Sandwich ELISAs with an anti-apo(a) trapping antibody and either an anti-apolipoprotein B or anti-apolipoprotein(a) reporting antibody, were used to measure the concentrations of Lp(a) in 230 plasma samples that were either freshly drawn or stored at -20 degrees C for 4-6 weeks. The assays produced significantly different results for the fresh and frozen samples, however, the magnitudes of these differences were small, about 8% higher for the frozen samples, and independent of total cholesterol, HDL cholesterol, triglyceride, apolipoprotein B or Lp(a) concentration or assay configuration. A similar difference was seen for a freshly drawn plasma sample assayed at the time as the fresh and frozen samples, indicating the differences were due to inherent differences in the assays at the times the assays were performed. The assay configuration was an important factor in determining the Lp(a) concentrations for identically treated samples. ELISAs using the apoB reporting antibody yielded concentrations that were significantly less than those determined by ELISAs using the anti-apo(a) reporting antibody. The assay differences did not correlate with total cholesterol, HDL cholesterol, triglyceride, or apoB concentration. However, the magnitude of the difference did correlate well with Lp(a) amount. Low Lp(a) concentrations produced greater assay differences than high Lp(a) concentrations.