A European perspective on population studies of dementia

The change in the world’s age demographics and the predicted rise in the incidence of age-related diseases, including dementia, is a source of major public health concern. Major research effort in both the United States and Europe has been targeted
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Perspectives A European perspective on population studies of dementia Carol Brayne a, *, Blossom C. M. Stephan a , Fiona E. Matthews b a  Department of Public Health and Primary Care, Institute of Public Health, University of Cambridge, Cambridge, United Kingdom b  Medical Research Council (MRC) Biostatistics Unit, Institute of Public Health, University of Cambridge, Cambridge, United Kingdom Abstract  The change in theworld’s age demographics and the predicted rise in the incidence of age-relateddiseases,includingdementia,isasourceofmajorpublichealthconcern.Majorresearcheffortinboththe United States and Europe has been targeted toward understanding the pathogenesis and epidemi-ology of dementia. This article presents a general overview of the history of dementia research in Eu-rope and how it compares with that in the United States. The review highlights the common issueswhich both U.S. and European researchers have identified and attempted to tackle. To maximize in-formation gained from studies across the world, better harmonization of methodology is needed, asinformed from current research practice.  2011 The Alzheimer’s Association. All rights reserved. Keywords:  Alzheimer’s disease; Dementia; Mild cognitive impairment; Diagnostic criteria; Prevalence; Incidence; Popula-tion-based 1. Introduction The conference that led to the development of this specialissueof   Alzheimer’sandDementia wasfocusedonU.S.prev-alence studies. All the areas of research discussedat the con-ference are relevant towider audiences. The issues that wereraised, including diagnostic methods and approaches, preva-lence, incidence, case-finding, risk factors, and public healthimplications, have also been the subject of research outsidethe United States. Dementia has long been a concern in Eu-rope,wheretheestimatedprevalenceis6.2%[1]andtheeco-nomic effect significant, costing a total of U.S. $210.12billion in Western Europe alone [2]. This article attempts tocombine research experience gained in Europe over the lastfew decades as a complementary perspective to those articu-lated so eloquently in this special issue. 2. Brief history In the middle of the last century, there were a small num-ber of population-based studies that reported on prevalenceof dementia, largely from Scandinavia and the United King-dom. The early studies were summarized in a seminal book byMortimerandSchuman[3],which highlightedthelackof epidemiologic attention to dementia at a time when researchwas largely oriented to heart disease and cancer. In the late1960s, Roth et al showed that, in clinically based studies(Newcastle, UK), older people with dementia had bothAlzheimer-type pathology and vascular pathology [4–8].These findings led to the awareness of Alzheimer’s disease(AD) as a major contributor to dementia in agingpopulations. Katzman et al took the findings frompopulation-based data and further extended the research fo-cus to the malignancy of progression [9]. Katzman is largelycredited with catching the attention of policy makers for aninvestmentin the specific area of AD, still verymucha focusof the dementia funding agenda of government and industry.Around the same time, therewas a group of early prevalencestudies,which startedin the1980sin Europe (includingGer-many, Finland, France, Italy, The Netherlands, Norway, Por-tugal, Spain, Sweden, and the United Kingdom).Individually, these had relatively little influence on policybut,together, through thevision of Hofman and Rocca alongwith European Union funding, created the first batch of Eu-ropean Studies of Dementia (EURODEM) reports [10–12].By this time, dementia prevalence rates for the 5-year agegroups ranging from 60 to 94 years had already beenestimated at 1.0%, 1.4%, 4.1%, 5.7%, 13.0%, 21.6%, and32.2%, respectively. Many European research agencies andgovernments had begun to become aware of demographic The authors have no conflicts to disclose.