The Significance of Measuring Non-HDL-Cholesterol

Glenn A. Hirsch, MD; Nidhi Vaid, MBBS; Roger S. Blumenthal, MD

In This Article

Abstract and Introduction

The third Adult Treatment Panel of the National Cholesterol Education Program has recently issued revised guidelines for the treatment of cholesterol in adults. Increased attention to the metabolic syndrome and diabetes, including the inaccuracy of the low-density lipoprotein cholesterol (LDL-C) measurement in these patients because of elevated triglycerides is highlighted. To overcome the inaccuracy of the Friedewald equation in calculating LDL-C when the triglycerides are elevated, measuring non-high-density lipoprotein (non-HDL-C) may provide a better means to follow these patients toward their treatment goals. Recently, non-HDL-C was shown to be a better predictor of cardiovascular death than LDL-C, even in patients with triglyceride levels below 200 mg/dL. The authors review the basis for using non-HDL-C as a treatment target for cholesterol, in comparison with other lipoproteins.

Coronary heart disease (CHD) is the most common cause of death among men and women in the United States, accounting for approximately 500,000 deaths per year.[1] Atherosclerosis is a complex process involving multiple interactions among immune, coagulation, hormonal, and vascular systems. Dyslipidemia is a major risk factor for CHD; thus, its diagnosis and management are key factors in the prevention of atherosclerotic plaque and development of cardiovascular disease (CVD) events. Decreases in cholesterol levels lead to diminished macrophage activity and improvement in endothelial function, and thus lead to increased levels of the antiatherosclerotic molecule, nitric oxide.[2,3] Improvements in endothelial function are associated with plaque stabilization and prevention of acute CHD events. Obtaining a standard lipid profile in a clinical setting involves measurement of the total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and triglyceride (TG) levels, while the low-density lipoprotein cholesterol (LDL-C) level is a calculated value.

Lipoproteins are particles containing cholesterol esters and TGs in their core, with their surface layers composed of apolipoproteins, phospholipids, and free cholesterol. The classification of apolipoproteins is based on density. The most dense, chylomicrons, contain the most TGs, followed by very low-density lipoproteins (VLDLs), LDLs that are cholesterol-rich, and HDLs, whose composition is almost one half apolipoproteins.[4]

The role of LDL-C in the pathogenesis of CHD has been established.[5,6,7,8,9] Nevertheless, there are some patients with CHD with plasma LDL-C levels within the normal range. The variation in size, density, and composition of the LDL-C particle governs its properties. The use of gradient gel electrophoresis has demonstrated the existence of two distinct LDL-C phenotypes.[10] The larger, less dense particles are known as pattern A and the smaller, denser particles are known as pattern B. These small, dense LDL-C particles are more prevalent in patients with the atherogenic metabolic syndrome (low HDL-C and high TG levels) and those with CHD.[11,12,13,14]

Among the different risk factors for CHD, increased LDL-C levels are a major contributory factor in atherogenic processes. Drug therapy aimed at reducing LDL-C levels significantly reduces the risk of coronary events. The 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, statins, are the mainstay of treatment to achieve target LDL-C levels[15]; in addition, they lower TG levels.

The expert panel on the Detection, Evaluation, and Treatment of Increased Blood Cholesterol -- the Adult Treatment Panel (ATP) III of the National Cholesterol Education Program (NCEP) -- has consistently targeted high LDL-C levels as a method of primary prevention of CHD. The latest report, published in May, 2001,[16] emphasized the need for more intensive LDL-C-lowering therapy in both the primary and secondary prevention of CHD. It also emphasized the need for primary prevention in patients with multiple risk factors.

The indications for therapy are based on the patient's risk status, including the measurement of LDL-C. Adults who are ≥20 years of age are recommended to obtain a fasting lipid profile once every 5 years. Lipid-lowering therapy consists of both therapeutic lifestyle changes and drug therapy. The type of lipid-lowering therapy initiated is dependent on the calculation of the Framingham 10-year risk of a coronary event. The Framingham risk score takes into account age, TC, cigarette smoking, HDL, and systolic blood pressure. In patients in the highest risk category -- that is, those with a 10-year risk of a coronary event >20%, the optimal LDL goal would be a value <100 mg/dL, with therapeutic lifestyle changes being initiated at LDL levels >100 mg/dL and drug therapy at >130 mg/dL.[16]

The NCEP emphasizes the need for optimization of LDL levels, but it has recently been suggested that non-HDL-C may be a better predictor of cardiovascular risk.[17,18] Non-HDL-C encompasses all cholesterol present in potentially atherogenic lipoprotein particles (VLDL, intermediate-density lipoprotein [IDL], LDL, and lipoprotein[a]). At present, the concentration of LDL-C is estimated using the Friedewald equation: LDL-C=TC-HDL-C-TG/5 mg/dL.[17] This equation requires measurement of TC, TG, and HDL-C. Although the LDL-C concentration estimated by this method provides a reasonable estimate of the amount of LDL-C, it also encompasses IDL-C and lipoprotein(a). As plasma TG concentration increases, as in patients with diabetes mellitus, the VLDL composition is altered and the estimation of LDL-C concentration becomes progressively less accurate. The Friedewald equation is generally considered to be less accurate with increasing TG levels and inapplicable at TG concentrations >400 mg/dL.[19] The advantages of using non-HDL-C as a screening tool include the fact that it requires measurement of only TC and HDL-C, both of which can be measured reasonably accurately in a nonfasting sample, as opposed to LDL-C measurement, which requires a fasting sample.[19,20,21] Non-HDL-C targets are calculated by adding 30 to the standard ATP III target LDL-C level, as demonstrated in the Table .

Although LDL-C has always been regarded as the most atherogenic of the lipoproteins, it has now become clearer that TG-rich lipoproteins, especially VLDLs, are also associated with the development of vascular disease.[20] Particles known as VLDL remnants contain more cholesterol and less TG than VLDLs, and it has been suggested that these particles may be particularly atherogenic.[17,21] IDL-C also has many of the same properties as VLDL remnants and therefore is thought to confer similar atherogenic potential. In addition, raised VLDL concentrations are associated with procoagulant and prothrombotic factors in plasma, which also contribute to the development of arterial disease.[18] Using the Friedewald equation for LDL-C ignores these important atherogenic VLDL remnants as targets for therapy. Meanwhile, these VLDL remnants, as well as the atherogenic IDL-C, LDL-C, and TGs, are accounted for using the simple non-HDL-C calculation.

Data are currently limited, although a few studies have demonstrated a correlation between elevated non-HDL-C and increased atherogenic risk. One recent study compared the efficacy of LDL-C and non-HDL-C as predictive factors in deaths from cardiovascular events.[22] Data from the Lipid Research Clinic program cohort study[22] were used to compare the predictive value of non-HDL-C as a risk factor for CVD mortality with the current "gold standard," LDL-C.

In this study, a total of 2406 men and 2056 women without pre-existing CVD at study onset, with ages ranging from 40-64 years at entry, were included. After a follow-up period averaging 19 years, there were a total of 234 male deaths and 113 female deaths attributable to CVD. In men, an increase in non-HDL-C level was shown to be associated with an increase in CVD mortality.[22] When men with non-HDL-C values of <160 mg/dL were compared to those with values between 190 and 220 mg/dL, the latter group was found to have a 43% increased risk of death from CVD.[22] This risk was further increased as the non-HDL-C level increased, and in men with non-HDL-C measurements of >220 mg/dL, the relative risk of CVD mortality was found to be 2.14 (95% confidence interval, 1.50-3.04).

As expected, an association was also found between the level of LDL-C and CVD mortality in males, with LDL-C measurements ≥190 mg/dL associated with a 77% increase in CVD deaths, as compared to men with LDL-C levels <130 mg/dL. An interesting observation in this group was that men with LDL-C levels <100 mg/dL had a higher risk of CVD mortality than those with values in the 100-130 mg/dL range. This increased risk, however, was confined to patients with TG levels >200 mg/dL. An inverse association was also noted between HDL-C levels and the risk of CVD mortality. A comparison of non-HDL-C vs. LDL-C as predictive risk factors revealed that both HDL-C and non-HDL-C levels were better predictors of CVD mortality than LDL-C levels, in both men and women (relative risk for non-HDL-C levels in men, 1.19; in women, 1.15; relative risk for LDL-C levels in men, 1.11; in women, 1.08).[22]

