Alzheimer’s & Dementia 11 (2015) 258-270

Prevalence of dementia in the oldest old: The Monzino 80-plus population based study Ugo Luccaa,*, Mauro Tettamantia, Giancarlo Logroscinob, Pietro Tiraboschic, Cristina Landid, Leonardo Saccoe, Mariateresa Garr
ıa, Sonia Ammessoa, Chiara Bertinottia, Anna Biottia, Elena Gargantinia, Alessandro Piedicorciaa, Alessandro Nobilif, Luca Pasinaf, Carlotta Franchif, Codjo Djignefa Djadef, Emma Rivaa, Angela Recchiaa a

Department of Neuroscience, Laboratory of Geriatric Neuropsychiatry, IRCCS – Istituto di Ricerche Farmacologiche “Mario Negri”, Milan, Italy b Department of Neurosciences and Sense Organs, University of Bari, Bari, Italy c Division of Neuropathology, Fondazione IRCCS Istituto Neurologico “Carlo Besta”, Milan, Italy d Division of Neurological Rehabilitation, European Foundation of Biomedical Research (FERB), Milan, Italy e Division of Neurology, Neurocentre of Southern Switzerland, Lugano, Switzerland f Department of Neuroscience, Laboratory of Quality Assessment of Geriatric Therapies and Services, IRCCS – Istituto di Ricerche Farmacologiche “Mario Negri”, Milan, Italy

Abstract

Background: Epidemiological studies commonly include too few of the oldest old to provide accurate prevalence rates of dementia in older age groups. Estimates of the number of those affected, necessary for healthcare planning, are thus flawed. The objective is to estimate the prevalence of dementia and levels of dementia severity in a very large population of oldest old and to investigate the relation between age and dementia prevalence in the extreme ages. Methods: The Monzino 80-plus is a population-based study among residents 80 years or older in Varese province, Italy. Dementia cases were identified using a one-phase design. The survey was conducted in the participant’s place of residence, whether home or institution. Both participants and informants were interviewed. Information was available for 2504 of the 2813 residents (89%). Results: In all, 894 individuals (714 women and 180 men) met the Diagnostic and Statistical Manual of Mental Disorders (fourth edition) criteria for dementia, for a standardized prevalence of 25.3% (95% confidence interval [CI]: 23.4, 27.2%), 28.5% (95% CI: 26.2, 30.9) in women and 18.6% (95% CI: 15.2, 21.9) in men. Age-specific prevalence estimates of dementia increased with age from 15.7% at age 80 to 84 years to 65.9% at age 100 years and higher. For women, prevalence continued to rise after age 100 years, from 64.8% at age 100 to 101 years to 76.1% at age 102 to 107 years. After age 85 years prevalence rates tended to rise linearly, on average 2.6% per year in women and 1.8% in men. About 80% of the cases were moderate or severe. The frequency of mild dementia decreased and that of severe dementia increased with age. Conclusion: One-quarter of 80-plus year olds are affected by dementia, mostly moderate or severe. Prevalence rates of dementia do not level off, but continue to rise gradually even in the extreme ages. Ó 2015 The Alzheimer’s Association. Published by Elsevier Inc. All rights reserved.

Keywords:

Dementia; Epidemiology; Prevalence; Population-based study; Oldest old

1. Introduction Conflict of interest disclosure: The authors have no potential or actual conflicts of interests to report. *Corresponding author. Tel.: 139-02-39014427; Fax: 139-02-39001916. E-mail address: [email protected]

The oldest old, variously defined as persons aged 80, 85, or even 90 years and older, are largely overlooked in research, in spite of being the fastest growing segment of the population and the one with the highest morbidity

1552-5260/$ - see front matter Ó 2015 The Alzheimer’s Association. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jalz.2014.05.1750

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

burden. Research in the oldest old does in fact pose serious challenges because the prevalence of health problems and sensory impairments, with the resulting functional disability, is high, frailty and fatigue are common, and the assessment of cognitive, occupational, and social functioning is problematic [1,2]. Despite the marked shift toward a much older population structure, the natural narrowing at the very top of the pyramid makes it very difficult to include sufficiently large numbers of oldest individuals even in sizeable population-based studies. As a result individuals aged 80 years and older are under-represented not only in clinical research but also in epidemiological surveys on dementia [3,4]. Many population-based studies on the prevalence of dementia have been conducted worldwide. Most have included a wide age range but only have small numbers of the oldest old and no or too few participants in the extreme age groups [1,2,4,5]. Prevalence and incidence estimates in the oldest old are therefore imprecise and consequently often inadequate to reach trustworthy conclusions. Accurate estimates of dementia in this age group are crucial for a better understanding of the heterogeneous pathological mechanisms of dementia and its relationship with aging [6,7]. These estimates are also indispensable for planning health and social policy [8], especially because, as the world population ages, the social and economic costs of dementia are bound to grow in the near future. The Monzino 80-plus is a prospective population-based study specifically designed to investigate cognitive decline and dementia in a very large, representative population of oldest old, including a considerable number of individuals in the extreme age groups. Aims of the present study were to estimate the prevalence of dementia and its severity levels in the oldest old and investigate the relationship between age and dementia prevalence in the extreme ages. 2. Methods Study design, methods, and measurements have already been described in detail [9]. Briefly, the Monzino 80-plus Study is an ongoing, prospective door-to-door populationbased study among the oldest old residents in the province of Varese, Italy. 2.1. Population and setting Lists of residents were obtained from the municipal registry offices. All individuals 80 years or older residing in Castellanza, Gorla Maggiore, Gorla Minore, Marnate, Olgiate Olona, and Solbiate Olona and 85 years or older residing in Fagnano Olona on February 12, 2002 were eligible for the study. To increase the number of individuals in the extreme age groups, the survey was subsequently extended to all individuals aged 90 years or older residing in the neighboring municipality of Gallarate on January 1, 2005 and, more recently, to all those aged 100 years or older

259

(N 5 272) residing in the remaining municipalities of the province of Varese in 2009. In view of the low number of men aged 95 years or over, the study was further extended to include a random sample of men aged 95 to 99 years (N 5 110), resident in the same municipalities as the centenarians, in the first 9 months of 2010. The study remained confined to the small area of Varese province, historically and culturally very homogeneous, particularly considering the age classes investigated [9]. No exclusion criteria were used, but all individuals in the age groups of interest resident in the study area were included. Study procedures were in accordance with the principles outlined in the Declaration of Helsinki of 1964 and amendments. The study protocol was approved by the local research ethics committee (Azienda Sanitaria Locale of Varese Province). Written informed consent was obtained from participants and informants. 2.2. Study design The survey was conducted in the participant’s place of residence, whether home or institution. All the information was gathered at first visit (one-phase study design) by trained, skilled psychologists using a standardized, reliable questionnaire [9–12]. A proxy informant interview was sought systematically and was available for 90% of participants. For 209 participants living alone and with no informant available, all the information was gathered from the participants. Of these 209 individuals 205 definitely had normal cognitive functioning (mean Mini-Mental State Examination [MMSE] score 27.4) and only four were affected by dementia (mean MMSE score 17.0). During the same visit, testable participants were also administered a multidomain cognitive test battery. For participants physically or mentally unable or unwilling to be tested and for those whose relatives did not consent to have them tested (N 5 354), the information was gathered mainly from the proxy informant and, for individuals in nursing homes, also from the institution’s records. 2.3. Measurements All tests and scales used are widely employed measures with good validity and reliability [9]. Test norms had been previously determined in a sizeable representative population which included large numbers of elderly in the extreme age groups [13]. Tests and scales were administered and scored by specifically trained psychologists, according to standardized criteria [9]. Instruments used in the Monzino 80-plus Study and their psychometric properties have already been reported [9] and will be briefly described here. 2.3.1. Cognitive performance and competence Global cognitive performance was evaluated with MMSE [14]. A set of cognitive tests, mainly from the Consortium to Establish a Registry for Alzheimer’s Disease

