One hundred eighty-three patients with endometrial carcinoma had both immunohistochemical and biochemical analysis of estrogen receptor (ER) and progesterone receptor (PR) content of the tumor. Biochemical analysis was done on a homogenate of uterine curettings; immunohistochemical stainings were done on formalin-fixed, paraffin-embedded sections. The biochemical method was scored quantitatively and the immunohistochemical, in a semiquantitative way. There was a significant correlation between the immunohistochemical categories of the malignant component of the tumor and the biochemical levels of receptor content (P = 0.0001). The sensitivity of the immunohistochemical method for the ER content was 78.5% and the specificity was 58.2%. The sensitivity for the PR was 52.5% and the specificity was 92.8%. Changing cut-off levels for positivity of the biochemical analysis changed the sensitivity and specificity. Survival was predicted by the immunohistochemical status of the ER (P = 0.001) and PR (P = 0.01). Similarly, it was predicted by the biochemical receptor status. Multivariate analysis of the immunohistochemical receptor status and the biochemical receptor status showed that the immunohistochemically determined estrogen status was the most significant predictor of survival (P = 0.001). The immunohistochemical analysis of sex steroid receptor status on formalin-fixed, paraffin-embedded tissue is not only possible and practical, but also predicts survival.