Diabetes and dementia – who is at higher risk?


9th May 2021, Dr Chee L Kho

When we think of diabetes-related complications, we usually think about microvascular and macrovascular complications. We don’t usually think about dementia. Type 2 diabetes (T2D) represents an increasing health burden world-wide and its prevalence in particularly higher in elderly population. While epidemiological evidence suggests an increased risk of vascular and Alzheimer’s dementia associated with T2D, the mechanisms underlying the association, however, remain unclear. You may think that dementia is just an old person disease but with the age of onset of diabetes increasingly getting lower, we will be seeing more dementia earlier. Who amongst our patients with T2D is at higher risk of developing dementia?

The most evaluated mechanisms linking both disorders in pre-clinical studies include an increase in neuronal insulin resistance, impaired insulin signalling, pro-inflammatory state, mitochondrial dysfunction and vascular damage which increase deposition of β-amyloid, tau proteins and GSK3β, leading to an earlier onset of dementia in individuals with impairment in the glucose metabolism. Neuroimaging and neuropathology evidence linking cerebro-vascular lesions, neuro-degeneration and particularly small-vessel disease in the onset of dementia is consistent with the increased risk of incident dementia in T2D, but consistent evidence of AD-related pathology is scarce.

A very recent study published in JAMA examined the association between age at onset of diabetes and incident dementia using data spanning midlife to old age. It brought out some really interesting data that might help us identify who is at higher risk of dementia. The Whitehall II study is an ongoing cohort study established in 1985-1988 among 10 308 persons, aged 35-55 years employed in London-based government departments. Patients had follow-up clinical examinations every 4 to 5 with each wave taking 2 years to complete, with ongoing data collection at the 2020-2021 wave.

Over the 30-year study, patients were categorised into T2D or free of diabetes according to fasting glucose readings or self-reported physician diagnosed T2D. Patients were categorised as prediabetes if the glucose was between (5.6-6.9 mmol/L). Their FINDRISC scores were also assessed. Dementia cases were ascertained by linkage to three national registries.

The association of T2D with dementia was examined according to age at T2D onset (at 55, 60, 65, and 70 years) with the reference group at each age being participants free of type 2 diabetes at that specific age. For example, at age 70 years the group defined as having no T2D at age 70 was compared with a group diagnosed with T2D at ≤ 5 years earlier, 6-10 years earlier, and > 10 years earlier.

The risk of dementia among participants with type 2 diabetes without CVD was compared with participants who had T2D and CVD co-morbidity (CHD, heart failure, and stroke).


Between 1985 and 2019, a total of 10 308 persons were recruited. Data was available for 10,095 participants. 6.3% of participants were diagnosed with dementia. These individuals were more likely to have chronic conditions over the follow-up period and to have a worse cardiovascular risk profile at age 60. Over the follow-up period, 16.9% developed diabetes, among whom 8.9% were subsequently diagnosed with dementia. Not expectedly:

Participants with diabetes were more likely to developed dementia compared participants who were free of diabetes:

  • At age 55, 5.06 per 1000 person years vs 3.14 per 1000 person years (Hazard ratio (HR) 2.14)
  • At age 70, 13.88 per 1000 person years vs 8.85 per 1000 person years (HR 1.58)

Diabetes at an earlier age was more associated with dementia in a linear fashion.

At 65 years, compared with participants who were free of diabetes:

  • Participants who had diabetes for 0-5 years were 1.53X more likely to develop dementia
  • Participants who had diabetes for 5-10 years were 2.03 X more likely to develop dementia

At 70 years, compared with participants who were free of diabetes:

  • Participants who had diabetes for 0-5 years were 1.11 X more likely to develop dementia
  • Participants who had diabetes for 5-10 years were 1.49 X more likely to develop dementia
  • Participants who had diabetes for >10 years were 2.12 X more likely to develop dementia

Participants who developed diabetes at a younger age were younger at onset of dementia

Prediabetes was not significantly associated with dementia, irrespective of the age at which prediabetes was assessed

The presence of cardiovascular co-morbidity (heart failure, coronary heart disease) significantly increases the risk of dementia:

  • Compared with participants with diabetes alone, diabetes accompanied by stroke increases the risk of dementia by 2.17 X while the presence of stroke and heart disease increases the risk of dementia by 4.99 X. The numbers were small, though.

The results of this study is not entirely surprising. We do see patients with diabetes who developed dementia at a higher rate than the general population without diabetes. This study reminds us that the younger the patient is at diagnosis of diabetes, the higher the risk is. Further, if the patient with diabetes also have cardiovascular co-morbidities, the risk of dementia is significantly higher.

If we can delay or prevent altogether the development of diabetes, we could reduce the patient’s risk of developing dementia. If we can work on preventing macrovascular and microvascular complications, we can further reduce those risks. There is work for us to do in primary care looking after patients with diabetes and those who are about to get diabetes.


Barbiellini Amidei C, Fayosse A, Dumurgier J, Machado-Fragua MD, Tabak AG, van Sloten T, Kivimäki M, Dugravot A, Sabia S, Singh-Manoux A. Association Between Age at Diabetes Onset and Subsequent Risk of Dementia. JAMA. 2021 Apr 27;325(16):1640-1649. doi: 10.1001/jama.2021.4001. PMID: 33904867; PMCID: PMC8080220.