Background:​

Older people with diabetes mellitus (DM) are at increased risk of Alzheimer’s disease (AD) and deficits in specific domains (e.g., executive function and psychomotor speed) of cognitive functioning. In addition, vitamin B12 deficiency is common in older people with DM. There is therefore a need to examine cognitive profiles, investigate the risk factors associated with cognitive impairment and elucidate the effect of vitamin B12 supplementation in older people with DM.

Objectives:​

1. Assess the effects of vitamin B12 supplementation to prevent cognitive decline and brain atrophy in older people coexists with DM and mild vitamin B12 deficiency.

2. Assess the clinical, biochemical and neuroimaging risk factors for cognitive decline in older people with DM.

3. Compare the profiles of cognitive deficits, nutritional status, cardiovascular conditions and brain abnormalities between subjects with and without DM in the presence of amnestic mild cognitive impairment (aMCI) or AD.

4. To investigate the potential mechanisms of cognitive impairment in older people with DM.

Subjects and Methods:​

The main study was a randomized controlled trial of vitamin B12 supplementation to prevent cognitive decline and brain atrophy in 271 older diabetic people with mild vitamin B12 deficiency. All subjects were randomly assigned to take either methylcobalamin 1000 microgram or two similar looking placebo tablets once daily for 27 months and in the following 18 months all subjects took methylcobalamin as an open label extension trial. All subjects were followed up at 9 monthly intervals.

Brain volumetric magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI) scans were performed to a subsample of subjects with consent at baseline, 27 months and 45 months. Seven lipid classes were analyzed by liquid-chromatography mass spectrometric (LC-MS) at baseline.

Data from two other studies was selected to make cross-sectional comparisons between DM and non-DM in older people with aMCI or AD.

Results:​

The randomized controlled trial with vitamin B₁₂ supplementation did not show significant group difference in the rate of brain changes and cognitive decline at follow up. At month 27, the rate of atrophy in whole brain and hippocampus per year in active group was 0.80% and 0.66%, respectively, and in placebo group it was 0.79% and 0.39%, respectively. At month 45, the rate of atrophy in hippocampus per year in the active group and placebo group was 0.45% and 0.11%, respectively. 23 (18.9%) and 15 (15.2%) active treatment group subjects and 18 (14.9%) and 16 (17.6%) placebo group subjects had cognitive decline (defined by any increase in clinical dementia rating scale (CDR) global scores) at month 27 and 45, respectively. As vitamin B12 supplementation did not have significant effect on cognitive function or brain structures, we considered all trial group subjects as a cohort to look for risk factors associated with cognitive decline in older people with DM.

Baseline serum high-density lipoprotein cholesterol (HDLC) level was positively associated with executive function in older people with DM at month 27. Lipidomic analysis of serum samples at baseline found that 8 metabolites (7 phosphatidylcholines (PC), 1 lyso-PC) were significantly decreased in subjects with cognitive decline at month 27 after adjusting for age, sex, education level and group assignment. Out of the eight lipids, 5 and 2 correlated significantly with HDL and statin use respectively. 6 lipids were significantly associated with cognitive decline independent of statin use.

MRI imaging markers of baseline higher white matter hyperintensities (WMH) and lower hippocampal volume ratio (hippocampal volume/intracranial volume) were associated with declining in executive function at month 45 and the association between WMH and executive function was not independent of hippocampal volume ratio (P>0.05). Both hippocampal and cingulate volume ratios were positively and independently associated with memory function at month 45. Higher novel DTI-based marker for white matter integrity (peak width of skeletonized mean diffusivity [PSMD]) at baseline was significantly associated with worse executive function at both month 27 and month 45 after controlling for confounding factors. However, this association was not independent of baseline hippocampal volume ratio (P=0.092). One unit of PSMD change could result in more executive dysfunction at month 45 compared with WMH (Beta was -3.97 and -0.51, respectively).

Cross-sectional comparison between aMCI with and without DM found that DM-aMCI group had more hypertension and history of stroke, whereas, lipids profile (except HDL-C level) was better due to the more users of statin than their counterparts without DM and also HDL-C was lower in subjects with DM. Serum homocysteine concentration was independently raised in DM-aMCI group. Non-DM aMCI had smaller whole brain and hippocampal volumes than the DM-aMCI group. Linear regression models found that executive function, memory function and daily function (measured by CDR-sum of box (SOB)) were impaired more in aMCI with DM compared to those without DM after controlling for age, sex, education levels, serum homocysteine, active vitamin B12,folate, creatinine, HDL-C levels and history of stroke. The relationship between executive dysfunction, CDR-SOB and DM was not independent of WMH while the association between memory function and DM was independent of hippocampal volume ratio.

In the cross-sectional comparison between DM-AD and non-DM AD, there was no any significant group difference in regard to cognitive profiles (MoCA, ADAS-cog and verbal fluency test), brain volumes and WMH. Whereas, subjects with DM got lower scores in Disability assessment for dementia (DAD) especially in the domain of initiation, when compared with their counterparts without DM. This association was independent of ADAS-cog and hippocampal volume ratio, but not independent of WMH.

Conclusion:​

Vitamin B₁₂ supplementation did not show any beneficial effects on cognitive function or brain atrophy in older people with diabetes mellitus and mild vitamin B12 deficiency over 27 months.

Overall, eight lipids at baseline were associated with global cognitive deterioration at month 27. Lower baseline serum HDLC level was associated with more cognitive decline in executive function in older people with DM. Lower hippocampal volume ratio was independently associated with decline in executive and memory function. White matter integrity was significantly associated with executive function. Taken together, both hippocampal atrophy and white matter integrity are important mediators of cognitive impairment in older people with DM. Furthermore, DM had an independent effect on memory function and DM could also affect executive function through WMH in the presence of aMCI. In contrast, DM did not appear to have significant influence on cognitive function in older people with AD, which may suggest DM affects cognitive function in the very early stage of AD. However, AD patients with DM were relatively more impaired in daily functioning especially in initiation. This was probably mediated by white matter disease.