Abstract With a two-site CK-MB assay, we screened serum samples from 1008 blood donors for the presence of antibodies to mouse monoclonal immunoglobulin. These antibodies were capable of cross-linking the labeled antibody with the solid-phase antibody in the two-site assay, thus generating a falsely high apparent CK-MB concentration. In 92 (9.12%) of the blood donors tested, apparent CK-MB concentrations of 10-1000 micrograms/L decreased to less than 3 micrograms/L when re-assayed with non-immune mouse serum (10 mL/L) included in the assay reagent. We tested the ability of non-immune sera from other animal species to lower the concentration of apparent CK-MB in 58 of the 92 samples. Bovine and ovine serum were almost as effective as mouse serum; feline, canine, and rabbit serum were less effective. Of the samples tested, 12% (1.1% of the original population screened) showed apparent CK-MB values that either were not depressed by bovine serum or were only partly depressed. We discuss the possible etiology of these antibodies in normal subjects and recommend that all mouse monoclonal two-site assays should contain non-immune mouse serum (or a suitable "irrelevant" mouse monoclonal antibody) to prevent false-positive results.