%, moderate in5%, and severe in 28%.

Statistical MethodsDifferences between participants with and without 1 or

more e4 alleles for individual descriptive variables wereassessed with Wilcoxon rank-sum and w2 tests. Anexploratory principal components analysis partialing outthe effects of age at death and years of education was usedto identify factors from the 7 pathologic measures obtainedin the study. Preliminary analyses showed that none of thecorrelations between measures exceeded 0.5, permitting theinclusion of ordinal and dichotomous variables.22 Logisticregression was used to examine the associations between thepresence of 1 or more APOE-e4 alleles and dementia,controlling for age at death and years of education. The

TABLE 1. Characteristics of Participants by APOE Genotype

Characteristic

Participants

Without e4Alleles

(n=200)

Participants

With 1 or

More e4Alleles

(n=67)

Education, mean (SD), years 15.7 (3.1) 16.2 (2.7)Age at death, mean (SD), years 90.8 (5.4) 90.0 (4.8)Percent demented at death 45.0 67.2*Number of neocorticalneurofibrillary tangles,mean (SD)

4.2 (10.1) 14.3 (15.3)**

Number of neocortical neuriticplaques, mean (SD)

3.3 (3.4) 4.9 (3.0)**

Percent who fulfilled studycriteria forneuropathologic AD

51.8 92.3**

Amyloid angiopathy severityrating, mean (SD)

2.08 (1.01) 3.24 (1.02)**

Rating of atherosclerosis in circleof Willis, mean (SD)

0.46 (0.26) 0.45 (0.27)

Percent with 1 or more largeinfarcts

17.5 19.4

Percent with 1 or more lacunarinfarcts

33.5 32.8

Percent with chronicmicroinfarcts

23.4 27.3

Statistical tests: Wilcoxon rank-sum tests were used to compare groupsfor age at death, number of neocortical neurofibrillary tangles, number ofneocortical neuritic plaques, amyloid angiopathy severity rating, and ratingof atherosclerosis in the circle of Willis; remaining variables were assessedwith the w2 test.

*P<0.01 **P<0.0001.APOE indicates apolipoprotein E.

Alzheimer Dis Assoc Disord � Volume 23, Number 2, April–June 2009 APOE-e4, Neuropathology and Dementia

r 2009 Lippincott Williams & Wilkins www.alzheimerjournal.com | 153

effect of indices representing the severity of Alzheimerpathology and vascular pathology was then investigated byadding these indices to the models with APOE-e4, age atdeath, and education, and comparing the strength ofassociations between APOE-e4 and dementia when theseindexes were present and absent. Additional models wererun in which individual pathologic measures were added toAPOE-e4, age at death and education. Finally, associationsof the 7 pathologic measures with each other were assessedby w2, rank biserial correlation (rrb), or Spearman correla-tion (rs), depending on the types of variables.

RESULTSTable 1 compares characteristics of the 200 participants

without an e4 allele and the 67 participants with 1 or more e4alleles. Those with 1 or more e4 alleles had higher meancounts of neocortical NFT, higher mean counts of neocor-tical NPs, and higher ratings of the severity of amyloidangiopathy. In addition, they were more likely to bedemented at death, and to fulfill study neuropathologiccriteria for AD. Other measures, including the 4 nonamyloidvascular measures, were not significantly different.

Principal components analysis of the 7 pathologicmeasures revealed 3 factors with eigenvalues greater than 1.The first factor had loadings of 0.62 (mean number ofneocortical NFT), 0.59 (mean number of neocortical NP),0.51 (rating of amyloid angiopathy severity), 0.07 (lacunarinfarcts), 0.07 (large infarcts), � 0.01 (presence of micro-infarcts), and � 0.06 (rating of severity of atherosclerosis inthe circle of Willis). The second factor had loadings of 0.61(large infarcts), 0.55 (lacunar infarcts), 0.45 (rating ofseverity of atherosclerosis in the circle of Willis), 0.33(presence of microinfarcts), 0.09 (rating of amyloid angio-pathy severity), � 0.06 (mean number of neocortical NFT),and � 0.10 (mean number of neocortical NP). The thirdfactor had only 1 variable loading greater than 0.40:0.76(presence of microinfarcts). Two indices were created bysumming standardized values of neocortical NFT, neocor-tical NP, and amyloid angiopathy to yield an ADpathology index; and by summing standardized values ofatherosclerosis and large and lacunar infarcts to yield avascular pathology index. Because the third factor con-sisted of a single variable, the presence of microinfarcts wasadded independently to the models.

