medwireNews: Findings from two studies published in JAMA Neurology suggest that glucose-lowering with glucagon-like peptide (GLP)-1 receptor agonists may reduce the risk for dementia in patients with type 2 diabetes, whereas the results were contradictory for sodium-glucose cotransporter (SGLT)2 inhibitors.
“While we have strong evidence demonstrating the glycemic control, weight loss, and cardioprotective benefits of GLP-1 [receptor agonists] like semaglutide, we are learning more about the neuroprotective effects of this class of medications,” observes Diana Thiara (University of California San Francisco, USA) in an accompanying editorial.
The first study, by Jingchuan Guo (University of Florida College of Pharmacy, Gainesville, USA) and colleagues, used a target trial emulation approach to assess electronic healthcare data on 92,160 adults aged 50 years or older with type 2 diabetes from the OneFlorida+ Clinical Research Consortium.
The patients, none of whom had a prior diagnosis of Alzheimer disease or related dementia (ADRD), were divided into three cohorts: GLP-1 receptor agonist use versus other glucose-lowering drugs (GLDs; n=33,858; mean age 65 years; 53.1% women); SGLT2 inhibitor use versus other GLDs (n=34,185; mean age 65.8 years; 50.7% men), or GLP-1 receptor agonist versus SGLT2 inhibitor use (n=24,117; mean age 63.8 years; 51.7% women).
The other GLDs included sulfonylurea, thiazolidinedione, dipeptidyl peptidase-4 inhibitor, α-glucosidase inhibitor, or meglitinide.
The patients were followed up until the onset of ADRD, death, or the end of the study, whichever came first, and this ranged from a mean of 1.95 years to 3.76 years.
The results showed that GLP-1 receptor agonist use was associated with a significant 33% lower risk for ADRD, identified using clinical diagnostic codes, than other GLDs, while SGLT2 inhibitor use was associated with a 43% lower risk.
Specifically, after adjusting for baseline covariates, the incidence rates of ADRD per 1000 person–years in patients using GLP-1 receptor agonists compared with other GLDs were 4.35 versus 6.60. For participants using SGLT2 inhibitors compared with other GLDs, they were a corresponding 4.19 versus 7.23.
“However, when directly compared, there was no significant difference in ADRD risk between GLP-1RA and SGLT2i users,” the team reports.
The incidence rates per 1000 person–years were 3.65 for the GLP-1 receptor group and 3.74 for the SGLT2 inhibitor group, giving a nonsignificant hazard ratio of 0.97.
Consistent findings were seen across sensitivity and subgroup analyses, “further strengthening the reliability of the results,” say Guo et al, and suggesting that “the potential neuroprotection of GLP-1 [receptor agonists] and SGLT2 [inhibitors] may be applicable across diverse patient populations.”
The second study was a systematic review and meta-analysis that included 26 randomized trials comparing the effect of cardioprotective glucose-lowering therapy versus placebo and included outcomes data on dementia or cognitive impairment or change in cognitive score.
A total of 164,531 participants (65.1% men) aged an average of 64.4 years were included. In all, 12 trials assessed SGLT2 inhibitors, 10 assessed GLP-1 receptor agonists, and one trial evaluated pioglitazone.
The results showed that glucose-lowering therapy overall was not associated with a significant reduction in dementia or cognitive impairment compared with placebo over a mean follow-up of 31.8 months (0.12 vs 0.14%). However, there was a significant 45% reduction with GLP-1 receptor agonists, but not with SGLT2 inhibitors (odds ratio [OR]=1.2) or pioglitazone (OR=0.83).
When the researchers looked at the risk for different subtypes of dementia, they found that vascular dementia event rates over a mean follow-up of 35.7 months were numerically but not significantly lower with glucose-lowering therapy versus placebo (0.01 vs 0.03%), whereas AD rates over a mean follow-up of 37.1 months were comparable (0.09 vs 0.09%). In both cases, the findings were consistent across the different types of glucose-lowering therapies.
Catriona Reddin (University of Galway, Ireland, UK) and colleagues note, however, that event rates for the dementia subtypes were low, due to “the average age of the study population, limited duration of follow-up and method of ascertainment (as adverse events rather than systematically sought) reducing power to detect differences in treatment effect.”
They say that the findings “may have implications for choice of glucose-lowering therapy in patients with diabetes and higher risk of dementia,” but add that large randomized clinical trials are needed to validate the findings.
Indeed, researchers on both studies recommend caution in interpreting the findings, with Guo highlighting that their findings should by “considered hypothesis generating rather than definitive.”
Thiara highlighted in her editorial the advent of “next-generation drugs” that incorporate dual or triple agonism, such as glucose-dependent insulinotropic polypeptide receptor agonists and glucagon agonists, alongside GLP-1 inhibitors. These could “further enhance brain health, improve vascular integrity, and potentially provide stronger protection against conditions like cognitive impairment, vascular dementia, and ADRD,” she says.
The editorialist concludes that it will be “crucial to study these newer medications individually rather than grouping them with older drugs in the same class, as their effects may differ significantly.”
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2025 Springer Healthcare Ltd, part of the Springer Nature Group
JAMA Neurol 2025; doi:10.1001/jamaneurol.2025.0353
