Logo image
Back
Is Estimation of Lp(a)-C Accurate Enough for Clinical Use?
Abstract

Is Estimation of Lp(a)-C Accurate Enough for Clinical Use?

Vasanth Sathiyakumar, Jihwan Park, Eliseo Guallar, Mariana Lazo, Raul Santos, Peter Toth, Steven Jones and Seth Martin
Circulation (New York, N.Y.), v 138(Suppl 1), 12647
06 Nov 2018

Abstract

Introduction Estimating the cholesterol concentration of lipoprotein(a) (Lp(a)-C) may be of clinical value, since calculated low-density lipoprotein-cholesterol (LDL-C) also contains Lp(a)-C. Recently, it has been proposed that Lp(a)-C may be estimated by simply dividing the particle concentration of Lp(a) in nmol/L by 2.4 to convert to Lp(a) mass in mg/dL, and then multiplying this value by 0.3 to reflect its cholesterol content. Hypothesis Given variability in Lp(a) particle composition, we hypothesized that estimated Lp(a)-C may frequently show large deviations from measured Lp(a)-C by ultracentrifugation. Methods All patients with complete lipid values including Lp(a)-C from the Very Large Database of Lipids study were identified. Lp(a)-C was measured through Vertical Auto Profile (VAP) - a rapid ultracentrifugation technique. Accuracy was expressed as percentage of VAP-measured Lp(a)-C falling in the appropriate category of estimated Lp(a)-C (<10, 10-19, 20-29, 30-39, 40-49, and ≥50 mg/dL). We further evaluated magnitude of error in estimated Lp(a)-C by assessing the percentage of patients whose estimated Lp(a)-C was <5, 5-9, 10-19, 20-29, and ≥30% of VAP-measured Lp(a)-C. Secondary analyses were stratified by high-density lipoprotein-cholesterol (HDL-C) values, given prior concern that HDL-C may interfere with precision in VAP-derived Lp(a)-C. Results A total of 177,885 patients with Lp(a)-C values were identified. The median age was 57 years and 42% were men. Accuracy in estimated Lp(a)-C decreased as estimated Lp(a)-C increased. In the estimated Lp(a)-C <10 mg/dL group, accuracy was 89%, but accuracy was only 1% in those with estimated Lp(a)-C ≥50 mg/dL. Error also increased as estimated Lp(a)-C increased although 72% of patients with estimated Lp(a)-C <10 mg/dL had errors >30% compared to measured Lp(a)-C, 99% of those with estimated Lp(a) ≥50 mg/dL had similar error. These findings were robust to stratification by HDL-C. Conclusions In a large, US cross-sectional sample, a proposed equation to estimate Lp(a)-C was largely inaccurate especially where accuracy matters most - those with high Lp(a)-C levels. These findings may have important clinical implications because they raise concern that inappropriate decisions may be made were this equation used in practice.

Metrics

9 Record Views

Details

Logo image