Disponible en ligne sur

www.sciencedirect.com

Annales d’Endocrinologie 73 (2012) 208–212

Original article

Risk factors of mild cognitive impairment in middle aged patients with type
2 diabetes: A cross–section study
Facteurs de risque du déficit cognitif léger chez les sujets d’âge moyen atteints de diabète de type 2 :
une étude transversale
Rui-Hua Chen a , Xiao-Zhen Jiang a,∗ , Xiao-Hui Zhao b , Yu-Lan Qin a , Zhe Gu a , Pei-Li Gu a ,
Bin Zhou a , Zhen-Hong Zhu a , Lin-Yan Xu a , Yu-Feng Zou a
a Department of Endocrinology, Pudong New Area People’s Hospital, Chuan Huan Nan Road No. 490, Shanghai, China
b Department of Neurology, Pudong New Area People’s Hospital, Shanghai, China

Abstract
The aim of this study was to evaluate the risk factors of mild cognitive impairment (MCI) in middle-aged patients with type 2 diabetes (T2DM).
Methods. – Montreal Cognitive Assessment (MoCA) was applied as cognition assessment implement. One hundred and fifty-seven middle-aged
type 2 diabetic patients were enrolled in this cross-section study (age 40∼69, mean age 55 ± 7). There were 93 patients with MCI (MoCA score < 26)
in MCI group and 64 with normal cognitive function (MoCA score ≥ 26) in control group. Information of history of disease, family history, data
of BMI, WHR, HbA1c, FINS, C-Peptide (C-P), SBP, DBP, blood lipid (TG, TC, LDL-C, HDL-C and carotid ultrasound (carotid IMT, carotid
resistance index [RI]) was collected. Results. – There were significant differences in the rate of patients with hypertension ([40.63 vs. 58.06%],
P = 0.026), duration of diabetes mellitus ([3.09 ± 4.04 y vs. 4.80 ± 4.94 y], P = 0.024), C-P ([2.79 ± 1.09 ng/ml vs. 2.26 ± 1.00 ng/ml], P = 0.008),
Max C-IMT ([0.81 ± 0.15 mm vs. 0.91 ± 0.15 mm], P < 0.001), Min C-RI (0.71 ± 0.06 vs. 0.68 ± 0.06, P < 0.05), and no significant differences in
the duration of hypertension and hyperlipidemia, BMI, WHR, HbA1c, SBP, DBP and blood lipid between control group and MCI group. MoCA
scores were positively correlated with C-P (r = 0.252, P = 0.005), and negatively correlated with the history of hypertension (r = −0.244, P = 0.002),
duration of DM (r = −0.161, P = 0.044), Max C-IMT (r = −0.253, P = 0.005) and Min C-RI (r = −0.183, P = 0.023). Multiple regression analysis
showed that history of hypertension (Beta = −0.267, P = 0.002), C-P (Beta = 0.281, P = 0.001) and Min C-RI (Beta = −0.221, P = 0.011) were
significantly independent determinants for the MoCA scores. Conclusions. – The longer duration of diabetes, history of hypertension, lower serum
C-P levels, thickened C-IMT and higher C-RI could be risk factors of MCI in type 2 diabetic patients. This finding could have an important impact
on the management of cognitive decline in diabetic patients.
© 2012 Elsevier Masson SAS. All rights reserved.
Résumé
L’objectif de cette étude est d’évaluer les facteurs de risque du déficit cognitif léger (mild cognitive impairment [MCI]) chez les sujets d’âge
moyen atteints de diabète de type 2. Méthodes. – Mesure du niveau de dysfonction cognitive grâce à la cotation du Montreal Cognitive Assessment
(MoCA). Étude transversale de 157 patients âgés de 40 à 69 ans (moyenne 55 ± 7) présentant un diabète de type 2. Il y avait 93 patients MCI
(MoCA < 26) et un groupe témoin de 64 patients sans dysfonction cognitive (MoCA ≥ 26) Paramètres étudiés : histoire de la maladie, antécédents
familiaux, BMI, WHR, HbA1c, Fins, CRP, SBP, DBP, bilan lipidique (TG, CT, LDL-C, HDL-C, échographie carotidienne (mesure de l’épaisseur
intima-média carotidienne [C-IMT] et de l’indice de résistance carotidienne [C-RI]). Résultats. – En comparant les sujets MCI au groupe témoin, les
résultats montraient des différences significatives pour les paramètres suivants : hypertension artérielle ([40,63 % vs 58,06 %], p = 0,026), durée de
la maladie diabétique ([3,09 ± 4,04 ans vs 4,80 ± 4,94 ans], p = 0,024), CRP ([2,79 ± 1 ng/mL, 09 vs 2,26 ± 1,00 ng/mL], p = 0,008), max C-IMT
([0,81 ± 0,15 mm vs 0,91 ± 0,15 mm], p < 0,001), min C-RI (0,71 ± 0,06 vs 0,68 ± 0,06, p < 0,05). Aucune différence n’était retrouvée pour la
durée de l’hypertension artérielle et l’hyperlipidémie, BMI, WHR, HbA1c, SBP, DBP et le bilan lipidique. Il y avait une corrélation positive entre
le MoCA et le CRP (r = 0,252, p = 0,005) et une corrélation négative entre le MoCA et une histoire d’hypertension (r = −0,244, p = 0,002), durée
du diabète (r = −0,161, p = 0,044), max C-IMT (r = −0,253, p = 0,005) et min C-RI (r = −0,183, p = 0,023). Selon l’analyse multifactorielle, une
histoire d’hypertension (Bêta = −0,267, p = 0,002), CRP (Bêta = 0,281, p = 0,001) et min C-RI (Bêta = −0,221, p = 0,011) étaient des déterminants
indépendants de la mesure de la dysfonction cognitive (MoCA). Conclusions. – Une durée plus longue de la maladie diabétique, une histoire
∗ Corresponding author.
E-mail address: [email protected] (X.-Z. Jiang).

0003-4266/$ – see front matter © 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ando.2012.04.009
 R.-H. Chen et al. / Annales d’Endocrinologie 73 (2012) 208–212 209

d’hypertension, une CRP plasmatique basse, une épaisseur intima-média carotidienne et un indice de résistance carotidienne élevés seraient des
facteurs indépendants du déficit cognitif léger chez les sujets présentant un diabète de type 2. Ces résultats pourraient avoir un impact important
sur la prise en charge du déclin cognitif des diabétiques.
© 2012 Elsevier Masson SAS. Tous droits réservés.

1. Introduction disease, hydrocephalus, brain neoplasm, epilepsy, multiple scle-

rosis, chronic subdural hematoma, severe head trauma, abnormal
Type 2 diabetes (T2DM) is associated with many chronic brain structure, schizophrenia, etc.); history of alcohol abuse or
complications including stroke and cardiovascular disease. In drug abuse; history of depression and took antidepressants from
additional, diabetes may also contribute to the development of 6 months before the recruitment; those who received medicines
cognitive dysfunction [1,2]. It has been reported that T2DM which can affect cognitive function 1 months before the study
is associated with decreases in psychomotor speed [3], verbal (e.g.: steroid, antiepileptic drug, sedative-hypnotic drug, anes-
memory [4], immediate recall, delayed recall [5], and verbal thetics, etc.).
fluency [5]. Research has been focusing on the study of mild
cognitive impairment (MCI), which is an intermediate clini- 2.2. Methods
cal condition between normal aging and dementia [6]. MCI
is defined by an impairment of memory and other cognitive Demographic characteristics and medical history: datas were
functions, not so serious as dementia [7]. collected including age, gender, years of education, duration of
In this cross-sectional study, we compare cases of MCI and diabetes, history and duration of hypertension and dyslipidemia,
normal controls (NC) of middle-aged diabetes adults in Pudong body mass index (BMI), waist hip ratio (WHR). Systolic blood
New Area People’s Hospital, Shanghai, China, in order to find pressures (SBP), diastolic blood pressure (DBP) were measured.
out risk factors of MCI in such population. Blood samples were obtained at fasting, Fasting glucose (FPG),
triglyceride (TG), total cholesterol (TC), low-density cholesterol
2. Subjects and methods (LDL-C), high-density cholesterol (HDL-C) (Cobas 8000 C701
C502 auto chemistry analyzer, Roche, enzyme method), HbA1c
2.1. Subjects (HLC-723G7 analyzer, TOSOH kabushiki kaisha, HPLC), fast-
ing insulin (FINS) and C-peptide (C-P) (Cobas 6000 analyzer,
A cross-sectional study was conducted in patients with a his- Roche, electrochemiluminescence immunoassay), were deter-
tory of T2DM recruited from the outpatient clinic of Endocrine mined using standard laboratory procedures.
department in Pudong New Area People’s Hospital. The study
group comprised of 157 subjects belonging to both genders in 2.2.1. Cognitive measures
the age group 40–69 years, and diabetes duration (0–20 years). MoCA scale was applied to assess the cognitive situation
The study group was divided into two groups (MCI Group: of all subjects. Here we give a detailed introduction to MoCA
MoCA score < 26, n = 93, normal cognitive function (NC) group: Beijing version. The final English version of the MoCA is a one-
MoCA score ≥ 26, n = 64). There were no significant differ- page 30-point 10-minute screening test to identify elderly people
ences in sex (male/female = 37/27 vs. 44/49, P > 0.05), mean age with MCI [8]. The MoCA Beijing version used in this study
([55.37 ± 7.15 years old vs. 55.33 ± 7.8.31 years old], P > 0.05) is based on MoCA and contains some cultural and linguistic
and years of education ([10.16 ± 3.07 vs. 9.21 ± 3.42 years] changes but carries the same meaning as the English version
P > 0.05) in both groups. Data collection started in December [9]. The instructions are:
2010 and completed in November 2011. Cognitive evaluation
was performed on all the subjects by using Montreal Cogni- • add an extra point if the individual has 12 years or fewer of
tive Assessment (MoCA) Beijing version. All subjects had the formal education;
ability to response the questionnaires and receive all the clini- • operators need to give all patients similar commands as far as
cal and blood test and inclusion criteria of MCI group were as possible to reduce errors and ensure that subjects do not get
follows: Complaints of hypomnesis; MoCA score < 26; Clini- any hint from the commands;
cal dementia rating (CDR) ≥ 0.5. Inclusion criteria of NC group • and control operating time strictly. In order to reduce the
was MoCA score ≥ 26. Exclusion criteria were as follows: those experimental error, the results are excluded if the comple-
with diabetic ketoacidosis or other acute diabetic complications tion time of the MoCA is longer than 20 minutes. Patients
in recent 3 months, with severe heart failure, chronic renal fail- with scores less than 26 were determined as MCI.
ure, lung disease, or had the history of infection in central
nervous system, stroke, cerebral hemorrhage or other clinical 2.2.2. Ultrasonography of carotid arteries
evidences of central nervous damages; persons with history of Detailed B-mode images of the right and left common carotid
auditory disorders and psychological disturbances, which might artery (CCA), carotid bifurcation, and the first 1.5 cm of the inter-
interfere with the MoCA test; patients with history of chronic nal carotid artery (ICA) were obtained at each ultrasound visit
disease which could cause dementia (including Alzheimer syn- using a Sequoia scanner (SIEMENS, Germany) equipped with a
drome, Parkinson syndrome, vascular dementia, Huntington’s 4–6 MHz linear array imaging probe. IMT and resistance index
210 R.-H. Chen et al. / Annales d’Endocrinologie 73 (2012) 208–212

(RI) of both sides of CCA were recorded. To measure the aver- Table 2
age IMT of each segment, lines were electronically drawn along Cinical characteristics in NC group and MCI group.
1-cm segments of the lumen-intima interface and the media- NC MCI T/Z P
adventitia interface of the near and far walls of the distal CCA BMI (kg/m2 ) 24.69 ± 4.40 25.38 ± 2.48 −1.128 0.262
and along the far walls of the carotid bulb and ICA. The aver- WHR 0.89 ± 0.13 0.91 ± 0.05 −1.088 0.340
age of these was recorded for each location. The mean of all FPG (mmol/L) 8.33 ± 2.79 8.96 ± 3.30 0.905 0.367
average readings across the eight locations (four on each side) HbA1c (%) 7.11 ± 1.61 7.05 ± 1.71 0.515 0.819
was calculated. The site of the greatest thickness including a FINS (mIU/L) 8.20 ± 7.64 8.47 ± 9.04 −0.185 0.853
C-P (ng/ml) 2.79 ± 1.09 2.26 ± 1.00** 2.691 0.008
plaque lesion was sought along both near and far walls bilaterally SBP (mmHg) 132.19 ± 23.42 132.03 ± 15.15 0.050 0.960
(max-IMT). The RI was calculated by subtracting end-diastolic DBP (mmHg) 82.42 ± 14.61 84.01 ± 9.01 −0.836 0.404
Doppler-shifted frequency from peak-systolic-shifted frequency TG (mmol/L) 2.06 ± 1.16 1.71 ± 1.32 0.897 0.384
and dividing this value by peak-systolic-shifted frequency. Data TC (mmol/L) 4.85 ± 1.28 4.93 ± 1.05 −0.399 0.691
of RI in both sides was classified into max and min RI. HDL-C (mmol/L) 1.16 ± 0.42 1.20 ± 0.32 −0.741 0.460
LDL-C (mmol/L) 2.82 ± 0.86 2.99 ± 0.86 −1.158 0.249
IMT (L) (mm) 0.79 ± 0.14 0.87 ± 0.16** −3.804 < 0.001
3. Statistical analyses IMT (R) (mm) 0.77 ± 0.16 0.87 ± 0.16** −3.390 0.001
Max IMT (mm) 0.81 ± 0.15 0.91 ± 0.15** −3.810 0.001
All continuous data were tested for normality with Max RI 0.72 ± 0.07 0.74 ± 0.05* −1.927 0.056
Min RI 0.68 ± 0.06 0.71 ± 0.06* −2.188 0.030
Kolmogorov-Smirnov (KS) test. Statistical analysis was per-
formed for comparison between groups (variables) with BMI: body mass index; WHR: waist hip ratio; FPG: fasting plasma glucose;
Independent t-test (normally distributed) and Mann-Whitney U HbA1c: glycated hemoglobin; FINS: fasting insulin; C-P: C peptide; SBP: sys-
tolic blood pressure; DBP: diastolic blood pressure; TG: triglyceride; TC: total
test (non-normally distributed) for continuous data. The Chi2 cholesterol; LDL-C: low-density cholesterol; HDL-C: high-density cholesterol;
test was used for comparison of categorical variables. Series IMT: intima-media thickness; RI: resistance index.
of Pearson correlation and Spearman rank correlation methods * P < 0.05, ** P < 0.01.

were used to determine the variables that correlated with MoCA.

Multiple linear regression was used to create a prediction model
family history of diabetes, the durations of hypertension and
for correlations of variables and MoCA scores. These analy-
hyperlipidemia. Lower C-P levels were observed in MCI
ses were performed using the Statistical Package for the Social
group (2.26 ± 1.00 ng/ml vs. 2.79 ± 1.09 ng/ml, P = 0.008). No
Science (SPSS Graduate Student Version 16.0 for Windows). A
significant differences were found in BMI, WHR, FPG, HbA1c,
P-value of less than 0.05 was considered to be significant.
SBP, DBP, TG, TC, HDL-C, LDL-C (Table 2).
Ultrasonography of carotid arteries showed an increase of
4. Results IMT in both sides (left: 0.87 ± 0.16 vs. 0.79 ± 0.14, P < 0.001;
right: 0.87 ± 0.16 vs. 0.77 ± 0.16, P = 0.001) and max-IMT
Leaner correlation analysis showed that sex (r = −0.092, (0.91 ± 0.15 vs. 0.81 ± 0.15, P < 0.001) in MCI group. An
P = 0.198) and age (r = −0.048, P = 0.550) were not associated increase in min-RI (0.71 ± 0.06 vs. 0.68 ± 0.06, P = 0.030) was
with MoCA scores in these subjects. So the difference of also found in MCI group (Table 2).
cognitive function induced by sex and age was excluded in We evaluated the relationships between MoCA scores and
this study. Comparing with the NC group, the incidence of the duration of diabetes, history of hypertension, levels of C-
hypertension was higher in MCI group (58.06 vs. 40.63%, P, max-IMT and min-RI. A positive correlation was observed
P = 0.026). The durations of diabetes were longer in MCI between MoCA scores and C-P (r = 0.252, P = 0.005). MoCA
group than in NC group (4.80 ± 4.94 y vs. 3.09 ± 4.04 y, scores was negatively correlated with the history of hyper-
P = 0.024) (Table 1). No significant differences were found tension (r = −0.244, P = 0.002), duration of DM (r = −0.161,
between the two groups in the incidence of hyperlipidemia, P = 0.044), max C-IMT (r = −0.253, P = 0.005) and min C-
RI (r = −0.183, P = 0.023). On multiple regression analysis to
Table 1 which MoCA scores as a dependent variable, and the duration
Demographics in NC group and MCI group. of diabetes, history of hypertension, levels of C-P, max-IMT
NC MCI T/Z/χ2 P and min-RI as operands were included as independent varia-
bles. The significant relationship between MoCA scores and
Sex (male/female) 64 (37/27) 93 (44/49) 2.549 0.280
Mean age (y) 55.06 ± 7.28 55.34 ± 8.35 −0.218 0.827 duration of hypertension (␤ = −0.267, P = 0.002) and C-P lev-
Years of education (y) 10.16 ± 3.07 9.21 ± 3.42 1.631 0.105 els (␤ = 0.281, P = 0.001), as well as between MoCA scores and
History of hypertension 26 (40.63%) 54 (58.06%)* 4.933 0.026 min-RI (␤ = −0.221, P = 0.011) was found.
History of hyperlipidemia 43 (67.19%) 61 (65.59%) 0.043 0.835
Family history of T2DM 21 (32.81%) 35 (37.63%) 0.384 0.535
Duration of hypertension (y) 3.26 ± 7.00 4.10 ± 5.54 −0. 842 0.401 5. Discussion
Duration of hyperlipidemia (y) 0.94 ± 1.72 1.40 ± 3.61 −0.949 0.344
Duration of T2DM (y) 3.09 ± 4.04 4.80 ± 4.94* −2.287 0.024 MCI is a syndrome defined as cognitive decline greater than
NC: normal cognitive function; MCI: mild cognitive impairment. expected for an individual’s age and education level but that
* P < 0.05.
does not interfere notably with activities of daily life. It is, thus,
 R.-H. Chen et al. / Annales d’Endocrinologie 73 (2012) 208–212 211

distinct from dementia, in which cognitive deficits are more make contribution to cognitive function in this population. We
severe and widespread and have a substantial effect on daily also found RI, especially minimum RI in MCI group was greater
function [10]. Many studies have reported cognitive deficits in than NC group, and was associated with MoCA scores. RI was
type 2 diabetics when compared to a non-diabetic population introduced by Pourcelot [25] in 1974, to detect peripheral vascu-
[11,12]. Some studies showed that the incidence of cognitive lar disease. It is calculated from blood flow velocities in vessels
impairment was higher in patients with T2DM than non-diabetic during the cardiac cycle by a pulsed-wave Doppler ultrasound,
subjects [4,13]. About 10.8%∼17.5% [14] diabetic patients and could represent the stiffness and resistance of carotid vessel.
turned to cognitive deficits and presented as mild to moderate The higher the values, the greater is the impedance to blood flow.
cognitive dysfunction and decline of learning and/or memory. Previous studies showed RI was closely related to atherosclero-
Among such patients, about 10–15% would develop to dementia sis and CVD [26]. Our study revealed that min RI was positively
[15]. correlated with duration of DM (r = 0.363, P < 0.001) and max
MoCA is one of the most sensitive neuropsychological tests IMT (r = 0.363, P < 0.001) but not history (r = 0.132, P = 0.104)
in distinguishing those with MCI from healthy control subjects or duration of hypertension (r = 0.138, P = 0.089), suggested that
[16]. The cut-off value determined by the developers of the min RI of patients in this study was associated with DM and
MoCA was 25/26, which suggested a probable case of MCI. atherosclerosis but not hypertension. Although no else study
To find out DM patients with MCI could help to identify those indicated relationship between RI and cognitive function, we
with impaired cognitive function so that take intervention mea- still conjectured that stiffness of artery increases and atheroscle-
sures to them as early as possible. Thus, prevent the decline of rosis intensifies along with diabetic course prolonging, which
quality of life. attribute to the increase of arterial RI and decrease of blood flow
The mechanism responsible for cognitive decline associated of brain. All above would cause hypoxic ischemia of cerebral
with diabetes was still unclear. A number of possible mecha- and MCI ultimately.
nisms have been raised. Firstly, hyperglycemia appears to be We found low level of serum C-P as a potential risk factor
related to abnormalities in cognitive function in patients with of MCI in this study by comparing all the demographic char-
T2DM [17]. In other studies, hyperinsulinemia [18] was sup- acteristics. C-P is a 31–amino acid peptide that is cleaved from
posed to be another risk. At present study, we found cognitive proinsulin during biosynthesis of insulin [27]. Recent studies
function was not associated with duration of hypertension and have shown that C-P possesses physiological functions other
hyperlipidemia, history of hyperlipidemia, BMI, WHR, levels than providing structural support for proinsulin cleavage. C-P
of FPG, HbA1c, FINS, SBP, DBP, TG, TC, HDL-C or LDL- improves renal function, reduces urinary albumin excretion and
C. But we demonstrated several risk factors possibly associated glomerular filtration, and decreases blood retinal barrier leak-
with MCI in patients with T2DM. age [28]. Although Olivia et al. [29] thought Higher levels of
We found DM duration was important in the pathogene- C-P in those without diabetes was related to decline in gen-
sis of cognitive impairment. Durations of DM in MCI group eral cognition and verbal memory, some other study [30] found
were longer than in NC group. Pearson analysis showed reverse C-P replacement prevented oxidative stress, endoplasmic retic-
relationship between duration of DM and MoCA scores. This ulum, nerve growth factor receptor p75, and poly (ADP-ribose)
finding was in agreement with other studies, which reported polymerase-related apoptotic activities, thus, prevent the cogni-
the correlation between duration of DM and cognitive function tive dysfunction in type 1 DM. Our study showed C-P level was
[19,20]. lower in MCI group than NC group, and were positively cor-
A greater decline among the hypertensive diabetes patients related with MoCA scores, indicating that patients with higher
would have been expected. We found that the incidence of hyper- level of C-P keeps more cognitive function. Whereas patients
tension was higher in MCI group than NC group ([58.06% vs. with shorter duration of DM were with high insulin secretion
40.63%], P = 0.026). Spearman analysis and multiple regres- and C-P level, whether the result was caused by C-P’s own
sion showed that history of hypertension was correlated with physiological function or bias by duration of DM is unclear. Fur-
cognitive impairment. This finding confirms other data [21,22]. ther study is needed to find out the relationship and mechanism
We didn’t find the relationship between MoCA score and SBP between them.
or DBP. We speculated that may be caused by the effect of Unlike some other study, we didn’t find any directly correla-
hypotensor. tions of cognitive function with blood glucose, blood pressure,
Macrovascular (including cardiovascular, cerebrovascular blood lipids. We speculated the causes were as follow: some
and peripheral vascular) atherosclerosis was considered as the patients with high MoCA score and high blood glucose were
most common complications caused by diabetes. Patients with new diagnosed, which were with short duration of DM; patients
diabetes were more likely to have chronic cerebrovascular dis- were with different recognition of their own disease condition
ease comparing to the control group [13]. In Thus, a relationship and underwent different therapies; subjects in our study had dif-
between cognitive changes and diabetes may be based on pro- ferent age range from other studies, which maybe with different
gression of cerebrovascular disease. Many study proved that level of cognitive function.
IMT was a common marker of atherosclerosis [23]. Recently, Our study has a number of limitations. Firstly, patients
Gaetano et al. [24] revealed that IMT was also associated with recruited in this study received different therapy that we didn’t
cognitive function. In our study, Mean IMT of MCI group was exclude the effect of medication to the cognitive function. Next,
higher than that of NC group, indicating that atherosclerosis did our findings of negative relationships may be a consequence
212 R.-H. Chen et al. / Annales d’Endocrinologie 73 (2012) 208–212

of having evaluated individuals with modest cognitive deficits [11] Shimada H, Miki T, Tamura A, Ataka S, Emoto M, Nishizawa Y. Neu-
who had few co-morbid conditions and were in relatively good ropsychological status of elderly patients with diabetes mellitus. Diabetes
Res Clin Pract 2008;87:224–7.
metabolic control at study entry. Furthermore, the data collec-
[12] Murthy SB, Jawaid A, Schulz PE. Diabetes mellitus and dementia: advo-
tion of some factors including duration of DM, hypertension, cating an annual cognitive screening in patients with diabetes mellitus. J
hyperlipidemia and family history were on self-report and the Am Geriatr Soc 2008;56:1976–7.
medical record which may have lead to recall bias. Furthermore, [13] Petrova M, Prokopenko S, Pronina E, Mozheyko E. Diabetes type 2, hyper-
sample in this study was not larger enough to represent all dia- tension and cognitive dysfunction in middle age women. J Neurol Sci
2010;299:39–41.
betic patients. However, this sample size was considered to be
[14] Bruce DG, Casey GP, Grange V, Clarnette RC, Almeida OP, Foster JK,
sufficiently powered to examine the relationship between C-P et al. Cognitive impairment, phycical disability and depressive symptoms
levels, macrovascular complications and cognition in diabetic in older diabetic patients: the Fremantle Cognition in Diabetices Study.
patients. Diabetes Res Clin Prac 2003;61:59–67.
In summery, we found that duration of DM, history of [15] Petersen RC, Thomas RG, Grundman M, Bennett D, Doody R, Ferris
S, et al. Alzheimer’s Disease Cooperative Study Group. Vitamin E and
hypertension, low serum C-P level and subclinical macrovascu-
donepezil for the treatment of mild cognitive impairment. N Engl Med
lar complications including thickened C-IMT and higher C-RI 2005;352:2379–88.
were closely associated with MCI in patients with T2DM. We [16] Lee JY, Dong Woo Lee, Cho SJ, Na DL, Hong Jin Jeon, Kim SK, et al.
supposed this finding could have an important impact on the Brief screening for mild cognitive impairment in elderly outpatient clinic:
management of cognitive decline in diabetic patients. Further validation of the Korean version of the Montreal Cognitive Assessment. J
Geriatr Psychiatry Neurol 2008;21:104–10.
confirmation of the association is required in enlargement of the
[17] Mortimer JA, Borenstein AR, Ding D, Decarli C, Zhao Q, Copenhaver C,
number of samples and the form of a randomized, controlled trial et al. High normal fasting blood glucose is associated with dementia in
of intensive versus less intensive glycemic control with cognitive Chinese elderly. Alzheimers Dement 2010;6:440–7.
endpoints. [18] Luchsinger JA, Reitz C, Patel B, Tang MX, Manley JJ, Mayeux R. Relation
of diabetes to mild cognitive impairment. Arch Neurol 2007;64:570–5.
Disclosure of interest [19] Ebadya SA, Aramib MA, Shafigha MH. Investigation on the relationship
between diabetes mellitus type 2 and cognitive impairment. Diabetes Res
Clin Prac 2008;82:305–9.
The authors declare that they have no conflicts of interest [20] Xiuying Qi, Xianhai Meng, Yun Xie, Dan Qian, Yuying Zhou, Xia Zhao,
concerning this article. et al. Risk factors and predisposing genes for mild cognitive impairment in
patients with type 2 diabetes mellitus. Alzheimers Dement 2010;6:S453.
Acknowledgements [21] Hassing LB, Hofer SM, Nilsson SE, Berg S, Pedersen NL, McClearn
G, et al. Comorbid type 2 diabetesmellitus and hypertension exacer-
bates cognitive decline: evidence from a longitudinal study. Age Ageing
This work was supported by a joint research project of Shang- 2004;33:355–61.
hai health bureau (2009214) and science committee of Shanghai [22] Restelatto LMF, Varela JS, Costa AL, Bombardelli K, Rosin E, Silva RS,
Pudong New Area (PKJ2010-Y25). et al. Cardiovascular risk factors are directly correlated with the degree of
cognitive decline in Mild cognitive impairment and Alzheimer’s disease
patients. Alzheimers Dement 2011;7:S325.
References [23] Franek E, Januszkiewicz-Caulier J, Błach A, Napora M, Jedynasty K,
Budlewski T, et al. Intima-media thickness and other markers of atheroscle-
[1] Biessels GJ, Deary IJ, Ryan CM. Cognition and diabetes: a lifespan per- rosis in patients with type 2 diabetes and periodontal disease. Kardiol Pol
spective. Lancet Neurol 2008;7:184–90. 2012;70:7–13.
[2] Okereke OI, Kang JH, Cook NR, Gaziano JM, Manson JE, Buring JE, [24] Gorgone G, Ursini F, Altamura C, Bressi F, Tombini M, Curcio G, et al.
et al. Type 2 diabetes mellitus and cognitive decline in two large cohorts Hyperhomocysteinemia, intima-media thickness and C677T MTHFR gene
of community dwelling older adults. J Am Geriatr Soc 2008;56:1028–36. polymorphism: a correlation study in patients with cognitive impairment.
[3] Ryan CM, Geckle MO. Circumscribed cognitive dysfunction in middle- Atherosclerosis 2009;206:309–13.
aged adults with type 2 diabetes. Diabetes Care 2000;23:1486–93. [25] Pourcelot L. Applications clinique de l’examen Doppler transcutané. In:
[4] Arvanitakis Z, Wilson RS, Li Y, Aggarwal NT, Bennett DA. Diabetes Peronneau P, editor. Symposium: velometric ultrasonnordoppler. Paris:
and function in different cognitive systems in older individuals without Inserm; 1974. p. 213–40.
dementia diabetes. Diabetes Care 2006;29:560–5. [26] Grima P, Guido M, Chiavaroli R, De Donno A, Tana M, Zizza A. Com-
[5] Grodstein F, Chen J, Wilson RS, Manson JE. Type 2 diabetes and cog- parison of intima-media thickness and ophthalmic artery resistance index
nitive function in community-dwelling elderly women. Diabetes Care for assessing subclinical atherosclerosis in HIV-1-infected patients. Car-
2001;24:1060–5. diovasc Ultrasound 2011;1:9.
[6] Petersen RC. Mild cognitive impairment: current research and clinical [27] Li ZG, Qiang X, Sima AAF, Grunberger G. C-peptide attenuates pro-
implications. Semin Neurol 2007;27:22–31. tein tyrosine phosphatase activity and enhances glycogen synthesis in L6
[7] Weiner MF, Rosenberg RN, Womack KB, Svetlik DA, Fuller C, Fields myoblasts. Biochem Biophys Res Commun 2001;280:615–9.
J, et al. Atherosclerosis risk factors in American Indians with Alzheimer [28] Zhen-Guo Li, Sima AAF. C-peptide and central nervous sys-
disease: preliminary findings. Alzheimer Dis Assoc Disord 2008;22:245–8. tem complications in diabetes. Experimental Diabetes Res 2004;5:
[8] Luis CA, Keegan AP, Mullan M. Cross validation of the Montreal Cog- 79–90.
nitive Assessment in community dwelling older adults residing in the [29] Okereke OI, Pollak MN, Hu FB, Hankinson SE, Selkoe DJ, Grod-
Southeastern US. Int J Geriatr Psychiatry 2009;24:197–201. stein F. Plasma C-peptide levels and rates of cognitive decline in older,
[9] Zhao S, Guo C, Wang M, Chen W, Wu Y, Tang W, et al. A clinical memory community-dwelling women without diabetes. Psychoneuroendocrino
battery for screening for amnestic mild cognitive impairment in an elderly 2008;33:455–61.
chinese population. J Clin Neurosci 2011;18:774–9. [30] Sima AAF, Zhen-Guo Li. The effect of C-peptide on cognitive dys-
[10] Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, et al. function and hippocampal apoptosis in type 1 diabetic rats. Diabetes
Mild cognitive impairment. Lancet 2006;367:1262–70. 2005;54:1497–505.