Androgen deprivation therapy (ADT) is a mainstay of prostate cancer treatment to improve overall survival for intermediate- and high-risk localized disease as well as metastatic disease. While ADT improves survival, it can cause significant morbidity and a decrement in quality of life. In particular, ADT is associated with decrease in bone mineral density (BMD) and increased risk of fracture. Although current guidelines recommend continuous androgen deprivation therapy (CAD) as standard therapy for high-risk disease, there has been increasing recognition of adverse effects from CAD. Since 1986, intermittent androgen deprivation therapy (IAD) as alternative therapeutic strategy for prostate cancer has been proposed to delay development of castration resistance and to reduce the side effects of ADT. While both CAD and IAD are commonly used in real clinical practice, no prior study examined BMD change after CAD or IAD, and assessed whether bone loss would recover during off-treatment of IAD. The investigators therefore determine the rate of change in BMD induced by ADT (CAD versus IAD) in men with prostate cancer.

Objective: To determine the rate of bone mass loss induced by two therapeutic strategies of ADT (CAD versus IAD) in men with prostate cancer. Design, setting, and participants: the investigators will perform randomized, open label clinical trial. Men aged over 50 yrs old with prostate cancer (localized, locally advanced, metastatic prostate cancer) who are treated with primary ADT for newly diagnosed prostate cancer or salvage ADT at biochemical recurrence following radical prostatectomy will be included. Participants will be randomly assigned to one of the following treatment arms: Arm 1 (CAD): ADT without any discontinuation during study period (12 months). Arm 2 (IAD): ADT for the first 6 months of study period, if the prostate-specific antigen (PSA) reaches its nadir (\< 4 ng/dL) and serum testosterone reaches castration level (\< 50 ng/dL). Outcomes: Primary outcome: change of L-spine total BMD. Secondary outcomes: change of femur neck BMD, incidence rate of osteoporosis, risk of 10 year major osteoporotic fracture, quality of life based on Expanded Prostate Cancer Index (EPIC) questionnaire. Timing of outcome measurement: at baseline and up to 12 months after randomization. Statistical analyses: student's t test for continuous outcomes and Fisher's exact or chi-square test for dichotomous outcomes.