The aim of this study was to compare the diagnostic utility of a new assay that measures all forms of prostate-specific antigen complexed (cPSA) to serum proteins except alpha(2)-macroglobulin with the assay of free PSA (fPSA) and the corresponding ratios to total PSA (tPSA) to improve the differentiation between benign prostatic hyperplasia (BPH) and prostate cancer (PCa). Serum samples were collected from 91 men without prostate disease and with normal digital rectal examination (controls), 144 untreated patients with PCa, and 89 patients with BPH. tPSA and cPSA were measured using the Bayer Immuno 1 system; fPSA and the additional tPSA were measured with the Roche Elecsys system. The median cPSA/tPSA, fPSA/tPSA, and fPSA/cPSA ratios were significantly different between patients with BPH and PCa (78.7% vs 90.7%, 25.5% vs 12.1%, and 36.8% vs 14.3%, respectively; P <0.001). No correlations of cPSA and its ratios to tumor stage and grade were found. ROC analysis showed that cPSA was not different from tPSA (areas under the curve, 0.632 vs 0.568), whereas the cPSA/tPSA ratio was similar to the fPSA/tPSA ratio in increasing discrimination between BPH and PCa patients with tPSA concentrations in the tPSA gray zone between 2 and 10 microg/L (areas under the curve, 0.851 vs 0.838). Compared with tPSA, the fPSA/tPSA and cPSA/tPSA ratios both improve the differentiation between BPH and PCa comparably and are similarly effective in reducing the rate of unnecessary biopsies, whereas cPSA alone does not have any effect.