The presence of behavioural and psychological symptoms

and progression to dementia in the cognitively impaired

older population
Rianne M. van der Linde
1
, Blossom C. M. Stephan
1
, Fiona E. Matthews
2
, Carol Brayne
1
and George M. Savva
1,3
on
behalf of the Medical Research Council Cognitive Function and Ageing Study
1
Department of Public Health and Primary Care, Institute of Public Health, University of Cambridge, Cambridge, UK
2
MRC Biostatistics Unit, Institute of Public Health, University of Cambridge, Cambridge, UK
3
The Irish Longitudinal Study on Ageing, Trinity College Dublin, Dublin 2, Ireland
Correspondence to: R. M. van der Linde, E-mail: [email protected]
Objective: Behavioural and psychological symptoms (BPS) are common in the older population and
may be an indication of early dementia. We explored the predictive effect of the presence of BPS on
the 2-year progression to dementia in a cognitively impaired population aged 65 years and over without
dementia at baseline.
Methods: Twelve symptoms were measured in 2024 participants without dementia at baseline as part of
a population-based longitudinal study of ageing. The risk of progression to dementia was predicted in
those with cognitive impairment for each individual BPS and using a BPS composite score.
Results: Wandering and persecution were independently associated with progression to dementia after
adjustment for socio-demographic factors, cognitive domains and other BPS. When stratifying by
cognitive function, those with low cognition (MMSE 018) and 4 or more BPS were more likely to
progress to dementia than those without BPS.
Conclusions: We have shown that some psychiatric symptoms are associated with increased short-term
progression to dementia in those with low cognition. The predictive effect of BPS in dementia progression
has implications for risk stratication of those at high risk of progression to dementia, along with memory
impairment, other cognitive impairment and health variables. Copyright #2012 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this article.
Key words: dementia; longitudinal study; population-based study; behavioural and psychological symptoms; cognitive
impairment
History: Received 25 October 2011; Accepted 18 July 2012; Published online 10 August 2012 in Wiley Online Library
(wileyonlinelibrary.com)
DOI: 10.1002/gps.3873
Introduction
Behavioural and psychological symptoms (BPS) are
common in people with dementia (Savva et al., 2009).
These symptoms include depression, anxiety, psycho-
sis, wandering, agitated behaviour and sleep disorders,
and are collectively known as the behavioural and
psychological symptoms of dementia (BPSD). BPS are
associated with cognitive and functional decline and
institutionalisation in dementia patients (Harwood
et al., 2000; Black and Almeida, 2004). In previous
cross-sectional analyses of data from the Medical
Research Council Cognitive Function and Ageing
Study (MRC CFAS), we have shown that BPS are not
only common in those with dementia but are also
found in the non-demented older population, with
most symptoms more common in cognitively impaired
sub-populations (van der Linde et al., 2010). Not all
cognitive impairment in older people leads to demen-
tia, and it is important that those whose cognitive
impairment does indicate early dementia can be identi-
ed. That BPSD are common in dementia and the fact
that psychiatric symptoms such as depression and psy-
chosis have been shown to be risk factors for dementia
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
RESEARCH ARTICLE
suggest that BPS in those with cognitive impairments
may indicate early dementia. On the other hand, cogni-
tive impairment may be secondary to an underlying
psychiatric disorder such as late-onset depression. It is
therefore unclear whether individuals meeting the cog-
nitive criteria for mild cognitive impairment (MCI) are
more likely to develop dementia if they are also suffer-
ing from BPS. Previous clinical studies of such cohorts
have reported mixed ndings (Monastero et al., 2009).
Here, we investigate the effect of the presence of BPS
on 2-year dementia incidence in a population aged
65 years and above with cognitive impairments but
without dementia at baseline, using data from the
MRC CFAS.
Methods
Study participants
The Medical Research Council Cognitive Function
and Ageing Study is a large multi-centre longitudinal
study of ageing that is fully described elsewhere
(Brayne et al., 2006). Between 1990 and 1991, a
random sample of individuals aged 65 years and older
living in ve centres representative of rural and urban
areas in England and Wales (Cambridgeshire,
Gwynedd, Newcastle, Nottingham and Oxford) were
contacted from a sampling frame of n =123 691 (The
Medical Research Council Cognitive Function and
Ageing Study (MRC CFAS), 1998). The size of the
random sample (approximately 2500 per centre) was
calculated to estimate differences in prevalence with
sufcient precision (if difference is twofold, the 95%
condence interval has 1.5 as a lower bound) and in-
cidence rates with a high degree of power (80% for a
5% signicance level). Sample selection allowed for
loss before interview of 1200 people. At baseline,
13 004 participants completed the screening interview
(response rate 80%). A subgroup of those screened
and willing to be interviewed again was selected on
the basis of the Mini Mental State Examination
(MMSE) and age, including all of those with severe
cognitive impairments (GMS-AGECAT organicity
rating case level of 3 or above), and a stratied
subsample of the remainder also completed a more
detailed assessment (n =2640; 20%). This included se-
lected items from the Geriatric Mental State Auto-
mated Geriatric Examination for Computer Assisted
Taxonomy (GMS-AGECAT) (Copeland et al., 1986),
the MMSE (Folstein et al., 1975) and Cambridge Men-
tal Disorders of the Elderly Examination (CAMDEX)
(Roth et al., 1986). The History and Aetiology
Schedule, which accompanies the GMS, was con-
ducted with caregivers for n =2197 participants.
