New research this week is the latest to suggest that some drugs meant to treat type 2 diabetes might also help prevent dementia in high-risk groups. The study found that older people who used a relatively older class of antidiabetic medication were less likely to develop dementia than those who used other common antidiabetic drugs. The findings might warrant new clinical trials to confirm the untapped potential of these drugs, the authors say.

Dementia is the progressive and ultimately fatal loss of cognitive function, which steadily impairs a person’s ability to remember, reason, and eventually perform basic tasks like eating. There are several different forms of dementia and many different factors that make people more vulnerable to developing it, including our genetics. One major suspected contributor to dementia risk is type 2 diabetes. And that’s led some researchers to wonder if drugs that can successfully manage diabetes can also lower a person’s risk of dementia.

Unfortunately, the research on this connection has been mixed, with some studies showing that popular diabetes drugs like metformin are linked to a lower risk of dementia, but other studies show no potential benefit. In this new research, the researchers decided to look at a class of diabetes drug that’s received less attention in the field of dementia research, drugs known as thiazolidinediones, or TZDs.

The team analyzed the medical records of over a half million people with type 2 diabetes aged 60 and older who were given care through Veterans Affairs, the largest interconnected healthcare system in the U.S. They compared the long-term outcomes of people who took at least one year of TZDs to people taking other diabetes drugs like metformin and another class of drugs called sulfonylureas. These people were diagnosed free of dementia before they started the medication, and were tracked for an average length of nearly seven years. People taking only metformin were used as the control group since the drug is the most common front-line option for type 2 diabetes and people who haven’t been prescribed medication for their diabetes tend to be healthier than the typical patient.

Over the study period, those taking TZDs alone were 22% less likely to be diagnosed with any form of dementia when compared to those taking only metformin. They also found that these people were 11% less likely to be diagnosed specifically with Alzheimer’s disease, the most common form of dementia. Additionally, patients were 57% less likely to develop vascular dementia, which accounts for about 10% of cases. The findings were published Tuesday in BMJ Open Diabetes Research & Care.

There can be many different causes of dementia, and it’s likely that any preventive effect from TZDs or other drugs would be complicated as well. But circulation problems are common for diabetes patients and are a known risk factor for vascular dementia and Alzheimer’s, so the researchers argue that TZDs might mainly work to prevent dementia by helping promote healthier circulation.

These results show a correlation between taking TZDs and a lower risk of dementia, but not a definitive cause-and-effect. And TZDs aren’t free of their own concerns. These drugs were developed in the 1990s and are still often used today—but most often as a second-line option. In the mid-2000s, some research suggested that certain TZDs could raise the risk of heart attacks, which led to the Food and Drug Administration attaching a warning label to the TZD drug rosiglitazone. Later studies failed to confirm this connection and the warning label was eventually removed. But more recent research has reignited this debate, and the drugs are still not recommended for patients at higher risk of heart failure.

All that said, the authors say that the findings should spur more research into the potential benefits of these drugs for dementia, including clinical trials. And assuming that they’re validated, the research could inform how doctors treat diabetes patients at high risk of dementia. The team found, for instance, that people taking sulfonylureas alone were actually more likely to develop dementia than those taking metformin or TZDs. If these drugs do somehow increase the odds of dementia, the researchers say, then it might be worth pairing them with metformin or TZDs to offset that risk. The possible effect of TZDs in preventing dementia was also larger in people who were overweight or obese, suggesting that these patients would benefit the most. Other, newer diabetes drugs are also being studied for their potential in preventing dementia.

“Our findings provide additional information to aid clinicians’ selection of [antidiabetic medications] for patients with mild or moderate type 2 diabetes and are at high risk of dementia,” the authors wrote.