Experimental Gerontology 75 (2016) 42–47

Contents lists available at ScienceDirect

Experimental Gerontology

journal homepage: www.elsevier.com/locate/expgero

Mild cognitive impairment is associated with falls among older adults:

Findings from the Irish Longitudinal Study on Ageing (TILDA)
Stefanos Tyrovolas a,b,⁎, Ai Koyanagi a,b, Elvira Lara a,b, Ziggi Ivan Santini a, Josep Maria Haro a,b
a
Parc Sanitari Sant Joan de Déu, Universitat de Barcelona, Fundació Sant Joan de Déu, Dr. Antoni Pujadas, 42, Sant Boi de Llobregat, Barcelona 08830, Spain
b
Instituto de Salud Carlos III, Centro de Investigación Biomédica en Red de Salud Mental, CIBERSAM, Monforte de Lemos, 3-5, Pabellón 11, 28029, Madrid, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: The role of mild cognitive impairment (MCI) on falls among older adults remains under-
Received 17 August 2015 investigated. The aim of this study was to evaluate the association between MCI and number of falls or occurrence
Received in revised form 15 December 2015 of non-accidental falls among older adults.
Accepted 16 December 2015 Methods: Data from the first wave of the Irish longitudinal Study on Ageing (TILDA) was analysed. The analytical
Available online 18 December 2015
sample consisted of 5364 individuals aged ≥ 50 years. MCI was defined as: Montreal Cognitive Assessment
Section Editor: Holly M. Brown-Borg
(MoCA) score b 26; the presence of subjective cognitive complaints; Mini-Mental State Examination (MMSE)
score ≥ 14; and no limitations in activities of daily living (ADL). Multivariable Poisson and logistic regression anal-
Keywords: yses were conducted to assess the association between MCI and number of falls or the presence of non-accidental
Mild cognitive impairment falls in the past 12 months.
Falls Results: The prevalence of MCI was 10.1%. In the fully-adjusted model, MCI was associated with a higher rate of
Gait speed falls (PR = 1.41 95%CI = 1.05–1.89) and odds for non-accidental falls in the past 12 months (OR = 1.67
Muscle strength 95%CI = 1.07–2.61). Muscle strength and performance indicators, and medical health conditions were influential
factors in the association between MCI and falls but did not fully explain the association.
Conclusion: MCI is related with higher rates of falls and the occurrence of non-accidental falls among older adults.
Future studies are warranted to clarify the underlying mechanism linking MCI and falls, and to establish interven-
tions targeting MCI to reduce the risk of falls.
© 2015 Elsevier Inc. All rights reserved.

1. Introduction
The European commission has recognized population aging as one of
the most challenging policy issues of this century in Europe (European
Commission 2006). Advanced age is accompanied by various co-
morbidities that affect health status and quality of life including falls
(Janssen et al., 2002; Landi et al., 2013; Newman et al., 2006). Falls are
a major health care problem for the elders. Almost 30% of the older pop-
ulation have been reported to experience a fall accident at least once per
year (Muir et al., 2012). Moreover, falls are associated with a higher risk
of loss of independence, autonomy, and confidence. Falls are one of the
major contributors to the increased need for specialized care and hospi-
talization among older adults, while it is also associated with higher
rates of morbidity, mortality, and institutionalization (Tinetti et al.,
1995; Tinetti and Williams, 1997). Additionally, the cost of falls for
the public health services is high. For example, in the UK, the cost of
fall-related hospitalizations among older adults is almost £1 billion per
year (Scuffham et al., 2003).
⁎ Corresponding author at: Parc Sanitari Sant Joan de Déu, Fundació Sant Joan de Déu,
CIBERSAM, Dr. Antoni Pujadas, 42, 08830 Sant Boi de Llobregat, Barcelona, Spain.
E-mail address: s.tyrovolas@pssjd.org (S. Tyrovolas).
Various factors such as vision and hearing problems, abnormal blood
pressure, mobility limitation, neuropsychiatric disorders, sarcopenia,
and frailty have been associated with falls (Tinetti et al., 1986; Robbins
et al., 1989; Shumway-Cook et al., 1997; Vellas et al., 1997; Mühlberg
and Sieber, 2004). Among neuropsychiatric disorders, decline in cogni-
tive function has been related with greater risks of falls in the older pop-
ulation. Recent studies have reported an increased frequency of falls
with lower Mini-Mental State Examination (MMSE) scores (i.e., loss of
global cognitive ability) (Gleason et al., 2009). Impairments in attention
(Amboni et al., 2013), processing speed (Chen et al., 2012), and execu-
tive functions (Banich, 2009) have been proposed as a set of interrelated
factors in the pathway between cognitive impairment and falls. Based
on these previous findings, some researchers have proposed that fall
and injury prevention strategies may benefit from focusing on the
early prevention of cognitive decline (Montero-Odasso et al., 2009). In
particular, in recent years, mild cognitive impairment (MCI), which is
considered an intermediate state between normal aging and dementia,
is gaining further attention from the viewpoint of prevention of demen-
tia or cognitive decline. However, despite the potentially important
role that cognitive function plays in the occurrence of falls, the associa-
tion between MCI and falls among older adults still remains under-
investigated (Delbaere et al., 2012).

http://dx.doi.org/10.1016/j.exger.2015.12.008
0531-5565/© 2015 Elsevier Inc. All rights reserved.
 S. Tyrovolas et al. / Experimental Gerontology 75 (2016) 42–47
Given the rapid aging occurring in Europe, the scarcity of studies on
MCI and falls, and a complete lack of studies on this topic from Ireland,
the aim of the present work was to evaluate the associations between
MCI and frequency of falls or occurrence of non-accidental falls in a
large, nationally-representative sample of non-institutionalized older
Irish individuals.
2. Methods
2.1. Study design and sample
Data from the first wave of the Irish Longitudinal Study on Ageing
(TILDA) was analyzed. The full description of the survey and the sam-
pling procedures can be found elsewhere (Cronin et al., 2013). Briefly,
TILDA was an Irish nationally-representative, cross-sectional study on
the economic, health, and social status of the non-institutionalized
population, and was conducted between 2009 and 2010 by Trinity
College in Dublin (Cronin et al., 2013). The sample included a total of
8504 people [individuals aged ≥50 years (n = 8175) and their spouses
or partners younger than 50 years (n = 329)]. Of these individuals,
5895 completed a health assessment. Information was obtained
by face-to-face interviews conducted by trained professionals using
Computer Assisted Personal Interviewing (CAPI). The response rate
was 62% (Whelan and Savva, 2013).
The Trinity College Dublin approved the design and procedures of
the study. Informed consent was obtained from all participants. Individ-
uals were not eligible for inclusion if they reported a doctor's diagnosis
of dementia. Furthermore, individuals who were not able to consent
personally because of severe cognitive impairment (at interviewer’s
discretion) were also excluded.
2.1.1. Number of falls and the presence of non-accidental falls
The number of falls in the past 12 months was assessed by the ques-
tion “How many times have you fallen in the last year?” Information on
the presence of non-accidental falls in the past 12 months was assessed
by the question “Were any of these falls non-accidental, i.e., with no
apparent or obvious reason?” among those who had fallen in the past
12 months. The answer options were “Yes” or “No”.
2.1.2. Mild cognitive impairment
The case definition of MCI was based on the core criteria outlined by
the National Institute on Aging-Alzheimer's Association (Albert et al.,
2011):
1) Concern about a change in cognition: subjective cognitive com-
plaints were assessed by the question “How would you rate your
day-to-day memory at present time?” with answer options: excel-
lent, very good, good, fair, and poor. Those who replied fair or poor
were considered to have subjective cognitive complaints.
2) Objective evidence of impairment in one or more cognitive domains,
typically including memory: cognitive function was assessed with
the Montreal Cognitive Assessment (MoCA) (score range: 0–30).
This tool has been demonstrated to be sensitive to mild cognitive
deficits when applied in cognitively intact older adults (Kenny
et al., 2013), and includes measures of executive function, language,
memory, attention, orientation, calculation, and visuospatial ability.
Cognitive impairment was defined as a MoCA score b 26 (Freitas
et al., 2013).
3) Preservation of independence in functional abilities: the participants
were presented with a list of six basic standard ADLs on dressing,
walking, bathing, eating, getting in or out of bed, and using the toilet
(Katz et al., 1963), and were asked if they have difficulty with these
activities. They were also asked to exclude any difficulties that are
expected to last for less than three months. Those who claimed
to have difficulty with any of the six abovementioned ADLs were
excluded from the analysis.
43
4) Not demented: individuals who obtained a score b14 on the MMSE
were excluded from the analytical sample (Shigemori et al., 2010).
2.1.3. Sociodemographic and lifestyle characteristics
Sociodemographic and lifestyle characteristics included age (50–59,
60–69, 70–79, ≥80 years), gender, education (primary, secondary, ter-
tiary), wealth, living arrangement (alone or not), residence [urban
(Dublin city or county/another town or city) or rural], physical activity,
and problem drinking. Wealth (financial strain) was assessed by the
statement “shortage of money stops me from doing the things I want
to do” with answer options never, rarely, sometimes, and often. Physical
activity was measured using the short form of the International Physical
Activity Questionnaire, which converts levels of physical activity of var-
ious domains into predicted kilocalories expended per week (Craig
et al., 2003). Problem drinking was assessed by the CAGE screening
test with scores of ≥ 2 being used as a cut-off for problem drinking
(Mayfield et al., 1974).
2.1.4. Muscle strength and performance
Handgrip strength and gait speed were considered indicators of
muscle strength and performance respectively (Tyrovolas et al., 2015).
Grip strength was assessed using a dynamometer. Two readings from
the dominant hand were taken, and the mean strength was calculated.
Gait speed was measured using the GAITRite portable electronic walk-
way system (CIR Systems, Inc., Havertown, PA). Participants walked at
their usual pace along a 4.88-m (16 ft) walkway with an extra 2.5 m
at each end to allow for acceleration and deceleration. Gait speed was
then calculated as meters per second and then transformed to centime-
ters per second.
2.1.5. Obesity and medical health conditions
Weight and height were measured using standard procedures. Body
mass index (BMI) was calculated as weight in kilograms divided by
height in meters squared. Obesity was defined as BMI ≥ 30 kg/m2. The
presence of medical conditions was assessed by asking the respondents
about whether they were ever told by a doctor that they have angina,
arthritis (including osteoarthritis and rheumatism), congestive heart
failure, diabetes or high blood sugar, heart attack (including myocardial
infarction and coronary thrombosis), stroke (cerebral vascular disease),
or Parkinson's disease. Heart disease referred to having at least one of:
angina, congestive heart failure, and heart attack. Depression was mea-
sured with the 20-item Center for Epidemiologic Studies Depression
(CES-D) (Radloff, 1977) based on symptoms experienced in the past
week, and was defined as a CES-D score of ≥16 (Beekman et al., 1997).
2.2. Statistical analysis
A descriptive analysis was conducted to characterize the study sam-
ple (overall and by the presence of MCI). The differences in sample char-
acteristics by the presence of MCI were tested by chi-squared tests and
student's t-tests for categorical and continuous variables respectively.
Poisson and logistic regression analyses were done with number of
falls and the presence of non-accidental falls in the past 12 months as
the outcome respectively. MCI was the main covariate of interest.
Since it is possible that the inclusion of different blocks of control vari-
ables in the model affects the association between MCI and falls in dif-
ferent ways, we conducted hierarchical analyses where three different
models were constructed for each outcome: Model 1 — adjusted for
sociodemographic and lifestyle characteristics; Model 2 — adjusted for
covariates in Model 1 and grip strength and gait speed; Model 3 —
adjusted for covariates in Model 2 and obesity and medical health con-
ditions. All variables were included in the models as categorical vari-
ables with the exception of grip strength and gait speed (continous
variables). The selection of the covariates was based on past literature
(Muir et al., 2012; Tinetti et al., 1995; Tinetti et al., 1986; Robbins
et al., 1989). In order to assess the influence of multicolinearity, we
44 S. Tyrovolas et al. / Experimental Gerontology 75 (2016) 42–47

calculated the variance inflation factor (VIF) value for each independent 12 months is shown in Table 2. In the model adjusted for
variable. The highest VIF was 2.44, which is much lower than sociodemographic and lifestyle characteristics (Model 1), MCI was asso-
the commonly used-cut off of 10 (O'Brien, 2007), indicating that ciated with a higher rate of falls 1.51 (95%CI 1.15–1.97). After the addi-
multicolinearity was unlikely to be a problem in our analyses. The anal- tion of hand grip strength and gait speed, the PR (95%CI) became 1.47
yses were done with Stata version 13.1 (Stata Corp LP, College Station, (95%CI 1.11–1.95). This association remained significant even after fur-
Texas). In order to generate nationally-representative estimates, in all ther adjustment for obesity and other medical health conditions [PR
analyses, the sample weighting and the complex study design were 1.41 (95%CI 1.05–1.89)]. In the fully-adjusted model (Model 3), gender,
taken into account with Taylor linearization methods. Prevalence ratios grip strength, and arthritis were also significantly associated with falls.
(PR), odds ratios (OR), and 95% confidence intervals (95%CI) are The association of MCI and other covariates with non-accidental falls
reported. The level of statistical significance was set at P b 0.05. in the past 12 months is shown in Table 3. MCI was associated with 1.93
(95%CI 1.28–2.93) times higher odds for non-accidental falls in the
3. Results model adjusted for sociodemographic and lifestyle factors (Model 1).
Further adjustment for grip strength and gait speed (Model 2), or obe-
The analytical sample consisted of 5364 individuals aged ≥50 years sity and medical health conditions in addition to grip strength and gait
with no limitations in ADL and a MMSE score of ≥14. The prevalence speed (Model 3) attenuated the ORs [Model 2: OR 1.68 (95%CI 1.09–
of MCI was 10.1%. Overall, 18.9% had fallen at least once and 4.6% had 2.60); Model 3: OR 1.67 (95%CI 1.07–2.61)] when compared to Model
non-accidental falls in the past 12 months. The characteristics of the 1, but results were still statistically significant in both Models 2 and 3.
study participants (overall and by MCI status) are summarized in In the fully-adjusted model (Model 3), apart from MCI, the factors sig-
Table 1. The following characteristics were significantly associated nificantly associated with non-accidental falls were financial strain,
with MCI: older age, lower levels of education, higher levels of financial grip strength, gait speed, and arthritis.
strain, living alone, rural residence, low physical activity, weaker hand
grip strength, slower gait speed, obesity, and the presence of medical 4. Discussion
conditions (arthritis, stroke, heart disease, and depression). The associ-
ation of MCI and other covariates with number of falls in the past The present work revealed a strong association between MCI and
number of falls or presence of non-accidental falls in the past 12 months
Table 1
in the older Irish population. Muscle strength and performance, and
Sample characteristics (overall and by presence of mild cognitive impairment). health conditions were influential factors in the association between
MCI and falls but did not fully explain the association. While there is a
Characteristics Categories Overall Mild cognitive P-valuea
impairment
growing literature on cognitive decline and falls, studies specifically
on the topic of MCI and falls are scarce. To date, a small number of stud-
No Yes
ies (Liu-Ambrose et al., 2008; Borges et al., 2015; Dealbaere et al., 2012;
Age (years) 50–59 44.7 46.4 29.7 b0.0001 Uemura et al., 2014) have assessed this association. However, with the
60–69 32.8 32.5 35.5
exception of one study (Uemura et al., 2014), the rest had very small
70–79 17.1 16.0 26.6
≥80 5.4 5.1 8.1
sample size and were conducted in limited geographical areas. To the
Gender Female 51.8 52.0 50.2 0.4883 best of our knowledge, this is the first study that examined this associ-
Male 48.2 48.0 49.8 ation using a large nationally-representative dataset of the older
Education Primary 31.4 28.9 53.3 b0.001 European (Irish) population.
Secondary 46.7 47.6 38.6
Our findings on the association between MCI and falls are in line
Tertiary 21.9 23.4 8.0
Financial strain Never 22.4 22.2 23.5 0.0017 with previous studies. For example, Liu-Ambrose et al. (2008), in a sam-
Rarely 23.2 23.7 18.6 ple of 158 older Canadians, reported that females with MCI had higher
Sometimes 37.4 37.7 34.4 physiological risk of falling and increased postural sway compared to
Often 17.1 16.4 23.4
females without MCI. Additionally, Dealbaere et al. (2012) analyzed a
Living alone No 80.9 81.6 74.5 0.0004
Yes 19.1 18.4 25.5
sample of 419 non-demented community-dwelling adults in Sydney,
Residence Rural 48.0 46.6 60.5 b0.0001 and reported that MCI was associated with a 1.72 (95%CI 1.03–2.89)
Urban 52.0 53.4 39.5 times greater risk for falls. Also, Borges et al. (2015), in a sample of
Physical activity Low 28.4 27.7 35.5 0.0015 104 community-dwelling elders in Brazil, showed that the prevalence
Medium 35.2 35.4 33.1
of falls in MCI was higher than in cognitively healthy older adults. Finally
High 36.4 36.9 31.5
Problem drinking No 87.1 87.0 87.5 0.8132 Uemura et al. (2014) analyzed a sample of 4474 community-dwelling
Yes 12.9 13.0 12.5 older Japanese adults and concluded that MCI has an effect on fear of
Grip strength (kg) Mean (SD) 26.7 (10.1) 26.9 (10.1) 25.2 (9.6) 0.0013 falling.
Gait speed (cm/s) Mean (SD) 11.9 (2.5) 12.0 (2.4) 10.9 (2.5) b0.0001 Our analysis revealed a consistent association between MCI and the
Obesity No 66.4 66.9 61.3 0.0185
(BMI ≥ 30 kg/m2) Yes 33.6 33.1 38.7
number of falls or the presence of non-accidental falls in the past
Arthritis No 75.0 76.0 65.5 b0.0001 12 months. Furthermore, the inclusion of different blocks of covariates
Yes 25.0 24.0 34.5 in the models attenuated the association but even after inclusion of all
Parkinson's disease No 99.7 99.7 99.3 0.2018 potentially influential variables, the association between MCI and falls
Yes 0.3 0.3 0.7
remained significant. The exact mechanism linking MCI with falls is un-
Stroke No 98.7 98.8 97.2 0.0038
Yes 1.3 1.2 2.8 clear but attention deficit, impaired psychomotor processing, problem-
Diabetes No 93.3 93.6 91.3 0.0657 solving, and spatial awareness, which are characteristics associated with
Yes 6.7 6.4 8.7 MCI, have been reported to be related with balance control and conse-
Heart disease No 92.4 92.9 87.5 b0.0001 quently with falls (Alexander and Hausdorff, 2008). Differences in
Yes 7.6 7.1 12.5
Depression No 92.2 93.4 81.1 b0.0001
brain structures between those with and without MCI have also been
Yes 7.8 6.6 18.9 reported. For example, individuals with MCI may have reduced integrity
of the posterior regions of the brain, and their medial temporal
Abbreviation: BMI: body mass index.
Data are percentage (%) unless otherwise stated. lobe, insula, and thalamus may constitute of reduced gray matter
a
P-values were calculated by Chi-square tests and Student's t-tests for categorical and (Medina et al., 2006). Since these regions of the brain are known to be
continuous variables respectively. associated with attention and balance control (Zimmerman et al.,
 S. Tyrovolas et al. / Experimental Gerontology 75 (2016) 42–47 45

Table 2
Correlates of number of falls in the past 12 months assessed by multivariable Poisson regression.

Characteristics Categories Model 1 Model 2 Model 3

PR 95%CI PR 95%CI PR 95%CI

Mild cognitive impairment Yes vs. No 1.51⁎⁎ [1.15,1.97] 1.47⁎⁎ [1.11,1.95] 1.41⁎ [1.05,1.89]
Age (years) 50–59 1.00 1.00 1.00
60–69 1.07 [0.87,1.31] 1.00 [0.82,1.23] 0.97 [0.79,1.19]
70–79 1.28 [1.00,1.64] 1.08 [0.81,1.43] 1.05 [0.80,1.38]
≥80 1.47 [0.98,2.20] 1.13 [0.72,1.78] 1.07 [0.68,1.69]
Gender Female 1.00 1.00 1.00
Male 1.08 [0.91,1.29] 1.37⁎ [1.05,1.79] 1.37⁎ [1.04,1.81]
Education Primary 1.00 1.00 1.00
Secondary 1.06 [0.85,1.32] 1.10 [0.88,1.38] 1.11 [0.88,1.39]
Tertiary 1.04 [0.82,1.31] 1.12 [0.87,1.43] 1.09 [0.85,1.41]
Financial strain Never 1.00 1.00 1.00
Rarely 1.27 [0.96,1.68] 1.26 [0.95,1.67] 1.27 [0.95,1.69]
Sometimes 1.26 [0.97,1.62] 1.24 [0.95,1.60] 1.23 [0.94,1.61]
Often 1.34⁎ [1.01,1.78] 1.31 [0.98,1.75] 1.26 [0.93,1.70]
Living alone Yes vs. No 1.28⁎ [1.04,1.57] 1.24⁎ [1.00,1.53] 1.23 [0.98,1.53]
Residence Rural 1.00 1.00 1.00
Urban 1.23⁎ [1.02,1.48] 1.20 [1.00,1.46] 1.20 [0.99,1.45]
Physical activity Low 1.00 1.00 1.00
Medium 1.18 [0.95,1.48] 1.20 [0.96,1.51] 1.21 [0.97,1.52]
High 0.95 [0.76,1.19] 0.99 [0.79,1.24] 1.01 [0.80,1.27]
Problem drinking Yes vs. No 1.11 [0.87,1.42] 1.10 [0.86,1.42] 1.09 [0.85,1.39]
Grip strength (kg) 0.98⁎ [0.97,1.00] 0.98⁎ [0.97,1.00]
Gait speed (cm/s) 0.97 [0.92,1.02] 0.97 [0.92,1.03]
Obesity (BMI ≥ 30 kg/m2) Yes vs. No 0.99 [0.81,1.19]
Arthritis Yes vs. No 1.24⁎ [1.02,1.53]
Parkinson's disease Yes vs. No 2.45 [0.55,10.97]
Stroke Yes vs. No 1.39 [0.72,2.71]
Diabetes Yes vs. No 1.10 [0.84,1.46]
Heart disease Yes vs. No 1.02 [0.70,1.48]
Depression Yes vs. No 1.24 [0.86,1.77]

Abbreviations: PR: prevalence ratio; CI: confidence interval; BMI: body mass index.
The models are mutually adjusted for all covariates in the respective columns.
Grip strength and gait speed were included in the models as continuous variables.
⁎ P b 0.05.
⁎⁎ P b 0.01.

2006), individuals with MCI may be more likely to fall due to impair-
ment in attention and equilibrium.
Apart from MCI, factors such as handgrip strength, gait speed and ar-
thritis, were also associated with falls. Muscle strength and performance
have been reported to be significant predictors of falls (Mühlberg
and Sieber, 2004). Furthermore, an interrelated pathway among
gait speed, balance control, cognitive decline and falls in older adults
has been proposed. In addition, falls commonly occur in patients with
arthritis (Kaz Kaz et al., 2004), and the role of arthritis on the epidemi-
ology of falls has previously been reported. Since falls are associated
with high healthcare expenditures (Ambrose et al., 2013), the preven-
tion of MCI and other co-morbidities may be important to promote
healthy aging and minimize disability and their resulting distress and
costs.
4.1. Strengths and limitations
The present study has several strengths. It is one of the few studies
that evaluated the association between MCI and falls (especially with
large sample size), while it is also the first study from Ireland on this
topic. In terms of limitations, since some data were obtained using
self-reported measures, reporting-bias may be present. Also, because
the survey was not designed to generate clinical diagnoses for dementia,
some individuals with dementia may have been included in our analyt-
ical sample. However, we made use of information on dementia diagno-
sis obtained from the participant or family members, and excluded
those with MMSE score b 14, which is a cut-off for dementia that has
been used previously (Shigemori et al., 2010). Furthermore, there are
currently no standard definitions for the identification of MCI and a va-
riety of definitions have been used in previous studies. In our study,
objective cognitive impairment (one of the MCI criteria) was based on
a MoCA score b 26 which has been reported to have high sensitivity
and specificity for detecting MCI (Nasreddine et al., 2005). However,
as with most population-based studies, a thorough clinical evaluation
to detect MCI was lacking. Next, we did not have information on spatial
and temporal aspects of gait (e.g., step length and time). Conducting
analyses with more detailed information could have lead to a better
understanding of the association between MCI and falls. Finally, due
to the cross-sectional design, we were unable to establish causal
relationships.
5. Conclusions
The present work evaluated the role of MCI on falls among older
Irish adults. In view of rapid population aging accompanied with an in-
crease in the prevalence of cognitive impairment, it is of major interest
nowadays to study the negative health outcomes of cognitive impair-
ment. In our study, the potential importance of early detection and
management of MCI to prevent falls among the elderly has been
highlighted. Also the prevention or treatment of arthritis and muscle
function decline, which are becoming increasingly prevalent among
the elders, may also constitute effective means for reducing falls and
to promote healthy aging among the older population.
Conflicts of interest
The authors report no relationships that could be construed as a con-
flict of interest.
46 S. Tyrovolas et al. / Experimental Gerontology 75 (2016) 42–47

Table 3
Correlates of non-accidental falls in the past 12 months assessed by multivariable logistic regression.

Characteristics Categories Model 1 Model 2 Model 3

OR 95%CI OR 95%CI OR 95%CI

Mild cognitive impairment Yes vs. No 1.93⁎⁎ [1.28,2.93] 1.68⁎ [1.09,2.60] 1.67⁎ [1.07,2.61]
Age (years) 50–59 1.00 1.00 1.00
60–69 1.30 [0.89,1.89] 1.13 [0.77,1.66] 1.04 [0.70,1.54]
70–79 1.78⁎ [1.12,2.82] 1.12 [0.67,1.85] 1.05 [0.63,1.75]
≥80 2.35⁎ [1.18,4.71] 0.97 [0.43,2.23] 0.86 [0.36,2.04]
Gender Female 1.00 1.00 1.00
Male 0.62⁎⁎ [0.45,0.86] 1.22 [0.78,1.91] 1.22 [0.76,1.95]
Education Primary 1.00 1.00 1.00
Secondary 1.25 [0.83,1.91] 1.41 [0.91,2.17] 1.41 [0.91,2.19]
Tertiary 0.89 [0.58,1.39] 1.08 [0.69,1.71] 1.01 [0.64,1.60]
Financial strain Never 1.00 1.00 1.00
Rarely 1.49 [0.95,2.35] 1.44 [0.91,2.28] 1.48 [0.93,2.37]
Sometimes 1.46 [0.96,2.22] 1.39 [0.91,2.12] 1.39 [0.89,2.16]
Often 1.88⁎ [1.16,3.04] 1.81⁎ [1.12,2.94] 1.83⁎ [1.11,3.02]
Living alone Yes vs. No 1.30 [0.89,1.91] 1.23 [0.83,1.80] 1.24 [0.84,1.84]
Residence Rural 1.00 1.00 1.00
Urban 1.33 [0.95,1.88] 1.30 [0.92,1.85] 1.30 [0.92,1.85]
Physical activity Low 1.00 1.00 1.00
Medium 1.17 [0.81,1.69] 1.22 [0.84,1.77] 1.19 [0.81,1.73]
High 0.83 [0.56,1.23] 0.90 [0.61,1.34] 0.90 [0.60,1.34]
Problem drinking Yes vs. No 1.09 [0.69,1.72] 1.09 [0.68,1.74] 1.11 [0.70,1.77]
Grip strength (kg) 0.95⁎⁎⁎ [0.92,0.97] 0.95⁎⁎⁎ [0.93,0.98]
Gait speed (cm/s) 0.90⁎⁎ [0.84,0.96] 0.90⁎⁎ [0.84,0.97]
Obesity (BMI ≥ 30 kg/m2) Yes vs. No 0.90 [0.63,1.29]
Arthritis Yes vs. No 1.58⁎⁎ [1.15,2.16]
Parkinson's disease Yes vs. No 3.41 [0.73,15.99]
Stroke Yes vs. No 1.45 [0.56,3.74]
Diabetes Yes vs. No 1.09 [0.63,1.87]
Heart disease Yes vs. No 0.75 [0.39,1.44]
Depression Yes vs. No 0.82 [0.49,1.37]

Abbreviations: OR: odds ratio; CI: confidence interval; BMI: body mass index.
The models are mutually adjusted for all covariates in the respective columns.
Grip strength and gait speed were included in the models as continuous variables.
⁎ P b 0.05.
⁎⁎ P b 0.01.
⁎⁎⁎ P b 0.001.

Source of funding
Stefanos Tyrovolas received a scholarship from the Foundation for
Education and European Culture (IPEP) to undertake his post-doctoral
research, of which this work is a part. Ai Koyanagi's work was supported
by the Miguel Servet contract financed by the CP13/00150 project, inte-
grated into the National R + D + I and funded by the ISCIII — General
Branch Evaluation and Promotion of Health Research — and the
European Regional Development Fund (ERDF-FEDER). Elvira Lara's
work is supported by the FPU predoctoral grant (FPU13/03573) from
the Spanish Ministry of Education, Culture and Sports. Ziggi Ivan
Santini's work has received funding from the People Programme
(Marie Curie Actions) of the European Union's Seventh Framework Pro-
gramme FP7/2007–2013 under REA grant agreement no. 316795.
Acknowledgments
The authors are particularly grateful to the men and women who
participated in the TILDA survey.
References
Albert, M.S., DeKosky, S.T., Dickson, D., Dubois, B., et al., 2011. The diagnosis of mild cog-
nitive impairment due to Alzheimer's disease: recommendations from the National
Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for
Alzheimer's disease. Alzheimers Dement. 7, 270–279.
Alexander, N.B., Hausdorff, J.M., 2008. Linking thinking, walking, and falling. J. Gerontol. A
Biol. Sci. Med. Sci. 63, 1325–1328.
Amboni, M., Barone, P., Hausdorff, J.M., 2013. Cognitive contributions to gait and falls:
evidence and implications. Mov. Disord. 28, 1520–1533.
Ambrose, A.F., Pau, G., Hausdorff, J., 2013. Risk factors for falls among older adults: a
review of the literature. Maturitas 75, 51–61.
Banich, M.T., 2009. Executive function: the search for an integrated account. Curr. Dir.
Psychol. Sci. 18, 89–94.
Beekman, A.T., Deeg, D.J., Van Limbeek, J., Braam, A.W., et al., 1997. Criterion validity of the
Center for Epidemiologic Studies Depression Scale (CES-D): results from a
community-based sample of older subjects in The Netherlands. Psychol. Med. 27,
231–235.
Borges, S.M., Radanovic, M., Forlenza, O.V., 2015. Fear of falling and falls in older adults
with mild cognitive impairment and Alzheimer's disease. Neuropsychol. Dev. Cogn.
B Aging Neuropsychol. Cogn. 22, 312–321.
Chen, T.Y., Peronto, C.L., Edwards, J.D., 2012. Cognitive function as a prospective predictor
of falls. J. Gerontol. B Psychol. Sci. Soc. Sci. 67, 720–728.
Craig, C.L., Marshall, A.L., Sjostrom, M., 2003. International physical activity questionnaire:
12-country reliability and validity. Med. Sci. Sports Exerc. 35, 1381–1395.
Cronin, A., O'Regan, C., Finucane, C., Kearney, P.M., et al., 2013. Health and aging: develop-
ment of the Irish longitudinal study on ageing health assessment. J. Am. Geriatr. Soc.
61, S269–S278.
Delbaere, K., Kochan, N.A., Close, J.C., Menant, J.C., Sturnieks, D.L., et al., 2012. Mild cogni-
tive impairment as a predictor of falls in community-dwelling older people. Am.
J. Geriatr. Psychiatr. 20, 845–853.
European commission, 2006. The Demographic Future of Europe — From Challenge to
Opportunity. (COM(2006)0571). EC, Brussels.
Freitas, S., Simões, M.R., Alves, L., Santana, I., 2013. Montreal Cognitive Assessment:
validation study for mild cognitive impairment and Alzheimer disease. Alzheimer
Dis. Assoc. Disord. 27, 37–43.
Gleason, C.E., Gangnon, R.E., Fischer, B.L., Mahoney, J.E., 2009. Increase risk for falling
associated with subtle cognitive impairment: secondary analysis of a randomized
clinical trial. Dement. Geriatr. Cogn. Disord. 27, 557–563.
Janssen, I., Heymsfield, S.B., Ross, R., 2002. Low relative skeletal muscle mass (sarcopenia)
in older persons is associated with functional impairment and physical disability.
J. Am. Geriatr. Soc. 50, 889–896.
Katz, S., Ford, A.B., Moskowitz, R.W., Jackson, B.A., et al., 1963. Studies of illness in the
aged. The index of ADL: a standardized measure of biological and psychosocial
function. J. Am. Med. Assoc. 185, 914–919.
Kaz Kaz, H., Johnson, D., Kerry, S., Chinappen, U., et al., 2004. Fall-related risk factors
and osteoporosis in women with rheumatoid arthritis. Rheumatology 43,
1267–1271.
Kenny, R.A., Coen, R.F., Frewen, J., Donoghue, O.A., et al., 2013. Normative values of cogni-
tive and physical function in older adults: findings from the Irish Longitudinal Study
on Ageing. J. Am. Geriatr. Soc. 61, S279–S290.
 S. Tyrovolas et al. / Experimental Gerontology 75 (2016) 42–47
Landi, F., Cruz-Jentoft, A.J., Liperoti, R., Russo, A., et al., 2013. Sarcopenia and mortality risk
in frail older persons aged 80 years and older: results from ilSirenTe Study. Age Age-
ing 42, 203–209.
Liu-Ambrose, T.Y., Ashe, M.C., Graf, P., Beattie, B.L., et al., 2008. Increased risk of falling in
older community-dwelling women with mild cognitive impairment. Phys. Ther. 88,
1482–1491.
Mayfield, D., McLeod, G., Hall, P., 1974. The CAGE Questionnaire: validation of a new
alcoholism screening instrument. Am. J. Psychiatry 131, 1121–1123.
Medina, D., DeToledo-Morrell, L., Urresta, F., Gabrieli, J.D., et al., 2006. White matter
changes in mild cognitive impairment and AD: a diffusion tensor imaging study.
Neurobiol. Aging 27, 663–672.
Montero-Odasso, M., Wells, J.L., Borrie, M.J., Speechley, M., 2009. Can cognitive enhancers
reduce the risk of falls in older people with mild cognitive impairment? A protocol for
a randomised controlled double blind trial. BMC Neurol. 9, p. 42
Mühlberg, W., Sieber, C., 2004. Sarcopenia and frailty in geriatric patients: implications for
training and prevention. Z. Gerontol. Geriatr. 37, 2–8.
Muir, S.W., Gopaul, K., Montero Odasso, M.M., 2012. The role of cognitive impairment in
fall risk among older adults: a systematic review and meta-analysis. Age Ageing 41,
299–308.
Nasreddine, Z.S., Phillips, N.A., Bédirian, V., Charbonneau, S., et al., 2005. The Montreal
Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment.
J. Am. Geriatr. Soc. 53, 695–699.
Newman, A.B., Kupelian, V., Visser, M., Simonsick, E.M., et al., 2006. Strength, but not mus-
cle mass, is associated with mortality in the health, aging and body composition
study cohort. J. Gerontol. A Biol. Sci. Med. Sci. 61, 72–77.
O'Brien, R.M., 2007. A caution regarding rules of thumb for variance inflation factors. Qual.
Quant. 41, 673–690.
Radloff, L.S., 1977. The CES-D scale: a self-report depression scale for research in the
general population. Appl. Psychol. Meas. 1, 385–401.
Robbins, A.S., Rubenstein, L.Z., Josephson, K.R., Schulman, B.L., et al., 1989. Predictors of
falls among elderly people: results of two population-based studies. Arch. Intern.
Med. 149, 1628–1633.
47
Scuffham, P., Chaplin, S., Legood, R., 2003. Incidence and costs of unintentional falls in
older people in the United Kingdom. J. Epidemiol. Community Health 57, 740–744.
Shigemori, K., Ohgi, S., Okuyama, E., Shimura, T., et al., 2010. The factorial structure of the
Mini-Mental State Examination (MMSE) in Japanese dementia patients. BMC Geriatr.
10, 36.
Shumway-Cook, A., Baldwin, M., Polissar, N.L., Gruber, W., 1997. Predicting the probability
for falls in community-dwelling older adults. Phys. Ther. 77, 812–819.
Tinetti, M.E., Doucette, J., Claus, E., Marottoli, R., 1995. Risk factors for serious injury during
falls by older persons in the community. J. Am. Geriatr. Soc. 43, 1214–1221.
Tinetti, M.E., Williams, C.S., 1997. Falls, injuries due to falls, and the risk of admission to a
nursing home. N. Engl. J. Med. 337, 1279–1284.
Tinetti, M.E., Williams, T.F., Mayewski, R., 1986. Fall risk index for elderly patients based
on number of chronic disabilities. Am. J. Med. 80, 429–434.
Tyrovolas, S., Koyanagi, A., Olaya, B., Ayuso-Mateos, J.L., et al., 2015. The role of muscle
mass and body fat on disability among older adults: a cross-national analysis. Exp.
Gerontol. 69, 27–35.
Uemura, K., Shimada, H., Makizako, H., Doi, T., et al., 2014. Effects of mild and global
cognitive impairment on the prevalence of fear of falling in community-dwelling
older adults. Maturitas 78, 62–66.
Vellas, B.J., Wayne, S.J., Romero, L., Baumgartner, R.N., et al., 1997. One-leg balance is an
important predictor of injurious falls in older persons. J. Am. Geriatr. Soc. 45,
735–738.
Whelan, B., Savva, G.M., 2013. Design and methodology of the Irish longitudinal study on
ageing. J. Am. Geriatr. Soc. 61, S265–S268.
Zimmerman, M.E., Brickman, A.M., Paul, R.H., Grieve, S.M., et al., 2006. The relationship
between frontal gray matter volume and cognition varies across the healthy adult
lifespan. Am. J. Geriatr. Psychiatr. 14, 823–833.