This study compares the diagnostic utility of complexed prostate-specific antigen (cPSA), total PSA (tPSA) and their ratios with free PSA (fPSA) for benign prostatic hyperplasia (BPH) and prostate cancer. This is though to be the first study to evaluate cPSA in the ethnic population of Saudi Arabia. Serum samples were collected from 54 patients (aged over 50) and assayed for tPSA, cPSA and fPSA. Thirty-five patients were histologically and clinically proven to have BPH and 19 patients were proven to have cancer. Sensitivity, specificity and ROC curves were calculated. With a cPSA cut-off of 4 ng/mL the sensitivity was 79%, the specificity was 34%, and the positive and negative predictive values (PPV and NPV) were 39% and 75%, respectively. At the same cut-off for tPSA, the sensitivity was 84%, the specificity was 29%, and the PPV and NPV were 39 and 77%, respectively. The sensitivity for both tests was lower at a cut-off of 20 ng/mL but the specificity increased to 77% for cPSA and 69% for tPSA. The areas under the receiver operating characteristic (ROC) curves were found to be 0.608 for tPSA and 0.559 for cPSA (P = 0.69). The incidence of prostate cancer in the Saudi population may be lower than that in Western populations. The data presented show little advantage in using cPSA over tPSA for discriminating BPH and prostate cancer in the population studied.