260

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

neuropsychological assessment battery [15], was used to investigate diverse cognitive domains. Daily-life cognitive competence was assessed with the Social Interactions section of the Spontaneous Behavior Interview rating scale (SBI-SI) and the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE). The SBI is a multidimensional interview with a proxy informant used to measure the whole spectrum of dementia symptoms in the everyday environment and consists of three subscales assessing basic activities of daily living (bADL), social interactions (SI), and behavioral disturbances (BD) [16]. The IQCODE rates change in everyday cognitive function [17]. When used retrospectively (on average 22 months after death), the IQCODE was shown to be a valid instrument to identify cognitive impairment in individuals with neuropathologically defined Alzheimer’s disease (AD) [18]. To give a global idea of the cognitive status of individuals who died before the planned visit or who could not be tested because of serious sensory deficits, acute medical condition, or refusal, a MMSE score was calculated (convertedMMSE: c-MMSE) by converting the score on SBI-SI following a previously described method [9]. This method of estimating a c-MMSE score was empirically tested in a subpopulation of oldest old for whom the actual scores of MMSE and SBI-SI were available (N 5 1268) and proved to be highly accurate (mean actual score: 22.44, mean estimated score: 22.90) [9]. 2.3.2. Functional disability The bADL section of the SBI assesses five domains of basic self-care ADL [16]. The Instrumental Activities of Daily Living scale (IADL) investigates more complex daily tasks such as the ability to use the telephone, prepare meals, handle finances, etc. [19]. 2.3.3. BD and mood Presence and frequency during the previous month of common BD in dementia were investigated with the BD section of SBI (SBI-BD), whereas depressive symptoms were assessed with ad hoc instruments (the Geriatric Depression Scale-10 and, when cognitive impairment was apparent, the Cornell Scale for Depression in Dementia). 2.4. Diagnosis Dementia was diagnosed according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) [20]. Impairment in social and occupational functioning was investigated mainly with IADL, IQCODE, SBI-SI, and with those sections of the questionnaire devoted to hobbies, participation in leisure and social activities, and to the diary of a typical day. At the end of each interview, field psychologists described and evaluated the cognitive status of the elderly person. Subsequently each study evaluator (a neurologist, a neuropsychologist, and a psychologist) reviewed all the information

collected and reached independently a preliminary diagnosis of dementia syndrome according to DSM-IV criteria. When differences arose in diagnosis, the evaluators reviewed and discussed all available information and reached a final consensus on the presence or absence of dementia at a subsequent diagnostic conference. Diagnoses of dementia made at first visit but not confirmed at later visits were reclassified as “nondementia”. Severity of dementia was staged according to levels of cognitive impairment as measured by the MMSE or c-MMSE score: mild
18, moderate 9 to 17, and severe 0 to 8 [9]. 2.5. Statistical analysis Prevalence was calculated by strata of sex and 5-year age groups with 95% confidence intervals (CIs) obtained with exact binomial methods. The age- and sex-standardized prevalences of dementia were calculated using the direct standardization method with reference to residents in the province of Varese in 2011. In the groups for which a stratification was used (i.e. women 801, men 801, and all 801, and women plus men 95–99 years), CIs were computed considering the proportion of that strata in the population [21]. Difference in prevalence between sexes, correcting for age and education, was modeled using logistic multivariable regression, reporting odds ratios (ORs) and 95% CIs. Relations between dementia prevalence and age (separately by sex and correcting for education in years) or sex (correcting for age and education in years) were modeled using logistic multivariable regressions, reporting ORs and 95% CIs. The Cochran-Armitage trend test was used to look for increases in the prevalence of dementia with age and for the trend of dementia severity levels with age. Survival of participants and nonparticipants and between individuals with and without dementia were compared with a Cox proportional regression model, adjusting for age at prevalence day (i.e. the prespecified time at which the presence of dementia was investigated). Total number of persons affected by dementia in Italy was estimated by multiplying dementia prevalence rates specific for each quinquennia by the number of individuals in the corresponding age range. Cognitive, functional, and behavioral profiles were described using both means and standard deviations (SDs) and medians and upper and lower quartiles to give a comprehensive view of the distribution of the variables, particularly those that were not normally distributed. Because the IADL scale covers eight activities for women and five for men, raw scores were converted to a new score indicating the percentage of functional loss. This new score ranges between 0% and 100% for both men and women with the lowest score indicating no degree of dependence. Differences of baseline cognitive, functional and problem behavior scores among levels of cognitive impairment were tested by analysis of covariance, correcting for age, sex, and years of education. Rates of cognitive and functional decline are significantly higher in elderly individuals with dementia than in those

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

without. Thus, to test longitudinally the accuracy of dementia diagnosis at first visit, we prospectively investigated in the two cognitively close groups in which a diagnosis of dementia could be problematic (i.e. in the mild dementia group and in the nondemented group with MMSE/c-MMSE  26), the changes over time of those scores expected to decline significantly in the presence of a progressive dementing disorder and that were not near the lower limit of the score range at first visit (floor effect) as for IQCODE and IADL. Annual rates of change for MMSE/c-MMSE, SBI-SI, and SBI-bADL scores were calculated subtracting the score at baseline from the score at first follow-up and dividing the result by the number of intercurrent years between the two visits. Statistical significance of differences in rates of change between mild demented participants and participants with a possible mild cognitive impairment was calculated using multiple linear regressions adjusting for age, sex, and education. Data were analyzed using JMP Pro v9.0.3 (SAS Institute Inc., Cary, NC) and STATA/IC v12.0 (StataCorp, College Station, TX).

3. Results 3.1. Population Of the 2813 individuals 80 years or older registered in the municipalities, 127 could not be traced despite numerous attempts. Of the 2686 traced, 488 refused and 2198 agreed to participate. A parallel study with the general practitioners of the same population provided information, in an anonymous form (the only demographic data were sex and age at prevalence day), on the medical history of a further 306 elderly persons among those who refused or could not be found. Thus 2504 individuals (89.0% of the registered population; 80–84: 83.6%, 85–89: 88.6%, 90–94: 90.5%, 95–99: 98.5%, 1001: 93.4%) were included in the main analyses (Supplementary Fig. 1 shows the flow chart of the Monzino 80-plus Study). Information was available also for the 176 individuals who died between the prevalence day and the first planned visit. Demographic characteristics of the original population, the population included in the analyses, and the group of elderly with no information available are reported in Table 1. The percentage of women (about 73%) and mean age (88.9 and 89.1 years) were very similar between the registered and the investigated populations, whereas individuals with no information available were on average a little younger (87.3 years). Socioeconomic and clinical characteristics of the population by dementia status are described in Tables 2 and 3. In this population, mean level of education was about 5 years and three fourths had practiced a manual occupation before retirement. As expected, the percentages of participants living alone were lower among oldest old with dementia than in those without dementia whereas the percentages of participants living in institutions were higher

261

Table 1 Baseline demographic characteristics of the resident population, the population included in the analyses, and the elderly individuals without available information Individuals with information Residents

Available

All, n (%) 2813 (100) 2504 (89) Men, n (%) 768 (27.3) 685 (27.4) Women, n (%) 2045 (72.7) 1819 (72.6) Mean age at 88.9 (6.3) 89.1 (6.4) prevalence day, all (SD) Age groupsy, no. (row %; column %): 80–84 825 (100; 29.3) 690 (83.6; 27.6) 85–89 738 (100; 26.2) 654 (88.6; 26.1) 90–94 705 (100; 25.1) 638 (90.5; 25.5) 95–99 259 (100; 9.2) 255 (98.5; 10.2) 1001 286 (100; 10.2) 267 (93.4; 10.7)

Not available 309 (11) 83 (26.9) 226 (73.1) 87.3 (6.3)*

135 (16.4; 43.7) 84 (11.4; 27.2) 67 (9.5; 21.7) 4 (1.5; 1.3) 19 (6.6; 6.2)

*Approximate estimates of the standard deviation. y Age at interview for resident population and individuals with information; age at prevalence day for individuals without information.

among oldest old with dementia than in those without dementia. Oldest old with dementia were more likely to have a history of heart failure, stroke, and serious hearing impairment. They were also more likely to use more medications, including drugs for dementia and nootropics, but differences were very limited and the proportion of subjects with dementia treated with cholinesterase inhibitors or memantine was very small. 3.2. Prevalence of dementia and levels of dementia severity In all, 894 (180 men and 714 women) 801 year olds met the DSM IV criteria for dementia, for a standardized prevalence of 25.3% (95% CI: 21.3–29.3). For three persons (two only with information from their general practitioners) no diagnosis could be reached. Age-specific prevalence estimates of dementia by sex and levels of severity are presented in Table 4. Prevalence was higher in women (28.5%, 95% CI: 24.6–32.5) than men (18.6%, 95% CI: 14.4–22.7): age and education adjusted OR 1.60 (95% CI: 1.29–2.00, P , .0001). The standardized prevalence of dementia in the 176 individuals who died between the prevalence day and the first planned interview was higher (36.6%) than in the population alive at interview (24.8%). Of the 74 who died between the prevalence day and the planned interview and were classified as having dementia at prevalence day, only six had mild dementia and 25 had already had a diagnosis of dementia during their lives. If individuals who died are excluded from the analyses, the age-specific prevalence estimates are similar: 14.7% (80–84 years), 32.8% (85–89 years), 42.2% (90– 94 years), 56.1% (95–99 years), and 66.7% (1001 years). If the 20 individuals with a mild dementia and a serious sensory impairment were reclassified as nondemented, prevalence estimates would remain almost the same.

