Various blood constituents can interfere with immunoassays, usually by binding the Fc portion of antibodies. Our previously developed assays for intact free prostate-specific antigen (PSA), free human kallikrein 2 (hK2), and total hK2 frequently yielded falsely high results despite including an excess of scavenger antibodies. We investigated whether this interference could be eliminated by replacing monoclonal capture or tracer antibodies with F(ab')2 or recombinant Fab fragments. Female heparin plasma samples (n = 1092), which should have negligible PSA and hK2, and male samples (n = 957) were analyzed to identify samples manifesting interference, which then were used to optimize protocols for the immunoassays. We compared original assays (monoclonal antibodies) versus optimized assays (F(ab')2 fragments: denatured mouse IgG added as scavenger) using another set of EDTA plasma (n = 113), heparin plasma (n = 160), and serum samples (n = 171). With the original assays, the frequency of falsely elevated hK2 and intact free PSA was 15 and 13%, respectively. The optimized assays eliminated 70-85% of these falsely elevated results and importantly reduced the magnitude in the remainder. F(ab')2 fragmentation was the most important factor in reducing interference. The optimized intact free PSA, free hK2, and total hK2 assays manifested high accuracy close to the lower limit of detection.