Cholesterol and triglycerides were measured in plasma samples from patient with cancer of the prostate before and after 3 months treatment with either Premarin, Provera, Provera and diethylstilbestrol, or diethylstilbestrol alone. Cholesterol was also measured before and after one of three doses of diethylstilbestrol or placebo. Pretreatment cholesterol levels at 196 +/- 1.3 mg per 100 ml (X +/- SE, N = 1093) were significantly lower than these reported for similar age group noncancer controls. Significant increases occurred with some of the estrogen treatments. Pretreatment cholesterol levels showed a significant negative correlation with age in Stage III and IV patients of both studies and a positive correlation with hemoglobin in Stage III patients of both studies. Pretreatment triglyceride levels at 120 +/- 1.9 mg per 100 ml (X +/- SE, N = 1089) were similar to levels reported for noncancer controls of similar age. Estrogen treatment produced a significant increase in triglyceride levels. Serum triglycerides were significantly correlated with hemoglobin, weight, and cholesterol and negatively correlated with age, Analysis of covariance for both cholesterol and triglycerides showed highly significant treatment effects, but no stage effects and no stage-treatment interactions. It showed that the pretreatment value is of extreme importance for predicting or explaining the 3-month value. Death rates were calculated by level of pretreatment cholesterol or pretreatment triglycerides for all Stage II and IV patients, all treatments combined, and for Study 2 and Study 3 separately. No consistent trends were evident for cholesterol. Spearman correlation coefficients between category of initial triglyceride value and rank of death rate were computed to test for a quadratic effect. When the absolute values of the initial triglyceride values minus the overall mean were correlated with the death rate, a significant negative correlation was found for all causes of death and for deaths due to cardiovascular disease and prostatic cancer. These results indicate that the death rate is highest near the overal mean for initial triglyceride values and decreases as the initial values deviate above or below the mean. Initial triglyceride levels appear to have potential as indicators of risk of death in patients with prostatic cancer. The percentage of patients dead at 1 year by initial triglyceride levels, measured only in Study 3, revealed a pattern similar to that observed for the death rate, that is, the highest percentages were associated with values near the overall mean.
