Lipoprotein(a) particle concentration and lipoprotein(a) cholesterol assays yield discordant classification of patients into four physiologically discrete groups
Peter P. Toth
Steven R. Jones
There is little known about the relative predictive value of different lipoprotein(a) [Lp(a)] assays in clinical use, although each has been shown to predict similar incremental risk over conventional clinical and lipid risk factors. Thus, we examined the classification behavior of two commonly used Lp(a) assays and their associations with other lipid parameters. Serum lipid and Lp(a) concentrations were measured in 144 primary and secondary prevention patients. Lp(a) cholesterol [Lp(a)-C] was measured with the Vertical Auto Profile (upper limit of normal, 10 mg/dL). Lp(a) particle concentrations [Lp(a)-P] were measured with an isoform-independent molar assay (upper limit of normal, 70 nmol/L). The subjects were divided into the following four groups on the basis of their Lp(a)-C and Lp(a)-P levels: normal Lp(a)-P and Lp(a)-C; high Lp(a)-P and normal Lp(a)-C; normal Lp(a)-P and high Lp(a)-C; and high Lp(a)-P and Lp(a)-C. The proportion of subjects with values above the upper limit of normal was similar with both assays (P = .15). However, the Lp(a)-C and Lp(a)-P assays discordantly classified 23% of the study's subjects. In addition, the four Lp(a)-defined groups displayed differences in their relationships with other lipoproteins. The two groups with elevated Lp(a)-C showed significant associations with higher high-density lipoprotein cholesterol, apolipoprotein AI, and high-density lipoprotein cholesterol/apolipoprotein AI ratios. Triglycerides were also noted to be above normal in discordant and normal within concordant Lp(a) groups. Finally, the amount of cholesterol per Lp(a) particle [Lp(a)-C/Lp(a)-P] varied widely across the four groups. These findings suggest that the four Lp(a)-defined groups are physiologically discrete. Further investigation is warranted to assess which parameters among Lp(a)-P, Lp(a)-C, and Lp(a)-C/Lp(a)-P can be used to more accurately characterize Lp(a)-associated cardiovascular risk.