The purpose of this study was to assess how the reporting of biochemical failure (BF) rates would be affected by the application of three different definitions. Three hundred and fifteen men with localized prostate cancer underwent I-125 brachytherapy (n=109), conformal three-dimensional radiation therapy (n=99), or radical prostatectomy (n=107). No patient received adjuvant or neoadjuvant hormone therapy in this study. BF rates at 12, 24 and 36 months were assessed using three definitions: (1) prostate-specific antigen (PSA) nadir >0.5 ng/ml; (2) PSA rise by 0.5 ng/ml; and (3) three consecutive PSA rises. Median follow-up for the brachytherapy group, external beam radiotherapy group, and the radical prostatectomy group was 27, 30 and 36 months respectively. The applied definition influenced reporting of failure rates in two of the three groups. I-125 brachytherapy group: BF rates at 24 months: 46%-definition 1, 35%-definition 2, and 4%-definition 3 (P<0.05). Radiation therapy group: BF rates at 24 months: 39%-definition 1, 17%-definition 2 and 3%-definition 3 (P<0.05). No patient in the radical prostatectomy group had a BF by any applied definition. A more universal definition of BF is needed to compare the efficacy of treatments for localized prostate cancer.