Table 2 summarizes multiple logistic regression modelswith dementia as the outcome. APOE-e4 was stronglyassociated with dementia [odds ratio (OR)=2.98; 95%confidence interval (CI), 1.62-5.48)] adjusted for age at deathand years of education (Model 1). When the AD pathologyindex was added to the model, the association of APOE-e4with dementia was reduced and lost its statistical signifi-cance (OR=1.38; 95% CI, 0.68-2.80) (Model 2). Bycontrast when the vascular pathology index was added,the association of e4 with dementia changed only slightly(Model 3). The addition of microinfarcts (Model 4) led to asmall increase in the OR for APOE-e4. Finally, in the fullmodel (Model 5), both the AD and vascular pathologyindices were independently associated with dementia.

To assess the role of individual pathologies, a series oflogistic regression analyses were performed in which theAPOE-e4 association with dementia was evaluated withaddition of each pathologic measure alone (Table 3).Addition of the mean count of neocortical NFT eliminated

TABLE 2. Multiple Logistic Regression Models for Odds of Dementia

Model 1 Model 2 Model 3 Model 4 Model 5

Variable

OR

(95% CI) POR

(95% CI) POR

(95% CI) POR

(95% CI) POR

(95% CI) P

e4 allele 2.98 (1.62-5.48) <0.001 1.38 (0.68-2.80) 0.37 2.85 (1.54-5.27) <0.001 3.17 (1.71-5.88) <0.001 1.29 (0.63-2.68) 0.49Age at death 1.09 (1.04-1.15) <0.001 1.09 (1.03-1.15)<0.01 1.10 (1.04-1.16) <0.001 1.10 (1.04-1.16) <0.001 1.08 (1.02-1.15)<0.01Years ofeducation

0.90 (0.82-0.99) <0.05 0.90 (0.82-0.99)<0.05 0.92 (0.84-1.00) 0.07 0.90 (0.82-0.99) <0.05 0.92 (0.83-1.01) 0.08

Alzheimerindex

1.40 (1.21-1.62)<0.0001 1.43 (1.23-1.67)<0.0001

Vascularindex

1.15 (1.02-1.30) <0.05 1.17 (1.02-1.35)<0.05

Microinfarcts 0.92 (0.70-1.20) 0.54 0.89 (0.66-1.19) 0.42

CI indicates confidence interval; OR, odds ratio.

TABLE 3. Multiple Logistic Regression Analyses* for Odds ofDementia With Addition of Individual Pathologic Measures to aModel With APOE-e4

Pathologic Measure

Added Variables OR (95% CI) P

None e4 allele 2.98 (1.62-5.48) <0.001Mean count ofneocorticalneurofibrillary tangles

e4 allele 1.10 (0.52-2.34) 0.80pathology 1.14 (1.08-1.21) <0.0001

Mean count ofneocortical neuriticplaques

e4 allele 2.22 (1.18-4.19) <0.05pathology 1.17 (1.08-1.27) <0.0005

Rating of amyloidangiopathy

e4 allele 2.90 (1.45-5.76) <0.01pathology 1.05 (0.80-1.36) 0.73

Presence of large corticalinfarct(s)


Presence of lacunarinfarct(s)


Degree of atherosclerosisin circle of Willis

e4 allele 2.82 (1.53-5.21) <0.001pathology 0.89 (0.33-2.38) 0.81

Presence of microinfarcts e4 allele 3.17 (1.71-5.88) <0.001pathology 0.78 (0.35-1.73) 0.54

*All models adjusted for age at death and years of education.CI indicates confidence interval; OR, odds ratio.

Mortimer et al Alzheimer Dis Assoc Disord � Volume 23, Number 2, April–June 2009

154 | www.alzheimerjournal.com r 2009 Lippincott Williams & Wilkins

the significant association of APOE-e4 with dementia; noneof the other 6 measures eliminated this significant associa-tion. However, inclusion of the mean count of neocorticalNPs reduced the magnitude of the APOE-e4 OR from 2.98to 2.22.

