Monday, February 16, 2004

Breast Cancer and Antibiotic use: responsible reporting!

Study Suggests Breast Cancer Is Linked to Use of Antibiotics
Frequent use of antibiotics has been linked to a greater risk of breast cancer, say researchers who studied thousands of American women and found that those who took the drugs most often had twice the risk of the disease.

The study uncovered a relationship between greater use of antibiotics and a heightened risk of breast cancer, but researchers sought to temper their findings by cautioning that they had only highlighted an association, not a causal link.

Astoundingly this was reported responsibly! The journalist made clear this was an association, and that most researchers don't think the antibiotics are causing breast cancer. Instead the interest is what's special about these women that they use so many antibiotics and/or that they are so susceptible to infection requiring antibiotic therapy.

Since I first read this news report I actually did read the JAMA article:JAMA -- Abstracts: Velicer et al. 291 (7): 827. They found a very strong statistical association between increasing use of oral antibiotics and increased risk of breast cancer -- about doubling risk -- even after "ruling out" other associated factors.

This is roughly similar to the efffect of other things known to increase breast cancer risk, such as frequent periods (infertility, etc). However there are lots of issues and oddities with this study:

1. These association studies don't have a great track record. Over the past 20 years I've guess more than 50% don't pan out. It usually turns out that the there was a missed cofactor that was the real agent.

2. WHY were these women using SO many antibiotics? Did they all have something wrong with their immune system? Did they have some "bad habits" -- like smoking or alcohol use that predisposed them to infection?

3. Because of where they got their data, the researchers have no information on either smoking or alcohol use. Both of these will increase antibiotic use, and both are risk factors for breast cancer. Smoking is getting more attention lately.

4. ALL the antibiotics had very similar effects, despite being very different medicines with very different actions. This suggests that the real cause was not the antibiotics, but something related to these women's need for them.

5. Do we see a relationship in animal models between antibiotic use and breast cancer?

My gut suspicion is that this is spurious, and and that smoking and/or alcohol use are the real actors that are producing an association between breast cancer and antibiotic use.

No comments: