Cancer is one of the leading causes of death, and as such, cancer treatment and prevention research has been a large focus of medical professionals worldwide. Over the past few decades, several studies have proposed that metformin—a medication widely used for type II diabetes management—is a potential preventative measure for cancer. 

Unfortunately, the truth is not so glamorous. A recent study published in the Journal of Clinical Epidemiology and led by Samy Suissa, a professor in McGill’s Departments of Epidemiology and Biostatistics and Medicine, sought to put misleading claims about metformin to rest. 

“Many studies were appearing reporting that metformin, a treatment for diabetes, could lower the rates of several different cancers,” Suissa wrote to The Tribune. “What got us interested in this issue is that the reported reductions in cancer in these studies were spectacular, simply too good to be true.”

But what caused the misinformed results from these studies in the first place? Suissa traced the misleading results to a study error known as ‘immortal time bias,’ which can inflate the perceived effectiveness of a drug. He pointed to misinformation surrounding metformin and breast cancer as an example.

“Immortal time bias occurs when follow-up of breast cancer patients starts before they initiate metformin treatment. So those that receive metformin are ‘immortal’ between the start of cancer and the start of metformin; they must be alive to receive the metformin.” Suissa wrote. “The problem with immortal time bias is that all these studies started at breast cancer, but then looked at metformin use in the future, thus inherently giving a survival advantage to the women who start metformin.”

In other words, not every breast cancer patient survives long enough to actually reach the point of beginning metformin medicine. However, researchers may retroactively look at metformin use in breast cancer, and not account for the time in the interim when patients stayed alive while not using metformin. In this case, it could seem as though the metformin had a hand in treating the cancer, when in reality it was only useful for those who were able to live long enough to reach the point of taking it. This bias mischaracterizes the effectiveness of metformin as a breast cancer treatment.

Most earlier studies on metformin were observational, with no randomization occurring, and researchers only monitored the health progression of their patients without intervening. As such, Suissa stressed the importance of randomized trials—when participants are distributed by chance into different treatment groups—for producing accurate, unbiased results.

“Many scientists started to conduct randomized trials of metformin as a treatment for cancer, with the largest involving 3,600 women with breast cancer,” Suissa wrote. “Half received the usual treatment for breast cancer along with metformin, while the other half received the usual treatment. They were followed for more than five years and the study found no benefit for the metformin group.” 

Suissa has already begun examining potential misinformation in other major pharmaceutical drugs, and emphasized that both observational and randomized trials are necessary to reduce medical and scientific misinformation.

“Many such flawed studies are now appearing in many medical journals on the potential extracurricular benefits of GLP1 receptor agonists, the family of drugs that the popular Ozempic belongs to,” Suissa wrote. “The studies are also suggesting that these drugs are very effective at reducing all kinds of diseases, including cancer [….] Before expensive randomized trials are undertaken to test this hypothesis, we are undertaking properly done observational studies using rapid cutting-edge methods to confirm that the hypothesis is tenable. We wish to ensure that any time-consuming randomized trials are based on solid data and avoid a repetition of the metformin failure.”

Suissa’s studies will prevent future misinformation regarding drugs such as Ozempic. This will ensure that Ozempic and similar pharmaceuticals are only prescribed for conditions that they are completely able to alleviate or prevent.