Significant associations were found between thepresence of large and lacunar infarcts (w2=29.3,P<0.0001), large infarcts and ratings of atherosclerosisin the circle of Willis (rrb=0.30, P<0.01), and lacunarinfarcts and atherosclerosis in the circle of Willis(rrb=0.21, P<0.05). However, there were no significantassociations between ratings of amyloid angiopathy and theoccurrence of large (rrb=0.13, NS) or lacunar (rrb=0.09,NS) infarcts, between ratings of amyloid angiopathy andatherosclerosis (rs= � 0.06, NS), between ratings ofamyloid angiopathy and microinfarcts (rs=0.14, NS),between microinfarcts and either lacunar (w2=1.16, NS)or large (w2=2.41, NS) infarcts, or between microinfarctsand atherosclerosis (rrb=0.13, NS). The rating of amyloidangiopathy was significantly correlated with both the meanNFT count in the neocortex (rs=0.36, P<0.0001) and themean NP neocortical count (rs=0.33, P<0.0001).

DISCUSSIONConsistent with the previous research,3 presence

of 1 or more APOE-e4 alleles was a strong predictorof dementia. The addition of the index of severity ofAlzheimer pathology to the model reduced the strength ofthis association from 2.98 to 1.38, suggesting that the effectof APOE-e4 on dementia is largely accounted for by moresevere AD neuropathology. Addition of an index ofcerebrovascular disease, including the degree of athero-sclerosis in the circle of Willis and the presence of large andlacunar infarcts, as well as addition of chronic micro-infarcts, had little effect on the association between the e4allele and dementia.

Among the 3 AD-related pathologies, only the meancount of neocortical NFTs when added to a modelpredicting dementia from APOE-e4, age at death, andeducation substantially reduced the association between e4and dementia. This finding suggests that APOE-e4 influ-ences the occurrence of dementia primarily through itsassociation with more severe NFT expression, and isconsistent with the results of many clinicopathologic studiesincluding the Nun Study,23 which found a strongerassociation of dementia with the frequency of NFTs thanwith the frequency of NPs.

Our findings for AD neuropathology are in generalagreement with those of the ROS,9 which used a similarapproach to analyze data. In that study, the likelihood ofclinical AD was strongly associated with the e4 allelein logistic regression models and the association wasreduced approximately 50% by addition of a summarymeasure of global AD pathology reflecting diffuse plaque,NP and NFT densities. However, in contrast to the findingsof that study, we did not see a large reduction in theassociation of e4 with dementia by addition of a measureof mean neocortical NP count. Although the mea-sures of NP density were similar between the 2 studies,there was a large difference in age at death. Participants inthe present study were on average more than 5 years olderat the time of their death than those in the ROS. Whereasthe ROS investigators reported similar associations withdementia of normalized measures of NFTs (OR=3.59; 95

CI, 1.69-7.60) and NPs (OR=3.90; 95% CI, 1.99-7.61);normalized measures of NFTs and NPs in the present studydemonstrated much stronger associations of NFTs withdementia (OR=5.17; 95% CI, 2.76-9.71) as comparedwith NPs (OR=1.89; 95% CI, 1.42-2.50). It is possiblethat the difference in findings may be accounted for byincreased progression of the AD pathologic process,24 withNFTs playing a more important role in those who survivedto older ages.

A recent publication based on data from the ROS25

examined the extent to which a dichotomous variable(presence or absence of chronic cerebral infarction)mediated the effect of APOE-e4 on continuous measuresof cognitive function. In agreement with the results of thepresent study, the authors did not find a significantmediation effect of chronic infarction on the associationbetween APOE-e4 and global cognition.

Autopsy-based studies have shown significant correla-tions between the e4 allele and the density of NPs,26–28

NFT,27,29–31 and amyloid deposition.29–34 However, somestudies have not found an association with amyloidplaques35–37 or NFT.26,32,35,37,38 With few exceptions, thosestudies not finding an association were restricted toindividuals with clinically diagnosed dementia, whereasthose showing an association included substantialnumbers of nondemented elderly. The lack of an associa-tion seen in some studies may represent a ceiling effectachieved with increasing disease severity.39 The Catholicsisters that are the focus of this report cover a widespectrum of both cognitive performance and neuropathol-ogy, which may have facilitated finding stronger associa-tions between e4 and AD pathology.