Data from the initial prevalence screen, rst assess-
ment and 2-year follow-up interviews were used in
this analysis. Persons diagnosed with dementia at
baseline (n =587), with unknown dementia status
(n =3) or with Parkinsons disease (n =26) were
excluded. Therefore, of the 2640 who undertook the
assessment interview, 2024 persons were included.
Informed consent and condentiality
The Medical Research Council Cognitive Function
and Ageing Study has Multi-centre Research Ethics
Committees approval and ethical approval from the
relevant local research ethics committees. All partici-
pants gave informed consent, and patient condential-
ity was not breached.
Dening dementia
Dementia status was derived using the full GMS-
AGECAT diagnostic algorithm (Copeland et al., 1986).
Dementia was dened as an GMS-AGECAT organicity
rating case level of 3 or above and has been reported
as equivalent to dementia as diagnosed using Diagnostic
and Statistical Manual of Mental Disorders, Third
Edition, Revised (DSM-III-R) (American Psychiatric
Association, 1980; Copeland et al., 1990). In total, 133
persons (4.4% weighted for attrition) had a case level
diagnosis of dementia at follow-up.
Assessment of behavioural and
psychological symptoms
Behavioural and psychological symptoms assessed
included depression (dened by the GMS-AGECAT
algorithm), apathy (lack of interest in things generally
or in hobbies and previous interests; severe listlessness;
severe slowing in thinking), anxiety (suffering from
fear or panic attacks, often trembles or interviewer
observes severe bodily features of anxiety), feelings of
persecution (participant believes that someone is try-
ing to upset or harm them or that people are laughing
at them; interviewee makes false accusations or is sus-
picious or mistrusting; excluded if interviewer feels
that beliefs are likely to be true), hallucination (hear-
ing voices; visions; smells or gustatory hallucinations),
agitated behaviour (inappropriate verbal, vocal or
motor activity; observed restlessness during interview),
elation (uncontrollable bouts of laughter; infectious
701 Neuropsychiatric symptoms and dementia progression
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
gaiety; inappropriate elated mood), irritability (partici-
pant has been more angry or irritable lately, inter-
viewer observes severe hostility or angry responses),
sleep problems (difculty in getting to sleep and stay-
ing asleep at night, or problems with falling asleep dur-
ing the day), wandering (getting lost or wandering has
become a problem), confabulation (untrue responses
to questions without deliberate intention to mislead)
and misidentication (tendency to misidentify people
or objects; participant feels that special messages are
being sent on the television or radio or that their mind
or body are being controlled in other ways). Symptoms
were assessed using questions from GMS-AGECAT,
History and Aetiology Schedule and CAMDEX. Inter-
views were conducted with both study participants
and their informants, and combined with observations
made by the interviewer during the course of the
interviews. The questions used to assess each symptom
have been used previously in demented and non-
demented populations (Savva et al., 2009; van der
Linde et al., 2010), and questionnaires are available
online (http://www.cfas.ac.uk).
Assessment of cognitive function
Cognitive function was assessed in domains typically
used to make diagnoses of MCI including general
cognitive function, objective memory impairment,
subjective memory complaint and objective non-
memory impairment.
General cognitive function was assessed using the
MMSE with four groups dened including: severely
impaired (018), intermediately impaired (1922),
mildly impaired (2326) and normal or slightly
impaired (2730). Cut-points were based on previous
analyses in MRC CFAS comparing the predictive
probability of each MMSE score to detect those who
have dementia by 2 years (Huppert et al., 2005;
Matthews et al., 2008; Stephan et al., 2010). Those
who could not complete the MMSE were assumed to
be severely impaired (Folstein et al., 1975).
Objective memory impairment was dened using a
score below the 16th age-specic percentile on one or
more memory domain of the CAMDEX Cambridge
Cognitive Examination (CAMCOG), a short neuro-
psychological battery within the CAMDEX covering
seven broad areas of cognition: orientation, language,
memory (learning, recent and remote), attention and
calculation, praxis, abstract thinking and perception.
Non-memory impairment was dened using a score
of below the 16th age-specic percentile on any of the
other non-memory domains measured by CAMCOG
including orientation, language, attention/calculation,
praxis, abstract thinking or perception. CAMCOG
cut-off scores were age adjusted using ve 5-year
age groups including 6569, 7074, 7579, 8084 and
85+years.
Subjective memory complaint was dened as a
report of memory problems by either the respondent
or an informant at either the screening or the assess-
ment interview.
Assessment of socio-demographic factors and
health variables
Age was reported continuously. Level of education was
measured in number of years of education and was
dichotomised into two groups including less than
10 years and 10 or more years of education, reecting
basic versus higher education as 9 years was the statu-
tory time for this generation. Social class was coded
according to Ofce of Population Censuses and
Surveys standard occupational codes using CASOC
(Elias et al., 1993), coding each persons main occupa-
tion for most of their working life and the occupation
of their most recent husband for married, widowed,
divorced or separated women. Classes I and II include
all those who had professional and managerial
occupations; classes III to V includes class IIIN (skilled
non-manual, e.g. clerical and secretarial), class IIIM
(skilled manual, e.g. craft and related occupations
and foremen), class IV (partly skilled occupations,
e.g. street traders and scaffolders) and class V (unskilled
occupations, e.g. cleaners and farm labourers). Institu-
tionalisation was divided into two groups including
independent (living alone or in a warden controlled
at) or dependent (living in a residential home,
nursing home or hospital).
