UC San Diego

Background: Lipoprotein(a) is a risk factor for cardiovascular events and modifies the benefit of pcsk9 inhibitors (pcsk9i). Lipoprotein(a) concentration can be measured with immunoassays reporting mass or molar concentration or a reference measurement system employing mass spectrometry. Whether the relationships between lipoprotein(a) concentrations and cardiovascular events in a high-risk cohort differ across lipoprotein(a) methods is unknown. We compared the prognostic and predictive value of these types of lipoprotein(a) tests for major adverse cardiovascular events (mace). Methods: The odyssey outcomes trial compared the pcsk9i alirocumab with placebo in patients with recent acute coronary syndrome (acs). We compared risk of mace in the placebo group and mace risk reduction with alirocumab according to baseline lipoprotein(a) concentration measured by siemens n-latex nephelometric immunoassay (ia-mass, mg/dl), roche tina-quant® turbidimetric immunoassay (ia-molar, nmol/l), and a non-commercial mass spectrometry-based test (ms, nmol/l). Lipoprotein(a) values were transformed into percentiles for comparative modeling. Natural cubic splines estimated continuous relationships between baseline lipoprotein(a) and outcomes in each treatment group. Event rates were also determined across baseline lipoprotein(a) quartiles defined by each assay. Results: Among 11,970 trial participants with results from all 3 tests, baseline median (q1, q3) lipoprotein(a) concentrations were 21.8 (6.9, 60.0) mg/dl, 45.0 (13.2, 153.8) nmol/l, and 42.2 (14.3, 143.1) nmol/l for ia-mass, ia-molar, and ms, respectively. The strongest correlation was between ia-molar and ms (r=0.990), with nominally weaker correlations between ia-mass and ms (r=0.967) and ia-mass and ia-molar (r=0.972). Relationships of lipoprotein(a) with mace risk in the placebo group were nearly identical with each test with estimated cumulative incidences differing by ≤0.4% across lipoprotein(a) percentiles, and all were incrementally prognostic after accounting for ldl-c (all spline p≤0.0003). Predicted alirocumab treatment effects were also nearly identical for each of the three tests, with estimated treatment hazard ratios (hrs) differing by ≤0.07 between tests across percentiles and nominally less relative risk reduction by alirocumab at lower percentiles for all three tests. Absolute risk reduction with alirocumab increased with increasing lipoprotein(a) measured by each test, with significant linear trends across quartiles. Conclusions: In patients with recent acs, three lipoprotein(a) tests were similarly prognostic for mace in the placebo group and predictive of mace reductions with alirocumab at the cohort level.