Age and Ageing Advance Access published July 7, 2014 Age and Ageing 2014; 0: 1–6 doi: 10.1093/ageing/afu085

© The Author 2014. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For Permissions, please email: [email protected]

Dementia prediction for people with stroke in populations: is mild cognitive impairment a useful concept? BLOSSOM C. M. STEPHAN1, THAIS MINETT2, GRACIELA MUNIZ TERRERA3, FIONA E. MATTHEWS4, CAROL BRAYNE2 1

Address correspondence to: B. C. M. Stephan. Tel: (+ 44) 0191 222 3811; Fax: (+44) 0191 208 6043. Email: [email protected]

Abstract Background: criteria for mild cognitive impairment (MCI) capture an intermediate cognitive state between normal ageing and dementia, associated with increased dementia risk. Whether criteria for MCI are applicable in the context of stroke and can be used to predict dementia in stroke cases is not known. Objectives: to determine the prevalence of MCI in individuals with stroke and identify predictors of 2-year incident dementia in stroke cases. Methods: individuals were from the Medical Research Council Cognitive Function and Ageing Study. MCI prevalence in individuals with stroke was determined. Logistic regression, with receiver operating characteristic curve analysis, was used to identify variables associated with risk of dementia in stroke cases including MCI criteria, demographic, health and lifestyle variables. Findings: of 2,640 individuals seen at the first assessment, 199 reported stroke with no dementia. In individuals with stroke, criteria for MCI are not appropriate, with less than 1% of stroke cases being classified as having MCI. However, in individuals with stroke two components of the MCI definition, subjective memory complaint and cognitive function (memory and praxis scores) predicted 2-year incident dementia (area under the curve = 0.85, 95% CI: 0.77–0.94, n = 113). Conclusion: criteria for MCI do not appear to capture risk of dementia in the context of stroke in the population. In stroke cases, subjective and objective cognitive performance predicts dementia and these variables could possibly be incorporated into dementia risk models for stroke cases. Identifying individuals with stroke at greatest risk of dementia has important implications for treatment and intervention. Keywords: stroke, dementia, mild cognitive impairment, risk, epidemiology, older people

Introduction Stroke is associated with an increased risk of cognitive impairment and dementia [1, 2]. How best to capture those individuals with mild impairment in cognition and at increased risk of dementia secondary to stroke in populationbased cohorts is not currently known. The term vascular cognitive impairment no dementia (VCIND) refers to the mildest form of cognitive impairment related to cerebrovascular diseases and is associated with elevated risk of dementia [3, 4]. However, there are currently no standard criteria for diagnosing VCIND across the many different vascular

conditions such as stroke and hypertension. Further, there is wide variation in how different studies classify such individuals and this usually depends on the data available (e.g. given health condition, neuropsychological test, whether brain MRI was undertaken). Lack of consensus impedes cross-study comparability and undermines the validity of this concept. Stratification of stroke cases into more homogeneous groups, for example in terms of cognitive functioning (e.g. impaired versus not-impaired) and risk of dementia (at-risk versus not-at risk), will be important for clinical trial enrolment within the context of VCIND as well as for improving routine care.

1

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

Institute of Health and Society, Newcastle University, The Baddiley-Clark Building, Richardson Road Newcastle upon Tyne, NE2 4AX, UK 2 Department of Public Health and Primary Care, Institute of Public Health, Cambridge University, Forvie Site, Robinson Way, Cambridge CB2 0SR, UK 3 MRC Lifelong Health and Ageing Unit at UCL, 33 Bedford Place, London, WC1B 5JU, UK 4 MRC Biostatistics Unit, Cambridge Institute of Public Health, Cambridge University, Cambridge CB2 0SR, UK