Studies of clinically diagnosed vascular dementiapatients have been equivocal with regard to associationswith APOE-e4. Some have shown elevated e4 frequen-cies,10–13 whereas others demonstrated normal frequenciesof this allele.40–44 Similarly, in clinical studies the associa-tion of the e4 allele and stroke has been mixed, with somestudies reporting a positive association between APOE-e4and stroke,45–47 whereas others have failed to confirm thisassociation.48–53 Studies in which strokes were identified byautopsy are relatively few. Two studies54,55 found that thefrequency of cerebrovascular lesions was not increased inAPOE-e4, whereas another group14 reported that APOE-e4significantly increased the odds of both cortical andsubcortical infarctions. In the present study, we did notfind a statistically significant association of APOE-e4 withthe presence of large infarcts or lacunar infarcts, the degreeof atherosclerosis in the circle of Willis or the presenceof chronic microinfarcts. However, the degree of athero-sclerosis in the circle of Willis was significantly associatedwith the presence of both large and lacunar infarcts, andboth types of infarcts contributed to the expression ofdementia as we have shown previously for the samecohort.16

The association between cardiovascular disease, ather-osclerosis and AD has been controversial.56–59 Our resultssuggest that cerebrovascular and AD lesions do not shareAPOE-e4 as a common risk factor. Rather they support thehypothesis that vascular lesions increase the clinicalexpression of an AD process that is accelerated bypossession of one or more e4 alleles. An important excep-tion is amyloid angiopathy, which was strongly related toAD neuropathology, but had no association with theseverity of atherosclerosis or the occurrence of infarcts.



In summary, the present results suggest the effect of e4on dementia is mediated almost entirely by the severity ofAD pathology. Although infarcts contribute significantly tothe occurrence of dementia, this contribution seemsunrelated to APOE genotype.

This study has several strengths. These include theability to compare dementia status before death to detailedmeasures of AD and vascular pathology at autopsy in alarge population genotyped for APOE. Little attritionoccurred during the study and the autopsy rate was high,limiting these sources of bias. Because approximately halfof the sisters died without becoming demented and a widespectrum of both AD and vascular pathology was presentat autopsy, there was significant power to assess associa-tions between the presence of e4 alleles, dementia, andvascular and AD lesions. Blinding of the neuropathologistto all clinical data eliminated bias due to expectations basedon premorbid status.

Limitations include the restriction of AD neuropatho-logic measures to mean counts of neocortical NFTs andNPs and ratings of severity of amyloid angiopathy. Becausewe did not know the pathologic burden from sisters whenthey were alive, the sequence of events from APOEgenotype to clinical presentation must be inferred fromcross-sectional regression analyses. Finally, the study wasconducted entirely on women with a high degree ofeducation and lifestyles that limited exposureto tobaccoand alcohol as well as other confounders. Because highereducation reduces expression of dementia,60 a less well-educated sample might be more likely to become dementedwith less severe neuropathology, resulting in somewhatdifferent associations between APOE-e4, AD neuropatho-logy, and dementia. Our findings, although largely con-sistent with those of another study of well-educatedCatholic nuns, priests, and brothers,9 need to be replicatedin a sample of lay persons with a greater range ofeducational attainment.

ACKNOWLEDGMENTS

This study would not have been possible without thespirited support of the members, leaders, and healthcareproviders of the School Sisters of Notre Dame religiouscongregation.

REFERENCES

1. Xuereb JH, Brayne C, DuFouil C, et al. Neuropathologicalfindings in the very old. Results from the first 101 brains of apopulation-based longitudinal study of dementing disorders.Ann NY Acad Sci. 2000;903:490–496.

2. White L, Small BJ, Petrovitch H, et al. Recent clinical-pathologic research on the causes of dementia in late life:Update from the Honolulu-Asia Aging study. J GeriatrPsychiatry Neurol. 2005;18:224–227.