Functional disability was assessed using questions
from the Modied Townsend Disability Scale, with
three additional items (Townsend, 1979; Bond and
Carstairs, 1982). With the use of information on a
hierarchy of activities of daily living/instrumental
activities of daily living (ADL/IADL), individuals are
classied into one of three groups. The rst group
included those individuals who showed no evidence
of impairment in ADL or IADL on items including
washing, cooking hot meals, putting on shoes and
socks, completing heavy housework or shopping and
carrying heavy bags, and the individual can get around
outside. The second group included individuals with
impairments only in IADLs, including those indivi-
duals who require help on items including heave
housework or shopping and carrying heavy bags. The
702 R. M. van der Linde et al.
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
third group included those individuals with decits in
basic ADL, including individuals who require help at
least several times per week on items relating to wash-
ing, cooking and dressing or if they are house/chair
bound. Here, we have dened severe impairment as
having both ADL/IADL impairment (third group),
and we have combined the rst and second groups,
similar to the functional disability criteria used in
MCI denitions (Stephan et al., 2010).
In addition, participants were asked if they had a
history of emotional problems (Have you ever
consulted a doctor about emotional problems, or
problems with your nerves?), about their current and
previous smoking behaviour and to rate their own
health as excellent, good, fair or poor. Data were
missing in fewer than 2% of individuals for each of
these variables and missing data was assumed to be
missing at random.
Statistical analyses
All analyses were undertaken using STATA 11.0
(StataCorp, College Station, TX, USA). Differences in
socio-demographic factors and health factors between
MMSE groups assessed at baseline were described. The
prevalence of each BPS in participants was estimated
in those who progressed to dementia and in those
who did not, stratied by baseline MMSE score.
Prevalence estimates were weighted to adjust for
over-sampling in the study population of individuals
aged 75 years or older and the subsequent stratied
sampling for the assessment interview. Longitudinal
analyses were further weighted to adjust for attrition
due to drop out between waves.
The percentage of participants with MMSE 026
who progressed to dementia by presence of BPS was
plotted. The effect of each symptom on risk of
progression to dementia in each of the three cognitive
groups was tested using logistic regression. Statistical
signicance was set at a p-value of less than 0.05 or a
95% condence interval not containing 1.0. Four
models were applied for the relationship between each
symptom and progression to dementia: model 1
including only the symptom of interest; model 2
adjusting for demographic factors of age, sex and so-
cial class; model 3 additionally adjusting for baseline
cognitive function and disability; and model 4 addi-
tionally adjusting for health factors and other BPS.
In a secondary analysis, the association between
dementia progression and the number of BPS was
assessed, the number of BPS was simply counted and
four categories dened depending on total scores in-
cluding 0, 1, 23, and 4 or more BPS. Progression to
dementia in groups dened by number of BPS were
investigated with logistic regression, stratied by
general cognitive function and adjusting for socio-
demographic factors, cognitive function and health
factors as previously.
Results
Baseline socio-demographic and health characteristics
for each MMSE group are shown in Table 1. Those
with lower MMSE scores were more likely to be older,
female, less educated, having a history of psychiatric
symptoms, having a history of vascular disease, in
poor or fair health and of a lower social class. Of the
2024 participants at baseline, 210 (7.0% weighted for
Table 1 Socio-demographic and health factors in groups assessed at baseline
MMSE18N (%)
a
MMSE 1922N (%)
a
MMSE 2326N (%)
a
MMSE 27+ N (%)
a
N 298 (8.5) 405 (10.9) 713 (32.9) 608 (47.7)
Age75year 205 (54.4) 248 (57.1) 346 (43.9) 197 (29.4)
Female sex 203 (68.2) 289 (65.3) 435 (57.2) 348 (54.5)
Institutionalised 50 (11.0) 24 (4.3) 13 (0.8) 7 (0.7)
History of psychiatric disorder 68 (35.0) 99 (23.8) 207 (28.2) 189 (32.1)
History of emotional problems 38 (15.2) 76 (20.3) 154 (22.3) 123 (18.6)
History of stroke, heart attack or diabetes 93 (28.1) 122 (28.4) 172 (22.1) 112 (18.5)
More than 9years education 70 (23.0) 81 (18.9) 217 (32.9) 274 (46.6)
Regular contact with family 156 (85.1) 231 (80.6) 409 (77.5) 344 (75.2)
Poor self-rated health 140 (49.3) 181 (41.1) 247 (33.1) 168 (25.6)
ADL impairment 147 (39.9) 117 (27.8) 104 (9.9) 55 (7.7)
Lower social class 215 (64.4) 299 (73.8) 491 (67.1) 330 (50.3)
Drinking problem 21 (6.8) 18 (7.8) 41 (6.7) 48 (10.4)
Smoking 52 (18.9) 84 (21.8) 160 (23.8) 128 (21.1)
a
Percentages backweighted to normal population.
703 Neuropsychiatric symptoms and dementia progression
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
attrition) participants died during the follow-up
period, 450 (20.3% weighted for attrition) refused to
participate and 28 (1.2% weighted for attrition) could
not be contacted because they had moved. Therefore,
1336 persons were included in the follow-up study,
of whom 133 (4.4% weighted for attrition) persons
developed dementia during the follow-up period.