262

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Table 2 Baseline socioeconomic characteristics of participants with and without dementia Women

No. Mean age (SD) Mean education (SD) Main occupation*, % Manual workers Crafts and trades Professionals and clerks Marital status, % Never married Married Widowed Living alone,% Living in institutions, % Economic condition, insufficient, % Principal informant, child, or spousey, % Informant’s mean agey (SD) Informant’s mean educationy (SD)

Men

All

Without dementia

With dementia

Without dementia

With dementia

Without dementia

With dementia

935 88.7 (5.6) 5.0 (2.0)

663 92.9 (6.0) 4.4 (2.0)

439 88.5 (5.6) 6.2 (3.4)

160 91.9 (5.4) 5.6 (3.6)

1374 88.7 (5.6) 5.4 (2.6)

823 92.7 (5.9) 4.6 (2.5)

78.7 11.7 8.3

82.8 8.9 5.9

60.1 18.5 17.4

61.9 15.0 16.9

72.8 13.9 11.2

78.6 10.1 8.0

9.6 7.6 82.5 49.4 6.1 5.6 72.4 60.3 (10.3) 8.8 (3.7)

8.7 3.2 88.5 18.3 34.1 13.8 75.9 61.1 (9.6) 8.5 (3.8)

3.0 55.6 40.7 22.1 6.4 6.6 87.2 66.5 (13.8) 8.5 (4.1)

5.7 46.5 47.8 9.7 28.8 12.3 83.8 62.6 (11.9) 8.5 (4.3)

7.5 23.0 69.1 40.7 6.2 6.0 77.5 62.5 (12.0) 8.7 (3.8)

8.1 11.6 80.2 16.5 33.1 13.5 77.3 61.4 (10.0) 8.5 (3.9)

Abbreviation: SD, standard deviation. *Organizing the major groups of the International Standard Classification of Occupations (ISCO-08) in three classes. y Doctors and nurses of geriatric institutions were excluded from the analysis.

The 2-year survival rate from prevalence day was significantly lower in individuals with dementia than in those without in the whole population (age and sex adjusted hazard ratio [HR] 1.91, 95% CI: 1.63–2.25, P , .0001) and in each age group (specifically, in 80–84 years: age and sex adjusted HR 3.04, 95% CI: 1.92–4.71, P , .0001). To see whether the inclusion of the 309 individuals with no information influenced the estimated prevalence of dementia, we compared the 2-year survival rate from the prevalence day for participants and for individuals who refused or could not be traced: the risk of mortality was not significantly different (age and Table 3 Main clinical features of the population with available information (2502) by dementia status at baseline

No. Hypertension, % Diabetes, % Myocardial infarction, % Heart failure, % Chronic pulmonary disease, % Stroke, % Serious hearing impairment, % Serious visual impairment, % Osteoarthrosis, % Hospital admission, previous year, %y Drug use, % Mean number of drugs per users, (SD) Cholinesterase inhibitors and memantine, % users Nootropics, % users

Without dementia

With dementia

P-value*

1539–1606 58.9 12.1 7.5 25.8 16.9 7.9 3.9 5.8 68.6 28.4 89.0 3.3 (1.9) 0.3

872–893 56.1 11.9 8.1 38.4 21.3 21.5 12.3 8.6 73.0 28.0 91.6 3.6 (1.9) 2.3

.0783 .1556 .1284 ,.0001 .0153 ,.0001 ,.0001 .4999 .2053 .0604 .7395 .0007 ,.0001

0.8

1.8

Abbreviation: SD, standard deviation. *Age and sex adjusted (covariance analysis). y n 5 2197 (no information on hospital admissions from GPs).

.0285

sex adjusted HR 0.97, 95% CI: 0.86–1.23, P 5 .7440). This result did not change when the analyses were restricted to the age groups 80 to 84 years (age and sex adjusted HR 0.99, 95% CI: 0.67–1.44, P 5 .9690) or 85 to 90 years (age and sex adjusted HR: 1.07, 95% CI: 0.75–1.49, P 5 .6907). If 306 individuals with information available only from their general practitioners are excluded, the overall prevalence (26.1%) and the age-specific prevalence rates, though slightly higher (10.8% on average), do not change substantially: 16.4% (80–84 years), 33.9% (85–89 years), 43.2% (90–94 years), 56.0% (95–99 years), and 66.3% (1001 years). Almost half the cases were in a moderate stage of dementia (46.8%), whereas 21.1% were in a mild and 32.1% in a severe stage. The percentage of mild dementia cases decreased with age (80–84 years: 34.3%; 85–89 years: 26.5%; 90–94 years: 22.6%; 95–99 years: 14.4%; 1001 years: 9.1%; CochranArmitage trend test: P , .0001), whereas that of severe dementia increased with age (80–84 years: 22.2%; 85–89 years: 23.3%; 90–94 years: 27.4%; 95–99 years: 34.3%; 1001 years: 50.6%; Cochran-Armitage trend test: P , .0001) (Supplementary Fig. 2). Comparable results were obtained when subjects with severe sensory impairment in mild (N 5 24; 12.7%), moderate (N 5 84; 20.1%), and severe (N 5 50; 17.4%) dementia groups were excluded from the analysis (Cochran-Armitage trend tests: P , .0001). 3.3. Relationship between age and dementia prevalence The prevalence rose progressively with age (Cochran-Armitage trend test P ,.0001) in both men (education adjusted one-year OR 1.11, 95% CI: 1.07–1.14, P , .0001) and women (1.13, 95% CI: 1.11–1.15, P , .0001). In women,

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

263

Table 4 Prevalence of dementia by age, sex, and severity in the elderly population of the Monzino 80-plus Study Dementia Age groups, yr Women 80–84 85–89 90–94 95–99 1001 All (801) Men 80–84 85–89 90–94 95–99 1001 All (801) All 80–84 85–89 90–94 95–99 1001 All (801)

Mild dementia

Moderate dementia

Severe 1 dementia

Population no.

Cases no.

Prevalence, % (95% CI)

461 485 497y 129 245 1817

81 168 225 76 164 714

17.6 (14.2–21.4) 34.6 (30.4–39.1) 45.3 (40.8–49.8) 58.9 (49.9–67.5) 66.9 (60.7–72.8) 28.5 (24.6–32.5)*

5.6 (3.7–8.2) 8.3 (6.0–11.1) 8.7 (6.3–11.5) 8.5 (4.3–14.8) 5.3 (2.9–8.9) 7.0 (4.4–9.6)*

7.6 (5.3–10.4) 17.5 (14.3–21.2) 22.9 (19.3–26.9) 31.8 (23.9–40.6) 27.4 (21.9–33.4) 13.8 (10.9–16.8)*

4.3 (2.7–6.6) 8.9 (6.5–11.8) 13.7 (10.8–17.0) 18.6 (12.4–26.4) 34.3 (28.4–40.6) 7.7 (5.8–9.7)*

228z 169 139 126 22 684

27 47 45 49 12 180

11.8 (8.0–16.8) 27.8 (21.2–35.2) 32.4 (24.7–40.8) 38.9 (30.3–48.0) 54.5 (32.2–75.6) 18.6 (14.4–22.7)*

4.8 (2.4–8.5) 10.1 (6.0–15.6) 13.0 (7.9–19.7) 5.6 (2.3–11.1) 13.6 (2.9–34.9) 7.0 (4.3–9.6)*

5.3 (2.8–9.0) 13.6 (8.8–19.7) 15.1 (9.6–22.2) 12.7 (7.4–19.8) 18.2 (5.2–40.3) 8.6 (5.6–11.6)*

1.8 (0.5–4.4) 4.1 (1.7–8.4) 4.3 (1.6–9.2) 20.6 (13.9–28.8) 22.7 (7.8–45.4) 3.0 (1.1–5.0)*