*Corresponding author. Tel.: 01223 330334; Fax: 01223 330330.E-mail address: 1552-5260/$ - see front matter  2011 The Alzheimer’s Association. All rights reserved.doi:10.1016/j.jalz.2010.12.003Alzheimer’s & Dementia 7 (2011) 3–9  aging, and research bodies were asked to consider the keyareas on which to focus. Those countries represented inthe original EURODEM brought forth the accumulatedexperience of what opportunities were available and thebest current methodological practice for dementia researchand together led to the development of the hugelyinfluential Rotterdam Study, in addition to other Europeanstudies as outlined in Table 1. These allowed the creationof harmonized prevalence and incidence rates for dementiain Europe [13–18]. Such estimates have been widely usedin research, policy, and capacity building.The outcomes have also been influential in the develop-ment of new standardized approaches to dementia researchused within countries, such as for the Medical ResearchCouncil’s Cognitive Function and Ageing Study (CFAS)and the new CFAS II, including CFAS II Wales (UK)( ), and internationally, such as the 10/66 De-mentia Research Group ( ). Im-portantly, these will increase understanding of the effect of cross-cultural and generational differences on disease inci-dence and prevalence. 3. Diagnostic approaches In the 1970s, prevalence estimates were derived fromstudies reporting clinical diagnosis and diagnosis recorded Table 1Examples of longitudinal studies of aging in Europe (ordered by start date)Study Country Start N (baseline) Age (years)Iceland Birth Cohort Study Iceland 1957 3704 61–64Bonn Longitudinal Study of Aging (BOLSA) Germany 1965 221 60–75Gothenburg Study (G € oteborg) Sweden 1971 1148 70Southampton Ageing Project (SAP) UK 1977 340 65 1 Bangor Longitudinal Study of Ageing(BLSA)Wales, UK 1979 534 65 1 Melton Mowbray Ageing Project UK 1981 3000 75 1 Nottingham Longitudinal Study of Activityand Ageing (NLSAA)UK 1983 1042 65 1 Cambridge City over 75’s Cohort (CC75C) UK 1985 2165 75 1 Personnes Ag  ees Quid (PAQUID) France 1987 3777 65 1 Kungsholmen Project Sweden 1987 1810 75 1 LUND 80 1  Sweden 1988 211 80ALPHA UK 1989 5222 65 1 Pamplona Spain 1989 1127 70 1 AMSTEL Amsterdam 1990 4051 65–84Berlin Ageing Study (BASE) Germany 1990 516 70–105Rotterdam The Netherlands 1990 7983 55 1 Medical Research Council (MRC) CognitiveFunction and Ageing Study (CFAS)UK 1991 13004 65 1 Etude du Vieillissement Art  eriel (EVA) France 1991 1389 60–70Vantaa 85 1  Finland 1991 553 85 1 Italian Longitudinal Study on Aging (ILSA) Italy 1992 5632 65–84Longitudinal Aging Study Amsterdam(LASA)The Netherlands 1992 3017 55–85Maastricht Aging Study (MAAS) The Netherlands 1992 2043 24–81Zaragoza Spain 1992 1080 64 1 Odense Denmark 1993 3346 65–84Longitudinal Study of Aging Danish Twins(LSADT)Denmark 1995 2401 75 1 The German Aging Survey Germany 1996 5000 40–85Leipzig Longitudinal Study of the Aged(LEILA 75 1 )Germany 1997 1378 75English Longitudinal Study of Ageing(ELSA)England 1998 12100 50 1 The Danish 1905-Cohort Study Denmark 1998 2262 93In Chianti Study Italy 1998 1453 20–1023-City Study France 1999 9294 65 1 ZARADEMP Spain 2000 4803 55 1 The Swedish National study of Aging andCare (SNAK-K)Sweden 2001 3089 60 1 The Norwegian Life Course, Ageing andGenerations Study (NorLAG)Norway 2002 5589 40–79Newcastle 85 1 Study UK 2006 1042 85CFAS II UK 2008 Recruiting 65 1 CFAS II-Wales Wales, UK 2011 Planned 65 1 C. Brayne et al. / Alzheimer’s & Dementia 7 (2011) 3–9 4  in medical records. However, diagnostic methods acrosssources were not standardized. At that time, dementia wasthe preferred diagnosis; this was before the unification of se-nile dementia not otherwise explained and early onset AD.Asawarenessoftheimportanceofmentalhealthandpsychi-atric disorders grew, national multicenter studies were initi-ated. To make cross-study comparisons, standardization of methodology, including diagnostic assessment, was pro-moted, not just for researchers but for other reasons suchas the development of service payment for diagnostic group-ingfor insurance purposes. Inmainstreampsychiatry, astan-dardized instrument emerged in the United Kingdom for theexamination of present state (Present State Examination).This was then adapted by Copeland et al in 1976 to becometheGeriatricMentalState(GMS)Examination.Fromthisanautomated computer algorithm was created to provide quasidiagnostic groupings that were validated against the clinicaldiagnostic process (Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised) [19]. The aimwasnotonlytotrytomakeclinicaldiagnosislesshaphazard,as many studies had revealed much variability between cli-nicians, but also to make diagnosis applicable by nonclini-cians. Thus, quasi clinical diagnosis is possible in a studywith a large number of individuals, with a single interview,rather than two phases. Many international studies haveused and continue to use this methodology or some adapta-tionofit.However,nostudyintheUnitedStateshasadoptedthis algorithm. In the United States, a different traditionemerged, one that was also used in many European settings,of multidisciplinary consensus diagnosis, with or withoutstandardized components contributing to it. The diagnosticalgorithm was advocated on the basis that clinical expertisewould not be required because interviewers could be trainedto administer questions or conduct examinations in a stan-dardized manner. Indeed, the main difficulties with clinicaldiagnosis were, and are still, between-clinician variabilityand drift over time with diagnostic boundaries in syndromesthat do not have definite neurobiological or pathological un-derpinnings. 3.1. Subtype diagnosis The primary research focus has been on AD with knowl-edge that therewere mixed forms of dementia effectively ig-nored for many years despite articulated concern [20]. Thiscan be partly explained by a strong belief that the patholo-gies demonstrated in clinicopathologic studies of early onsetdementia (i.e., predominately AD) also underlie the mostcommon dementias observed in people aged   75 years.However, the population-based neuropathology cohorts ini-tiatedinEurope(EpidemiologicalClinicopathologicalStud-ies in Europe) [21] along with other critical studies reportedin this special issue have found mixed subtypes to be themost common [22]. Further, although theworld was focusedon beta amyloid, other avenues of exploration continued inEurope. For example, 20 years ago, tau was investigated asa key pathophysiologic process in Europe in the laboratoryof molecular biology by Wischik et al [23,24] and relatedto dementia, including its mildest forms. 3.2. Mild forms and continua Although most output from clinical and population-basedstudies has been on people fulfilling diagnostic criteria, inthe last three decades in Europe, there has been a steady fo-cusoncognitivechangeandtrajectoriesinpopulations.Usu-ally these research programs have proceeded independently,but there are exceptions. Examples include the East Cam-bridgeshire Study, The Cambridge City over 75s Cohort[25], the Personnes Ag  ees Quid Study [26], The SwedishTwin Registry (OCTO) Study [27], The Gerontologicaland Geriatric Population Studies in Gothenburg (H70)[28], and The Swedish Adoption/Twin Study of Aging[29]. In addition to cognitive decline, core research themeshave included terminal decline [30], reliable change indices[31,32], and continuous distributions of change [33,34]. Analytically, there has been a focus on statisticalepidemiology in aging, including issues of individualtrajectories and multistate modeling. These are importantfor informing prediction and individualized medicine,most particularly for research.Mild forms of cognitivedecline have alsobeen identified.Historically, these were captured in the terms “age-associ-ated cognitive decline” [35] and “benign senescent forget-fulness” [36]. Here, cognitive change was not consideredto be preclinical but rather a result of normal aging. By con-trast, in the last 2 decades, the research focus has shifted to-wardamoreseverepreclinicalstageofdiseaseascapturedinthe term “mild cognitive impairment (MCI)” [37]. In theearly1990s,TheCambridgeMentalDisorders oftheElderlyExamination (CAMDEX) instrument captured this in itsminimal dementia category with criteria similar to the cur-rent adaptations of Petersen criteria described in this specialissue [38,39]. This was driven largely by a change inintervention focus, possibly as a result of the difficulty infinding a suitable target for dementia. As an opportunityfor identifying individuals at increased risk of developingdementia, MCI is an important concept. However, ascompared with U.S. research focus, in Europe, a moreconservative approach to dementia risk prediction has beenadopted, with a particular emphasis on validation andrepresentativeness of proposed models. This will haveimportant implications for public health policy. Theseareas have been extensively investigated in Europe [40–45]. 