Selections from current literature: abortion, breast cancer and causal inferences.

The association between abortion and breast cancer was first reported in 1981 by Pike. Since that time there have been over 35 publications worldwide that have evaluated this association in an attempt to definitively answer the question of whether abortion is an aetiologic factor in the multicausal web of breast cancer. The results of these studies, both case control and cohort, have been mixed. Despite these contradictory results, the problem of a rising incidence and prevalence of breast cancer remains. In women, breast cancer is the most common type of cancer in the USA and the second leading cause of cancer mortality. In 1995, it is estimated that there will be 182 000 new cases of breast cancer and 46 000 deaths as a result of breast cancer in America. Among the industrialized countries of the world, the age adjusted death rate from breast cancer is in excess of 20 per 100 000 population. However, perhaps more controversial than the mixed results of the studies evaluating the association between abortion and breast cancer, is the debate over abortion rights. As often happens in politics, good science gets twisted to favour one position over another. The net result of these manipulations can be bad public policy based on poor logic, by individuals not versed in the subtleties of epidemiology or scientific reasoning. In this Selections, five recent articles addressing the issue of abortion and breast cancer will be reviewed. Following this, a classic paper outlining how causal inferences should be made will be summarized. Then, without considering the politics of abortion, but instead focusing only on the available scientific evidence, a judgement will be made on whether or not abortion has been shown to have a causative role in breast cancer.