B. C. M. Stephan et al.

Methods

Neuropsychological evaluation

Global cognition was assessed at each interview using the Mini-Mental State Examination (MMSE) [10]. A MMSE score of 21 or less was taken to reflect impairment when defining MCI. Domain-specific cognitive function was assessed using the Cambridge Cognitive Examination (CAMCOG) [11] (augmented in that items on tactile recognition of coins were excluded). For the CAMCOG, a total score assessing overall ability was calculated (range: 0–103). Sub-scale scores were also derived for memory including learning, recent and remote memory and non-memory domains including orientation, language comprehension and expression, perception, praxis, abstraction and attention and calculation. Summing scores only from the memory or nonmemory sub-scales created composite scores. As the CAMCOG scores (total, sub-scale and composite memory and non-memory scores) were not normally distributed, impairment for each domain and the composite memory and non-memory scores was defined using a cut-off score of 1 SD below the mean (16th centile), adjusted for age.

Participants

This study was embedded within the Medical Research Council Cognitive Function and Ageing Study (CFAS), a population representative cohort of individuals aged 65 years and older recruited from registers of primary care physicians in five UK centres using identical methodology. Full details of the study design have been published previously [7]. In brief, baseline interviewing began in 1991 and in total, 13,004 individuals were recruited (82% response rate). A 20% (n = 2,640) stratified sample, selected based on cognitive ability, age and centre completed a more detailed assessment interview, with re-interviewing at 2 years. Information on demographics (age, gender and years of education), self-reported health (including chronic conditions), cognition and functional ability were collected by computerised interviews conducted at the participant’s place of residence by a trained interviewer. Data from the initial prevalence screen, first assessment and 2-year followup interviews (Data Version 9.0) were used in this analysis.

Ethics

All participants gave informed consent before the interview. When participants could not consent for themselves, for example due to severe cognitive impairment, consent was obtained by their informant.

Dementia diagnosis

Dementia was diagnosed using the Geriatric Mental State Examination Automated Geriatric Examination for ComputerAssisted Taxonomy (AGECAT) diagnostic algorithms [8] and defined as an AGECAT organic symptom level of 3 or greater. This has been found to be comparable to dementia as diagnosed using the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, Revised (DSM-III-R) [9].

2

Physical function

Performance on activities of daily living (ADL) and instrumental activities of daily living (IADL) was assessed using items from the Modified Townsend Disability Scale [12]. Using information on the hierarchy of ADL and IADL disability, individuals were classified as (i) severely impaired (if the person is housebound or requires help at least several times per week with washing, cooking and dressing), (ii) mildly impaired (if the person requires help regularly with heavy housework or shopping and carrying heavy bags) and (iii) not-impaired (if no help is needed with washing, hot meals, shoes and socks, heavy housework or shopping and carrying heavy bags and the individual can get around outside). Stroke cases

Stroke was assessed at the baseline interview and again at the first assessment interview using a single question—baseline interview: ‘Have you ever had a stroke that required medical attention’, and assessment interview: where there is observation of paralysis ‘What did your doctor say was wrong with your .....? Was the possibility of a stroke mentioned?’. Of the 2,640 individuals seen at the first assessment, 283 reported stroke (at either the baseline or assessment interview), 2,232 were stroke free and 125 individuals had missing stroke status. When excluding individuals with dementia (n = 587), 199 individuals reported stroke and 1,843 were stroke free. Diagnosing MCI

As the cognitive deficit in stroke cases is yet to be fully elucidated, in this study we chose to map three different definitions of MCI capturing both domain-specific and multi-domain impairment including (i) Mayo Clinic criteria for aMCI [5], (ii) criteria for nMCI (nMCI: where there is no evidence of memory

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

Within the framework of Alzheimer’s disease, preclinical cognitive impairment is usually captured using the term mild cognitive impairment (MCI) as defined by Petersen et al. [5] and includes the following: subjective memory complaint (SMC), normal general cognitive function, objective cognitive deficit, maintenance of function and no dementia. Typically, the most prominent feature of MCI is an impairment in memory performance (e.g. amnestic MCI: aMCI), but nonmemory (e.g. non-amnestic MCI: nMCI) and multi-domain sub-types (e.g. multi-domain MCI: mMCI) also exist [6]. The applicability of MCI criteria for capturing cognitive impairment and risk of dementia associated with stroke is not known and is explored here. We also test the predictive accuracy of conventional risk factors to predict 2-year incident dementia in stroke cases including demographic, cognitive, health, lifestyle and physical functioning variables.

