Background: We sought to compare two common salvage strategies for radio-recurrent prostate cancer: androgen deprivation therapy (ADT: PR7 RCT NCT00003653) or local salvage ablation using cryotherapy (CRYO: single institution study Williams, Eur Urol. 2011;60(3):405). Methods: Pre-salvage therapy prognostic variables common to the two datasets (Gleason score at initial treatment, time from original RT, use of ADT at time of original RT, PSA at time of salvage, patient age) were used for propensity matching between patients from previously published ADT (1) and CRYO (2) datasets. Progression free survival (PFS, defined as time from initial treatment to development of castrate resistance or death); Disease Specific Survival (DSS, defined as time from salvage to prostate cancer related death) and Overall Survival (OS, defined as time from salvage to death from any cause) were compared between the propensity matched cohorts using Log-Rank and Cox PH regression statistics. Raw linear propensity scores included in the PH model to account for residual variability. A planned subset analysis examined the effect of neoadjuvant ADT among the CRYO cohort (no CRYO patients had adjuvant ADT). Results: Overall, 1119/1386 (ADT) and 172/187 (CRYO) patients were included in the propensity matched analysis. Median follow up was 6.7 yrs (ADT) and 18.7 yrs (CRYO). Median PFS (95% CI) was 10.7 yrs (9.5, 12.3) for CRYO vs. 7.0 yrs (6.1, 10.0) for ADT (HR 0.63 (0.44, 0.89), p = 0.009). Median OS was also longer for CRYO vs. ADT: 12.3 (11.0, 13.8) vs. 10.2 (9.4, not reached) yrs (HR 0.69; p = 0.02). 10 year DSS event rate was 16.5% CRYO vs. 18.5% ADT but was not statistically different. Neoadjuvant ADT did not affect outcomes in CRYO. Conclusions: A 3-year PFS and 2-year OS benefit was noted for the CRYO vs. ADT cohorts while no difference was noted in DSS. Potential explanations include residual bias not corrected for in the propensity scoring, variable follow-up duration, adverse effects from differing cumulative exposure to ADT or a combination of these factors. Prospective comparisons are required to control for these potential biases and compare other important outcomes such as side effects and quality of life.