HDL Cholesterol Not As Useful for Predicting MI in Minority Patients

June 23, 2020

While often referred to as "Good cholesterol" new research suggests HDL measurements do little for predicting risk of heart attack in Black patients

Anand Rohatgi, MD

Despite being considered a powerful predictor of cardiovascular disease risk in many patients, new research from the University of Texas South Western (UTSW) suggests high-density lipoprotein (HDL) cholesterol levels may not be effective for predicting cardiovascular events in minority patients. 





With previous research indicates patients with higher levels of HDL cholesterol tend to have lower rates of cardiovascular disease, this new study examined the predictive value of HDL cholesterol concentration (HDL-C) and HDL particle (HDL-P) levels and found HDL-P may be a better predictor, but neither were significantly associated with myocardial infarction among Black patients.

“If you’re white, low HDL cholesterol is still a powerful predictor of heart attack and stroke risk, and that has not changed,” said lead investigator Anand Rohatgi, MD, an associate professor of internal medicine at UTSW, in a statement. “But if you’re not white, it’s not that straightforward.”

To further investigate how HDL levels might help predict myocardial infarction in underrepresented populations, including women and minority patients, Rohtagi and a team of colleagues designed the current study using data from the Dallas Heart Study (DHS), the Multi-Ethnic Study of Atherosclerosis (MESA), the Atherosclerosis Risk in Communities Study (ARIC), and the Prevention of Renal and Vascular End-stage Disease (PREVEND).

The ultimate goal of their analyses was to investigate potential associations between the markers HDL-P, HDL-C, HDL-C combined HDL-P, and HDL size combined with HDL-P and the outcomes of myocardial infarction and stroke as well as overall atherosclerotic cardiovascular disease (ASCVD) with adjustment for gender and race/ethnic background.

“Previous studies have looked at HDL levels in the population as a whole,” Rohatgi said. “But we know that sometimes biology differs by gender and race, so we thought it was important to separately tease apart what’s happening in those populations, as well as how HDL is associated with stroke, which has been understudied.”

From these 4 studies, investigators identified a total of 15,784 patients for inclusion in their study. The median age of this cohort was 56 years, 46% were and 22% were Black. Specifically, the study population included 8550 women and 3520 Black participants. Additionally, the median HDL-C was 48 mg/dL, median HDL-P was 32.5 µmol/L, and median HDL-size was 9.1 nm.

During the follow-up period, which lasted a mean of 8-12 years, 515 fatal/non-fatal MI events, 321 fatal/non-fatal ischemic stroke events, and 1,242 overall ASCVD events occurred in the study cohort.

Upon analysis, investigators found HDL-P (HR for Q4 vs Q1, 0.64; 95% CI, .52-.78) and HDL-C (HR for Q4 vs Q1, 0.76; 95% CI, 0.61-0.94) levels were inversely associated with the combined outcome of myocardial infarction and stroke in models adjusted for cardiometabolic risk factors. Adjustment for HDL-C did not appear to reduce the inverse relationship between HDL-P and ASCVD while adjustment for HDL-P mitigated all associations between HDL-C and events of interest.

In regard to race/ethnic background, HDL-P was inversely associated with myocardial infarction among white patients but not among their Black counterparts (HR Q4 vs Q1 for White, 0.49; 95%CI, .35-.69; for Black, 1.22; 95% CI, 0.76-1.98; P for interaction=.001). Additionally, HDL-C appeared to be inversely associated with myocardial infarction among white patients (HR Q4 vs Q1, 0.53; 95% CI, .36-.78) but investigators noted a weak direct association with incidence among Black patients (HR Q4 vs Q1, 1.75; 95% CI, 1.08-2.83; P for interaction <.0001).

This study, “Associations Between HDL Particles and Ischemic Events by Vascular Domain, Gender, and Ethnicity: A Pooled Cohort Analysis,” was published in Circulation.


x