CONTEXT: Androgen deprivation therapy (ADT) for prostate cancer (PCa) leads to a selective loss of leg muscle function during walking. Rodent models of ADT have demonstrated that the levator ani is exquisitely androgen sensitive. OBJECTIVE: To determine whether the high androgen responsiveness of the levator ani muscle documented in rodents is evolutionarily conserved and ADT is associated with a selective loss in leg muscle volume. DESIGN: Prospective longitudinal case-control study. SETTING: Tertiary referral hospital. PARTICIPANTS: Thirty-four men newly beginning ADT and 29 age-matched controls with PCa. MAIN OUTCOME MEASURES: The muscle volumes in liters of the levator ani and primary muscles involved in walking (iliopsoas, quadriceps, gluteus maximus, gluteus medius, calf). RESULTS: Compared with controls, during a 12-month period, men receiving ADT experienced a mean reduction in total testosterone from 14.1 to 0.4 nmol/L and demonstrated greater decreases in levator ani [mean adjusted difference (MAD), -0.005 L; 95% CI, -0.007 to -0.002; P = 0.002; -16% of initial median value], gluteus maximus (MAD, -0.032 L; 95% CI, -0.063 to -0.002; P = 0.017; -5% of initial median value), iliopsoas (MAD, -0.005 L; 95% CI, -0.001 to 0.000; P = 0.013; -5% of initial median value), and quadriceps (MAD, -0.050 L; 95% CI, -0.088 to -0.012; P = 0.031; -3% of initial median value). No substantial differences were observed in the gluteus medius and calf muscles. CONCLUSIONS: The androgen responsiveness of the levator ani appears to be evolutionarily conserved in humans. ADT selectively decreases the volume of muscles that support body weight. Interventional strategies to reduce ADT-related sarcopenia and sexual dysfunction should assess whether targeting these muscle groups, including the pelvic floor, will improve clinical outcomes.