Mild Cognitive Impairment: Epidemiology and Dementia Risk in
an Elderly Italian Population
Giovanni Ravaglia, MD,?Paola Forti, MD,?Fausta Montesi, MD,?Anna Lucicesare, MD,?
Nicoletta Pisacane, MD,?Elisa Rietti, MD,?Edoardo Dalmonte, MD,?Marisa Bianchin, MD,w
and Patrizia Mecocci, MDz
OBJECTIVES: To investigate prevalence and incidence of
mild cognitive impairment (MCI) andits risk of progression
to dementia in an elderly Italian population.
SETTING: Population-based cohort aged 65 and older
resident in an Italian municipality.
PARTICIPANTS: A total of 1,016 subjects underwent
baseline evaluation in 1999/2000. In 2003/04, information
about cognitive outcome was collected for 745 participants
who were free of dementia at baseline.
MEASUREMENTS: MCI (classified as with or without
impairment of the memory domain), dementia, Alzheimer’s
dementia (AD), and vascular dementia (VaD) diagnosed
according to current international criteria.
RESULTS: Overall prevalence of MCI was 7.7% (95%
confidence interval (CI)56.1–9.7 %) and was greater with
older age and poor education. During 4 years of follow-up,
155 incident MCI cases were diagnosed, with an incidence
rate of 76.8 (95% CI566.8–88.4) per 1,000 person-years.
Approximately half of prevalent and incident MCI cases
had memory impairment. Compared with normal cognition,
multivariable-adjusted risk for progression from MCI with
memory impairment to dementia was 4.78 (95% CI52.78–
8.07) for any dementia, 5.92 (95% CI53.20–10.91) for
AD, and 1.61 (95% CI50.37–7.00) for VaD. No associa-
tion with dementia risk was found for MCI without memory
impairment. Approximately one-third of MCI cases with
memory impairment did not progress to dementia.
CONCLUSION: MCI occurs often in this elderly Italian
cohort and is associated with greater risk of AD, but only
when the impairment involves the memory domain. How-
ever, a substantial proportion of MCI cases with memory
impairment do not progress to dementia. J Am Geriatr Soc
Key words: mildcognitiveimpairment;elderly;epidemio-
logy; Alzheimer’s dementia
logical testing in the absence of clinically overt dementia.1
This condition is of interest for identifying prodromal stag-
es of Alzheimer’s disease (AD) and other dementias of older
age. Indeed, in series from memory clinics, more than half
of MCI cases progress to dementia within 5 years.1More-
over, elderly persons with MCI are at greater risk of other
negative outcomes such as death and institutionalization.2
Therefore, knowledge of MCI epidemiology at the popu-
lation level is of importance.
Italy is currently has one of the oldest populations in
the world, with approximately 18% of the total population
aged 65 and older.3However, only two population-based
studies have investigated prevalence of MCI in the Italian
population,4,5and only one had a longitudinal component
allowing estimation of MCI incidence and rates of progres-
sion to dementia.6
In the present study, MCI prevalence, incidence, and
dementia progression rates were investigated in an elderly
population-based Italian cohort with 4 years of follow-up.
ild cognitive impairment (MCI) is a condition of ob-
jective cognitive impairment based on neuropsycho-
Analyses were undertaken as a part of the Conselice Study
of Brain Ageing (CSBA), a population-based study of el-
derly Italian individuals for investigation of cognitive im-
pairment epidemiology and risk factors. CSBA design and
methods have been described in detail elsewhere.7,8The
institutional review board of the Department of Internal
Medicine, Cardioangiology, and Hepatology, University of
Address correspondence to Giovanni Ravaglia, MD, Department of Internal
Medicine, Cardioangiology, and Hepatology, University Hospital S. Orsola-
Malpighi, Via Massarenti 9, 40138 Bologna, Italy. E-mail: ravaglia@med.
From the?Department of Internal Medicine, Cardioangiology, and Hepato-
logy; University Hospital S. Orsola-Malpighi, Bologna, Italy;wHealth District
of Lugo,Local Health Unit, Ravenna,Italy; andzInstituteof Gerontologyand
Geriatrics, Perugia University Hospital, Perugia, Italy.
r 2007, Copyright the Authors
Journal compilation r 2008, The American Geriatrics Society
Bologna, approved the study; written informed consent was
obtained from all participants.
Briefly, in 1999/2000, 1,016 (75%) of the 1,353 indi-
viduals aged 65 and older residing in the Italian munici-
pality of Conselice (province of Ravenna, Emilia Romagna
region) participated in the prevalence study. Since the
1950s, Conselice has been a wealthy urban area; its econ-
omy is based mainly on industry and handicraft, but the
older inhabitants were raised and lived in a rural environ-
ment most of their youth. The occupation most often held
in this cohort was farmer or housewife (47.8%), followed
by blue collar (34.3%) and white collar (17.9%). Thirty-
three of the 34 subjects of the study population known to
live in nursing homes participated in the prevalence study.
A two-phase procedure was used during 1999/2000, con-
sisting of a screening phase and an extensive clinical as-
sessment of those positive at screening to confirm a
diagnosis of MCI or dementia. The screening phase includ-
ed a standardized personal interview for collection of data
on sociodemographic characteristics, lifestyle, medical his-
tory, and functional status (activities of daily living9and
instrumental activities of daily living10), along with cogni-
tive screening using the Italian version of the Mini-Mental
State Examination (MMSE);11a standardized medical and
neurological examination; and a routine biochemical blood
and urine analysis along with genotyping for apolipopro-
tein E (APOE) epsilon 4 allele.12Previous medical records
were also reviewed. If a subject was unable to answer be-
cause of physical or mental impairments, information was
obtained from relatives and general practitioners. Subjects
with an MMSE score less than 24 were considered to be
positive at screening. To give people with different ages and
levels of education an equal probability of being detected as
a case, MMSE scores were adjusted according to standard-
ized age- and education-specific coefficients previously val-
idated in a rural Italian population.13Participants positive
at screening underwent detailed neuropsychological testing
with the Mental Deterioration Battery (MDB).14The MDB
includes tests for evaluation of memory (immediate and
delayed recall of Rey’s 15 words), language (sentence con-
struction), frontal function (phonological word fluency),
abstract reasoning (Raven’s 47 progressive colored matri-
ces), and visuospatial abilities (freehand copying of draw-
ings and copying of drawings with landmarks). Memory
was additionally tested using the prose memory test.15All
tests are provided with standardized thresholds for the defi-
nition of impairment in the corresponding domain (score
?1.5 standard deviations (SDs) below the mean for a ref-
erence adult population-based Italian cohort14,15), and age-
and education-specific coefficients to be applied to the sub-
ject’s raw score before comparison with the corresponding
Subjects with an MMSE score less than 10 did not re-
ceive further neuropsychological testing. Whenever recent
neuroradiological data were not available, the subject was
scheduled for a noncontrast computed tomography brain
scan. Standardized information about functional and men-
tal status of these subjects was also obtained from a col-
lateral informant (a relative or any other person with
reliable knowledge of the individual, including the subject’s
According to the latest international consensus crite-
ria,16MCI was defined as age- and education-adjusted
score 1.5 SDs or fewer below the reference threshold on any
of the tests used for detailed neuropsychological testing; no
evidence of functional dependency, defined as need for su-
pervision or external help to perform any of the activities of
daily living9and instrumental activities of daily living10
(functional dependency due to physical impairment was not
considered); and absence of clinical criteria for dementia.
For incident MCI cases, decline in any neuropsychological
test was deemed relevant only if score at follow-up assess-
ment crossed the established threshold for impairment; cur-
rent age at the time of follow-up assessment was used for
Differently from the consensus criteria, confirmation of
subjective memory complaint was not required, because the
application of this criterion has been shown to be ques-
tionable in population-based investigations.17According to
neuropsychological testing results, MCI was further cate-
gorized into two subtypes: MCI with memory impairment
(m1MCI), if there was objective impairment in memory
(alone or with other cognitive domains), and MCI without
memory impairment (m?MCI), if objective impairment
was restricted to one or more nonmemory cognitive do-
mains. The consensus criteria16distinguish m1MCI with
memory impairment only (amnestic MCI, aMCI) from
m1MCI with impairment of memory plus other cognitive
domains, but there were too few aMCI cases in this cohort
for separate analyses, and they were considered together
with the other m1MCI subjects.
Dementia was defined using the Diagnostic and Statis-
tical Manual of Mental Disorders, Fourth Revision, clinical
criteria.18AD was diagnosed according to the National In-
stitute of Neurological and Communication Disorders and
Stroke/Alzheimer’s Disease and Related Disorders Associ-
ation criteria.19Vascular dementia (VaD) was diagnosed
according to the National Institute of Neurological Disor-
ders and Stroke/ Association Internationale pour la Re-
cherche ´ et l’Enseignement en Neurosciences criteria.20
Probable and possible cases were grouped together.
Subjects affected by majorsensory-motor deficits or any
psychiatric condition other than dementia deemed to ham-
per a reliable cognitive assessment and for whom it could
not be ascertained whether there had been a decline from a
previously higher functioning level were diagnosed as cog-
nitively unclassifiable and excluded from the present anal-
Two physicians independently made diagnoses based
on all available information. A final diagnostic conference
with a third senior physician was held to discuss each case.
Identification and differential diagnosis of MCI and
dementia incident cases in 2003/04 followed the same pro-
cedures used to identify prevalent cases. A fall in cognitive
test scores at follow-up was not considered relevant unless
they crossed the threshold used for identification of prev-
alent cases in participants of similar age and education. Five
incident dementia cases in subjects with adjusted MMSE
scores of 24 or greater were identified because the patients
had already received a diagnosis of mild dementia in the
Cronos Project, a program founded after 2000 by the Italian
RAVAGLIA ET AL.
JANUARY 2008–VOL. 56, NO. 1 JAGS
Ministry of Public Health for promoting diagnosis and care
of AD. Additionally, for 141 participants who did not un-
dergo follow-up examination, adequate information for es-
tablishingor excludinga diagnosisofdementiawas gathered
from relatives, general practitioners, and whenever possible,
an informal interview with the subjects themselves.
A flow chart detailing the derivation of the sample used
in this study is reported in Figure 1.
Student t test and chi-square test were used for comparison
between groups. Age-, sex-, and education-specific preva-
lence rates of MCI and its subtypes were calculated as cases
per 100 population on baseline participants after exclusion
of prevalent dementia cases.21
The incidence cohort included all participants without
MCI at baseline who underwent cognitive reexamination in
2003/04 or did not undergo formal screening at follow-up
but for whom a reliable dementia diagnosis could be es-
tablished. Age-, sex-, and education-specific incidence rates
of MCI per 1,000 person-years were calculated as the num-
ber of new cases divided by the number of person-years at
risk. Age-standardized incidence rates were calculated us-
ing direct standardization to the 2003 Italian population.22
Figure 1. Flow chart detailing the derivation of the study sample. MCI, mild cognitive impairment; MDB, mental deterioration
battery; MMSE, Mini-Mental State Examination.
MILD COGNITIVE IMPAIRMENT IN ELDERLY ITALIANS
53 JAGS JANUARY 2008–VOL. 56, NO. 1
The 95% confidence intervals (95% CIs) were based on the
Poisson distribution. Person-years of observation for cog-
nitively normal persons were calculated as the time from
baseline study until follow-up or death. Because of lack of
information on MCI and dementia onset, the midpoint of
the time from baseline study until follow-up, informal re-
assessment or death were used to calculate person-years for
incident cases of MCI and dementia.
Cox proportional hazard multivariate models includ-
ing sociodemographic variables and APOE genotype as
covariates were used to estimate the risk of progression to
dementia, AD, and VaD associated with MCI and its sub-
types. Age was entered in all models as a continuous vari-
able. Because of the low educational level of the CSBA
cohort (4.3 ? 2.3 years), subjects were categorized has
having 4 or more versus 3 or fewer years of formal edu-
cation. (Three years of education corresponds to the first
level of elementary education provided for in the old Italian
school system.) Based on their APOE genotype, subjects
were dichotomized as carriers or not carriers of an epsilon 4
allele. Occupation, hypertension, cardiovascular disease,
and history of stroke did not significantly affect results of
the Cox models and were not reported in this paper.
Statistical analyses were performed with SYSTAT10
(SPSS Inc., Chicago, IL).
Prevalence of MCI
Of 937 participants free of dementia at baseline, 72 were
diagnosed with MCI (8.0%). Forty-five percent of all sub-
jects positive at screeningwere prevalentMCIcases.(Ofthe
subjects positive at screening, 95% had impairment
according to neuropsychological testing; 54% had no func-
tional dependency, and 57% had no clinical evidence of
dementia.) Approximately 52.8% of prevalent MCI cases
were m1MCI. Only fivesubjects hadaMCI. Onlytwo MCI
cases were diagnosed in participants living in nursing
homes. As shown in Table 1, independent of subtype, MCI
was associated with older age, lower education, and lower
MMSEscore thanincognitivelynormalsubjects. Thelower
MMSE score was expected because of the MCI definition.
No association was found for sex or APOE. Prevalence was
7.7% for overall MCI, 4.1% for m1MCI, and 3.6% for
m?MCI. Age-, sex-, and education-specific prevalence
rates are reported in Table 2.
Risk of Progression to Dementia in MCI
A diagnosis of dementia at follow-up could be reliably es-
tablished or excluded for 60 of the 72 prevalent MCI cases
(83.3%) and for 797 of the 865 subjects who were cogni-
tively normal at baseline (92.1%), including 17 subjects
with MCI and 124 cognitively normal subjects who did not
undergo formal reassessment. Subjects with unknown cog-
nitive outcome were older than those with known cognitive
outcome but did not differ in terms of sex and education
(data not shown).
During 3.8 ? 0.8 years of follow-up, there were 25 in-
cident dementia cases in subjects with MCI (?14% per
year) and 90 incident dementia cases in cognitively normal
subjects (?4% per year). Length of follow-up did not differ
between subjects with MCI and cognitively normal subjects
(P5.75). Of subjects who did not undergo formal re-eval-
uation at follow-up, two with a baseline diagnosis of MCI
and 10 with a baseline diagnosis of normal cognition had
incident dementia. Subjects who developed dementia were
generally older (77.3 ? 6.0 vs 73.0 ? 6.1, Po.001); had
lower MMSE scores (25.5 ? 3.1 vs 28.0 ? 2.0, Po.001);
and were more often female (64.3% vs 51.6%, P5.001),
poorly educated (49.6% vs 27.8% with ?3 years of edu-
cation, Po.001), and APOE epsilon 4 carriers (23.9% vs
14.9%, P5.02; missing data for 31 who did not develop
dementia and 2 who did) than those who did not. Approx-
imately 76% of all progressions from MCI to dementia
occurred in m1MCI subjects, but approximately 30% of
m1MCI cases did not progress to dementia during follow-
up. Of 20 baseline MCI participants who underwent full
reassessment in 2003/04 and did not progress to dementia,
13 still had MCI (65%), and seven (35%) had reconverted
to normal cognition. Analyses of the predictive power of
impairment in specific neuropsychological tests showed a
significant association only between deficit in any memory
test and conversion to dementia (odds ratio54.75, 95%
confidence interval51.48–15.21). No association was
found for reversion to normal cognition, but because of
the small number of cases, a significant association may
have been missed. As shown in Table 3, multivariable-ad-
justed risk of overall dementia in subjects with MCI was
three times the risk in those with normal cognition. Mul-
tivariable-adjusted risks of any dementia and AD in
m1MCI subjects were five and three times, respectively,
the corresponding risk in those with normal cognition.
No association with dementia risk was found for m?MCI.
The lack of significant associations between VaD and
MCI must be interpreted with caution because of the small
number of cases.
Incidence of MCI
The final MCI incidence cohort included 685 participants
of the 865 at risk, including 12 subjects without formal
reassessment at follow-up but for whom a dementia diag-
nosis could be established. During 3.8 ? 0.8 years of fol-
low-up, there were 155 incident cases of MCI (22.6%).
Incident MCI cases represented 64% of all subjects positive
at screening. (Of subjects positive at screening, 94% had
impairment according to neuropsychological testing, 70%
had no functional dependency, and 69% had no clinical
evidence of dementia.) Approximately 51% of all incident
MCI cases were m1MCI. Only 15 aMCI cases were iden-
tified. None of the 13 subjects living in institutions at fol-
low-up received a diagnosis of incident MCI. Age-, sex-,
and education-specific incidence rates for MCI and its sub-
types are reported in Table 4. Overall MCI incidence stan-
dardized to the 2003 Italian population was 76.8 per 1,000
person-years, corresponding to approximately 812,200
new cases of MCI expected per year in Italy. Incidence
tended to be higher in subjects with low education, whereas
no obvious pattern of association was found for age and
sex. Incident MCI cases had lower baseline MMSE scores
than subjects who maintained a normal cognitive function
(27.7 ? 1.6 vs 28.7 ? 1.1, Po.001) but did not differ with
respect to number of APOE epsilon 4 carriers (14.4% vs
RAVAGLIA ET AL.
JANUARY 2008–VOL. 56, NO. 1 JAGS
15.4 %, P5.80). Similar results for MMSE scores and
APOE genotype were found when considering m1MCI and
m?MCI separately (data not shown).
In this elderly population-based Italian cohort, MCI prev-
alence was 7.7%, with cases almost equally distributed be-
tween the m1MCI and m?MCI subtypes. This matches
the 7.5% prevalence found for a comparable diagnostic
construct in the Italian Longitudinal Study of Aging
(ILSA),4a large multicentric survey of people aged 65 to
84. A prevalence of 9.3% was reported in a study of 1,600
elderly subjects living in a district of Tuscany, central Italy,5
in which MCI assessment was also performed in subjects
with MMSE scores between 25 and 26 and a clinical di-
agnosis of questionable cognitive impairment. Overall MCI
prevalence estimate also agrees with results from Finnish23
and Swedish24elderly cohorts using comparable assessment
procedures. Higher rates were reported in an Italian cohort
of 400 elderly community dwellers (16.2%)25and in the
population-based cohort of the Leipzig Longitudinal Study
of the Aged (19.7%).26In these studies, all participants
were tested for MCI, but 1 SD was used instead of 1.5 SD as
the cutoff for defining impairment according to neuropsy-
chological testing. Although increasing sensitivity, this ap-
proach reduces specificity and may overestimate MCI
prevalence. In the Italian cohort,25moreover, selection bi-
as due to subjects’self-referral to a geriatric unit as a part of
a comprehensive geriatric assessment program may have
additionally influenced prevalence data.
MCI prevalence rates ranging from 4.6%27to 28.3%28
were reported in U.S. elderly population-based cohorts, but
the differences in MCI diagnostic criteria do not allow a
direct comparison with the results of the current study, al-
though results from the only U.S. population-based study
investigating MCI subtypes support the finding from the
current study that m?MCI is as frequent as m1MCI.29
The prevalence estimate of approximately 4% for
m1MCI agrees with results for a comparable diagnostic
entity in ILSA (3.2%)6and in an elderly French cohort
Prevalence of aMCI in the CSBA cohort was less than
1%. In the Tuscany cohort,5aMCI prevalence was as high
Table 1. Baseline Characteristic of the Dementia-Free Cohort According to Cognitive Status
MCI with Memory
MCI without Memory
Age, mean ? SD
Women, n (%)
Education ?3 years, n (%)
Mini Mental State Examination score,
mean ? SD?
Apolipoprotein E epsilon 4 allele
carrier, n (%)w
73.6 ? 6.1
28.1 ? 1.5
78.1 ? 8.3 o.001
43 (59.7) o.001
21.4 ? 2.1 o.001
77.4 ? 8.0
21.2 ? 2.2
78.9 ? 8.0
21.6 ? 2.1
132 (15.9)12 (17.1).79 9 (23.7) .203 (9.4) .32
Based on t tests and chi-square test as appropriate. P-value refers to comparison with cognitively normal subjects.
?Mini-Mental State Examination scores are adjusted for age and education.
wMissing data for 35 subjects with normal cognitive function and two subjects with MCI without memory impairment.
MCI5mild cognitive impairment.
Table 2. Prevalence of Mild Cognitive Impairment (MCI) in Study Participants without Dementia at Baseline
MCI with Memory
MCI without Memory
n (%) 95% Confidence Interval
93772 (7.7)6.1–9.738 (4.1) 2.3–5.634 (3.6) 2.6–5.1
MILD COGNITIVE IMPAIRMENT IN ELDERLY ITALIANS
55JAGS JANUARY 2008–VOL. 56, NO. 1
as 4.9%, but rates as low as those of the current study were
reported in an elderly Canadian population.17
The higher MCI prevalence in the older and less-edu-
cated CSBA participants agrees with findings from previous
Italian,4,5European,23,25and U.S. cohorts.21,28The lack of
a sex effect is in contrast to previous reports of an associ-
ation between MCI prevalence and female sex in the Italian
cohorts4,5but agrees with results from other European23,25
and U.S. studies.21,31
In the CSBA cohort, the progression rate from MCI to
dementia was approximately 14% per year. This figure
agrees with findings from an elderly Swedish cohort24and
from U.S. memory clinic samples1and population-based
studies.27A dementia progression rate of 3.8% was report-
ed in the ILSA cohort6but referred to the m1MCI subtype
only. Moreover, although numerically larger than CSBA,
ILSA did not include individuals aged 85 and older, who are
at the highest risk for dementia.10
Risk of developing dementia and AD in CSBA subjects
with MCI was three times the risk of cognitively normal
participants and increased to five times when considering
onlythe m1MCI subtype. However, more thanhalf of MCI
cases remained stable or had converted to normal at follow-
up. The ILSA investigators did not report estimates of de-
mentia risk for MCI participants,6but the figures from the
current study replicate findings from other European26,30
and U.S. population-based studies32independent of the
MCI operational definition used. Moreover, data from a
Dutch memory clinic sample33support the finding that
m1MCI is a better predictor of dementia and AD risk than
m?MCI. Caution is needed in interpreting the negative
findings about VaD because of the small number of cases.
Finally, the findings that approximately one third of the
m1MCI cases did not progress to dementia confirm the
great variability in cognitive outcomes of the MCI subjects
found in ILSA6and other European24,30and U.S. cohorts.32
Overall MCI incidence in the CSBA cohort was as high
as 74.5 per 1,000 person-years (almost twice the incidence
of dementia in the same population8). The estimate for
m1MCI from the current study is almost twice the ILSA
estimate for the same subtype (21.5 per 1,000 person-
years).6The lower educational level of the CSBA sample
than of the ILSA sample and the inclusion of subjects aged
85 and older are possible explanations for this difference,
because older Italian subjects are also those with the lowest
Only two other population-based studies have reported
MCI incidence data.In a Swedish cohort, m1MCI incidence
was 25.9 per 1,000 person-years, but subjects were younger
and better educated than the CSBA sample.35By contrast, in
the Leipzig cohort (aged ?75),25incidence for a diagnostic
entity corresponding to the general definition of MCI in the
current study was 77.0 per 1,000 person-years, which is
similar to the findings from the current study.
Data from previous studies do not show any consistent
pattern of association between incident MCI and sociode-
mographic variables or APOE.3,35
The cognitive reserve hypothesis,36assuming that peo-
ple with higher education are better able to cope with the
early effects of brain-damaging processes, might explain the
trend for association between less education and MCI in-
cidence found in the CSBA cohort. Alternatively, the lack of
Table 3. Association Between Mild Cognitive Impairment (MCI) and Its Subtypes and Risk of Dementia and Its Subtypes in the Conselice Study of Brain Aging Cohort
Cognitive Status at Baseline
Cases, n (%)
HR (95% CI)
Cases, n (%)
HR (95% CI)
Cases, n (%)
HR (95% CI)
Cognitively normal (n5797)
MCI with memory impairment (n527)
MCI without impairment memory (n533)
?Adjusted for age, sex, education, and apolipoprotein epsilon 4 carrier status.
HR5hazard ratio; CI5confidence interval.
RAVAGLIA ET AL.
JANUARY 2008–VOL. 56, NO. 1JAGS
association found between MCI incidence and other ac-
knowledged demographic and genetic risk factors for de-
mentia support the view that people with MCI are a
heterogeneous group with respect to dementia risk.
Strengths of this study are the longitudinal population-
based design, the inclusion of the older segment of the pop-
ulation, and the effort made to collect information for sub-
jects who did not undergo formal assessment at follow-up.
The study has also several limitations. First, prevalence
and incidence of MCI may have been seriously underesti-
mated, because subjects with an MMSE score of 24 or
higher did not receive further testing. This subgroup in-
cluded five incident cases of mild dementia identified only
through a previous clinical diagnosis, so it may also have
included a significant proportion of undetected MCI cases.
Moreover, CSBA participants were evaluated only twice,
and some subjects might have gone from normal to having
dementia without being detected in an MCI phase.
Alternatively, testing only subjects with an MMSE
score less than 24 for MCI might raise the doubt that the
MCI cases included individuals with mild dementia, al-
though the CSBA design provided a careful dementia clin-
ical examination for all subjects positive at cognitive
screening. Moreover, MMSE is not a diagnostic test of de-
mentia,37and alternative MMSE thresholds have been pro-
posed for the Italian population38as low as 18 for women
and 20 for men for less-educated persons aged 75 and older.
A second major limitation relates to the neuropsycho-
logical instrument used for testing subjects positive at cog-
nitive screening. MDB has been validated for use in elderly
and poorly educated Italian subjects living in rural areas,
but its normative data were derived from an Italian sample
aged 20 to 90 living in urban areas and with a higher ed-
ucational level than the CSBA cohort.13The 7.7% found in
CSBA for prevalent MCI cases agrees with the 7% of the
population statistically predicted to be 1.5 SDs below the
mean in any test, but the 17% found for incident MCI cases
is far above this figure. This might raise the doubt that the
adjustments provided in MDB did not fully correct for the
effects of age and education on the performance of CSBA
participants. However, it must be taken into account that
only CSBA participants with a MMSE score less than 24
underwent neuropsychological testing, and there is a known
direct relationship between MMSE and BDM perfor-
mance.39Therefore, the percentage of subjects with impair-
ments might be higher than expected if the tests had been
administered to participants in the whole range of MMSE.
The optimal way to address this concern would have
been to administer MDB to all (or a substantial part) of the
CSBA participants without clinical dementia, independent-
ly of their MMSE score and then to calculate age- and ed-
ucation-specific MDB average scores for the CSBA cohort
to compare them with the original MDB norms, but the
small number of participants and budget restraints did not
permit this study design to be used.
A third possible limitation is an overestimation of MCI
related to the choice not to include subjective memory
complaints among the MCI diagnostic criteria. However,
this criterion has been shown to be unnecessarily restrictive
when applied to subjects from community-based studies.17
Indeed, different from what has been observed in subjects
self-referring to memory clinics, most memory-impaired
community dwellers think of their decline in memory as
‘‘aging-related’’and donot considerit relevant.40However,
a bias selection at baseline, with subjects with cognitive
concerns being more interested in participating in the study
than subjects without awareness of slight cognitive impair-
ment, cannot be excluded, whereas incident cases may in-
clude a higher proportion of the latter subjects. Fourth, an
existent statistical association might have been missed
becauseof limited statistical
drawn from a single center, the study sample could not be
power. Finally, being
Table 4. Incidence Rates of Mild Cognitive Impairment (MCI) and Its Subtypes in the Conselice Study of Brain Aging
MCI with Memory
MCI without Memory
Person-YearsCases, n Rate (95% CI)?
Cases, nRate (95% CI)?
Cases, n Rate (95% CI)?
?Per 1,000 person years.
wStandardized to the 2003 Italian population.
2,174.0 15576.8 (66.8–88.4)w
76 36.3 (29.6–44.5)w
MILD COGNITIVE IMPAIRMENT IN ELDERLY ITALIANS
57 JAGSJANUARY 2008–VOL. 56, NO. 1
representative of the genetic heterogeneity of the Italian Download full-text
This study shows that MCI is common in this elderly Italian
sample and might represent an important societal burden,
because subjects with MCI are at greater risk of dementia
and AD in the short term, especially when the impairment
involves memory. However, MCI is a heterogeneous diag-
nostic category and lacks definite association with ac-
knowledged sociodemographic and genetic risk factors for
AD. Therefore, its prognostic reliability is limited.
Authors thank the staff and participants in the Conselice
Study of Brain Ageing and the Conselice municipal admin-
Conflict of Interest: The study was supported by basic
oriented research grants from the Italian Ministry of Edu-
cation, University and Research. The authors have no fi-
nancial arrangements to disclose.
Author Contributions: GR and PF: study concept and
design, analysis and interpretation of data, and preparation
of the manuscript. FM and AL: study concept and design,
acquisition of subjects and data, analysis and interpretation
of data, and preparation of the manuscript. NP and ER:
study concept and design and interpretation of data. ED,
MB, and PM: study concept and design and analysis and
interpretation of data.
Sponsors’s Role: None.
1. Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med
2. Tuokko H, Frerich R, Graham J et al. Five-year follow-up of cognitive im-
pairment with no dementia. Arch Neurol 2003;60:577–582.
3. Kinsella K, Velkoff VA. U.S. Census bureau, Series P95/01-1, An Aging World:
2001. Washington, DC: U.S. Government Printing Office, 2001.
4. Di Carlo A, Baldereschi M, Amaducci L et al. Cognitive impairment without
dementia in older people: Prevalence, vascular risk factors, impact on disabil-
ity. The Italian Longitudinal Study on Aging. J Am Geriatr Soc 2000;48:775–
5. Tognoni G, Ceravolo R, Nucciarone B et al. From mild cognitive impairment
to dementia: A prevalence study in a district of Tuscany, Italy. Acta Neurol
6. Solfrizzi V, Panza F, Colacicco AM et al. Vascular risk factors, incidence of
MCI, and rates of progression to dementia. Neurology 2004;63:1882–1891.
7. Ravaglia G, Forti P, Maioli F et al. Education, occupation, and prevalence of
dementia: Findings from the conselice study. Dement Geriatr Cogn Disord
8. RavagliaG,FortiP,Maioli Fetal.Incidence and etiologyofdementiainalarge
elderly Italian population. Neurology 2005;64:1525–1530.
9. Katz S, Downs TD, Cash HR et al. Progress in development of the index of
ADL. Gerontologist 1970;10:20–30.
10. Lawton MP, Bordy EM. Assessment of older people: Self maintaining and
instrumental activities of daily living. Gerontologist 1969;9:179–185.
11. Valente C, Maione P, Lippi A et al. Validation of the Mini-Mental State Ex-
amination (MMSE) as a screening instrument for dementia in an Italian pop-
ulation. Giornale di Gerontol 1992;40:161–165.
12. Licastro F, Pedrini S, Ferri C et al. Gene polymorphism affecting alpha-1-
antichymotrypsin and interleukin-1 plasma levels increases Alzheimer’s dis-
ease risk. Ann Neurol 2000;48:388–391.
13. Magni E, Binetti G, Bianchetti A et al. Mini-mental state examination: A
normative study in an Italian elderly population. Eur J Neurol 1996;3:1–5.
14. Carlesimo GA, Caltagirone C, Gainotti G. The mental deterioration battery:
Normative data, diagnostic reliability and qualitative analyses of cognitive
impairment. The group for the standardization of the mental deterioration
battery. Eur Neurol 1996;36:378–384.
15. Italian Group on the Neuropsychological Study of Aging: Italian standard-
ization and classification of neuropsychological tests. Ital J Neurol Sci
16. Winblad B, Palmer K, Kipivelto M et al. Mild cognitive impairment. Beyond
controversies,towards a consensus: Report of the international workinggroup
on mild cognitive impairment. J Intern Med 2004;256:240–246.
17. Fisk J, Merry HR, Rockwood K. Variations in case definition affect prevalence
but not outcomes of mild cognitive impairment. Neurology 2003;61:1179–
18. American Psychiatric Association. Diagnostic and Statistical Manual of Men-
tal Disorders, 4th Ed. Washington, DC: American Psychiatric Association,
19. McKhann G, Drachman D, Folstein M et al. Clinical diagnosis of Alzheimer’s
disease: Report of the NINCDS-ADRDA Work Group under the auspices of
department of health and human services task force on Alzheimer’s disease.
20. Roman GC, Tatemichi TK, Erkinjuntti Tet al. Vascular dementia: Diagnostic
criteria for research studies. Report of the NINDS-AIREN International
Workshop. Neurology 1993;43:250–260.
21. Unverzagt FW, Gao S, Bayewu O et al. Prevalence of cognitive impairment.
Data from the Indianapolis Study of Health and Aging. Neurology 2001;
22. Italian National Institute of Statistic (ISTAT). Demography in Figures [on-
line]. Available at http://demo.istat.it/index.html Accessed October 1, 2006.
23. Kipivelto M, Helkala EL, Hanninen T et al. Midlife vascular risk factors and
late-life mild cognitive impairment. Neurology 2001;56:1683–1689.
24. Palmer K, Wang HX, Backman L et al. Differential evolution of cognitive
impairment in non demented older persons: Results from the Kungsholmen
Project. Am J Psychiatry 2002;159:436–442.
25. Zanetti M, Ballabio C, Abbate C et al. Mild cognitive impairment subtypes
and vascular dementia incommunity-dwellingelderly people: A 3-year follow-
up study. J Am Geriatric Soc 2006;54:580–586.
26. Busse A, Bischkopf J, Riedel-Heller SG et al. Mild cognitive impairment:
Prevalence and incidence according to different diagnostic criteria. Results of
the Leipzig Longitudinal Study of the Aged. Br J Psychiatr 2003;182:449–454.
27. Tschanz JT, Welsh-Bohmer KA, Lyketsos CG et al. Conversion to dementia
from mild cognitive disorder. The Cache County Study. Neurology 2006;67:
28. Manly JJ, Bell-McGinty S, Tang MX et al. Implementing diagnostic criteria
and estimating frequency of mild cognitive impairment in an urban commu-
nity. Neurology 2005;62:1739–1746.
29. Lopez OL, Jagust WJ, DeKosky STet al. Prevalence and classification of mild
cognitive impairment in the Cardiovascular Health Study Cognition Study.
Arch Neurol 2003;60:1385–1389.
30. Larrieu S, Letenneur L, Orgogozo JM et al. Incidence and outcome of mild
cognitive impairment in a population-based prospective cohort. Neurology
31. Lopez OL, Jagust WJ, Dulberg C et al. Risk factors for mild cognitive im-
pairment in the Cardiovascular Health Study Cognition Study. Arch Neurol
32. Ganguli M, Dodge H, Shen C et al. Mild cognitive impairment, amnestic type.
An epidemiologic study. Neurology 2004;63:115–121.
33. Rasquin SMC, Lodder J, Visser PJ et al. Predictive accuracy of MCI subtypes
for Alzheimer’s disease and vascular dementia in subjects with mild cognitive
impairment: A 2-year follow-up study. Dement Geriatr Cogn Disord
34. Sestito P, Tangorra R, Freguija C. Poverty and social exclusion in Italy. Annex
to the National Action Plan against poverty and Social Exclusion 2003 [on-
line]. Available at http://europa.eu.int/comm/employment_social/news/2001/
jun/it_statannex_napincl_2003_Final_en.pdf Accessed October 21, 2006.
35. Tervo S, Kivipelto M, Hanninen T et al. Incidence and risk factors for mild
cognitive impairment: A population-based three-year follow-up study of cog-
nitively healthy elderly subjects. Dement Geriatr Cogn Disord 2004;17:196–
36. Coffey CE, Saxton JA, Ratcliff G et al. Relation of education to brain size in
normal aging: Implications for the reserve hypothesis. Neurology 1999;53:
37. Tombaugh TN, McIntyre NJ. The mini-mental state examination: A compre-
hensive Review. J Am Geriatr Soc 1992;40:922–935.
38. Grigoletto F, Zappala G, Anderson DWet al. Norms for the mini-mental state
examination in a healthy population. Neurology 1999;53:315–320.
39. Pasqualetti P, Moffa F, Chiovenda P et al. Mini-mental state examination and
mental deterioration battery: Analysis of the relationship and clinical impli-
cations. J Am Geriatr Soc 2002;50:1577–1581.
40. Jungwirth S, Fisher P, Weisgram S et al. Subjective memory complaints and
objective memory impairment in the Vienna-Transdanube aging community.
J Am Geriatr Soc 2004;52:263–268.
RAVAGLIA ET AL.
JANUARY 2008–VOL. 56, NO. 1JAGS