In a population-based randomized study comparing 150 patients with advanced prostatic cancer treated with orchiectomy or estrogen, some possible prognostic factors were analyzed. The observation period was 78 to 114 months. M category, T category, tumor grade, performance status, pain, prostatic acid phosphatase, sedimentation rate, hemoglobin and serum creatinine level were all statistically significantly related to the interval to progression and to disease-specific death on univariate analyses. Variables that were statistically significant on multivariate analyses were M category, T category, sedimentation rate and patient age. The sedimentation rate predicted the intervals to progression and to disease-specific death, with the relative hazard and 95% confidence interval for the latter end point being 1.018 (range 1.010 to 1.027) for each millimeter increase in sedimentation rate per hour. An analysis that was stratified according to the extent of the disease as measured on a bone scan showed that the sedimentation rate was the only prognostic factor of statistical significance with an estimate of relative hazard of 1.52 if the sedimentation rate was elevated 20 mm. per hour. Progression-free survival but not disease-specific survival was statistically significantly better in the estrogen group (relative hazard 0.47), as assessed by multivariate analysis in which all variables were taken into account.