Lp(a) (Lipoprotein[a]) Concentrations and Incident Atherosclerotic Cardiovascular Disease: New Insights From a Large National Biobank

Abstract

Objective: Lp(a) (lipoprotein[a]) concentrations are associated with atherosclerotic cardiovascular disease (ASCVD), and new therapies that enable potent and specific reduction are in development. In the largest study conducted to date, we address 3 areas of uncertainty: (1) the magnitude and shape of ASCVD risk conferred across the distribution of lipoprotein(a) concentrations; (2) variation of risk across racial and clinical subgroups; (3) clinical importance of a high lipoprotein(a) threshold to guide therapy. Approach and Results: Relationship of lipoprotein(a) to incident ASCVD was studied in 460 506 middle-aged UK Biobank participants. Over a median follow-up of 11.2 years, incident ASCVD occurred in 22 401 (4.9%) participants. Median lipoprotein(a) concentration was 19.6 nmol/L (25th-75th percentile 7.6-74.8). The relationship between lipoprotein(a) and ASCVD appeared linear across the distribution, with a hazard ratio of 1.11 (95% CI, 1.10-1.12) per 50 nmol/L increment. Substantial differences in concentrations were noted according to race-median values for white, South Asian, black, and Chinese individuals were 19, 31, 75, and 16 nmol/L, respectively. However, risk per 50 nmol/L appeared similar-hazard ratios of 1.11, 1.10, and 1.07 for white, South Asian, and black individuals, respectively. A high lipoprotein(a) concentration defined as ≥150 nmol/L was present in 12.2% of those without and 20.3% of those with preexisting ASCVD and associated with hazard ratios of 1.50 (95% CI, 1.44-1.56) and 1.16 (95% CI, 1.05-1.27), respectively.

Conclusions: Lipoprotein(a) concentrations predict incident ASCVD among middle-aged adults within primary and secondary prevention contexts, with a linear risk gradient across the distribution. Concentrations are variable across racial subgroups, but the associated risk appears similar.

Keywords: atherosclerosis; cardiovascular disease; lipoprotein(a); myocardial infarction; secondary prevention.

FIGURE 1:

Lipoprotein(a) concentrations according to race, sex, and atherosclerotic cardiovascular disease at time of enrollment Boxplots showing log-transformed lipoprotein(a) distribution by (A) race (Kruskal-Wallis test p<2.2 × 10⁻¹⁶), (B) sex (Mann-Whitney U-test p<2.2 × 10⁻¹⁶), and (C) presence or absence of atherosclerotic cardiovascular disease (ASCVD) at time of enrollment (Mann-Whitney U-test p<2.2 × 10⁻¹⁶). Dimensions of the box capture the 25th to 75th percentiles, and whiskers capture an additional one interquartile range.

FIGURE 2:

Incidence and risk of atherosclerotic cardiovascular disease according to lipoprotein(a) concentration A: Incidence rates per 1000 person-years with corresponding 95% confidence intervals of atherosclerotic cardiovascular disease events grouped by percentile of the lipoprotein(a) distribution. B: Smoothed adjusted hazard ratio and 95% confidence intervals of individuals with a given lipoprotein(a) concentration with respect to the risk in an individual with the median lipoprotein(a) concentration in the population (19.6 nmol/L), estimated using a Cox proportional hazards regression model with covariates of enrollment age, sex, self-reported race, and lipoprotein(a) concentration modeled using cubic natural splines.

FIGURE 3:

Risk associated with increased lipoprotein(a) according to racial and clinical subgroups Hazard ratios with corresponding 95% confidence intervals and p values for risk of atherosclerotic cardiovascular disease (ASCVD) events per 50 nmol/L increase in lipoprotein(a) are reported based on a Cox proportional hazards regression models with covariates of enrollment age, sex, and self-reported race. Subgroup analyses were performed with cohorts splitting based on presence or absence of variable of interest. Additionally, an interaction p value was computed to examine the significance of interaction between variable of interest and lipoprotein(a) concentration in influencing ASCVD risk. Median estimated unadjusted low-density lipoprotein (LDL) cholesterol concentration: 143.2 mg/dL.

FIGURE 4:

Prevalence and clinical importance of high lipoprotein(a) concentrations Histogram of counts of individuals per given lipoprotein(a) [Lp(a)] concentration bin, with red shaded region designating above cutoff of 150 nmol/L (~70 mg/dL) in individuals without (A) and with (C) prior atherosclerotic cardiovascular disease (ASCVD) at enrollment (distribution truncated at 400 nmol/L). Cumulative incidence of initial (B) or subsequent (D) ASCVD events over length of follow-up stratified by lipoprotein(a) cutoff of 150 nmol/L, estimated using Cox proportional hazards regression model standardized to the average of the enrollment age and sex, for individuals without (B) and with (D) prior ASCVD at enrollment, respectively.

FIGURE 5:

Lipoprotein(a) association analysis by concentration, composite and individual component endpoints Hazard ratios with corresponding 95% confidence intervals and p values for composite atherosclerotic cardiovascular disease (ASCVD) and component endpoints, comparing individuals with lipoprotein(a) concentrations above and below cutoff of 150 nmol/L (~70 mg/dL), calculated using Cox proportional hazards regression models with covariates of enrollment age, sex, and self-reported race. MI: Myocardial infarction.

FIGURE 6:

Lipoprotein(a) association analysis by racial subgroup Hazard ratios with corresponding 95% confidence intervals and p values for first incidence of composite atherosclerotic cardiovascular disease (ASCVD) end point across different racial subgroups, computing risk of disease for every 50 nmol/L increase in lipoprotein(a) [Lp(a)], comparing individuals with lipoprotein(a) concentrations above and below cutoff of 150 nmol/L and race-specific 90th percentiles (white: ≥168.2 nmol/L, South Asian: ≥139.5 nmol/L, and black ≥211.7 nmol/L), calculated using Cox proportional hazards regression models with covariates of enrollment age, and sex. Owing to incident ASCVD events in only 27 of 1,415 individuals of Chinese ancestry, reliable effect estimation was not possible in this subgroup.