Stroke, MCI and dementia

Results Sample demographics

The prevalence of stroke stratified by study centre, age group and gender has been published previously for the total CFAS population [17], and the demographics of the stroke (n = 283) and no-stroke (n = 2,232) groups are shown in Table 1. Individuals with stroke were more likely to have dementia, fewer years of education, lower MMSE scores, greater functional impairment and increased presence of disease comorbidity related to CHD, diabetes, hypertension and PVD. Age did not differ between the stroke and no-stroke groups. MCI criteria in the context of stroke

MCI is rare in individuals with stroke. Less than 1% of individuals with stroke fulfil criteria for aMCI, nMCI or mMCI. Rather, more individuals with stroke have impairment outside the MCI range, falling into the OCIND category (prevalence of OCIND in stroke cases = 3.4%, 95% CI: 2.7–4.4%). As shown in Table 2, when breaking down the components of the MCI definition over half of the individuals with stroke would be excluded from a MCI diagnosis due to failure to meet the requirement of high physical functioning (46.7% Table 1. Characteristics of stroke and no-stroke cases at the first assessment interview No stroke (n = 2,232)

Stroke (n = 283)

17.4 (388)

29.7 (84)

76.9 (7.7) 9.6 (2.1) 64.5 (1,439) 22.8 (5.4)

77.3 (7.1) 9.3 (1.9) 58.3 (165) 20.9 (5.6)

25.6 (560) 19.9 (422) 6.1 (131) 24.3 (519) 3.8 (80)

63.2 (175) 31.1 (82) 9.5 (26) 43.6 (116) 8.4 (20)

........................................ Analyses

All analyses were completed using Stata version 12 (StataCorp., College Station, TX, USA). Differences in demographics between the stroke and no-stroke groups were compared using t-tests for continuous variables and Chi-squared test for categorical variables. The prevalence of MCI (amnestic, non-amnestic and multi-domain) in stroke cases without dementia was calculated using back weighting for study design. Logistic regression was used to predict the risk of 2-year incident dementia in the stroke group. The following variables were tested in univariable models: demographic [sex, age group (young-old 9 years)], physical function (ADL/IADL disability), psychiatric status (depression, anxiety), disease co-morbidity [coronary heart disease (CHD) defined as the presence of self-reported heart attack or angina based on the Rose Diagnostic Scale [15], peripheral vascular disease (PVD) defined using the Rose Scale [15], self-reported hypertension and diabetes], lifestyle (current smoking and alcohol consumption) and cognitive function [self or informant memory complaint, MMSE score (dichotomised: ≤24 versus ≥25) [16], CAMCOG total and sub-domain scores (age adjusted)]. Backward stepwise logistic regression (P = 0.2; including only significant variables from the univariable analyses) was used for final model selection. Discriminative accuracy of the final model was assessed using the area under the receiver operator characteristic curve (AUC).

Percentage of dementia (n)a Demographics Mean age years (SD) Mean education years (SD)a Percentage of women (n)a Mean MMSE (SD)a Health status, % (n) Severe functional difficultya CHDa Diabetesa Hypertension (medicated)a PVDa

a Significant difference between the stroke and no-stroke groups (P < 0.05). CHD, coronary heart disease; PVD, peripheral vascular disease; SD, standard deviation.

Table 2. Percentage (n) of individuals without dementia in the stroke and no-stroke groups who fulfil the different criteria required for a MCI diagnosis No stroke/no dementia Stroke/no dementia (n = 1,843) (n = 199)

........................................ SMC Normal general cognitive function (MMSE ≥ 22) None/mild ADL impairment Memory impairment Non-memory impairment

48.7 (986) 78.0 (1,361)

68.7 (136) 66.5 (121)

81.8 (1,492) 34.7 (608) 52.8 (953)

46.7 (92) 45.3 (81) 64.2 (122)

ADL, activities of daily living; MMSE, Mini-Mental State Examination; SMC, subjective memory complaint.

3

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

impairment) and (iii) criteria for mMCI (mMCI: where both memory and non-memory are impaired) [6]. While memory deficits are typically associated with neurodegenerative rather than vascular pathologies, we chose to include aMCI as this is one of the commonly applied definitions of MCI in clinical and research practice. Full details of the operationalisation of each sub-type of MCI in CFAS have been reported in detail previously [13]. For this analysis, no medical exclusions were applied when mapping MCI. At the first assessment, individuals without dementia were classified into three groups including (i) no cognitive impairment (NCI), (ii) MCI or (iii) other cognitive impairment no dementia (OCIND). For classification of MCI, individuals had to fulfil the following criteria: (i) no dementia, (ii) subjective or informant complaint of memory loss, (iii) normal general cognitive function (MMSE ≥ 22), (iv) no or only mild functional impairment and (v) objective memory or non-memory impairment (defined using percentiles (16th centile) to approximate 1 SD below the mean, adjusted for age, derived from the CAMCOG composite memory or non-memory scores). To be classified as NCI, individuals had to fulfil the following criteria: (i) no dementia, (ii) normal general cognitive functioning, (iii) no severe functional impairment and (iv) normal memory and non-memory test performance. The OCIND group included all nonnormal, non-demented individuals who failed to fulfil one or more diagnostic criteria for MCI (e.g. individuals without a study diagnosis of dementia, but a MMSE score 10 years) 0.5 (0.2–1.4) 0.191 Lifestyle Alcohol use (none/mild versus severe) 2.5 (0.8–7.5) 0.100 Smoking status (never versus ever) 0.8 (0.2–3.2) 0.809 Physical function ADL (none/mild versus severe) 3.3 (1.1–9.9) 0.031 Mental health Anxiety (absent versus present) 1.4 (0.1–13.3) 0.768 Depression (absent versus present) 3.0 (1.0–9.2) 0.054 Physical health CHD (absent versus present) 0.8 (0.3–2.3) 0.657 Diabetes (absent versus present) 0.4 (0.0–3.2) 0.383 Heart attack (absent versus present) 0.7 (0.1–3.1) 0.594 Hypertension (medicated) (absent versus present) 0.6 (0.2–1.5) 0.243 PVD (absent versus present) 2.2 (0.4–12.2) 0.352 Cognition SMC (no complaint versus complaint) 5.0 (1.1–22.9) 0.037 MMSE (≤24 versus >25) 4.3 (1.2–16.1) 0.029 CAMCOG total scoreb 12.5 (1.6–99.1) 0.017 CAMCOG learning memoryb 10.5 (2.8–40.2) 0.001 CAMCOG recent memoryb 4.4 (1.2–15.5) 0.021 CAMCOG remote memoryb 1.4 (0.3–5.5) 0.645 0.6 (0.1–3.1) 0.581 CAMCOG abstractionb CAMCOG attention and calculationb 3.3 (1.0–10.7) 0.043 4.3 (1.4–13.3) 0.011 CAMCOG language comprehensionb CAMCOG language expressionb 1.4 (0.4–4.6) 0.561 CAMCOG orientationb 1.8 (0.4–7.2) 0.420 8.0 (2.3–27.5) 0.001 CAMCOG praxisb CAMCOG perceptionb 1.9 (0.5–6.6) 0.334

