Regular aspirin use might reduce the risk of cancer by as much as 38 percent, according to a big new review of research on the issue. But "might" is the key word here, other scientists say. And even if it works, that benefit comes with costs, including an increased risk of ulcers and internal bleeding.

Three articles published in The Lancet and Lancet Oncology looked at several hundred studies of people taking aspirin to reduce the risk of heart disease and stroke to see if aspirin use was associated with less cancer.

They found that people who took aspirin daily had a 15 percent lower risk of death by cancer, and a 38 percent reduction in the risk of colorectal and gastrointestinal cancer. Metastasis, or spread of cancer, was 38 to 40 percent less common.

Earlier studies have found that taking aspirin reduces the risk of colon cancer, so that's not a big surprise. What's generating headlines with this group, led by Peter Rothwell at the University of Oxford in England, is that they're saying the benefit extends to other forms of cancer, too.

But people's risk of serious internal bleeding doubled in the first three years of taking aspirin daily. And the reduced cancer risks didn't start to show up until three years out. So even if these results stand the test of time, patients would have to weigh the known risks of bleeding and ulcers against potential benefits.

And other studies that weren't included in the Oxford scientists' review found no cancer reduction benefit from aspirin.

The Women's Health Study followed almost 40,000 women who took low-dose aspirin every other day for 10 years. And the Physicians' Health Study tracked 22,017 men who took regular-strength aspirin every other day for five years. Neither study showed any reduction in cancer risk or death, even for colon cancer.

It's "plausible" that such differences in dosing could explain why the cancer-reducing effect wasn't seen in the Women's and Physician's health studies, write Andrew Chan and Nancy Cook, in a comment in The Lancet. But that's "far from conclusively established," they write. Cook is a biostatistician who worked on both the women's and physicians' studies.

Despite these cautions, Chan and Cook write, "Rothwell and colleagues show quite convincingly that aspirin seems to reduce cancer incidence and death."

Copyright 2016 NPR. To see more, visit