REVIEW
SA JOURNAL OF DIABETES & VASCULAR DISEASE
144
VOLUME 7 NUMBER 4 • NOVEMBER 2010
Diabetes mellitus
The association between diabetes mellitus and cognitive decline
is supported by evidence from biochemistry, neuroimaging and
pathology. The hyperglycaemia-induced mitochondrial superoxide
overproduction leads to glucose-mediated microvascular damage.
8
A dysfunction of the insulin-degrading enzyme may result in both
hyperinsulinaemia and accumulation of cerebral amyloid protein
b
, thus providing an association between diabetes and dementia.
9
Diabetes often develops in the context of the metabolic syndrome,
leading to the indirect ischaemic effects of diabetes-associated
cerebrovascular disease.
10
MRI studies revealed a higher frequency of
brain atrophy and a reduced volume of memory-relevant structures
(hippocampus, amygdala) in diabetic patients.
11,12
A large autopsy
study detected more microvascular infarcts and an activation of
neuro-inflammation in demented patients with diabetes.
13
Studies
Initial evidence regarding an association between diabetes and
cognitive dysfunction was derived from several smaller case-control
studies. Though these studies had some methodological problems,
they demonstrated poorer performance in diabetic patients in at
least one aspect of cognitive function, mostly verbal memory.
14
A systematic review of 14 longitudinal population-based studies
assessing the incidence of cognitive decline revealed a higher
incidence of dementia in the majority of studies but also criticised
the lack of relevant confounders such as hypertension, stroke and
glycaemic control.
15
Additional prospective studies accounting for
these deficits have been published and add further evidence that
diabetes is an independent factor in cognitive decline (Table 1).
In the Osteoporotic Fractures Research Group which included
9 679 older women, controlled for several confounding factors
and showed that diabetes was associated with greater cognitive
decline.
16
The Nurses’ Health Study followed 16 596 women, aged
70–81 years, for two years and included details about treatment
and duration of diabetes. The study showed an increased risk of
poor cognition among women who had had diabetes for
>
15
years (OR 1.52; 95% CI 1.15–1.99) and for women not using any
medication (OR 1.45; 95% CI 1.04– 2.02).
17
Use of oral antidiabetic
therapy may reduce risk of cognitive decline as diabetic women
treated with these agents performed similarly to women without
diabetes (OR 1.06 and 0.99).
17
In a sub-analysis of data from a
four-year randomised trial of raloxifene among 7 027 osteoporotic
postmenopausal women the risk of developing cognitive impairment
among diabetic women was almost doubled.
18
The Physicians’
Health Study II with 5 907 men and the Women’s Health Study
with 6 326 women revealed that participants with longer duration
of diabetes (
>
5 years) had generally a greater cognitive decline.
19
More recent trials like the Washington Heights–Inwood Columbia
Aging Project focused on the relationship of diabetes to MCI and
found a higher risk of amnestic MCI among diabetic participants.
20
Hypoglycaemic episodes are an important barrier to the
achievement of optimal glycaemic control. A large cohort study
with 16 667 elderly diabetic people showed an association with
the number of severe hypoglycaemic episodes and increased risk
of dementia.
21
In contrast, a smaller study found no contribution
of hypoglycaemic episodes to cognitive impairment, but reported a
higher risk of hypoglycaemia in patients with dementia.
22
Table 1.
Summary of major prospective studies of the association of diabetes mellitus and cognitive performance.
Study
Participants Mean
age
(years)
Follow-
up
(years)
Cognitive test
Assessment of
diabetes
Covariates
Result (OR/
HR; 95% CI)
for cognitive
decline in
diabetes
Osteoporotic
Fractures Research
Group
16
9 679
♀
>
65 3–6
Digit symbol test,
Trails B and MMSe
at baseline
Self-report
Age, education, depression,
stroke, visual impairment,
heart disease, hypertension,
physical activity, oestrogen use
and smoking
1.63; 1.20–2.23
Nurses’ Health
Study
17
16 596
♀
74.2 2
Telephone
interview for
cognitive status
Self-report
Age, education, duration and
medication of DM
1.34; 1.14–1.57
Multiple Outcomes
of Raloxifene
Evaluation
18
7 027
♀
66.3 4
Five cognitive tests
at baseline and
after four years
Self-report
+
antidiabetic
medication
+
increased
fasting glucose
Age, education, race,
depression
1.79; 1.14–2.81
Physicians’ Health
Study
+
Women’s
Health Study
19
5 907
♂
+
6 326
♀
♂
74.1;
♀
71.9
♂
2;
♀
4
Telephone
interview for
cognitive status
Self-reported
questionnaire
Age, education and hormone
use (women), baseline score,
BMI, hypertension, cholesterol,
depression, smoking, alcohol,
physical activity
–0.74; –1.05
to –0.43 mean
difference of
decline in DM
Washington
Heights– Inwood
Columbia Ageing
Project
20
918
75.9 6
Complete
neuropsychological
testing
Self-report or
antidiabetic medication
Age, sex, education, ethnic
group, apo E, hypertension,
LDL, smoking, stroke, heart
disease
1.4; 1.1–1.8
apo E
=
apolipoprotein E; BMI
=
body mass index; CI
=
confidence interval; DM
=
diabetes mellitus; HR
=
hazard ratio; LDL
=
low-density lipoprotein; MMSe
=
mini-mental state examination; OR
=
odds ratio;
♂
=
men;
♀
=
women