Results where P < 0.05 are in bold. Significant predictors were used in the stepwise model and therefore correction for multiple comparisons was not undertaken at the univariate modelling stage. b Not-impaired versus impaired, with impairment defined using percentiles (16th centile) to approximate 1 SD below the mean, adjusted for age. ADL, activities of daily living; CAMCOG, Cambridge Cognitive Examination; CHD, coronary heart disease; CI, confidence interval; MMSE, Mini-Mental State Examination; OR, odds ratio; PVD, peripheral vascular disease; SMC, subjective memory complaint. a

4

Discussion In this study, we found that most individuals with self-reported stroke failed to fulfil criteria for MCI, including aMCI, nMCI or mMCI, mainly due to restrictions on functional ability. In individuals who reported a history of stroke, the best predictors for 2-year incident dementia were SMC and scores on the CAMCOG sub-domains of learning memory and praxis. In CFAS, health status was self-reported and therefore disease prevalence estimates may reflect bias in memory or awareness of a condition. However, it should be noted that self-reported and objective disease status has previously been found to be in high agreement for the diseases included in this analysis [18, 19]. Further, brain imaging was not available and we are therefore unable to determine the prevalence of silent strokes, and whether features of stroke such as locality or sub-type (e.g. haemorrhage or infarct) influence cognitive profiles and risk of dementia. MCI was defined using Mayo Clinic criteria [5] for aMCI and criteria for nMCI and mMCI [6]. More inclusive definitions of MCI that cover a broader range of cognitive (e.g. executive deficits and deficits in global cognitive functioning) and functional deficits, such as the definition of cognitive impairment no dementia (CIND) [20], may be more useful for capturing cognitive impairment in the context of stroke. However, we tested the sensitivity of different MCI sub-types, including aMCI, nMCI and mMCI. This had little effect on the estimated prevalence. Further, MCI criteria were never designed to be used within the context of stroke and, as suggested by the results, alternative methods are needed to capture VCIND associated with stroke. Additional risk modelling was also undertaken to avoid the limitations of attempting to fit predefined criteria that were not developed nor validated for dementia prediction in the context of vascular disease. It was found that current criteria for MCI are not the best approximation for cognitive impairment following stroke. Individuals with stroke appear to have a more extreme pattern of impairment that falls outside the MCI range. Their cognitive profile reflects impairment in global function, memory and non-memory domains. Therefore, definitions that rely only on poor memory or non-memory performance will be too restrictive. Further, a high proportion of stroke cases had severe functional impairment and this is typically an exclusion when diagnosing MCI across the many different definitions [13]. If a pre-dementia state is to be mapped within the framework of stroke, a broader concept than MCI is needed. Results from the risk factor analysis suggest that impairments in domain-specific cognitive scores, including nonmemory (e.g. praxis) and memory (e.g. learning memory)

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

Table 3. Results from the univariate logistic regression analysis for predicting 2-year incident dementia in the stroke group

the domains of learning memory, recent memory, attention and calculation, language comprehension and praxis). Results from the backward stepwise regression model found that the most parsimonious model for predicting 2-year incident dementia in the stroke group included SMC and the CAMCOG sub-domain scores of learning memory and praxis (AUC = 0.85, 95% CI: 0.77–0.94, n = 113, R 2 = 23.1).

Stroke, MCI and dementia

Conclusion How do these results help in defining those individuals with stroke at increased risk of dementia? The results suggest that criteria for VCIND should include both subjective and objective cognitive function, not exclusively focused on nonmemory performance. Prediction of who will develop cognitive impairment and dementia in individuals with stroke is a major challenge. This has important clinical implications and may allow for better estimation of prognosis, and improved care and future planning for stroke patients. Future studies need to determine whether cognitive or other interventions in people with stroke reduce dementia risk.

Key points • VCIND is an umbrella term that captures cognitive deficits associated with vascular disease that falls short of a dementia diagnosis. • Identification of VCIND is a major challenge due to a lack of operational criteria. • The question of the validity of MCI outside of Alzheimer’s disease is timely and important. • MCI criteria fail to capture cognitive impairment and risk of dementia in the context of stroke. • In stroke cases, SMC and objective cognitive performance (memory and non-memory) predict dementia and these variables could possibly be incorporated into criteria for VCIND.

