Friday , February 20, 2004
Antibiotic use increases breast cancer risk, according to a new study. But after scrutinizing the study, I’m leaning toward linking grant-hungry researchers and a publicity-hungry medical journal with reprehensible sensationalism.
The study in the Feb. 18 issue of the Journal of the American Medical Association (search) reported that all levels of antibiotic use were associated with increased risk of breast cancer and death from breast cancer.
The study triggered an avalanche of “Antibiotics Linked to Breast Cancer” news reports, most of which were sensibly tempered with “don’t panic,” “don’t stop taking antibiotics” and “more research is needed” caveats.
Researchers compared antibiotic use among 2,266 women with breast cancer (search) and 7,953 women without breast cancer, all of whom belonged to a large health plan in western Washington state. Data on antibiotic use was obtained from the health plan’s records.
Antibiotic use for 1-50 cumulative days reportedly was associated with a 50 percent increase in breast cancer risk. Antibiotic use for more than 1,000 cumulative days reportedly was associated with a 100 percent increase in risk.
The results sound scary, but here’s why they’re not.
No evidence was presented that antibiotics (search) were the biological cause of any of the cases of breast cancer considered in the study. This is no surprise since no demonstrable biological explanation exists as to why antibiotics would cause cancer in the first place.
Without a plausible biological link between antibiotic use and breast cancer, the researchers relied exclusively on statistical analysis, a potentially useful tool provided the data to be analyzed are of reasonable quality. These study data, however, fall way short.
The average study subject was about 60 years old. But study subjects who had taken antibiotics had been enrolled in the health plan for only about 20 years on average. Since the sole source for data on antibiotic use was the health plan’s records, about 40 years of data, on average, about potential antibiotic use were missing for each study subject.
Even more data is missing for the 18 percent of the study subjects who supposedly never took antibiotics. These women had only been in the health plan for about 10 years on average.
But just because they didn’t take antibiotics while they were enrolled in the health plan, doesn’t mean they didn’t take antibiotics before enrollment. Indeed, some of the supposed “never users” could actually have been extremely heavy users of antibiotics prior to enrollment in the health plan.
Since the vast majority of Americans have taken antibiotics at some point, it’s difficult to believe that so many of the study subjects had never taken antibiotics.
This is a crucial data gap since the researchers claim that even a single day of antibiotic use increased breast cancer risk. The absence of complete data on lifetime antibiotic use renders comparisons between antibiotic users and “non-users” meaningless.
The study data are also faulty in terms of level of exposure to antibiotics. The researchers assumed exposure to antibiotics could be measured either by number of antibiotic prescriptions written or by the number of days prescribed for antibiotic use according to prescription records.
But patients commonly fail to complete courses of antibiotics prescribed by their doctors. Patients with a prescription for 10 days of antibiotics may feel better after just a few days and cease taking their medicine. A 10-day prescription, therefore, doesn’t necessarily mean 10 days of use. It may, in fact, mean much less use.
So the researchers really can’t say that more antibiotic use increases breast cancer risk because they really don’t know who took more antibiotics.
Though the researchers acknowledged their study doesn’t prove antibiotics cause breast cancer and that many other possible explanations for their reported results could exist, they called for more research ― that is, more taxpayer-funded research like this study.
That’s pretty twisted, though. The researchers used obviously deficient data to stir up a frightening, but dubious controversy that they hope to milk in terms of continued research funding.
Antibiotics have been used since the 1940s ― to the great benefit of public health. More than 100 million prescriptions are written in the U.S. every year. If antibiotics were a cancer risk, it’s very likely that such a link would have been observed long ago.
Finally, the Journal of the American Medical Association should be ashamed for scaring the public about antibiotics with such a flimsy study. Five years ago Journal editor George Lundberg (search) was fired for trying to garner headlines with another sensational but absurd study.
Heads should roll again.
Steven Milloy is the publisher of JunkScience.com, an adjunct scholar at the Cato Institute and the author of Junk Science Judo: Self-Defense Against Health Scares and Scams (Cato Institute, 2001).
Respond to the Writer