The publication in The BMJ is exciting news and provides yet another avenue for the prevention of dementia, in particular Alzheimer's and vascular dementia.

This impressive and very large study from Korea of 110,885 pairs of adults compared how two anti-diabetic drugs influenced the likelihood of developing dementia over eight years in adults aged 40-69 years with type 2 diabetes.

People on an SGLT-2 inhibitor had a 35% lower rate of dementia compared to those on a DPP-4 inhibitor, a different anti-diabetic drug. The reductions were similar regardless of whether cases of any dementia or just those with Alzheimer's or vascular dementia were considered.

People who took the SGLT- 2 inhibitor for more than two years had a stronger benefit than people who took it for two years or less. Suggested mechanisms of how this could work are direct neuroprotective effects and fewer deposits of toxic proteins implicated in Alzheimer's in the brain.

As this is an observational study, not a trial, the findings have limitations. It is premature we should be prescribing semaglutide to prevent dementia. We now need a large randomised controlled trial to determine how useful this will be. For example, we don't know if it works in older people, people without diabetes, or people who already have early signs of dementia.

This is a very interesting study looking at the potential for SGLT2 drugs, a newer type of diabetes medication, to prevent dementia. Using an observational approach, the authors found that people taking SGLT2s were quite a bit less likely to get dementia than people taking a somewhat older type of medication called a DPP4i. 

Because of the observational nature of this study, it’s hard to know if this means that SGLT2s prevent dementia or if there are other factors at play. The study was impressively robust, but even very good observational studies may have unseen issues that impact their results. There are some indications in this paper that the effect (reducing dementia) may in part be due to better diabetes control rather than some direct impact of SGLT2s.

However, the findings do raise a very interesting possibility for preventing dementia, which is at the moment both hard to prevent and treat. The authors recommend looking into this further in randomized trials, which seems like an extremely good idea.

People with type 2 diabetes mellitus (T2DM) have an increased risk of dementia. Many factors may contribute to this strong association, including increased risk of stroke, narrowing and thickening of the small vessels in the brain and the cross-association of mid-life obesity and diabets. Efforts to treat vascular risk factors, both high blood pressure and diabetes, have been proposed as a means to reduce global dementia burden. There is now evidence that some drugs commonly used in the treatment of diabetes may be protective against cognitive decline.

Diabetes drugs are usually referred to by their class and include sodium-glucose cotransporter-2 (SGLT-2) inhibitors, like empagliflozin/Jardiance®, dipeptidyl peptidase-4 inhibitors (DPP-4, gliptins) like sitagliptin/Januvia®, and the ones most in the news for weight loss, the glucagon-like peptide-1 (GLP-1) receptor agonists, like semaglutide/Ozempic®. We’ve known for a while that SGLT-2 inhibitors can have benefits beyond controlling blood sugars.

The authors looked at Korean National Health Insurance Service data on people aged 40-69 years who were taking SGLT-2 inhibitors – what’s called a pharmaco-epidemiological approach. They compared those on the SGLT-2 inhibitors to those taking DPP-4 inhibitors. They found that the risk of any type of dementia was reduced in people on these drugs, as well as the specific risk for Alzheimer’s dementia and vascular dementia. Overall, they found a 35% relative reduction in dementia risk, but it also looked like the longer you were on the drugs, the greater the effect.

The study is observational, meaning that future randomised controlled trials will be needed. There is evidence that other classes of drugs, like the glucagon-like peptide-1 (GLP-1) receptor agonists, and blood pressure control tablets, may also protect cognition. Whilst we are at the beginning of medications to treat dementia, current drugs are far from a cure. This makes this exciting news for the more than 400,000 people in Australia with dementia. It doesn’t mean that if you take this pill that you will never get dementia, but it gives further weight to the belief that risk factor reduction – good control of diabetes and blood pressure – coupled with a ‘brain health optimised’ lifestyle – great diet, lots of exercise and social engagement – goes a long way towards protecting your brain.

Previous studies on the effects of anti-diabetic drugs on dementia risks have been inconclusive e.g. metformin, the frontline anti-diabetic drug, is reported to increase the risk of Alzheimer’s in Asian populations (JAD 2022 Anling Luo et al.). The study is really important in providing strong evidence that the use of SGLT-2 is associated with a substantially reduced risk of dementia. The use of a robust, large dataset gives a reasonable confidence in the accuracy of the outcomes.

A report by Brown et al indicates that if interventions delay both disease onset and progression by a modest one year, there would be nearly 9.2 million fewer cases of the disease in 2050 world-wide, so the impact could be remarkable.

However, there are a few limitations to point out. As noted in the Editorial, the data only follows up for less than a two year period, whilst ‘silent’ dementia pathologies are known to progressively develop over years and decades. Also, the study does not consider the dosage and duration of these drugs. It would have made the study much stronger if the authors could indicate the association with risk reduction against the ‘net exposure (cumulative dose + duration factors)’.

Finally, also noted in the Editorial too but in these studies, the accuracy of diagnosis and its timing can be a concern. However in Korea, unlike the UK or Australia, the GP system is not common and ICD-10 indicated diagnoses are normally given by specialist, which increases the accuracy compared to healthcare system with GP.