To establish whether absolute prostate-specific antigen (PSA) value after androgen deprivation (AD) is prognostic in metastatic (D2) prostate cancer (PCa).

D2 PCa patients with baseline PSA of at least 5 ng/mL received 7 months induction AD. Patients achieving PSA of 4.0 ng/mL or less on months 6 and 7 are randomly assigned to continuous versus intermittent AD on month 8. Eligibility for this analysis required a prestudy PSA with at least two subsequent PSAs and that patients be registered at least 1 year before analysis date. Survival was defined as time to death after 7 months of AD. Associations were evaluated by proportional hazards regression models.

One thousand one hundred thirty four of 1,345 eligible patients achieved a PSA of 4 ng/mL or less. At end of induction, 965 patients maintained PSA of 4 or less and 604 had a PSA of 0.2 ng/mL or less. After controlling for prognostic factors, patients with a PSA of 4 or less to more than 0.2 ng/mL had less than one third the risk of death (ROD) as those with a PSA of more than 4 ng/mL (P < .001). Patients with PSA of 0.2 ng/mL or less had less than one fifth the ROD as patients with a PSA of more than 4 ng/mL (P < .001) and had significantly better survival than those with PSA of more than 0.2 to 4 ng/mL or less (P < .001). Median survival was 13 months for patients with a PSA of more than 4 ng/mL, 44 months for patients with PSA of more than 0.2 to 4 ng/mL or less, and 75 months for patients with PSA of 0.2 ng/mL or less.

A PSA of 4 ng/mL or less after 7 months of AD is a strong predictor of survival. This data should be used to tailor future trial design for D2 prostate cancer.