Table 2 shows the prevalence of each BPS stratied
by baseline MMSE and by whether or not the partici-
pant was diagnosed with dementia at follow-up. For
most symptoms, the prevalence at baseline was higher
in those who progressed to dementia than in those
who did not, although this could not be seen for MMSE
score 1922. In order to focus on the population with
cognitive impairment and because of a striking lack of
BPS in the high functioning group (MMSE27) as well
as a low progression to dementia, this group was
excluded from further analyses (n =608).
Figure 1 shows the percentage of participants with
MMSE6 who progressed to dementia stratied by the
Table 2 Prevalence of BPS in those who progressed to dementia, compared with those who did not, stratied by MMSE score
Progressed to dementia
MMSE
18 1922 2326 27+ Total
BPS n (%)
a
Depression 11 (21.6) 7 (14.7) 2 (3.9) 0 (0) 20 (11.5)
Apathy 17 (28.5) 7 (17.5) 9 (26.2) 1 (18.4) 34 (24.0)
Agitation 5 (9.1) 0 (0) 1 (1.8) 0 (0) 6 (3.2)
Persecution 16 (27.4) 8 (15.1) 9 (21.8) 0 (0) 33 (20.4)
Anxiety 5 (9.8) 2 (4.2) 2 (11.8) 0 (0) 9 (8.6)
Hallucination 9 (16.6) 3 (5.0) 3 (13.8) 0 (0) 15 (11.4)
Elation 5 (8.4) 2 (4.2) 1 (1.7) 0 (0) 8 (4.1)
Irritability 11 (17.9) 4 (7.4) 11 (33.7) 2 (30.4) 28 (21.9)
Sleep problems 24 (39.1) 15 (37.5) 16 (36.4) 1 (18.4) 56 (36.6)
Wandering 4 (5.5) 0 (0) 3 (4.6) 0 (0) 7 (3.3)
Confabulation 0 (0) 1 (1.8) 1 (0.7) 0 (0) 2 (0.8)
Misidentification 4 (8.2) 2 (3.9) 1 (1.7) 1 (33.2) 8 (5.5)
Not progressed to dementia
MMSE
18 1922 2326 27+ Total
BPS n (%)
a
Depression 16 (8.7) 39 (16.0) 44 (9.3) 34 (6.9) 133 (8.6)
Apathy 17 (11.1) 53 (22.8) 56 (12.1) 39 (7.9) 165 (10.7)
Agitation 8 (6.9) 17 (7.2) 20 (3.3) 14 (2.7) 59 (3.5)
Persecution 12 (6.4) 30 (16.0) 35 (6.7) 31 (6.1) 108 (7.2)
Anxiety 6 (8.5) 15 (7.0) 21 (4.8) 35 (8.1) 77 (7.0)
Hallucination 11 (5.0) 14 (5.0) 15 (2.8) 19 (4.2) 59 (3.8)
Elation 2 (1.0) 12 (4.1) 13 (3.1) 18 (3.5) 45 (3.3)
Irritability 19 (17.4) 46 (18.3) 56 (13.6) 54 (10.4) 175 (12.5)
Sleep problems 53 (52.9) 102 (51.4) 204 (42.4) 197 (42.6) 556 (43.9)
Wandering 0 (0) 1 (0.4) 0 (0) 3 (0.4) 4 (0.3)
Confabulation 2 (3.4) 1 (0.4) 2 (0.3) 1 (0.0) 6 (0.4)
Misidentification 2 (1.1) 12 (5.8) 18 (4.2) 9 (2.5) 41 (3.2)
Lost to follow up
MMSE
18 1922 2326 27+ Total
n (%)
a
Participated, progressed to dementia 47 (9.7) 40 (8.1) 41 (3.8) 5 (0.3) 133 (3.1)
Participated, not progressed to dementia 105 (48.3) 213 (56.2) 433 (65.5) 452 (76.7) 1203 (68.4)
Died 68 (18.1) 47 (8.9) 69 (7.0) 26 (4.7) 210 (7.0)
Refused 73 (23.0) 101 (26.2) 161 (22.5) 115 (16.9) 450 (20.3)
Moved 5 (0.9) 4 (0.6) 9 (1.2) 10 (1.4) 28 (1.2)
Total 298 405 713 608 2024
a
Percentages backweighted to normal population.
704 R. M. van der Linde et al.
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
presence of each BPS. In univariate analyses, apathy
(OR=1.7, 95% CI 1.12.7), persecution (OR=3.0,
95% CI 1.94.8), hallucination (OR=2.4, 95% CI
1.34.4) and wandering (OR=43.4, 95% CI 5.3
355.8) were associated with 2-year progression to
dementia (Table 3). Adjustment for age, sex, education
and social status had little effect, whereas adjustment for
cognitive function attenuated the relation between BPS
and dementia progression. After full adjustment for
baseline cognition and disability, socio-demographic
factors, health variables and other BPS, only wandering
(OR=30.5, 95% CI 3.1298.3) and feelings of persecu-
tion (OR=2.3, 95% CI 1.34.2) were signicantly
associated with dementia progression.
Figure 1 Percentage of participants with MMSE 026 who progressed to dementia by presence of BPS.
