Measurement error does not explain the persistence of a body mass index association with endometrial cancer after adjustment for endogenous hormones

Abstract

Identified risk factors for endometrial cancer are accepted as operating through estrogen exposure. In a recent analysis, the effect of risk factors such as body mass index (BMI) was not explained by circulating estrogen concentrations. In the present analysis, we correct for measurement error associated with obtaining only one blood sample per subject. Applying regression calibration ideas, we found that error correction of log estrone had little impact on estimates of the BMI effect, suggesting that hormone measurement error does not account for the residual importance of BMI. The biologic mechanism for the increased risk associated with BMI remains to be explained.