3. Henderson AS, Easteal S, Jorm AF, et al. Apolipoprotein Eallele epsilon 4, dementia, and cognitive decline in a populationsample. Lancet. 1995;346:1387–1390.

4. Myers RH, Schaefer EJ, Wilson PW, et al. Apolipoprotein Ee4 association with dementia in a population-based study: TheFramingham Study. Neurology. 1996;46:673–677.

5. Evans D, Beckett L, Field R, et al. Apolipoprotein E e4 andincidence of Alzheimer’s disease in a community population ofolder persons. JAMA. 1997;277:822–824.

6. Farrar LA, Cupples LA, Haines JL, et al. Effects of age, sex,and ethnicity on the association between apolipoprotein Egenotype and Alzheimer’s disease: a meta analysis. JAMA.1997;278:1349–1356.

7. Breitner JCS, Jarvik GP, Plassman BL, et al. Risk ofAlzheimer disease with the e4 allele for apolipoprotein E in apopulation-based study of men aged 62-73 years. Alzheimer DisAssoc Disord. 1998;12:40–44.

8. Tang MX, Stern Y, Marder K, et al. The APOE-e4 allele andthe risk of Alzheimer disease among African Americans, whitesand Hispanics. JAMA. 1998;279:751–755.

9. Bennett DA, Wilson RS, Schneider JA, et al. Apolipoprotein Ee4 allele, AD pathology, and the clinical expression ofAlzheimer’s disease. Neurology. 2003;60:246–252.

10. Shamano H, Ishibashi S, Murase T, et al. Plasma apolipopro-teins in patients with multi-infarct dementias. Atherosclerosis.1989;79:257–260.

11. Frisoni GB, Calabresi L, Geroldi C, et al. Apolipoprotein Eepsilon 4 allele in Alzheimer’s disease and vascular dementia.Dementia. 1995;5:240–242.

12. Engelborghs S, Dermaut B, Goeman J, et al. ProspectiveBelgian study of neurodegenerative and vascular dementia:APOE genotype effects. J Neurol Neurosurg Psychiatry.2003;74:1148–1151.

13. Davidson Y, Gibbons L, Purandare N, et al. Apolipoprotein Ee4 allele frequency in vascular dementia. Dement Geriatr CognDisord. 2006;22:15–19.

14. Schneider JA, Bienias JL, Wilson RS, et al. The apolipoproteinE e4 allele increases the odds of chronic cerebral infarctiondetected at autopsy in older persons. Stroke. 2005;36:954–959.

15. Snowdon DA, Kemper SJ, Mortimer JA, et al. Linguisticability in early life and cognitive function and Alzheimer’sdisease in late life: findings from the Nun Study. JAMA.1996;275:528–532.

16. Snowdon DA, Greiner LH, Mortimer JA, et al. Braininfarction and the clinical expression of Alzheimer disease.JAMA. 1997;277:813–817.

17. Morris JC, Heyman A, Mohs RC, et al. The Consortium toEstablish a Registry for Alzheimer’s Disease (CERAD). Part I.Clinical and neuropsychological assessment of Alzheimer’sdisease. Neurology. 1989;39:1159–1165.

18. Kuriansky J, Gurland B. The performance test of activities ofdaily living. Int J Aging Hum Develop. 1976;7:343–352.

19. Potvin AR, Tourtellotte WW, Dailey JS, et al. Simulatedactivities of daily living examination. Arch Phys Med Rehab.1972;53:476–489.

20. Welsh KA, Butters N, Mohs RC, et al. The Consortium toEstablish a Registry for Alzheimer’s Disease (CERAD). PartV. A normative study of the neuropsychological battery.Neurology. 1994;44:609–614.

21. Saunders AM, Hulette C, Welsh-Bohmer KA, et al. Specificity,sensitivity, and predictive value of apolipoprotein E genotyp-ing for sporadic Alzheimer’s disease. Lancet. 1996;348:90–93.

22. Wildt AR, Mueller CW, Kim JO. Introduction to FactorAnalysis: What It Is and How To Do It. Thousand Oaks, CA:Sage Publications Inc; 1978.

23. Riley KP, Snowdon DA, Markesbery WR. Alzheimer’s neuro-fibrillary pathology and the spectrum of cognition function:findings from the Nun Study. Ann Neurol. 2002;51:567–577.

