diabetes research and clinical practice 109 (2015) 299–305

Contents available at ScienceDirect

Diabetes Research
and Clinical Practice
journ al h ome pa ge : www .elsevier.co m/lo cate/diabres

The characteristic of cognitive function in Type 2

diabetes mellitus

Yuxia Gao a, Yanyu Xiao c, Rujuan Miao b, Jiangang Zhao c,

Wenwen Zhang b, Guowei Huang c,**, Fei Ma b,*
a
Department of Cardiology, General Hospital of Tianjin Medical University, Tianjin 300052, China
b
Department of Epidemiology and Biostatistics, School of Public Health, Tianjin Medical University, Tianjin 300070,
China
c
Department of Nutrition and Food Science, School of Public Health, Tianjin Medical University, Tianjin 30070, China

article info abstract

Article history: Aim: To identify characteristics of neuropsychological function among elderly individuals
Received 8 August 2014 with Type 2 diabetes mellitus with mild cognitive impairment (T2DM-MCI) and evaluate
Received in revised form domain-specific effects of T2DM on cognition.
2 February 2015 Methods: This was a cross-sectional study conducted in Tianjin, China. MCI subjects
Accepted 3 May 2015 (n = 246) and controls were identified in elderly individuals with diabetes, and groups were
Available online 14 May 2015 matched in a 1:1 ratio for sex, age and educational level. Cognitive function was assessed
using WAIS-III (block design, digit span), Trail Making Test A, Trail Making Test B, WMS-III
Keywords: (word list learning, logical memory), verbal fluency and MMSE. We used multivariable
Type 2 diabetes mellitus logistic regression to find diabetic factors associated with MCI.
Dementia Results: The mean MMSE score was 22.73  2.32 in subjects with T2DM-MCI, versus
Mild cognitive impairment 26.71  2.43 in subjects cognitive normal (P < 0.001). Executive and visuospatial functions
Cognition were more impaired in individuals with T2DM-MCI than in those without, as assessed using
Cross-sectional study block design (P < 0.001), digit span test (P < 0.001), and Trails B (P < 0.001). For memory,
subjects with T2DM-MCI did worse than those cognitive normal on the word learning list
delayed recall (P = 0.015). Diabetic-related factors such as longer duration of T2DM, higher
HbA1c, insulin treatment was associated with a lower level of cognitive functioning using
MMSE, block design, delayed recall and Trails B test.
Conclusions: T2DM should be considered a risk factor for MCI. This risk may be associated
with duration of diabetes, use of glucose-lowering medications, degree of glucose control.
To decrease risk of MCI, it is important to monitor glucose control and adjust medications
appropriately in elderly patients.
# 2015 Elsevier Ireland Ltd. All rights reserved.

* Corresponding author. Tel.: +86 22 83336619; fax: +86 22 83336603.

** Corresponding author. Tel.: +86 22 83336618; fax: +86 22 83336603.
E-mail addresses: huangguowei@tmu.edu.cn (G. Huang), mf2002245mf@aliyun.com (F. Ma).
http://dx.doi.org/10.1016/j.diabres.2015.05.019
0168-8227/# 2015 Elsevier Ireland Ltd. All rights reserved.
300 diabetes research and clinical practice 109 (2015) 299–305
1. Introduction
Cognition is a collective term for a range of higher brain
functions, including memory, perception, language, and rea-
soning; impaired cognitive function could be a first sign for
developing different forms of dementia. Mild cognitive im-
pairment (MCI) can occur several years before a clinical
diagnosis of dementia and increases risk of developing
dementia at an early stage [1].
Epidemiologic studies suggest that MCI may be one of the
many complications experienced by elderly persons with Type
2 diabetes mellitus (T2DM) [2–5]; T2DM doubles risk of
dementia [6]; moreover, numerous studies report T2DM is
associated with poor performance on various cognitive tests,
especially among elderly individuals. However, the underlying
cause for this association is not clear. Thus, the relationship of
MCI with T2DM has been the subject of much speculation in
recent years.
A number of studies have found that certain aspects of
cognitive function are impaired in elderly patients with T2DM
compared to age-matched non-diabetic controls [2–5,7–12].
However, a number of methodological problems, including
sample size and selection of neuropsychological tests, may
account for differences in findings across studies examining
cognition and T2DM [13,14].
In view of this, by addressing MCI in patients with T2DM,
we assessed cognitive function in people with T2DM-MCI,
described the pattern of cognitive impairment in T2DM,
sought diabetic factors that might be associated with cognitive
function, and elucidated possible mediating mechanisms
potentially underlying this association.
Fig. 1 – Flow diagram of the study population in the project.
2. Research design and methods
2.1. Study population and recruitment
In March 2010, a cross-sectional study was conducted among
community-dwelling residents aged 65 years. Using random
cluster sampling, six geographically convenient communities
with a high proportion of elderly residents who are all native
Chinese speakers were selected from the entire city of Tianjin,
China. An eligible person was aged 65 years in 2010 and self-
designated as an ethnic Han. This recruitment procedure
yielded an almost representative sample for the respective
communities. Details about the recruitment information have
been published elsewhere.
Of the original 8608 elderly subjects, 395 participants were
excluded: (1) those with difficulties in daily living activities
(n = 199), (2) those with inability to communicate with the
interviewer (n = 41), (3) those with cerebral trauma and intoxica-
tion (n = 63), and (4) those refusing to complete the tests (n = 92).
After exclusions, 8213 elderly Chinese were eligible for investi-
gation and included in the study. Each participant underwent
an in-person interview of general health and function at the
time of study entry, followed by a standard assessment,
including medical history, physical and neurological examina-
tion, and a neuropsychological test battery. Fig. 1 presents an
overview of the study design and sampling procedures.
In compliance with the Helsinki declaration, all subjects
were informed of the concept of the study and signed an
informed consent prior to their participation. The study was
approved by the Ethics Committee of the Tianjin Health
Service (protocol number: TMUhMEC201220).
 diabetes research and clinical practice 109 (2015) 299–305
2.2. Measurement of T2DM
Subjects with T2DM were selected according to World Health
Organization recommendations recently modified by the
American Diabetes Association’s Report of the Expert
Committee on the Diagnosis and Classification of Diabetes
Mellitus [15]. Participants were considered to have T2DM if
they were taking insulin (serum C-peptide determinations
were used to document Type 2 diabetes) and/or oral
hypoglycemic agents. They also were considered to have
T2DM if their fasting serum glucose level was 7.0 mmol/l
(126 mg/dl) or their 2-h value in an oral glucose tolerance test
was 11.1 mmol/l (200 mg/dl). Finally, 1109 participants
underwent diabetic testing and also completed the following
cognitive interview.
2.3. MCI diagnostic criteria
According to the Petersen’s diagnostic standard on MCI [16],
results from the neurologic, psychiatric, and neuropsycholog-
ical examinations were reviewed in a consensus conference
comprising physicians, neurologists, neuropsychologists,
and psychiatrists. According to this review, 1109 eligible
subjects with T2DM were evaluated for MCI based on the
following criteria for MCI including (1) memory complaint,
preferably corroborated by an informant; (2) objective
memory impairment adjusted for age and education level;
(3) preservation of general cognitive functioning; (4) no or
minimal impairment of daily life activities; and (5) not
fulfilling the revised Diagnostic and Statistical Manual of
Mental Disorders, 3rd edition (DSM-III-R) criteria of dementia
[17]. Of 1109 diabetic subjects, 690 subjects had MCI, and 287
had normal cognition; the remaining subjects suffered from
dementia.
2.4. Matching cases and controls
For each eligible case, a control subject matched for sex, birth
year, education (<10 years or 10 years of education) was
randomly selected from the subjects with normal cognition.
Controls were assigned the same index date as their
corresponding case, and each control was included only
once. In order to justify how many cases and controls are
necessary, We suppose H0: OR = 1 at a = 0.05 (two-sided) with
power = 80% using a 1:1 ratio of cases to controls while
looking for an odds ratio of 2. The prevalence of exposure in
the source population (controls) is 25% [18]. EpiCalc 2000
determines the Basic Sample Size 179 cases and 179 controls
in the current study. Among the 1109 diabetic subjects, 246
pairs were selected to participate in the assessment reported
here and formed the final study group. Analyses are based on
the match-paired groups.
2.5. Data collection
Interviews and assessments were conducted in the parti-
cipants’ homes over the course of three home visits. The
neuropsychological battery was accomplished in 1.5–2.0 h
during the first home visit. At the second home visit, blood
was drawn with participants having fasted for at least 7 h.
301
During the third home visit, information on medical
conditions and medication usage was obtained. Neuropsy-
chological testing was administered by trained and super-
vised research assistants who were blind to subject
diagnoses.
2.6. Cognitive assessment
Research assistants, trained by a board-certified neuro-
psychologist, must be fluent in oral English and Chinese
mandarin. There are a number of confounding demographic
factors that can influence this equation including age,
education, gender, and race/ethnicity. Research assistants
can use standard Chinese mandarin to administer the
cognitive tests. Eligible matched pairs were subsequently
examined using the following neuropsychological battery.
(1) Wechsler Adult Intelligence Scale (WAIS)-III Block Design. This
test requires construction of abstract designs from colored
blocks. The total raw score was recorded to assess
visuospatial and executive functions. All WAIS-III subtests
were scored according to the WAIS-III Manual [19].
(2) Trail Making Test A. This is another measure of attention
and information processing speed with a visuomotor
component requiring rapidly connecting consecutively
numbered circles. The time to connect numbers was
recorded, with a maximum time of 301 s [20].
(3) Trail Making Test B. This is a measure of cognitive flexibility
and complex conceptual tracking requiring rapidly con-
necting numbered and lettered circles in an alternating
sequence. The subject was asked to alternate between
numbers and letters, with a maximum time of 301 s. This
test was used to measure executive function [20].
(4) WAIS-III Digit Span. This requires repeating digits forward
and in reverse sequence; a total raw score was recorded.
This test was used to evaluate attention/concentration,
working memory, and executive function.
(5) Weschler Memory Scale (WMS)-III Word List Learning. A list of
12 words was read to the subject four times, and the
subject was asked to recall the list of words each time. The
recall total score was calculated for immediate recall. After
30 min, the subject was asked to recall the same list of
words again; this comprised the delayed recall score. The
delayed score was divided by the number of words the
subject learned by the fourth time to calculate percentage
retention. These tests were used to measure verbal
learning and memory.
(6) WMS-III Logical Memory. Two stories were read aloud to the
subject; the subject was then asked to repeat each story for
immediate recall. After 30 min, the subject was asked to
repeat both stories; this comprised the delayed recall score.
The ratio of delayed recall score to immediate recall score
was used to calculate percentage retention. These tests
measured a different aspect of memory from word list
learning.
(7) Verbal fluency (controlled oral word association test). The
subject was asked to give words beginning with three
different letters, and the total number of correct responses
was recorded. This test was used to measure language
ability related to executive function.
302 diabetes research and clinical practice 109 (2015) 299–305

2.7. Collection of other information
Interviews were administered to obtain information on socio-
demographic characteristics, family composition, lifestyle
(smoking, consumption of alcohol, and use of drugs), duration
of diabetes, and treatment of diabetes. Past and current
cigarette smoking status was self-reported. Weight and height
were measured using standardized instruments and protocols
during the home visit. Body mass index (BMI) was calculated as
measured weight in kilograms divided by measured height in
meters squared. Blood pressure was assessed twice at the home
visit, and the averages of the two measurements for systolic and
diastolic blood pressure were used. Systolic pressure > 140 mm
Hg, diastolic pressure > 90 mm Hg, or reported use of hyper-
tension medication was defined as hypertension. Severe
hypoglycaemia was defined as self-reported episodes of
hypoglycaemia requiring external help. Self-reported history
of severe hypoglycemia (history of hypoglycemia) was recorded
including the number of episodes experienced in total. Reports
of milder episodes of hypoglycaemia (those not requiring
external assistance for treatment) were not collected, as the
validity of self-reports of such episodes is questionable [21,22].
All subjects, except those receiving insulin, underwent an
oral glucose tolerance test with a 75 g glucose load. Venous blood
samples for the determination of plasma glucose and insulin
levels were taken 1 and 2 h after the glucose load. Glucose levels
were determined from samples by the glucose oxidase method
(Glucose Auto & Stat HGA 1120 Analyzer, Daichii Kyoto, Japan).
Insulin levels were determined by radioimmunoassay (Phase-
deph Insulin RIA 100, Pharmacia, Uppsala, Sweden). Glycated
hemoglobin (HbA1c) was measured using high performance
liquid chromatography (A. Menarini Diagnostics, Firenze, Italy)
based on an ion-exchange reverse phase partition method. The
genetic status of apolipoprotein E (ApoE) alleles was character-
ized using polymerase chain reaction.
2.8. Statistic analysis
Statistical analysis was performed using SPSS version 11.0
(SPSS Inc., Chicago, IL, USA). All continuous data were tested
for normality with the Kolmogorov–Smirnov test. Comparison
between groups (variables) was performed with an indepen-
dent t-test (normally distributed) and Mann–Whitney U test
(non-normally distributed) for continuous data. The chi-
square test was used for comparison of categorical variables
(for expected frequencies [i.e. cell counts <5] we used Fisher’s
exact test), when appropriate. A P-value <0.05 was considered
to be significant. Multiple logistic regression analysis was used
to examine association between diabetic factors and MMSE,
Trails B, digit span, and block design.
3. Results
According to MCI diagnostic criteria, 246 pairs of diabetic MCI
subjects and cognitively normal controls were identified and
matched in a 1:1 ration for sex, age, and educational level.
Mean age (MCI, 73.92 years [SD 2.17] vs. control, 74.15 [SD
2.43]), the average years of formal education (10.00 years [SD
3.92] vs. 10.00 years [SD 3.87], respectively), and female
proportion of the sample (14.72% vs. 14.63%, respectively)
were nearly identical in both groups.
The demographic variables are shown in Table 1. Com-
pared with cognitively normal subjects, subjects with MCI

Table 1 – Demographic, health, and diabetes-related data for T2DM-MCI and T2DM-normal cognition.
Characteristics T2DM-MCI (n = 246) T2DM-normal cognition (n = 246) t P-value
c
Duration from onset of diabetes (years) 12.31  3.72 10.91  4.32 3.478 <0.001
BMI (kg/m2)a 28.72  4.86 29.43  4.17 1.739 0.051
Systolic blood pressure (mm Hg) 138.52  23.27 138.17  22.35 0.195 0.860
Diastolic blood pressure (mm Hg) 76.43  13.15 75.92  12.77 0.430 0.475

Fasting plasma insulin (pmol/l) 129.82  12.31 92.93  14.75 30.125 <0.001
2
Characteristics T2DM-MCI T2DM-normal cognition x P-value
c
Current smoker (%) 36.18 32.52 0.904 0.296
User of alcohol (>30 g/week) (%)b,c 63.01 56.10 2.437 0.178
ApoE e4 (%) 28.86 19.92 5.334 0.019
HbA1c (%) 10.26 0.59 7.09  0.53 44.329 <0.001

No. of hypoglycaemic episodes

0 86.27 91.82 4.077 0.037
1 11.23 5.74 5.102 0.022
2 4.73 1.35 5.570 0.014
3 or more 2.36 1.09 5.539 0.017

Diabetic treatment: n (%)c

Diet 16.72 18.23 0.713 0.483
Oral hypoglycemic agent 53.33 57.32 0.703 0.503
Insulin 29.95 24.45 5.379 0.017

Values are means + SD or value (percentage) Student’s t-test or the chi-square test was used when appropriate.
a
BMI (weight in kg divided by squared height in meter).
b
Alcohol: grams of pure alcohol consumed calculated from glasses of beer, wine, and spirits regularly consumed per week.
c
Variables derived from self-reported data.
 diabetes research and clinical practice 109 (2015) 299–305 303

were more likely to have a longer duration from onset of
diabetes; higher frequency of hypoglycaemic episodes; higher
levels of glucose, insulin, HbA1c, and ApoE e4; and to use
insulin (all P < 0.05). Important non-significant variables
included BMI, current smoker, user of alcohol, and systolic
and diastolic blood pressures (all P > 0.05).
Table 2 shows that the mean MMSE score was 22.73  2.32
in subjects with T2DM-MCI, compared to 26.71  2.43 in
subjects with normal cognition (P < 0.001). Using a more
sensitive test in the neuropsychological examination, subjects
with T2DM-MCI had significantly lower scores on the block
design test (MCI, 16.97  8.31 vs. control, 21.24  7.45,
P < 0.001). Trails B test was also used to evaluate executive
function. Subjects with T2DM-MCI needed longer to complete
Trails B than those with normal cognition (171.82  23.73 vs.
158.76  31.57, P < 0.001, respectively). Those with T2DM-MCI
had significantly lower scores than cognitively normal con-
trols on the digit span test (13.28  3.19 vs. 14.30  2.17,
P < 0.001, respectively), which tests for executive function and
working memory. For memory, subjects with T2DM-MCI had
lower scores than those with normal cognition on the word list
learning delayed recall (3.01  2.87 vs. 3.6 6 2.79, P = 0.011,
respectively), although no statistically significant differences
were found between groups on the word list immediate recall
or the story recall tasks. There was no difference in verbal
fluency, a measurement of language ability, between T2DM
subjects with and without MCI.
To further understand the relationship between T2DM and
cognitive impairment, multivariate regression was undertak-
en to analyze potential confounding factors associated with
each individual outcome. Factors included were based on a
univariate regression analysis and clinical knowledge and
included MMSE, Trails B, digit span, and block design. Table 3
shows that performance on the MMSE, Trails B, digit span, and
block design was lower with increasing duration of T2DM. All
outcomes were associated with a higher HbA1c level and
insulin treatment. As expected, ApoE e4 status did not
significantly affect scores for cognitive outcomes, nor did
BMI or glucose concentration.
After adjusting for confounders, diabetic factors such as
longer duration of T2DM, higher HbA1c, higher frequency of
hypoglycaemic episodes, insulin treatment was associated
with a lower level of cognitive functioning using MMSE, Trails
B, digit span, and block design.
4. Discussion
Several cross-sectional and prospective studies have shown
that T2DM may increase the risk of dementia, including
Alzheimer’s disease and vascular dementia. However, there is
a lack of strong evidence explaining the relationship between
T2DM and cognitive function in the absence of dementia or at
an early stage of dementia [23]. Some large population-based
studies have demonstrated that T2DM is associated with
cognitive impairment, whereas others have reported no such
relationship [14]. In the current study, when the sensitive
neuropsychological tests were applied in subjects with T2DM,
more specific cognitive impairments that were not indicated
by the MMSE were exposed.
In this study, all subjects with a diagnosis of dementia
according to DSM-III-R criteria were excluded, whereas those
individuals with MCI, i.e., those who did not yet meet the
criteria for dementia, were included. Our results showed
impaired performance with T2DM-MCI in MMSE, Trails B, digit
span, and block design tests than in the cognitively normal
control group (Table 2). Despite the ceiling effect of Trails B,
subjects with T2DM-MCI spent more time to complete it than
did controls, which was a more conservative evaluation. Of
memory tests, immediate recall (word list and story) and the
delayed logical memory test did not show a difference
between groups (Table 2).
That older people with MMSE scores 10 were excluded
may explain why significant visuospatial/executive dysfunc-
tion but not severe memory impairment was observed in those
with T2DM-MCI, as was observed in the other studies [24,25].
In summary, these combined results suggest that a pattern of
cognitive dysfunction (significant executive and visuospatial

Table 2 – Results of global cognitive functioning tests and neuropsychological tests in different cognitive domains in MCI
subjects and cognitively normal controls among diabetic patients.
Neuropsychological tests T2DM-MCI (n = 246) T2DM-normal cognition (n = 246) t P-value
MMSE 22.73  2.32 26.71 + 2.43 48.977 <0.001
Block design 16.97  8.31 21.24  7.45 6.065 <0.001
Trail Making test A (s) 58.72  13.73 57.43  15.87 0.731 0.135
Trail Making test B (s) 171.82  23.73 158.76  31.57 5.212 <0.001
Digit span 13.28  3.19 14.30  2.17 4.608 <0.001

Word learning list

Immediate recall 22.99  6.57 23.97  6.2 1.736 0.065
Delayed recall 3.01  2.87 3.66  2.79 2.419 0.015

Logical memory
Immediate recall, 33.14  10.66 33.72  10.58 0.631 0.651
Delayed recall 17.37  8.97 18.74  9.08 0.735 0.539
Verbal fluency 24.85 (%) 26.86 (%) 0.265 0.613
Values are means  SD.
MMSE, mini-state mental examination.
304 diabetes research and clinical practice 109 (2015) 299–305
Table 3 – Results of multiple logistic regression analysis.
Characteristic
MMSE
Duration of T2DM 0.09 (0.02) <0.001
HbA1c (%) 0.13 (0.01) <0.001
Self-reported hypoglycaemia 0.13 (0.02) <0.001
episode (%)
Diabetic treatment: n (%)
Diet 0.36 (0.54) 0.327
Oral hypoglycemic agent 0.32 (0.50) 0.267
Insulin (with or without oral 1.09 (0.41) 0.012
hypoglycemic agent)
Note: All variables were included in the regression model after adjusting for ApoE e4, BMI, systolic blood pressure, diastolic blood pressure,
current smoker, and user of alcohol.
dysfunction) but less memory impairment is associated with
T2DM. It is possible that visuospatial and executive dysfunc-
tion precedes memory decline in T2DM. This pattern fits the
clinical picture of frontal-subcortical syndrome, which is seen
in microvascular disease of the brain.
After adjustment for potential confounders, multiple
logistic regression analyses revealed that diabetic factors,
including duration of T2DM, HbA1c levels, frequency of
hypoglycaemic episodes and insulin treatment remained
significantly associated with a decline in general cognition,
executive and visuospatial function, and memory in commu-
nity-dwelling elderly individuals. A high insulin level, but not
an elevated glucose level, was associated with poor memory
function, which is consistent with other studies [26–29].
Epidemiological and laboratory studies have demonstrated
that hypoglycemia results in neuroglycopenia with subse-
quent cognitive impairment [30–32]. But the direction of
association between episodes of severe hypoglycemia and
T2DM-MCI remains uncertain. The current study shows that
self-reported history of severe hypoglycaemia was associated
with T2DM-MCI independent of diabetes duration and vascu-
lar disease. Potential mechanisms by which severe hypogly-
caemia could cause cognitive impairment include neuronal
cell damage, vascular insufficiency via increased platelet
aggregation and fibrinogen formation, impaired nutrient
delivery to the brain, increased accumulation of neurotoxin
and impaired permeability of blood–brain barrier [33,34]. The
association of severe hypoglycaemia with T2DM-MCI requires
confirmation in prospective studies using measures of actual
cognitive change over time, and the direction of the associa-
tion has yet to be determined. Confirmation of a true causal
relationship between exposure to severe hypoglycaemia and
subsequent cognitive impairment will have important impli-
cations for the management of older people with T2DM, in
whom highly intensive glycaemic control, using anti-diabetic
agents that promote hypoglycaemia, may be contraindicated.
The present study and a previous one indicate that T2DM
itself, rather than preclinical syndrome or complications of it,
is related to cognitive impairment. Brain function in non-
demented patients with non-insulin-dependent diabetes may
be mildly affected. In our study, memory dysfunction was not
present. An elderly subject with T2DM-MCI, who does not
show signs of overt dementia, is unlikely to suffer significant
memory dysfunction when compared with cognitively normal
b Estimate (standard error) P-value
Block design Digit span Trails B
0.14 (0.03) 0.012 0.82 (0.21) <0.001 3.59 (0.54) <0.001
0.38 (0.03) <0.001 0.67 (0.13) <0.001 2.77 (0.33) <0.001
0.34 (0.04) <0.001 18.34 (11.15) 0.086 0.54 (0.15) <0.001
0.72 (1.24) 0.539 7.02 (4.27) 0.083 18.72 (9.13) 0.066
0.73 (1.22) 0.528 7.08 (4.09) 0.107 17.23 (10.13) 0.087
3.17 (1.25) 0.011 10.37 (3.97) 0.008 33.21 (9.57) 0.003
subjects. Diabetic factors, including longer duration of T2DM,
higher HbA1c, and insulin treatment are associated with
impaired cognitive function in older T2DM subjects; however,
these impairments may be of only minor significance in daily
life.
Confidence in these findings is strengthened by three
factors. First, it is the first study to explore the association of
T2DM and cognitive impairment in a representative sample of
the Chinese elderly population. Criteria for disease diagnosis
were standardized, and all participants underwent structured
clinical evaluations to identify persons with diabetes and
other major chronic conditions. Second, using a large
cognitive battery, we assessed a broad range of cognitive
domains (e.g., global, attention, and memory tasks). Detailed
neuropsychological data were available and summarized to
yield previously established composite measures of five
cognitive domains and global cognition. Third, we controlled
for most of the traditional behavioral, demographic, biological,
and physical risk factors for cognitive impairment.
This study also has several limitations. First and impor-
tantly, this was a cross-sectional study, and a causal
relationship between T2DM and cognitive function cannot
be concluded. Second, the analysis did not include magnetic
resonance imaging or autopsy, so the relationship between
brain pathology and diabetes mellitus cannot be causally
linked. These factors may bias results toward the null and
mask the modest associations of diabetes to specific cognitive
domains.
Longitudinal data from a large group demographically and
clinically representative of the older population is needed to
examine possible underlying neurobiological mechanisms
linking diabetes to impaired cognition and contribute to a
better understanding of these associations.
In summary, MCI in elderly individuals with T2DM is
associated with the duration of diabetes, use of glucose-
lowering medications, and degree of glucose control. It is
therefore important to monitor glucose levels in elderly
patients to decrease risk of MCI.
Conflict of interest
No potential conflicts of interest relevant to this article were
reported.
 diabetes research and clinical practice 109 (2015) 299–305
Acknowledgments
This study was supported by a National Natural Science
Foundation of China (grant number: 81130053). The authors
thank Prof. Wang Jianhua and Liu Xinmin for helpful
discussions during the preparation of this manuscript.
references
[1] Evans JG, Sastre AA. Effect of the treatment of Type II
diabetes mellitus on the development of cognitive
impairment and dementia. Cochrane Database Syst Rev
2002;4:CD003804.
[2] Arvanitakis Z, Wilson RS, Li Y, Aggarwal NT, Bennett DA.
Diabetes and function in different cognitive systems in
older individuals without dementia. Diabetes Care
2006;29:560–5.
[3] Luchsinger JA, Reitz C, Patel B, Tang MX, Manly JJ, Mayeux
R. Relation of diabetes to mild cognitive impairment. Arch
Neurol 2007;64:570–5.
[4] Logroscino G, Kang JH, Grodstein F. Prospective study of
type 2 diabetes and cognitive decline in women aged 70–81
years. BMJ 2004;328:548.
[5] Yaffe K, Blackwell T, Kanaya AM, Davidowitz N, Barrett-
Connor E, Krueger K. Diabetes, impaired fasting glucose,
and development of cognitive impairment in older women.
Neurology 2004;63:658–63.
[6] Cukierman T, Gerstein HC, Williamson JD. Cognitive
decline and dementia in diabetes – a systemic overview of
prospective observational studies. Diabetologia
2005;42:2460–5.
[7] Croxson SC, Jagger C. Diabetes and cognitive impairment: a
community-based study of elderly subjects. Age Aging
1995;24:421–4.
[8] Gregg EW, Yaffe K, Cauley JA, Rolka DB, Blackwell TL,
Narayan KM, et al. Is diabetes associated with cognitive
impairment and cognitive decline among older women?
Arch Intern Med 2000;160:174–80.
[9] Fontbonne A, Ducimetiere P, Berr C, Alperovitch A. Changes
in cognitive abilities over a 4-year period are unfavorably
affected in elderly diabetic subjects. Diabetes Care
2001;24:366–70.
[10] Munshi M, Grande L, Hayes M, Ayres D, Suhl E, Capelson R,
et al. Cognitive dysfunction is associated with poor
diabetes control in older adults. Diabetes Care
2006;29:1794–9.
[11] Arvanitakis Z, Wilson RS, Bennett DA. Diabetes mellitus,
dementia, and cognitive function in older persons. J Nutr
Health Aging 2006;10:287–91.
[12] Okereke OI, Kang JH, Cook NR, Gaziano JM, Manson JE,
Buring JE, et al. Type 2 diabetes mellitus and cognitive
decline in two large cohorts of community dwelling older
adults. J Am Geriatr Soc 2008;56:1028–36.
[13] Sinclair AJ, Girling AJ, Bayer AJ. Cognitive dysfunction in
older subjects with diabetes mellitus: Impact on diabetes
self-management and use of care services. All Wales
Research into Elderly (AWARE) Study. Diabetes Res Clin
Pract 2000;50:203–12.
[14] Atiea JA, Moses JL, Sinclair AJ. Neuropsychological function
in older subjects with non-insulin-dependent diabetes
mellitus. Diabetes Med 1995;12:679–85.
305
[15] The Expert Committee on the Diagnosis and Classification
of Diabetes Mellitus. Report of the Expert Committee on the
Diagnosis and Classification of Diabetes Mellitus. Diabetes
Care 1997;20:1183–97.
[16] Petersen RC, Doody R, Kurz A, Mohs RC, Morris JC, Rabins
PV, et al. Current concepts in mild cognitive impairment.
Arch Neurol 2001;58:1985–92.
[17] Association American Psychiatric. Diagnostic and
statistical manual of mental disorders. 3rd ed. revised
Arlington: American Psychiatric Association; 1987.
[18] Fei M, Yan Ping Z, Ru Juan M, Ning Ning L, Lin G. Risk
factors for dementia with type 2 diabetes mellitus among
elderly people in China. Age Ageing 2013;42:398–400.
[19] Wechsler D, Wechsler. Adult intelligence scale. 3rd ed. San
Antonio, TX: The Psychological Corporation; 1997.
[20] Lezak MD. Neuropsychological assessment. 3rd ed. New
York: Oxford University Press; 1995.
[21] Amiel SA, Dixon T, Mann R, Jameson K. Hypoglycaemia in
Type 2 diabetes. Diabetes Med 2008;25:245–54.
[22] Akram K, Pedersen-Bjergaard U, Carstensen B, Borch-
Johnsen K, Thorsteinsson B. Prospective and retrospective
recording of severe hypoglycaemia, and assessment of
hypoglycaemia awareness in insulin-treated Type 2
diabetes. Diabetes Med 2009;26:1306–7.
[23] Coker LH, Shumaker SA. Type 2 diabetes mellitus and
cognition: an understudied issue in women’s health. J
Psychosom Res 2003;54:129–39.
[24] Ryan CM, Geckle M. Why is learning and memory
dysfunction in type 2 diabetes limited to older adults?
Diabetes Metab Res Rev 2000;16:308–15.
[25] Messier C, Awad N, Gagnon M. The relationships between
atherosclerosis, heart disease, type 2 diabetes and
dementia. Neurol Res 2004;26:567–72.
[26] Lowe LP, Tranel D, Wallace RB, Welty TK. Type II diabetes
and cognitive function. A population-based study of Native
Americans. Diabetes Care 1994;17:891–6.
[27] Elias PK, Elias MF, D’Agostino RB, Cupples LA, Wilson PW,
Silbershatz H, et al. NIDDM and blood pressure as risk
factors for poor cognitive performance. The Framingham
Study. Diabetes Care 1997;20:1388–95.
[28] Grodstein F, Chen J, Wilson RS, Manson JE, Nurses’ Health
Study. Type 2 diabetes and cognitive function in
community-dwelling elderly women. Diabetes Care
2001;24:1060–5.
[29] Wu JH, Haan MN, Liang J, Ghosh D, Gonzalez HM, Herman
WH. Impact of diabetes on cognitive function among older
Latinos: a population-based cohort study. J Clin Epidemiol
2003;56:686–93.
[30] Suh SW, Hamby AM, Swanson RA. Hypoglycemia, brain
energetics, and hypoglycaemic neuronal death. Glia
2007;55:1280–6.
[31] Bree AJ, Puente EC, Daphna-Iken D, Fisher SJ. Diabetes
increases brain damage caused by severe hypoglycemia.
Am J Physiol Endocrinol Metab 2009;297:E194–201.
[32] Ryan CM, Geckle M. Why is learning and memory
dysfunction in Type 2 diabetes limited to older adults?
Diabetes Metab Res Rev 2000;16:308–15.
[33] Wright RJ, Frier BM. Vascular disease and diabetes: is
hypoglycaemia an aggravating factor? Diabetes Metab Res
Rev 2008;24:353–63.
[34] Kristensen PL, Hoi-Hansen T, Boomsma F, Pedersen-
Bjergaard U, Thorsteinsson B. Vascular endothelial growth
factor during hypoglycemiain patients with type 1 diabetes
mellitus: relation to cognitive function and renin–
angiotensin system activity. Metabolism 2009;58:1430–8.