Ischemic Heart Disease Risk and Remnant Cholesterol Levels

Peter W.F. Wilson, MD; Alan T. Remaley, MD, PHD


J Am Coll Cardiol. 2022;79(24):2398-2400. 

In this issue of the Journal of the American College of Cardiology, Doi et al[1] report on the predictive capability of elevated remnant cholesterol (REM-C) and risk for myocardial infarction and ischemic heart disease. The report is based on >40,000 adult White Copenhagen City Population participants without diabetes or ischemic heart disease and not taking statins at baseline. The participants were followed for 10 years for the occurrence of ischemic heart disease, myocardial infarction, and coronary revascularization. The major finding is that elevated REM-C was an important predictor of ischemic coronary disease outcomes when Bayesian reclassification analysis was undertaken after multivariable prediction.

REM-C is typically calculated in nonfasting blood specimens by measuring total cholesterol (TC) minus low-density lipoprotein cholesterol (LDL-C) minus high-density lipoprotein cholesterol (HDL-C). Equations 1–5 illustrate the calculation of REM-C from nonfasting lipid specimens from TC, LDL-C, HDL-C, and triglyceride (Trig):

As shown in equations 1–5, when the Friedewald formula is used with nonfasting lipoprotein cholesterol data, the REM-C estimate is simply the Friedewald constant multiplied by the Trig concentration (0.20 × Trig [mg/dL] or 0.45 × Trig [mmol/L]). Similar REM-C calculations apply when the Martin or Sampson LDL-C estimating equation is used to estimate LDL-C.[2,3] Implicit in this REM-C calculation on the basis of nonfasting lipid information, the LDL-C estimate is likely to be less accurate than classically estimated LDL-C from the Friedewald, Martin, or Sampson method on the basis of fasting lipid levels.

The investigators' primary model was atherosclerotic cardiovascular disease (ASCVD) risk in adults without diabetes and not on statins predicted by age, sex, smoking, systolic blood pressure, nonfasting LDL-C, and nonfasting REM-C. Notably, HDL-C was not included in the core prediction model, as the analyses emphasized atherogenic lipids. Additionally, C statistics were not provided to gauge overall prediction performance. The development of diabetes and statin initiation after the baseline also were not assessed as risk modulators in the 10-year follow-up analyses. The investigators evaluated a variety of multivariable prediction models, including additional adjustments for plasma Trig, apolipoprotein B, and non-HDL cholesterol levels, and these analyses did not materially affect their overall results.

The investigators report that blood specimens with very elevated REM-C estimated from nonfasting lipid measurements of cholesterol, HDL-C, and Trig provide added value to predict ischemic cardiovascular events over and above nonfasting LDL-C estimations in a multivariable risk model that did not include HDL-C information. Overall, elevated REM-C estimates helped to reclassify individuals beyond their standard ASCVD risk assessment using an adaptation to U.S. guidelines for absolute risk for initial vascular disease events.[4] Stroke was not included as one of the clinical ASCVD outcomes.

The reported findings rekindle interest in atherogenic nonfasting lipid measurements and emphasize an important role for elevated nonfasting REM-C, estimated from nonfasting Trig, as a value-added predictor of ischemic events. Elements of this story have been reported in the past with multivariable analyses that included fasting LDL-C and HDL-C with inclusion of Trig or log Trig information to aid in the prediction of ASCVD.[5–7] Significant effects were observed for Trig in those analyses, and the statistical significance was typically greater for logarithmically transformed Trig than for Trig analyzed as a linear term.

The Bayesian reclassification analyses in the study by Doi et al[1] show that nonfasting specimens can be used to predict initial ischemic cardiovascular disease events in Danish adults, and elevated REM-C >61 mg/dL (Trignonfasting >305 mg/dL) was associated with approximately 20% reclassification of heart disease ischemic events. Reclassification estimates reported in the past have been approximately 10% for elevated C-reactive protein[8,9] and 25% for coronary artery calcification in MESA (Multi-Ethnic Study of Atherosclerosis).[10] Reclassification analyses for C-reactive protein and coronary artery calcification have typically been used with classical prediction models that included age, sex, blood pressure, blood pressure treatment, smoking, diabetes, and fasting lipids (LDL-C, TC, or HDL-C).

The Danish investigators report relatively simple adaptations to standard lipoprotein testing: nonfasting specimen collection and the use of Trig level information. There is no detailed information concerning the number of hours fasting or antecedent food or beverage intake. For instance, eating a meal typically raises Trig levels 10%-30%, depending on the baseline Trig levels and the fat and carbohydrate content of the meal. Alternatively, study participants may have eaten but consumed few calories, and their nonfasting lipid values were relatively similar to the fasting state, in which case REM-C or elevated Trig levels would not be expected to be markedly elevated. Nevertheless, such variability would tend to bias toward a null result, and the investigators report positive results.

How should researchers and clinicians interpret and potentially use these Danish results? First, both fasting and nonfasting lipid values provide useful information for ASCVD risk estimation,[11] and elevated nonfasting Trig or REM-C appears to help identify persons at greater risk for an initial cardiovascular ischemic event. Second, very elevated levels (above the 75th percentile) of nonfasting REM-C, or nonfasting Trig, deserve further evaluation as a potentially valuable "modifier of ASCVD risk," as discussed in the 2019 U.S. guideline on the management of blood cholesterol.[4] Third, replication of the findings, especially in adults without ischemic heart disease, without diabetes, and not on lipid-altering therapy, could move these findings forward to potentially improve prognostication and care for patients at risk for ischemic heart disease events.