24. Bennett DA, Schneider JA, Wilson RS, et al. Amyoid mediatesthe association of apolipoprotein E e4 allele to cognitivefunction in older people. J Neurol Neurosurg Psychiatry.2005;76:1194–1199.

25. Li Y, Schneider JA, Bennett DA. Estimation of the mediationeffect with a binary mediator. Statist Med. 2007;26:2298–3414.

26. Olichney JM, Hansen LA, Galasko D, et al. The apolipopro-tein E e4 allele is associated with increased neuritic plaques andcerebral amyloid angiopathy in Alzheimer’s disease and Lewybody variant. Neurology. 1996;47:190–196.

27. Warzok RW, Kessler C, Apel G, et al. Apolipoprotein E4promotes incipient Alzheimer pathology in the elderly.Alzheimer Dis Assoc Disord. 1998;12:33–39.

28. Yip AG, McKee AC, Green RC, et al. APOE, vascularpathology, and the AD brain. Neurology. 2005;65:259–265.

Mortimer et al Alzheimer Dis Assoc Disord � Volume 23, Number 2, April–June 2009

156 | www.alzheimerjournal.com r 2009 Lippincott Williams & Wilkins

29. Ohm TG, Kirca M, Bohl J, et al. Apolipoprotein Epolymorphism influences not only cerebral senile plaque loadbut also Alzheimer-type neurofibrillary tangle formation.Neuroscience. 1995;66:583–587.

30. Polvikoski T, Sulkava R, Haltia M, et al. Apolipoprotein E,dementia, and cortical deposition of beta-amyloid protein.NEJM. 1995;333:1242–1247.

31. Ohm TG, Scharnagl H, Marz W, et al. Apolipoprotein Eisoforms and the development of low and high Braak stages ofAlzheimer’s disease-related lesions. Acta Neuropathol.1999;98:273–280.

32. Gomez-Isla T, West HL, Rebeck GW, et al. Clinical andpathological correlates of apolipoprotein E e4 in Alzheimer’sdisease. Ann Neurol. 1996;39:62–70.

33. Morishima-Kawashima M, Oshima N, Ogata H, et al. Effect ofapolipoprotein E allele e4 on the initial phase of amyloid beta-protein accumulation in the human brain. Am J Pathol.2000;157:2093–2099.

34. Walker LC, Pahnke J, Madauss M, et al. Apolipoprotein E4promotes the early deposition of Ab42 and then Ab40 in theelderly. Acta Neuropathol. 2000;100:36–42.

35. Heinonen O, Lehtovirta M, Soininen H, et al. Alzheimerpathology of patients carrying apolipoprotein E e4 allele.Neurobiol Aging. 1995;16:505–513.

36. Morris CM, Massey HM, Benjamin R, et al. Molecularbiology of APO E alleles in Alzheimer’s and non-Alzheimer’sdementias. J Neur Transm Suppl. 1996;47:205–218.

37. Landen M, Thirsell A, Wallin A, et al. The apolipoprotein Eallele epsilon 4 does not correlate with the number of senileplaques or neurofibrillary tangles in patients with Alzheimer’sdisease. J Neurol Neurosurg Psychiatry. 1996;61:352–356.

38. Mukaetova-Ladinska EB, Harrington CR, Roth M, et al.Presence of the apolipoprotein E e4 allele is not associated withneuro-fibrillary pathology or biochemical changes to tau protein.Dement Geriatr Cogn Disord. 1997;8:288–295.

39. Hansen LA, Galasko D, Samuel W, et al. Apolipoprotein Eepsilon-4 is associated with increased neurofibrillary pathologyin the Lewy body variant of Alzheimer’s disease. Neurosci Lett.1994;182:63–65.

40. Betard C, Robitaille Y, Gee M, et al. Apo E allele frequenciesin Alzheimer’s disease, Lewy body dementia, Alzheimer’sdisease with cerebrovascular disease and vascular dementia.Neuroreport. 1994;5:1893–1896.

41. Palumbo B, Parnetti L, Nocentini G, et al. Apolipoprotein-Egenotype in normal aging, age-associated memory impairment,Alzheimer’s disease and vascular dementia patients. NeurosciLett. 1997;231:59–61.

