To compare the distribution of total and complexed prostate-specific antigen (cPSA) in men with and without prostate cancer with another studied population and to ascertain whether cPSA could enhance the detection of prostate cancer in Japanese men.

A total of 760 men whose serum total PSA (tPSA) values ranged from 1.0 to 100 ng/mL were enrolled. Serum samples for tPSA and cPSA (ADVIA Centaur) were obtained in all cases. The area under the curve was calculated for comparison of the tPSA and cPSA values. We calculated the number of cancers missed and false-positive results at various cutoff values of cPSA compared with the conventional tPSA threshold of 4.0 ng/mL.

Prostate cancer was detected in 268 (35.3%) of 760 patients. cPSA was greater than 8.3 ng/mL (equivalent to 10.0 ng/mL tPSA) in 46.6% of the men with cancer. The area under the curve for cPSA (0.741) was significantly better than that for tPSA (0.721, P <0.001). At a sensitivity of 85% to 95%, significant differences were found in the corresponding specificity between tPSA and cPSA. cPSA at a 3.0-ng/mL threshold detected an identical number of cancers as a tPSA cutoff of 4.0 ng/mL; however, it decreased the false-positive results by 28 cases.

To our knowledge, this is the first report of the distribution of cPSA in Japanese men using a urologic referral population. cPSA can be an alternative to tPSA as the first screening test. A substantial number of men in Japan with prostate cancer are currently diagnosed with a tPSA value greater than 10.0 ng/mL.