Lp(a) and Long-Term CV Risk in a Pooled Cohort

Quick Takes

  • When pooling five multi-ethnic US longitudinal studies with average follow-up of 21 years, there was graded association of Lp(a) with ASCVD events. Compared to those below the 50th percentile (median 3.6 mg/dL), those in the highest decile (median 52.6 mg/dL) were 46% more likely to have future ASCVD events.
  • While diabetics are less likely to have a high Lp(a), risks were strongest in diabetes, but similar in female vs. male and by race/ethnicity.
  • Lp(a) predicted individual ASCVD endpoints of MI, stroke, revascularization, and CHD death but not total mortality.

Study Questions:

What is the long-term implication of lipoprotein(a) [Lp(a)] elevation in a pooled multi-ethnic US cohort?

Methods:

The authors used data on Lp(a) and development of atherosclerotic cardiovascular disease (ASCVD) in five US longitudinal studies with a good representation of age range, sex, and race/ethnicity. Lp(a) levels were classified based on cohort-specific percentiles with Lp(a) categorized by percentile within each cohort (<50th, 50th-<75th, 75th-<90th, and ≥90th percentile) followed by pooling of the cohort-specific groupings. Multivariable Cox regression related Lp(a) with composite incident ASCVD events by risk group and diabetes status. CV risk factors and lipid-lowering medication were assessed at the same time as the Lp(a) levels and used as covariates. Carotid intima-media thickness (IMT) and coronary artery calcium scores were available in some cohorts for secondary analysis to determine whether the relationship of Lp(a) and ASCVD events was mediated by subclinical atherosclerosis.

The 10-year risk of ASCVD was obtained in those aged 40-79 years without known ASCVD using the Pooled Cohort Equations and classified as low/intermediate if <20% and high risk if ≥20%. For those aged 20-<40 years, high risk was ≥2 of the major risk factors with above normal values, current smoking, and diabetes. Primary ASCVD outcomes included nonfatal and fatal myocardial infarction (MI), stroke, revascularization, and coronary heart disease (CHD) death.

Results:

The five studies included 27,756 persons with Lp(a) levels obtained at various exam/years ranging from 1983–2004. Age ranged from 20–79 years (mean 51 ± 12.4 years), 55% female, 35.6% Black, and <5% Hispanic and Asian. Mean follow-up was 21.1 ± 9.9 years for events. Overall, 82.4% (n = 22,871) were at low/intermediate risk, 17.6% (n = 4,885) at high risk, and 7.6% (n = 2,100) had diabetes. Median Lp(a) levels were 3.6, 13.5, 25.9, and 52.6 mg/dL for the <50th, 50-<75th, 75th-<90th, and >90th percentile groupings, respectively. Compared to those <50th percentile, Lp(a) levels in the 50-<75th, 75-<90th, and >90th percentiles had adjusted hazard ratios (95% confidence interval [CI]) of 1.06 (0.99-1.14), 1.18 (1.09-1.28), and 1.46 (1.33-1.59), respectively for ASCVD events. Elevated Lp(a) predicted similarly by risk group, sex, and race/ethnic groups, but more strongly in those with versus without diabetes (interaction p = 0.0056) with HRs (95% CI) for Lp(a) ≥90th percentile of 1.92 (1.50-2.45) and 1.41 (1.28-1.55), respectively.

Lp(a) also individually predicted MI, revascularization, stroke, CHD death, but not total mortality. Higher Lp(a) levels were consistently associated with ASCVD event risk across low-density lipoprotein cholesterol (LDL-C) categories, especially in those with LDL-C >70 mg/dL. There were similar relationships of Lp(a) with ASCVD events according to the presence of obesity, smoking, and hypertension, and hazard ratios remained essentially unchanged based upon carotid IMT and coronary artery calcium, indicating that Lp(a) relationships with ASCVD events do not appear to be mediated by subclinical atherosclerosis.

Conclusions:

In a large US pooled cohort with over 20 years of follow-up, higher Lp(a) levels are associated with increased ASCVD risk, in particular those with diabetes, but similar among low/intermediate versus high-risk primary prevention, female versus male, across race/ethnicity, as well as across the range of LDL-C levels.

Perspective:

This longitudinal study of risk associated with high levels of Lp(a) supports the recommendation that this highly atherogenic lipid moiety be measured at least once in adults, particularly in persons with premature ASCVD, and if above 50 mg/dL, be used as a risk enhancer in primary prevention (particularly in diabetes) and an indication for intensification of lipid-lowering therapy in secondary prevention. Considering that Lp(a) is an independent significant risk factor for ASCVD, hopefully data from this pooled study of persons without ASCVD, the primary prevention statin trials, and biobank studies may be used to add Lp(a) levels to risk predicting algorithms.

The mean Lp(a) levels in the study performed over 20 years ago were lower than found in more recent population studies likely related to older methods, which may have resulted in underestimation of hazard ratios.

Clinical Topics: Dyslipidemia, Advanced Lipid Testing, Lipid Metabolism, Prevention

Keywords: Atherosclerosis, Lipoprotein(a)


< Back to Listings