Metformin, a common option for patients with type 2 diabetes, had previously been shown in observational studies to be associated with improved survival of cancer patients. Those studies mostly involved older patients with cancer who also had diabetes.
These findings have led to trials of the use of metformin for patients with cancer who do not have diabetes, but two lung cancer trials found no effect on survival.
Now this latest trial in breast cancer, which included 3,649 patients with hormone receptor–positive or –negative disease – who did not have diabetes – also found that metformin had no effect on survival.
These results “tell us that metformin is not effective against the most common types of breast cancer and any off-label use [of] this drug for the treatment of these common types of breast cancer should be stopped,” lead investigator and medical oncologist Pamela Goodwin, MD, a breast cancer researcher at the Lunenfeld-Tanenbaum Research Institute in Toronto, said in a press release.
The negative results “underscore the need for well-conducted randomized trials” before observational studies are put into practice, Dr. Goodwin and her team said.
However, the investigators cautioned against extrapolating their results to patients with diabetes, noting that “because metformin is effective in type 2 diabetes, the results ... should not affect the use of metformin” in breast cancer patients who have diabetes.
The study was published online in JAMA.
Patients were enrolled from 2010 to 2013 while undergoing adjuvant treatment – chemotherapy, radiotherapy, hormone therapy, and/or others – following complete resection of T1-3, N0-3 tumors. They were almost exclusively women (mean age, 52.4 years), and almost 90% were non-Hispanic White. They were primarily from the United States and Canada, with some patients from the United Kingdom and Switzerland.
Patients were randomly assigned equally to receive either metformin 850 mg twice daily or placebo for 5 years. Median follow-up was about 8 years.
Among 2,533 patients with estrogen receptor– and/or progesterone receptor–positive disease, the incidence of invasive disease–free survival events was 2.78 per 100 patient-years in the metformin group, vs. 2.74 per 100 patient-years in the placebo arm (hazard ratio [HR], 1.01, P = .93). There were 1.46 deaths per 100 patient-years with metformin, vs. 1.32 with placebo (HR, 1.10, P = .47).
Metformin was stopped early at about 3 years for the 1,116 hormone receptor–negative patients after futility was declared on interim analysis. The incidence of invasive disease–free survival events was 3.58 with metformin, vs. 3.60 with placebo per 100 patient-years (HR, 1.01, P = .92). There were 1.91 deaths per 100 patient-years in the metformin arm, vs. 2.15 in the group that received placebo (HR, 0.89, P = .46).
However, the findings were different and suggested a signal among the small subset of patients (17% of the total) who had HER2-positive disease. There were 1.93 disease-free survival events with metformin per 100 patient-years, vs. 3.05 events with placebo (HR, 0.64, P = .03), and 0.78 deaths in the metformin arm, vs. 1.43 deaths per 100 patient-years in the placebo arm (HR, 0.54, P = .04).
The benefit seen in this HER2-postive subgroup was limited to patients with any C allele of the rs11212617 single-nucleotide variant.
This was an exploratory analysis, so the results need to be confirmed in a randomized trial, but it’s possible that metformin “could provide an additional treatment option for HER2-positive breast cancer,” Dr. Goodwin said.
Grade 3 or higher adverse events were more common with metformin (21.5% vs. 17.5%). The most common such events were hypertension (2.4% vs. 1.9%), irregular menses (1.5% vs. 1.4%), and diarrhea (1.9% vs. 0.8%).
The study was conducted by the Canadian Cancer Trials Group and was funded by the Canadian Cancer Society, the National Cancer Institute, and others. Dr. Goodwin has disclosed no relevant financial relationships. Several coauthors reported ties to Pfizer, Eli Lilly, Roche, and a number of other companies. One coauthor is an AstraZeneca employee.
A version of this article first appeared on Medscape.com.