The percentage of free PSA in serum is currently used to better discriminate between patients with prostate cancer and patients with benign prostatic hyperplasia, in prostate cancer screening programs. We measured using non-competitive immunological techniques, the total PSA and free PSA in post-surgical serum of prostate cancer patients who underwent radical prostatectomy and then relapsed. We compared these data with those of a group of 40 age-matched men with no evidence of prostatic disease. Although in general, patients with prostate cancer had lower percentage of free PSA in serum in comparison to the controls, a subset of these patients (approximately 20%) had percent free PSA significantly higher than the levels considered as exclusive of prostate cancer in screening programs. We also found that percent free PSA does not correlate significantly with most of the standard clinical or pathological indicators of prostate cancer aggressiveness. Only a weak negative association with Gleason Score was observed. The percent free PSA in serum of relapsing prostate cancer patients varies within a relatively wide range and does not correlate significantly with indicators of cancer aggressiveness. The use of percent free PSA for excluding prostate cancer in screening programs must be approached with caution until the mechanism of low percent free PSA in the majority but not all prostate cancer patients is elucidated.