A retrospective study conducted by Dr. Vahit Ozmen and his colleagues at the Istanbul Medical Faculty and Magee-Women's hospital reported a statistically significant 66% increase in breast cancer risk among women who'd had any abortions.
According to Joel Brind, professor of endocrinology at Baruch College, City University of New York and a director at the Breast Cancer Prevention Institute, Ozmen's team most likely underestimated the breast cancer risk associated with abortion because of a flaw known as "selection bias."
Selection bias would also explains their team's unusual findings - significantly decreased risks for women who use oral contraceptives (OCs) and hormone replacement therapy (HRT). The World Health Organization and the NCI acknowledge that use of combined (estrogen + progestin) OCs and combined HRT increase risk.
Selection bias is a flaw in the study because only hospital or clinic patients were selected as study subjects, and they were therefore not representative of the general population. According to Brind's hypothesis, a disproportionate number of "modern" women were likely represented among the controls, a group more likely to use HRT and OCs, have abortions and visit the hospital often for minor complaints. By contrast, a disproportionate number of "traditional" women were represented among the patients; women less likely to use HRT and OCs, have abortions and visit the hospital (except in cases of serious illness, like breast cancer).
To their credit, Dr. Ozmen et al. did acknowledge the likelihood of selection bias in their study, although they were not specific in attributing any effects on their results to it.