Table 3 Odds of progression to dementia for BPS adjusted for socio-demographic factors, health factors, MCI criteria and other BPS in those with
MMSE score 26
Symptom
present
Symptom
not
present Model 1 Model 2 Model 3 Model 4
BPS N
N
dem N
N
dem OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Depression 119 20 760 108 1.2 (0.72.1) 1.3 (0.82.3) 1.0 (0.61.9) 1.2 (0.62.5)
Apathy 159 33 720 95 1.7 (1.12.7) 1.5 (1.02.4) 1.2 (0.71.9) 1.0 (0.61.8)
Agitation 51 6 828 122 0.8 (0.31.8) 1.0 (0.42.4) 0.7 (0.31.8) 0.7 (0.22.1)
Persecution 110 33 769 95 3.0 (1.94.8) 3.2 (2.05.2) 2.5 (1.54.2) 2.3 (1.34.2)
Anxiety 51 9 828 119 1.3 (0.62.7) 1.2 (0.62.7) 0.9 (0.42.1) 1.0 (0.42.5)
Hallucination 55 15 824 113 2.4 (1.34.4) 2.3 (1.24.4) 1.4 (0.72.9) 0.9 (0.42.2)
Elation 35 8 844 120 1.8 (0.84.0) 2.0 (0.94.7) 1.7 (0.74.2) 1.0 (0.33.0)
Irritability 147 26 732 102 1.3 (0.82.1) 1.3 (0.82.1) 1.0 (0.61.8) 0.9 (0.51.6)
Sleep problems 414 55 465 73 0.9 (0.51.5) 0.9 (0.51.6) 0.8 (0.51.5) 0.8 (0.51.2)
Wandering 8 7 871 121 43.4 (5.3355.8) 45.9 (5.4392.4) 44.6 (4.6430.3) 30.5 (3.1298.3)
Confabulation 7 2 872 126 2.4 (0.512.3) 1.7 (0.39.1) 1.1 (0.26.2) 0.6 (0.06.2)
Misidentification 39 7 840 121 1.3 (0.63.0) 1.3 (0.63.2) 1.1 (0.42.9) 1.2 (0.43.9)
Model 1 Unadjusted.
Model 2 Adjusted for age, sex, education and social class.
Model 3 Adjusted for age, sex, education, social class, MMSE score, subjective memory complaint, objective memory impairment and impairment of
activity of daily living.
Model 4 Adjusted for age, sex, education, social class, MMSE score, subjective memory complaint, objective memory impairment, impairment of
activity of daily living, other BPS, institutionalisation, history of emotional problems, history of stroke, heart attack or diabetes, self-rated health,
smoking history.
705 Neuropsychiatric symptoms and dementia progression
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
Figure 2 shows the percentage of those who
progressed to dementia by the number of BPS (0, 1,
23, and 4 or more), stratied by MMSE score.
The effect of BPS on dementia progression varied
across cognitive groups. Among those with very low
cognition (MMSE 18) 67% of participants with 4
or more BPS progressed to dementia, whereas in those
with 0, 1 or 23 symptoms, this was only 17%, 13%
and 14%, respectively. After adjusting for socio-demo-
graphic factors, health factors, MCI component
criteria (i.e. functional ability and subjective memory
complaint) and other BPSD, those with low cognition
and 4 or more BPS remained signicantly more likely
to progress to dementia than those without symptoms
(OR=15.4, 95% CI 1.7136.3) although no relation
was seen for those with fewer than 4 BPS (Table 4).
In those with MMSE 1922, BPS were generally
associated with lower risk of progression (23 vs 0
BPS OR=0.4, 95% CI 0.11.0). In those with MMSE
2326, having 2 or 3 symptoms was associated with
progression to dementia (OR=2.9, 95% CI 1.17.8)
although no association was seen in other groups
including those with 1 symptom and those with more
than 4 symptoms.
When stratifying the odds of progression to demen-
tia for individual BPS by MMSE score, similar patterns
to those aforementioned were observed, although
numbers are too small for meaningful statistical
comparison (see supplementary tables).
Figure 2 Percentage of participants with MMSE 026 who progressed
to dementia by number of BPS stratied by MMSE score.
Table 4 Odds of progression to dementia for number of BPS adjusted for socio-demographic factors, health factors, MCI criteria and other BPS in
those with MMSE score 026
Model 1 Model 2 Model 3 Model 4
N N (%)
a
dem OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
MMSE18
Number of BPS
0 34 7 (17.3) Ref. Ref. Ref. Ref.
1 48 15 (13.3) 1.8 (0.64.9) 1.6 (0.64.8) 1.1 (0.33.7) 0.9 (0.23.2)
23 56 15 (13.6) 1.4 (0.53.9) 1.3 (0.43.6) 1.0 (0.33.7) 1.0 (0.34.0)
4 14 10 (67.9) 9.6 (2.340.2) 9.4 (2.241.0) 7.4 (1.245.4) 15.4 (1.7136.3)
MMSE 1922
Number of BPS
0 67 15 (19.0) Ref. Ref. Ref. Ref.
1 80 12 (12.3) 0.6 (0.31.4) 0.6 (0.31.5) 0.5 (0.21.4) 0.4 (0.21.2)
23 84 10 (9.1) 0.5 (0.21.1) 0.5 (0.21.2) 0.4 (0.11.1) 0.4 (0.11.0)
4 22 3 (7.7) 0.5 (0.12.1) 0.5 (0.12.2) 0.3 (0.11.6) 0.5 (0.12.6)
MMSE 2326
Number of BPS
0 166 9 (1.9) Ref. Ref. Ref. Ref.
