A cheap and relatively safe diabetes drug, metformin, might have cancer-fighting properties, according to an international study - but findings fall short of proving it actually can stave off cancer.

Researchers, whose findings were published in the Journal of Clinical Oncology, found that women with diabetes who took the medication had a 25 percent lower risk of developing breast cancer over more than a decade of follow-up.

"Metformin use in postmenopausal women with diabetes was associated with lower incidence of invasive breast cancer," wrote lead researcher Rowan Chlebowski, at the Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center.

"These results can inform future studies evaluating metformin use in breast cancer management and prevention."

Metformin, also sold under the brand name Glucophage, has been on the market for many years and is generally considered safe, although five to 10 percent of patients experience side effects like nausea and bloating.

It is used by millions of type 2 diabetics every day to help control their blood sugar, and studies have shown it also shrinks lung and breast tumors in mice. Several reports show people taking it for diabetes appear to develop cancer less often.

The new study used data from about 68,000 postmenopausal women who took part in the U.S. government-funded Women's Health Initiative clinical trials.

Over nearly 12 years of observation, there were more than 3,200 new cases of breast cancer among the women.

Every year, 0.42 percent of women without diabetes developed breast cancer, compared to 0.40 percent of diabetics on metformin and 0.47 percent of diabetics taking other drugs.

After taking into account risk factors for breast cancer, the gap between women without diabetes and diabetics on drugs other than metformin vanished. But diabetics on metformin turned out to have a 25 percent lower cancer risk than their diabetes-free peers.

"This is an area of great excitement," said Pamela Goodwin, a breast cancer expert at Mount Sinai Hospital in Toronto, Canada, who wrote an editorial that appeared with the study.

"The evidence is coming together that metformin may actually have a clinically-relevant effect, but none of this is good enough to chance clinical practice just yet."

But she added that while the study is the best of its kind so far, it relies on observations instead of an actual experiment in which women are randomly selected to take metformin or not.

Goodwin and her colleagues are currently running a trial to test whether metformin can help ward off new tumors in women getting breast cancer treatment. Results are expected in three to four years.