Author contributions B.C.M.S. contributed to design, analysed the data, drafted the first version of the manuscript and was responsible for revisions. T.M. and G.M. contributed to design and analysis and revised the manuscript. F.M. provided intellectual input into the manuscript. C.B. is the PI of the CFAS study and contributed to design and editing of the manuscript. All authors read and approved the final manuscript.

Acknowledgements We would like to thank the CFAS population, their families and carers for their participation.

Conflicts of interest None declared.

Ethical approval CFAS was approved by the local ethics committee and has multicenter ethics committee approval (ethics approval reference 05/mrec05/37).

Funding MRC CFAS has been funded by the Medical Research Council (G9901400) and Department of Health. The funding bodies had no role in the study design, data collection, analysis or decision to publish. Funding was obtained from the Guarantors of Brain to present these results as part of a symposium on VCIND at the International Stroke Conference, New Orleans, 2012. Graciela Muniz Terrera was funded by the National Institute on Aging (NIH/NIA 1P01AG043362).

References 1. Savva GM, Stephan BC. Epidemiological studies of the effect of stroke on incident dementia: a systematic review. Stroke 2010; 41: e41–6. 2. Pendlebury ST, Rothwell PM. Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: a systematic review and meta-analysis. Lancet Neurol 2009; 8: 1006–18. 3. Narasimhalu K, Ang S, De Silva DA, Wong MC, Chang HM, Chia KS et al. Severity of CIND and MCI predict incidence of dementia in an ischemic stroke cohort. Neurology 2009; 73: 1866–72. 4. Stephan BC, Matthews FE, Khaw KT, Dufouil C, Brayne C. Beyond mild cognitive impairment: vascular cognitive impairment, no dementia (VCIND). Alzheimers Res Ther 2009; 1: 4. 5. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999; 56: 303–8. 6. Winblad B, Palmer K, Kivipelto M, Jelic V, Fratiglioni L, Wahlund LO et al. Mild cognitive impairment—beyond

5

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

combined with SMC (self or informant), predict 2-year dementia progression best in stroke cases. These results support recent findings highlighting the importance of poor cognitive function in predicting dementia in the context of stroke [3, 21–24] extending them to suggest that only limited domains need to be tested, but tests of memory and nonmemory performance should be included. Few variables may be needed to predict dementia in stroke cases as stroke itself already confers a high risk for dementia [1]. Indeed, demographic factors (including age and education), lifestyle variables (e.g. alcohol use and smoking), mental health (e.g. depression and anxiety) and disease co-morbidity were also not found to be predictive of incident dementia in stroke cases. Health-related co-morbidity has been previously associated with higher risk of dementia in stroke cases in some (over 8-year follow-up) [21], but not all studies [5, 25]. In a recent review [1], recurrent stroke rather than cardiovascular risk factors was found to be associated with dementia. Determining cases where disease co-morbidity is associated with greater risk of dementia has important implications for treatment. Further studies investigating whether other variables associated with dementia [e.g. genetics such as apolipoprotein (APOE) e4 status] improve dementia risk prediction in stroke cases and whether MRI derived stroke-related factors (e.g. such as laterality, sub-type) mediate risk prediction are needed.

B. C. M. Stephan et al.

7.

8.

9.

10.

12.

13.

14.

15.

16.

17.

6

18.

19.

20.

21.

22.

23.

24.

25.

Research Council Cognitive Function and Ageing Study. The Analysis Group. J Epidemiol Community Health 1997; 51: 494–501. Corser W, Sikorskii A, Olomu A, Stommel M, Proden C, Holmes-Rovner M. Concordance between comorbidity data from patient self-report interviews and medical record documentation. BMC Health Serv Res 2008; 8: 85. Simpson CF, Boyd CM, Carlson MC, Griswold ME, Guralnik JMFried LP. Agreement between self-report of disease diagnoses and medical record validation in disabled older women: factors that modify agreement. J Am Geriatr Soc 2004; 52: 123–7. Graham JE, Rockwood K, Beattie BL, Eastwood R, Gauthier STuokko H et al. Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet 1997; 349: 1793–6. Allan LM, Rowan EN, Firbank MJ, Thomas AJ, Parry SW, Polvikoski TM et al. Long term incidence of dementia, predictors of mortality and pathological diagnosis in older stroke survivors. Brain 2011; 134: 3713–24. Ballard C, Rowan E, Stephens S, Kalaria R, Kenny RA. Prospective follow-up study between 3 and 15 months after stroke: improvements and decline in cognitive function among dementia-free stroke survivors >75 years of age. Stroke 2003; 34: 2440–4. Melkas S, Oksala NK, Jokinen H, Pohjasvaara T, Vataja R, Oksala A et al. Poststroke dementia predicts poor survival in long-term follow-up: influence of prestroke cognitive decline and previous stroke. J Neurol Neurosurg Psychiatry 2009; 80: 865–70. Knopman DS, Roberts RO, Geda YE, Boeve BF, Pankratz VSCha RH et al. Association of prior stroke with cognitive function and cognitive impairment: a population-based study. Arch Neurol 2009; 66: 614–9. Reitz C, Bos MJ, Hofman A, Koudstaal PJ, Breteler MM. Prestroke cognitive performance, incident stroke, and risk of dementia: the Rotterdam Study. Stroke 2008; 39: 36–41.

Received 24 March 2014; accepted in revised form 14 May 2014

Downloaded from http://ageing.oxfordjournals.org/ at Istanbul University Library on November 5, 2014

11.

controversies, towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Intern Med 2004; 256: 240–6. Brayne C, McCracken C, Matthews FE. Cohort profile: the Medical Research Council Cognitive Function and Ageing Study (CFAS). Int J Epidemiol 2006; 35: 1140–5. Copeland JR, Dewey ME, Griffiths-Jones HM. A computerized psychiatric diagnostic system and case nomenclature for elderly subjects: GMS and AGECAT. Psychol Med 1986; 16: 89–99. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 3rd edition, revised (DSM-III-R). Washington, DC: APA; 1987. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–98. Huppert FA, Brayne C, Gill C, Paykel ES, Beardsall L. CAMCOG—a concise neuropsychological test to assist dementia diagnosis: socio-demographic determinants in an elderly population sample. Br J Clin Psychol 1995; 34(Pt 4): 529–41. Bond J, Carstairs V. Services for the Elderly: A Survey of the Characteristics and Needs of A Population of 5,000,000 Old People. Scottish Home and Health Studies No 42. Edinburgh: Scottish Home and Health Department. 1982. Matthews FE, Stephan BC, Bond J, McKeith I, Brayne C. Operationalization of mild cognitive impairment: a graphical approach. PLoS Med 2007; 4: 1615–9. Stephan BC, Brayne C, McKeith IG, Bond J, Matthews FE. Mild cognitive impairment in the older population: who is missed and does it matter?. Int J Geriatr Psychiatry 2008; 23: 863–71. Rose G, McCartney P, Reid DD. Self-administration of a questionnaire on chest pain and intermittent claudication. Br J Prev Soc Med 1977; 31: 42–8. Stephan BC, Savva GM, Brayne C, Bond J, McKeith IG, Matthews FE. Optimizing mild cognitive impairment for discriminating dementia risk in the general older population. Am J Geriatr Psychiatry 2010; 18: 662–73. Parker CJ, Morgan K, Dewey ME. Physical illness and disability among elderly people in England and Wales: the Medical

Dementia prediction for people with stroke in populations: is mild cognitive impairment a useful concept?

criteria for mild cognitive impairment (MCI) capture an intermediate cognitive state between normal ageing and dementia, associated with increased dem...
121KB Sizes 0 Downloads 8 Views