1 167 13 (6.0) 1.5 (0.63.5) 1.8 (0.74.3) 1.5 (0.63.9) 1.6 (0.64.3)
23 114 18 (11.7) 3.3 (1.47.6) 3.5 (1.58.3) 2.5 (1.06.6) 2.9 (1.17.8)
4 27 1 (2.1) 0.7 (0.15.5) 0.8 (0.16.7) 0.6 (0.15.2) 0.6 (0.16.7)
Model 1 Unadjusted.
Model 2 Adjusted for age, sex, education and social class.
Model 3 Adjusted for age, sex, education, social class, MMSE score, subjective memory complaint, objective memory impairment and impairment of
activity of daily living.
Model 4 Adjusted for age, sex, education, social class, MMSE score, subjective memory complaint, objective memory impairment, impairment
of activity of daily living, institutionalisation, history of emotional problems, history of stroke, heart attack or diabetes, self-rated health,
smoking history.
a
Percentages backweighted to normal population.
706 R. M. van der Linde et al.
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
Discussion
In this large, representative study of the cognitively
impaired population, we have reported the association
between BPS and 2-year progression to dementia.
Wandering was strongly associated with progression
to dementia, although wandering at baseline was only
reported for eight participants included in our study,
of which seven received a study diagnosis of dementia
at 2 years. Wandering was ascertained through the
proxy interview by the question Has wandering or
getting lost been a problem for him/her? and there-
fore covers two potentially separate phenomena that
we are unable to further distinguish. It is possible that
those who reported wandering at baseline were
misclassied by the AGECAT dementia algorithm
employed in our study. However, the MMSE scores
for those with problematic wandering were not
substantially lower than those with other symptoms,
and the association with incident dementia remained
after adjusting for baseline cognitive function includ-
ing MMSE and other components of MCI criteria.
Excluding those with an MMSE of 18 or lower did
not change the results for persecution (OR=2.0;
95%CI 0.94.2) and wandering (OR=26.4; 95%CI
2.0340.0). Presence of wandering in dementia has
been shown to be associated with cognitive and
functional decline, disease duration and institutionali-
sation (McShane et al., 1998; Holtzer et al., 2003;
Scarmeas et al., 2007). It is likely therefore that
wandering or getting lost in individuals with cognitive
impairments does suggest an increased risk of a subse-
quent dementia diagnosis.
Feelings of persecution at baseline were also associ-
ated with dementia at follow-up. Persecution is not
typically specically included in assessments of BPSD
(e.g. the Neuropsychiatric Inventory). In the present
study, persecutory delusions were determined through
any positive response to any one of four questions to
the respondent and their informant. The respondent
was asked if they feel people are laughing at you or
Is anyone trying deliberately to annoy you or harm
you. The informant was asked Does he or she ever
wrongly accuse you of things? and Is there a tendency
to be more suspicious or mistrusting? A positive
response was excluded if the interviewer felt that the
beliefs of persecution were likely to be true. Our
results suggest that this specic psychiatric symptom
should be included in future BPSD research and
clinical assessments as well as more general delusional
ideation.
When stratifying by baseline MMSE, we found that
BPS were most strongly associated with progression to
dementia in those with poorest cognitive function. In
our study, dementia was diagnosed as an AGECAT
organicity rating case level of 3 or above, which has
been previously validated against a DSM-III-R clinical
diagnosis (Copeland et al., 1990). We have previously
reported that relatively few people classied as
demented using this algorithm have MMSE scores
above 18 (The Medical Research Council Cognitive
Function and Ageing Study (MRC CFAS), 1998).
However, there were a substantial number of people
with very low MMSE (18) who were not given a
study diagnosis of dementia, and it is it is possible that
individuals in this category may have been misclassi-
ed. Nevertheless, our ndings show that in this group
of individuals with very poor cognitive function but
not meeting the criteria for a diagnosis of dementia,
a high number of BPS indicate a very high chance of
receiving a dementia diagnosis within 2 years.
In those with MMSE score between 19 and 22,
there was some evidence that those with BPS were less
likely to progress to dementia. It is possible that in
some members of this group, cognitive impairment
was to some extent secondary to the presence of a
psychiatric symptom and therefore did not indicate
early dementia. However, a signicant proportion of
this group did receive a dementia diagnosis at 2-year
follow up suggesting that those with BPS do still
remain at risk of dementia progression.
Our results should be considered in the light of
the strengths and weaknesses of the study. We used
data from a population-representative sample and
took into account a wide range of possible confoun-
ders. The initial response rate to the study was high
(80%) (The Medical Research Council Cognitive
Function and Ageing Study (MRC CFAS), 1998),
and follow-up data was available for 74% of the sur-
viving baseline sample. We have previously shown
that loss to follow up through refusal to participate
is associated with poor cognitive ability and fewer
years of education (Matthews et al., 2004). Weights
were applied to adjust these for baseline determi-
nants of attrition, and these factors were included
in multivariable analyses. This analysis was restricted
to a 2-year follow-up, but this period is of interest
in those with MCI, and previous longitudinal stud-
ies have concluded that the risk of progression to
dementia in those with MCI appears to diminish
with time (Mitchell and Shiri-Feshki, 2009). How-
ever, it is also possible that certain symptoms, for
example, depression and psychosis, are independent
risk factors for cognitive decline and dementia over
longer periods, and this is the subject of further
planned investigation.
707 Neuropsychiatric symptoms and dementia progression
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
Associations between BPS and dementia incidence
in patients with MCI have been previously reported al-
though associations with specic symptoms have been
mixed (Copeland et al., 2003; Palmer et al., 2007;
Edwards et al., 2009). Previous studies have typically
been based on smaller clinical samples and report a
higher prevalence of BPS and a higher progression rate
to dementia than was seen in our study (Matthews
et al., 2008). Edwards et al. followed 270 patients from
memory disorder clinics diagnosed with MCI for a
mean period of 1.6 years (Edwards et al., 2009). After
adjustment for age, MMSE score, number of comor-
bidities and Blessed Roth score, those with at least 4
BPS had a 2.4 times higher risk of progression to
dementia compared with those with less than four
symptoms. Copeland et al. recruited 165 participants
through print media with a Clinical Dementia Rating
score of 0.5 (Copeland et al., 2003). The 22 subjects
that progressed to dementia during the 3-year follow-
up period were more likely to have symptoms of
personality change, such as agitation and passivity,
and mild depressive symptoms at baseline than those
who did not progress to dementia. Palmer et al.
compared 3-year progression to dementia in 47
persons with MCI from a population-based sample.
Mood-related depressive symptoms were high in both
persons with MCI who progressed to dementia and
those who did not, but anxiety symptoms predicted
dementia in persons with MCI (Palmer et al., 2007).
Our assessment of symptoms is study specic, and
although it has been previously used to report the
prevalence and correlates of BPSD and MCI (van der
Linde et al., 2010), it is difcult to directly compare
our results with previous studies.
In many situations individuals with depression or
other psychiatric conditions are excluded from a diag-
nosis of MCI (Salloway et al., 2004; Visser et al., 2005;
Doody et al., 2009). We have shown that having at
least 4 BPS is associated with increased short-term
progression to dementia in those with very low cogni-
tion. Given that the association was unchanged after
controlling for subjective and objective evidence of
cognitive impairment at baseline, it is unlikely that
the BPS were a reaction to loss of cognitive function
at baseline. These ndings along with a higher preva-
lence of BPS among the more cognitively impaired
(van der Linde et al., 2010) are consistent with the hy-
pothesis that BPS may be a representation of early de-
mentia, and cognitive and non-cognitive symptoms of
dementia may have similar causal pathways. For ex-
ample, studies have suggested associations between
BPS and the neurobiological degeneration aspect of
dementia (Shinosaki et al., 2000).
Conclusion
We have shown that BPS, in particular wandering and
persecution, are associated with increased short-term
progression to dementia in those with low cognition.
This has implications for predicting individuals at high
risk of progression to dementia, which may enable early
treatment where available and could give patients and
their families the opportunity to make preparations for
the future. Many of the population with cognitive
impairments also have BPS, and these individuals
should not be excluded froma diagnosis of MCI or from
inclusion into trials aimed at dementia prevention. Fur-
ther, non-cognitive symptoms currently offer greater
opportunities for intervention and management than
do cognitive symptoms. To date, studies in this area
have been limited and results have been mixed. Other
population-representative longitudinal studies exist and
have generated datasets that could be examined to shed
further light on the role of BPS in early dementia. Spe-
cic questions for future research are how non-cognitive
symptoms can be used alongside memory, other cogni-
tive impairment and health variables to identify those
individuals at increased risk of dementia incidence in
the general population; how different assessments of
BPS affect reported relationships; and what are the
mechanisms by which these relationships occur.
Key points

Behavioural and psychological symptoms

(BPS) are common in the older population
and may be an indication of early dementia.

Wandering and persecution are independently

associated with progression to dementia after
adjustment for socio-demographic factors,
cognitive domains and other BPS.

Those with low cognition (MMSE 018) and

4 or more BPS are more likely to progress to
dementia than those without BPS.
Acknowledgements
We would like to thank all participants and investigators
of the MRC Cognitive Function and Ageing Study. The
Medical Research Council Cognitive Function and
Ageing Study is funded by the Medical Research
Council (grant number G9901400). BCMS is funded
by the Joint European Post-Doctoral Programme:
The European Research Area in Ageing (ERA-AGE)
Network FLARE Programme. R. L. received a stipend
from Alzheimer Nederland.
708 R. M. van der Linde et al.
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.
Conict of interest
None of the sponsors inuenced the design or conduct of
the study or the analysis or interpretation of the ndings.
References
American Psychiatric Association. 1980. Diagnostic and Statistical Manual of Mental
Disorders (DSM III). American Psychiatric Association: Washington, DC.
Black W, Almeida OP. 2004. A systematic review of the association between the
behavioral and psychological symptoms of dementia and burden of care. Int
Psychogeriatr 16(3): 295315.
Bond J, Carstairs V. 1982. 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. Scottish Home and Health Department: Edinburgh.
Brayne C, McCracken C, Matthews FE. 2006. Cohort prole: the Medical Research
Council Cognitive Function and Ageing Study (CFAS). Int J Epidemiol 35(5):
11401145.
Copeland MP, Daly E, Hines V, et al. 2003. Psychiatric symptomatology and prodro-
mal Alzheimers disease. Alzheimer Dis Assoc Disord 17(1): 18.
Copeland JR, Dewey ME, Grifths-Jones HM. 1986. A computerized psychiatric
diagnostic system and case nomenclature for elderly subjects: GMS and AGECAT.
Psychol Med 16(1): 8999.
Copeland JRM, Dewey ME, Grifths-Jones HM. 1990. Dementia and depression in
elderly persons: AGECAT compared with DSM III and pervasive illness. Int J
Geriatr Psychiatry 5(1): 4751.
Doody RS, Ferris SH, Salloway S, et al. 2009. Donepezil treatment of patients with MCI:
a 48-week randomized, placebo-controlled trial. Neurology 72(18): 15551561.
Edwards ER, Spira AP, Barnes DE, Yaffe K. 2009. Neuropsychiatric symptoms in mild
cognitive impairment: differences by subtype and progression to dementia. Int J
Geriatr Psychiatry 24(7): 716722.
Elias P, Halstead K, Prady K. 1993. Computer Assisted Standard Occupational
Coding. HMSO: London.
Folstein MF, Folstein SE, McHugh PR. 1975. Mini-mental state. A practical method
for grading the cognitive state of patients for the clinician. J Psychiatr Res 12(3):
189198.
Harwood DG, Barker WW, Ownby RL, Duara R. 2000. Relationship of behavioral
and psychological symptoms to cognitive impairment and functional status in
Alzheimers disease. Int J Geriatr Psychiatry 15(5): 393400.
Holtzer R, Tang MX, Devanand DP, et al. 2003. Psychopathological features in
Alzheimers disease: course and relationship with cognitive status. J Am Geriatr
Soc 51(7): 953960.
Huppert FA, Cabelli ST, Matthews FE. 2005. Brief cognitive assessment in a UK
population sampledistributional properties and the relationship between the
MMSE and an extended mental state examination. BMC Geriatr 5: 7.
van der Linde R, Stephan BC, Matthews FE, Brayne C, Savva GM. 2010. Behavioural
and psychological symptoms in the older population without dementia
relationship with socio-demographics, health and cognition. BMC Geriatr 10(1): 87.
Matthews FE, Chateld M, Freeman C, McCracken C, Brayne C. 2004. Attrition and
bias in the MRC cognitive function and ageing study: an epidemiological investiga-
tion. BMC Public Health 4: 12.
Matthews FE, Stephan BC, McKeith IG, Bond J, Brayne C. 2008. Two-year progres-
sion from mild cognitive impairment to dementia: to what extent do different
denitions agree? J Am Geriatr Soc 56(8): 14241433.
McShane R, Gedling K, Keene J, et al. 1998. Getting lost in dementia: a longitudinal
study of a behavioral symptom. Int Psychogeriatr 10(3): 253260.
Mitchell AJ, Shiri-Feshki M. 2009. Rate of progression of mild cognitive impairment
to dementiameta-analysis of 41 robust inception cohort studies. Acta Psychiatr
Scand 119(4): 252265.
Monastero R, Mangialasche F, Camarda C, Ercolani S, Camarda R. 2009. A systematic
review of neuropsychiatric symptoms in mild cognitive impairment. J Alzheimers
Dis 18(1): 1130.
Palmer K, Berger AK, Monastero R, et al. 2007. Predictors of progression from mild
cognitive impairment to Alzheimer disease. Neurology 68(19): 15961602.
Roth M, Tym E, Mountjoy CQ, et al. 1986. CAMDEX. A standardised instrument for
the diagnosis of mental disorder in the elderly with special reference to the early
detection of dementia. Br J Psychiatry 149: 698709.
Salloway S, Ferris S, Kluger A, et al. 2004. Efcacy of donepezil in mild cognitive
impairment: a randomized placebo-controlled trial. Neurology 63(4): 651657.
Savva GM, Zaccai J, Matthews FE, et al. 2009. Prevalence, correlates and course of
behavioural and psychological symptoms of dementia in the population. Br J
Psychiatry 194(3): 212219.
Scarmeas N, Brandt J, Blacker D, et al. 2007. Disruptive behavior as a predictor in
Alzheimer disease. Arch Neurol 64(12): 17551761.
Shinosaki K, Nishikawa T, Takeda M. 2000. Neurobiological basis of behavioral and
psychological symptoms in dementia of the Alzheimer type. Psychiatry Clin
Neurosci 54(6): 611620.
Stephan BC, Savva GM, Brayne C, et al. 2010. Optimizing mild cognitive impairment
for discriminating dementia risk in the general older population. Am J Geriatr
Psychiatry 18(8): 662673.
The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS).
1998. Cognitive function and dementia in six areas of England and Wales: the
distribution of MMSE and prevalence of GMS organicity level in the MRC CFAS
Study. Psychol Med 28(2): 319335.
Townsend P. 1979. Poverty in the United Kingdom: a Survey. Pelican: Harmondsworth, UK.
Visser PJ, Scheltens P, Verhey FR. 2005. Do MCI criteria in drug trials accurately
identify subjects with predementia Alzheimers disease? J Neurol Neurosurg
Psychiatry 76(10): 13481354.
709 Neuropsychiatric symptoms and dementia progression
Copyright #2012 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2013; 28: 700709.