American Association for Clinical Chemistry
Better health through laboratory medicine
November 2011 Clinical Laboratory News: Lipids and Lipoproteins

CLN Banner Logo

November 2011: Volume 37, Number 11

Lipids and Lipoproteins
The Value of the CDC Lipid Standardization Program

By Mary M. Kimberly, PhD

Lipid Standardization

The Centers for Disease Control and Prevention (CDC) has worked for more than 50 years to standardize lipid and lipoprotein measurements to improve public health. Early efforts focused on developing a reference system, including reference measurement procedures and reference materials to use in the Lipid Standardization Program. Not only has this important infrastructure program permitted comparisons of clinical and epidemiological studies, but it has also allowed researchers to recognize long-term population trends in lipids and lipoproteins. Furthermore, the Lipid Standardization Program has provided a uniform basis for patient selection and data interpretation in clinical and epidemiological studies.

Most importantly, however, these efforts have enabled the National Heart, Lung, and Blood Institute’s (NHLBI) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults, known as the Adult Treatment Panel (ATP), to develop guidelines used by physicians to diagnose, treat, and monitor patients with cardiovascular disease (CVD). First issued in 1988 (ATP I), major revisions of ATP occurred in 1993 (ATP II) and 2001 (ATP III) (1,2).

In anticipation of the forthcoming ATP IV guidelines, this review will focus on the key aspects of the Lipid Standardization Program and will compare the important parameters of the ATP guidelines.

The ATP III Guidelines

Low-density lipoprotein cholesterol (LDL-C) is the primary target for therapeutic treatments of adults with high blood cholesterol in the ATP III guidelines issued in 2001. The expert panel revised the guidelines in 2004 based on additional clinical trial evidence, lowering the LDL-C target of therapy level for people at high risk for CVD (Table 1).

Table 1
Revised ATP III Classification of Serum LDL-C
Risk Category
Risk Definition
LDL-C Goal
Initiate TLC
Consider Rx
High With CHD or CHD
risk equivalents; 
FRS >20%
<100 mg/dL
Optional <70mg/dL
≥100 mg/dL
≥100 mg/dL
Moderately High ≥2 risk factors; 
FRS 10 to 20%
<130 mg/dL
≥130 mg/dL
≥130 mg/dL
Moderate ≥2 risk factors; 
FRS <10%
<130 mg/dL
≥130 mg/dL
≥160 mg/dL
Lower 0–1 risk factors
<160 mg/dL
≥160 mg/dL
≥190 mg/dL
Abbreviations: coronary heart disease, CHD; Framingham Risk Score, FRS; therapeutic lifestyle changes, TLC.

While ATP III places the greatest emphasis on LDL-C, high-density lipoprotein cholesterol (HDL-C) also plays an important role. The expert panel determined that <40 mg/dL HDL-C was undesirably low and that ≥60 mg/dL counted as a negative risk factor for CVD (Table 2). Elevated triglyceride (TG) levels are also cited as a risk factor for CVD in ATP III, because the panel found new evidence suggesting that TGs were a marker of elevated atherogenic remnant lipoproteins. Consequently, ATP III recommends lower medical decision points for TG compared to previous guidelines (Table 3).

Table 2
ATP III Classification of Serum HDL-C
HDL-C Concentration
Risk Category
<40 mg/dL
≥60 mg/dL

Low (negative risk factor)

Table 3
ATP III Classification of Serum Triglycerides
TG Concentration
Risk Definition
<150 mg/dL
150–199 mg/dL
Borderline High
200–499 mg/dL
≥500 mg/dL
Very High

ApoB: Missing in Action?

One marker that is notably absent from ATP III is apolipoprotein B (apo B).

A measure of all atherogenic particles in blood, apo B is the primary apolipoprotein of LDL-C and includes remnant lipoproteins that are rich in TG. Considerable evidence demonstrates that LDL-C and apo B are highly correlated in persons with normal TG levels, yet levels of apo B are disproportionately higher in people with hypertriglyceridemia.