689z 654 636y 255 267 2501y,z

108 215 270 125 176 894

15.7 (13.0–18.6) 32.9 (29.3–36.6) 42.5 (38.6–46.4) 55.3 (53.2–57.5)* 65.9 (59.9–71.6) 25.3 (21.3–29.3)*

5.4 (3.8–7.3) 8.7 (6.7–11.1) 9.6 (7.4–12.2) 8.0 (6.9–9.1)* 6.0 (3.5–9.6) 7.0 (4.4–9.6)*

6.8 (5.1–9.0) 16.5 (13.8–19.6) 21.2 (18.1–24.6) 28.4 (26.6–30.2)* 26.6 (21.4–32.3) 12.1 (9.2–15.1)*

3.5 (2.2–5.1) 7.7 (5.7–10.0) 11.6 (9.3–14.4) 19.0 (17.2–20.7)* 33.3 (27.7–39.3) 6.2 (4.3–8.1)*

Abbreviation: CI, confidence interval. *Standardized to the 2011 population of Varese province. y Plus two women (90–94 years). z One man (80–84 years) with no possible diagnosis.

the prevalence continued to rise even after the age of 100, from 64.8% at age 100 to 101 years (N 5 199) to 76.1% at age 102 to 107 years (N 5 46); there were too few men aged 102 years or more (N 5 4) to be analyzed separately (Fig. 1). After age 85 years, prevalence rates tended to slow but still rose linearly, with an average increase of 2.5% per year (95% CI: 1.8–3.2), 1.8% (95% CI: 0.5–3.1) in men and 2.6% (95% CI: 1.6–3.5) in women. 3.4. Estimate numbers with dementia Using age-specific prevalence of dementia from the pooled results of Italian studies for the younger age groups (55–79 years) [4] and from the present study for the older groups (801 years), there were about 1,330,000 individuals with dementia in Italy in 2011, 68% were 80 years or older and 45% were 85 years or older. If age-specific prevalence rates remain stable using the Istituto Nazionale di Statistica’s middle estimates for Italian population growth, in 2050 more than 2,900,000 Italians will have dementia, 79% of them aged 80 years or more and 60% 85 years or older (Fig. 2). 3.5. Accuracy of dementia diagnosis 3.5.1. Cross-sectional evaluation Table 5 shows the baseline cognitive, functional, and problem behavior scores of participants in relation to demen-

tia and cognitive status. Functional disability and the presence of BD increased steadily with decreasing cognitive performance in all groups (age, sex, and education adjusted P . .0001). At first visit, ratings in the mild dementia group were all noticeably worse than in the nondemented group with MMSE/c-MMSE  26, i.e. the group including most of the oldest old with a mild cognitive impairment (Table 5). 3.5.2. Longitudinal evaluation Of the 835 individuals classified in the two cognitively close groups in which a diagnosis of dementia could be problematic (i.e. in the mild dementia group and in the nondemented group with MMSE/c-MMSE  26) at initial visit, 710 were alive at first follow-up visit and 666 (93.8%) had available information. Mean (SD) annual rates of cognitive and functional decline between the initial visit and first follow-up (1.7 years on average) were higher in the group with mild dementia compared with the group with a possible mild cognitive impairment, respectively: 23.5 (4.0) versus 21.1 (3.4) on MMSE/c-MMSE (age, sex, and education adjusted P , .0001); 3.6 (4.7) versus 2.2 (4.0) on SBI-SI (age, sex, and education adjusted P 5 .0017); and 3.7 (5.1) versus 2.1 (3.6) on SBI-bADL (age, sex, and education adjusted P , .0001). 4. Discussion 4.1. Summary of the findings This sizeable population-based study showed that about one fourth of the 80-plus year olds and more than one-third

264

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Fig. 1. Prevalence (95% confidence intervals) of dementia by age and sex in the Monzino 80-plus Study.

of the 85-plus year olds (36.9%; 95% CI: 34.2–39.6%) suffered from dementia. About 80% were in a moderate or severe stage. The prevalence of dementia rose gradually with age and from the large number of very old cases it was clear that the prevalence does not level off in the extreme ages, in either women or men. The number of women aged 102 years or more was large enough (N 5 46) to explore the prevalence of dementia in this extreme age group as well: estimates went on increasing even after 101 years, with more than three fourths of women affected in the102-plus group (67% of the 245 aged 100-plus, 70%

of the 100 aged 101-plus, and 76% of the 46 aged 102plus). The high prevalence of disease and disability among the oldest old and their naturally limited occupational functioning can sometimes make the diagnosis of mild dementia problematic. On the other hand, the diagnosis of mild dementia at first visit proved rather accurate both at crosssectional and longitudinal evaluations in the groups at risk of misdiagnosis. Furthermore, the decreasing frequency of mild dementia with age reduces the risk of misclassification. But even reclassifying all cases previously classified as mild

Fig. 2. Estimated numbers of cases of dementia in Italy in 2011 and 2050. *Percentages of subjects with dementia 80 years or older on the total number of dementia cases.

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

265

Table 5 Mean (SD) and median (25th–75th percentiles) baseline cognitive, functional, and problem behavior scores by diagnosis and cognitive status of participants Dementia No dementia

Mild

Moderate

Severe1


18

9–17

0–8

142 19.8 (1.5) 19 (18–21) 4.4 (0.4) 4.4 (4.1–4.7) 11.7 (3.8) 12 (9–14) 9.4 (6.8) 8 (4–15) 79.1 (21.5) 85 (67–96) 1.6 (2.3) 0 (0–3)

403 13.6 (2.6) 14 (12–16) 4.7 (0.4) 4.9 (4.6–5.0) 17.0 (4.7) 17 (14–20) 15.4 (7.5) 16 (10–22) 91.1 (16.1) 100 (90–100) 2.4 (3.2) 1 (0–4)

276 3.7 (3.2) 4 (0–7) 4.9 (0.2) 5.0 (5.0–5.0) 29.3 (4.7) 30 (26–34) 25.5 (5.2) 28 (23–30) 99.1 (4.5) 100 (100–100) 2.7 (3.8) 1 (0–4)

MMSE Test or scale (score range: best-worst) MMSE or converted-MMSE (30–0) Informant Questionnaire on Cognitive decline in the elderly (1–5) Spontaneous Behavior Interview—social interaction (0–34) Spontaneous Behavior Interview—basic activities of daily living (0–30) Instrumental Activities of Daily Living, % of lost functional ability (0–100%) Spontaneous Behavior Interview—behavioral problems (0–27)

27–30

26

No. 682 691 2194 28.2 (1.0) 23.7 (2.3) 28 (27–29) 24 (22–25) 2011 3.2 (0.3) 3.6 (0.4) 3.1 (3–3.3) 3.5 (3.3–3.9) 2184 1.0 (1.6) 4.4 (3.1) 0 (0–1) 4 (2–7) 2190 2.0 (3.8) 4.8 (5.9) 0 (0–2) 2 (0–8) 2183 19.0 (19.6) 43.1 (28.2) 15 (0–29) 38 (22–65) 2177 0.1 (0.5) 0.4 (1.1) 0 (0–0) 0 (0–0)

Abbreviations: SD, standard deviation; MMSE, Mini-Mental State Examination. NOTE. For all comparisons (analysis of variance) both among all groups and between the no-dementia group with MMSE or converted-MMSE 26 and the mild dementia group (MMSE or converted-MMSE
18), age, sex, and education adjusted P values ,.0001.