Analysis of the female subjects in the study also demonstrated an increased risk of CVD mortality associated with an increase in non-HDL-C levels. A comparison between women with non-HDL-C levels <160 mg/dL and those with levels between 190 and 220 mg/dL demonstrated 61% increased CVD mortality, with a relative risk of 1.61 in the latter group. This risk was increased to 2.43 in women with non-HDL-C levels >220 mg/dL. Of significant interest was the finding that only non-HDL-C and HDL-C levels significantly predicted CVD death in females, since, surprisingly, there was found to be no significant correlation between LDL-C levels and CVD deaths. In women, as in men, LDL-C levels were found to be the least reliable predictor of CVD mortality, with the best being HDL-C and non-HDL-C.[22] Based on the results of this study, the acceptance of non-HDL-C level as a principal CVD risk factor may result in a more effective approach to cardiovascular risk reduction.

Supporting evidence for the above suggestion comes in the form of a recent study using data from the Atorvastatin Comparative Cholesterol Efficacy and Safety Study (ACCESS).[23] This study looked at the effect of statins on lipid and apolipoprotein levels, including non-HDL-C. Specifically, it compared the relationships of baseline apolipoprotein (apo) B levels with LDL-C and non-HDL-C levels, as well as those relationships after treatment with statin therapy. Apolipoproteins are the chief structural components of lipoproteins, and apo B, in particular, is integral to the production of chylomicrons and VLDLs. Since total plasma apo B concentration reflects the number of LDL-C- and TG-rich particles, it is an approximate guide to the amount of atherogenic particles in plasma.[24] Therefore, while all of the lipoproteins discussed above have atherogenic potential, measurement of apo B may be more useful than calculation of LDL-C in terms of identifying modifiable risk factors. This is supported by the Quebec Cardiovascular Study,[25] which established apo B as a more powerful predictor of CHD than LDL-C. Because LDL-C was calculated using the Friedewald equation, patients with a TG level >400 mg/dL were excluded to avoid inaccurate LDL-C levels, although this would not have affected the accuracy of non-HDL-C measurements. Patients were then randomized to one of five statins over a period of 54 weeks, with lipid levels assessed regularly at 6-week intervals. Apo B was also measured at weeks 0, 6, and 54.[23]

Measurements at weeks 6 and 54 showed that atorvastatin was the most effective drug in the lowering of LDL-C; however, HDL-C change did not differ significantly among the different statins. Non-HDL-C was strongly correlated with apo B across CHD risk categories (week 0, r=0.914; week 54, r=0.938). This correlation was found to be much stronger than that between LDL-C and apo B, although the latter association was still statistically significant. The correlation between LDL-C and apo B was especially weak in patients with CHD. Although the correlation between non-HDL-C and apo B remained consistently strong with variations in TG levels, the correlation between LDL-C and apo B became weaker as TG levels or CHD risk increased.[23]

Although atorvastatin was the most effective statin at lowering both LDL-C and non-HDL-C, fewer patients reached non-HDL-C targets than LDL-C targets for all the statins studied. Comparison of baseline non-HDL-C with LDL-C levels showed that in patients with CHD, both non-HDL-C and apo B were significantly elevated relative to LDL-C, suggesting that greater reduction in non-HDL-C than LDL-C would be required for optimal risk factor management.

The difficulty with measurement of apo B, although recently improved, lies in the lack of standardization across centers. Within the realms of the standard lipid profile, non-HDL appears to be the parameter correlating best with apo B.[22,23] It therefore appears prudent to establish non-HDL-C as a target for modification of CVD death risk.

The studies described above provide evidence supporting the rationale of using non-HDL-C as a target for lipid-lowering therapy. In both males and females, non-HDL-C predicted CVD death better than LDL-C, with increasing levels of non-HDL-C corresponding to an increased risk of CVD mortality. In addition, for female patients, only HDL-C and non-HDL-C significantly predicted CVD death, while the currently targeted lipoprotein, LDL-C, did not correlate with outcomes. In fact, LDL-C levels were the least reliable predictor of CVD deaths in women, when compared with non-HDL-C and HDL-C.