In 1998, researchers introduced a value called non-HDL-C as a surrogate for apo B, which was derived from the equation: non-HDL-C = total cholesterol – HDL-C (3). While the ATP III guidelines did not include direct assessment of apo B, the panel recommended that labs calculate non-HDL-C from this equation. They included this secondary treatment target for individuals with high TG (≥200 mg/dL) for three reasons: 1) the apo B assays available at the time were not well-standardized; 2) there were insufficient prospective studies of individuals with high TGs that showed greater predictive power for apo B compared to non-HDL-C; 3) the expense of measuring apo B in addition to routine lipid analysis was too great. Researchers subsequently showed that labs could obtain qualitatively similar results for non-HDL-C and apo B without regard to TG concentration (4), lending further support to the expert panel’s decision not to include apo B in the guidelines. But in spite of the fact that non-HDL-C can be calculated from two standard lipid and lipoprotein laboratory markers, physicians and clinical laboratories have been slow to adopt it.

Reference Measurement Procedures

In order to create the ATP III treatment guidelines, NHLBI’s expert panel relied upon research studies from laboratory measurements that were standardized through the Lipid Standardization Program. The National Cholesterol Education Panel (NCEP) convened an additional panel of laboratory experts, the Working Group on Lipoprotein Measurement, which also advised NHLBI on method performance, and recommended selecting CDC’s reference measurement procedures to standardize the methods in clinical laboratories (5).

At that time, CDC used the Abell-Kendall method for cholesterol, which involves saponification of cholesterol ester, extraction of cholesterol, followed by developing the chromophore and measuring it (6). For HDL-C, the reference measurement procedure involved ultracentrifugation to separate the very low-density lipoproteins (VLDL) from HDL-C. The method also removed apo B-containing lipoproteins by precipitation with a reagent composed of heparin and Mn+2 ions. In the final step, the Abell-Kendall method quantitates the remaining cholesterol that is associated with HDL.

To measure LDL-C, CDC used a beta-quantification procedure that was exactly the same as the HDL-C method. The method included an additional measurement of cholesterol in the bottom fraction (BF-C) of the centrifugation step but before the precipitation. LDL-C was then calculated from the equation: LDL-C = BF-C – HDL-C.

Historically, CDC has used the chromotropic acid method as the reference measurement procedure for TG. This method was recently discontinued in favor of an isotope dilution-gas chromatography/mass spectrometric method, which will soon be published by CDC.

Homogeneous Direct Method Issues

In contrast, most of the lipid and lipoprotein epidemiologic studies and clinical trials have employed chemical precipitation methods for measuring HDL-C. They either use the Friedewald equation to calculate LDL-C or measure it indirectly using a beta-quantification approach. In such studies, LDL is defined as the lipoprotein fraction with a density of 1.019–1.063 g/mL.

In practice, however, the beta-quantification reference method procedure for LDL-C uses ultracentrifugation to separate LDL at 1.006 kg/L, the native density of serum. This fraction includes: intermediate-density lipoprotein (IDL); some, but not all, of lipoprotein(a); and possibly larger apolipoprotein E (apo E)-containing HDL particles. It is worth noting that by recommending the CDC beta-quantification method be the accuracy point, the Working Group on Lipoprotein Measurement effectively defined LDL as a broad band of densities that includes these other apo B-containing lipoproteins (5,6).

Implementing Direct Homogeneous Assays

Since the ATP III guidelines were published, commercial homogeneous methods that measure HDL-C and LDL-C directly have become more widely available. The CDC’s Cholesterol Reference Method Laboratory Network (CRMLN), which uses reference measurement procedures traceable to the CDC, has standardized these methods.

Several published studies assessed the reliability and accuracy of these assays, and in some cases labs have observed discrepant results with the homogeneous direct HDL-C (dHDL-C) methods. Although comparison studies have been published that evaluated dLDL-C and/or dHDL-C methods, these studies usually were performed with one or only a few of the commercially methods available, often did not include comparison to the reference measurement procedure, and usually did not include serum from dyslipidemic subjects. Therefore, labs that want to replace the Friedewald calculation with a homogeneous dLDL-C assay should first carefully evaluate new assays (7).

A good starting point for such evaluation studies is a paper published by Miller and colleagues in which they evaluated the results of all commercially available dLDL-C and ≤dHDL-C assays (8). The team collected samples from 37 subjects without disease and from 138 subjects with established CVD or other conditions that affect lipoprotein measurement and conducted the analysis within 48 hours of sample collection. One collaborating laboratory measured dHDL-C and dLDL-C, while CDC conducted the reference measurement procedure analysis.

The researchers used NCEP performance goals to evaluate the performance of each homogeneous direct method tested. They used a total error (TE) goal for HDL-C of ≤13% versus the reference method procedure and ≤12% for LDL-C. Accuracy goals were bias ≤5% versus the reference measurement procedure for HDL-C and ≤4% versus for LDL-C.

For the samples from non-diseased patients, six of eight HDL-C methods met the TE goal, while five of eight LDL-C methods met the TE goal. None of the HDL-C or LDL-C methods met the TE goals with samples from diseased patients. Error components analysis revealed that sample-specific effects dominated the random error components in six of eight HDL-C and LDL-C methods for non-diseased samples and in all HDL-C and LDL-C methods for diseased samples. Furthermore, these sample-specific effects were larger for the diseased group than for the non-diseased group, leading the researchers to conclude that the effects may be caused by any component in the serum that influences the specificity of the reagents for the HDL or LDL lipoprotein particles.

The error analysis also showed that the mean bias met the accuracy goal for all of the HDL-C methods when non-diseased samples were analyzed. In contrast, the NCEP accuracy goal was exceeded for three of eight HDL-C methods in the diseased group. The mean bias also exceeded the accuracy goal for three of eight LDL-C methods for both non-diseased and diseased groups, and the bias was correlated with TG concentration, which reflects abnormalities in lipoprotein composition that may influence the specificity of the dLDL-C and dHDL-C methods. In fact, high TG concentration may be a surrogate for the presence of abnormal lipoproteins that can influence these methods.

In many cases, the differences between the dHDL-C and dLDL-C method results and the reference measurement procedure were of sufficient magnitude that clinical management of patients could be affected. Interestingly, several patients in the study had a genetic disorder of lipid metabolism. Their dHDL-C and dLDL-C results differed from the reference method procedure so markedly that they could have been misdiagnosed. Although all of these methods were standardized through CDC’s CRMLN, the samples typically collected for certification are from normolipidemic people and have TG <200 mg/dL. Therefore, labs should retest patient samples that show discrepancies between the dLDL-C results and clinical observations using an ultracentrifugation method.

Effect of Method on Risk Assessment

Given the variations observed in the method comparison study, how accurately can test results assess CVD risk? van Deventer and colleagues recently published additional findings from the same cohort sample described above in which they assessed patients’ CVD risk classification according to the ATP III medical decision points (9). In the study, the researchers compared results from the dLDL-C method, the Friedewald-calculated LDL-C (cLDL-C), and the reference measurement procedure (rLDL-C). For each manufacturer’s dHDL-C assay, they used common measurements for TC and TG to calculate the cLDL-C and compared non-HDL-C calculated using results from the dHDL-C methods to non-HDL-C calculated from rHDL-C.

For the CVD risk assessment, the researchers divided the study cohort into two groups based on TG concentration: normotriglyceridemic subjects with TG <200 mg/dL and hypertriglyceridemic subjects with 200–400 mg/dL TG. They then looked at patients’ CVD risk classifications based on dLDL-C and cLDL-C results versus those obtained from rLDL-C.

Among subjects with TG <200 mg/dL, dLDL-C methods did not improve the accuracy of risk classification compared to cLDL-C, and dLDL-C tended to classify more subjects into a lower, rather than higher, risk category. In hypertriglyceridemic samples, the dLDL-C methods performed better than cLDL-C in part due to the fact that the dHDL-C methods tend to perform poorly on samples that have both low HDL-C and high TG, which leads to errors in the Friedewald calculation. In spite of performing better than cLDL-C for samples with TG between 200 and 400 mg/dL, the dLDL-C methods performed worse with hypertriglyceridemic samples than they did with normotriglyceridemic samples. Finally, non-HDL-C calculated using dHDL-C methods compared well with non-HDL-C calculated using rHDL-C, and non-HDL-C outperformed both dLDL-C and cLDL-C in correctly classifying both normotriglyceridemic and hypertriglyceridemic subjects.

As mentioned earlier, the LDL-C reference method procedure has a broad functional definition of the LDL fraction. Therefore, if the dLDL-C methods are more specific than the reference measurement procedure for LDL particles and not to other apo B-containing lipoproteins such as IDL or Lp(a), then the method will tend to have negative bias to rLDL-C. This means that the method will tend to classify subjects into lower risk categories.

One thing to keep in mind about CVD risk classification is that extensive research suggests that other apo B-containing lipoproteins included in the rLDL-C measurement are also proatherogenic. Since these other lipoproteins contribute to the cLDL-C and non-HDL-C calculations, using them to classify individuals should lead to better performance in risk classification compared to dLDL-C.

