New Kink in the Link Between GLP-1 Drugs and Cognition

Adults whose type 2 diabetes was treated with GLP-1 receptor agonists were more than likely to develop cognitive impairment over 10 years than their counterparts not treated with GLP-1 agents, a propensity-matched retrospective study of nearly 65,000 patients suggested.

Durable cognitive impairment — defined as vascular dementia, Alzheimer’s disease, or mild cognitive impairment — occurred twice as frequently in diabetes patients who used GLP-1 drugs (2.6% vs 1.3%, HR 2.74, P<0.0001), but mortality risk was lower with the drugs (3.9% vs 8.2%, HR 0.68, P<0.0001), according to Isaac Thorman, ScM, an epidemiology researcher at the Johns Hopkins University School of Medicine and a third-year medical student at New York Medical College in Valhalla, and colleagues.

On a compound outcome that assessed both cognitive impairment and mortality, there was no significant difference between GLP-1 receptor agonist users and non-users (6.1% vs 9.1%, HR 0.98, P=0.39), Thorman reported at a late-breaking science session at the American Academy of Neurology annual meeting in Chicago.

Overall, GLP-1 analogs were associated with an increased risk of cognitive impairment, secondary to a larger, protective effect against mortality, Thorman noted.

“We interpret this to mean that GLP-1 analog recipients lived significantly longer than non-recipients, and that they lived long enough for them to develop cognitive impairment,” he said.

The findings come on the heels of two phase III trials showing that Alzheimer’s patients treated with the GLP-1 agent semaglutide (Rybelsus) had no significant improvement in cognitive or functional decline over 2 years compared with placebo, Thorman noted.

“The apparent survival paradox demonstrated here, plus our unprecedented sample size and long-term follow-up, may explain the non-significance found in the randomized controlled trials,” he said.

The idea of repurposing semaglutide to treat Alzheimer’s was driven by real-world data and preclinical research. In observational analyses, GLP-1 receptor agonists were tied to a reduced risk of cognitive impairment and dementia, including a lower risk of Alzheimer’s diagnoses. The drug class also demonstrated neuroprotective and anti-inflammatory effects in animal models of Alzheimer’s disease.

“Our objective was to assess the long-term risk of cognitive impairment associated with GLP-1 analog use in older adults with type 2 diabetes,” Thorman said.

Thorman and co-authors studied people ages 50 and older with type 2 diabetes in the TriNetX dataset of patients from 115 healthcare organizations in five countries. People with pre-existing cerebrovascular disease or durable cognitive impairment were excluded from the study. Patients were followed for up to 10 years.

From a sample of 404,084 middle-age and older adults, the researchers matched 64,530 diabetes patients on more than 170 variables, including demographics, vital statistics, comorbidities, and medications prescribed.

A critical distinction between this study and other observational data “is likely in how we accounted for overall patient health,” Thorman told MedPage Today. “At baseline and prior to matching, patients receiving GLP-1 analogs were in significantly worse health, with significant elevations in almost every lab value, diagnosis, and medication. After matching, there were few statistically significant differences, and almost all of these differences were devoid of clinical significance.”

In women, the mortality protection of GLP-1 agents balanced cognitive impairment risk; whereas in men, it outweighed the risk of cognitive impairment. The researchers also found no overall protective effect of GLP-1 drugs in diabetes patients ages 80 and older.

When patients are in their 80s and are being followed up for another 10 years into their 90s, other comorbidities may influence cognitive deterioration, observed Paul Edison, MD, PhD, of Imperial College London, who wasn’t involved with the study.

Whether robust measurements were used to assess cognitive function also is a question, Edison pointed out. “A thorough investigation into these factors is important before reaching a conclusion,” he told MedPage Today.

Long-term, prospective surveillance and results from randomized trials remain critical in assessing the risk-benefit ratio of these drugs, Thorman noted. “Caution is advised when interpreting these findings, as causality cannot be inferred from this retrospective analysis,” he said.