Sex differences in the outcome of first episode psychosis

Exogenous estrogen use was determined by interview in 206 endometrial cancer patients. 191 women with nonmalignant gynecologic disorders, and 199 women other illnesses. Seperate comparisons of these patients with gynecolgic and otherillnesscontrol subjects provided estimates of the risk of emdometrial cancer among estrogen users, adjusted for the effects of cofounding variables, of 1.5 (95% confidence limits 0.9 and (2.7 and 8.4), respectively, in both comparisons risk increased with increasing duration of use; regardless of duration of use, risks were higher among users of estrogen within 12 months of diagnosis than among past users. The effects of bias, including surveillance bias, on the results were considered and it was concluded that the true risk was not less than case/gynecologic control estimate and was likely to approach the case/other control estimate (Am. J. Obstet Gynecol. 141:658, 1981.)