Triglyceride Measurements

Another risk factor for CVD in the ATP III guidelines is elevated TGs; however, routine TG measurements have their own analytical problems. The long-established enzymatic methods used by clinical labs to assess TG are actually not specific for TGs. The lipases used in these methods sequentially release all of the fatty acids of TGs, diglycerides, and monoglycerides to produce one glycerol per molecule. Glycerol reacts further with other enzymes in the assay to produce molecules that can be measured spectrophotometrically. Most individuals have low levels of diglycerides and monoglycerides, but this is dependent upon many factors, including the age of the sample. In addition, most native patient samples are low in free glycerol. Clinical chemists have debated the need for blanking TG analyses for free glycerol; however, in a large study of hospitalized patients, no significant clinical consequences resulted for the average individual when a glycerol blank was not used (10). The exception was patients with abnormally high glycerol levels.

Serum glycerol blanking becomes extremely important, however, when labs attempt to standardize TG measurements. Glycerol content in commercially prepared materials, such as proficiency testing survey samples, controls, and calibrators, can be quite variable, depending on the matrix of the source material (frozen versus fresh serum) and the type of fortification used to increase lipid concentrations. This makes standardizing TG measurements across clinical laboratories difficult and assessing the accuracy of results very challenging. However, when these materials are prepared from fresh off-the-clot human serum using CLSI’s guideline C37-A, the levels of free glycerol present can be reduced to the range found in most patient specimens. Laboratories would be well advised to evaluate each lot of material for free glycerol before using it for calibration or accuracy evaluation because error in the TG measurement contributes to error in cLDL-C derived from the Friedewald equation.

Future Guidelines

NHLBI’s ATP IV panel has been working on an update to the previous guidelines and is expected to release them soon for comment. Among the key considerations are the results from the recent dHDL-C and dLDL-C method evaluations, as well as the growing body of epidemiologic studies and clinical trials on biomarkers of CVD risk. No doubt clinical labs will be at the forefront of providing the most appropriate tests in the new guidelines to diagnose, treat, and monitor patients.


  1. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285:2486–97.
  2. Grundy SM, Cleeman JI, Bairey CN, Merz H, et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III Guidelines. Circulation 2004;110:227–39.
  3. Frost PH, Havel RJ. Rationale for use of non–high-density lipoprotein cholesterol rather than low-density lipoprotein cholesterol as a tool for lipoprotein cholesterol screening and assessment of risk and therapy. Am J Cardiol 1998;81:26B–31B.
  4. di Angelantonio E, Sarwar N, Perry P, Kaptoge S, et al. Major lipids, apolipoproteins, and risk of vascular disease. JAMA 2009;301:1993–2000.
  5. Bachorick PS, Ross JW. National Cholesterol Education Program recommendations for measurement of low-density lipoprotein cholesterol: executive summary. Clin Chem 1995;41:1414–20.
  6. Warnick GR, Kimberly MM, Waymack PP, Leary ET, et al. Standardization of Measurements for Cholesterol, Triglycerides, and Major Lipoproteins. Lab Med 2008;39:481–490.
  7. Nauck M, Warnick GR, Rifai N. Methods for Measurement of LDL-Cholesterol: A Critical Assessment of Direct Measurement by Homogeneous Assays versus Calculation. Clin Chem 2002;48:236–54.
  8. Miller WG, Myers GL, Sakurabayashi I, Bachmann LM, et al. Seven direct methods for measuring HDL and LDL cholesterol compared with ultracentrifugation reference measurement procedures. Clin Chem 2010;56:977–86.
  9. van Deventer HE, Miller WG, Myers GL, Sakurabayashi I, et al. Non-HDL cholesterol shows improved accuracy for cardiovascular risk score classification compared to direct or calculated LDL cholesterol in a dyslipidemic population. Clin Chem 2011;57:490–501.
  10. Jessen R, Cass C, Eckfeldt J. Do enzymatic analyses of serum triglycerides really need blanking for free glycerol? Clin Chem 1990;36:1372–1375.

Mary Kimberly
Mary M. Kimberly, PhD, is former chief of the Lipid Reference Laboratory, Clinical Chemistry Branch, National Center for Environmental Health, Centers for Disease Control & Prevention, Atlanta, Ga. She is now retired. Email: