Interpreting the non-Lp(a) apoB to Lp(a) ratio in coronary plaque phenotypes

To the Editor, We read with great interest the study by Tsimikas et al. examining non-Lp(a) apolipoprotein B and Lp(a) across PROSPECT II and CASABLANCA.1 Separating apoB-containing particles into Lp(a) and non-Lp(a) pools is conceptually appealing, and the integration of intracoronary imaging with clinical outcomes is a strength. Several aspects of the ratio may nevertheless merit further clarification before it is interpreted as a measure of the relative biological contribution of the two particle classes. The ratio appears to track Lp(a) more strongly than a balance between its two components. In PROSPECT II, the reported ratio tertiles showed no clear gradient in non-Lp(a) apoB, whereas Lp(a) differed substantially across tertiles. More importantly, the earlier PROSPECT II analysis found that Lp(a) itself was associated with plaque burden ≥70%, maxLCBI4mm ≥324.7, and their combination.2 In the present analysis, the ratio was associated with plaque burden ≥70% and the composite endpoint, but not with maxLCBI4mm alone.1 This pattern raises the possibility that constructing the ratio attenuates part of the information carried by Lp(a) rather than adding information from non-Lp(a) apoB. Comparing models containing Lp(a) alone, non-Lp(a) apoB alone, and both components jointly with the ratio model would clarify whether the ratio provides incremental information beyond its denominator.