4. Populations and prevalence estimation In Europe, access to whole populations has generallybeen easier than in other parts of the world, although stillvariable, with the United Kingdom and Scandinavia beinghistorically more straightforward than continental Europe.Population registers exist, or have existed, which allow for C. Brayne et al. / Alzheimer’s & Dementia 7 (2011) 3–9  5  the invitation of whole or randomly selected representativepopulations to participate, including individuals in institu-tions, the frail, and the oldest-old.As noted previously, although the importance of clinicalexamination has never been doubted, it has since longbeen recognized that it is important to standardize the pro-cess of data collection, including cognitive function, activi-ties of daily living, other factors such as physical healthwhich influence the two, and other features of the dementiasyndrome such as behavioral and psychological symptomsof dementia. However, for population studies, many of thesewould not have been measured in the first interview withpopulation participants, but rather in a more detailed assess-ment at a second stage. Such a second stage would involvea substantial standardized interview, a clinician, or a combi-nationofthetwowithconsensusdiagnosis.Whateverthede-sign,thiswouldleadtotheneedforeitherassumptionsaboutthe ability of the screening instrument to be fully sensitiveortoaselection from thefullspectrumofthepopulation for theestimation of prevalence. The development of diagnostic al-gorithms for the older population in Europewas done by twoseparate groups—GMS and CAMDEX [46]. The GMS wasinitially semi-structured, but fully structured forms are nowavailable and embedded in CFAS and 10/66. The CAMDEXwas fully structured at outset and can be administered bytrained interviewers, but the algorithms srcinally describedhavenotbeen aswidelyused astheGMS, although elementshave been used internationally (e.g., in the RotterdamStudy). The GMS is now used internationally and 15 organ-icity questions were incorporated into the core of the largeEuropean population-based incidence studies designed aspart of the EURODEM series noted previously. 5. Dealing with attrition Estimation of prevalence, incidence, and risk is poten-tiallybiasedwhenresponseratesandattritionarenotconsid-ered. In addition to the ability to undertake total populationsampling frames, high response rates for studies of agingwere found to be possible across European countries (e.g.,95% in the Hughes Hall Project for Later Life, now knownas the Cambridge City over 75s Cohort [47]). Indeed, allstudies contributing to the srcinal EURODEM prevalenceestimate had relatively high response rates, with secondwave incidence studies, in general, having rates lower thanbut still between 60% and 80% (including the Three-CityStudy [48] and the Italian Longitudinal Study of Aging[49]).Unlessinformationisavailableonthosenotparticipat-ing,itisdifficulttoassesstheeffectofevensmallamountsof refusal. In studies where this has been possible [50], refusalhas been found to be associated with increased cognitiveim-pairment and lack of perceived personal relevance of a studyof health in aging, that is, the very healthy older population.Attrition through dropout and death is easier to track and ac-count for. In parallel with methods for estimation of preva-lence has been the development and application of analytical methods to account for informative missingnessas opposed to missing at random in longitudinal designs.Factors associated with dropout have been systematicallyanalyzed to inform such methods [51,52]. However, fewstudies report findings derived from analytical methodsthat fully reflect the study designs and those that do tendto show quite marked differences between observed andestimated findings [53]. 