Response

Extract

Prompted by substantial regional differences in the association between induced abortion and risk of breast cancer, we attributed the overall 90% increased risk in our study (1) largely to underreporting of abortion by healthy subjects. Dr. Brind and colleagues argue that the small number of subjects exposed to induced abortion (12 of 225 case patients and one of 230 control subjects) in the southeastern regions does not justify this conclusion.

We agree with them that subgroup analyses based on small numbers increase the probability of chance findings. However, the choice for comparing the two regions was not arbitrary. Rather, it was based on a sound hypothesis: Populations with different religions and attitudes toward induced abortion may differ in their willingness to report induced abortions. Indeed, we ended up with small numbers in the southeastern region, but precisely these numbers were found to have a large impact on the estimated relative risk (RR) of breast cancer after induced abortion (all regions RR = 1.9 and 95% confidence interval [CI] = 1.1-3.2, versus western regions RR = 1.3 and 95% CI = 0.72.6).