June 16, 2000 -- Ever since some studies in animals showed that antidepressants might increase the risk of breast tumors, researchers have been looking for a similar link in humans. Now a group of Canadian researchers reports that its study of more than 5,000 women found that those who had taken the drugs for at least two months had a greater probability of having breast cancer than those who had not.

This finding, which the researchers discussed at a meeting of the Society for Epidemiologic Research in Seattle, adds support to a study done by the same researchers published this spring in the American Journal of Epidemiology. That study looked at a much smaller number of women. In it, researcher Nancy Kreiger and her colleagues at the University of Toronto reported that women who took an older class of drugs, called tricyclic antidepressants, for two years had twice the risk of breast cancer. And those taking Paxil, one of a class of drugs called serotonin reuptake inhibitors (SSRIs), had seven times the risk. But because of the small number of patients in that study, the true effect of the drugs couldn't be confirmed.

In the more recent study, a substantially smaller link was seen between antidepressants and breast cancer.

"The reason we looked at this is that a psychiatrist, who is also one of the co-investigators, noticed that in mouse studies, antidepressants may cause mammary tumors," one of the researchers, Michelle Cotterchio, PhD., tells WebMD.

"The results could be the result of chance alone. That's why future studies need to be done. It takes five to 10 studies in different subgroups before people start believing the findings."

Indeed, several critics have pointed out what they consider to be shortcomings in the studies, or problems in the interpretation of the results.

The animal studies were prompted by the recognition that the antidepressant drug molecule is similar to that of other chemicals that are known to cause cancer. These agents appear to inhibit an enzyme associated with cancer, explains Judith Parsells Kelly, PhD. The next step was to do so-called epidemiologic studies, in which the incidence of cancer in women who had used antidepressants was compared with that in those who hadn't used the medications.

"Epidemiology is really a rather crude tool, and it's effective for identifying increases in risk when the magnitude of the risk is quite high," Kelly says. But that is not the case with antidepressants and breast cancer. Kelly was not involved in the Canadian study, but she has studied links between drugs such as SSRIs, tricyclic antidepressants, and antihistamines and cancer.

Several physicians who spoke to WebMD raised the question of whether depression alone could have been responsible for the increased risk Krieger and colleagues found.

"There's a general theory that depression may make people susceptible to cancers in general, because it can decrease immune function. There has not been any strong data to support that. On the other hand, it hasn't been looked at in a well-designed study," says Bruce Trock, PhD, associate professor of medicine and oncology at Georgetown University School of Medicine in Washington.

Krieger's findings raise other several questions, Trock says, such as: How does the number of women who took antidepressants and didn't get cancer compare to the general population? Were the researchers able to eliminate the impact of known breast-cancer risk factors?

For these reasons, more studies need to be done. At this point, Kelly says, "Women should not avoid taking antidepressants because they're afraid of developing breast cancer. The evidence is not strong enough to justify that type of action."

Says Charles Loprinzi, MD, an oncologist at the Mayo Clinic in Rochester, Minn., who has studied the use of Prozac to relieve hot flashes in breast cancer survivors: "The bottom line must be that you can't put much credence in [these findings]. The numbers are small and consistent with what might have occurred by chance. This is the same sort of risk that people have talked about with alcohol use. It's very, very small indeed."