6. Differential mortality, incidence, and prevalence As noted in other articles in this series, it is possible to es-timate one population indicator when two others are present,assuming population stability. Differential mortality be-tween demented and nondemented individuals has beenshown repeatedly and does vary across cultures. Thus, if prevalence and mortality including differentials are known,then incidence can be estimated and vice versa. This ap-proach was replicated in Europe to explore variations intheestimation ofprevalence acrosscountries,whendifferentassumptions about differential mortality are made [54,55].This approach shows that the continuing debate aboutwhether prevalence of dementia continues to rise inextreme old age was effectively explained some time ago,although rarelyconsidered intheage-related, age-dependentcontroversy because mortality at extreme ages is so greatthatincidencehastorisesub-exponentially todrivetheprev-alence estimates seen in the older age groups. 7. The nature of dementia in the oldest age groups Population-based studies in Europe have contributed tothe understanding of the clinicopathologic basis of dementiafrom an epidemiological perspective, particularly in the old-est age groups. Studies with brain donation have informedthe debate on mixed dementia and identified changes inthe relationship between neuropathology and dementia byage [22,56–59]. Protective factors have also beenidentified, including increased education and the absenceof apolipoprotein E  3 4 [60]. In addition, these studies havealso brought to prominence the focus on vascular elementsin dementia. These can be used to inform age-specific treat-ment targets. 8. Dementia at the end of life and terminal decline Several decades ago, terminal decline was a topic of in-tense interest. With the emergence of the AD research focus,interestinterminaldeclinewaned.InEurope,researchcontin-ued on continuous trajectories and methods to analyze these,including work on terminal decline. This is continuing nowwith the International Network for the Integrative Analysisof Longitudinal Studies on Aging ( ),whichbringstogetherlongitudinalstudieswithrepeatedmea-sures. Some European studies have followed up populationsto “extinction.” These provide a valuable resource that C. Brayne et al. / Alzheimer’s & Dementia 7 (2011) 3–9 6  complements dementia research. Although these approacheshave beenintegratedinasmallnumberofstudiesworldwide,the data collected in many of the studies represented in thisvolume would add considerably to the more population andcontinuous approach, rather than the dementia/subtype ap-proach, which has dominated the last decades. Indeed, whena retrospective approach to dementia is taken for whole pop-ulation cohorts after extinction, it is clear that for many indi-viduals, dementia is a terminal event [61]. For those who donotbecome demented beforedeath,large proportions experi-encemoderatetoseverecognitivedecline[61].Thishasledtoa much greater focus on end of life and dementia, but there ismuch more to be done that should include geriatrics and pal-liative care beyond specialist Alzheimer’s settings. In fact,many people dying with severe cognitive impairment, in thecontext of comorbidity, will never receive a diagnosis of de-mentia, nor would it necessarily benefit them. 9. Assumption that all current research has relevance toother populations The paradigm at present with dementia research, includ-ing population-based research, is that outcomes are likely tobe relevant to all other populations. However, this has notbeen shown to be always true. Transcultural work has iden-tified variability in the expression of dementia across popu-lation groups, even though those with longer life expectancyandgreateraffluencetendtoshowahighdegreeofsimilaritywhen compared with other variable chronic diseases such ascancer and vascular diseases. Although cohort comparisonshave been attempted (including in this volume), it is still notpossible to examine cohort effects from within a single co-hort study because of attrition and dropouts. Comparisonacross time using the current favored method of consensusdiagnosis is subject to drift because the introduction of new categories causes boundary changes as in MCI, aswell as new types of recognized subtypes, such as Lewybody,extensionofthevascular dementiacategory,and intro-duction of new testing modalities. One method to establishcohort differences is if standardized and fixed methods canbe used. However, as noted by Knopman et al [62] in thisspecial issue, there are enormous changes in population re-sponsiveness to the civic contribution that taking part in re-search requires and these changes may profoundly affect ourability to compare across time. By no means can we be surethat our estimation of genetic influence and gene–environ-ment relations will remain stable across future generationswhere the environment is changing and may change in thepresence of factors such as climate change quite dramati-cally. 