Lipoprotein(a) levels and association with myocardial infarction and stroke in a nationally representative cross-sectional US cohort

Abstract

Lipoprotein(a) (Lp(a)) has not been well-studied in a nationally representative US cohort. The objective of this study was to investigate the distribution of Lp(a) and its associations with nonfatal cardiovascular events in a nationally representative cohort. Cross-sectional analysis using the National Health and Nutrition Examination Survey III cohort (1991-1994). We compared Lp(a) levels across demographics and tested the associations between Lp(a) and patient-reported nonfatal myocardial infarction (MI) and/or stroke using multivariate logistic regression. Median Lp(a) was 14mg/dL (interquartile range [IQR]: 3, 32) (n=8214). 14.7% (95% CI: 13.6%-15.9%) had Lp(a) ≥50mg/dL. Women had slightly higher median Lp(a) than men (14mg/dL [IQR: 4, 33] vs 13 [(IQR: 3, 30], P=.001). Non-Hispanics blacks had the highest median Lp(a) (35mg/dL [IQR: 21, 64]), followed by non-Hispanic whites (12mg/dL [IQR: 3, 29]) and Mexican Americans (8mg/dL [IQR:1, 21]). In multivariate analysis, Lp(a) was associated (odds ratio per SD increase [95% CI], P-value) with MI (1.41 [1.14-1.75], P=.001), but not stroke (1.14 [0.91-1.44], P=.26). Lp(a) associated with MI in men (1.52 [1.13-2.04], P=.006), non-Hispanic whites (1.60 [1.27-2.03], P<.001), and Mexican Americans (2.14 [1.29-3.55], P=.003), but not women or non-Hispanic blacks. Lp(a) was not associated with stroke among any subgroups. In a nationally representative US cohort, 1 in 7 had Lp(a) ≥50mg/dL, the guidelines-recommended threshold to consider Lp(a) a risk enhancing factor. Lp(a) was associated with nonfatal MI but not stroke, although there were differential associations by sex and race/ethnicity. Future nationally representative cohorts should test Lp(a) to get an updated estimation.