Smoking and breast cancer:
an uncertain link

Alcohol could be the spoiler in this massive study

By Marilyn Ostrosky

There aren't many diseases for which smoking is not yet a proven risk factor, but breast cancer is one of them. Some previous research has even suggested a small protective effect from tobacco. But women who smoke may breathe a little less easily if they read the results of a large prospective study from California, which found that smoking increases the risk of breast cancer by 32%.

The study, led by the California Department of Health Services, is published in the Journal of the National Cancer Institute. It analysed data from 116,544 female members of the California Teachers Study cohort, none of whom had a breast cancer diagnosis at the study's outset in 1995. Over the next five years, 2005 of these women developed breast cancer.

The rate among women who were current smokers was 32% higher than among those who had never smoked. Heavy smokers, women who started smoking before the age of 20, and those who started at least five years before their first pregnancy appeared to be most at risk. The authors speculate that prior smoking may diminish the known protective effect of breastfeeding.

There's no obvious biological mechanism to explain how smoking could provoke breast cancer, but it is possible that toxins produced by tobacco smoke are stored in the fatty tissues of the breast. If so, the damage does not appear to be irreversible. The data showed no significant elevation of breast cancer risk in former smokers.

The researchers also looked for an effect among non-smokers who listed themselves as being exposed to household passive smoking, but they found no significant difference between this group and those who considered themselves completely smoke-free. In fact never-smokers who were exposed to passive smoke were fractionally less likely to develop the disease than those who were not exposed.

With over 116,000 participants, the study has plenty of brute statistical power, but it's received a lukewarm reception from many experts, who say it fails to control for a crucial confounding factor, namely, the heavier drinking typically observed in smokers.

Cancer Research UK, the organization formed in 2002 from the amalgamation of Britain's biggest cancer charities, conducted a meta-analysis a year ago of 53 international breast cancer studies. They deliberately set out to separate the effects of alcohol and tobacco.

Sir Richard Doll, the man who first postulated a link between smoking and cancer in a famous 1950 paper, was a co-author of this research, published in the British Journal of Cancer. He said: "For the first time we have undertaken a study large enough and detailed enough to look at the separate effects of tobacco and alcohol reliably. When we did this we found that drinking, but not smoking, increases the risk of breast cancer."

Professor Valerie Beral, of Cancer Research UK, said: "Smokers tend to drink more alcohol than non-smokers, the risk of breast cancer is clearly related to alcohol consumption, and the analyses of the association between breast cancer and smoking among women who drink are inextricably confounded by the effect of alcohol. Analyses restricted to non-drinkers avoid confounding by alcohol and show no increase in the risk of breast cancer among smokers." She cautioned, however, that even if the breast cancer link is a red herring, smoking can still be blamed for one-third of all cancers.