demented as nondemented, the rates of moderate-severe dementia would continue to steadily increase with age in parallel with those reported for mild-moderate-severe dementia (see also Table 2). Thus, the increasing prevalence of dementia in the oldest age segments cannot be explained by a higher misclassification rate with age. Many studies have described the sharp rise in dementia prevalence with age as exponential, with rates roughly doubling every 5 years at least until the age of 85 years. But this exponential model has problems, because it predicts that, contrary to the evidence, around age 100 years everyone should have dementia. In this large population of oldest old, although the prevalence rates of dementia continued rising steadily, estimates after age 85 years did not appear to double every 5 years but tended to increase linearly. Thus, according to this model, all women would have dementia around age 115 years and all men around age 125 years. However, although a 115-year-old woman without dementia and minimal neuropathology has been described [22], men and women reaching these ages are so rare that presently any conclusion would be unsound. 4.2. Comparison with previous studies in the oldest old 4.2.1. Prevalence of dementia Although dementia is primarily a condition of advanced age, trustworthy information in this crucial age group has come from a limited number of studies as highlighted in several recent reviews [1,2,5]. Inconsistency of dementia prevalence estimates between studies tends to increase with individuals’ ages [23,24]. The oldest old investigated in most studies, when present, have been too few to reliably estimate the prevalence of dementia in this

segment of the population. Even when the number of patients who were 80 years and higher or 85 years and higher was relatively large or the studies deliberately focused on the oldest old, extreme ages have mostly been poorly represented (just four studies included more than 100 individuals 95 years and older [25–28]). Among studies that reported age-specific prevalence rates of dementia, only four included at least 500 individuals aged 85 years and older, with at least 90 of them aged 95 years and older [25,26,28,29]; two other studies which included more than 600 90-plus year olds, unfortunately reported only the overall prevalence of dementia in this age, without any subgroup analyses [30,31]. In these studies (see Fig. 3), estimates in the oldest age groups do not appear to fan out, whereas all the agespecific estimates rise roughly parallel within a rather narrow range of differences between studies. This relatively small variability of estimates (approximately 10%) might be due to methodological differences, but also to real differences between populations. In this regard, it is intriguing that populations with comparable levels of formal education gave very similar results, those with a lower educational level (like the G€oteborg and Monzino 80-plus studies) in the upper part of the range and those with a higher educational level (like the Kungsholmen-St Goran and 901 studies) in the lower part. 4.2.2. Prevalence of severity levels of dementia Few studies reported the prevalence of severity levels of dementia. Differences in methodology and age composition and size of the populations investigated make comparisons between studies challenging. Notwithstanding the different criteria used for classifying the severity levels,

266

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Fig. 3. Age-specific prevalence rates of dementia in studies with no less than 500 individuals aged 851 and with at least 90 of them aged 951. G€oteborg Study: prevalences were estimated from published data. The number of 100 years or older subjects included in the age group 951 was 8 in the Kungsholmen-St G€oran Study and 13 in the Canadian Study of Health and Aging (CSHA).

the prevalence of mild dementia in the present study was similar to that in other studies for ages 801 [32] and 851 [33]. Among cases, the frequency of mild dementia in the present study declined whereas moderate-severe dementia rose with age, a finding in agreement with that of the other studies reporting the severity levels of dementia by age group in individuals aged 80/85 years or older notwithstanding the different staging criteria used (DSM-III-R, Clinical Dementia Rating, and their own criteria) [28,29,32–37]. Thus, the declining frequency of mild dementia with age does not appear to be related to the type of severity criteria used, whether staging cognitive impairment (as the MMSE), daily function (as the DSMIII-R), or both (as the Clinical Dementia Rating). The large differences found between mild and moderate dementia groups in daily function abilities (see Table 5), frequency of institutionalization, and the number of daily hours of care (data not shown) also suggest that had we used DSM-III-R criteria to stage oldest old with dementia, results would not have noticeably changed from those reported. It should be also noted that studies using DSMIII-R staging criteria often reported both the highest and the lowest estimates of mild dementia prevalence in the same age group [28,34,36,37]. This indicates that, as predictable from the imprecision in the definitions of the severity stages, DSM-III-R criteria for severity of dementia are differently interpreted and operationalized by diverse raters. On average, only one out of five cases 85 years and older had mild dementia, a rate comparable with other reports: 20.7% in the CSHA, and around 10% in the Vantaa 851 study [25,38]. This proportion dropped to 1 in 10 for dementia sufferers aged 95 years and older,

in line with the G€oteborg (5.2%) and Kungsholmen-St Goran (8.3%) figures [28,29]. 4.3. Strengths To improve the accuracy and reliability of estimates in the most advanced ages, this sizeable representative population of oldest old included a considerable number of individuals in the extreme age groups. The analyses included a very high proportion of the original total population (89%), exceeding 95% in the 95 to 99 years and 100-plus age groups. Only a small number of individuals had no information, so a nonresponse bias is unlikely and would in any case have very limited influence. Moreover, in contrast with the consistent and robust association between shorter survival and dementia observed in the present study as well, participants and individuals who refused or could not be traced showed very similar mortality rates. A one-phase design made it possible to avoid potential bias due to the lack of sensitive and specific screening tools for the oldest old and to nonresponse in the second phase of the study [39,40]. The validity and reliability of scales and tests used and norms for cognitive tests pertaining to the oldest old were available. Relying on a wide array of tools to assess the multifaceted manifestations of dementia has also been proven appropriate and effective in a population with high prevalence of morbidity, disability, and cognitive impairment such as the oldest old. Incomplete or missing neuropsychological testing is not a random occurrence in this age segment, but is strongly associated with increasing age and the presence and severity of dementia [9,26,41]. Hence,

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

the importance of measuring cognitive and functional abilities in the everyday environment as well to get more complete and reliable information in a field survey of the very old. The informant interview is a valid measure of cognitive decline in the impaired elderly and a sensitive instrument for detecting the antemortem presence of dementia [18,42,43]. Moreover, by measuring observable behaviors, informant-based instruments help establish the impact of current deficits in cognitive tests on everyday functioning and assess changes from the premorbid level. This longitudinal approach proved to be a valid aid in the diagnostic process. The consistency and accuracy of dementia diagnoses and consequently of the estimated prevalences was strengthened by the possibility to confirm the diagnosis made at first visit at later visits. Moreover, using prospective data made it possible to further test the accuracy of the initial diagnostic classification: as one would have expected in the case of subjects with and without dementia, the rates of cognitive and functional decline over time were much higher in the group with mild stage dementia compared with the cognitively close group with a possible mild cognitive impairment at the initial visit. 4.4. Limitations (1) In general, findings in a defined geographical area, inevitably reflecting the specific characteristics of the population under study, may not necessarily be replicated in other population-based surveys. (2) To be able to include a large number of individuals in the extreme age groups, the economic burden of the study had to be partitioned over time and some 7 to 8 years elapsed between the inclusion of the 80 years and older residents in the initial study area and the 100-plus-year-old and 95- to 99-year-old-men groups residing in the remaining municipalities of the small province of Varese. However, the stability and cultural homogeneity of the population in this age group is very high and, over such a short period changes in life expectancy for the age groups investigated were negligible. (3) Participation in the study was lower in the younger than in the older age groups. This may have contributed to differences in dementia prevalence between groups and consequently to the relationship between age and prevalence of dementia. However, in the 80- to 84-year age group the prevalences of dementia in individuals with information available (83.6% of residents) and in participants were similar (15.7% and 16.4%, respectively). Moreover, persons with dementia have markedly decreased survival rates compared with those without dementia, whereas survival rate in participants and nonparticipants was almost the same in the younger age groups. Thus, it is very likely that prevalence rates of dementia in the younger age groups of nonparticipants are fairly similar to those estimated. (4) For a small number of individuals information on health status could only be obtained from their personal general practitioners, but their exclusion from the analyses would not substan-

267

tially change the estimated prevalence rates. (5) No proxy informant was available for a relatively small proportion of participants living alone, however, almost all these individuals (98%) were completely independent and definitely showed normal cognitive functioning. (6) The information on cognitive status of participants physically or mentally unable or unwilling to be tested and of those whose relatives did not consent to have them tested came mainly from the proxy informant interview. However, participants without cognitive testing are in fact older, more often institutionalized, and have a much worse cognitive, functional, and behavioral profile than those with [9,26,41]. Thus the failure to include these individuals would underestimate the prevalence of dementia in the general population. (7) Serious sensory impairments are common among the oldest old and can be an obstacle to the assessment of cognitive performance in surveys investigating old age populations. Informant-based measures of cognitive competence and decline represent an appropriate strategy to try to overcome this difficulty and obtain valid and reliable information. But even assuming that all individuals with mild dementia and a serious sensory impairment were actually nondemented, the estimates of the prevalence would remain almost unchanged and the relationship between age and dementia prevalence would not be modified. It is also rather unlikely that dementia cases were increasingly misclassified as having more severe dementia with age only because they scored lower on MMSE due to the increasing rates of severe sensory impairments with age. In fact, the frequency of mild dementia continues to decrease and that of severe dementia to increase with age in the same fashion also when the oldest old with severe sensory impairment are excluded from the analysis. Moreover, informant-based cognitive, functional, and behavioral ratings of subjects without and with severe sensory impairment in each stage group are very similar (data not shown). 4.5. Public health impact and implications for future research The estimated percentage of persons with dementia in the entire Italian population in 2011 was quite high (2.2%) and very similar to that for the US population in 2010 (1.6% for AD, thus approximately 2.3% for dementia considering that AD accounts for about 70% of all dementia cases in the United States) [44,45]. Not considering the possible increase in survival of persons with dementia due to improvements in care, by 2050 this already high percentage is expected to double (4.6%; 4.3% in the US) [44,45]. But developing countries are aging rapidly as well and according to a Delphi consensus study, 60% of all dementia sufferers lived in developing countries in 2001, a percentage that will rise to 71% by 2040 [46,47]. Despite their limitations, these estimates give a good idea of the present and future magnitude of the problem. In a rapidly aging world, enormous resources will be required

268

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

to care adequately for persons with dementia [48]. At present, more than two thirds of all dementia cases are 80 years and older and almost half are 85 years and older (43% of the AD population in the United States) [44]. By 2050 persons with dementia in these age groups will be the vast majority of dementia cases (80-plus: 79%; 85-plus: 60%). However, “dementia research is systematically biased toward patients who are young relative to the many who have dementia” [3, p. 629]. This age gap may be an important source of bias and an obstacle to appropriate generalizability: findings in the younger old might in fact not apply to the oldest old [3,49]. The relationship between neuropathological features and dementia has also been reported to change with age [50]. 5. Conclusions Health and social policy makers need accurate estimates of the numbers of people who currently have dementia to plan adequate medical care and health services [8]. Differences in prevalence rates would significantly affect the projections of social and economic costs of dementia. Moreover, “people with dementia at every stage and age will have greatly different needs” [49, p. 1442]. The Monzino 80-plus is one of the largest prospective populationbased studies specifically aimed at thoroughly investigating cognitive decline and dementia in the oldest old. The present study shows what a large proportion of the oldest old have dementia and that at this advanced age four out of five cases are in a moderate or severe stage. Prevalence rates of dementia do not level off but continue to rise even in the extreme ages, although after age 85 the rate does not double every 5 years but tends to slow down. Individuals included in basic, clinical, and epidemiological research on dementia need to appropriately reflect the population at risk if successful strategies to postpone cognitive decline and thus reduce the burden of dementia on the individual and society are to be developed. Acknowledgments The authors are very grateful to all the participants who made this investigation possible and to the “Monzino 80plus Study” Group (full list in the supplementary material). The Monzino 80-plus Study is being supported by a research grant from Italo Monzino Foundation, Milan, Italy. The Italo Monzino Foundation, a nonprofit organization, had no role in the conception and design of the study; the collection, management, analysis, and interpretation of data; and the preparation and writing of the report or in the decision to submit the manuscript for publication. UL, MT, and MG conceived and designed the study. MT established the database. UL, MT, PT, CL, LS, MG, SA, CB, AB, EG, AP, AN, LP, CF, CDD, ER, AR were responsible for data collection and management. MT and AR did the statistical analysis. UL, MT, GL, PT, MG, and AR analyzed and

interpreted the data. UL drafted the manuscript. All authors critically revised the manuscript and approved its final version. UL, MT, and AR are the guarantors. We are grateful to J.D. Baggott for language editing.

RESEARCH IN CONTEXT

1. Systematic review: We searched PubMed with the terms “dementia” and “prevalence OR epidemiology”. We also identified studies from the references of the included papers and from the authors’ personal files. There are usually too few oldest old included in epidemiological studies to provide accurate prevalence rates of dementia in the older age groups. Estimates of the numbers required for healthcare planning are thus flawed. Many studies have described the steep increase in dementia prevalence with age as exponential, with rates roughly doubling every 5 years until about the age of 80 to 85 years. It is not yet clear whether the prevalence of dementia continues to increase and if so whether the pattern of growth is still exponential after the age of 85 years. 2. Interpretation: Prevalence rates of dementia do not level off in the oldest old, but continue to rise gradually even in the extreme ages. One-quarter of the over-80s suffer from dementia, mostly moderate, or severe. More than two thirds of all dementia cases are 80 years and older and this proportion is destined to grow markedly over the coming decades. 3. Future directions: Further similar population-based studies including large numbers of oldest old are needed. More generally, individuals included in basic, clinical, and epidemiological research on dementia need to appropriately reflect the population at risk if successful strategies to postpone cognitive decline and thus to reduce the burden of dementia on the individual and society are to be developed.

References [1] Mayeux R, Reitz C, Brickman AM, Haan MN, Manly JJ, Glymour MM, et al. Operationalizing diagnostic criteria for Alzheimer’s disease and other age-related cognitive impairment—part 1. Alzheimers Dement 2011;7:15–34. [2] Brumback-Peltz C, Balasubramanian AB, Corrada MM, Kawas CH. Diagnosis dementia in the oldest-old. Maturitas 2011;70:164–8. [3] Schoenmaker N, Van Gool WA. The age gap between patients in clinical studies and in the general population: a pitfall for dementia research. Lancet Neurol 2004;3:627–30.

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270 [4] Lucca U, Tediosi F, Tettamanti M. Epidemiologic dimension and economic impact of dementia in Italy. [Dimensione epidemiologica e impatto economico delle demenze in Italia]. Milano: Emme Edizioni; 2001. [5] Kawas CH, Corrada MM. Alzheimer’s and dementia in the oldest-old: a century of challenges. Curr Alzheimer Res 2006;3:411–9. [6] Ritchie K, Kildea D. Is senile dementia “age-related” or “ageingrelated”?—evidence from meta-analysis of dementia prevalence in the oldest old. Lancet 1995;346:931–4. [7] McGee MA, Brayne C. The impact on prevalence of dementia in the oldest age groups of differential mortality patterns: a deterministic approach. Int J Epidemiol 1998;27:87–90. [8] Knapp M, Prince M, eds. Dementia UK. A report into the prevalence and cost of dementia. London: Alzheimer’s Society; 2007. [9] Lucca U, Garr
ı M, Recchia A, Logroscino G, Tiraboschi P, Franceschi M, et al. A Population-based study of dementia in the oldest old: the Monzino 80-plus Study. BMC Neurol 2011;11:54. [10] Lucca U, Tettamanti M, Mosconi P, Apolone G, Gandini F, Alessandro Nobili, et al. Association of mild anemia with cognitive, functional, mood and quality of life outcomes in the elderly: the “Health and Anemia” study. PLoS ONE 2008;3:e1920. [11] Riva E, Tettamanti M, Mosconi P, Apolone G, Gandini F, Nobili A, et al. Association of mild anemia with hospitalization and mortality in the elderly: the Health and Anemia population-based study. Haematologica 2009;94:22–8. [12] Lucca U, Nobili A, Riva E, Tettamanti M. Cholinesterase inhibitor use and age in the general population. Arch Neurol 2006;63:154–5. [13] Lugli A, Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). CERAD. Normative study of the neuropsychological assessment battery in an Italian speaking population. Postgraduate dissertation, in Italian. [Studio normativo della batteria di valutazione neuropsicologica su una popolazione di lingua italiana]. Pavia, Italy: Universit
a degli Studi di Pavia, Scuola di Specializzazione in Psicologia; 2002. [14] Folstein MF, Folstein SE, McHugh PR. “Mini-Mental State”. A practical method for grading the cognitive state of patients for the clinician. J Psychiat Res 1975;12:189–98. [15] Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G, et al., the CERAD investigators. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD): Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology 1989;39:1159–65. [16] Lucca U, Tettamanti M, Martelli P, Lucchelli F, Alberoni M, Spagnoli A. Spontaneous behaviour interview: from test performance to daily life-based Alzheimer patient evaluation [abstract]. The challenge of the dementias. Edinburgh, UK: The Lancet; 1996, 58. [17] Jorm AF, Jacomb PA. The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): socio-demographic correlates, reliability, validity and some norms. Psychol Med 1989;19:1015–22. [18] Thomas LD, Gonzales MF, Chamberlain A, Beyreuther K, Masters CL, Flicker L. Comparison of clinical state, retrospective informant interview and the neuropathologic diagnosis of Alzheimer’s disease. Int J Geriatr Psychiatry 1994;9:233–6. [19] Lawton MP, Brody EM. Assessment of older people. Self-maintaining and instrumental activities of daily living. Gerontologist 1969; 9:179–86. [20] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. Fourth edition. Washington, DC: American Psychiatric Association; 1994. DSM-IV. [21] Cochran WG. Sampling techniques. Third Edition. New York: John Wiley & Sons; 1977. [22] den Dunnen WF, Brouwer WH, Bijlard E, Kamphuis J, van Linschoten K, Eggens-Meijer E, et al. No disease in the brain of a 115-years-old woman. Neurobiol Aging 2008;29:1127–32. [23] Lobo A, Launer LJ, Fratiglioni L, Andersen K, Di Carlo A, Breteler MM, et al. Prevalence of dementia and major subtypes in Europe: a collaborative study of population-based cohorts. Neurologic

