Oral contraceptive use and the risk of breast cancer [letter]

The findings of the cancer and steroid hormone study suggested no increased risk of breast cancer associated with oral contraceptive (OC) use but some potential sources of bias were not discussed by the authors. A possible protective effect of OC use may have been obscured by such biases. Bias in the detection of breast cancer may occur because OC users generally undergo closer medical surveillance than women who do not use OCs. If asymptomatic breast cancer is more likely to be detected in OC users because they are more frequently and carefully examined then OC users will be overrepresented in a group of women with breast cancer as compared with a group without that diagnosis. If this source of bias applies to the current study the odds ratio is erroneously high -- the actual risk may be significantly less than unity. The cases and controls in the study did not undergo equal diagnostic examinations. Another concern is potential bias in the ascertainment of exposure to OCs. A standardized questionnaire was used to interview subjects but the interviewers may have been aware of the subjects status as cases or controls as well as of the study hypothesis. The exclusion of certain cases and controls also may have affected the results although the direction of potential bias so introduced is unknown.