Metformin Linked With Less Dementia in People With Diabetes

Marlene Busko, for Medscape

July 28, 2022

The study covered in this summary was published on medRxiv as a preprint and has not yet been peer reviewed.

In an observational study of more than 112,000 US residents aged 50 years or older with type 2 diabetes, new users of metformin had about a 20% lower rate of incident dementia compared to new users of sulfonylurea during 5 years of follow-up.
The new users of metformin also had about a 20% lower rate of Alzheimer’s disease (AD) and vascular dementia than new users of sulfonylurea, but they were not at lower risk of mild cognitive impairment (MCI) or Parkinson’s disease (PD).

Type 2 diabetes is a well-established risk factor for dementia and PD, suggesting that metformin might reduce the risk of these neurodegenerative diseases, but findings from prior studies have been inconsistent, and these studies also had important limitations.
The authors said that this is the largest observational study to investigate the association between metformin and multiple neurodegenerative outcomes, that it used a variety of statistical methods to reduce bias and to validate the robustness of the results, and that it provides stronger evidence of the potential neuroprotective qualities of metformin regarding risk of dementia.
Were future randomized clinical trials and genetic studies to confirm the findings and help elucidate the mechanisms of the neuroprotective effect, it might lead to repurposing metformin to treat dementia as well as helping to develop novel therapies for dementia.
The findings highlight the need for careful consideration of age, race, and renal function when selecting participants in clinical trials that address the benefits and safety concerns associated with repurposing of metformin.

The authors used observational data from the Optum electronic health record dataset of patients in the United States. They focused on patients with diabetes who were at least 50 years old. They chose patients for whom at least 1 year of follow-up data were available for the period 2006–2008, as well as at least 1 year of follow-up data before starting treatment with either metformin or a sulfonylurea. The patients had not been diagnosed with dementia, MCI, or PD at the time of this treatment onset.
Patients included in the analysis had to continue taking either metformin or a sulfonylurea for at least 2 years. They could receive other antidiabetic agents during this period, but the study excluded those who added the comparator agent during follow-up.
The study’s primary outcomes were the incidence during follow-up of all-cause dementia and PD. Secondary outcomes were the incidence of AD, vascular dementia, and MCI.
The researchers applied propensity score adjustment that incorporated 33 demographic and clinical variables.

The study population included 96,140 new users of metformin and 16,451 new users of sulfonylurea.
The average age of the patients was about 66 years; the mean A1c level was about 6.8%; 48% were men; and about 83% were White.
During a 5-year follow-up, 2.3% of metformin users and 5.8% of sulfonylurea users developed all-cause dementia, and 0.7% of metformin users and 0.8% of sulfonylurea users developed PD.
In the propensity score–adjusted analysis, compared with new users of sulfonylurea, new users of metformin had a significantly lower risk of all-cause dementia (hazard ratio [HR], 0.80) but not PD (HR, 1.00).
New users of metformin also had a significantly lower risk of AD (HR, 0.81). The rate of vascular dementia was reduced by a similar amount (HR, 0.79); this was on the cusp of significance. The adjusted incidence of MCI showed a point estimate reduction, which was not significant.
Similar associations appeared in subgroup analyses of patients who were younger (50 to <75 years old), White, or those with moderate renal function. They were consistent across the spectrum of glycemic control.

Physical activity, education, diet, and other unmeasured confounders may have influenced the findings.
The findings may not be generalizable to other countries.
Dementia and PD tend to be underdiagnosed and underrecorded. The researchers used diagnosis and prescription records to define outcomes, and this may have led to some misclassifications, owing to off-label drug use.
The study used an intent-to-treat analysis and did not include add-on drugs or consider changes in antidiabetic treatment. This could have introduced confounding or limited the study’s generalizability.
There is a long preclinical period for dementia and PD. Excluding cases of less than 1 year’s duration may not have been adequate to prevent a reverse causation bias.
Metformin improves survival for patients with type 2 diabetes, so the competing risk of death might have influenced the findings.

The study was supported in part by Janssen Pharmaceuticals.
One author received funding from Janssen Pharmaceuticals, GlaxoSmithKline, and Ono Pharma. Another author is an employee of Janssen and may hold equity in Johnson & Johnson. The other authors have disclosed no relevant financial relationships.