[24] [25]

[26]

[27] [28] [29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39] [40] [41]

[42]

[43]

269

Diseases in the Elderly Research Group. Neurology 2000;54(11 Suppl 5):S4–9. Berr C, Wancata J, Ritchie K. Prevalence of dementia in the elderly in Europe. Eur Neuropsychopharmacol 2005;15:463–71. Ebly EM, Parhad IM, Hogan DB, Fung TS. Prevalence and types of dementia in the very old: results from the Canadian Study of Health and Aging. Neurology 1994;44:1593–600. Corrada MM, Brookmeyer R, Berlau D, Paganini-Hill A, Kawas CH. Prevalence of dementia after age 90. Results from The 901 Study. Neurology 2008;71:337–43. Mathillas J, L€ovheim H, Gustafson Y. Increasing prevalence of dementia among very old people. Age Ageing 2011;40:243–9. B€orjesson-Hanson A, Edin E, Gislason T, Skoog I. The prevalence of dementia in 95 years olds. Neurology 2004;63:2436–8. von Strauss E, Viitanen M, De Ronchi D, Winbland B, Fratiglioni L. Aging and the occurrence of dementia: findings from a populationbased cohort with a large sample of nonagenarians. Arch Neurol 1999;56:587–92. Saunders PA, Copeland JR, Dewey ME, Gilmore C, Larkin BA, Phaterpekar H, et al. The prevalence of dementia, depression and neurosis in later life: The Liverpool MRC-ALPHA Study. Int J Epidemiol 1993;22:838–46. Llibre J de J, Fernandez Y, Marcheco B, Contreras N, Lopez AM, Otero M, et al. Prevalence of dementia and Alzheimer’s disease in a Havana municipality: a community-based study among elderly residents. MEDICC Rev 2009;11:29–35. O’Connor DW, Pollitt PA, Hyde JB, Fellows JL, Miller ND, Brook CP, et al. The prevalence of dementia as measured by the Cambridge Mental Disorders of the Elderly Examination. Acta Psychiatr Scand 1989;79:190–8. Graham JE, Rockwood K, Beattie BL, Eastwood R, Gauthier S, Tuokko H, et al. Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet 1997; 349:1793–6. Heeren TJ, Lagaay AM, Hijmans W, Roojmans HG. Prevalence of dementia in the ‘oldest old’ of a Dutch community. J Am Geriatr Soc 1991;39:755–9. Juva K, Sulkava R, Erkinjuntti T, Valvanne J, Tilvis R. Prevalence of dementia in the city of Helsinki. Acta Neurol Scand 1993; 87:106–10. Boersma F, Eefsting JA, van den Brink W, Koeter M, van Tilburg W. Prevalence of dementia in a rural Netherlands population and the influence of DSM-III-R and CAMDEX criteria for the prevalence of mild and more severe forms. J Clin Epidemiol 1998;51:189–97. Riedl-Heller SG, Busse A, Aurich C, Matschinger H, Angermeyer MC. Prevalence of dementia according to DSM-III-R and ICD-10. Results of the Leipzig Longitudinal Study of the Aged (LEILA751) Part I. Br J Psychiatry 2001;179:250–4. Sobel E, Louhija J, Sulkava R, Davanipour Z, Kontula K, Miettinen H, et al. Lack of association of apolipoprotein E allele epsilon 4 with lateonset Alzheimer’s disease among Finnish centenarians. Neurology 1995;45:903–7. McNamee R. Efficiency of two-phase designs for prevalence estimation. Int J Epidemiol 2003;32:1072–8. Prince M. Commentary: two-phase surveys. A death is announced; no flowers please. Int J Epidemiol 2003;32:1078–80. Stephan B, Brayne C. Prevalence and projections of dementia. In: Downs M, Bowers B, eds. Excellence in dementia care: research into practice. Berkshire, England: Open University Press; 2008. p. 9–34. Ellis RJ, Jan K, Kawas C, Koller WC, Lyons KE, Jeste DV, et al. Diagnostic validity of the dementia questionnaire for Alzheimer disease. Arch Neurol 1998;55:360–5. Marioni RE, Matthews FE, Brayne C, MRC Cognitive Function and Ageing Study. The association between late-life cognitive test scores

270

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

and retrospective informant interview data. Int Psychogeriatr 2011; 23:274–9. [44] Hebert LE, Scherr PA, Bienias JL, Bennett DA, Evans DA. Alzheimer disease in the US population. Prevalence estimates using the 2000 census. Arch Neurol 2003;60:1119–22. [45] Plassman BL, Langa KM, Fisher GG, Heeringa SG, Weir DR, Ofstedal MB, et al. Prevalence of dementia in the United States: the aging, demographics, and memory study. Neuroepidemiology 2007;29:125–32. [46] Kinsella K, Wan H. An Aging World: 2008. US Census Bureau, International Population Reports, P95/09-1. Washington, DC: US Government Printing Office; 2009.

[47] Ferri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, Ganguli M, et al. Global prevalence of dementia: a Delphi consensus study. Lancet 2005;366:2112–7. [48] Brookmeyer R, Johnson E, Ziegler-Graham K, Arrighi HM. Forecasting the global burden of Alzheimer’s disease. Alzheimers Dement 2007;3:186–91. [49] Brayne C, Davis D. Making Alzheimer’s and dementia research fit for populations. Lancet 2012;380:1441–3. [50] Savva GM, Wharton SB, Ince PG, Forster G, Matthews FE, Brayne C, et al. Age, neuropathology, and dementia. N Engl J Med 2009; 360:2302–9.

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Supplementary Acknowledgments The authors are very grateful to all the participants who made this investigation possible and to the “Monzino 80plus Study” Group: survey psychologists: Sonia Ammesso, Maria Caterina Bellomo, Laura Bergamin, Chiara Bertinotti, Anna Biotti, Fabiana Colombo, Elena Gargantini, Marilena Maragna, Lorenza Marino, Alessandro Piedicorcia, Sonia Scordari, Tiziana Tentorio, Barbara Viale; survey neurologists: Cristina Landi, Leonardo Sacco, Pietro Tiraboschi; genetic analyses: Diego Albani, Gianluigi Forloni; laboratory analyses: Laboratorio Milano: Giangiacomo Conrado, Adriana Ripamonti, Vito Catillo; AEU neurologists: Massimo Franceschi, Michele Perini, Daniele Porazzi, Fabio Reverberi, Patrizia Stefanoni (geriatrician), Marta Zuffi; ASL, physicians: Maria Grazia Buzzi, Carmelo Di Fazio, Antonella Scampini, Bianca Zocchi; ASL, civil servants: Ileana Ceriotti, Carmine Forese, Anna Gallazzi, Barbara Levati; mayors or town councillors: Paolo Albe, Franco Azimonti, Graziano Baschirotto, Nicola Bruno, Rosalia Chendi, Carlo Chierichetti, Antonello Colombo, Livio Frigoli, Valerio Mola, Federico Simonelli; voluntary association: Giuseppe Senaldi; GPs and nursing home geriatricians: Paola Aceti, Mauro Albertalli, Enrico Alberti, Barbara Albertoli, Mustaf
a Alchahabi, Donatella Aloigi Luzzi, Abdul Razzak Al Oum, Mohsen Anbarafshan, Roberto Andriani, Marcella Androni, Marco Angelini, Salvatore Annunziata, Patrizia Anzani, Giorgio Apollonia, Giovanni Arangio Febbo, Franceco Asquer, Roberta Azimonti, Paolo Azzimonti, Rosalia Bandera, Giuseppe Banfi, Franco Baranzini, Roberto Baratelli, Catullo Barbazza, Marina Barbazza, Maria Chiara Bareggi, Mario Barel, Fulvio Beati, Pamela Bergamo, Francesco Bertelli, Enrica Bertoni, Filippo Bianchetti, Rosario Bill
e, Giorgio Bisaccia, Michele Bisaccia, Giuseppe Blumetti, Giulio Bollini, Donata Bombelli, Donata Bordoni, Renato Borgomaneri, Ettore Bortolozzi, Paolo Bossi, Silvana Bossi, Vanna Bott
a, Franco Bottini, Marina Bove, Marco Budelli, Stefano Cambria, Carlo Campiglia, Alfredo Candiani, Feliciano Carnelli, Domenico Capaccioni, Elena Cassani, Valeria Cassani, Silvano Castiglioni, Matteo Catenazzi, Rita Catenazzi, Gianfranco Cavi, Amedeo Celora, Luca Cenacchi, Maria Cervini, Donatella Chiaravalli, Sergio Chironi, Annunziata Ciani, Giovanna Ciceri, Adelio Colombo, Cinzia Colombo, Giorgio Colombo, Walter Colombo, Anna Cordani, Marino Corio, Giovanni Battista Costantini, Tiziana Cotti, Alessandro Cova, Giovanni Cova, Angelica Cozzi, Mirto Crosta, Carmelo Curcio, Leonardo Curreri, Giuseppe D’Alicandro, Anna Pia Dalla Villa, Angela De Bernardi, Fernando De Eguia, Gialuigi De Molli, Alfredo De Nigris, Luigi Angelo Demori, Giandomenica Derisi, Annunziata Patrizia Di Fonte, Sergio Dimori, Giovanni Dodaro, Lucia Doria, Franca Dozio, Damiana Maria Dulbinska, Alberto Edalli, Adnan El Dogum, Barbara Elsi, Claudio Ercoli, Mario Daniele Etro, Livio Felloni, Paolo Ferioli, Sergio Ferioli, Chiara Ferrari, Giuseppe Ferrari, Mauro Enrico Nicola Ferrari, Enrico Ferrario, Giuseppina

270.e1

Filippini, Pier Mario Forni, Celeste Franchi, Nicola Franco, Eugenia Franzese, Marco Gaiazzi, Cesare Gallazzi (engineer), Renzo Gallazzi, Paolo Genoni, Maria Grazia Ghidini, Giorgio Ghittori, Michele Giacomino, Grazia Giardina, Angelo Giudici, Daniele Giudici, Paola Giugnoli, Giovanna Gobbi, Laura Greco, Sergio Greco, Valerio Greco, Urszula Grzelakowska, Carlo Guardascione, Fiorella Gussoni, Aldo Incarbone, Ezia Iorio, Alberto Lachi, Marina Eva Landoni, Patrizia Landoni, Franco Lattuada, Natalia Lattuada, Giorgio Lia, Paolo Licini, Claudio Locatelli, Luigi Angelo Luce, Roberta Luoni, Fabrizio Lupi, Lina Macchi, Maurizio Macchi, Enrico Maccianti, Luigi Giovanni Maggioni, Franco Magnaghi, Ermanna Mainini, Elisabetta Mairani, Roberto Manfredi, Alessandro Manzoni, Gabriele Mara, Rosalba Marazzi, Maria Angela Marcian
o, Paolo Cesare Martinenghi, Alberto Martini, Ada Marziale, Michele Marzorati, Luciana Mattaini, Chiara Mazzetti, Emilia Mazzucchelli, Paolo Mazzucchelli, Alessandra Medea, Ombretta Michetti, Alberto Antonio Milani, Sergio Miotti, Giovanni Montano, Maria Luisa Monticelli, Milvia Monzini, Edoardo Morelli, Anna Maria Moroni, Claudio Morosi, Agha Haytham Mourad, Roberto Nardi, Daniela Nasuelli, Donatella Nava, Giacomo Navarra, Maria Nerviani, Bruno Nicola, Sandro Noto, Tiziana Olgiati, Angelo Oliva, Onofrio Olivadotti, Massimo Amedeo Orsini, Claudia Osculati, Roberto Pala, Gianfranco Palamone, Gabriella Paoletti, Luigi Parassoni, Gianni Pariani, Marina Parini, Maurizio Parisi, Giorgio Parolini, Enrica Pastore, Luisa Patitucci, Paola Pellegrini, Grazia Pepe, Maria Ester Perego, Roberto Pericoli, Rosalia Peverelli, Elena Piantanida, Giovanni Pingue, Patrizia Piran, Maria Pisciotta, Paolo Gian Maria Porrini, Daniele Porro, Franco Porzio, Bruno Pozzi, Moreno Praderio, Marco Predazzi, Maria Gabriella Premoselli, Giovanni Puricelli, Gianmarco Quadrelli, Nicola Quaroni, Quirino Quisi (psychiatrist), Chiara Raineri, Francesco Rapetti, Emanuela Rastelli, Luigi Recchia, Enrichetta Rellecati, Marzia Revezzani, Orlando Vivaldo Rinaldi, Rosa Carmela Ripoli, Giuseppe Riva, Daniela Rocca, Giuseppe Ronzi, Donata Rossi, Nardo Rossi, Pier Giuseppe Rovellotti, Lanfranco Roviglio, Gabriele Russo, Massimo Sala, Silvio Salmoiraghi, Maurizio Sammarelli, Clara Sapini, Carmelo Sardo Infirri, Mariangela Scaglioni, Roberto Scalabroni, Rosanna Scancarello, Giorgio Roberto Sciarini, Giovanna Scienza, Enrica Sciocco, Eliseo Scrofani, Luigi Simonetta, Claudia Ornella Soldati, Claudio Sommaruga, Rossella Sommaruga, Massimo Speranza, Ernesto Stella, Roberto Stella, Paolo Storni, Pietro Susia, Emanuele Tacchini, Enrico Tosca, Maurizio Tovaglieri, Annarita Traini, Roberto Tramontano, Rita Triacca, Antonio Trimarchi, Enrico Turuani, Enrica Vanoli, Valerio Vassallo, Valerio Mario Vigoni, Giuseppe Villano, Daniela Zacchini, Pierangelo Zaffaroni, Antonio Zammarelli, Laura Zanzi, Giuliano Carmelo Zappal
a, Romano Zuccoli; nurses: Daria Ferrari, Pasquina Giani, Zivka Janevska, Matilde Martinoia, Luana Maria Negrisolo, Maurizio Rusca; nursing homes: ASFarM Azienda Speciale Gestione Farmacia e Servizi Sociosanitari,

270.e2

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Associazione Casa San Giorgio, Azienda di Servizi alla Persona “Il Girasole”, Casa Albergo “Villa Puricelli”, Casa Albergo “Villaggio del Fanciullo”, Casa Albergo Gilda, Casa di Riposo Calicantus, Casa di Riposo Cardinal Colombo, Casa di Riposo “Gianetti Giulio”, Casa di Riposo per Anziane “Maria Immacolata”, Casa San Gaetano, Casa San Giuseppe, Casa Sant’Agnese, Casa Santa Maria Annunciata, Fondazione Casa di Riposo “C e G Bellini”, Fondazione Casa di Riposo “D Bernacchi”, Fondazione Casa di Riposo Intercomunale FOCRIS, Fondazione Centro Assistenza Anziani “G. Moroni”, Fondazione Centro di Accoglienza per Anziani, Fondazione Centro Residenziale per Anziani “Menotti Bassani”, Fondazione Don C. Gnocchi, Fondazione G e G Ronzoni, Fondazione Giovent
u Nova

-Villa Rovera Molina, Fondazione Istituto Madonna della Croce, Fondazione Molina, Fondazione Raimondi Francesco, Fondazione Velini Casa Famiglia, Ispra Relais, Istituto Cav. F. Menotti, Istituto La Provvidenza, Residenza ai Pini, Residenza Prealpina, RSA “Fondazione A. Poretti e A Magnani”, RSA “Pineta”, RSA “San Luigi Gonzaga”, RSA Casa di Riposo Corte Cova, RSA Casa Sant’Angelo, RSA Centro Polifunzionale per la Terza Et
a, RSA Fondazione Bellora, RSA Istituto Palazzolo “Casa San Giuseppe”, RSA Mons. Gerolamo Comi, RSA Residenza al Lago, RSA Sant’Andrea, RSA Societ
a Cooperativa Sociale “Il Melo”, RSA Villa Fiammetta, SANAS Residenza Sanitario Assistenziale, “3SG” Azienda Servizi Socio Sanitari Gallarate, Villa Rovera Molina.

U. Lucca et al. / Alzheimer’s & Dementia 11 (2015) 258-270

Supplementary Fig. 1. Flow chart of the Monzino 80-plus Study.

Supplementary Fig. 2. Percentage of mild, moderate, and severe dementia in each age group in the Monzino 80-plus Study.

270.e3