42. Wehr H, Bednarska-Makaruk M, Lojkowska W, et al.Differences in risk factors for dementia with neurodegenerativetraits and for vascular dementia. M Dement Geriatr CognDisord. 2006;21:1–7.

43. Orsitto G, Seripa D, Panza F, et al. Apolipoprotein Egenotypes in hospitalized elderly patients with vasculardementia. Dement Geriatr Cogn Disord. 2007;23:327–333.

44. Kim KW, Youn JC, Han MK, et al. Lack of associationbetween apolipoprotein E polymorphism and vascular demen-tia in Koreans. J Geriat Psychiatry Neurol. 2008;21:12–17.

45. McCarron MO, Delong D, Alberts MJ. APOE genotype as arisk factor for ischemic cerebrovascular disease. Neurology.1999;53:1308–1311.

46. Kim JS, Han SR, Chung SW, et al. The apolipoprotein E e4haplotype is an important predictor for recurrence in ischemiccerebrovascular disease. J Neurol Sci. 2003;206:31–37.

47. Jin ZQ, Fan YS, Ding J, et al. Association of apolipoproteinE4 polymorphism with cerebral infarction in Chinese Hanpopulation. Acta Pharmacol Sin. 2004;25:352–356.

48. Couderc R, Mahieux F, Bailleul S. Apolipoprotein E allelefrequency, ischemic cerebrovascular disease and Alzheimer’sdisease. Stroke. 1994;24:1416–1417.

49. Kuusisto J, Mykkanen L, Kervinen K, et al. ApolipoproteinE4 phenotype is not an important risk factor for coronaryheart disease or stroke in elderly subjects. Arterioscler ThrombVasc Biol. 1995;15:1280–1286.

50. Basun H, Corder EH, Guo Z, et al. Apolipoprotein Epolymorphism and stroke in a population sample aged 75years or more. Stroke. 1996;27:1310–1315.

51. Kuller LH, Shemanski L, Manolio T, et al. Relationshipbetween apoE, MRI findings, and cognitive function in theCardiovascular Health Study. Stroke. 1998;29:388–398.

52. Frikke-Schmidt R, Nordestgaard BG, Thudium D, et al.APOE genotype predicts AD and other dementia but notischemic cerebrovascular disease. Neurology. 2001;56:194–200.

53. MacLeod MJ, De Lange RP, Breen G, et al. Lack ofassociation between apolipoprotein E genotype and ischaemicstroke in a Scottish population. Eur J Clin Invest. 2001;31:570–573.

54. Olichney JM, Hansen LA, Hofstetter CR, et al. Associationbetween severe cerebral amyloid angiopathy and cerebrovas-cular lesions in Alzheimer disease is not a spurious oneattributable to apolipoprotein E4. Arch Neurol. 2000;57:869–874.

55. Chui HC, Zarow C, Mark WJ, et al. Cognitive impact ofsubcortical vascular and Alzheimer’s disease pathology. AnnNeurol. 2006;60:677–687.

56. Sparks DL, Scheff SW, Liu H, et al. Increased density of senileplaques (SP), but not neurofibrillary tangles (NFT), in non-demented individuals with the apolipoprotein E4 allele:comparison to confirmed Alzheimer’s disease patients. J NeurolSci. 1996;138:97–104.

57. Irina A, Seppo H, Arto M, et al. b-amyloid load is notinfluenced by the severity of cardiovascular disease in aged anddemented patients. Stroke. 1999;30:613–618.

58. Roher AE, Esch C, Kokjohn TA, et al. Circle of Willisatherosclerosis is a risk factor for sporadic Alzheimer’s disease.Arterioscler Thromb Vasc Biol. 2003;23:2055–2062.

59. Beeri MS, Rapp M, Silverman JM, et al. Coronary arterydisease is associated with Alzheimer disease neuropathology inAPOE4 carriers. Neurology. 2006;66:1399–1404.

60. Mortimer JA, Snowdon DA, Markesbery WR. Head circum-ference, education and risk of dementia: findings from the NunStudy. J Clin Exp Neuropsychol. 2003;25:671–679.



apoe in dementia

Documents