10. Summary There is a rich tradition of population-based studiesaround the world to inform our understanding of aging anddementia. It is important that we combine such research soas to benefit from knowledge and conceptual developmentsacross time and culture, necessary for future work and pol-icy. Acknowledgment Blossom CM Stephan is supported by the ERA-Age Fu-ture Leaders of Ageing (FLARE) Postdoctoral Fellowship. References [1] Alzheimer’s Disease International. World Alzheimer Report 2009.London, United Kingdom: Alzheimer’s Disease International; 2009.[2] Alzheimer’s Disease International. World Alzheimer Report 2010.London, United Kingdom: Alzheimer’s Disease International; 2010.[3] Mortimer JA, Schuman LM, eds. The Epidemiology of Dementia.New York, NY: Oxford University Press; 1981.[4] Roth M, Tomlinson BE, Blessed G. The relationship between quanti-tative measures of dementia and of degenerative changes in the cere-bral grey matter of elderly subjects. Proc R Soc Med 1967;60:254–60.[5] Roth M, TomlinsonBE, Blessed G. Correlationbetween scores for de-mentiaandcountsof“senileplaques”incerebralgreymatterofelderlysubjects. Nature 1966;209:109–10.[6] Blessed G, Tomlinson BE, Roth M. The association between quantita-tive measures of dementia and of senile change in the cerebral greymatter of elderly subjects. Br J Psychiatry 1968;114:797–811.[7] Tomlinson BE, Blessed G, Roth M. Observations on the brains of non-demented old people. J Neurol Sci 1968;7:331–56.[8] Tomlinson BE, Blessed G, Roth M. Observations on the brains of de-mented old people. J Neurol Sci 1970;11:205–42.[9] Katzman R, Terry R, DeTeresa R, Brown T, Davies P, Fuld P, et al.Clinical, pathological, and neurochemical changes in dementia: a sub-group with preserved mental status and numerous neocortical plaques.Ann Neurol 1988;23:138–44.[10] Hofman A, Rocca WA, Brayne C, Breteler MMB, Clarke M,CooperB, et al. The prevalence of dementiain Europe: a collaborativestudy of 1980-1990 findings. Int J Epidemiol 1991;20:736–48.[11] Rocca WA, Hofman A, Brayne C, Breteler MMB, Clarke M,Copeland JRM, et al. The prevalence of vascular dementia in Europe:facts and fragments from 1980-1990 studies. Ann Neurol 1991;30:817–24.[12] Lobo A, Launer LJ, Fratiglioni L, Andersen K, Di Carlo A,Breteler MM, et al. Prevalence of dementia and major subtypes in Eu-rope: a collaborative study of population-based cohorts. NeurologicDiseasesintheElderlyResearchGroup.Neurology2000;54(11Suppl.5):S4–9.[13] Lobo A, Saz P, Marcos G, Dia JL, De-la-Camara C, Ventura T, et al.Prevalence of dementia in a southern European population in two dif-ferent time periods: the ZARADEMP Project. Acta Psychiatr Scand2007;116:299–307.[14] The Medical Research Council Cognitive Function and Ageing Study.Cognitive function and dementia in six areas of England and Wales:the distribution of MMSE and prevalence of GMS organicity levelin the MRC CFA Study. The Medical Research Council CognitiveFunction and Ageing Study (MRC CFAS). Psychol Med 1998;28:319–35.[15] Skoog I, Nilsson L, Palmertz B, Andreasson LA, Svanborg A. Apopulation-based study of dementia in 85-year-olds. N Engl J Med1993;328:153–8.[16] Helmer C, Peres K, Letenneur L, Guttierez-Robledo LM,Ramaroson H, Barberger-Gateau P, et al. Dementia in subjects aged75 years or over within the PAQUID cohort: prevalence and burdenby severity. Dement Geriatr Cogn Disord 2006;22:87–94.[17] Obadia Y, Rotily M, Degrand-Guillaud A, Guelain J, Ceccaldi M,Severo C, et al. The PREMAP Study: prevalence and risk factors of  C. Brayne et al. / Alzheimer’s & Dementia 7 (2011) 3–9  7


Mar 19, 2018


Mar 19, 2018
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks