Peripheral microvascular complications in diabetes are associated with concurrent cerebrovascular disease. As detailed cognitive assessment is not routinely carried out among diabetic patients, the aim was to establish whether the presence of clinical “peripheral” microvascular disease can identify a subgroup of patients with early evidence of cognitive impairment. Detailed psychometric assessment was performed in 23 diabetic patients with no microvascular complications (Group D), 27 diabetic patients with at least one microvascular complication: retinopathy, neuropathy, and/or nephropathy (Group DC), and 25 healthy controls (Group H). Groups D and DC participants had significantly lower scores on reaction time (
Diabetes mellitus is a common and serious disorder. The prevalence of the disease is projected to continue to increase significantly worldwide over the next few decades [
In recent years, interest has also been directed towards another potential complication of diabetes, namely, cognitive decline. Increasing epidemiological evidence has linked diabetes with cognitive decline and dementia [
Many of the clinical complications of diabetes are caused by small and large vessel pathology [
The association between diabetes and cognitive impairment is important not just from an aetiological perspective, but also from the standpoint of day to day practical management of the disorder. The presence of dementia is likely to have a significant impact on the self-care and independency of patients [
Microvascular complications in diabetes arising outside the brain are associated with concurrent brain pathology. However, it remains unclear whether this translates into early evidence of cognitive abnormalities beyond what is observed in diabetic patients free from such complications. Hence, the main aim of this study was to assess whether the presence of “peripheral” microvascular complications in patients with diabetes pointed towards evidence of early cognitive decline. Such data is important to assess whether the clinical diagnosis of microvascular complications in organ systems other than the brain identifies a subgroup of diabetic patients showing early sign of cognitive failure. In addition, such information may also contribute to increasing our understanding on the role of microvessel disease in the genesis of cognitive dysfunction in diabetes.
Seventy-five participants were recruited into three study groups: 25 healthy nondiabetic controls (Group H), 27 diabetic patients without the clinical presence of microvascular disease complications (Group D), and 23 diabetic patients with at least one peripheral microvascular complication such as nephropathy, retinopathy, and/or neuropathy (Group DC).
Group H participants were recruited from diabetic case-group relatives or those who were attending community health clinics but who met all the inclusion and exclusion criteria and did not have diabetes. Groups D and DC participants were recruited from a mixture of community and hospital clinics. All diabetic participants in groups D and DC had a documented diagnosis of diabetes and were under the care of hospital diabetes service. The diagnosis of microvascular complications was obtained and classified using clinical, blood, and imaging investigations. In addition to a diagnosis of diabetes, group DC had a documented diagnosis of at least one of the following microvascular disease complications: retinopathy, nephropathy, and/or peripheral neuropathy. Such a diagnosis was established prior to the commencement of this study and was made or confirmed by experts. Hence, the diagnosis of retinopathy was made by ophthalmologists and the diagnosis of nephropathy was made by nephrologists. The diagnosis of peripheral neuropathy was confirmed by peripheral limb being insensate to 10 g microfilament, which is the gold standard assessment tool currently. Those who had uncertain diagnosis of vascular complications or were still under investigations to confirm such a diagnosis were not recruited to this study. Likewise, those with suspected preclinical microvascular disease but with no clear clinical manifestations were also excluded. Participants who had evidence of rheumatoid arthritis, cerebrovascular accidents, myocardial infarction, peripheral arterial disease, alcohol dependency, and depression were excluded from the study. None of the participants had a diagnosis of dementia in any of its stages or subtypes. Approval for this study was sought and obtained from the local Ethics and Research and Development committees.
Cognitive testing was carried out by using a touch-screen-computerised battery of psychometric tests. The Cambridge Neuropsychological Test Automated Batteries (CANTAB) [
Immediately before testing, participants were interviewed for current demographic data and assessed for the clinical diagnosis of a small vessel disease. Patients were included in either the D or DC groups solely based on the presence or absence of clear clinical diagnosis of microvascular complications. Data was recorded on the data collection sheet annotated with the confidential project number.
Quantitative data was collected from each participant and from each CANTAB test. The data exhibited clearly nonnormal distributions (predominantly a left-tail skew) in consequence of which nonparametric test procedures were used throughout. The Kruskal-Wallace analysis of variance by ranks was used initially to test for between-group differences in cognitive function [
Demographic data and social characteristics of participants.
Characteristics | Healthy group (H) | Diabetics without complications (D) ( | Diabetics with complications (DC) ( |
---|---|---|---|
Range | Range | Range | |
Mean age (years) | 53 (41–68) | 60 (45–80) | 61 (42–80) |
Male : Female | 11 : 14 | 14 : 11 | 14 : 11 |
Ethnicity (non-Caucasian) | 0 | 0 | 0 |
School years | 11 | 10.6 | 10.9 |
Smokers | 0 | 0 | 0 |
Diabetes type 1 : 2 | 0 | 6 : 19 | 17 : 9 |
Diagnosed diabetic (years) | 0 | 7.4 (1–27 years) | 20.8 (1–46 years) |
The sample size was estimated using the industry standard sample size calculation software package nQuery Advisor Release 4.0. The nonparametric size module was used for independent group comparisons (Mann-Whitney and Kruskal-Wallis test procedures) assuming a two-sided test significance level of 5%, and a specified test power of 80%. A difference between group-median values of at least 10% of scale range was proposed as the effect of interest to be detected, should such a difference exist.
Seventy-five individuals took part in the study and were subsequently divided into 3 groups. Demographic characteristics of participants are presented in Table
The results obtained on the 7 different CANTAB test domains are presented in Table
CANTAB test data for the three study groups across the various test domains. Pattern recognition memory (PRM), reaction time (RT), and rapid visual information processing (RVP) differed among the study groups with only RT and RVP remaining significantly different after adjusting for age.
Test | Group | min | 25th centile | Median | 75th centile | max | 95% CI (median) | |
---|---|---|---|---|---|---|---|---|
PRM | H | −2.048 | −0.338 | 0.495 | 0.919 | 1.286 | −0.179 to +0.581 | |
D | −2.5 | −1.452 | −0.282 | 0.495 | 1.023 | 0.024 | −1.300 to +0.319 | |
DC | −3.1 | −0.879 | −0.338 | 0.495 | 1.343 | −0.752 to +0.441 | ||
SRM | H | −4.136 | −0.639 | 0.0278 | 0.317 | 1.923 | −0.560 to +0.275 | |
D | −2.747 | −1.139 | −0.140 | 0.360 | 1.878 | 0.86 | −1.062 to +0.360 | |
DC | −4.136 | −1.360 | −0.275 | 0.360 | 1.415 | −0.897 to +0.344 | ||
PAL | H | −1.652 | −0.492 | 0.214 | 0.530 | 1.271 | −0.292 to +0.447 | |
D | 10.419 | −2.928 | 0.0976 | 0.695 | 1.017 | 0.63 | −0.250 to +0.695 | |
DC | 11.182 | −2.831 | 0.000 | 0.508 | 1.017 | −1.187 to +0.390 | ||
MTS | H | −3.808 | −1.351 | −0.307 | 0.686 | 1.170 | −1.027 to +0.431 | |
D | −6.271 | −0.997 | 0.322 | 0.521 | 1.176 | 0.89 | −0.903 to +0.467 | |
DC | −4.001 | −1.046 | −0.247 | 0.940 | 1.170 | −0.841 to +0.505 | ||
RTmove | H | −2.890 | −0.003 | 0.466 | 1.060 | 1.435 | +0.094 to +0.965 | |
D | −2.407 | −0.641 | 0.217 | 1.025 | 1.573 | 0.43 | −0.639 to +0.812 | |
DC | −331.3 | −0.572 | 0.434 | 1.010 | 1.970 | −0.330 to +0.838 | ||
RT reaction time | H | −1.476 | −0.148 | 0.515 | 0.797 | 1.847 | +0.076 to +0.766 | |
D | −9.092 | −1.150 | −0.354 | 0.248 | 1.578 | 0.003 | −1.307 to −0.018 | |
DC | −4.299 | −1.496 | −0.525 | −0.156 | 1.199 | −1.462 to −0.210 | ||
RVP | H | −1.773 | −0.846 | 0.025 | 0.689 | 1.595 | −0.702 to −0.577 | |
D | −1.904 | −1.713 | −1.037 | 0.217 | 1.221 | 0.001 | −1.698 to −0.151 | |
DC | −2.644 | −1.621 | −1.007 | −0.344 | 1.782 | −1.538 to −0.410 |
Group H differed significantly from both diabetic groups on two of CANTAB domains, namely, RT and RVP. Interestingly, there was no significant difference between Group D and Group DC on these two or any other CANTAB domains (Table
Analysis of group differences for RT and RVP. Group H differed significantly from Groups D and DC on both measures. However, there was no significant difference between Group D and Group DC on either of the cognitive tests.
RT | |
Mann-Whitney | Significance |
H versus D | |
H versus DC | |
D versus DC | |
RVP | |
Mann-Whitney | Significance |
H versus D | |
H versus DC | |
D versus DC |
Diabetes is a common disorder with a range of serious and potentially life-threatening complications. Many of these complications are clearly mediated through the disease’s toxic effects on blood vessels. To date, however, the neuropathological mechanisms which may contribute to the observed cognitive impairment in diabetes remain unclear.
There is growing body of evidence linking non-brain microvascular complications with cerebrovascular disease findings in patients with diabetes [
The results obtained in this study have shown that patients with a diagnosis of diabetes, but who do not suffer from a dementia illness, have some early evidence of cognitive deficits. Patients with diabetes, with and without microvascular complications, performed significantly worse on reaction times and on rapid visual processing which require focussed attention for a prolonged period. These results support the notion that diabetes negatively affects some aspects of cognition, a process which may lead in some to dementia. These results add weight to the various epidemiological and retrospective studies which have shown a link between diabetes in general and dementia.
MRIs of the brain are not routinely ordered as part of assessment of diabetes. In fact, diabetic patients are not routinely evaluated for cognitive outcome [
As no significant difference in cognitive scores existed among diabetes patients with and without microvascular complications, it is tempting to speculate about the role of microvascular pathology in the development of dementia among diabetic patients. A recent study has shown that medial temporal lobe atrophy was associated with diabetes independently of the amount of small vessel disease [
This study has several limitations which need to be considered. The relatively small number of participants in each of the 3 study groups did not allow for meaningful subgroup evaluation in relation to diabetes types 1 and 2. In this study, diabetes patients had both types 1 and 2, as has been the case in many other studies [
Both hypertension and diabetes are known risk factors for vascular disease [
There has been interest in recent years in establishing whether the degree of glycemic control may influence cognitive function. Cukierman-Yaffe et al. (2009) [
The main aim of this study was to assess whether the presence of microvascular complications arising outside the brain proper identifies a subgroup of diabetic patients with a worse cognitive profile which would not necessarily be detected by the widely used brief cognitive screening tests [
Notwithstanding some of the limitations in the current study, the data presented here are important in two aspects. Firstly, clinically relevant microvascular complications arising outside the brain are not associated with a poorer cognitive function among diabetic patients. Hence, based on these results, which will need to be confirmed by other studies, no additional cognitive investigations are warranted in such patients solely based on the presence of such microvascular complications; that is, when no cognitive impairment is clinically suspected. Secondly and indirectly, the results add weight to the notion that mechanisms not restricted to microvascular pathology may be responsible for the associated cognitive impairment observed in diabetes.
The authors thank Paul Fosbray and Sarah Bowditch for their assistance with CANTAB equipment.